Gosta Esping-Andersen

Why Deregulate Labour Markets?
This page intentionally left blank

Why Deregulate Labour Markets?
Edited by
Gsta Esping-Andersen
and
Marino Regini
Great Clarendon Street, Oxford OX2 6DP
Oxford University Press is a department of the University of Oxford
It furthers the University's objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Bangkok Buenos Aires Cape Town Chennai
Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata
Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi
So Paulo Shanghai Taipei Tokyo Toronto
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States by
Oxford University Press Inc., New York
the several contributors 2000
The moral rights of the authors have been asserted
Database right Oxford University Press (maker)
First published 2000
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographicsrights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose this same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Why deregulate labour markets? / edited by Gsta Esping-Andersen and Marino Regini.
Includes bibliographical references.
1. Labor marketEuropeCase studies. 2. UnemploymentEuropeCase studies. 3. Labor laws and
legislationEuropeCase studies. I. Esping-Andersen, Gsta, 1947II. Regini, Marino, 1943
HD5764.A6 W48 2000 331.12042094dc21 99-057196
ISBN 0-19-829681-9
ISBN 0-19-924052-3 (Pbk.)
To Paula and Maria Luisa
Preface
The public debate on the consequences of labour market regulation in Europe is, at present, very intense and,
occasionally, ideologically polarized. In turn, the scientic literature is riddled with ambiguities and contradictory
claims. Our aim in this book is to sort out the issues, both conceptually and empirically. What can we conclude from
the myriad existing studies and, equally importantly, from the concrete empirical experience of distinct countries? In
addressing these questions, this volume provides a critical review of the different national approaches to exibility and
deregulation, and an assessment of the evidence on their impact on aggregate unemployment trends and
unemployment structure. The rst chapter frames the debate within the broader context of change in the regulatory
regimes of advanced economies; in so doing, it helps clarify the main concepts and analytical tools used in the scientic
and policy debate on labour market trends, policies and performance. The three comparative chapters that follow
review current trends in labour market policies as well as the different approaches taken to exibilization. They assess
the available cross-country evidence regarding their effects on labour market performance. Finally, these chapters
highlight convergence and divergence between the countries studied and depict some alternative scenarios for
European labour markets.
The comparative chapters are followed by detailed studies of eight European countries, selected as representative of
distinct regulatory patterns and of different types of performance. These country studies review the national debate,
current trends, and evidence on outcomes, using a common framework which allows for close comparison among
different patterns of change.
This volume stems from a report on the effects of labour market deregulation on unemployment, commissioned to
Ires Lombardia by the European Commission, Directorate-General XII. The report was based on research co-
ordinated by Marino Regini. We gratefully acknowledge the nancial and organizational support given by the EC,
DGXII, and by Ires Lombardia. Also a special thanks to John Martin at the OECD for his helpful criticisms on parts
of this book. The purpose of the report was not to provide blunt conclusions or policy prescriptions, but rather to
identify more clearly the trade-offsand associated
viii Preface
policy implicationsthat emerge in alternative systems of labour market regulation. In writing this book, we have been
guided by the same overall purpose.
and Marino Regini
Barcelona and Milano,
June 1999
Contents
List of Figures xi
List of Tables xiii
List of Contributors xv
Introduction Gsta Esping-Andersen and Marino Regini 1
PART I. LABOUR MARKET REFORM IN EUROPE
1 The Dilemmas of Labour Market Regulation Marino Regini 11
2 The Dynamics of Labour Market Reform in European Countries Manuela Samek Lodovici 30
3 Who is Harmed by Labour Market Regulations? Quantitative Evidence Gsta Esping-Andersen 66
4 Regulation and Context: Reconsidering the Correlates of Unemployment Gsta Esping-Andersen 99
PART II. NATIONAL VARIATIONS
5 River Crossing or Cold Bath? Deregulation and Employment in Britain in the 1980s and 1990s
Simon Deakin and Hannah Reed 115
6 Going Different Ways: Labour Market Policy in Denmark and Sweden Anders Bjrklund 148
7 The Dutch Miracle? Cees Gorter 181
x Contents
8 Germany: A Regulated Flexibility Susanne Fuchs and Ronald Schettkat 211

9 France: The Deregulation that Never Existed Miguel A. Malo, Luis Toharia, and Jerome Gauti 245
10 Italy: The Long Times of Consensual Re-regulation Manuel Samek Lodovici 271
11 The Spanish Experiment: Pros and Cons of the Flexibility at the Margin Luis Toharia and
Miguel A. Malo 307
Conclusions Gsta Esping-Andersen and Marino Regini 336
Index 343
List of Figures
2.1 Stringency of employment protection and wage-setting regulation and expenditure on labour
market policies 34
6A Unemployment rates in Denmark and Sweden 19871997 175
6B The evolution of unemployment in Denmark since 1994. Monthly data. 1624 years,
and 2559 years 176
6C Stylized patterns of conditional exit-rates out of unemployment in Sweden during the
recession in the 1990s according to two recent studies. 177
7.1 Employment growth in The Netherlands, 19871997 182
7.2 Unemployment rate in The Netherlands, 19801997 195
7.3 Long-term unemployment, 19801990 195
7.4 Development of the absolute number of long-term unemployed, 19901995 196
7.5 Employment ows in The Netherlands, 19711991 197
7.6 Unemployment, vacancies, and matches, The Netherlands, 19801993 200
7.7 Labour market efciency over the business cycle, The Netherlands, 19801993 201
8.1 The West-German Beveridge curve 217
8.2 Flow-duration curve of unemployment and vacancies in West Germany 218
8.3 Employment elasticity in German and US rms 234
9.1 Labour force, employment, and unemployment in France 246
9.2 Unemployment rates: gender, youth, and long-term unemployed, France, 19851997 262
9.3 Unemployment rate by level of education, France, 19911997 264
10.1 Inter-sectoral wage dispersion, Italian industry, 19701995 284
10.2 Employment evolution in Italy by geographic area, 19811997 290
10.3 Employment of unemployed with previous jobs and equivalent workers on CIG benets
in industrial rms, 19701998 (quarterly moving averages) 297
xii List of Figures
10.4 Part-time and xed-term contracts by geographic area, July 1993April 1997
(quarterly moving averages) 298
11.1 Labour force, employment, and unemployment, Spain, 19701996 308
11.2 Contractual and empirical employment stability, Spain, 19871997 324
11.3 The employment situation of temporary workers 327
List of Tables
2A Employment protection legislation: evolution in the 1990s 55

2B Incidence of atypical work and non-dependent employment, non-agricultural sector, 19851996 56
2C Incidence of accompanying labour policies, 19961997 57
2D Working time: variability of hours worked among full-time employees 58
2E Wage-setting: main trends since the early 1980s 59
2F Tax and contributive wedge, 1985 and 1994 60
2G Unemployment benets 61
2H Summary of action taken to change regulation in the 1990s 62
3.1 Divergent patterns of employment and unemployment in Europe and the USA, 1980s1990s 68
3.2 Unemployment and social benets 79
3.3 Unemployment and wage structure 81
3.4 Unemployment, wages, and labour market regulation 88
3.5 The impact of employment protection on unemployment 89
3A Trends in employment and unemployment across business cycles 94
3B Indicators of employment stability 94
3C Unemployment ows 95
4.1 Causes of women's employment 104
4.2 Explaining cross-sectional employment growth, 19791994 107
4.3 Explaining unemployment levels, 1996 108
4.4 Explaining unemployment outows, 1994 109
4.5 Explaining unemployment structure 111
5.1 Legislative developments affecting labour market exibility, UK 117
5.2 Workless households, 19771998 119
5.3 Labour market status data 129
5.4 Number of people making at least one claim for unemployment-related benets 131
xiv List of Tables
5.5 Redundancies and redundancy rate 132

5.6 EPI joblessness indicators, 19841997 133
5.7 Probability of movement between employment states (%) 134
5.8 The changing composition of all employment, 19791997 136
6A Predicted unemployment and underlying factors 172
6B Hourly wage inequality, Denmark and Sweden, 19801997 173
6C Union membership rates, 19801996 174
7A Institutions and deregulation, The Netherlands 206
7B Selected international studies on employment ows 207
8.1 Unemployment trends in West Germany 213
8.2 Unemployment rates, labour participation, and employment-population ratios by skill level,
Germany and the USA 214
8.3 Employment-population ratios for West Germany and the USA, 19701996 216
8.4 Labour market mobility in West Germany 220
8.5 East German labour market trends 221
9A Institutional changes in France 267
9B Distribution of employment by sector, France, 19851995 267
9C Evolution of employment by industry, France, 19921996 268
9D Atypical forms of work, France, 19851997 (% of total employment) 268
10.1 Italy: main labour market indicators, 1997 272
10.2 Institutional labour market exibility developments in Italy 277
10.3 Incidence of regular and non-regular labour units on total standard full-time labour units 292
10.4 Unemployment ows and long-term unemployment 19791991 294
10.5 Quarterly and annual exit routes from dependant employment, 19951996 (%) 299
10.6 Labour market conditions for young rst job seekers after one year from rst survey 19951996 (%) 301
11A Institutional changes in Spain's labour market: impact on exibility 331
11B Characteristics of permanent and temporary workers, private sector, Spain, 1992 and 1997 332
11C Cohorts aged 2530: the situation ve years later, Spain, 1992 333
List of Contributors
Marino Regini, Professor of Sociology, University of Milano, and Director of IRES Lombardia
Gsta Esping-Andersen, Professor University Pompeu Fabra, Barcelona and University of Trento, Italia
Manuela Samek Lodovici, Economist and Researcher at the Institute for Social Research (IRS), Milano
Anders Bjorklund, Professor of Economics, Stockholm University and Institute for Social Research,
Stockholm
Simon Deakin, Centre for Business Research, Department of Applied Economics, University of Cambridge
Hanna Reed, Centre for Business Research, Department of Applied Economics, University of Cambridge
Ronald Shettkat, Professor of Economics, Department of Social and Institutional Economics, University of
Utrecht
Cees Gorter, Department of Regional Economics, Free University of Amsterdam Luis Toharia, Professor of
Economics, University of Alcala de Henares, Spain
Miguel-Angel Malo, Economist, University of Salamanca, Spain
Introduction
Gsta Esping-Andersen and Marino Regini

The debate on labour market deregulation occasionally hardens into two opposing positions, neither of which enjoys
broad academic credibility nor political acceptance. It is a debate that, too often, pits a exible and dynamic United
States against a sclerotic, fossilized Europe. It is easily ignored that Europe is hardly homogeneous on any count of
labour market regulation, or on any measure of labour market performance. Ideological fervour tends to reduce
phenomena to excessive simplicity. We therefore risk glossing over the truly complicated and multi-faceted meanings
of exibility.
Many proponents of labour market deregulation take the tout court view, arguing that the closer we approach pure,
undisturbed market clearing, the better we shall be able to optimize welfare, efciency, and employment goals. Within
labour markets, this entails deregulating hiring and ring rules, wage setting, working time and, as an unavoidable
accompaniment, we must accept signicant reductions in social benetsthe reservation wage. In other words, the
more exibility of all types, the better.
The critics of deregulation highlight both its associated risks, and the potential advantages that go with regulation. A
dominant position, echoed in many European Community (EC) documents and manifestos, is that labour market
deregulation can jeopardize the social consensus and unleash conict because of the inequalities that it, inevitably, will
catapult. They also maintain that rigidities is the wrong diagnosis of Europe's poor employment performance. The
unemployment problem, say the critics, has less to do with worker protection than with the restrictive macro-economic
policies adopted by European monetary authorities and governments. Hence, this position calls for a package of
macro-economic stimulation policies.
It is perhaps unavoidable that such a debate aligns itself closely to the reigning political climate, thus additionally
politicizing and simplifying the issues. To foster a more serious and policy-relevant discussion, we need to return to
basics and examine more carefully the relative merits
2 Gsta Esping-Andersen and Marino Regini
of deregulation. We need to be able to answer the following simple question: what exactly does what kind of labour
market deregulation achieve? If we lay aside ideology, we soon recognize that no serious scholar, or even politician,
genuinely believes that exible labour markets resolve everything. This appears, in fact, to be mirrored in actual
practice. As we shall learn from the chapters in this book, the real world of regulatory and deregulatory reform is one
overwhelmingly dominated by partial and piecemeal exibilization. As Saint-Paul (1996) argues, this may simply reect
stubborn political realities and the power of veto blocks; whether or not far-reaching deregulation is actually called for,
marginal adjustments are the best one can hope for. The absence of dramatic departures from the status quo may also
reect a genuine and broadly shared commitment to maintaining egalitarian traditions. In any case, there is little doubt
that European governments have refrained from radical deregulation for the basic reason that most remain sceptical
about its virtues, and fearful of its consequences.
What, then, is the strongest case in favour of exibilizing labour markets or, even better, what kinds of deregulation
promote what kinds of benecial results? In contrast, what is the associated price of such policies? Under what
conditions can deregulatory reforms be Pareto-improving? Unfortunately, the lion's share of existing scholarly research
does not furnish very clear answers to this way of posing the question. As we shall examine in Chapter 2, it is only very
recently that some, indeed only a few, economists have begun to realize that labour market rigidities simply cannot
explain the European unemployment problem per se. There is very little solid evidence that rigidities of hiring and
ring, or of wage structures, have anything to do with Europe's chronically high unemployment levels, nor with
Europe's apparent incapacity to fuel job growth. There is, however, substantially more credible evidence that rigidities
of this sort inuence who are the unemployed.
We have to be very clear about our goals. Are we primarily concerned to bring down aggregate unemployment levels?
In this case, we should probably shift our attention from deregulation to other strategies. Of possibly greater relevance
is the debate on macro-economic shocks and suppressed aggregate demand that some economists, such as Blanchard
(1998), and politicians, such as Europe's new Center-Left governments, are attempting to re-stimulate.
We may, of course, also be concerned with unemployment structure, and probably rightly so. The insider-outsider
theory of unemployment (Lindbeck and Snower 1988) has, in recent years, moved from academia to the political
terrain, largely because of the growing preoccupation in Europe with social marginalization. If, as is certainly the case
Introduction 3
in many European countries, unemployment is concentrated in certain visible groupslike youth or the low-
skilledand if it also produces long-term exclusion, we face scenarios of social division that are reminiscent of the
troubled pre-war era. A battle to democratize the unemployment experience, even if it will hardly lower unemployment
levels, is certainly not a trivial political goal. Here, as we shall suggest, the proponents of deregulation do seem to have
a stronger case. There are some regulatory practices that do seem to systematically bias the chances of being
unemployed towards youth, women or low-skilled workers. Here some exibilization might be called for. There are,
however, other regulatory practices that apparently have little bearing on this bias.
Identifying more clearly these somes and these others is a pressing task in our ideologically charged political climate.
Partial measures, be they called deregulatory or otherwise, that effectively stimulate employment among youth are
crucial for our economic future. We now have at our disposal quite an impressive accumulation of individual national
experiences with deregulatory policies ranging from liberalizing temporary work contracts to lowering social benets.
The bulk of this book is, in fact, devoted to a careful examination of such reforms and of their employment
consequences. What, on occasion, emerges is how important it can be to have a clear idea of the somes and others.
In some cases, a reform intended as a exibilizing measure actually provokes the oppositea rigidication. In other
cases, a minor adjustment of existing policies (like reducing the duration of unemployment benets) may sufce to
achieve intended goals.
As stressed by the sceptics of wholesale deregulation, worker protection is not just a matter of welfare, but may also be
conducive to efciency. There are hard trade-offs involved. Firmsand entire nationswhich choose to compete on
quality rather than mere price need a qualied, dependable and cooperative workforce. Cheap labour cannot guarantee
such qualities. This is exactly why, even in highly exible labour markets like the USA, core economy rms operate
with efciency wage systems and implicit, possibly informal, rigidities. As a growing literature attests, markets alone are
not very capable of assuring adequate workforce training. Hence, overly deregulated labour markets may engender a
low-skill equilibrium, the long-run result of which is productivity lag and loss of competitiveness (Soskice 1990;
Snower 1997).
The insider-outsider theory rightly focuses on the exclusionary consequences that come from the privileged treatment
of the insidersthe core workers. But it tends to ignore that a stable core workforce is a necessary economic asset for
economies in high value-added markets. In order to compete, these kinds of rms must rely on the cooperation,
skills, and functional exibility of their employees. In the larger picture it is increasingly the performance of these kinds
of rms that is decisive for a country's long-term economic performanceand, hence, for the sustainability of its
welfare goals.
The typical high-level conference on unemployment in Europe usually arrives at the following sort of conclusion:
what we must nd is the right balance between social guarantees and exibility. Not surprisingly, just about everyone
can agree. No one is, of course, even distantly able to pinpoint where such a balance may be found. Worse, this is
sloppy thinking; the wrong way to approach the question. The issue may not be how far to push the deregulation lever
without provoking harm to the social fabric, but rather how to recombine existing regulatory elements so that they
help us pursue what we desire. And here we arrive at the crux of the matter. Just like it makes absolutely no sense to
discuss Pareto optimalities without rst having decided on a criterion of justice, it makes no sense to discuss regulatory
reform without a precise idea of what we want to accomplish and what we want to avoid. If deregulating labour
markets by easing layoffs and hiring procedures implies, at the same time, an erosion of workers' and unions'
willingness to cooperate in technological change, have we gained? What price do we put on either?
The kind of thinking that dominates high-level ministerial conferences and academic writing in equal measure suffers
from an additional shortcoming: it remains xated on our manufacturing economy, continuing to assume that this is
where the jobs and unemployment problem is buried. In reality, any hope of returning to full employment must be
premised on services. These are much less understood; they do not necessarily follow the same rules as industry, and it
is far from clear whether all service sectors follow a similar logic as far as the exibility-jobs nexus is concerned. It is
often believed that since they are more sheltered from international competition, there is less need for deregulation of
service labour markets. To be sure, some services are subject to intense global competitionairlines and banks at one
end, domestics and nannies at the other end. None the less, what most forget is that a huge amount of existingand a
large share of potentialservice jobs compete, not with American consulting rms or Philippine nannies, but instead
face-to-face with household self-servicing, which is produced gratis, a truly formidable competitor.
Promoting employment growth through low-end consumer and social services may require wage deregulation and
exibilities or, alternatively, subsidies. But, again, is this desirable? If, concomitantly, our aim is to stimulate high-value-
added manufacturing, are we willing to accept a dual economy with deregulated, low-wage services and stable,
Introduction 5
high-paying manufacturing jobs; a sectorally divided, two-class labour market?

Our high-level conference is usually quite aware of some of the dilemmas and trade-offs involved. Indeed, the call for a
balance is usually spirited by the perennial search for a third way, a win-win strategy. Today this almost invariably sees
education and training as the solution. Whether they are right or not is one question, but what this signals is the acute
awareness of how inter-connected are the labour market and the institutions that surround it.
One such formidable institution is collective bargaining and the trade unions; for better or worse, they cannot be
assumed away. Going against trade unions' established prerogatives and principles for the sake of exibilization may be
feasible if the trade unions, themselves, can be emasculated. If not, the only realistic analysis is one that takes their
power into account; one that departs from the premise that some bargained consensus is a precondition for reform.
And even if emasculation were possible, is it desirable to crush strong institutions for the sake of exibility? Under
what conditions?
There is an alternative to the emasculation strategy: to nurture an industrial relations system in which the social
partners will be capable of cooperation, and motivated to cooperate. In some of our country analyses, it appears that
the social partners can be persuaded into the kind of cooperation that results in win-win outcomes. Indeed, Europe's
two best employment performers in the 1990s, the Netherlands and Denmark, pursued strategies that counted on
active trade union participation.
In most European countries the trade unions and collective bargaining institutions are much more than powerful
constraints that somehow must be co-opted or tranquilized in order to get things done. Their role in production as
well as in labour market regulation is typically ambivalent, at once acting as a constraint and as a resource for exibility.
Far from simply being a conventional agent of redistribution, of more for their members, European trade unions
nowadays perform crucial productive functions. They organize cooperation in the workplaces, help develop human
resources, and strike social pacts in the economic policy arena. These functions may be far more instrumental to
overall growth and employment in the long run than would labour market deregulation premised on weakened
unionism.
The family is a second formidable social institution. When wages are threatened, it is often not just the individual
worker's welfare that is at stake but an entire family's. In some instances the job and earnings of a partner, parents or
children can cushion wage erosion and worker insecurity. In Southern Europe, however, the prevailing scenario is that
families typically count on the pay and social entitlements of one breadwinner. As such, you hardly even need powerful
unions to oppose exibilization because the median voterand thus the winning political coalitionis bound to do
the same. No serious economic analysis can afford to assume away the family because this is the real decision unit of
labour supply and product consumption. How can families be persuaded to support a strategy of deregulation? What
kinds of exibilization, under which conditions, are compatible with the reigning family structure of the country in
question?
A tentative answer to the dilemmas that these institutions provoke, is that both trade unions and families can play a
positive role in labour market exibilization under two conditions. The rst is that exibility should not be the
outcome of generalized deregulation but rather of controlled exception to the prevailing regime. The country chapters
show several examples of exibility at the margin: selective measures aimed at increasing the employability of the
weaker members of families or trade union associationsmeasures that basically leave the prerogatives of the core
labour force untouched. This, in essence, is the kind of implicit strategy that several European countries are following.
It may not be the most effective medicine against mass unemployment, as Bertola and Ichino (1995) argue. Yet, as all
available evidence suggests, neither is deregulation in general because, as noted, strict worker protection affects
primarily the structure and not the levels of unemployment. The second condition has more to do with equity. Trade
unions and families may only be inclined to support labour market exibilization if the actual or perceived losers of
reform can be persuaded that their losses are temporary; that they stand a good chance of beneting in the long run. It
is not difcult to tolerate substandard employment among young workers if, that is, they receive adequate training and
stand a good chance of subsequently moving into the core labour force. Companies in less developed areas may be
allowed to adopt a more exible labour market regime if this temporary advantage is used to build a stronger
competitive positionone compatible with the general regime.
Trade unions and families are not necessarily the enemies of deregulation. What matters are assurances that the
associated ills and the possible inequities will not linger or, worse, become a permanent feature of society. If there is an
emerging European middle way between exibility and social guarantees, this seems to be it: partial deregulation,
typically targeted at the outsiders while leaving many prerogatives of the core workforce more or less intact. This is
certainly the prevailing picture that emerges from the country chapters in this book. It is, to be sure, a middle way that
hardly satises the radical deregulator. But
Introduction 7
then it is exceedingly difcult to imagine where political support for the radical approach would come from. In the
unlikely event that such support were forthcoming, it would probably be futile to implement it; the evidence shows
that radical deregulation would not solve Europe's unemployment problem.
References
Bertola, G., and Ichino, A. (1995), Crossing the river: A comparative perspective on Italian unemployment dynamics.
Economic Policy, 21: 359420.
Blanchard, O. (1998), Thinking about unemployment, Paolo Baf Lecture on Money and Finance, 16 October, 1998.
Rome: University of Rome.
Lindbeck, A., and Snower, D. (1988), The Insider-Outsider Theory of Employment and Unemployment. Cambridge, Mass: M.I.
T. Press.
Saint-Paul, G. (1996), Exploring the political economy of labour market institutions. Economic Policy, 23: 263316.
Snower, D. (1997), The low-skill, bad-job trap, in A. Booth and D. Snower (eds.) Acquiring Skills. Market Failures, Their
Symptoms and Policy Responses. Cambridge: Cambridge University Press, 10926.
Soskice, D. (1990), Wage determination: the changing role of institutions in advanced industrialized countries, Oxford
Review of Economic Policy, 6/4.
Part I Labour Market Reform in Europe
1 The Dilemmas of Labour Market Regulation
Marino Regini
Labour Markets Between Regulation and Flexibility

Labour markets have traditionally been among the most heavily regulated areas of the Western economies, whether by
law and the administrative activity of the state; by collective bargaining; by corporatist bodies which involve the social
partners in labour policy formation and implementation; or by networks of strong or weak ties which constrain the
behaviour of labour market actors. Indeed, it is widely recognized that nowhere has any such thing as a market for
labour ever existedand not just for the obvious reason that this is a commodity with very special, indeed unique,
features (Solow 1990). Even more importantly, labour market actors are very seldom guided by price signals alone,
since they are embedded in a network of social relations which, even in the absence of strong institutional regulation,
largely shapes their behaviour and determines the outcome (Granovetter 1985).
Institutional regulation has additionally helped restrict the scope for pure market clearing. In all mature economies,
institutions have come to govern several areas and aspects of the labour market. Some of these areas and aspects are
especially relevant to the current debate, not just because they have been a major focus of institutional regulation, but
more importantly because they are the main targets of attacks against the rigidity of European labour markets, and are
arguably therefore in need of greater exibility.
The rst of these areasthe one that most markedly differentiates continental European labour markets from the
USAis employment protection, namely the complex set of legislative measures and collective agreements which
constrain employers' ability to re and hire at will. A second area is welfare state benets for those involuntarily leaving
the labour market, and especially such income security measures as the various forms of unemployment insurance. A
third aspect of the
12 Marino Regini
labour market traditionally subjected to institutional regulation is working time. Finally, wage levels and wage structures
have been a primary object of collective bargaining, and, in several countries, of statutory regulation or tripartite
agreements as well.
Several factors account for the emergence of labour market regulation. Economists themselves are ready to
acknowledge that some forms of the institutional regulation of labour markets may have positive economic functions.
The following are frequently mentioned: to off-set market failures (Williamson 1975, Lindblom 1977), because labour
markets are especially imperfect and failures may stem from uncertainty and asymmetrical information; to increase
efciency, because labour standards compel rms to raise efciency or go under (Sengenberger and Campbell 1994);
to lower costs, because the welfare state greatly assists in socializing the costs of labour reproduction (O'Connor 1973);
and to foster mass consumption, because collective bargaining or minimum wage legislation help to raise wages in low
skill-based mass production (Boyer 1986).
However, economic reasons alone are not enough to explain the widely different extent of labour market regulation in
the advanced economies. Nor can they account for the resilience of such regulation. Sociological and political
factorssuch as the extension of social citizenship rights (Marshall 1964), or the need for consensus by a highly
organized and mobilized labour movement (Pizzorno 1978)must be cited as well. Job and income security for all
workers has always been of paramount importance to the European trade unions, as opposed to American-style
business unionism, which is more narrowly concerned with wage demands for unionized workers alone (Crouch
1993). Hence, objectives of social equity and redistribution in order to off-set the inequalities of income and power
produced by the market have been at the centre of the social contract that, in one way or another, has characterized
most European countries since the Second World War (Esping-Andersen 1990). These objectives have been pursued
primarily through the construction of the modern welfare state; but the institutionalization of collective bargaining on
wages and working conditions, as well as legislation protecting jobs and basic employees' rights, have also been part of
the deal.
In the 1980s and 1990s, however, profound changes have taken place in labour markets and labour market policies;
changes which have threatened the existing social contract and have called for a renegotiation. At the macro-level, the
crisis of the Keynesian welfare state has substantially decreased labour demand. In fact, macro-economic policies based
on counter-cyclical public expenditure had been instrumental in sustaining demand and achieving full employment.
Also, the
The Dilemmas of Labour Market Regulation 13
welfare state, whose initial objective was simply to reduce social insecurity, had become a major employer in its own
right, since its func-tioning required large amounts of public employees. However, the very achievement of these goals
provoked a series of unintended consequences, which eventually weakened the entire edice of public intervention
(Regini 1995). Politically induced full employmentnot induced, that is, by the natural workings of the marketthus
became one of the chief targets of criticism during the 1980s, and an objective which even the left-wing European
governments found increasingly difcult to pursue, with the temporary exception of Sweden (Scharpf 1984).
At the micro-level, on the other hand, industrial adjustment to increased competition and the ensuing reorganization
of production have bred a growing need for the exible deployment of labour. The organization of the Fordist factory
was based on a system of rules which applied in relatively uniform manner to all productive units and work
relationships (Piore and Sabel 1984, Boyer 1986). Its corresponding system of industrial relations consisted, in practice,
of bargaining over these rules according to equally uniform and standardized criteria, or of setting the price for their
acceptance. With the crisis of the Fordist system, however, this uniformity has faded and company strategies have
diversied according to the type of production re-organization that each individual rm has decided to implement.
Moreover, the human resource policies adopted by individual companies do not apply equally to all categories of the
workforce. As a consequence, the production of uniform rules at the central-national level to regulate wages, working
hours, mobility, or labour market entry is regarded by companies as increasingly less appropriate to their problems.
Flexibility is now considered to be a key element in the organization of production, and greater labour exibility, in
particular, has become of vital importance to rms.
Finally, at the policy level, the new phenomenon of unemployment hysteresis or low responsiveness to the business
cycle has prompted reconsideration of the effects of labour market regulation, which is increasingly re-conceptualized
as labour market rigidity. Since this is the specic topic of this volume as a whole, I shall not elaborate the point
further.
What should be noted here, however, is that the second and third changes mentioned above have gradually become
the focus of attention, while the rst has been largely overlooked. As a result, the issues of labour exibility and labour
market deregulation have moved to the centre of both scientic and policy debate, and alsoin some casesto the
top of the policy-makers' agendas. However, the economic and
14 Marino Regini
sociological literature has developed sharply different approaches to these issues. Not only do the assumptions on
which it rests diverge widely, but more importantly, the evidence so far produced is rather fragmentary and
contradictory. Moreover, comparative analyses in this area often suffer from over-simplication and reductionism.
Indicators of labour market rigidity are used which have different meanings and different implications in the various
countries concerned, and the role of social institutions other than state and marketfor example, family and social
networksin producing different unemployment outcomes is frequently neglected.
Although neither the scientic and policy debate nor actual labour market trends have followed uniform paths in the
European countries, there is a general tendency to equate the concepts of exibility and deregulation. A close
association is usually discerned between employers' demands for greater labour exibility and a perceived need to
remove legal restraints on the employment of labour by the extension or regaining of managerial prerogatives (EIRR
1985: 24). Or, to quote Ralf Dahrendorf (1995: 24), exibility means eliminating rigidities, hence. . .deregulating and
limiting governments interference. . .The word exibility has become practically synonymous with the relaxation of
constraints on the labour market's operation: greater freedom in hiring and ring, a possibility to increase or decrease
wages, more part-time and temporary work, higher labour turnover [translation provided1]. However, as we shall see
below, the two concepts of exibility and deregulation address quite different phenomena. Further, these
phenomena are not necessarily interrelated, since labour exibility may be the outcome of labour market deregulation
but more often results either from a change in the regulatory regime or from informal adjustments to new pressures
which leave the level of formal labour market regulation unchanged.
Types of Labour Flexibility: Trade-Offs and Contradictions

Flexibility seems to have become the catch-all word for everything that employers nd desirable and for every recipe
prescribed by policy advisers to ght unemployment and the loss of competitiveness. Not surprisingly, therefore, any
serious discussion of exibility usually starts
1
Translation into English and page number are from the Italian edition, published as Quadrare il cerchio. Benessere economico, coesione sociale e libert politica (Bari: Laterza, 1995).
with the warning that exibility is a somewhat vague and multipurpose concept; indeed exibility in economic activities
and in labour markets may be dened in several ways.2
Some authors dene organizational exibility as the ability to use machines and workers in different combinations in
order to adapt to changing market conditions (Piore and Sabel 1984). By extension, labour exibility may be taken to
mean the ability to combine the different components of the employment relation (wage levels, skills, working time, job
security, etc.) in different ways. There are two other meanings of exibility which are probably of greater signicance
for the analysis of labour markets. Both public policies and industrial relations systems tend to regulate labour markets
in ways that are both sticky, in the sense that their re-negotiation is difcult and takes time (Olson 1982), and uniform
in that they tend to be identical or similar for different sectors, regions, rms and categories of workers (Streeck 1987).
Hence, when advocating labour exibility, some commentators (for example Accornero 1992), as well as policy-makers
place special stress on the components of non-standardization, diversication, and nonuniformity of labour
performance and conditions. In this sense, demand for greater exibility may arise both from companies seeking to
destandardize labour contracts or wages and employees who are more often interested in the diversication of working
time or work roles. Other authors, drawing on a seminal work by Stigler (1939) on the different ways in which
companies react to market volatility, equate exibility in labour markets with adaptability and versatility: the ability to
adjust quickly to change. More generally, exibility is then seen as the capacity of enterprises to reorganize in close
response to uctuations in their environment.
Under all the above meanings, labour exibilitywhether understood as the ability to combine resources, to accept
diversied regulation, or to adjust to the company's changing needsis crucial if new modes of production are to
function properly. Both post-Fordist manufacturing and the organization of services require far more labour exibility
than does traditional mass-production industry. But such exibility may relate to four different areas of the
employment relationship: entry and exit from employment; the assignment of tasks and horizontal and vertical
mobility; the level and structure of compensation; and working time. Put differently, exibility may concern both the
form in which labour is contracted and the scope provided to
2
Given the purpose of this review, the focus is of course on labour exibility. However, for companies, the demands for more exible systems of employment are (or should
be) inseparable from the simultaneous search for higher exibility of product ranges, technology, capital equipment, methods of nance, and relationships with suppliers
(Streeck 1987).
16 Marino Regini
management by the contract of employment for varying labour inputs. There is wide consensus in the literature (Piore
1986, Atkinson 1987, Streeck 1987) that different types of labour exibility correspond to these areas. Although both
the number of types and the terminology used vary, they are most often called numerical or external, or employment
exibility; functional or internal, or technical-organizational exibility; wage or nancial exibility; and temporal or
working-time exibility, respectively.
Numerical exibility refers to the ease with which the numbers of workers employed can be adapted to meet uctuations
in demand (EIRR 1985: 26) or technological innovation. More generally, it denotes managerial capacity to dismiss
employees in order to allow downsizing or replace workers whose skills have become obsolete, and to use new forms
of employmentsuch as atypical, temporary, or contingent workwhen hiring new workers. All countries have some
forms of social control over lay-off decisions; even in the USA, where employers can, in theory, discharge employees
without cause, the courts have set increasing constraints on this practice and forced policy-makers to reconsider the
need for regulation. On the other hand, exibility in hiring meets with less opposition and, in the 1990s, has greatly
increased in Europe.
Functional exibility refers to the ease with which the tasks carried out by employees can be adapted to changes in
demand (EIRR 1985: 26). At stake here is the employers' ability to move employees from one task or department to
another, or to change the content of their jobs. This type of exibility refers to employers' ability to reorganize their
existing workforce so as to adapt to new technologies and new ways of utilizing their human resources by means such
as job rotation, multi-skilling, retraining, mobility and so on (Kalleberg 1990: 18). This ability is based on mutual
acceptance of a broad denition of the tasks implicit in a job, as well as on the existence of internal labour markets or
incentives to horizontal and vertical mobility within the company. From this perspective, Germany and Japan enjoy a
high functional exibility, as opposed to Britain and the USA, where job demarcation and strict seniority rules are
common. Important aspects of this are redeployment and retraining, which require institutions and payment systems
that motivate workers to take over new tasks, acquire new qualications, and generally accept continuous fast
adjustments in the organisation of work (Streeck 1987: 67). Although external and internal exibility are the most
frequently discussed forms, this latter point suggests the importance of two other types of labour exibility.
Wage exibility concerns the extent to which management is free to
alter wages and wage systems in response to changing labour market or competitive conditions; that is, to adopt pay
structures and pay levels not fully determined by collective agreements or statutory regulation. Wage exibility may
work upwards, as in the case of incentivesthe pay systems that motivate workers to take over new tasks referred to
aboveor downwards, when there are no minimum wages set by the law or by collective agreements, or, more
frequently, when derogations from such minimum wages are allowed for specic areas or occupational groups. This
type of exibility affects the extent to which wage differentials are governed unilaterally by employers or are instead the
object of institutional regulation. From a macro perspective, however, such institutional regulationor the various
functional equivalents that allow for a high degree of wage coordination in some economies (Soskice 1990)has
positive effects on real wage adjustment to changing conditions of international competition and on the ability to create
employment, as highlighted by the Dutch case.
Finally, temporal exibility refers to the possibility of adjusting the amount of labour utilized in accordance with cyclical
or seasonal shifts in demand by varying the number of hours worked in a day, week or year, rather than the number of
people employed or the degree of stability of their employment (Adam and Canziani 1998). Well-known examples are
overtime, shiftwork, exible part-time, and weekly or annual working time arrangements. This form of exibility is
often viewed as a variant of functional exibility, since both types imply adjustment in the work performance of
employees, as opposed to variation in their wages or their security of employment. However, adjusting working time is,
from many points of view, a matter quite distinct from adjusting tasks and skills. The two options are likely to be used
to differing degrees in different countries, industries and companies, depending on their skill structure but also on their
institutions and social norms.
The Varying Role of Flexibility in Companies' Strategies

Each type of labour exibility varies in its importance to a company according to its product market strategy; that is,
according to the specic manner in which the company chooses to compete. This therefore requires a brief discussion
of the competitive strategies available to companies,3 in order to assess the role that the various forms of labour
exibility come to play. In the post-Fordist era, rms are able to choose among a wide range of organizational and
market strategies. While price
3
For which I draw extensively on Regini (1997).
18 Marino Regini
constraints remain important to all of them, some compete chiey on quality or design, others on product
diversication, and yet others on organizational exibilityin the sense of versatility and rapid adjustment to changing
demand. For the purposes of this discussion, I shall call the rst of these market strategies diversied quality
production (DQP), borrowing Streeck's (1991) term but giving it a somewhat different meaning. The second I shall
call exible mass production (FMP), taking up the distinctions drawn by Boyer (1987). The third is exible
specialization (FS), although this concept, in its original formulation (Piore and Sabel 1984), was wider in its scope.
In the terminology used here, a DQP strategy is adopted by rms which set out to compete on the qualityand to a
lesser extent on the diversicationof their products, rather than solely on their price. Their aim is to avoid
competition from low-wage economies by targeting higher market segments, or by responding to the greater
sophistication and volatility of demand through product customization. Quality thus dened is made possible by
factors such as signicant organizational and coordination capacities, but a crucial role is played by the high and broad
skilling of all occupational groups in the workforce. The workers must be able to integrate several tasks in the
performance of work, to learn new tasks rapidly, and to be involved in corporate objectives of constant improvement
and incremental innovation. This corresponds to a pattern of human resources utilization where a large proportion of
the workforce receives extensive vocational trainingboth basic and company-specicand is encouraged to develop
such social skills as initiative, a problem-solving attitude, and an ability to work with others, as well as to develop a high
degree of identication with the company. In sum, high functional exibility is required of the whole workforce, and
this is matched by low levels of numerical, wage, and temporal types of exibility.
An FMP strategy is instead based on the mass production of a variety of goodsrather than the standard goods of
classic Fordismin order to meet changing demand and volatile markets while holding prices down. Programmable
automation enables the company to compete simultaneously on both price and product diversication, by
massproducing a wide range of products and drastically reducing demand for medium-to-low skilled personnel (blue-
collar production workers and administrative staff), as well as emphasizing adaptability to change and cooperation
rather than technical abilities. On the other hand, demand for high-level skills is concentrated on a few key
occupational groups, especially middle management, technicians, and personnel in the commercial area (sales,
marketing, customer relations). The pattern of human resources utilization deriving from these features exhibits a
polarization between highly-skilled personnel belonging to these occupational groups and the low-skilled, as well as
workers whose skills have become obsolete. Hence, the entire workforce is required to display a high level of exibility;
but this is mainly of the functional (polyvalence) and wage (incentives) types for the core occupational groups, while it
is mainly numerical (temporary work, training and work contracts, possibility to lay-off) for the low-skilled groups.
An FS strategyin which the principal weapon in a company's competitive armoury is its versatility and rapid
adjustment to changes in demandis particularly common among small rms, although it is not conned to these. In
fact, small-rm systems typied by low costs for entry into the market and exit from it, extremely low organizational
costs, and low costs for production plan selection and errors, are best equipped to respond rapidly to quantitative and
qualitative changes in demand, or even to anticipate them. Diffuse-production systems, in fact, enable a wide range of
products to be placed on the market, leaving it up to the market to select the ttest. The crucial human resource in
this type of rm is the entrepreneurand his/her assistantswho must be able to carry out a wide range of
functions, using external consultants but often with very few employees. The entrepreneur and his/her assistants must
possess broad and variedif not necessarily very sophisticatedtechnical expertise, and social skills which are
difcult to acquire through formal training. The small number of employees in these rmsusually enjoying less
institutional protectionknow that they are naturally subject to numerical exibility. Yet it is crucial that they exhibit
high temporal and functional exibility, pragmatic adaptability, and a general willingness to cooperate, since they must
be able to shift among different machines and adjust their workload to changing demand.
The discussion so far has been restricted to the industrial sector. Though extremely important for employment
creation, in fact, services have received less attention as far as their competitive strategies are concerned, and stylized
discussion of what types of exibility they most need is difcult. The available generalizations (e.g. Esping-Andersen
1990) may be too crude from this point of view, because they fail to disaggregate the two broad categories of personal
and business-related services in a sufciently detailed way. The only general statement that seems to be supported by
the empirical evidence is that the dualism between low-skilled and high-skilled workforceswhich has been discussed
as characterizing FMP strategiesis especially relevant in services. Even more than in industrial companies, wage
exibility appears to be the only way to create employment for low-skilled personnel in services, whereas functional
and temporal exibilityand the
20 Marino Regini
availability of social-relational skills to an even greater extentare generally demanded of the high-skilled.
Contradictions Between the Different Types of Flexibility

Except perhaps for the FS-based rms, there seems to be a contradiction among the four types of labour exibility, in
that the determined managerial pursuit of some forms may make it impossible to achieve the others. The well known
trade-off between numericaland wageexibility on the one hand, and functionaland temporalexibility on
the other, is of especial relevance here. Too much of the former usually prevents achievement of a sufcient amount of
the latter, in that it implies a lack of trust and cooperation, an unwillingness to share information, a disincentive against
long-term investment in human resources, and greater resistance to technological and organizational change (Gutchess
1985, Deakin and Wilkinson 1996).
Cooperation is required for exibility in both the substantive areas such as information sharing and in the
procedural ones involving change in work rules. The biggest obstacle in either case is the fear that the other party
will not reciprocate and will seek to take advantage of the rst party's cooperative stance and grasp all the fruits for
itself. . .It is very difcult to achieve mutual cooperation without the presence of trust, which we might dene here
as a mutual expectation of cooperative behaviour, and without some social or institutional framework to sustain
that trust. (Marsden 1995: 767)
The most obvious supporting framework is of course one which guarantees a fair level of protection against job and
income losses.
This problem has long applied to traditional workplaces with a low degree of employment security, but similar
observations have been made about the increasing use of temporary and atypical workers. As the Wall Street Journal put
it as early as 1987, some corporations aren't sure they like temps. That very lack of identication with the corporation
and its aims and the absence of loyalty to the corporation and fellow workers is seen as a business handicap.4 Hence,
where numerical exibility is high, rms tend to under-invest in human resource development, and employees are not
encouraged to undertake rmspecic training and to show commitment to the organization.
On the other hand,
the conditions that lead to functional exibilitylong-term employment relations, strong attachment [to the
company]mean that labor becomes a xed cost for the organization. Labor as a xed cost, in turn, decreases
employers' wage and employment exibility. There is thus a contradiction between increasing
4
Quoted in Kalleberg (1990: 21).
commitment and functional exibility, on the one hand, and obtaining greater employment/wage exibility, on the
other. (Kalleberg 1990: 1819)
The most common way out of this dilemma is of course segmentation of the workforce. The Japanese recipe has long
been high functional and temporal exibility for a core group of permanent employees and extreme numerical and
wage exibility for the large periphery of temporary, contingent jobs. Yet this recipe has become decreasingly
successful as companies have found it more difcult to guarantee employment security to the core workers even in
recessionary periods, and as demands for equity and exposure to other cultures have made sharp labour market
dualisms unsustainable in the long run.
Aside from Japan, continental European and Anglo-saxon countries have long displayed different responses to the
trade-off between numerical and wage exibility on the one hand and functional and temporal exibility on the other.
However, it seems increasingly necessary to consider the combined effects of the different types of exibility by taking
this trade-off into account and nding more balanced solutions than those tried to date. This may require deep
changes in labour market regulatory regimes; changes which involve the mechanisms and the levels of regulation and
which may require the imposition of benecial constraints on economic actors (Streeck 1987), thus giving rise to a
different type of regulation, rather than just less of it. Which brings us to a second set of analytical clarications.
Regulatory Regimes and Deregulation in Advanced Economies

The debate on deregulation reveals frequent misunderstandings of how advanced economies are regulated. Some
conceptual clarications and analytical discussion are therefore needed if actual trends in the late 1990s are to be
properly assessed.
First, only too often in the policy debate is the meaning of regulation or governance of the economic system reduced
to that of state intervention as opposed to the workings of the market, or any aspect that prevents the market from
operating properly. However, in the political economy literature, regulation of the economy denotes the various ways
in which the set of activities and relationships associated with the production and distribution of economic resources
is coordinated, [these] resources allocated, and the related conicts, whether real or potential, structured (that is,
prevented or reconciled) (Lange and Regini 1989: 4).
22 Marino Regini
Second, reference to the various modes or forms of regulation of the economy by itself suggests the existence of both
a plurality of institutions with a regulatory role in advanced economies and a plurality of levels at which economic
activities may be coordinated. Let us rst briey recall what the institutions in question actually are (for the classic
analysis of this issue, see Polanyi 1944).
1. The market is the regulatory institution of the economy par excellence. Where it predominates, the coordination
of activities, alloc-ation of resources and structuring of conicts occurs as the outcome of exchanges based on
prices determined by the interaction of supply and demand under conditions of competition. In the idealized
market, competition is widely dispersed and is not inuenced by normative linkages or by the exercise of
power and authority. Of course this rarely occurs in reality, especially as far as labour markets are concerned.
2. The state can coordinate activities, allocate resources and structure conicts primarily through the exercise of
its authority, which, in the last analysis, is based on its monopoly of legitimate coercion. In this case, regulative
activity occurs primarily by means of laws and administrative rulings which are binding on the actors involved.
3. Where community institutions and social networks predominate, the coordination of activities and the
allocation of resources takes place primarily through forms of normative compliance or spontaneous solidarity.
This solidarity can be rooted in norms, habits, or values shared by members of the community, and is based on
respect, trust, or simply identication with the community and thus with its rules and hierarchy.
4. Finally, a wide range of economic activities appear to be regulated by accords reached among a few large
interest associations with quasi-monopoly of representation of functional interests and a high level of
disruptive power, and which, as a result, are able to obtain privileged recognition from other associations and
political authorities (Streeck and Schmitter 1985).
Moreover, economic activities may be coordinated at numerous levels, among which the micro levels of the rm and
of the region are of increasing importance. In several European countries, new patterns of micro-concertation,
whether formal or informal, have emerged, demonstrating their ability to encourage economic adjustment. These are
all forms of consensual coordination at the company level, and sometimes at the territorial level as well, based on the
shared acceptance of goals, constraints and compatibilities (Regini 1995).
Third, as far as the labour market is concerned in particular, the decline of both rigid state legislation and tripartite
concertation at the macro-national level in the late 1980s has not, in general, led to a crisis tout court in its regulation.
This is partly because hidden exibilities traditionally co-exist alongside the ofcial regidities, so that when
opportunities to break such rigidities materialize, few actors try to exploit them to achieve full deregulation (see below).
On the other hand, hidden rigidities preventing a potential excess of exibility are often observed in the most
deregulated countries like the USAas various studies on the non-union sector or on the green-eld sites in that
country show. After all, even the Wild West needed some rules, and the failure of legislation and collective bargaining
to provide them may paradoxically result in a rigidication of labour relations through the enhanced role of such
mechanisms as informal negotiation, practices or conventions, normative bonds and values shared by the working
communitynot to mention court rulings on such issues as dismissals.
Hence, the regulatory regime of any advanced economy is a complex one. Opening re against some modes or levels
of regulation may simply lead to their substitution by other modes or levels. As regards labour markets in particular,
very rarely can changes in their regulatory regime be understood as simply an extension or a restriction of the role of
statutory and associational regulation to the advantage or disadvantage of market mechanisms.
Fourth, any regulatory regime has multiple, and often contradictory, effects on the economy. Depending on their
interests, economic actors are likely to focus on some effects while overlooking the othersor to conceptualize them
in consistently positive or negative wayswhereas the overall picture is often mixed and even blurred. At the most
general level, a regulatory regime creates rules on individual behaviour and institutions which embody these rules.
However, some of these rules and institutions may workor may be understoodas constraints, burdens, or even
rigidities, on economic action, whereas others may provideor be perceived as providingresources, opportunities,
or even collective goods. In other words, institutional regulationwhether public or associationalmay be
facilitative rather than restrictive in character. This is especially true of the role of institutions vis--vis rms; a role that
does not stop at the simple imposition of constraints on rms, since institutions often full the crucial function of
providing rms with resources, advantages and opportunities, so that they structure, as it were, the economy in which
the rms operate.
24 Marino Regini
Thus the very meaning of deregulation of the economy, and of labour markets specically, is multi-dimensional and
basically ambiguous. More importantly, any process of intended deregulationor any change in the regulatory regime,
for that mattermay have uncertain or unintended effects. At one extreme, deregulation is conceived by its
proponents as an attempt to minimize all rules on individual behaviour and all the functions performed by state and
associational institutions, in order to pursue the ideal of a pure market. The essence of the deregulatory approach is
to increase the autonomy of individual rms and workers to respond to shifting markets. Within labour markets, this
has usually been taken to mean that the institutional structures regulating them, be they of state or collective-bargaining
origin, should be dismantled (Marsden 1995: 83). In this case, however, the usual counter-argument of market failures
involving information, sunk costs, and the likely under-provision of public goods, fully applies. This is probably the
main reason why attempts to dismantle associational and public regulation have been far more limited than the change
in power relations that occurred in the 1980s, and the weakness of political opposition to such change, would have
allowed.
Alternatively, the proponents of deregulation may pursue the less ambitious objective of removing only those rules and
institutions which, from a comparative point of view, may be shown to impose excessive rigidities on economic
activity. However, as discussed above, the same institutions that set constraints may on the other hand provide
resources and competitive advantages, so that the cumulative effect of even limited deregulation is not easy to
determine, or may be largely unintended. A well-known example in the labour market area is the role of trade unions,
collective bargaining, works councils and other industrial relations institutions. All such mechanisms for collective
representation and collective voice usually act as a constraint on managerial authority and prevent some forms of
exibility, especially (downward) wage exibility. On the other hand, in several countries they have greatly enhanced
temporal and functionaland in some cases even numericalexibility, not to mention their crucial role for rms in
providing workforce cooperation.
Finally, at the other extreme, deregulation is often used as a term which denotes far less dramatic changes. This is the
case of processes which scale down the role of some instruments of economic regulationsuch as the law, or tripartite
concertationto the advantage of othersdecentralized collective bargaining, or informal agreements. The concept
of change in the regulatory regime rather than deregulation is, in such instances, all the more appropriate. In fact, what
is involved here is not an attempt to restore the free market, unfettered
by rules and constraints, but rather to create a new, possibly more exible, regulatory system which allows for greater
diversication and faster adjustment to change. As far as labour markets are specically concerned, the objective in
this case is not so much downward adjustment in the terms of employment, as the decentralization of regulatory
mechanisms with a subsequent variety of results (Streeck 1987: 64). To be sure, the workplace-specic regulation of
employment relations may be instrumental in lowering the overall level of constraints on labour market behaviour. But
the main effect of the decentralization of negotiations and decisions is to permit a custom-made organization of work
that ts the specic needs of rms rather than the requirements of general regulations (Streeck 1987: 65), namely that
road to competitiveness which allows rms to create jobs without relying too heavily on numerical and wage exibility.
Changes in Regulatory Regimes and Labour Market Policies

What general conclusions can be drawn from the foregoing examination of the validity and potential of the concepts
and analytical tools used in the debate on labour markets?
Let us rst summarize briey. The regulatory regime of any advanced economy is a highly complex system, since it
involves different institutions, levels and mechanisms, and has multiple, and often contradictory, economic effects.
Hence, outright deregulation of labour markets may be neither feasible nor benecial in terms of greater exibility.
Firstly, attempts to dismantle some institutions or rules may simply lead to their replacement by other regulatory
modes and mechanisms, or they may generate hidden rigidities. Secondly, since the institutions that set constraints may
also provide resources and competitive advantages, the cumulative effect of such deregulation may be negative.
On the other hand, the term deregulation is often used simply to denote a change in the regulatory mechanisms,
especially their decentralization in order to achieve a greater variety of results, or the introduction of controlled
exceptions to the general rules. These have been the forms most frequently adopted in European reforms in the
1990s. Greater exibility has thus been injected into European labour markets. However, labour exibility may
concern different areas of the employment relationship: labour market entry and exit, the assignment of tasks and
internal mobility, the level and structure of compensation, and working time. Each country has focused on some of
these areas
26 Marino Regini
more than on others, depending on their viability in terms of consensus but also on the country's production structure,
which requires different forms of labour exibility. None of them has managedor even attempted, for that
matterto achieve a great deal of exibility in all these areas at the same time, mainly because of the contradiction
discussed above.
Even at this preliminary, analytical level, then, it seems that exibility and deregulation are far from being simple
recipes with univocalwhether positive or negativeeffects on labour market performance. Not only do exibility
and deregulation have numerous meanings, more importantly they may have a range of consequences which should be
carefully considered before radical action is taken. While the exibilization trends of the 1990s in Europe have had
little positive impact on unemployment, the foregoing discussion suggests that further or more drastic moves to labour
market deregulation may hinder, rather than foster, national competitiveness, thereby exacerbating the problem of
unemployment in the long run.
Basically, two alternative roads to competitivenesswith several variants in betweenhave been followed in Europe,
and they largely correspond to the varieties of capitalism discussed in a growing body of literature (Albert 1991,
Berger and Dore 1996, Crouch and Streeck 1997). At one extreme we nd a change in the regulatory regime aimed at
decentralization, the exercise of unilateral authority, and deregulation in the stronger sense set out above; at the other
we nd a change based on wage coordination, investment in training and consensual re-regulation (Soskice 1989).
The national studies in this volume discuss the extent to which individual countries have followed one of these two
roads, and how they have dealt with the ensuing trade-off between greater exibility and the adequate provision of
collective goods such as social cohesion, wage coordination, trust and cooperation, availability of human capital. The
actual solutions to this trade-off comprise variations and mixed processes. But neither the low road to
competitiveness based on a low-wage, low-skill, low-involvement, and low-quality equilibrium; nor the high road
entailing high wages, high skill, high co-operation and high product quality, have proved fully able to overcome
problems related to them. This applies to national competitiveness as an outcome of regulatory change, as well as to
the issue of solving the joint problem of unemployment and inequality.
Hence, what types of labour market policies does the above discussion suggest as possible alternatives, or even as
benchmarks against which to assess developments in Europe in the 1990s and
References 27
their impact on the structure of employment and unemployment described in the chapters that follow? More generally,
what types of change in the regulatory regimes of labour markets are envisaged as enabling a labour market
performance which combines high employment levels, greater equity, preservation of social cohesion, and company
competitiveness? Detailed discussion of the various options, of the trade-offs in terms of costs and benets, as well as
of their actual or likely outcomes, is conducted in the chapters that follow, which draw systematically on the evidence
presented in the country studies.
The two extreme approaches to labour market reform can be summarized by two opposing slogans. The rst: the
more exibility of all types you achieve the better, and you need labour market deregulation to achieve it. The second:
exibilization of the labour market is irrelevant or even dangerous, since the main solution to the problem of
unemployment is macro-economic stimulation, scal policies, expansion of services, etc.. The chapters that follow
show that both these polar positions are either wrong or only partially true, and the aim of this chapter has been to
provide the analytical tools to understand why. Between these two poles, in fact, lie a series of less dramatic alternatives
constituting the menu from which European policy-makers have, more or less hesitantly, drawn. This menu can, and
probably will, be enriched and made more sophisticated, but before calling for radically new dishes it is necessary to
understand better what the various ingredients are, whether they go together well, and what the likely side effects are.
References
Accornero, A. (1992), Are we going towards a society of activity or work?, Labour, 6/3: 89107.
Adam, P., and Canziani, P. (1998), Partial de-regulation: xed-term contracts in Italy and Spain. London: Centre for Economic
Performance, Discussion paper No. 386.
Albert, M. (1991), Capitalisme contre capitalisme. Paris: Editions du Seuil.
Atkinson, J. (1987), Flexibility or fragmentation? The United Kingdom labour market in the eighties, Labour and
Society, 12: 87105.
Berger, S., and Dore, R. (1996) (eds.), National Diversity and Global Capitalism. Ithaca, NY: Cornell University Press.
Boyer, R. (1986) (ed.), La exibilit du travail en Europe. Paris: La Dcouverte.
28 References
Boyer, R. (1987), Labour exibilities: many forms, uncertain effects, Labour and Society, 12: 10729.
Crouch, C. (1993), Industrial Relations and European State Traditions. Oxford: Clarendon Press.
and Streeck, W. (1997) (eds.), Political Economy and Modern Capitalism: Mapping Convergence and Diversity. London:
Sage.
Dahrendorf, R. (1995), Economic opportunity, civil society and political liberty, paper presented at the United Nations
conference on Rethinking social development, Copenhagen, 1112 March.
Deakin, S., and Wilkinson, F. (1996), Contracts, Co-operation and Trust: The Role of the Institutional Framework, in
D. Campbell and P. Vincent-Jones (eds.), Contract and Economic Organisation. Dartmouth: Aldershot.
EIRR (1985), European Industrial Relations Conference, European Industrial Relations Review, 138/July: 247.
Esping-Andersen, G. (1985), Politics against Markets. Princeton, NJ: Princeton University Press.
(1990), The Three Worlds of Welfare Capitalism. Princeton University Press.
Granovetter, M. (1985), Economic action and social structure. The problem of embeddedness. American Journal of
Sociology, 91/3: 481510.
Gutchess, J. F. (1985), Employment Security in Action: Strategies that Work. New York, NY: Pergamon Press.
Kalleberg, A. (1990), Commitment and exibility: changing employment relations in industrial societies, paper presented at
workshop on The nature of consensus at the place of production, Siena, 1415 June.
Lange, P., and Regini, M. (1989) (eds.), State, Market and Social Regulation. Cambridge: Cambridge University Press.
Lindblom, C. (1977), Politics and Markets. New York: Basic Books.
Marsden, D. (1995), Deregulation or Co-operation? The Future of Europe's Labour Markets, Labour special issue:
S6791.
Marshall, T. H. (1964), Class, Citizenship, and Social Development. Garden City, NY: Doubleday.
O'Connor, J. (1973), The Fiscal Crisis of the State. New York: St. Martin's Press.
Olson, M. (1982), The Rise and Decline of Nations. Economic Growth, Stagation and Social Rigidities. New Haven, Conn.: Yale
University Press.
Piore, M. (1986), Perspectives on labor market exibility. Industrial Relations, 25: 14666.
and Sabel, C. (1984), The Second Industrial Divide. New York: Basic Books.
Pizzorno, A. (1978), Political Exchange and Collective Identity in Industrial Conict in C. Crouch and A. Pizzorno
(eds.), The Resurgence of Class Conict in Western Europe Since 1968. London: Macmillan, Vol 2: 27798.
Polanyi, K. (1944), The Great Transformation. New York: Holt, Rinehart & Winston.
Regini, M. (1995), Uncertain Boundaries. The Social and Political Construction of European Economies. Cambridge: Cambridge
University Press.
(1997), Social Institutions and Production Structure. The Italian Variety of
References 29
Capitalism in the 1980s in C. Crouch and W. Streeck (eds.), Political Economy and Modern Capitalism: Mapping Convergence
and Diversity. London: Sage, 10216.
Scharpf, F. (1984), Economic and Institutional Constraints of FullEmployment Strategies: Sweden, Austria and W.
Germany 19731982 in J. Goldthorpe (ed.), Order and Conict in Contemporary Capitalism. Oxford: Clarendon Press,
25790.
Sengenberger, W., and Campbell, D. (1994) (eds.), Creating Economic Opportunities: The Role of Labour Standards in Industrial
Restructuring. Geneva: International Institute for Labour Studies.
Solow, R. (1990), The Labor Market as a Social Institution. Cambridge, Mass.: Blackwell.
Soskice, D. (1989), Reinterpreting Corporatism and Explaining Unemployment: Coordinated and Non-Coordinated
Market Economies in R. Brunetta and C. Dell'Aringa (eds.), Markets, Institutions and Corporations: Labour Relations
and Economic Performance. London: Macmillan, 170211.
(1990), Wage determination: the changing role of institutions in advanced industrialized countries. Oxford Review
of Economic Policy, 6/4: 3661.
Stigler, G. J. (1939), Production and distribution in the short run. The Journal of Political Economy, 7.
Streeck, W. (1987), The uncertainties of management in the management of uncertainty. International Journal of Political
Economy, 17/3: 5787.
(1991), On the institutional conditions of diversied quality production in E. Matzner and W. Streeck (eds.),
Beyond Keynesianism. Aldershot: Edward Elgar, 2161.
and Schmitter, P. (1985), Community, Market, Stateand Associations? The Prospective Contribution of
Interest Governance to Social Order. European Sociological Review, 1/2: 11938.
Williamson, O. (1975), Markets and Hierarchies: Analysis and Antitrust Implications. New York: Free Press.
2 The Dynamics of Labour Market Reform in
European Countries
Manuela Samek Lodovici
Introduction
The high unemployment record of European Union (EU) countries since the mid-1970s can be largely attributed to
supply-side shocks and to the ensuing restrictive stance of macro-economic policies. It is however argued in the
current literature that the effects of these shocks were intensied in many countries by institutional features. The
greater rigidity of labour markets, due to employment protection rules and higher unionization rates in European
countries relative to nonEuropean ones, together with the presence of less competitive product markets, are
considered to be important explanations for the high level and persistence of European unemployment rates and of
the low capacity of European countries to create jobs.
Employment protection rules and real wage rigidity increase the adjustment costs of rms, and may reduce labour
demand for outsiders, giving rise to an inequitable distribution of job opportunities. Generous social benets, in
addition, may reduce the willingness of the unemployed to accept available jobs and increase the duration of out-of-
jobs spells. The competitive position of rms and their productive strategies may also be affected: investments may be
disincentived and, if there is a minimum size threshold for the application of employment protection rules, there may
be an incentive for rms not to increase their size, to decentralize production to smaller rms or to ones located
elsewhere, and to circumvent regulations in the underground economy.
On the other hand, as discussed in Chapter One, employment protection may provide strong incentives for functional
exibility within a rm, for investment in training and human capital, and for wage moderation. Labour market
regulation may also reduce the individual and social costs and problems linked to the existence of market failures in the
denition of labour contracts, that is, uncertainty and imperfect,
The Dynamics of Labour Market Reform 31
asymmetric information, heterogeneity of workers and rms,5 and high income inequality.
There are, however, costs associated with a government's failure to provide adequate regulation frameworks: costs
which should be minimized. Policy failure may be the result of design failurescaused for example by imposing rigid
thresholds, the superimposition of social and distributional objectives, the existence of vague standards and rules
creating legal uncertainty, or the combination of measures which may result in excessive rigidityimplementation
failuresthat is, inconsistent labour court decisions or the inconsistent implementation of legal norms by public
agencies, the lengthiness of authorisation or arbitra-tion procedures, the lack of co-ordination and consistency between
different measures and incentivesor the structural inability to adjust rules according to changing environments.
In order to compare regulation systems across countries and to assess their impact on labour market performance, it is
necessary to consider the different combinations of employment protection, welfare provisions and wage bargaining,
the coverage of regulations, current practices and the presence of measures that facilitate adjustment. In European
countries, often strict hiring and ring regulations are accompanied by measures aimed at facilitating labour adjustment
through softer mechanisms such as the adjustment of working hours, early retirement systems, and incentives to
voluntary redundancies.
The trade-offs between labour market regulation and jobs may have intensied in recent years due to greater need for
adjustment, in addition specic institutional combinations of regulation are likely to have become increasingly
inefcient. In the 1960s and 1970s the better employment performance of Scandinavian countries was explained by the
extensive use of active labour market policies, the centralization of their industrial relation systems and the creation of
jobs in the public sector instead of generous unemployment benets and employment protection. The recent success
stories of The Netherlands, Denmark and Great Britain have been ascribed to the introduction of exibilities in their
job security provisions and in wage bargaining arrangements, and to the introduction of work incentives in their
income support systems.
Since the mid-1980s, enhanced market volatility due to increased international competition and accelerated exogenous
technological change has raised the costs to employers of employment security. Moreover, ring restrictions have
become more binding for rms, given the low levels of quits due to adverse economic conditions, increased future
5
Implicit contract theories show that there is an incentive for the use of long-term contracts between the parties even in the absence of external regualtions.
32 Manuela Samek Lodovici
uncertainty and the reduced spending capacity of the public sector. Adjustment speed becomes an important, if not a
core, prerequisite for international competitiveness, and according to some commentators (Buchtemann and Walwei,
1996), the logic of internal labour markets that underlies most dismissal protection regimes has become outdated, and
institutional inertia creates inefciencies.
Centralized and solidaristic wage bargaining has also come under pressure: market forces and technological innovation
give rise to more idiosyncratic production processes, while they require more individualised reward systems between
and within rms and greater wage differentials. The shift of demand towards services has increased the role of white-
collar workers, who also demand greater wage differentials, and it reduces the size of average production units, thus
contributing to the decline in unionization and in union bargaining power (Soskice 1990). In addition, the slowdown in
ination, which was one reason for centralised bargaining, has reduced the need for this level of bargaining (Freeman
and Gibbons 1995).
Demographic and social changes have applied further pressure on the regulatory framework established in the late
1960s: the ageing of the population has increased the nancial burden of pensions on the welfare state, especially in
countries where early retirement and/or disability pensions have been widely used to support external exibility (Italy,
Germany, France, The Netherlands, Sweden). The support for continental European welfare systems has also been
eroded by the crisis of stable life-employment for core workers and by the growing instability of families.
In this chapter, a comparison is conducted between the patterns of labour market regulation that prevail in European
countries, providing evidence of differences of approaches and changes that have occurred in recent years, on the basis
of the recent comparative literature and the country reports presented in Part II.
Regulatory Approaches in European Countries in the Late 1980s

Comparisons of regulatory regimes, in order to derive information on the level of exibility in the functioning of the
labour market, should consider their coverage, the role of legislation versus collective agreements, judicial
interpretations, industrial relations systems, and current practice. Similar levels of regulation may involve very different
practices, which often explain differences in outcomes linked to similar
changes in the regulatory framework; for example the liberalization of atypical contracts in the mid-1980s had much
more marked effects in Spain than it did in Germany, a result that highlights the importance of the employers' attitude
and of the broader incentive structure of the latter country.6
Most comparative studies on labour market regulation have been restricted by the large number of aspects to be
considered when evaluating regulatory frameworks and the degree of exibility. Classication tables of the stringency
of regulation have been constructed, mainly focusing on hiring and ring regulations (OECD 1994). European
countrieswith the exception of the UKare usually presented as the most rigidly regulated of OECD countries, but
variation in the institutional combination of labour market regulation is higher within Europe than between European
and non-European countries. EU member states seem to choose a higher degree of workers' protection relative to
non-EU countries, but within Europe such protection is achieved by different combinations of job security provisions,
income support during unemployment and wage bargaining patterns as shown in Fig. 2.1 (Buti et al. 1998).
Recent discussion has paid close attention to the complex interactions among different dimensions of labour market
exibility, and the impact of regulations on labour market performance is now considered in terms of its real
stringency. For example it has been shown (Abraham and Houseman 1994) that European countries with stringent
employment protection rules adjust working hours rather than workers to changes in the economic cycle. Moreover, a
dynamic approach prevails, whereas the effects of regulatory constraints are considered in relation to the cycle or
external economic conditions (see chapter 3).
In order to compare regulatory regimes within Europe EspingAndersen's classication has been adopted, which
appears to be sufciently comprehensive to consider all the elements of interest in this context. Three main regulatory
regimes are considered: the liberal regime of the UK and Ireland, the social-democratic or corporatist regime of the
Scandinavian countries, and the conservative regime of the continental and Southern European countries (Esping-
Andersen, 1990).
Liberal welfare regimes are characterised by a limited role of the state which only supports, through social guarantees,
bad riskssuch as long-term unemployment, poverty and social exclusionwhile all other risks are left to the
operation of the market and private resources. Employment protection is governed by common law, which relates only
to the violation of individual rightssuch as unfair dismissal.
6
In Germany two-thirds of establishments did not use xed-term contracts because they were interested in long-term employment relationships.
Fig. 2.1. Strigency of employment protection (EPL) and wage-setting regulation and expenditure on labour market
policies
Notes: (a) average between EPL ranking and excess coverage (coverage-minus union density), (b) labour market policy
expenditure in % of GDP in 199697/unemployment rate 1997.
Source: own calculation on OECD (1998a) and Buti et al. (1998)
Decentralized and uncoordinated industrial relations at the rm/plant level govern wage determination.
The regimes of the Scandinavian countries are, conversely, based on the notion of the citizen's right to comprehensive
and generous risk coverage through the provision of public supportin the form of both transfers benets and
services. Active labour market policies (ALMP), public services and employment maximization are the main forms of
labour market intervention, while wage-setting and regulation of working conditions are left to centralized bargaining
between unions and employers' associations. The sectoral reallocation of labour does not operate through wage signals,
but mainly through quantity signals and ALMP, given the pursuit of solidarist wage policies by centralized
bargaining (Pissarides and Moghadam 1989). Placement in labour market programmes and relief work prevent entry
into long-term unemployment, while the outow from unemployment is supported by ALMP aimed at nding jobs or
providing training opportunities for all those unemployed for more than a year.
The conservative regimes of Continental and Southern Europe are based on a combination of high employment
protection regulation, high institutional support to union bargaining and relatively low support during non-
employment. Welfare schemes are usually fragmented according to occupational status and the support of male
breadwinners. Income support during unemployment is usually guaranteed for core workers, while families provide
care services and support for secondary components of the labour force. These countries also present coverage rates
of collective agreements usually much higher than unionization rates, due to the mandatory extension of union
negotiations to the non-union sector.
In Southern Europe, employment protection is especially high for core workers, while external exibility is obtained,
partly through the use of short time and early retirement schemes usually nanced by the state, and partly through the
large diffusion of small rmswhich do not comply with employment regulationsand self-employment. The result
is a highly segmented labour market with young people and women usually excluded from protected employment. The
Southern European countries are also characterized by the even stronger role played by families in inuencing labour
market outcomes: the reservation wage of secondary workers is closely linked to family conditions.
Employment Protection Regulation and External Flexibility

Table 2A in the Appendix provides a summary ranking of employment protection across OECD countries drawn up
by the OECD Jobs Study and indicates relative to recent changes. The country ranking describes the strictness of the
legal framework concerning hiring and ring rules and the regulation of xed-term contracts. It suffers from all the
shortcomings associated with such rankings, which are unable to take account of the effective stringency of regulatory
standards in labour market practice. It is based on indicators of regular procedural inconveniencesprocedures and
delays to start of noticeperiod of notice and severance pay for individual dismissals, and the difculty of
dismissalsdenition of unfair dismissal, trial period and reinstatement rules. European countries appear on average
to be more regulated than non-European onesthe USA and Japan in this case.
Some further aspects are worth emphasizing:
1. The differences within Europe are larger than is usually thought: the difference between the levels of regulation
in the Southern European countries and in Britain, Ireland and Denmark is greater than the difference
between levels in the USA and those of the UK and the Nordic countries.
2. These differences are the result of different regulatory approaches, which must be considered as a whole and in
terms of their interrelations. Statutory rights against unfair dismissals exemplify the different approaches taken
by European countries to labour market regulation. While the continental and, especially, southern European
countries rely heavily on statutory provisions designed to prevent job losses for core workers, the accent in the
UK and, especially, Scandinavian countries is on supporting mobility rather than on guaranteeing jobs, so that
greater weight is placed on nancial compensation for job loss and active labour market policies than on job
preservation.7
3. There is closer similarity among European countries in their regulation of collective dismissals than in their
regulation of individual dismissals. This is due to their adoption of a common EU directive on collective
dismissals in all member states, which provides for prior consultation with the employees' representatives and
requires advance notice to be given to the labour market authoritieswhich in many cases provide
accompanying measures. All countries provide for statutory monetary compensation. Differences across states
relate mainly to the notion of mass dismissals, to which special regulations apply. The Southern European
states and France have stricter denitions than the EC minimum.
4. Fixed-term and part-time work have long been restricted to specic casespublic sector work and seasonal
workin Continental and Southern European countries.8 It is especially the regulation of xed-term and
agency contracts which differs greatly among European countries and appears to be closely linked to the
regulation of dismissals. Regulation usually involves the denition of specic situations in which such contracts
are permitted, the maximum length of time available, and the number of renewals allowed. The tightest
restrictions on xed-term and agency contracts are to be found in Continental and, especially, Southern
Europe. However, in these countries, specic types of xed-term contracts targeted on rst-time
7
There are, of course, differences among countries within each category. For example, although Sweden and Denmark display similar basic characteristics, they differ in their
regulation of unfair dismissal, which is relatively restrictive in Swedenproviding for lengthy notice periods and seniority rulesbut very liberal in Denmark.
8
The low use of part-time work in Southern Europe is not due only to the regulatory framework but also to its scant appeal for small industrial rms and low-income
workers.
job seekers or the long-term unemployed have been liberalized or incentivated for labour market policy purposes and
casual or seasonal work by independent workers is widespread and little regulated (Table 2B).
5. Non-European countries, notably the USA and Japan, do not have specic regulations on rings. In Japan, the
only constraint is thirty days advance notice, but a particular ethical code in large rms sets restrictions on
ring similar to those prevailing in Europe: the reasons for dismissals must be stated, alternatives must be
considered, and severance payments are high since they also have to substitute for the lack of a public
retirement scheme. In the USA, outside the union sector and the public sector, employment regulation is
governed by common law on the basis of the employment at will doctrine to the effect that the employer can
legally terminate an individual open-ended contract without cause. These principles have been attenuated since
the 1970s as a result of state legislation and the courts, which have recently recognised exemptions in order to
protect workers exercising statutory rights and workers refusing to act unlawfully (Addison and Hirsch 1997:
22). In addition, some states in the USA have recently introduced just cause legislation and since 1988 the
federal law requires sixty days advance notice for rms with at least 100 employees. It is interesting to note that
unjust dismissal legislation is regarded by employers to be an acceptable political compromise with respect to
court rulings, which are much less predictable and usually involve large compensatory and punitive damage
awards (Krueger 1991).
6. Exemptions from coverage, public labour market programmes subsidizing work-sharing as an alternative to
redundancies, and xed-term and/or temporary contracts are all measures adopted in European countries to
reduce the adjustment costs related to employment protection regulation within rms. Other accompanying
measures that may facilitate voluntary quits or negotiated settlements are unemployment insurance and/or
early retirement schemes, redundancy payments, retraining and placement schemes for workers at risk or laid
off.
Except in the Scandinavian countries, employment regulation is usually incomplete in its coverage: employees with less
seniority and young people are usually less protected because of the presence of qualifying period thresholds and of a
system of compensation tied to seniority. Part-time workers are usually less protected against dismissals than are
permanent full-time workers, owing to the existence of minimum thresholds relating to hours worked or earnings. In
addition, all
countries allow for easy individual dismissals in the case of xed-term or temporary contracts. Only Nordic countries
and The Netherlands provide some form of specic protection for part-time and temporary workers.
The self-employed and the small rm sectorrms with fewer than ten or twenty employeesare usually excluded or
exempted from hiring and dismissal protection regulations. Low coverage is a specic characteristic of the highly
regulated Southern European countries, where self-employment, small rms and the underground economy account
for a large percentage of the employed.9 It is indeed the stringency of regulation that explains, at least in part, the large
share of these working patterns, and, in the case of Italy, the high incidence of self-employment and the low incidence
of part-time and xed-term contracts.
Policies facilitating adjustment processes include short-time schemes, early retirement and disability benet schemes, and
ALMP. Temporary dismissals and short-time schemes reduce the potential social costs of mass dismissals, and facilitate
temporary downward labour adjustment. Most OECD countries, including the USA,10 have introduced such measures,
which represent a signicant public subsidy for workers and rms during cyclical downturns. According to estimates
by Mosley and Kruppe (1993: 56), dismissals rates would have been ca. 30 per cent higher in the absence of short run
downward exibility in working time. The use of short-time schemes depends not only on its benets for workers and
employers but also on the stringency of other regulations. Empirical comparative evidence provided by Mosley and
Kruppe (1996) indicates that the use of short-time work to manage structural redundancies is a widespread practice in
Italy, Spain, Belgium and also in Germany since reunication. Firms prefer to use short-time work rather than mass
dismissals because they can select the employees affectedthus retaining human capitaland because they can avoid
the costs associated with redundancy payments and social plans. Compared with unemployment compensation, they
prove more attractive to redundant workers because there are no dismissals, and because benets are higher and of
longer duration. The rigidly regulated Southern European countries provide generous public benets for targeted
casesin terms of sector and rms' sizewhile the Nordic and
9
Recent estimates for Italy, indicate a coverage of employment protection regulation amounting to only 40% of total employment (Ichino 1997).
10
Within Europe, only the UK, Ireland and Portugal do not have such schemes. In the USA, temporary lay-off schemes guarantee access to publicly nanced unemployment
benets. For this reason, the scheme has been accused of encouraging lays-offs at the expense of the public nances.
Continental countries require higher contributions from employers and have less targeted benets.11 Italy provides the
highest level and duration of benets, and recent employers' surveys show that the possibility of using CIGS makes
collective redundancies easier to manage in Italy than in France, Spain and Germany (Fiat 1998).
Active labour market policies and unemployment benets are further instruments with which to alleviate the burden of
employment protection, because they reduce workers' and unions' opposition to lay-offs and social conict. All
European countries display levels of active labour market expenditure much higher than those of the USA or Japan
(Table 2C) and generous unemployment benets (Table 2G). ALMP are largely utilized in Scandinavian countries, in
order to facilitate re-employment and to limit the duration of unemployment spells. The composition of these policies
also differs across European countries: whilst in the UK and Southern Europe they are targeted on problem
groupsyoung people and the long-term unemployedin other countries, and especially Scandinavia, they are
available to all job-losers.
Early retirement schemes and disability benets were widely used throughout Europe in the 1980s to facilitate labour
adjustment and age turnover within rms, even if with great variation. Some of these programmes were entirely
nanced by the state, while others were company schemes subsidised in part out of public funds. A major difference
across countries was the presence, or otherwise, of a recruitment requirement for public subsidies: while the Nordic
countries and the UK were strictest in this respect, France, Belgium and Italy were very loose in their requirements.12
Working Time Regulation

The regulation of working time and of working conditions within rms is governed by collective agreements in most
European countries, while in the UK and Denmark it is left entirely to rm-level or individual agreements. In the UK,
minimum standards relating to working time have usually been dened for health and safety reasons. In Continental
Europe and in Sweden there are statutory limitations on normal working hours and overtime, although the legal
maximum is usually far
11
Employers' costs are lowest in Spain because all costs are assumed by the state and there is no experience rating mechanism. In Italy, rms bear only 8% of the costs of the
cyclical programme (CIGO) and 4.5% of the structural one (CIGS), and they do not have to pay social contributions. On the other hand, such programmes are expensive
for employers in Germany because besides supporting their share of the costs, they must also pay contributions for short-time workers and, according to some collective
agereements, must top up the public benet. (Mosley and Kruppe 1996).
12
For a detailed description and evaluation of such programmes see Casey (1996).
in excess of the actual hours worked under collective agreements. France has the most closely regulated system, with
working time dened by law. Other countries in which legislation plays an important role in regulating working time are
Spain, Belgium, Portugal and Greece. Most legislation and/or collective agreements set quantitative limits on the use
of overtime and on the amount of extra-wage to be paid. Overtime premiums are generally higher in Southern
European and Scandinavian countries (where they average around 75 per cent), and also in the USA (59 per cent), than
in Continental Europe (around 25 per cent).
Some specic working time reduction initiatives have been introduced in Continental Europe in order to promote
employment. Whilst in France and Belgium these initiatives were taken directly by the government through legislative
provisions, in Germany and the UK they have been activated by the trade unions. These measures increase the
regulative stance and are not usually regarded favourably even by workers in the Southern European countries, owing
to their relatively low labour incomes. In fact, in these countries the use of overtime is relatively high, and it provides a
means to support labour incomes.
The UK and The Netherlands have the highest variability in working hours, although even closely regulated countries
such as France and Germany have relatively large percentages of employees who usually work more than 45 hours a
week. This means that the regulation of working time is less stringent than it might appear (Table 2D). Flexible
working practicesespecially night- and shift-workare widespread in the more regulated Southern European
countries.
Wage Determination and Industrial Relations Systems

The characteristics of the industrial relations systems and wage-setting across countries are considered to be important
factors in inuencing wage differentials and the exibility of wages relative to labour market conditions (real wage
exibility). While the literature of the late 1990s does not nd empirical evidence for a precise relation between
economic performance and bargaining systems,13 there are indications of a relationship between bargaining structures
and wage inequalities. More centralized and co-ordinated bargaining systems display fewer wage inequalities than
decentralized/uncoordinated ones (OECD 1997a).
Some specic factors are important in inuencing wage exibility and wage dispersion:
13
The effects of the bargaining structure may depend on the kind of shocks affecting the economic system: co-ordinated bargaining systems may have favourable effects in the
case of aggregate shocks, while in the case of micro-economic shocks, union centralization may be detrimental (Buti et al. 1998).
1. The degree of co-ordination in wage bargaining systems: according to Soskice (1990), co-ordination is
particularly important in guaranteeing the consistency of wage results and economic conditions, and ensures
consensus in bargaining over macro-economic objectives. Co-ordination is not necessarily associated with the
centralization of wage bargaining: decentralized bargaining may in fact be highly co-ordinated, as in Germany,
Denmark and Japan.
2. The level(s) at which bargaining takes place inuences the structure of wage differentials and, in situations of
low ination, is an important determinant of real wage exibility.
3. Union density and the coverage rate of collective bargaining affect wage differentials across industries, regions
and skills. Buti et al. (1998) stress the importance of the presence of legally supported union strength, through
the administrative or mandatory extension of union bargaining coverage, in determining wage rigidity. There
seems to be a strong positive correlation between employment protection rigidity and such excess coverage.
4. The presence of binding minimum wage oors.
According to these indicators, the shared characteristics of most European countrieswith the exception of the UK
and Irelandare high levels of unionization and/or extensive coverage rates, often imposed by legislation, which
imply a strong role for unions and employers' organizations in dening wages. The role of government is also relevant
in so far as it directly intervenes in the bargaining process or plays a facilitating and mediating role. The existence of
multi-level bargaining is another characteristic common to most EU countries, while differences mainly concern the
level of co-ordination and the degree of centralization (Table 2E).
The UK is representative of completely decentralized wage bargaining with little co-ordination and low levels of union
density and bargaining coverage. State intervention in the 1970s was limited to statutory wage controls and minimum
wage oors. At the other extreme, in the 1970s and early 1980s the Scandinavian countries and Austria had high levels
of co-ordination and centralization of wage bargaining, high union density and coverage of collective bargaining.
Centralized wage setting and solidarist polices allowed wage moderation and reduced the risk of an upward wage
spiral, but they also gave rise to a compressed wage structure. Another feature of the Scandinavian model of industrial
relations is the neutral role of government, which mediates labour conicts and provides the accompanying measures
necessary to maintain consensus over the exchange between wage moderation and employment.
Continental and Southern European countries are characterized by lower unionization rates and unions which are
organized sectorally or regionally. Co-ordination is guaranteed by the adoption of common framework conditions, high
mandatory coverage and internal coordination by unions and employers' organizations. The Southern European
countries have multi-level bargaining systems with lower levels of co-ordination than in the Nordic and Continental
countries, and higher industrial conict. Interest organizations in these countries are strong enough to impose their
conditions in the negotiating process, but are not encompassing enough to sustain the social costs of their actions.
France is rather atypical in Europe: collective bargaining has been promoted by legislation, given the low level of union
density. In addition, France is one of the few European countries14 with a legally dened minimum wage; in all other
countries minimum wages are set by collective agreements, usually at the sectoral level.
The weight of the tax wedge on total labour costs is another factor considered important in explaining real wage rigidities in
European countries. High non-wage labour costsemployers' social security contributionsmay act as a disincentive
to hiring workers in the short term, especially low-skilled workers. The high level of the tax and contributive wedge on
labour costs may also explain the low growth of low-skilled labour intensive services in European countriesexcept
the UK. This type of rigidity is accentuated if high non-wage labour costs coexist with minimum wages and non-
competitive labour markets.15
European countries on average place higher tax and contribution wedges on labour (Table 2F), generally related to the
nancing of high expenditure on welfare policies. The highest rates of employers' social security contributions are to be
found in France, Italy and Belgium.16 Very low levels are recorded in Denmark because the nancing of social
protection expenditures is based on general taxation in that country, and in the UK because of the limited role of the
public sector in providing social protection.
Social Security and Unemployment Benets

Generous unemployment benets (UB) of long duration may reduce the search intensity of unemployed workers and
restrict labour mobility,
14
The others are The Netherlands, Spain, Portugal and, since 1998, again the UK.
15
Minimum wage thresholds and highly unionized labour markets do not permit the shifting of payroll taxes onto workers, which could occur in the long run via a reduction
in wages (Daveri-Tabellini 1997).
16
In the Southern European countries, however, these contributions are often reduced by state subsidies for particular regions and/or sectors in order to induce job creation.
thereby attenuating the pressure applied by outsiders for wage moderation.

The negative effects of unemployment benets do not stem from the existence of unemployment benets per se. To a
large extent they derive from the way that they have been designed: negative effects may arise if benets are too high in
comparison with the current wage rate, if they last too long, and if they are not accompanied by job search
requirements and measures to help the unemployed to nd jobs. Any analysis of the unemployment compensation
effects on the labour market must take account of all their institutional features, and of their effects not only on
unemployment, but also on labour market participation. Any analysis of the incentive/disincentive effects of the
benet system should also consider the taxation of benets and earnings, the means-tested nature of unemployment
assistance benets, child-care allowances, benets for one-parent families, and the benets paid to those with low
earnings (Atkinson and Micklewright 1991). The work-incentive/ disincentive effects of benets are closely linked to
the taxation system: if benets are not taxable, unlike income from work, it is possible (as in the case of the UK, for
example) that net income from unemployment benets will be higher than income from low wage part-time jobs.
In many countries, unemployment benets account for only a small proportion of the total cash benets paid to
people of working age. Often more substantial, in terms of both expenditure levels and number of beneciaries, are
non-employment benets paid to enable people to support themselves without workingsuch as invalidity benets,
sickness benets, and early retirement pensions. Recent estimates show that the number of invalidity benets recipients
outnumbered the registered unemployed in 1990 in nine out of fteen EU countries, and that their number has
increased since 1980. Invalidity benets and early retirement schemes remove a substantial part of the working age
population from the labour force, and they help to explain Europe's low participation rates compared with the USA
(OECD 1997b). Usually, these benets have higher replacement rates than unemployment benets, and they induce
people to leave the labour force. In the case of early retirement schemes, this is an explicit objective. Table 2G presents
net replacement ratios of unemployment benets in OECD countries for an average production worker with children,
their coverage and duration. Most of the countries in Europewith the exception of Italyhave generous
unemployment benet systems, with high replacement rates and relatively long durations, which generally reect the
use of different kinds of benets in sequence as the duration of unemployment increases. The differences mainly relate
to coverage and taxation systems.
The Scandinavian countries provide generous benets for all the
unemployed, while job search activity is supported by widespread recourse to ALMP and the provisions of community,
social, and personal services. The aim is to prevent long unemployment durations through the provision of temporary
jobs in the public sector, training placements and employment services. These measures can aid job search and skills
acquisition by the unemployed, but they can also be used to detect abuse in the use of unemployment benets and to
operate a self-selection mechanism whereby only the truly needy take up such opportunities (Jackman 1995). For this
reason, unemployment benets are often made conditional on participation in such schemes after a certain period of
unemployment. Increasing public budget constraints and the risk of displacing regular workers in private and public
sectors (Forslund and Krueger 1994), have however led to a reduction in temporary public placements. In addition,
recent evaluation results show that workers involved in training or temporary placements schemes often develop a
dependency pattern where participation in such schemes alternates with spells of unemployment.
The Southern European countries, by contrast, rely on the substantial protection of permanent workers against
dismissals, and on family solidarity. Generous income support measures are available for dismissed workersin the
form of unemployment benets, early retirement pensions or disability benetswhile the majority of rst-job seekers
are usually not supported and are left to family support and networks. The provision of services is marginal. Besides
cultural and historical factors, a major reason for these differences between Southern Europe and Scandinavian
countries is the capacity of the public sector to act as the provider of employment services in addition to income
transfers. The provision of services requires a different attitude towards users and the possession of different skills and
competencies by public ofcers. Recent research shows that the provision of in-kind services is positively correlated
with employment rates. Both Scandinavian countries and Anglo-Saxon countries present high levels of expenditure in
in-kind services and employment in services. This suggests that an orientation of social expenditure toward services
may offset the potential negative effect of a large welfare state on employment (Buti et al. 1998).
The other European countries lie between these two extremes. The UK presents a horizontal welfare oriented system,
more similar to the Scandinavian approach, where assistance benets are available, as well as the provision of
community, social, and personal services. In Continental European countries, such as Germany, Belgium, France and
The Netherlands, greater emphasis is placed on unemployment benets; while social assistance is left to other targeted
measuressuch as invalidity benets, early retirement, short time working.
Recent Changes in Labour Market Regulation in Eu Countries

The need for changes in the regulatory approach to the labour market and the welfare state has increased in the 1990s,
and the European countries have responded in ways that reect their combination of welfare state and labour market
regulatory regimes. Most countries have operated at the margin, without undergoing a comprehensive reform of their
labour and product markets. Many have encountered difculties in tackling the structural aspects of labour market
regulation, especially those that have set out to make radical changes in the structure of the welfare state.
Radical changes in the regulatory framework can be discerned only in the UK. France, on the other hand, appears to
be the only European country seeking to maintain the traditional model of strong regulation of hiring and ring,
working hours and wage determination, while leaving exibility to temporary contracts. Other countries have
introduced changes which preserve the prevailing regulatory framework while introducing elements of exibility at the
margin. The endeavour is to nd a third way between complete deregulationsuch as in the USAand complete
regulation. The pattern chosen is usually one based on four main dimensions:
(1) the basic system of employment protection is preserved in order to provide protection against unfair
behaviour;
(2) atypical contractsxed-term, temporary and part-time contractsare liberalized and often incentivated,
though with the provision of minimum protection in terms of social security benets;
(3) the regulation of wages and working time is left to collective bargaining, which in turn becomes more
decentralized, hence allowing for greater exibility;
(4) the level and duration of income support to the unemployed is reduced, eligibility conditions tightened and
resources shifted toward active labour market measures in order to enhance the job search intensity.
This approach is considered risky by Bertola and Ichino (1995) and Coe and Snower (1997), because the costs of
increased exibility are borne only by particular groups of the labour supplysee the Italian, Spanish and French cases
in Chapter 4and because the lack of credibility of a lengthy and patched process increases the time required and
raises the transition costs associated with the transition process.
However, the feasibility of radical deregulation in European countries
appears to be limited by the fact that it would require radical changes in their social cultures and practices. Saint-Paul
(1996) explains the low political viability of radical deregulation in Europe by pointing to the low coalition capacity of
the unemployed in most European countries, as opposed to the strong social and political power of the employed.
Examples cited by Saint-Paul are: the French attempts to reduce minimum wages for young workersattempts which
were blocked by social unrest; the Swedish effort to reduce the unemployment replacement ratio which was thwarted
by lost elections; the German government, which was forced by the western unions to raise effective wages in the
eastern Lnder despite large productivity differentials.
The need to maintain social consensus and political support there-fore explains the two-tier or selective deregulation
approach adopted by most European countrieswith the exception of the UK, which, however, is culturally and
socially more similar to the USA than to other European countries. It also explains the acceleration towards changes in
regulation that occurred in many countries in the early 1990s: in times of rising unemployment, incumbent workers are
more exposed to unemployment and more willing to accept regulatory reforms aimed at increasing labour market
exibility.
Common Trends in the Regulatory Framework

The main changes in the regulatory frameworks of the European countries are outlined below and summarized in
Table 2H.
1. Temporary and xed-term contracts have been liberalised either by making changes to legislationSpain,
France, Germany and Italyor through collective bargaining. All countries have also supported self-
employment and the legalization of private employment agencies. Most of the net additional jobs created in the
1990s in the European Union were in fact part-time (71 per cent of net additional jobs for men and 85 per cent
for women). Temporary jobs accounted for the entire increase in employment among men in the EU and for
almost 50 per cent of the rise of employment among women (EC 1996).
2. A trend toward more decentralized wage xing is also evident, albeit within a framework still characterized by a
high level of wage co-ordination at the central level, which has indeed been reinforced in some Southern
European countriesItaly, Portugal, Greeceand in Ireland and Norway. The only European countries to
have embarked on major reform of the bargaining system are the UK and Sweden. In other countries state
regulation continues to play a crucial role in
determining bargaining arrangements. In most countries there is evidence of increasing wage differentials,
although they are still much lower than in the UK and the USA.
3. Increasing exibility in working time and internal practices is one result of the decentralization of bargaining
over working conditions: more exible working hours have been achieved mainly through the annualization of
working hours, while night-work and weekend work have increased throughout Europe. The average length of
the working week for full-time workers actually increased between 1990 and 1995, and the reduction in the
average working week has been largely due to the increase in part-time work.
4. In many countries, measures were adopted in the late 1990s in an attempt to limit the negative and disincentive
effects of social shock absorbers without radically reducing replacement rates. The OECD Jobs Study (OECD
1994, 1997b) summarizes the most important of these measures:
(a) targeted reform of replacement rates, usually reducing the amount of benets for young peopleDenmark
and The Netherlands;
(b) restrictions of excessively easy entitlements to unemployment and related benets. Longer contribution periods
before the receipt of unemployment benets (UB) are now required in Sweden, The Netherlands and Spain.
Early retirement schemes and disability benets have been reduced as a result of budget concerns. Since the
early 1990s, some early retirement provisions have been abolished in Sweden and Germany, while other
countriesThe Netherlands, the UK, and more recently Italyhave also attempted to reduce the growth of
disability benets by imposing stricter eligibility criteria. Expenditure on early retirement schemes for labour
market reasons fell from 0.06 per cent of GDP in 19923 to 0.02 per cent in 19945 in Sweden, from 0.59 per
cent to 0.06 per cent in Germany, and from 0.40 per cent to 0.36 per cent in France (OECD 1997b);
(c) the controls on search requirements have been tightened, and links have been created between UB
administration and active policies. Limits on the duration of passive benets have been matched by alternative
income support provided through entitlement to places on active labour market schemes requiring full-time
attendance in the UK, Denmark, Spain and The Netherlands. Many countriesAustria, The Netherlands,
Sweden and the UKhave given stricter denition to availability for work and imposed tougher sanctions on
those refusing reasonable job offers, or, in Sweden, a place on training programmes;
(d) out-of-work benets have been reduced by lowering replacement rates and by reducing benet entitlements,
with an increase in the period of receipt in The Netherlands and Denmark;
(e) higher returns on being in work have been introduced in order to forestall the poverty trap17 by increasing the
value of child-care benets for those in work, providing tax-cuts for low-wage earners, extending the payment
of in-work benets, using part of the benet as a re-employment bonus, improving access to care services, and
promoting part-time work for older workers eligible for early retirement pensionsUK, Ireland and Belgium;
(f) systems to increase the nancial contribution of employers to short-term schemes and early retirement have
also been introduced in Italy and Germany;
Less has been done in the area of non-wage labour costs: indeed, the tax wedge on labour has increased in Continental and
Southern Europe. Reform has proved difcult, being constrained by the level of social security spending in some
countries and by the low scal base in others due to the growth of the underground economy. Finland, Spain, Portugal,
and Italy have cut social security contributions; while in the UK, Ireland, Belgium, France, and The Netherlands, cuts
have focused on the low paid. France and Portugal have also broadened the contributions base by shifting some
elements of the nancing of the social protection system to value-added tax (VAT)in Portugaland to
contributions levied on other sources of income in order to relieve the burden on labour costs in France. This has
reduced the adverse effects of the nancing system on the employment rate of low-wage earners. Targeted reductions
in social contributions have been introduced in Belgium, Ireland, Portugal and Spain in order to favour young people
and rst-time workers, the long-term unemployed in Portugal and the UK, part-time work in Italy, and the small and
medium-sized enterprise (SME) sector in France.
Differences Across European Countries

While these trends appear to be common to the European countries, the actual steps taken, the approach and the
aspects considered are different, and so too are the changes made to the regulatory framework.
The UK is the country that, during the Thatcher era, moved much
17
The poverty trap arises from the fact that additional earnings from work lead to the withdrawal of benets, higher tax and social security payments. People receiving benet
income are thus discouraged from accepting jobs that do not offer a substantial increase in labour incomesuch as part-time jobs.
closer to deregulation and laissez-faire. This was achieved through extensive statutory interventionsee Deakin and
Reed in Chapter 5which progressively dismantled the unions' bargaining power, employment protection norms and
minimum wages, with the abolition of Wage Councils.18 The welfare system was also substantially revised in order to
reduce public spending and to increase work incentives. The UK system is now characterized by the virtual absence of
legal minimum standards, the replacement of collective employment rights with individual ones, and the
decentralization of bargaining at the plant and individual level. The proportion of employees covered by multi-
employer agreements has accordingly declined noticeablyfrom 70 per cent in 1980 to 57 per cent in 1990 (OECD
1994). Interestingly, exceptions to this trend have been imposed by EC directives requiring the extension of
employment rights to part-timers and maternity rights.
In the Scandinavian countries, due to increasing budget constraints, intervention has sought mainly to tighten and lower
social security benets and to increase employers' contributions to sickness benets. In Denmark, the so-called
activation programme (1996) reduced the duration of UB from seven to ve years and entitlements for young people,
imposed tight controls on UB entitlements, and introduced enforcement rules relating to the acceptance of job offers
and participation on training or educational programmes. Paid leave schemes have also been extended in order to
encourage skills acquisition.19 Sweden has also reduced the replacement rate, and the duration and coverage of UB. In
addition, Sweden has introduced revisions to ease the regulation of employment protection, which, unlike in Denmark,
had been strictly regulated. Changes to economic conditions and technology, on the other hand, account for the
weakening of the corporatist credo in industrial relations. In the 1990s, the wage bargaining system has become less
centralized in Sweden and other Scandinavian countries, while the role of sectoral level bargaining has increased, with
the result that wage differentials have widened. The centralized solidarist model of industrial relations in Sweden
entered crisis in the 1980s, when it became less able to maintain the high level of consensus that previously existed
among the social partners, eroded by slowing growth rates, increasingly unsustainable public decit levels, and growing
international competition. These factors have reduced room for manuvre in the distribution of benets among the
social partners. Moreover, the growing importance of white-collar and public workers has led to the formation of
bargaining arrangements outside the domain of the central
18
Minimum wages have been recently restored by the Blair government.
19
In the long term these schemes may have negative effects, however, because of the reduction in labour supply.
union and the employers' associations, whose co-ordination capacity diminished during the 1980s, so that collective
bargaining oscillated between the central and sectoral level. In 1991, the employers' central associates (SAF) withdrew
from centralized wage bargaining and from all corporatist bodies. The current scenario is one of sectoral bargaining
with covert co-ordination by central associations.
Continental and Southern European countries are characterized by a partial deregulation of atypical contractsespecially
part-time and temporary workand by the tightening and lowering of social security benets. Strong protection is still
provided for full-timers however, especially in France, Italy and Spain, and the deregulation of atypical contracts is only
directed at certain categories of workers exempt from employment protection. Only in more recent years, have some
countries introduced changes in the regulation of dismissals: in 1996, Germany changed the period required for notice
of dismissal and its unfair dismissal legislation, increasing the employer's right to re and hire; in 1994, Spain
introduced a new type of permanent contract with lower severance payment in case of dismissal, while also simplifying
procedures.
Important changes have taken place in wage-setting, and many countrieswith the exception of France and
Spainexhibit a trend towards more decentralized wage-setting within a framework still characterized by a high level
of wage co-ordination at the central level. The precursor was The Netherlands in the early 1980s, and many countries
followed in the 1990sespecially Italy, but also Portugal and Greecewhen central level bargaining assumed an
increasingly important role in setting the general co-ordination framework for sectoral, rm or territorial level
bargaining. In Germany, unication of the country has increased pressures for wages to be set at the rm or plant
levels in the Eastern Lnder, which has resulted in increasing recourse to opening clauses allowing wage reductions in
case of severe difculties.
In Italy, the 1993 tripartite agreement reformed the wage-setting system, converting it into a more closely co-ordinated
two-level framework. At the central level, sectoral associations set contractual wages according to ination targets
dened by the government; at the rm or plant level, wages may be revised in relation to productivity and economic
performance. In order to improve wage exibility, the duration of wage agreements has been reduced from three to
two years. The structural economic and social problems of the South and other areas affected by de-industrialization
problems have also induced the experimentation of territorial pacts as a mode of micro-regulation at the local level.
Regulatory reforms have also reduced or removed built-in clauses of wage nominal indexation to pricesin Italy and
Belgiumand they have increased the opportunities to obtain exemptions from minimum wages.
Minimum wages have been reformed in Spain for certain target groupsthe young and the long-term
unemployedthus maintaining the two-tier approach followed in dismissal regulation. An attempt to introduce a
similar reform failed in France because of the opposition raised by workers. The success of the Spanish reform and the
failure of the French one have been explained by differences in labour market conditions: higher levels of
unemployment and temporary work in Spain increased support for the reform; the less binding level of minimum
wage in Spainbecause only a small proportion of workers are paid at or around such a wage; and the fact that the
Spanish reform was part of a global package which envisaged different forms of exibilityreduced minimum wages
were traded for increased restrictions on temporary contractsand was thus more acceptable to the unions (Saint Paul
1996).
Changes have also been made to the regulation and practice of working time in these countries, mainly through
collective bargaining. In France, agreements have been reached in the metalworking sector on a shorter working week
and a more exible shift system. Recent legislation, moreover, offers employers one-off reductions in social security
contributions in return for a cut in working hours and a matching rise in new hirings. In Germany more exible
working time has been introduced in the metalworking and electrical engineering sectors to secure employment during
downturns. These measures are part of a more general trend towards greater working time exibility.
A successful case in this group of countries is the Dutch experience based on the exchange of wage moderation for
employment (Wassenaar Agreements 1982), a revision of the welfare system which restricted the use of disability
benets and created work incentives, and the large-scale use of part-time work38 per cent of employment in 1996.
The option in the early 1980s for wage moderation has probably been the most important determinant of The
Netherland's current success, given the length of the adjustment process. On the other hand, critics emphasize the
scant additional job creation in the Dutch economy, given that most new jobs are part-time.
The UK and The Netherlands, together with Denmark and Ireland, are the European countries which show the
highest increases in employment rates in the 1990s; they present different regulatory combinations showing that there
is more than one institutional mix which
is favourable to employment growth. The experience of The Netherlands and Denmark, which present a regulative
model closer to European tradition and preferencesbased upon the combination of low employment protection,
high welfare provisions and high coordination/low legal intervention in wage bargainingappears to be as favourable
to employment growth as the UK/USA model based on low employment protection, low welfare provisions and wage
setting decentralization.
Some Preliminary Conclusions

The regulatory system of European countries is a complex web of many interrelated aspects which are difcult to
change radically in a short time span. With the exception of the UK, the approach followed by most countries has been
to introduce selective changes that do not completely overturn the existing social contract. In addition, changes have
been implemented usually through consensual negotiations. In the nineties, increasing pressure from external factors
and the need to improve the adjustment capacity of labour markets have produced an acceleration in regulatory
changes and a convergence of sorts in the steps taken. The resulting regulatory framework which is now emerging
presents differences that are growing less marked.
This convergence pattern is most evident in industrial relations systems and in wage-setting, which is moving towards
co-ordinated decentralization in most countries. Other common features are the increasing exibility of working time
and atypical work contracts, increasing reliance on the activation of labour market policies as measures to facilitate the
management of increased external exibility,20 and reduced income support during unemployment.
However, there are still considerable differences among countries as regards the regulation of hirings and rings and
income support measures, which represent the most sensitive areas of employment relations. The regulative mix of
rigidly regulated employment protection systems, uncoordinated industry bargaining and high mandatory coverage,
low welfare provisionsboth in terms of transfers and
20
Evaluations of ALMP, however, raise doubts concerning the effectiveness of these measures in improving labour market conditions in situations of low growth and high
unemployment (Calmfors 1994; OECD 1993; Schmid 1997), and they stress the importance of policy design and implementation in the achievement of satisfactory
results. The Southern European countries appear to be particularly weak with regard to this latter aspect: the ability of their public sectors to deliver effective employment
services and labour market measures is very low, if not non-existent.
servicesof Continental and Southern European countries appears to be increasingly inefcient in a context of rapidly
changing economic conditions, because it hinders labour mobility and adjustment to changes. In addition its
prerequisite of a stable, wage-increasing employment path for primary workers and family support for secondary
workers is no longer viable: working experiences are more fragmented, while an ageing population and low fertility
rates prevent the nancing of generous pension systems and require increasing participation and employment rates
among secondary workers.
European Monetary Union (EMU) will reinforce these trends and will put greater pressure for labour market reforms.
Convergence in regulatory frameworks will be accelerated by increased competition and factor mobility, and by the
imposition of a tight scal discipline and a common monetary policy at the European level which, in absence of sound
structural policies, may increase the polarization between rich and poor regions within Europe. Given the current large
structural differences among member states and the still low labour mobility,21 both within and among member states,
labour market exibilityand especially wage exibilityis an important factor in cushioning the impact of
asymmetric shocks.22
Complete convergence in social policies will, however, be prevented by differences in national preferences, customs
and histories across member states. In addition, low labour mobility may prevent the rapid convergence of regional
income levels. Specic resources and structural policies together with a strong co-ordination and redistributive role at
the European level are therefore needed to facilitate real convergence and to address high structural unemployment,
without imposing a common harmonized institutional arrangement. As the experience of countries with large internal
territorial differences in economic conditions show, the imposition of a common institutional arrangement in presence
of large differences in productivity and economic development
21
Labour supply adjustments in European countries seem to affect labour participationthrough a discouragement effectrather than labour mobility. The low levels of
intra-EU migration are explained by the progressive rise in wealth and living standards, the spread of unemployment across European countriesdifferentials in
unemployment have been decliningand the development of national welfare states, together with language and administrative constraintstransaction costs, the limited
cross-border transferability of social protection rights and professional diplomas, the inefcient functioning of public employment services, social, family and cultural factors.
EMU is likely to increase labour mobility, due to the removal of some of the administrative obstacles, but in the short run there will be no major changes except in the highly
skilled categories of the labour supply, which have already shown increasing mobility.
22
The European Commission (ECDGII 1997) stresses the need to achieve greater wage exibility in terms of wage responsiveness to macro-economic conditions, since this
would avert the permanent perverse effects that labour and capital mobility might exert in equalizing wages and capital prices across countries with different structural
conditions.
54 Appendix
may have negative effects for the labour market in some areas, and so specic approaches and policies are needed
(Buti et al. 1998).
These concerns explain the increasing attention paid to Europe-wide employment and social policies. To date, the large
differences in social policy regimes across European countries have restricted the denition of a common framework
acceptable to all the member states. EC intervention in the area of social policy has been highly fragmented and
characterized by repeated interruptions and changes of approach. The attempt to establish a trilateral European system
of neo-corporatist political bargaining has also failed (Streeck 1992).
The new approach adopted by the European institutions in the late 1990s appears to be going in the right direction,
even if a stronger co-ordination and redistributive role at the European level is needed. A mutual recognition and
subsidiarity principle has been established, according to which, European institutions formulate a general frame of
reference for social and employment policies which member states are asked to follow. In addition, European level
decision-making inuences the policy-making of member states in the area of labour policy through the use of
European Structural funds. The general framework for social and employment policies was initially set by the Delors
White Paper. The guidelines recently dened following the Amsterdam Treaty support recent developments in labour
market policies inducing member states to strengthen active labour market policies, to support functional exibility
and consultation in decision making, and to reduce the scal and contributive pressure on labour. Little attention,
however, is paid toward reforming wage-setting practices and employment security regulation.
A stronger pressure for convergence in national wage, employment, and social policies will, however, be exerted by
economic forces and the monetary integration process. Spontaneous convergence will be led by multi-national rms,
which adopt common human resources management policies (Brittan 1994), and by capital and labour mobility. This
convergence process will be supported by institutional competition. As we have seen, industrial relations systems are
already showing signs of convergence across Europe. Danthine and Hunt (1994), and from an industrial relations
perspective, Streeck (1992), point out that European integration by itself will reduce the degree of corporatism and
centralization. Increased international integration and the institutional weakness of the European level are going to
reduce the role of national and European centralization in determining economic performance. The Calmfors-Drifl
hump-shaped curve will atten as integration increases and the degree of centralization becomes less marked. The
potential for divergent wage policies will be reduced by higher product market competition and converging prices.
Appendix
Appendix 55
Table 2A. Employment Protection Legislation: Evolution in the 1990s

Countries EPL stringency Actions taken during the 1990s to ease:
ranking OECD Prior authoriza- Severance pay- Dismissals Fixed term/tem-
average ranking tion+notice peri- ments porary contracts
based on four od
studies
(198593)
Scandinavian coun-
tries
Denmark 5
Finland 10 X X
Sweden 13 X X
Continental Europe
Austria 16
Belgium 17 X
France 14 X X
Germany 15 X X X
Netherlands 9 X X
Southern Europe
Greece 18
Italy 21 X X
Portugal 19
Spain 20 X X X
Anglo-Saxon coun-
tries
Ireland 12
UK 7
USA 1
Japan 8
Notes: X = new actions taken.
Sources: OECD (1994b: 74, table 6.7); OECD (1998b: Annex Tables).
56 Appendix
Table 2B. Incidence of Atypical Work and Non-Dependent Employment, Non-Agricultural Sector, 19851996
Countries Atypical forms of employment
% Part-time % Fixed term contracts % Self employment
workers
1985 1996 1985 1996 1985 1996
Scandinavian 19.2 19.7 11.6 13.7 11.1 11.3
countries
Denmark 24.3 21.5 12.3 11.2 9.9 8.3
Finland 8.2 11.6 10.5 17.3 13.4 15.1
Sweden 25.4 24.5 11.9 11.8 9.0 11.7
Continental Eu- 12.3 19.4 7.3 9.1 11.5 12.0
rope
Austria 7.0 14.9 n.a. 8.0 11.4 14.4
Belgium 8.6 14.0 6.9 5.9 15.9 15.4
France 10.9 16.0 4.7 12.6 12.6 11.3
Germany 12.8 16.5 10.0 11.1 9.2 9.6
Netherlands 22.7 38.1 7.5 12.0 9.1 11.2
Southern Europe 5.6 6.6 14.0 15.6 27.5 26.5
Greece 5.3 5.3 21.1 11.0 36.0 33.7
Italy 5.3 6.6 4.8 7.5 24.1 24.8
Portugal 6.0 8.7 14.4 10.6 26.2 26.8
Spain 5.8 8.0 15.6 33.6 22.6 21.5
Anglo-Saxon 13.9 18.1 7.2 8.6 16.5 16.9
countries
Ireland 6.5 11.6 7.3 9.2 21.5 19.8
UK 21.2 24.6 7.0 7.1 11.4 12.6
USA 18.4a 18.3 n.a. 2.2b 7.5 11.5c
Japan 16.2a 21.8 10.3a 10.4b 12.9 7.6c
a
1983.
b
1993.
c
1994.
Sources: ECOGII (1997); OECD (1996: 8, Table 1.6).
Appendix 57
Table 2C. Incidence of Accompanying Labour Policies, 19961997

Countries Expenditure on ac- Expenditure on pro- Actions taken in the 1990s
tive measures (% of visions promoting To increase role To tighten early re-
GDP) (19961997) early retirement for ALMP tirement schemes
labour market rea-
sons (% of GDP)
Scandinavian countries 2.21 0.64
Denmark 2.01 1.39 XX X
Finland 1.67 0.47 XX X
Sweden 2.97 0.05 XX
Continental Europe 1.12 0.30
Austria 0.35 0.13 X X
Belgium 1.25 0.70
France 1.28 0.39 X
Germany 1.36 0.27 X X
Netherlands 1.36 n.a. X X
Southern Europe 0.73 0.12
Greece 0.68 n.a. X
Italy 0.93 b
0.32b X
Portugal 0.71 0.16 X
Spain 0.60 n.a. X
Anglo-Saxon countries 1.11 0.05
Ireland 1.64 0.10a
UK 0.58 n.a. XX
USA 0.21 n.a.
Japan 0.09 n.a. X
a
1991.
b
1992.
Notes: n.a. = Nil or less than 0.005; XX = relevant new actions taken; X = new actions taken.
Source: OECD (1998b), OECD (1996: 20512, Table T); Kalisch et al. (1998: Annex Tables).
58 Appendix
Table 2D. Working Time: Variability of Hours Worked Among Full-Time Employees
Countries Average hours % with usual hours: Dispersion of Actions taken in
usually worked hours worked the 1990s (easing
per week about the mode, of exible work-
% ing time and
part-time work)
below 36 above 45
Scandinavian countries
Denmark 39.1 4 6 43 X
Finland 40 n.a. n.a. n.a. X
Sweden 40 n.a. n.a. n.a.
Continental Europe 39.4 5.5 5.3 30.8
Austria 40 n.a. n.a. n.a. XX
Belgium 38.1 8 3 15
France 39.9 8 8 35 X
Germany 40.3 1 7 23 X
Netherlands 39.1 5 3 50 X
Southern Europe 40.3 9.5 9.3 34.8
Greece 40.0 10 12 36
Italy 38.6 9 8 26 X
Portugal 42.0 14 9 39 X
Spain 40.7 5 8 38 XX
Anglo-Saxon coun- 42.1 12.5 20.5 52.5
tries
Ireland 40.5 16 11 40
UK 43.6 9 30 65 X
XX = relevant new actions taken; X = new actions taken.
Source: Grubb and Wells (1993: 223, tables 7, 8); OECD (1998b).
Appendix 59
Table 2E. Wage-Setting: Main Trends Since the Early 1980s

Coun- Union density and coverage Bargaining level Co-ordination Minimum wages
tries Union density Bargaining Cover- Institu- Change Union Employ- Ratio to Actions
age tional- in pre- ers average taken to
Mid-19- Change Mid-19- Change ized domi- Wages modify
90s (%) 90s (%) bargain- nant mini-
ing lev- bargain- mum
els ing lev- wages
els (b)
Den- 76 stable 69 stable 3, 2 32 high high 0.54
mark
Finland 81 increase 95 stable 1, 2, 3 32 medium high 0.52
Sweden 91 increase 89 increase 1, 2, 3 32 high high 0.52
Continental Europe
Austria 42 decrease 98 0 2, 3 2 high high 0.62
Belgium 54 decrease 90 0 1, 2, 3 2 medium medium 0.60 X
France 9 decrease 95 increase 1, 2, 3 2 medium medium 0.50 raised
Germa- 29 decrease 92 stable 1, 2 2 medium high 0.55
ny
Nether- 26 decrease 81 increase 1, 2, 3 2 medium medium 0.55 X
lands
Southern Europe
Italy 39 decrease 82 decline 1, 2, 3 2 1/ medium medium 0.71
3
Portugal 32 decrease 71 stable 1, 2, 3 2 2/ medium medium 0.45
3
Spain 19 increase 78 increase 1, 2, 3 2/3 medium low 0.32
2
Anglo-Saxon countries
Ireland 50 low low 0.55
UK 34 decrease 47 decline 1, 2 21 low low 0.40
USA 16 decrease 18 decline 1, 2 1 low low 0.39 raised
Japan 1, 2 1 high high
Notes: (a) 1 = Company/plant level; 2 = Sectoral level; 3 = Central level. (b) Direction of changes; X = new actions taken.
Source: OECD (1994: 175, Table 5.1); Buti et al. (1998); Nickell (1997); OECD (1998b).
60 Appendix
Table 2F. Tax and Contributive Wedge, 1985 and 1994

Countries Actions taken to revise taxes Overall tax wedge (% of APW Employer's contributions (%
on labour and social security earnings) of gross earnings)
contributions
Reduction in Targeted re- 1985 1994 1985 1994
payroll and ductions on
income taxes low-paid jobs
on labour in-
come
Scandinavian 61 59 13.3 11.3
countries
Denmark X 66 63 2.9 0
Finland X 54 55 6.1 3.8
Sweden 63 60 31.0 30.1
Continental 57 58 25.7 26.2
Europe
Austria X n.a. n.a. 22.8 23.6
Belgium X 61 61 25.6 34.1
France X 56 59 39.3 46.0
Germany X X 53 59 17.0 19.4
Netherlands 58 55 23.8 7.9
Southern 46 50 29.7 34.1
Europe
Greece n.a. n.a. 21.7 n.a.
Italy X 56 57 45.9 46.1
Portugal X 40 47 20.1 24.5
Spain 43 47 31.2 31.6
Anglo-Saxon 51 49 11.3 11.2
countries
Ireland X 55 55 12.2 12.2
UK X 48 44 10.4 10.2
USA 37 35 7.1 7.7
Japan 26 26 7.3 7.5
Notes: X = new actions taken. APW = Average Production Worker.
Source: OECD (1997b: 667, Tables 24, 25); OECD (1998b); Kalisch et al. (1998: Annex Tables).
Appendix 61
Table 2G. Unemployment Benets

Coun- Net re- Maxi- Cover- Actions taken in the 1990s
tries place- mum age rates
ment ra- dura-
tio 1995a tionb
APW (years) Tighten- Lower Shorter Stronger Tighter Longer Reduce
level ing con- replace- duration work eligibility waiting working
Couple, ditions ment UB availabil- condi- periods disin-
2 chil- for dis- rates UB ity con- tions UB centives
dren ability ditions UB
benets UB
Den- 77 5 84 X X X X X X
mark
Finland 87 1.5 n.a. X X X X X X
Sweden 85 1.2 n.a. X X X X
Continental Europe
Austria 71 1 n.a. X X X
Belgium 60 inde- 82 X X
nite
France 79 5 45 X X X
Germa- 80 0.53.5 75 X X X X
ny
Nether- 82 0.54.5 45 X X
lands
Southern Europe
Greece 27 11 7 X X X
Italy 47 0.53 7 X
Portugal 77 12.5 25 X X
Spain 76 2 32 X
Ireland 64 1.2 69 X X
UK 67 0.5 64 X X X X
USA 59 0.5 n.a. X X X
Notes: a
rst month of benet.
b
The presence of a range means that the maximum duration is determined by workers' characteristics (age, tenure,. . . ).
X = new actions taken.
Source: Kalisch et al. (1998: Annex Tables and Charts); OECD (1998b); Nickell (1997).
62 Appendix
Table 2H. Summary of Action Taken to Change Regulation in the 1990s

Coun- EPL Wage-setting ALMP and unemployment benets
tries Dismiss- Fixed Part-time Bargain- Change Change Measure Tighten- Lower Stronger Intro-
als term/te- work and ing cov- in main in mini- to in- ing con- replace- work duction
mporary working erage bargain- mum crease ditions ment availabili- of work
contracts time ing levels wages role of for dis- rates and ty condi- incen-
ALMP ability shorter tions tives
benets, duration
UI and UI
early re-
tirement
schemes
Scandina-
vian coun-
tries
Denmark X stable 32 X X X
Finland X X X stable 32 X X X X X
Sweden X X Increase 32 X X X X
Continen-
tal Europe
Austria X stable 2 X X
Belgium X stable 2 X X
France X X increase 2 raised X X X
Germany X X X stable 2 X X X
Nether- X X increase 2 X X X X
lands
Southern
Europe
Greece X X X
Italy X X X decline 21/3 X X X X
Portugal X stable 22/3 X X
Spain X X X increase 2/32 x X
Ireland X X
UK X decline 21 X X X X
USA decline 1 raised x X
Notes: see Tables 2A-2G.
Source: Table 2A-2G.
References 63
References
Abraham, K., and Houseman, S. (1994), Does Employment Protection Inhibit Labour Market Flexibility? Lessons
from Germany, France and Belgium, in M. R. Blank (ed.), Social Protection versus Economic Flexibility. Is There a Trade-
off?. Chicago, Ill: University of Chicago Press: 5994.
Addison, J. T., and Hirsch, B. T. (1997), The Economic Effects of Employment Regulation: What are the Limits? (mimeo).
Adnett, N. (1996), European Labour Markets: Analysis and Policy. London: Longman.
Atkinson, A. B., and Micklewright, J. (1991), Unemployment Compensation and Labor Market Transitions: A Critical
Review, Journal of Economic Literature, 24/4: 1679727.
Bertola, G. (1990), Job Security, Employment and Wages, European Economic Review, 34: 85186.
and Ichino, A. (1995), Crossing the river. A comparative perspective of Italian employment dynamics, Economic
Policy, 21: 359420.
Brittan, A. (1994), Labour Markets in Britain and Continental Europe, National Institute Economic Review, 3/94: 57.
Bchtemann, C. F., and Walwei, U. (1996), Employment Security and Dismissal Protection, in Schmid G. et al. (eds.)
International Handbook of Labour Market policy and Evaluation. London: Edward Elgar, 65293.
Buti, M., Pench, L. R., and Sestito, P. (1998), European Unemployment: Contending Theories and Institutional
Complexities, Paper presented at the JEOC Conference Benchmarking Labour Market Performances and
Policies, Rotterdam, July 1998.
Calmfors, L. (1993), Centralisation of wage bargaining and Macroeconomic Perfomance. A survey, OECD Economic
Studies, 21: 15991.
(1994), Active Labour Market Policy and UnemploymentA Framework for the Analysis of Crucial Design
Features, OECD Economic Studies, 22: 747.
Cannari, L., Nucci, F., and Sestito, P. (1997), Mobilit territoriale e costo delle abitazioni: un'analisi empirica per
l'Italia, in Banca d'Italia, Temi di discussione, No. 30.
Casey, B. (1996), Exit Options from the Labour Force, in Schmid G. et al. (eds.) International Handbook of Labour
Market policy and Evaluation. London: Edward Elgar, 379401.
Coe, D., and Snower, D. (1997), Policy Complementarities: The Case for Fundamental Labour Market Reform, IMF
Staff Working Papers, 44/1: 135.
Danthine, J., and Hunt, J. (1994), Wage Bargaining Structure, Employment and Economic Integration, Economic
Journal, 104: 52841.
Daveri, F., and Tabellini, G. (1997), Unemployment, Growth and Taxation in Industrial Countries. London: CEPR
Working paper 1681.
Decressin, J., and Fatas, A. (1995), Regional Labour Market Dynamics in Europe, European Economic Review, 39:
162755.
64 References
Dolado, J., Kramarz, F., Machin, S., Manning, A., Margolis, D., and Teulings, C. (1996), The Economic Impact of
Minimum Wages in Europe, Economic Policy, 23: 31772.
EC (1996), Employment in Europe, 1996. Brussels: European Commission.
EC-DGII (1997), Economic Policy in EMU, Economic Papers, 124 and 125.
EC-DGV Sysdem Network (1996), Labour market exibility overview. (mimeo) Brussels: EC.
EC European Economy (1995), Performance of the Labour Market. Results of an ad hoc labour market survey
covering employers and employees, Reports and Studies, 3.
Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Princeton University Press.
Fiat (1998), Relazioni Industriali. Flessibilit del lavoro. Confronto internazionaler e spunti per la realt italiana, mimeo, Torino,
Aprile 1998.
Forslund, A., and Krueger, A. B. (1994), An Evaluation of the Swedish Active Labour Market Policy: New and
Perceived Wisdom, NBER Working Paper series, 4802.
Freeman, R. B., and Gibbons, R. S. (1995), Getting Together and Breaking Apart: The Decline of Centralized
Collective Bargaining, in R. B. Freeman and L. F. Katz (eds.), Differences and Changes in Wage Structures. Chicago, Ill:
The University of Chicago Press.
Grubb, D., and Wells, W. (1993), Employment Regulation and Patterns of Work in EC Countries, Oecd Economic
Studies, 21: 758.
Heylen, F., and van Poeck, A. (1995), National Labour Market Institutions and the European Economic and
Monetary Integration Process, Journal of Common Market Studies, 33/4: 57395.
Ichino, P. (1997), Il lavoro e il mercato, Mondadori.
Jackman, R. (1995), What can Active Labour Market Policy Do, CEPR Discussion Paper, 226.
Kalisch, D. W., Tetsuya, A., and Buchele, L. A. (1998), Social and Health policies in OECD countries: A survey of
current programmes and recent development, OECD Labour market and social policy occasional papers, 33, Paris:
OECD.
Koedijk, K., and Kremers, J. (1996), sMarket opening, regulation and growth in Europe, Economic Policy, 23: 44367.
Krueger, A. B. (1991), The Evolution of Unjust Dismissal Legislation in the United States, Industrial and Labour
Relations Review, 44/4.
Lazear, E. P. (1990), Job Security Provisions and Employment, Quarterly Journal of Economics, August.
Martin, J. P. (1996), Measures of replacement rates for the purpose of international comparisons: a note., OECD
Economic Studies, 26/1: 99115.
McFarlane, M., and Oxley, H. (1996), Social Transfers, Spending Patterns, Institutional Arrangements and Policy
Responses, OECD Economic Studies, 27/2: 14794.
Mosley, H., and Kruppe, T. (1993), Employment Protection and Labour Force Adjustment. A Comparative
Evaluation, Wissenschaftzentraum-Berlin (WZB) Discussion Paper FS I 929.
References 65
(1996), Employment Stabilization through Short-Time Work, in Schmid G. et al. (eds.) International Handbook of
Labour Market policy and Evaluation. London: Edward Elgar, 594622.
Nickell, S. (1997), Unemployment and Labour Market Rigidities: Europe versus North America, Journal of European
Perspectives, 11/3: 5574.
OECD (Organization for Economic Co-operation and Development) (1993), Employment Outlook (July). Paris: OECD.
(1994), The OECD Jobs Study. Part 2, Paris: OECD.
(1996), OECD Economic Surveys. Denmark. Paris: OECD.
(1997a), OECD Economic Surveys. Sweden. Paris: OECD.
(1997b), The OECD Jobs Strategy. Making work pay. Paris: OECD.
(1998a), Employment Outlook (July). Paris: OECD.
(1998b), The OECD jobs strategy: progress report on implementation of country specic recommendations,
OECD Economic department working papers, 196.
Pissarides, C., and Moghadam, R. (1989), Relative Wage Flexibility in Four Countries, Discussion paper 331, London
School of Economics, Centre for Labour Economics.
Rogowski, R., and Schomann, K. (1996), Legal Regulation and Flexibility of Employment Contracts, in Schmid G. et
al. (eds.) International Handbook of Labour Market Policy and Evaluation. London: Edward Elgar.
Saint-Paul, G. (1996), Exploring the political economy of labour market institutions, Economic Policy, 23: 263315.
Scarpetta, S. (1996), Assessing the role of labour market policies and institutional settings on unemployment: a cross-
country study, OECD Economic Studies, 26/1: 4398.
Schmid, G. (1997), The Dutch Employment Miracle? A comparison of the Employment Systems in the Netherlands
and Germany, Misep Policies, 59: 2331.
, O'Reilly, J., and Schomann, K. (1996), Theory and Methodology of Labour Market Policy and Evaluation: An
Introduction, in Schmid G. et al. (eds.) International Handbook of Labour Market policy and Evaluation. London:
Edward Elgar, 133.
Schomann, K., and Rogoswski-Kruppe, T. (1995), Fixed-Term Contracts and Labour Market Efciency in the
European Union, WZB Discussion paper FS I 95207.
Soskice, D. (1990), Wage Determination: The changing role of institutions in advanced industrialized countries,
Oxford Review of Economic Policy, 6/4: 3661.
Streek, W. (1992), From National Corporatism to Transnational Pluralism: European Interest Politics and the Single
Market, in T. Treu (ed.), Participation in Public Policy Making. Berlin: De Gruyter.
Wallerstein, M., Golden, M., and Lange, P. (1997), Unions, employers' associations, and wagesetting institutions in
Northern and Central Europe, 195092, Industrial and Labour Relations Review, 50/3: 379401.
3 Who Is Harmed by Labour Market Regulations?
Quantitative Evidence 23
Introduction
We have travelled a long distance since the times when egalitarianism and efciency seemed in harmony; when
incomes rose, the welfare state grew, and virtually no one was unemployed. Now, at n de siecle, even die-hard socialists,
from Stockholm to Madrid, bowalbeit reluctantlyto a different gospel: if you want jobs, deregulate!
Deregulate what? Why? There is one stylized fact that underpins just about all analysis, namely that the United States
creates jobs and inequality while Europe maintains equality and creates unemployment. The diagnosis is usually some
variation on the Eurosclerosis theme; too much welfare, wage equality, excessive labour costs, and institutional
rigidity. The deregulatory remedy will, perhaps, breed insecurity and more inequality. The payoff is jobs, more
dynamism and greater mobility. No one denies that exibility creates losers as well as winners, but many insist that a
more dynamic labour market means more mobility and less entrapment. Hence, we will all benet in the end.
That politicians and academics are readily persuaded by the deregulatory message is surprising when we consider that it
is only tenuously anchored to economic theory and, as we shall see, empirical research. It is indisputably true that
European social protection and job regulation is more generous and strict than North Americanor, more generally,
Anglo-Saxonpractice. But do observable facts translate into cause and effect? If so, how do any such causal links
operate? What kind of harm, more precisely, does regulation cause? Does it explain lack of employment growth?
Levels of unemployment? The structure of unemployment? Labour market ows? Does regulation always and
invariably cause harmful effects, or only under certain conditions?
23
I would like to thank Paolo Barbieri, Marino Regini, and Luis Toharia for their useful comments on this chapter.
Who is Harmed by Labour Market Regulations? 67
These are the questions that guide this chapter. There is some evidence that supports the Eurosclerosis thesis; some
that does not. To give a preview of the principal conclusions, the following stand out: One, the evidence is hardly ever
robust: change a measurement here or a model specication there and the correlations either turn signicant or
insignicant. Based on the cumulative evidence, no serious econometrician would dare to venture any strong
conclusion. Two, to the extent that the data do line up behind a coherent story, it is this: a nations' strictness of
employment regulation has very little to do with its levels of unemployment, but probably something to do with the
composition of unemploymentmainly because strong regulations nurture more stable employment relationships and,
therefore, less worker ows. Scaling back welfare benets or deregulating labour markets, may diminish the security of
insiders and augment the chances of the outsiders, and this will help redistribute who is unemployed. But it is difcult
to see how it would, per se, bring about a return to full employment. And, three, if rigidities do matter, they do so
interactivelywhen they combine with other, possibly more powerful, impulses such as massive de-industrialization or
the nature of industrial relations. On balance, the Eurosclerosis thesis enjoys very weak support or, to quote Nickell
and Layard (1998), . . . Time spent worrying about strict labour market regulations, employment protection and
minimum wages is probably time largely wasted.
Comparative Labour Market Performance

The popular Europe-America comparison is alluring but awed. It hardly makes sense to speak of a European
average. As Nickell (1997) notes, close to one-third of Europe's population live in an environment with less
unemployment than in America.24 In fact, the number of countries with truly alarming levels of structural
unemployment and associated illssuch as a prevalence of long-term and youth unemploymentis quite small. That
there is a certain clustering comes out in the overview in Table 3.1. One extreme group combines low employment
rates and high unemployment, especially among youthItaly, France and Spain. Another extreme boasts high
participation levels and modest unemploymentDenmark, Norway, the UK, and Austria. Some nations, like Sweden,
exhibit high overall unemployment, but with no
24
It is for this very reason that the individual country case-studies in this volume are of major importance. If the prole of unemployment differs among nations, it is risky to
contemplate policy in terms of general all-embracing strategies. The causes of overall unemployment and of who, precisely, is unemployed may differ and, if this is so, we
clearly need to think in terms of nation-specic and not Europe-wide remedies.
68 Gsta Esping-Andersen
Table 3.1. Divergent Patterns of Employment and Unemployment in Europe and the USA, 1980s1990s
Employment popu- Total unemployment Youth unemploy- Low-skill unemploy-
lation ratio ment ment
Low Belgium Austria Austria Austria
France Netherlands Denmark Italy
Ireland Portugal Germany Netherlands
Italy Norway Norway
Spain USA Portugal
Sweden
Medium Austria Germany Ireland Belgium
Finland Denmark Netherlands Denmark
Germany UK Norway France
Netherland Portugal Germany
Portugal Sweden UK
UK
USA
High Denmark Belgium Belgium Finland
Norway France Finland Ireland
Sweden Italy France Spain
UK Sweden Italy USA
USA Ireland Spain
OECD
Mean 66.5 7.1 13.2 11.5
Notes: The Low/Medium/High groupings are based on the mean OECD unemployment rates for 1996, except for low-skilled workers
which refers to 1994. Low-skilled is here dened as less than secondary school education.
Source: OECD (1997).
particular biassuch as among youth or the low-skilled. Others, such as the Netherlands, have modest
unemployment, but also low employment levels. If the USA is the benchmark of a positive employment scenario, it is
rivaled by Austria, Norway, Portugal, the Netherlands and, possibly, Denmark.
Some supposedly regulated economies do well and, vice-versa, some supposedly deregulated economies do less well.
The United States is, in fact, the sole example of a truly convincing labour market performance. The UK, after two
decades of deregulation, has brought down unemployment levels, but not long-term, youth, and low-skilled
unemployment. This is basically also the case in Canada, Australia, and New Zealand. We should, in brief, abandon any
kind of analysis or policy advocacy that bases itself on the simple USA-Europe dichotomy.
Europe's unemployment appears predominantly structural; it hardly declinesand employment hardly growsdur-
ing economic upswings. The NAWRU has typically risen in Europe and declined in America since the early 1980s
(Elmeskov et al. 1998; OECD 1997).25 This inuences long-term unemployment which, in Europe, has risen from
about one-third of all unemployment in 1980 to almost half in the 1990saffecting mainly young, rst-time job seekers.
Yet again intra-European differences are notable. Indeed, the American NAWRU is easily matched by Portugal,
Switzerland, Austria, Norway, and the Netherlands.
In order to capture Eurosclerosis, we should also examine trends in job growth and unemployment across business
cycles, as well as labour force ows. The main argument for why rigidities cause structural unemployment is that they
impair lay-offs during slumps and new hirings during upswings. This, in turn, implies low turnover rates, especially
between unemployment and jobs, and the net result should be a rigidication of the outsiders. Trend data on job
growth and unemployment rates over recent cycles are shown in Appendix Table 3A; Appendix Table 3B provides
worker ow data; and Appendix Table 3C presents ows into and out of unemployment.
These data conrm Europe's generally inferior job growth. Trends in unemployment between 1983 and 1993 capture
two troughs. Here the picture is more diversied. Finland and Sweden have experienced a sharp jump in the 1990s, in
both cases for reasons that cannot be generalized; Italy and Spain also perform poorly; and only The Netherlands
shows major improvements. Changes in unemployment during 19936 show the most recent recovery. In France,
Italy, and Germany, unemployment actually worsens during the recovery; and in three countries there is some
improvement (particularly in Denmark).
The pattern of labour force ows, rst measured through average job tenure, is quite bi-modal: low tenure in the USA,
but also in Denmark and Spain; very long tenure in Germany, Italy, France, and Sweden. More surprising, at rst
glance, is the similarity of job-to-job ows between the USA and supposedly rigid Italy and France. Deregulated Britain
and highly regulated Germany appear equally immobile. Yet,
25
Again, we must be careful not to generalize. Denmark, the Netherlands, and the UK have all enjoyed a decline in structural unemployment during the 1990s. NAWRU, the
successor to the erstwhile NAIRU as a measure of the structural component of unemployment, stands for non-accelerating wage rate of unemployment. For data and a
discussion, see Economic Outlook, 62 (OECD 1997). Between 1990 and 1997, the NAWRU actually declined in Ireland, the Netherlands, the UK, Denmark, Australia,
Canada, and the USA. It rose sharply in Finland and Sweden, not because of a sudden burst of regulatory reform but for basically nationspecic reasons.
this is not as surprising as it seems because employment regulations should principally affect ows from
unemployment to jobs, not job-to-job shifts. It is very possible that protected workers change jobs often, either
because of promotions or job redenitions. Rates of employer change are therefore a more sensitive measure, and here it
is clear that most European countries, except Denmark, are much more immobile than North America.
Lay-off rates, especially during troughs, are key to the worker protection issue. Since many countries have introduced
exibilization reforms over the 1980smany related to dismissal rules or to temporary workwe would anticipate a
rise in lay-off rates in the 1990s trough. In terms of overall volumes of dismissals, Denmark and Spain match exible
America. Surprisingly, the Netherlands and the UK score very low. As to trends, there is some evidence that
exibilization has borne fruit since dismissal rates have generally risen across the cycles.
Low rates of exit from unemployment imply, all other things being equal, that long-term unemployment will prevail.
Again, there is hardly any European convergence on this measure. In Scandinavia, exit rates are not far below USA
levels, but the rest of Europe performs generally poorly, Spain in particular.
In brief, our review of labour market indicators afrms the need to treat Europe as a differentiated group of nations.
Such differences in unemployment rates, trends, and ows may simply mirror variations in regulatory severity, but, as
we shall see, this does not hold up. It would, accordingly, seem that the trade-off between equality and jobs is less
stark and straightforward than a Eurosclerosis type argument would suggest. Denmark and The Netherlands match
the United States on most indicators, yet without more inequality or weakened social safety nets. Alternatively the UK's
sharp rise in earnings inequalities, with notable reductions in social protection, has not improved much upon long-term
unemployment or labour market ows.
The Links Between Employment Performance and Rigidities

Popular notions such as Eurosclerosis and rigidities are not particularly helpful in serious analysis because they
combine, rather vaguely, a set of discrete regulatory practices which may, or may not, have similar effects. When we
speak of employment-relevant regulations we usually distinguish three basic forms. The rst has to do with income
guarantees provided by the welfare state, such as unemployment benets; in short, the reservation wage. The second
refers to wage-setting practices and, in particular, to the wage oor and earnings differentials. The third
addresses regulation proper, that is, limitations to employer freedom to hire and re at will.26
As discussed by Regini in Chapter 1, labour market deregulation may come at a price. If it helps reduce
unemployment, it may also impair employee performance. Regulatory frameworks are not just valued by workers, but
also by rms because they stimulate commitment, trust, co-operation, and long-term horizons in the employment
nexus. The pay-off is more investment in training, internal exibility, and enhanced productivity. Secure workers are
less likely to resist technological innovation and work-redesign. How these contradictory objectives may nd resolution
is hard to say, but we would certainly expect that many rms which operate in a deregulated environment will initiate
in-house regulatory environments (Booth 1997; Gregg and Manning 1997). In other words, deregulation will breed its
oppositeas efciency-wage theory argues.
Formalized regulatory environments, alternately, will invariably provoke informal exibilities. The problem is that we
can measure legislated and bargained regulations, while adaptive practice is, by denition, known only to its
practitioners. It follows that comparative rankings of formal strictness are almost certainly unrepresentative of reality.
Hence national differences, at least as far as the core workforce is concerned, may be trivial. More generally, the link
between employment and regulations is difcult to identify precisely due to four sets of problems:
1. Regulations form a more or less coherent package, that in turn is upheld by a broader institutional infrastructure.
Many studies recognize this, but are nonetheless forced into partial analyses of one or another regulatory practice
be it welfare state effects or the impact of wage minima. There are very few studies which systematically examine
the unemployment consequences of all regulatory practices and their institutional setting as an integrated
phenomenon, identifying the relative contribution of one or the other factor (Alogoskous et al. 1995; Scarpetta
1996; and Nickell and Layard 1998, are some exceptions). An even more troubling problem along these lines is
that regulations may not have much of an effect per se, but if they coincide with certain conditions, or shocks, their
effect may become substantial. This is essentially the argument stressed by Blanchard (1998).27
26
Product market regulation may have substantial inuence on labour markets but is typically excluded from the kinds of studies examined here.
27
This may also help explain the oft-noted puzzle of why regulation is harmful today when, apparently, it was not in the 1960s and 1970s: regulatory practice may be the same
or even scaled down, but a sudden exogenous shocklike the OPEC oil criseswill render it harmful. There are also major statistical reasons why it is close to impossible
to estimate the impact of all variables simultaneously. We usually have very few observations to work with; and have serious problems of multi-co-linearity because many of
the relevant explanatory variables are highly correlated among themselves. It is for this reason that the statistical estimations presented in this chapter are partial, examining selectively the
impact of regulatory variables.
2. A second characteristic of contemporary research is that different studies explain different phenomena: some
are mainly preoccupied with employment or unemployment levels, some with trends, and some with
compositionwho is most likely to be unemployed. A few (Scarpetta 1996; Nickell and Layard 1998) examine
both trends and structure.
3. Research is severely hampered by the lack of appropriate data. The lack of data on informal practice is
matched by the unavailability of over-time information on regulation. In reality, we need long time-series
because many nations have been actively deregulating over the past decade, and because it is not certain that
reforms will produce immediate results. Research is therefore constrained to examine the effects of labour
market regulations on a cross-section of only up to twenty nations.
4. A fourth fundamental problem is that it is difcult to interpret the effects of isolated, distinct regulations. Even
if we nd signicant correlations, they may very well be spurious. As is frequently recognized (Bertola and
Ichino 1995a; Freeman 1994; Alogoskous et al. 1995; Scarpetta 1996; Siebert 1997), any given regulation of
wages or jobs occurs on a backdrop that is unmeasured. Thus we may identify a rigidity, but it may have only
marginal consequences if it is off-set by institutional arrangements or countervailing exibilities. Worse, the
impact of regulations may be spurious if it is hidden structural forces which really account for national
difference in employment performance.
Moreover, the impact of regulations is probably much less monotonic and straightforwardmore exibility equals less
unemploymentthan most studies assume. To give an example, high wage costs or egalitarian wage distributions can
produce both high and low unemployment. The correlation between unemployment and regulation may be non-
monotonic (or even spurious) because there may exist some underlying third force that drives them both.
All this goes to say that: a) the real employment impact of regulation is not crystal clear; b) some forms of deregulation
may, due to the surrounding institutional environment, actually have effects that run counter to the aim of exibilizing
and job stimulation; and c) policy-makers are mistaken if they assume that an extra dose of deregulation will
automatically translate into an extra dose of employment. They are equally mistaken if they assume that a given
deregulatory practice that works in one country can be made to work similarly in another. As several of the country-
case analyses in this book demonstrate, deregulatory policies may, paradoxically, have the perverse effect of
strengthening
other rigidities. It is difcult to know with any certainty what needs to be reformed or deregulated. If the underlying
cause of unemployment is to be found in bargaining mechanisms, it would clearly be more effective to reform
industrial relations than to weaken, say, unemployment benets or dismissal rules. Or, as we shall discover, where the
unemployment problem is mainly driven by an unfortunate skillstructure it would appear more relevant to discuss
education policy.
The organizing principle that I choose here is to divide the discussion according to the three main regulatory
mechanisms: the welfare state; wage-setting mechanisms; and employment protection. The dependent variable that
concerns us most is unemployment in general, but I shall give special attention to its structural manifestations: duration
and population incidence.
The Welfare State and Unemployment

The claim that generous social protection creates unemployment is as old as social policy itself. During the
conservative 1920s, the USA government sent a team of experts to Europe to study evolving social security. Their
report spoke alarmingly of rampant pension hysteria and induced work-evasion. The past decade has seen a repeat of
such views both in the general debate and in academia (Lindbeck 1994). On the empirical level, the impact of social
protection on unemployment concentrates on two issues: work incentives and wage behaviourwhich, in turn, affect
the demand for labour.
Work Incentives
There is a rich, predominantly American, literature on negative work-incentives. The basic point is that generous
income guarantees reduce incentives to take employment. Both the level and duration of unemployment may therefore
be greater than necessary. The key programme is, of course, unemployment insurance, but in principle any kind of
income guarantee may affect labour supply. The key issues are net, after-tax, income replacement rates and benet
duration.28 American research has found that unemployment benetsand also Aid to Families with Dependent
Children (AFDC)prolong employment re-entry, especially among the low-paid and single mothersfor an
overview, see Brown et al. (1982), Moftt (1992), Pedersen and Westergaard-Nielsen
28
Unfortunately, most existing studies use gross measures of replacement rates.
(1993), and Bean (1994). In a broader, comparative context, Layard et al. (1991), OECD (1993; 1994), Scarpetta
(1996), and Nickell (1997) provide some evidence that replacement ratesand especially duration of benetspush
up unemployment.29 There is, for example, an often strong bunching of unemployment exits at, or near, the moment
that benet entitlements are exhausted. Such effects may be especially strong among low-wage workers due to a
smaller net difference between benets and expected earnings. The consequence, then, is reduced ows out of
unemployment and, hence, unemployment statistics, especially those for long-term unemployment, which are higher
than they otherwise would be.30 In Scarpetta's study, unemployment benet levels have an independent, signicant
effect on both overall and youth unemployment, even when controlling for degree of employment regulation, active
labour market policies, bargaining structure and macro-economic variables. But, strangely enough, he does not nd
any consistent effect of benet generosity on long-term unemployment.31 Scarpetta also nds that benet generosity is
positively related to rising structural unemployment in the 1980s. High benets with long duration account for three to
ve percentage points of the unemployment differential for Denmark, Belgium, The Netherlands and France, and a
rise in benet generosity is said to account for up to a three percentage point increase in structural unemployment in
Spain.32
There are of course as many sceptics as proponents of this relationship. A simple comparison shows that American
(and generally Anglo-Saxon-nation) unemployment benets are more modest than in Europe, cover fewer workers and
are of more limited duration. To take two extremes, normative duration in the USA is twenty-six weeks; in Belgium or
Denmark, it is, in principle, indenite. Might this then account for the superior Anglo-Saxon unemployment
performance? Alternative simple comparisons would cause doubt. The average level of American unemployment
compensation is not dramatically lower than the European average at 4050 per cent of mean wage; and in some
29
Similar results are found in Nickell and Layard (1998).
30
A recent Danish study estimates that 10% of the unemployed did not even desire a job, and that 40% had not been searching actively over the past month. It also estimates
that the negative work-incentive is quite widespread in that about 22% of the unemployed would notor hardly at allimprove their income status by working (Pedersen
and Smith 1997).
31
The strong, positive effects that Scarpetta nds for youth unemployment may be sensitive to measurement problems. He uses an average of benet replacement rates for
different categories of workerssingles, couples, and familieswhich may not reect the reality of unemployed youth in most of Europe. Another oddity of his ndings is
that unemployment benets in Italy have a positive effect on unemployment. This is difcult to reconcile with the fact that the lion's share of Italian unemployed receive
absolutely no benets.
32
This effect should disappear subsequently because, as Toharia and Malo show in Chapter 11 on Spain, benets were cut back.
European nations, Italy par excellence, the largest group of unemployedrst-time job seekerscannot even claim
benets.33 As several studies note, a more dynamic view leads also to some doubt: the level of unemployment benets,
eligibility, and duration has, in most of Europe, been stagnant or even declining since the early 1970s when
unemployment was low. If, then, there is any causal relationship between welfare guarantees and unemployment, it may
be because the same programme that once did not induce perverse effects does so now because something else has
changed (Siebert 1997; Alogoskous et al. 1995).34
Several comparative studies conclude that the impact of social protection on unemployment is, at best, marginal. The
Northern European studies in Atkinson and Mogensen (1993), the broader correlational study by Burtless (1992), the
country-case analyses in Blank (1994), and also Atkinson and Micklewright (1991), all conclude that the negative
incentive effect is unimportant except at the margin. And this margin may even be ambiguous: even if workers remain
on unemployment benets longer than necessary, this may be positive if it improves the job-matching process.
How do we reconcile these apparently contradictory views? Besides the possibility that contrasting results derive from
differences in data and methodology, one explanation is that unemployment insurance also induces people to take
employment so as to qualify for benets in the rst place (Holmlund 1997). A second explanation may lie in interactive
effects: welfare benets may, as such, not induce unneccesary unemployment but when they coincide with other factors
this may happen. For example, generous unemployment benets might encourage insider-bargaining: core workers feel
free to push up real wages in the knowledge that the outsidersthe unemployedare taken care of (Siebert 1997: 15).
If this happens, the long-run effect is stronger segmentation and more persistent unemploymentespecially among
youth.
Another kind of interactive effect has to do with the macro-economy. Where, as in many European countries, de-
industrialization has been exceedingly heavy, and especially where this has coincided with continuous exit from
agriculture, there is the particular problem of managing
33
Even if American unemployment benets are more or less in step with the European average, eligibility for benets is much lower in the United States. Myles (1996)
suggests that only about 30% of unemployed American workers in the late 1980s were de facto eligible.
34
There are a few cases where unemployment benet systems have undergone signicant changes. Scarpetta (1996) links Spain's rising structural unemployment levels to
improvements in benet generosity; Layard et al. (1991) suggest that the fall in British unemployment is related to the introduction of stricter eligibility rules. Denmark's
reforms of unemployment insurance in the early to mid 1990s are of special interest since they sharply reduced entitlement duration for youth especiallysee Chapter 6 on
Denmark in this volume.
a mass of redundant and difcult-to-absorb workers (Jimeno and Toharia 1994; Glyn 1995; Marimon 1996; Esping-
Andersen 1999). These largely unskilled masses could conceivably nd re-employment in low-wage jobs, as happens in
the USA, but this avenue is foreclosed in Europe because of wage compression and high labour costs. So, if they do
not qualify for early retirement, one strategy is to park them on long-term unemployment benetsas in Spain
(Toharia 1997). In these scenarios, social benets are perhaps the proximate, but not underlying explanation of mass
unemployment.
Yet another interactive, or offsetting, effect might come from active labour market policies that, instead of passively
maintaining the unemployed, induce training or job re-insertion. In this way, the unemployment rolls are reduced
especially if, as now in Denmark, acceptance of training or a job is mandatory. Layard et al. (1991) nd that active
policies reduce unemployment somewhat while Scarpetta (1996) and Nickell (1997), controlling for the unemployment
benet variable, nd no substantial reduction of unemployment due to active labour market programs.35 If we try to
adjust for the fact that participants in active programs are articially removed from unemployment statistics, it is
generally quite doubtful whether active policies help to reduce unemployment effectively. Evidence from Denmark
suggests that some of these participants end up in a carousel, rotating between unemployment, training, and job
creation programmes (Pedersen 1997). Active labour market programmes may indeed be fair-weather policies: they
work best when least needed.
The evidence on the impact of social benets is inconsistent, with the partial exception of the impact of benet
duration.36 If, as in Denmark and Belgium, duration is, or was, effectively unlimited, some unemployment is likely to be
very prolonged. The Danish case outlined in Chapter 6 suggests, in fact, that a simple reform of duration may have
positive effects on unemployment exits. At the other extreme, we must keep in mind that benet systems of very short
duration can have
35
The Layard et al. study has been critized for its classication of countries in terms of unemployment inows and duration (Jensen 1997: 44). More importantly, the Layard
study refers to the 1980s when active policies were, in fact, positively correlated with low unemployment. The problem is that this effect disappears in the 1990s. Why? There
are two vying hypotheses: one, traditionally high unemployment countries, like France, have augmented their active programmes dramaticallyhence there is less variation
in expenditures on active programmes; two, active programmes perform well during economic upswings like the late 1980s, but less so during recessions like the early 1990s.
Regressions by the author, not shown, suggest that the latter case is closer to the truth.
36
Bi-variate regressions of unemployment benet levels on unemployment are systematically insignicant except for two marginal cases: the relative probability of
unemployment of women compared to males (B = 0.045 (t = 2.17) and a very low R-squared (0.053); and the relative duration of youth unemployment compared to adult
males (B = 0.01 (t = 2.33) with R-squared = 0.31) (author's own estimate's).
adverse effects. A more leisurely process of job search should improve the matching process and thus productivity.
Labour Demand
Welfare benets inuence labour demand in two ways. They implicitly dene the minimum wage-oor and hence, if
overly generous, may indirectly price some workers out of the market. They also necessitate high taxation which,
directly or indirectly, inuences the cost of labour. This can be problematic since xed labour costs in many countries
have risen steeply due to slower real wage growth, and the worsening ratio of active workers to welfare dependents;
and also because we are now paying the bill for the upgrading of social benetsespecially pensionsthat occurred
in the 1960s and 1970s.37 In France and Italy, mandatory social levies account for close to 50 per cent of the total wage
bill. This means a high tax wedge for rms, but also for workers, who may discover that the net return to a wage
increase, to education, or to moving from unemployment to a job, is trivial.38
If social benets push up the wage oor, this should affect employer demand, biasing it towards adjusting hours rather
than the volume of personnel. Indeed, the OECD Jobs Study (1994) estimates that there is a one-to-one elasticity
between labour costs and private sector employment, that is for every one per cent rise in costs we will have one per
cent less jobs. The important point here is that the effect is not across-the-board but mainly concentrated among the
less skilled and youth: the effect is basically similar to the wage-effects that will be discussed below (Alogoskous et al.
1995). Indeed, there is much evidence in favour of this argument: for example, demand for young workers rose
substantially in Italy following the introduction in the mid-1980s of temporary apprenticeship contracts exempt from
full social contribution requirements (Adam and Canziani 1998).39
37
We can illustrate the rising nancial squeeze in terms of the average worker's career cycle. In the post-war decades, the standard worker would begin working around age 16
and remain employed until age 65, followed by an average 67 years' retirement. Today in many countries the age of rst job is delayed into the mid-20s while retirement
begins around age 5558. Male longevity has risen by an average 8 years in the meantime. The upshot is that we have cut the contributing years by 15, maybe 20, years while
extending the passive recipient years by at least the same amount.
38
This would imply that progressive taxation and high marginal tax rates should inuence labour supply and investment in human capital negatively. The evidence in favour of
this view is, however, rather mixedmainly because it has been shown that strongly progressive taxes have a dampening effect on wage growth (European Commission
1994; Pissarides 1995). Similar results have, incidentally, been found for Italy (Malcomson and Sartor 1987), for Denmark (Haagen Pedersen 1997), and for Britain
(Lockwood and Manning 1993).
39
What we do not know is whether the same young workers would have been hired in any case. However, survey data on Italian employers show that their propensity to hire
new workers would increase with a reduction of labour costs more than any other measure (ISTAT 1997: Table 3.3).
Once again, many are sceptical about the importance of the tax-wedge or reservation wage. If it mainly affects
unskilled workers and youth, how do we explain that unskilled worker unemployment is comparably much higher in
the USA and the UK, two countries with modest welfare benets and tax-wedge, than in Italy, where labour costs are
heavy.40
To explain such anomalies, we need to consider offsetting effects. European countries' recourse to early retirement is
one such offsetting mechanism: if workers are induced to leave the labour force altogether, they obviously do not
appear in unemployment statistics. Another offsetting mechanism is atypical work arrangements, such as informal
employment or self-employment, that exempt employers from social contributions. Self-employment is particularly
diffused and dynamic in countries with very high labour costs, like France and Italy, but oddly enough also in Britain
(OECD 1994).
There is also the possibility that the low-skilled are excessively unemployed because, simply, they are in excessive
supply.41 A large low-wage labour market might improve their employment prospects, but probably not their life
chances. If the underlying problem is an un-trainable mass, a low-wage economy may harden and perpetuate the real
problem which is skill formation (Alogoskous et al. 1995; OECD 1996). This kind of deregulatory practice may have
devastating effects on the incentive structure for training, and the net result could be a low-skill equilibrium (Soskice
1990; Snower 1997).
In Table 3.2, I present a series of regressions on both the directnet unemployment replacement rates and
durationand indirectthe tax wedgeimpact of social benets. The results support the sceptical view that, if
welfare benets matter at all, it is at the margin: the coefcients for benet levels are extremely low and signicant only
in the case of low-skilled unemployment; the effect of benet duration is clearly stronger, and especially powerful in
reducing outows from unemployment. Here, then, we nd conrmation of previous ndings. The indirect impact of
generous welfare benets, that is via a high tax wedge, seems
40
The least educated workers (elementary education only) are more than twice as likely to be unemployed in the USA and about 60% more likely in Great Britain and Canada,
while their rate of unemployment is about average in Italy, Portugal, and Spain (OECD 1994).
41
The skill-structure of a society cannot be measured reliably through standard educational attainment statistics. As the recent OECD (1996) Literacy and the Economy study
shows, about 15% of native-born, high school graduates in America do not even possess the minimum level of cognitive skills necessary to be trainable, a gure that is
more than twice the level in Germany or Sweden.
Table 3.2. Unemployment and Social Benets

Dependent Vari- Log (unemployment) 1996 Youth/Adult un- Low-skilled/all Unemployment
able (% monthly) I II employm. unemployment (1994) Outows
(ratio %) (ratio %)
Constant 0.972 0.931 1.803 1.635 30.025
(2.37)a (2.62) (1.22) (26.14) (3.62)
Benet 0.138 0.114 -0.020 0.038 -3.898
Duration (1.77) (1.68) (-0.07) (3.18) (-2.46)
Net Unem. -0.003 0.004 0.012 0.006 0.077
Benetb (-0.35) (0.60) (1.33) (5.30) (0.51)
Tax wedge 0.023 0.029 0.020 -0.013 -0.260
(2.24) (3.08) (0.43) (-8.04) (-1.24)
Bargaining -0.046
Centralization (-2.05)
Adj. R2 0.333 0.479 0.292 0.106 0.317
Notes: a
T-statistics are marked in parentheses.
b
as a percent of average production worker wage.
Source: OECD. All estimates based on robust regressions. N = 20.
more consistent but not especially powerful. To take into account the possible mediating effect of bargaining structure,
all models were re-estimated with a centralization variable. Centralization helps reduce unemployment levels, but it has
no bearing on the structure of unemploymentthe youth and low-skilled unemployment bias.42
Hence, as far as the welfare state effect is concerned, the main problem appears to lie in duration of benet eligibility
and not in generosity.43 It is hardly surprising that unemployment benets and duration do not inuence the youth bias
of unemploymentmost rst-time job seekers can claim little or anything from their welfare state. Neither is it
surprising that the welfare state variables are much more signicant for the low-skilled bias of unemploymentthey
are more likely to suffer from lay-offs and are less likely to be re-hired. Unemployment benets may, as in Germany, be
their pathway to retirement. Otherwise, if welfare state generosity is harmful, it is harmful indirectly, via a higher
42
I therefore omit the centralization variable for the estimations of youth and low-skilled unemployment. Unemployment level is logged due to the very skewed distribution of
international unemployment rates, quite similar to income distributions. In the case of low-skilled unemployment, the model was re-estimated with an early retirement
variable (% males, aged 5564 non-active), but it has no effect whatsoever on the results.
43
A confrontation with earlier studies cannot be done directly since I here use net benet replacement rates, while most studies use gross. For work-incentives, it is clearly net
benet that matters.
tax-wedge. In this model, a ten percentage point jump in the tax wedge implies a 23 percentage point rise in overall
unemployment.
Wage Structure and the Minimum Wage

Many, like Wood (1994), believe that well-paid jobs for the unskilled have disappeared because of international trade.
An American or European low-skilled worker competes in an increasingly global labour market. Even if, as many
believe, the real culprit is technological change, the effect is similar. In any case, a wage structure may cause
unemployment independently of exogenous forces if the distribution of earnings does not reect productivity
differentials; if real wages rise faster than productivity; or if wage rates do not adjust to rising unemployment levels
(hysteresis). For an overview, see Blanchard et al. (1987; 1998), OECD (1994), and Dolado et al. (1996). High labour
costs and the tax wedge, as we have discussed, should mainly affect youth and low-skilled unemployment, because a
high wage oor and compressed wages price out less desirable workers, and may even reduce overall labour demand
(Brown 1982; OECD 1994 and 1996; Dolado et al. 1996).
There is now a huge amount of research on the impact of minimum wages on unemployment. One problem of
interpretation stems from opposing theoretical starting points. In the standard competitive model, high minima should
produce unemployment; in the monopsony model, however, the effect can be opposite (Dolado et al. 1996). We may
therefore arrive at very different empirical estimates depending on how the problem is modelled.
There are some facts that all can agree upon: In Europe and the UK, real wages have continued to rise despite
unemployment; in the USA, they have been stagnant and the minimum wage declined throughout the 1980s. The
minimum wage for both adults and youth (whether legislated or de facto) is considerably higher in most European
economiesaround 5060 per cent of average wages among adults and 4050 per cent among youththan in the
USA or UK where it is around 40 per cent for both youth and adults.44 The share of workers employed in low-wage
jobsless than 67 per cent of the median wageis correspondingly much higher in the USA (25 per cent) and UK
(20 per cent) than in Europe, which ranges from 5 per cent in Sweden to about 13 per
44
Note, however, that apprenticeship wages in Europe are often much lower than the American minimum wage.
Table 3.3. Unemployment and Wage Structure

Dependent variable Log (unemployment) Youth/Adult Unem- Unskilled/All Un- Unemployment
1996 ployment Ratio (%) employment Ratio Outows (%
(%) monthly) 1994
Constant 6.573 8.719 0.082 -113.172
(2.30) (1.69) (0.17) (-2.04)
Minimum -1.136 -2.958 2.077 12.283
Wage (-0.79) (-0.41) (3.17) (0.44)
Wage -1.831 -2.180 0.314 59.534
inequality (-1.48) (-1.10) (1.74) (2.48)
Centralization -0.091 -0.050 -0.022 2.820
(-1.42) (-0.34) (-1.68) (2.26)
Adj. R2 0.015 0.367 0.133 -0.106
Notes: T-statistics in parentheses. Explanations see Table 3.2.
Sources: Table 3.2.
cent in Italy, France and Germany (OECD 1996: Table 3.3; Dolado et al. 1996).
These averages disguise specics. Spain, with Europe's highest unemployment, has a low minimum wage of 32 per
cent of the average wage for adults, 21 per cent for youth.45 Another puzzling fact is that net job growth in Europe
during the troubled 1980s is more or less identical to the golden 1960s when wage distributions were less egalitarian.
Both anomalies are best explained by changes in labour supply: supply exceeds demand due to the combination of the
post-war population boom and the growth in women's participation. Spain, as Jimeno and Toharia (1994) and
Marimon (1996) point out, is especially unfortunate because these pressures coincide with massive de-ruralization and
de-industrialization. In a competitive market-clearing model, this should result in downward wage adjustments. So we
are back to the initial problem.46
Early USA studies within the competitive model framework found a very strong elasticity: a 10 per cent increase in the
minimum wage lowers teenage employment by 13 per cent (Brown et al. 1982).47 Other studies estimate that a 1 per
cent slower yearly growth of labour costs among the unskilled will produce 0.5 per cent more job growth per annum
(Glyn 1995). If this were true, most European unemployment rates
45
The percentage of workers who actually get paid the minimum wage in Spain is apparently very low.
46
Yet another anomaly is that, in the USA. women's real earnings have improved compared to males' notwithstanding the simultaneous large rise in female labour supply.
47
This nding may be less pertinent today considering that most countries attempt to raise the school-leaving age.
could be cut in half within ten years by simply lowering minimum wages.
With the notable exception of Card and Krueger's (1995) quasi-experimental study of the MacDonalds economy in the
USAwhich nds that higher wage minimums actually encourage job growththe consensus that emerges from the
recent literature is that high minimum wages and egalitarian wage structures have little or no effect on unemployment
as such, but that, again, they do affect the job opportunities of youth and low-skilled workers adversely (Wood 1994;
Bertola and Ichino 1995b; Dolado et al. 1996; OECD 1996; Nickell 1997).
The problem, then, seems to be identical to that of labour costs. Firms' demand for the less skilled and youth is more
sensitive to relative price. Yet, as earlier, this leaves us with some unanswered questions: why is youth unemployment
so high in some European countries even if the youth minimum wage is on par with, or even lower, than the USA?
Why has it risen even if the minimum has been stagnant? Why is the relative unemployment rate of the unskilled much
higher in the USA, Canada and the UK than in Italy, Spain or France?
There are no unambiguous answers to these puzzles. There are, however, some relatively convincing hypotheses.
Firstly, as discussed earlier, wage exibility may be of much greater importance where the labour force contains a
surplus of un-trainable workers. At the same time, not even extremely wide and growing wage differentials might
guarantee that the bottom-end will be absorbed. Net job-creation today depends mainly on services where, even at
the low-end, a minimum of social skills are a precondition of employment. This may favour women and college
students, but it works against laid-off male lumberjacks. If low-end wages fall too far the effect may be perverse in the
sense that crime may become the preferred choice. As Bjrklund and Freeman (1994) suggest, American youth
unemployment statistics would look much less glorious if we also counted in the long-term prison population.
A second compelling hypothesis comes from the theory of the hump-shaped curvilinear relationship between wages
and unemployment (Calmfors and Drifll 1988; Calmfors 1993; Soskice 1990; Rowthorne 1992). Wage compression
and rising real wages are more likely to produce unemployment if unions are powerful but bargaining uncoordinated.
Alternatively, feeble union power, as in the US, or broad consensual corporatism might actually diminish
unemployment. This, at one extreme, helps explain Sweden's erstwhile capacity to contain unemployment at very low
levels and, at the same time, the recent jump towards typical EC level unemployment rates: during the 1980s, Sweden
has witnessed a marked decentralization of wage bargaining. At the
other extreme, it helps explain the eclipse of mass unemployment in the UK and the USA since the 1980s. American
trade unionism experienced a virtual collapse, and in the UK it has been severely weakened (Freeman 1990). And, in
the middle, it may explain why some countries, like France, Italy and Spain, suffer from exceptionally high
unemployment: they combine strong, or insider-, bargaining power with weak centralization and co-ordination.48
The impact of wage structure will, nally, depend on the sectoral distribution of low-wage workers. If, as Dolado et al.
(1996: Table 2) suggest, most minimum wage workers, especially youth and unskilled, are employed in the non-
tradables sector, they are not directly exposed to global wage competition. In France, the Netherlands and Spain,
6580 per cent of low-wage workers are in restaurant and personal services. This is also true in the USA where,
however, there also exists a large low-wage labour market in traditional manufacturing, such as textiles and shoes. But
even if they are sheltered from international competition, immigration may exert the same effect. Here lies a major
difference between the European and American economies: the inuence of immigration on wage-setting. A parallel
phenomenon can be argued for Scandinavia: the high employment rates even among mothers with small children have
been compatible with high wages because women's employment has been so concentrated in protected public welfare
state jobs (Pedersen and Smith 1995).
Let us repeat our regression exercise from the previous section, again distinguishing between the effects on
unemployment levels, structure, and ows. In terms of wage effects, there are two things that matter: the minimum and
wage inequalitieshere measured as the ratio between fth and rst decile earnings. It is clear that we need to estimate
the simultaneous effect of the bargaining structureagain our centralization measure (see Table 3.3).49
The results from Table 3.3 seem at rst contradictory. A high minimum wage reduces aggregate unemployment and
also the youth bias, and it increases outows while, simultaneously, it contributes substantially to the unskilled bias. Here
we should rstly note the very large error term in all except the low-skilled bias model. In any case, the only clear
evidence is that a high minimum wage adversely affects the low-skilled. The impact of wage distributions is more
consistent:
48
Most observers agree that the Dutch miracle is mainly the product of consensual real wage restraint. Italy has also experienced consensual wage restraint since 1993. It is,
however, probably too early to expect any real effects on employment. See Samek's chapter in this volume.
49
As in Table 5, I re-estimated the model for unskilled workers, including the early retirement variable. Again, the effect is not signicant.
wage inequality does help bring down unemployment and raise outows. Overall, there is some evidence that wage
structures, if egalitarian, help price out workers from the labour market and, possibly, contribute also to structural
unemployment. Bargaining structure (centralization) may alleviate unemployment and, in particular, accelerate
outows, but does not cancel out the wage effects.
In brief, we have found some evidence of a negative effect of equality on jobs, but how much does it actually matter?
The extremely low variance explained indicates that we may be considering effects that are signicant but which do not
matter much in the overall picture. In the case of outows, a negative R-squared indicates that whatever results emerge
are purely random. Wage structural variables explain virtually nothing. We would have thought that bargaining
structure should have mattered more than it does. If the hump-shaped curve hypothesis is correct, high egalitarian
wages might very well produce low unemployment, like in Scandinavia, if the social partners are capable of coordinated
bargaining, and if public sector jobs offset the potentially negative impact of wage equality. Yet, the wage coefcients
are systematically not affected by whether we include, or omit, our bargaining variable in the model.50
The general problem, like before, is that there are too many alternative and potentially important mediating factors. For
example, the less skilled in Southern Europe are not particularly worse hit by unemployment, arguably because of early
retirement. Early retirement, however, does not seem to affect the overall picture.51 There are also powerful exogenous
factors that may determine the outcome, such as the skill-base or sectoral composition of the labour force. The
absence of solid empirical ndings as regards both the impact of social policies, wage costs, and wage structure, may
simply be due to the fact that social reality is too complex to model.
Employment Protection and Unemployment

Although there is no shortage of assertions that employment protection causes unemployment, formal economic
theory makes this link only
50
In separate analyses of the wage models (not shown), the centralization variable does have a substantial negative effect on the NAWRU measure of structural unemployment
(B = 0.15; t = -2.3), and including the centralization variable brings the D5/D1 variable to almost signicance with a negative sign.
51
Separate analyses (not shown) indicate that a variable measuring levels of early retirement (% males, 5564 not active) has absolutely no effect (B = 0.0004) on the low-
skilled bias of unemployment.
indirectly. The direct causal line is from the costs of hiring/ring, to labour turnover across the demand cycle. The
inference is that slower labour adjustment leads to lower hiring rates during upswings but also lower ring rates during
downswings. Since the two may cancel each other out, the effect may be indeterminate.52 For a discussion, see Bentolila
and Bertola (1990), Buechtemann (1993), Blank (1994), OECD (1994), Alogoskous et al. (1995), Bertola and Ichino
(1995a), and Siebert (1997).
In France and Spain, employment protection is dominated by legislation; in Italy, Denmark and the UK, by contractual
bargaining. In either case, its history is quite recentdating back to the tail-end of the post-war boom and the
proliferation of labour unrest in the late 1960s. As Bchtemann (1993) observes, job protection was relatively
unproblematic as long as full employment obtained, but it may have brought about negative effects when, after the
mid-1970s, unemployment rose and labour adjustments became more urgent. It stands to reason that if workers begin
to feel more insecure, they are less likely to quit voluntarily. If this obtains, any given level of job protection will become
more binding and this should inuence employers' hiring strategies (OECD 1997).
Identiable Effects of Employment Protection

As noted, if job protection causes unemployment it is because it affects rms' strategies both in terms of speed and
strategy of labour adjustment. If they hire fewer workers in upswings and re fewer in downturns, there will be less
mobility but not necessarily more unemploymentbut only in the short to medium term. In the longer run, the lack
of rotation is likely to raise structural unemployment; in other words, one consequence may be unemployment
hysteresis because insiders and outsiders cease to be substitutable. Strong employment protection with segmentation is
likely to produce an exclusion effect among weaker groups with a lower marginal productivityonce again, primarily
youth, women and the low-skilled (Bentolila and Bertola 1990). The effects of employment protection via labour
turnover have, therefore, less to do with the level of aggregate unemployment and more to do with its nature: high and
persistent unemployment with low chances of mobility into jobs, especially among the weakest labour force groups.
Econometric research on the effects of employment protection adopts a relatively narrow measure of rigidities the
difculties of
52
Here we have a parallel to the effects of unemployment insurance: it may, at the same time, induce negative work-incentives and added incentives to take a job.
measurement are immense. Most studies employ an index that summarizes the difculty of dismissing workers: ring
costs such as severance pay, notication, priority rules such as seniority or civil status, and procedural obstacles. Some
measures also include restrictions against the use of temporary or atypical contracts. Most rankings contain a fair
amount of interpretation (such as Bertola 1990), and some are predominantly subjective in the sense that they are
based on employers' responses to questionairessuch as the 1989 EC survey or the 1985 International Organization
of Employers study (OECD 1994: Chapter 6).
In any case, the various indicies of rigidity are highly correlated with bi-variate correlations in the order of 0.8 and 0.9.
The OECD (1994: Table 6.7) has constructed a synthetic ranking based on a menu of attributes that, even if still rather
narrowly conceived, is highly consistent. In virtually all comparative rankings there emerges a clear pattern with the
USA and the Anglosaxon nations, together with Denmark and Switzerland, being very unregulated and Italy, Spain,
Greece, Portugal, Belgium and Germany being very rigid.
These indicies are one point in time, reecting conditions during the year of measurement. Most indicies capture the
mid-1980s (Bertola 1990) or the early 1990s (Grubb and Wells 1993; OECD 1994). This is problematic because many
nations have been deregulating their labour markets over the past ten years. Collective bargaining has been
decentralized in France, The Netherlands and Sweden, and unions have been severely weakened in the UK. Laws
against dismissal have been eased in Belgium, Germany and now also in Italy. Many countries, but especially Spain,
have made short-term, temporary employment contracts possible, and most countries have, in one way or another,
introduced incentives to hire youth.
Lacking genuinely dynamic data, most available research is unsurprisingly of the cross-sectional kind. From the
literature emerge a number of unambiguous results that regard turnover. Strong worker protection does bias rms'
behaviour towards hours, rather than personnel, adjustment (Bchtemann 1993; Bertola 1990). In a rare case of time-
series analysis, Houseman and Abraham's (1995) USA-Germany comparison shows that German rms mainly adjust
hours, while American rms adjust personnel (see also Schettkat 1995). Adam and Canziani (1998), likewise, nd that
Spanish employers manage labour ows with temporary contracts which now account for 33 per cent of all
employment, while Italian employers prefer to adjust hours.
It is, therefore, not surprising that most studies nd that rigidities cause fewer lay-offs and lower levels of hiring across
cycles (Lazear 1990; Bertola 1990; Alogoskous et al. 1995). But the connection to
unemployment remains unclear and even the turnover-effect may be spurious. Lazear's (1990), Scarpetta's (1996) and
Elmeskov et al.'s (1998) studies are among the few that argue a direct link with unemployment. Lazear's ndings may
be questioned because of his narrow measure of rigidities, and the effects that Scarpetta identify are mainly mediated:
strong unions in highly regulated systems produce unemployment, but the effect disappears with co-ordinated
bargaining. Hence, the real variable of interest here seems to be bargaining structure rather than rigidities as such. As
discussed above, indications are that the potential ills of regulation disappear in co-ordinated bargaining systems (see
also Layard et al. 1991). Elmeskov et al.'s study is not comparable because their dependent variable is the NAWRU.
There are two other major ndings in the literature that modify the picture substantially. Many studies fail to uncover
any signicant link whatsoever between employment protection and unemployment (Bertola 1990; Alogoskous et al.
1995; Bertola and Rogerson 1995; Nickell 1997; Nickell and Layard 1998). The picture is modied, however, when
rigidities are studied together with other institutional factors. Bertola and Rogerson (1995) nd that rings but not
hirings do increase if strong protection coincides with wage compression. If this is true it means that countries with
rigid protection plus wage equality present a worst-case scenario as far as unemployment is concerned. Scarpetta
(1996), as noted, also points to decisive interaction-effects. In his ndings, a worst-case scenario would combine strong
employment regulations with generous unemployment benets and uncoordinated bargaining. In both cases the
bottom line is that institutional features may, or may not, catalyze negative unemployment effects.53
This kind of combined effect between rigidities, wage structure and bargaining system nds partial conrmation in
Table 3.4, namely in the case of outows and of unskilled worker unemployment bias. That is, centralized bargaining
makes high minimum wages more harmful for the low-skilled and for outows than otherwise would be the case.54
Otherwise, bargaining structure makes no difference. Minimum wages have absolutely no signicant effect on overall,
or on youth unemployment, nor on outows.55 Note that in these equations, employment protection has a signicant,
albeit very weak, inuence on unemployment levels
53
Another worst-case scenario would combine liberalized dismissal regulations with unaltered obstacles to hiring. In this case, we would expect asymmetry across the cycle:
massive layoffs during downswings without commensurate hirings during upswings.
54
But note the huge error term in the outow model.
55
Their lack of effect on youth unemployment is fairly easy to explain, given that youth jobs are often heavily subsidized and/or a large share of youth workers may work
under apprenticeship contracts. The same equations were estimated with the D5/D1 earnings differential variable, which systematically failed to be signicant.
Table 3.4. Unemployment, Wages, and Labour Market Regulation

Dependent log (Unemployment) Youth/Adult unem- Unskilled/All unem- Unemployment Out-
Variable: 1996 ployment ratio (%) ployment ratio (%) ows 1994
I II I II I II I II
Constant 2.556 2.547 3.094 1.734 1.573 0.985 26.054 25.547
(5.02) (5.03) (1.57) (0.64) (5.68) (3.32) (2.54) (2.28)
Employ- 0.046 0.052 0.156 0.174 -0.030 -0.002 -1.113 -1.118
ment
Protectiona (2.41) (2.82) (2.31) (2.20) (-2.87) (-1.98) (-2.92) (-2.73)
Minimum -1.687 -1.452 -3.173 1.255 0.345 1.743 1.734 5.676
Wage (-1.61) (-1.21) (-0.81) (0.18) (0.61) (2.29) (0.08) (0.21)
Centraliza- -0.021 -0.102 -0.024 -0.151
tion
(-0.85) (-0.72) (-1.59) (-0.27)
Adj. R 2
0.228 0.182 0.297 0.645 0.374 0.324 0.390 0.344
a
The measure used here is the OECD synthetic indicator (OECD 1994: Table 6.7).
Notes: T-statistics in parentheses. For explanations see Table 3.2.
Source: see Table 3.2.
and on outows.56 Rigidities are much more decisive as far as youth unemployment is concerned.
The systematic nding that employment protection affects who rather than how many are unemployed, is not only
consistent with theory, but also with what we otherwise know about countries' regulatory practice. The kinds of
deregulation that have swept across Europe over the past decade have, by and large, been partial rather than across-
the-board. Spain's introduction of temporary contracts in the early 1980s is proto-typical. As shown by Bentolila and
Dolado (1994), such partial deregulation may actually produce the perverse effect of strengthening the insiders,
heightening segmentation, and reinforcing hysteresis (see also Blanchard and Summers 1987; Bertola and Ichino
1995a). As a consequence, if employment regulation is to have any noticeable effect on unemployment it is most likely
to be selective and, once again, concentrated among youth and low-skilled workers.
56
That is only the case when unemployment level is logged. Most studies do not log unemployment, although there is a good case for doing so given the very skewed cross-
sectional distribution of aggregate unemployment ratesespecially considering extreme cases, such as Spain; the OECD distribution of unemployment is, in fact, quite
analogous to typical income distributions. In any case, one should not attach too much importance to the fact that employment protection suddenly turns signicant when
unemployment is logged. Rather, it may be interpreted as additional evidence of the weakness of the ndings. In most specications, employment protection has no effect
on unemployment levels (see below).
There remains the question of what are spurious as opposed to genuine effects. The pervasiveness of inconsistent and
even contradictory ndings suggests that the direct relationship is generally not robust. If there is such a thing as an
emerging consensus, it is limited to two ndings: that strong worker protection affects the composition of unemployment
rather than overall levels; and that context matters (Bertola and Rogerson 1995; Bertola and Ichino 1995a;
Alogoskous et al. 1995). In brief, we need a more multi-faceted theory of regulation and unemployment because there
is no straightforward connection between the two.
Most of the individual econometric studies that I have examined assume that the relationship is monotonic. Since there
are many indications that such an assumption does not hold, we should re-specify the causal relationship. If regulation
matters only in certain circumstances, the interesting relationships should be curvilinear rather than monotonic. The
pressing issue is to identify the hidden variables that represent certain circumstances.
Table 3.5 below, is an attempt to test the non-linearity hypothesis. Model I tests whether the impact of rigidities on
unemployment is straightforward and monotonic, as most assume, or, alternatively hump-shaped, u-shaped or
quadratic. If this happens to be the case, rigidities inuence unemployment solely under certain circumstances. What
are these circumstances? The leading candidate in the literature is a collective bargaining structure, in particular
centralization and
Table 3.5. The Impact of Employment Protection on Unemployment
Log (unemployment) Youth Bias Unskilled Bias Unemployment Out-
1996 ows
I II I II I II I II
Constant 1.97* 1.93* 2.86* 2.14* 1.37* 1.56* 34.45* 24.30*
Employ- -0.01 0.03 -0.17 0.16* 0.03 -0.25* -3.40* -0.93*
ment Pro-
tection
Protection 0.00 0.02 -0.01* 1.08*
Squared
Centraliza- -0.02 -0.27* 0.00 -0.04
tion
Adj. R2 0.120 0.136 0.550 0.390 0.240 0.327 0.491 0.320
* indicates level of signicance superior to 0.05.
Notes: Two different causal specications are used here.
Source: Table 3.2.
levels of co-ordination. Model II, therefore, presents estimates of the impact of rigidities controlling for degree of
bargaining centralization. As earlier, I differentiate between levels, structure and ows of unemployment.
Table 3.5 reafrms that labour market regulation has no effect on levels of unemployment. Indeed, neither does
bargaining centralization.57 In this case, the hump-shaped curve hypothesis fails. Differences in overall unemployment
must therefore nd explanation elsewhere.
The results change when we, instead, examine ows and who bears the brunt of whatever unemployment prevails.
Employment regulation has a relatively strong negative impact on leaving unemployment, and it is now a signicant
explanation of the youth and unskilled bias, but mainly in a mediated way. The youth bias emerges when rigidities
combine with bargaining structure: union centralization has, as we would have predicted, a negative effect on the youth
biasmost probably because centralized bargaining is more solidaristic, taking into account the situation of the
outsiders. However, in the very same context, employment regulation works against youth. The prole that emerges
for the low-skilled bias seems more puzzling. While it is clear that the impact of employment protection is curvilinear,
it is less clear what, precisely, causes this to be soit is certainly not bargaining structure.58 It appears that our earlier
results stand: the low-skilled are mainly vulnerable to the wage effect. On balance, as theory would predict,
employment regulations primarily inuence ows. Their effect on the probability of exiting unemployment is clearly
non-monotonic, but this cannot be explained by centralizationalthough co-ordinated bargaining has some effect.59
We should, in summary, be quite cautious about the way that rigidities affect unemployment. The consequences of
worker protection may be offset or compounded by other contextual factors, but so far we do not have a very clear
idea of what, precisely these are. Bargaining structure appears to be a plausible explanation, but only for youth
unemployment. We therefore still face the challenge of identifying the real smoking gun. In Chapter 4, I attempt to re-
think the issue with this in mind.
57
However, in this case there is evidence that co-ordinated bargaining, rather than centralization, matters.
58
Note also the substantial change in variance explained between Models I and II in the youth-bias estimation. Re-estimating the same model with a variable for co-ordinated
bargaining rather than centralization changes nothing and, as before, the early retirement variable is insignicant.
59
While centralization has no effect on outows, this is not true for co-ordinated bargaining. Indeed, controlling for rigidities, the degree of co-ordination signicantly
increases unemployment exits.
Conclusions
The effects of all three types of regulation go in the same direction: they slow down ows out of unemployment, and
they contribute to the specic unemployment problem among youth and low-skilled workers. Beyond that there is
simply no consistent, nor convincing, relationship between employment protection and unemploymentat least as
conceptualized and measured in the quantitative literature. So far, Nickell and Layard (1998) hardly exaggerate when
they propose that we should cease worrying about labour market regulation.
When, as with structure and ows, rigidities do matter somewhat, the immediate reason is that they affect the relative
price of select groups of workers. The two key-groups here are youth and the low-skilled, but subject to very different
logics. At rst glance, the low-skilled conform to standard economic theorysocial benets and high minimum wages
harm their employment prospects. But then economic theory faces difculties because employment protection is
negatively related to the unskilled bias. In other words, there are other forces at work. What might these be? It is clearly
not corporatism, but possibly other unmeasured variablesbe it offsetting policies, such as early retirement, or
structural factors, such as an excess stock of low-skilled workers. These will be examined in Chapter 4.
Youth, in contrast, seem to be subject to a very different logicone that bears virtually no relationship to standard
economic theory. For obvious reasons unemployment benet schemes are irrelevant and, much less obviously, the
youth bias appears completely immune to wage compression and the minimum wage. But rigidities do seem to
matter, especially in a curvilinear way. Part of the hump-shaped relationship can be attributed to the bargaining
system: centralization and coordination help reduce the youth bias, but less than one would have expected. In brief,
our survey of the evidence raises as many questions as it manages to answer. We clearly require structural
explanations that go beyond a simple supply-demand function.
The currently favoured institutional explanationbargaining structureholds the advantage, in that it takes into
account non-linear relations. Economies blessed with centralized, co-ordinated bargaining may permit themselves
more equality and social protection without adverse employment effects; while the worst off are those where unions
are powerful but represent only the core workforce, thus creating an insider-outsider divide. In the latter case, equality
will be negative because all it does is uphold the privileges of insiders at the cost of others. Yet the bargaining variables
generally fail to explain much. Why? One reason
is that the causal direction is a completely open question. It may be that the kind of rigidities that cause unemployment
are produced by insider-bargaining systems in the rst place. Or, it may be that regulations, once in place, nurture and
reinforce insider-outsider effects. There is, to my knowledge, no research at all on this question.
Neither I, nor most studies, can identify any clear link between regulation and unemployment, not even when taking
into account bargaining institutions. This is possibly because we are barking up the wrong tree, so to speak. How, then,
might we nd the right tree to bark at? One promising clue comes from a puzzle repeated again and again in the
literature: why, despite waves of deregulation over the past decade, do we have more structural unemployment today
than earlier? Blanchard's (1998) suggestion that the clue lies in exogenous shocks is enticing, and he may be right that
the culprit is overly restrictive scal and monetary policy. There is, however, another candidate that suggests itself from
a second puzzle: Europe's unemployment record is certainly worse than in the golden age, but this is not the case with
regard to job creation. In fact, average annual rates of job growth in the 1960s (0.96) are only marginally superior to the
troubled 1980s (0.83). More puzzling is the fact that 1960s employment growth was more negatively correlated with real
wage growth than in the 1980s (-0.47 compared to -0.20)! When we regress job growth in the 1960s and the 1980s
with real wage growth and a simple dummy for EC membership, it turns out that the wage effect completely
disappears. For the 1960s, the EC dummy accounts for 42 per cent of the total variance explained; for the 1980s, 15
per cent.
Hence, we seem to be back to the Eurosclerosis explanation, but not in the version that we are normally served.
There may exist an idiosynchratically European problem, but it has evidently little to do with welfare statism, wage
equality or employment protection. What then?
Europe's massive unemployment problem, despite stable rates of job growth, suggests that population labour supply
may be one of those exogenous shocks we are looking for. Indeed, most EU member countries combine the coming
of age of the huge baby-boom generations with a delayed burst of female labour supply. A handful countries, basically
the worst-case scenarios, have the added disadvantage that this occurs in an unfortunate historical juncture: in a
period marked by heavy de-industrialization and continued de-ruralization.
To stimulate job growth and to compensate for mass employment losses, contemporary economies depend primarily
on the dynamics of services. The Eurosclerosis thesis and basically the entire literature on labour market regulation
suffers from one great fault: it appears to assume that the question of jobs today, like before, is a question of
manufacturing. Some authors do recognize that services dominate
contemporary job creation, but they consider it unproblematicindeed a favourable prospectfrom the point of
view of a jobs-equality trade-off because, it is argued, the service economy is sheltered from global competition. Still,
the service economy is subject to other, possibly more severe, constraints. These need to be better understood and are
the themes that I will examine in Chapter 4.
Appendix
Table 3A. Trends in Employment and Unemployment Across Business Cycles
Annual change, employment Change in unemployment Change in unemployment
198494 (%) rate, 198393 (%) rate, 199396 (%)
United States 1.6 -3.0 -1.3
Denmark 0.0 1.0 -4.1
Sweden -0.8 5.6 0.5
France 0.2 2.6 0.7
Germany 0.6 0.2 1.1
Italy -0.2 2.6 1.7
Netherlands 1.7 -3.1 -0.3
Spain 0.7 5.3 -0.6
UK 0.5 -0.6 -2.3
Source: OECD, Employment Outlook (various issues). Paris: OECD.
Table 3B. Indicators of Employment Stability

Median years tenure, 1995 % Job-Job turnover 198391 Change of employer (average
no. years) 198691
United States 4.2 23 0.8
Denmark 4.4 30 1.1
Sweden 7.8 29 n.a
France 7.7 27 0.6
Germany 10.7 17 0.3
Italy 8.9 23 0.5
Netherlands 5.5 n.a n.a
Spain 4.6 n.a n.a
UK 5.0 15 0.6
Table 3C. Unemployment Flows

Layoffs as % of total employment troughs Exit from Unemployment Monthly rates,
19931994 (%)
1980s 1990s ALL YOUTH
United States 5.1 4.0 38 46
Denmark 5.1 8.6 21 28
Sweden 5.9 18 28
France 3.2 5.9 3 4
Germany 1.6 4.3 9 14
Italy 1.5 2.7 10 10
Netherlands 3.4 2.5 6 11
Spain 7.7 14.8 3 4
UK 4.4 4.4 9 11
References 95
References
Adam, P., and Canziani, P. (1998), Partial de-regulation: xed-term contracts in Italy and Spain, Centre for Economic
Performance Discussion Paper, 386.
Alogoskous, G., Bean, C., Bertola, G., Cohen, D., Dolado, J., and Saint-Paul, G. (1995), Unemployment: Choices for
Europe. London: Centre for Economic Policy Research.
Atkinson, A., and Micklewright, J. (1991), Unemployment compensation and labor market transitions. Journal of
Economic Literature, 29: 1679727.
and Mogensen, G. V., (eds.) (1993), Welfare and Work Incentives. Oxford: Clarendon Press.
Bean, C. (1994), European Unemployment: A Survey, Journal of Economic Literature, 32: 573619.
Bentolila, S., and Bertola, G. (1990), Firing costs and labour demand: How bad is Eurosclerosis?, Review of Economic
Studies, 57/3: 381402.
and Dolado, J. (1994), Labour exibility and wages: Lessons from Spain, Economic Policy, 18: 53100.
Bertola, G. (1990), Job security, employment and wages, European Economic Review, 34/4: 85186.
and Ichino, A. (1995a), Crossing the River: a comparative perspective on Italian employment dynamics,
(1995b), Wage inequality and unemployment: US versus Europe, NBER Macroeconomics Annual 5. Cambridge,
Mass: MIT Press.
and Rogerson, R. (1995), Institutions and labor reallocation, Unpublished paper. Milano: Universita Bocconi.
96 References
Bjorklund, A., and Freeman, R. (1994), Generating Equality and Eliminating Poverty, the Swedish Way, NBER
Working Paper, 4945.
Blanchard, O. (1998), Thinking about unemployment, Paolo Baf Lecture on Money and Finance, 16 October 1998. Rome:
University of Rome.
Dornbusch, R., and Layard, R. (1986), Restoring Europe's Prosperity. Cambridge, Mass: MIT Press.
and Summers, L. H. (1987), Histerisis and the European Employment Problem, NBER Macroeconomics Annual 1.
Cambridge, Mass: MIT Press.
Blank, R. (1994), Social Protection versus Economic Flexibility. Chicago, Ill.: University of Chicago Press.
(1995), Changes in inequality and employment over the 1980s, Journal of Population Economics, 8: 121.
Booth, A. (1997), An analysis of ring costs and their implications for unemployment policy, in D. Snower and G. de
la Dehesa (eds.) Unemployment Policy. Cambridge: Cambridge University Press, 35987.
Bchteman, C. (1993), Employment Security and Labor Market Behavior. Ithaca, NY.: ILR Press.
Calmfors, L. (1993), Centralization of wage bargaining and macroeconomic performance: a survey, OECD Economic
Studies, no 21.
and Drifll, J. (1988), Bargaining structure, corporatism and macroeconomic performance, Economic Policy, 6:
1461.
Card, D., and Krueger, A. (1995), Myth and Measurement: The new economics of the minimum wage. Princeton, NJ.: Princeton
University Press.
Dolado, J., Kramarz, F., Machin, S., Manning, A., Margolis, D., and Teulings, C. (1996), The Economic impact of
minimum wages in Europe, Economic Policy, 23: 31972.
Elmeskov, J., Martin, J., and Scarpetta, S. (1998), Key lessons for labour market reforms: evidence from OECD
countries' experiences, Unpublished paper. Paris: OECD.
Esping-Andersen, G. (1999), Social Foundations of Postindustrial Economies. Oxford: Oxford University Press.
European Commission (1991), Ad hoc Survey, European Economy, no 47.
(1994), Growth, Competitiveness, Employment. The Challenges and Ways forward into the 21st Century. Brussels: EC.
Freeman, R. (1988), Labor market institutions and economic performance, Economic Policy, 3: 6380.
(1994), Working under Different Rules. New York: Russell Sage.
Glyn, A. (1995), Unemployment and Inequality, WZB Working Paper, FS I: 95303.
Gregg, P., and Manning, A. (1997), Labour market regulation and unemployment, in D. Snower and G. de la Dehesa
(eds.) Unemployment Policy. Cambridge: Cambridge University Press, 395419.
Haagen Pedersen, L. (1997), Progressiv arbejdsindkomstbeskatning og arbejdsloeshed, in L. Pedersen (ed.) Er der Veje
til Fuld Beskaeftigelse? Copenhagen: Socialforsknings Instituttet, 15788.
Holmlund, B. (1997), Unemployment insurance in theory and practice, Working paper series no. 25, Uppsala:
Department of Economics, Uppsala University.
References 97
Houseman, S., and Abraham, K., (1995), Labor adjustment under different institutional structures, in F. Buttler et al.
(eds.) Institutional Frameworks and Labor Market performance. London: Routledge, 285315.
ISTAT, (1997), Rapporto sull' Italia. Bologne: Il Mulino.
Jensen, P. (1997), Ledighedens dynamik, in L. Pedersen, (ed.) Er Der Veje til Fuld Beskaeftigelse?. Copenhagen:
Socialforsknings Instituttet, 3962.
Jimeno, J., and Toharia, L. (1994), Unemployment and Labor Flexibility: the Case of Spain. Geneva: ILO.
Layard, R., Nickell, S., and Jackman, R. (1991), Unemployment: Macroeconomic Performance and the Labour Market. Oxford:
Oxford University Press.
Lazear, E. (1990), Job security provisions and unemployment, Quarterly Journal of Economics, 105: 699726.
Lindbeck, A., and Snower, D. (1988), The Insider-Outsider Theory of Employment and Unemployment. Cambridge, Mass.: MIT
Press.
Lockwood, B., and Manning, A. (1993), Wage setting and the tax system. Theory and evidence for the United
Kingdom, Journal of Public Economics, 52: 129.
Malcomson, J., and Sartor, N. (1987), Tax push ination in a unionized labor market, European Economic Review, 31:
158196.
Marimon, R. (1996), Actual versus Virtual Employment. European University Institute Working Paper 96/21.
Moftt, R. (1992), Incentive effects of the US welfare system, Journal of Economic Literature, 30: 161.
Myles, J. (1996), When markets fail: Social welfare in Canada and the United States, in G. Esping-Andersen (ed.)
Welfare States in Transition. London: Sage, 11640.
Nickell, S. (1997), Unemployment and labour market rigidities: Europe versus America, Journal of Economic Perspectives,
11/3: 5574.
and Bell, B. (1995), The collapse in the demand for the unskilled across the OECD, Oxford Review of Economic
Policy.
and Layard, R. (1998), Labour market institutions and economic performance, Centre for Economic Performance
Discussion Paper, 407/ September.
OECD (1993), Employment Outlook. Paris: OECD.
(1994), The OECD Jobs Study. Paris: OECD.
(1996a), Employment Outlook. Paris: OECD.
(1996b), Literacy and the Economy. Paris: OECD.
(1997), Employment Outlook. Paris: OECD.
Pedersen, L. (ed.) (1997), Er Der veje til Fuld Beskaeftigelse? Copenhagen: Socialforsknings Instituttet.
Pedersen, P., and Smith, N. (1995), The welfare state and the labour market. CLS Working Paper, no. 95-17. Aarhus:
University of Aarhus.
(1997), Arbejdsudbud og incitamenter, in L. Pedersen (ed.) Er der Veje til Fuld beskaeftigelse?. Copenhagen:
Socialforsknings Instituttet, 91114.
and Westergaard-Nielsen, N. (1993), Unemployment: a Review of the Evidence from panel data, OECD
Economic Studies, no. 20.
98 References
Rosen, S. (1985), Implicit contracts: An overview, Journal of Economic Literature, 23: 114475.
Rowthorne, R. (1992), Centralization, employment and wage dispersion, The Economic Journal, 102.
Scarpetta, S. (1996), Assessing the role of labour market policies and institutional settings on unemployment: a cross-
country study, OECD Economic Studies, no. 26.
Schettkat, R. (1995), The macroeconomic performance of the German labor market, in Buttler et al. (eds.) Institutional
Frameworks and Labour Market Performance. London: Routledge, 31642.
(1997), Employment protection and labor mobility in Europe, International Review of Applied Economics, 11/1.
Siebert, H. (1997), Labor market rigidities and unemployment in Europe, Institut fur Weltwirtschaft Working Paper no.
787.
Snower, D. (1997), The low-skill, bad-job trap, in A. Booth and D. Snower, (eds.) Acquiring Skills. Market Failures, Their
Soskice, D. (1990), Wage determination: the changing role of institutions in advanced industrialized countries, Oxford
Review of Economic Policy, 6/4.
Toharia, L. (1997), The Labour Market in Spain. Unpublished manuscript, Universidad de Alcala (February).
Wood, A. (1994), North-South Trade, Employment and Inequality. Oxford: Clarendon Press.
4 Regulation and Context. Reconsidering the
Correlates of Unemployment
Gsta Esping-Andersen60
General Issues
As became clear in Chapter 3, the link between labour market regulations and employment is hard to pin down. In line
with most recent research, we have established that some people are adversely affected by some regulations, but not why
Europe suffers from mass unemployment.61 Failing to identify the smoking gun responsible for unemployment,
researchers have begun to hunt for additional, possibly more promising, explanations. Macro-economists, like
Blanchard (1998), turn to outside shocks and, in particular, overly restrictive scal and monetary policy. Nickell and
Layard (1998) introduce home-ownership as a supplementary cause of rigidities and low labour mobility.62 There exist,
so far, no unambiguous conclusions about the impact of worker protectionunless we ignore the data and embrace
ideology. Why are the results of research so stubbornly ambiguous?
There are too many European countries that do not t the bill of chronic structural unemployment, and they seem to
have very little in common by way of regulatory practice. The Netherlands, with massive unemployment in the mid-
1980s, is now the Dutch Miracle despite hardly any regulatory reform.63 Britain, once the sick man of Europe, is
60
I would like to thank John Martin for his very generous comments, not all of which, unfortunately, I was able to address adequately.
61
Both Nickell (1997) and Nickell and Layard (1998) discover that labour market regulation has only selective effectsunemployment benet duration being one notable
case.
62
A priori, this does not convince easily when we consider that most of Europe's unemployment is among youth and the low-skilled, two groups that are certain to be under-
represented among home-owners. Home-ownership rates are, however, also a corollary of the size of the rental market and is thereby an indirect measure of available cheap
housing. Regardless, in cross-section analyses home-ownership is systematically insignicant.
63
Besides the well-known liberalization of part-time employment, Dutch deregulation has been modest, although there have been some adjustments in the tax-benet system
and in minimum wages. See the country case-study in Part Two in this volume (Chapter 7).
the epitomy of across-the-board deregulation and has done as well as the Dutch. Danish unemployment has dropped
to about 6 per cent within a few years, primarily due to tightening eligibility and shortening duration of unemployment
insurance, while augmenting activation measures. And then there are Austria and Portugal, both highly regulated, with
stable modest unemployment.64
Quantitative research seems to generate more new puzzles than it answers. As I asked in Chapter 3, why do labour
market regulations affect unemployment structure but not levels? This may be the easiest to answer: strict worker
protection reduces labour ows, leading to a cumulation of people in the outsider class. A more intriguing puzzle
emerges from the lack of a linear, one-to-one relationship between regulation and unemployment. Various factors
must somehow come together to produce adverse or favourable employment results. But what are they? What are the
shocks or institutional preconditions that either neutralize or reinforce the impact of rigidities? In brief, what is the
hidden force that hides behind the regulation-unemployment relationship?
Blanchard's (1998) emphasis on tight scal and monetary policy appears convincing. Europe's struggle to restore price
stability and, generally, to satisfy the Maastricht criteria has repressed aggregate demand through high interest rates.
When this coincides with heavily regulated labour markets, the consequences may be especially severe. An alternative,
and widely cited, candidate for hidden force is more institutional: the impact of industrial relations systems. Following
the work of political scientists and economists, such as Bruno and Sachs (1988) and Calmfors and Drifll (1988), the
problem is one of collective action. Since they predict low unemployment with both weak and strong, co-ordinated
bargaining, we should expect a hump-shaped relationship. There is some evidence that the regulation-unemployment
nexus is non-monotonic but, as discussed in Chapter 3, it does not seem to be because of bargaining structure. What
then?
Re-Thinking the Causal Mechanisms

The underlying regulation-unemployment thesis is not always spelled out very well, certainly not in the Eurosclerosis
version. The same goes for the alleged modifying effect of centralized and co-ordinated bargaining. Why should co-
ordination diminish the negative effects of regulatory strictness? In theory this should mainly occur indirectly since
64
Although highly regulated in the formal sense, actual enforcement of regulations is probably much weaker in Portugal.
Regulation and Context 101
centralized and co-ordinated bargaining are a better guarantee of across-the-board wage moderation which, in turn,
should favour more job growth in the long run, all other things being equal. In effect, this is the most cited explanation
for the Dutch miracle.
However, and this is important, across-the-board real wage moderation is unlikely to inuence unemployment structure
unless it is applied selectivelywhich is not a realistic prospect.65 Structure will depend on policies or institutional
conditions that inuence ows from unemployment into employment or leaving the labour force. Employment
protection, the tax wedge, earnings differentials, and the minimum wage may, in fact, reduce ows among less
productive workers, but it is possible that their impact can be neutralized by offsetting exibilities. Easing older
workers out of the labor market via retirement is one such compensator; active labour market policy (ALMP) is
another, especially for youth.66
In brief, we need different causal specications in order to account for unemployment structure and levels.
Unemployment levels, in the logic described above, are an implicit consequence of job growth. Hence, this is where we
should begin. Unemployment structure must be explained by how the combination of rigidities and offsetting
exibilities affect ows. In either case, nonetheless, rigidities, offsetting exibilities, and collective bargaining
institutions operate within an economic backdrop which may be more or less favourable. Let us call this historical
circumstance.
There is one historical circumstance that makes the task of clearing markets unusually arduous regardless of policy
choice or strategy, namely excess stock or supply. Clearing the market may be hampered by two circumstances. One, from
the entry of women and baby-boomers into the labour force. In some countries, such as those in Northern Europe
and North America, female employment growth has been a longer and smoother process; in others, it has been
delayed, yet massive. By and large the same holds for cohort effects. Two, a huge stock of low-skilled workers may
cumulate where de-ruralizationjob loss in agriculture during the 1980sis delayed and still ongoinglike in Spain,
Italy, Portugal, and even Franceand when it coincides with heavy de-industrialization.67
65
Moreover, real wage moderation can be a successful policy only if it is not followed by all nations.
66
There are other possibilities: informal black market employment, which we cannot measure; a second is auto-exibilization via self-employment.
67
Of course, de-industrialization alone, if exceptionally dramatic, may sufce to produce an excess supply of workers difcult to reabsorb. This may be the case for the U.K.,
Belgium, and Sweden.
Massive labour force migration, such as de-ruralization, is obviously nothing new. In the past, many countries managed
quite successfully to reabsorb agricultural labour in dynamic mass production industries. This, of course, is hardly
possible today when services constitute the only realistic source of meaningful job expansion. Labour supply shocks,
like the ones described above, may add to the burden of clearing labour markets, but if services are dynamic the
problem will be only temporary. Rapid services growth will eventually absorb the surplus. Or will it?
Flexibility and the Service Economy

Too often we debate exibility as if the economy and the employment problem were still predominantly industrial. The
fact of the matter is that services dominate any prospects of net job growth. It is not the auto industry, but services, that
must absorb most of Europe's 18 million unemployed, let alone coming generations. Unfortunately, the tertiary sector
is much less understood, and less documented by hard data. It is also more complex and possibly more dualistic,
because it combines highly professionalized productionespecially in banking, nance and business serviceswith
routine, low-skilled jobs, especially in restaurants, personal and some social services. Most young, female, and low-
skilled job-seekers do not move straight into professional services; it follows that we must rely heavily on lower-end
services to make a major dent in today's unemployment.
Many regard the service sectors as unproblematic because they are sheltered from international competition, believing
that, even if wages or costs are uncompetitive, services should grow. This is not necessarily true. Immigrants may
compete with domestic workers: in the USA, for example, Hispanic immigrants constitute a very large share of
employment in personal services. Services are also cost-sensitive because the consumer is usually the household, and
households have budget constraints and the choice of self-servicing.
Demand for the kinds of services that generate a large number of easy-entry jobs, such as social and personal services,
is sensitive to three factors.68 One, household disposable incomethe well-known Engel demand curve: low-income
households will have difculty affording services. Two, relative prices: if the price of services is prohibitive, families will
opt for do-it-yourself solutions. Three, time constraints: double income families are more likely to purchase market
services to compensate for time constraints. The problem across
68
Here I draw heavily on Esping-Andersen (1999).
much of Europe is that high labour costs and the tax wedge make out-servicing expensive and, especially in
Continental Europe, less pressing because of low female employment rates. In a way, low female activity rates and,
hence, low overall employment-to-population rates depress job growth.69
Labour-intensive household services are in the long run potentially inationary since productivity growth is modest
compared to manufacturingthe Baumol cost-disease. Even if low-end services are less unionized, their sluggish
growth in most of Europe is almost certainly a function of relatively high wage costs and tax wedges. Personal services
account for only 5 per cent of European employment, compared to 12 per cent in the USA (Esping-Andersen 1993).70
Of course, high wage costs can be, and are, offset by direct government provision, or by hidden exibilities, such as
informal employment or self-employment. Services may also benet from regulatory differentiation. Most countries'
regulatory practices usually exempt small rms, and personal service establishments tend to be quite small.
If the dynamics of services are, at least in some part, directly related to the micro-economic behaviour of households,
we would expect two factors to become decisive. One, as noted, is the relative cost of services. A second is the
employment of women and mothers. The latter is particularly interesting because it doubles household demand for
outservicing: two earners mean more disposable income and less available time for unpaid self-servicing.
We can, in fact, arrive at some estimate of the household effect, at least as far as growth in the service sector is
concerned. Time-series estimates for three countries, France, the USA, and Spain, suggest that every 100 mothers that
stay in the labour force may create something like 10 extra personal service jobs (Esping-Andersen 1999). This cannot
be estimated in a full model that includes the effect of relative prices. However, separate estimates of the impact of
prices, measured in terms of wage differentials, and household demandthe percentage of family income spent on
personal servicessuggest that the effect of rising demand is substantial, but that relative prices exert an extremely
powerful inuence. For every 1 percentage point rise in earnings inequality we can, holding demand constant, expect a
3 point growth in personal services. Indeed, this 1 point rise in the wage ratio would reverse the
69
As is well known, long-term employment growth is positively correlated with population and latent supply growth. If populations are stagnant or even declining, as in much
of Europe, then encouraging additional supply within the existing, latent stock becomes all the more pressing.
70
Lack of growth in market services has, in many European economies, been offset by public sector services.
Table 4.1. Causes of Women's Employment

Women's Employment Rate
Model I Model II
Constant 54.590*** 45.634***
Employment protection -0.276
Cost of childcare -0.527*** -0.531***
Percent women with secondary
education or higher 0.240** 0.293*
Fertility rate 2.782
% Part-time jobs 0.063
Adjusted R2 0.723 0.693
Notes: Child care costs are measured as the price, relative to earnings, of day care for two children under the age of 6. The child care cost
estimates are net, taking into consideration the value of public direct or indirect subsidies. It is assumed that the husband earns 100% of
average production worker wage, and that the wife earns 66%.
Signicance levels: * = 0.10 level; ** = 0.05 level; *** = 0.01 level.
Source: OECD and Esping-Andersen (1990: chapter 4).
negative trend in personal service jobs, as indicated in the following formula:71
Personal Service Job growth = -2.514a + 0.160b (household spending) + 3.098b (wage inequality)
(-2.82) (2.20) (6.97)
The family-service economy nexus turns the exibility debate on its head. The really salient rigidities may have less to
do with the labour market as such, and more to do with the traditional family! Families will opt for self-servicing when
the full-time housewife is available. But the more that wives are employed, the more families will contribute to service
demand and, thus, job growth. Household consumption and labour supply are a potentially strong employment
multiplier. The question, of course, is how to facilitate women's employment, and here rigidities certainly may matter.
But which?
Table 4.1 indicates that the relative cost of child care may be a much more formidable rigidity than employment
protection. In separate analyses, not shown, social benets, the tax wedge, and wage differential
71
T-statistics in parentheses. N = 17 (adjusted R-squared = 0.865). Service job growth and household spending growth are measured as change in period 19801994. Wage
inequality (the ratio between 5th and 1st decile earnings among males) pertains to 1994, and is therefore not dynamic. Yet, except for the USA and the UK, there were few
changes in wage differentials over these years. Sources: OECD, (1995) National Accounts. Detailed Tables and OECD, (1996) Employment Outlook. Paris: OECD.
variables are all systematically insignicant. I include a measure of the availability of part-time jobs but, surprisingly, it is
far from signicant. Neither, equally surprisingly, is fertility. Women's educational attainment is, as we would expect, a
signicant correlate of employment levels, but it nevertheless remains that, for women, access to affordable child care
constitutes a principal barrier to employment. To avoid multi-colinearity problems, Table 4.1 presents two separate
estimates.
So far, then, there is some evidence that growth of service jobs is intrinsically connected to women's entry in the labour
market. The causal logic is not straightforward because, as we have seen, if women's employment growth fuels service
jobs this, in turn, will only happen if women have access to child care in the rst place. As the estimates in Table 4.1
indicate, a 50 per cent reduction in relative child care costs would generate a 25 percentage point increase in female
employment. In other words, services beget services.
Demographics and Labour Supply

The relevance of deregulation and the choice of what to exibilize depends also on labour supply. It obviously makes
an enormous difference whether the stock of low-qualied workers is large or small. In Britain, the superior
employment performance of women in the last decade may simply reect the fact that male joblessness is concentrated
among low-skilled mass production workers in coal, steel and vehicle manufacturing, all of whom are difcult to re-
employ in services or, more generally, if wage costs are high across-the board. There is also little doubt that Spain's,
and to a lesser extent Italy's and France's, massive unemployment problems stem from the unfortunate coincidence of
de-ruralization and de-industrialization over a fairly brief span of years. Again, an over-supply of low-skilled workers.
Youthand often also womenpose an analogous problem. Even if average years of schooling today are higher than
some decades ago, youth lack experience and most likely also relevant skills. Where, as in Germany, a dual vocational
training system is virtually universal, we nd very low youth unemployment; where, as in Italy, there is no link between
formal education and jobs, even educated youth might have to compete as unskilled labour.
In other words, the structure of rms and skills, the dynamism of services, and the degree of over-supply constitute
the real point of departure for any meaningful policy of deregulation. A selective curtailment of unemployment benet
duration might do the trick in Denmark, but this would seem wholly irrelevant in Italy, where most unemployed
receive no benet whatsoever. Radical exibilization of, say, ring regulations might be deemed necessary for low-end,
labour intensive rms but might also be counter-productive for the high-end core economy. Similarly, a massive
exibilization of wages might help service growth and thus clear markets of unemployed youth and women, but this
would certainly also imply a radical erosion of social welfare guarantees and of trade union powerin other words, not
a very realistic scenario for contemporary Europe.
If the underlying problem is that unemployment is concentrated within specic groups, targeted exibilization
measures are an alluring alternative. Early retirement for older workers, albeit costly, will reduce their risk of
unemploymentand may even improve job opportunities for youth. Government wage subsidies in favor of youth,
active training and job search assistance, or direct public creation of service jobs, as in the Nordic countries, would
undoubtedly enhance the ow of youth out of unemployment.
A Quantitative Reassessment of the Issue

The preceding discussion has tried to highlight issues and interactions that do not normally nd their way into
deregulation studies. The gist is to identify, rstly, the mechanisms that would convert rigiditiesand regulations, as well
as institutions, into labour market outcomes and, then, take account of the structural conditions and constraints within
which these actually operate.
As far as unemployment levels are concerned, we have already seen that employment protection variables have virtually
zero impact. However, if my previous reasoning makes sense, the causal connections ought to be different. The impact
of bargaining structure, that is centralization and co-ordination, should primarily express itself in across-the-board real
wage moderation which, in turn, inuences job-growth in the long run.
Among the battery of regulatory variables examined in Chapter 3, only employment protection and the tax wedge had
any visible effect on rates of employment growth from 1979 to 94. As we discussed earlier, the rise of women's labour
force participation should have a positive effect as well. Our real interest centres, however, on the impact of long-term
wage moderation or real wage growth, brought about through co-ordinated and centralized bargaining between the
social partners. Table 4.2 below estimates the impact of women's supply and real wage growth, controlling for the two
most important rigidities variables (tax wedge and employment protection).
Table 4.2. Explaining Cross-Sectional Employment Growth, 19791994

Employment growth, 19791994 (%)
Constant 1.692***
Employment protection -0.027
Tax-wedge -0.027***
Growth in women's labour supply 0.021***
Real wage growth -0.425***
Adjusted R2 0.736
Notes: *** indicates signicant at 0.001 level. N = 21.
Source: See Table 4.1.
The model provides a good t, explaining 74 per cent of the variance, even if it excludes the effects of macro-
economic policy.72 A high tax wedge subdues job growth, but the effect is quite small; similarly, the rise of women's
participation gives a push to employment growth, but, again, the effect is not huge. What really matters is wage
moderation. A two per cent real wage growth per year translates into almost one percentage point (0.85 per cent) less
job growth.
This will, in turn, inuence unemployment levels. To estimate this indirect, or double, causality on unemployment, I
introduce into the unemployment model in Table 4.3 a variable (r) empg which are the residuals obtained from
regressing employment growth on real wage growth. Table 4.3 estimates, additionally, the effect of restrictive
macroeconomic policy, via the change in real private consumption expenditures, 19905; of excess supply of low-
skilled workers, that is, the percent of the labour force with less than secondary education; and of women's
employment.
In Table 4.3 all the signs point in the expected direction. The indirect effect of real wage growth and employment,
(r)empg, is insignicant mainly because it is drowned out by the consumption variable.73 What emerges is that
unemployment levels are principally inuenced by macro-economic policy and by women's employment. The less
women work, and the more restrictive macro-economic policy is, the more unemployment. Labour market regulations
just do not matter for unemployment levels. Excess supply of low-skilled workers, a proxy for de-ruralization and de-
industrialization, adds to the unemployment
72
Our employment growth data cover a very long period, while typical macro-economic policy measures are short-to-medium term.
73
The two are intercorrelated (0.58). The variable (r)empg is signicant and quite strong (b = -0.351) when the consumption variable is omitted.
Table 4.3. Explaining Unemployment Levels, 1996

(log) Unemployment 1996
Constant 3.319***
(r) empga -0.228
Change in consumption -0.027**
Low-skilled stock 0.004
Women's employment -0.017**
Adj. R2 0.501
a
residuals from equation: [employment growth = 1.125a - 0.573b (real wage growth)], with t(b) = -3.39; and R2 = 0.418.
Notes: standard rigidities measures, such as the tax wedge, social benets, the minimum wage or earnings differentials are systematically
insignicant. ** indicates signicant at 0.05 level: *** at 0.01 level.
problem but its effect evaporates when held together with the female employment multiplier.74
What, then, can we conclude from this? Firstly, that Blanchard and, generally, the European Left may be right to claim
that overly restrictive macro-economic policies worsen Europe's unemployment problem. Secondly, that historical
circumstance plays a signicant role. Countries with an excess of low skilled labour can expect higher unemployment,
but this effect can be cancelled out if women work. The earlier conclusion seems still valid: encouraging housewives to
work by lessening the rigidities they face is probably a superior strategy to combat unemployment levels.
Explaining Unemployment Structure

So, rigidities have nothing directly to do with unemployment levels, but we already know that, at least in some cases,
they may inuence who happen to be the unemployed. In Chapter 3, I distinguished two groups who tend to be
disproportionally affected: youth and the low-skilled. The rigidities that inuenced the youth or low-skilled bias were,
for youth, employment protectionthe rigidities variable; and for the low-skilled: unemployment benet duration,
the tax wedge, the minimum wage, and employment protection (negatively).
74
Also women's employment and low-skilled stock are fairly highly correlated (0.55). The effect of low-skilled stock becomes signicant (b = 0.011) when women's
employment is omitted. De-composing variance explained indicates that women's employment is the single strongest determinant.
Table 4.4. Explaining Unemployment Outows, 1994

Unemployment outow
I-All II-All I-Youth II-Youth
Constant 46.692*** 45.126*** 47.899*** 59.444***
Employment protec- -1.008*** -0.977**
tion
Unemployment ben- -5.264*** -3.635***
et duration
Early retirement rate -0.208*** -0.335*** -0.370* -0.604***
Low-skilled worker -0.167** -0.272**
stock
Adjusted R2 0.641 0.660 0.458 0.490
Notes: Outows are measured as average monthly exit from unemployment, 1994.
Signicance levels: * = 0.10; ** = 0.05; *** = 0.01.
The theoretical problem we face is that some of these effects are most likely direct, while others should be indirect.
Long benet duration would be a direct effect, because it means that the unemployed stay so longer; the tax wedge and
the minimum wage similarly, because both will price out less productive workers. Employment regulation, however,
works indirectly, via its effect on ows between unemployment and employment, because in essence it provides added
protection to those already employed. Hence, it punishes the unemployed and rst-time job seekers. The key variable
here, in other words, is ows.
Table 4.4 predicts both total and youth unemployment outows. Minimum wages and the tax wedge are omitted
because they are systematically insignicant. The impact of unemployment benet duration is only relevant for
experienced workers, and has therefore been omitted for youth. Because of multi-colinearity problems, I estimate two
separate models in each case.
Employment protection depresses both total and youth outows considerably, as does the duration of unemployment
benets for the total.75 In fact, holding the other variables constant, each additional
75
Since employment protection lowers outows but has no effect on overall unemployment, it stands to reason that its effect on inows into unemployment should be quite
similar to outows. Separate analyses, not shown, suggest this to be roughly the case. The effect of employment protection on inows is signicant (at the 0.01 level) with a
coefcient of -0.07).
month of benet duration reduces monthly outows by 4 or 5 percentage points. As we would expect, outows
become systematically inferior the larger the stock of low-skilled workers isexcess supply makes it more difcult to
clear markets. The question then is whether offsetting exibilities, like early retirement, help. No, they clearly do not.
Indeed, rather than easing ows, early retirement adds to the problem! For every 10 percentage point additional male
workers, aged 5564, in early retirement we get 23 points less outow among the total unemployed, and 46 points
less among youth. It is possible that early retirement helps companies shed unwanted workers, but this obviously does
not mean that the unemployed will fare betterto the contrary.
Turning now to unemployment structure, and the low-skilled and youth bias, we proceed in a manner parallel to
earlier. Since the rigidity effect is primarily indirect, via the effect of employment protection on outows, we take the
residuals from a regression of outows on employment protectionseperately for the total and for youthand insert
them in our unemployment structure equations as the variable r(outow). As noted, unemployment benet duration, the
minimum wage, and the tax wedge were all potential direct effects in the case of the unskilled bias. I omit the two
former since they proved systematically insignicant in all estimations. In the case of the youth bias, I also omit the tax
wedge. In both estimations I include the early retirement variable, although its meaning differs in the two cases. Early
retirement should help limit the low-skilled bias because the potentially unemployed are, instead, moved out of the
labour market altogether. In the case of youth, early retirement might clear the labour market of older workers and
thus facilitate access to youth. But the main variable of interest in the youth model is clearly the offsetting exibility
that may derive from active labour market policies. Finally, each estimation includes the low-skilled stock variable.
The low-skilled model in Table 4.5 is clearly poorly specied. There are clearly unmeasured factors of substantial
importance, and the results are quite confusing. The indirect effect of rigidities via ows, and the direct effect of the tax
wedge are both basically nil. The underlying problem is that the bias, or lack thereof, of low-skilled unemployment is
driven by a logic that diverges from that of others. In fact, in Chapter 3 we found that high rigidities were related to low
levels of unskilled worker unemployment. More surprisingly, excess stock reduces the unskilled biasalthough the
elasticity is very smalland early retirement increases it. What exactly drives the low-skilled bias remains, therefore,
somewhat of a mystery.
There are two possible explanations, though. One lies in the precise composition of excess stock. Since, as we know, a
large share of services absorb low-educated workers, the fact that stock reduces the
Table 4.5. Explaining Unemployment Structure

Low-skilled unemployment bias Youth unemployment bias 1994
1994
Constant 1.968*** -2.479
r(outow)a -0.005 0.100**
Tax wedge -0.013
Low-skilld stock -0.009*** 0.042**
Early retirement 0.007 0.076**
ALMP (youth) 5.621
Adj. R2 0.368 0.539
a
residuals from equation: [unemployment outow (total) = 24.355a - 0.959b (rigidity)], with t(b) = -3.26); and R2 = 0.399. and equation:
[unemployment outow (youth) = 34.292a - 1.312b (rigidity)], with t(b) = -3.26); and R2 = 0.400.
Notes: ALMP is active labour market programme participation of youth as a percent of all in age group. ** = signicant at 0.05 level; *** at
0.01 level.
low-skilled bias may simply mean that they areat least in a number of countireswell-absorbed, probably at low pay
as happens in North America. In this case, a more relevant measure might be deruralizationjob loss in agriculture
during the 1980s. In fact, when we substitute agricultural job loss for low-skilled stock in our model, it is highly
signicant with b = 0.111.
The second possible explanation lies in the peculiar impact of employment protection on the unskilled. Their mobility
is reduced because protection implies long employment tenure. But even when formal employment protection is not
especially strong, low-skilled workers may none the less enjoy long tenure, either because of their age prole or because
they are concentrated in sectors with exceptionally strict seniority rules. If they cannot be laid off, there is of course the
early retirement option. But if early retirement is similar among the skilled and unskilled equally, the effect would be
neutralized. In fact, when we re-estimate the low-skilled unemployment bias model in Table 4.5 by substituting
tenuremeasured as the average number of years with same employerfor the tax wedge variable, the coefcient for
tenure (-0.080) is signicant, and the variable adds 12 per cent to the total variance explained.76
Turning to the youth bias, the results are at least closer to what one would expect: the indirect effect of rigidities is
signicant and fairly
76
I am unfortunately unable to estimate tenure and agriculture job loss simultaneously due to strong multi-colinearity.
112 References
strong, and an excess of low-skilled workers does seem to crowd out youthalthough not much. Again, early
retirement worsens the unemployment problem.77 The variable of principal interest, none the less, is active labour
market policy for youth, which measures the percentage of youth involved. The sign is oddly enough positive, but the
variable is not signicant. Hence, active labour market policiesas many others arguedo not appear to be an
effective antidote to youth unemployment.78
References
Blanchard, O. (1998), Thinking about Unemployment, Paolo Baf Lecture on Money and Finance, 16, October, 1998.
Rome: University of Rome.
Bruno, M., and Sachs, J. (1988), The Economics of Worldwide Stagation. Cambridge, Mass: Harvard University Press.
Calmfors, L., and Drifll, J. (1988), Bargaining structure, corporatism and macroeconomic performance, Economic
Policy, 6: 1361.
Esping-Andersen, G., (ed.) (1993), Changing Classes. Stratication and Mobility in Postindustrial Societies. London: Sage.
(1999), Social Foundations of Postindustrial Economies. Oxford: Oxford University Press.
Giersch, H. (1985), Eurosclerosis, Kiel Discussion Paper, no. 112. Institute fr Weltwirtschaftforschung, University of
Kiel.
Nickell, S. (1997), Unemployment and Labour Market Rigidities: Europe versus North America, Journal of Economic
Perspectives, 3: 5574.
and Layard, R. (1998), Labour market institutions and economic performance, Centre for Economic Performance
Discussion Paper, 407/September.
OECD, (1994), The OECD Jobs Study. Paris: OECD.
Saint-Paul, G. (1996), Exploring the political economy of labor market institutions, Economic Policy, 23: 263316.
Snower, D. (1997), The low-skill, bad-job trap, in A. Booth and D. Snower, (eds.) Acquiring Skills: Market Failures, Their
Stigler, G. (1956), Trends in Employment in Service Industries. Princeton NJ.: Princeton University Press.
Visser, J., and Hemerijck, A. (1997), A Dutch Miracle. Amsterdam: University of Amsterdam Press.
77
As would be expected agrarian job decline has no bearing on youth unemployment. Less obviously, neither does tenure.
78
This may be due to an endogeneity problem if exsisting high youth unemployment is what, in the rst place, drives governments to expand ALMPs. There is also the well-
known problem that activation programs artically remove people from unemployment statistics.
Part II National Variations
5 River Crossing or Cold Bath? Deregulation and
Employment in Britain in the 1980s and 1990s
Simon Deakin and Hannah Reed
Introduction
The British labour market has been a testing ground for economic experimentation since the late 1970s, when the
Conservative government of Margaret Thatcher set about dismantling the post-war consensus on economic and social
policy. The structural adjustment which then took place is often seen today as a necessary part of adaptation to the
changed circumstances of a globalized, service-orientated economy. Neo-liberal economic theories, which see
regulation of employment as a source of rigidities, support this contention. Empirical evidence for the claim was thin
for most of the 1980s and 1990s. However, the recent rise in the British employment rate, which occurred during a
period when employment was falling and unemployment rising in other European systems, has led some to reassess
the British experience. For them, Britain has crossed the river, and its example, similar to that of the USA, is one
which other EU countries might do well to emulate.
This chapter seeks to evaluate the British experience with specic reference to the economic impact of changes in
employment law and social security. The broad nature of regulatory reforms is outlined in section 2, The Dynamics of
Rigidity/Flexibility; and the nature of shifts in the structure of employment and unemployment in section 3, The
Outcomes: the Structure of Employment and Unemployment. To juxtapose legal and economic changes in this way is
one thing; as we shall see, establishing causal links between the two is a much more hazardous endeavour. The
conclusion nevertheless attempts an overview of what we know and do not know about the relationship between
regulation and economic outcomes.
116 Simon Deakin and Hannah Reed
Review of the National Debate and Literature

The issue of labour market exibility has had a high prole in Britain for some time. Over the last two decades,
successive governments have identied increased exibility in the labour market as being a key element in government
strategy to tackle unemployment and ination, and to enhance competitiveness (Department of Employment 1985).
Reversing many of the legislative interventions of the 1960s and 1970s, the Conservative government elected in 1979
embarked upon a wide-ranging programme of labour market reform which lasted until the mid-1990s (Table 5.1).
An important element within this policy was deregulation, in so far as that term is understood to refer to the removal
of certain protective rights which could be construed as rigidities. Such measures included the dismantling of
collective institutions which had previously underpinned the process of wage determination; the abolition of wage
councils which had been responsible for setting statutory minimum wages, and terms and conditions in low paying
sectors; the contraction of social insurance provisions for the unemployed and retired; and restrictions upon the scope
of employment protection legislation. However, in many ways deregulation is a positively misleading term to describe
what occurred in the 1980s and 1990s, since the effort to revive market forces involved extensive legal intervention,
albeit of a different kind from that traditionally used to regulate the labour market. Three forms of legal strategy have
been identied (Brown et al. 1997): a return to private law implied by the abolition of trade unions' immunities against
civil liability; the use of competition policy to discourage the setting and enforcement of labour standards through
collective bargaining; and the re-regulation of industrial action procedures through highly complex rules governing
strike notices and ballots.
The policies pursued during the 1980s and 1990s were strongly inuenced by neo-classical economic arguments
against labour market regulation. These led to calls for the dismantling of institutions of collective regulation, leaving
greater scope for the operation of the law of contract and tort (Hayek 1960, 1973; Epstein 1983, 1984). Social security
benets were perceived as raising the reservation wage above the market clearing rate for employment in the non-
union sector, thereby generating unemployment (Minford et al. 1985). Unfair dismissal laws were viewed as increasing
the costs for employers of hiring the unemployed and of increasing the power of the insiders, those in employment, to
exclusion of the outsiders, the unemployed (Lindbeck and Snower 1989). This led some commentators to advocate
the complete
Britain-River Crossing or Cold Bath? 117
Table 5.1. Legislative Developments Affecting Labour Market Flexibility, UK

Legislative Change Year Protective Deregulatory
Pre-1979
Regulation of notice of termination 1963
Statutory redundancy compensation 1965
Extension of earnings-related unem- 1965
ployment benet
Equal pay for men and women 1970
Unfair dismissal protection 1971
Regulation of agency work 1973
Prohibition of sex discrimination 1975
Strengthening of employment protection 1975
laws
Consultation over collective dismissals 1975
State earnings-related pension scheme 1975
Prohibition of race discrimination 1976
Post-1979
Extension of qualifying periods for 1979
employment protection
Restriction of industrial action 1980
Abolition of extension legislation 1980
Abolition of earnings-related supple- 1980
ment to unemployment benet
Widening of derogations for xed-term 1980
employment
Protection of employment on transfers 1981
of undertakings
Further restriction of industrial action 1982
Rescission of fair wages resolution 1982
Restriction of closed shops 1982
Extension of equal pay for men and 1983
women
Industrial action ballots 1984
Further extension of qualifying periods 1985
Limitation of powers to set minimum 1986
wages
Restriction of state earnings-related 1986
pension scheme
Increase in qualifying period for un- 1986
employment benet
Tightening of contribution conditions 1988
for unemployment benet
Compulsory competitive tendering in 1988
local government
Further restriction of post-entry closed 1988
shop
Repeal of working time controls 1989
Enlargement of disqualications for 1989
unemployment benet
Actively seeking work requirement for 1989
unemployment benet
Restriction of pre-entry closed shop 1990
Abolition of powers to set minimum 1993
wages
Strengthening of rights to consultation 1993
Further restrictions on industrial action 1993
Abolition of restrictions on Sunday 1994
trading
Job-seeker's allowance replaces unem- 1995
ployment benet
Extension of rights of part-time workers 1995
Prohibition of disability discrimination 1995
Implementation of EC Working Time 1998
Directive
Statutory national minimum wage 1998
removal of protective legislation and the restriction of social security payments (Minford et al. 1985; Addison and
Siebert 1991), but for various reasons this approach was never openly contemplated by either policy-makers or
politicians.
The overwhelmingly important measure of the success of deregulatory policies concerns the structure of employment
and unemployment, and, above all, the capacity of the system to generate additional employment opportunities. The
recent UK experience seems encouraging for proponents of exibility, because the UK economy has enjoyed
substantial job growth since 1993 in contrast to the pattern elsewhere in Europe. This pattern conforms to an image of
a free or deregulated labour market in which the employed may be less secure than they used to be, but where
opportunities to enter or re-enter the labour market are greater than would otherwise be the case. According to the
more optimistic scenarios, this expansion in labour market mobility should more than offset the growth in earnings
inequality which has accompanied the freeing up of market forces.
On closer inspection, it can be seen that the picture for the UK labour market is much more complex than this. For
example, unemployment, although fallingin Summer 1998 it stood at 6.3 per cent on the ILO denitionis
increasingly concentrated in workless households. In Spring 1998 nearly 18 per cent of non-pensioner households had
no adult in work, some 3.15 million households; in 1979 the equivalent gure was 1.2 million households (EPI 1999,
Table 5.2). There are also over 2 million individuals below pensionable age who, while not being classied as
unemployed, are economically inactive, either because they fail to meet the strict qualication criteria for
unemployment benets or because they are not seeking immediate work by reason of illness, disability or
discouragement (Table 5.3).
Moreover, the UK has also experienced growing earnings inequality since 1979 (Gregg and Wadsworth 1997a), with
the result that the gap between pay levels is currently greater than when comparable records were started at the end of
the last century (Joseph Rowntree Foundation 1995; Machin 1996; Marx 1999). For the lower paid, there is evidence of
barriers to upward mobility between occupational groups, and of limited opportunities for investment in skills and
training of the kind which a dynamic labour market might be assumed to generate (Gregg 1997).
There is not much doubt that, after nearly two decades of neo-liberal policies, the labour market environment is
favourable to employers in the sense that the legal or regulatory impact upon managerial decision-making is, by
international standards, a light one. Hence British employers are found to perceive labour market rigidities, such as
hiring rules and restraints on dismissal, to be much less of a problem than their
Table 5.2. Workless Households 19771998

Spring/Y- Workless households Working age adults in Children in workless All individuals workless
ear workless households households household
(Nos) (% of to- (nos) (%) (nos) (%) (nos) (%)
tal)
1977 1,189 8.2 1,687 5.8
1979 1,213 8.3 1,669 5.6
1981 1,567 10.9 2,365 7.8
1983 2,063 14.5 3,148 10.7
1984 2,285 15.3 3,481 11.3
1985 2,528 16.5 3,817 12.1
1986 2,576 16.7 3,872 12.2
1987 2,632 16.9 3,905 12.3
1988 2,538 16.0 3,653 11.4
1989 2,347 14.7 3,303 10.3
1990 2,364 14.6 3,298 10.2
1991 2,529 15.5 3,730 11.2
1992 2,863 17.5 4,049 12.7 2,373 18.7 6,422 14.6
1993 3,145 18.7 4,391 13.8 2,750 20.0 7,141 15.6
1994 3,275 19.1 4,525 14.2 2,869 20.7 7,394 16.2
1995 3,294 19.1 4,478 14.1 2,881 20.4 7,359 16.0
1996 3,337 19.3 4,575 14.2 2,887 20.4 7,462 16.0
1997 3,175 18.2 4,361 13.5 2,679 18.9 7,040 15.0
1998 3,148 17.9 4,273 13.2 2,705 19.0 6,978 15.1
Notes: = information not available.
Source: Labour Force Survey; EPI Employment Audit (various issues).
continental counterparts do (Emerson 1988; Grubb and Wells 1993; Beatson 1995). Conversely, the UK has until
recently scored very low in international evaluations of labour standards in the areas of minimum wages, working time
and job security (OECD 1994).
Independent assessments of the overall impact of regulatory change are more cautious, though, stressing the wide
range of factors which lie behind shifts in the structure of employment and unemployment, many of which are beyond
the immediate control of policy. Indeed, a large variety of methods and approaches have been adopted, such that there
is little or no consensus on the relationship between legal and regulatory change and economic outcomes. The
methods used range from formal modelling of equilibrium unemployment (Layard et al. 1991); longitudinal studies of
social and labour market mobility (Gregg 1997); case-study analyses of the behaviour of rms (Atkinson 1985; Pollert
1988; Hakim 1990; Hunter et al. 1993; Rubery and Wilkinson 1993) and individuals (McLaughlin et al. 1989); studies of
shifts in productivity (Nolan and Marginson 1990; Metcalf 1989; Nolan 1989); and analyses of the impact of legal
change on the industrial relations system and on economic performance (Brown and Wadwhani 1990; Deakin and
Wilkinson 1991; Beatson 1995; Brown et al. 1997). While there is general agreement that the labour market
experienced deep structural changes during a period when social policy itself was undergoing a major transformation,
there has been much less agreement on the relative importance of regulatory changes and on the direction of
causation.
It is too early to assess the effects of the policies adopted by the New Labour government elected in 1997, in part
because the policies themselves remain ambiguous. The government has formally committed itself to maintaining a
exible labour market, and, indeed, has undertaken to seize what opportunities it can to export UK-style exibility to
the rest of the EU. However, it seems that labour market exibility does not bear quite the same meaning for New
Labour as it did for the Conservative administrations which held ofce after 1979. In a reversal of neo-liberal policies,
the new government has put in place a statutory national minimum wage as well as introducing an Employment
Relations Act which promises to strengthen individual employment protection as well as reintroducing a legal
procedure for trade union recognition. The minimum wage is linked to a policy of making work pay by increasing the
differentials between incomes in and out of work, in which changes to the tax-benet system also have a role (Hills
1998: 27). The aim of improving the quality of the labour supply lies behind a number of government initiatives in the
area of education and training: Tony Blair's famous three priorities of education, education and education ow
explicitly from an analysis that both low productivity and growing inequality have roots in a workforce which is ill
equipped for the contemporary global economy (Hills 1998: 26). The single most important set of measures is
contained in the New Deal: a package of training and employment subsidization aimed at moving the longer-term
unemployedand in particular the younginto paid employment.
In short, what has replaced the market liberalism of the 1980s and 1990s is a third way the precise direction of which
remains, however, unclear. A leading exponent of the third way has described it as a framework of thinking and
policy-making that seeks to adapt social democracy to a world which has changed fundamentally over the past two or
three decades (Giddens 1998: 26). But at a concrete level, the Government has in fact left its options very wide,
without giving many clues as to where the third way will actually take us (Hills 1998: 33). Nevertheless, it could be
said that a common element in recent measures
is the use of a range of mechanismsregulation, changes to taxes and benets, and active labour market
policyaimed at steering the market through incentives, a somewhat different emphasis from the 1980s mantra of
lifting the burden.
The Dynamics of Rigidity/exibility: Trends in Labour Market

Regulation
The British system of labour law has traditionally provided a relatively restricted role for regulatory legislation in
comparison to collective bargaining. Hence even at the end of the 1970s, the high point of what Giddens (1998: 6) calls
old-style or classical social democracy, direct statutory regulation of wages and conditions of employment was
conned to certain groups, in particular the low-paid for whom a statutory minimum wage was in operation, set by
tripartite bodies, known as wages councils. Employment protection legislation, granting income and job security rights
to individual employees, was a late arrival: the principle of redundancy compensation was introduced only in 1965 and
unfair dismissal legislation as late as 1971the latter under a Conservative government.
The modern British system of social security has a longer history, dating from the legislative reforms of 19467 which
nally removed the disciplinary framework of the poor law. The 1940s reforms were based largely on the Beveridge
Report and established a form of social insurance provision. Social assistance was made available as a safety net for
those were not entitled to contributory benets. However, unlike models developed in certain other European states,
the British system of social insurance did not provide for earnings-related benets; in-stead, at-rate benets were paid
in return for at-rate contributions. Earnings-related supplements for unemployment and sickness benets and the
State Earnings-Related Pension Scheme (SERPS) were only introduced during the 1960s and 1970s.
To examine how this framework of regulation was altered in the 1980s and 1990s, we will consider a number of areas
in turn, beginning with employment protection legislation.
Employment Protection Legislation

The Conservative governments of the 1980s and 1990s introduced a number of measures aimed at reducing what they
saw as the negative effect of protective legislation on labour exibility. In 1980 legislation
removed a provision placing the burden of proving fairness in dismissal on the employer, and made it possible for an
employment tribunal to take into effect the size and administrative resources of the employer when assessing whether
dismissal was reasonable, thereby providing some protection for smaller rms. Employees in small rmsthose with
fewer than twenty employeesceased to have the right to receive written particulars of disciplinary procedures. The
period of qualifying service was extended from six months to one year in 1979 and to two years in 1985. However,
most of the substantive provisions of the unfair dismissal legislation remained intact throughout the period of
Conservative ofce; in fact, in this area the impact of deregulatory policies was curiously muted.
In part this was because the scope of employment protection legislation was limited to begin with. The self-employed
were completely excluded, as were, in practice, many casual workers and agency workers whose employment status
was unclear (Deakin and Morris 1998: ch. 3). Provision was made to allow xed-term employees to waive their
statutory rights to dismissal protection. Part-time employees were subject to exclusions: those working less than eight
hours per week did not qualify for basic protection at all, and those working between eight and fteen hours had to
have ve years continuous employment as opposed to the normal two.
It was only in 1994 that the courts ruled that the part-time work thresholds were indirectly discriminatory on grounds
of sexthe great majority of part-time workers affected by them were, and are, womenand so contrary to the EC
Directive on Equal Treatment in Employment of 1976. In 1995 they were formally repealed by statute, a major change
in employment law. The Employment Relations Act 1999 envisages a further broadening of the scope of employment
protection legislation to include certain economically dependent workers whose status as employees is in doubt, as
well as new restrictions on waiver clauses in xed-term contracts. It will also reduce the qualifying period for general
unfair dismissal protection from the present two years to one year.
Nevertheless, remedies for unfair dismissal remain weak. The principal remedy in practice for an unfair dismissal was
and is compensation, and amounts awarded are not high. In most cases there are statutory caps to awards, although
were recently raised considerably by statutory instrument. But even then, average awards have long been well below
the statutory limits. This is a reection of the low earnings expectations of successful unfair dismissal applicants, and
the absence of an exemplary or punitive component within the compensation ordered by tribunals. For example, in
199596 the median award for
unfair dismissal cases was 2,499, at a time when the maximum poss-ible award for a case not involving discrimination
or anti-union grounds was around 17,000. The median weekly wage for full-time workers in the same year was 313
which translated into a yearly equivalent of 16,276. Re-employment, while available in principle to the tribunal as an
option, is awarded only in around 5 per cent of all cases of successful unfair dismissal claims and in only 1 per cent of
those which proceed to a contested hearing.
Related legislation requires employers to pay redundancy compensation to employees dismissed on economic grounds,
and these grounds are restrictively dened. However, this has not been much of a burden on employers since, until the
late 1980s, statutory redundancy payments have been subsidized by the state through the social security system.
Moreover, contractual redundancy payments which amount to more than the statutory minimum continue to qualify
for corporation tax relief.
It is generally agreed that the effect of redundancy compensation legislation was to encourage voluntary or agreed
redundancies in the 1970s and 1980s, speeding up the process of restructuring in coal, steel, shipbuilding and other
rapidly declining areas of heavy industry. In this sense, the legislation may be said to have enhanced exibility. By
contrast, legislation which implemented the EC Collective Redundancies Directives of 1975 and 1992, requiring
employers to inform and consult with representatives of the workforcenormally but not always ofcials of a
recognized trade unionwhen proposing large-scale redundancies, has had little impact on core managerial
prerogatives (Daniel 1985).
Hiring and the Form of the Employment Contract

The evidence concerning the impact of employment protection legislation on hiring processes and in particular on the
growth of exible or non-standard forms of employment suggests that there are a number of complex and
overlapping effects. It is necessary, rstly, to consider the role played by the generally permissive regime for hiring.
Employment legislation in Britain has made few inroads on the principle that the parties to the employment
relationship are free to choose from a number of different forms. There is no tradition, for example, of requiring
employers to present a formal justication for the adoption of part-time or xed-term working arrangements. There
has been no ban on the use of agency labour, nor are agency workers required to have employee status for the purpose
of employment protection legislation. To this extent, we would expect the British system to exhibit a higher degree
of use of exible employment than other European systems where regulation is more restrictive.
However, this is offset by a second general feature of the British system, which is the generally weak level of protection
granted by legislation to the standard or traditional employment relationship. As we have just seen, protection is
largely procedural rather than substantive, and reinstatement of dismissed employees is extremely rare. On this basis,
we would expect employers to have only limited incentives to adopt forms of employment which fall on the margins of
employment protection law.
Policy-makers have come to accept that there is a complex relationship between micro- and macro-level rigidity, and
that the two may be involved in a trade-off. Hence an Employment Department (ED) study (Beatson 1995) found
evidence of greater micro-level exibility in the UK in the form of an increase by rms of part-time, temporary and
self-employment, as well as an increase in functional exibility, or the adaptability of workers within enterprises, during
the 1980s. At the macro-level, by contrast, the study concluded that there was still evidence of substantial wage rigidity,
and in particular a lack of responsiveness of wages to changes in levels of unemployment. The ED's analysis also
accepted that systems like Germany, Japan and the Nordic countries succeeded in achieving a high level of macro-level
wage exibility, despite the presence of apparent rigidities at the micro-level of the employment relationship.
Wage Determination
Several steps were taken during the 1980s and 1990s to remove the statutory regulation of wage determination. Firstly,
fair wages legislation and minimum wage legislation was repealed. Arbitration procedures of a kind similar to extension
laws in other systems, which required non-unionised enterprises to observe the basic rates laid down in sector-level
collective agreements, were repealed in 1980, and the Fair Wages Resolution, which bound government contractors to
observe minimum labour standards, was rescinded in 1982. The powers of the wages councils were restricted in 1986
before they were completely repealed in 1993. Prior to their abolition, the wages councils had covered 2.5 million
workers in various trades. Their repeal left the Agricultural Wages Boardcovering 220,000 agricultural workersas
the only body with the power to set a legally binding minimum wage. However, with the election of a new Labour
government in 1997 this policy went into reverse. A Minimum Wage Act was adopted in 1998 and from April 1999
a new statutory minimum wage came into effect, set initially at 3.60 an hour for adult workers with lower rates for
younger workers.
The second change made in the 1980s and 1990s related to statutory developments aimed at limiting the capacity of
trade unions to organize industrial action in defence of terms and conditions of employment. The effect of these
changes was to reduce the power of trade unions to use economic pressure to regulate the external labour market; it
was no longer possible, for example, to organize industrial action with the aim of putting pressure on a supplier or
customer of the primary employer to recognize a trade union (Brown et al. 1997).
Thirdly, steps were taken to encourage the decentralization of collective bargaining. Local government authorities and
health authorities were required to outsource certain services (CCT), which in practice often led to the modication or
abandonment of sectoral collective bargaining; in the public health and education sectors, the introduction of quasi-
markets had a similar effect. In the private sector, changes to industrial relations law indirectly hastened the break-up
of sectoral agreements, by making it more difcult for trade unions for mount national-level disputes.
Fourthly, employers were given greater leeway to make their own arrangements for pay determination at the company
or establishment level. Effective enforcement of the closed shop or union monopoly was made impossible by changes
to the law relating to unfair dismissal. All industrial action in support of the closed shop, including action against the
primary employer, became presumptively unlawful. Litigation claried the right of employers to introduce personalized
or individualized contracts of employment following the abrogation of collective agreements.
In all these ways, the 1980s and 1990s represented a turning point in the system of wage determination, which came to
resemble much more the decentralized and fragmented US system of workplace bargaining than the predominant
European model of sectoral bargaining. The 1999 Employment Relations Act, while it reintroduces a form of statutory
recognition procedure which could assist the growth of collective bargaining, is only capable of having this effect at the
workplace or company level; it would do nothing to reinstate multi-employer bargaining.
Working Time
There is only minimal statutory regulation of the working day, week and year in Britain. Legislation inherited from the
nineteenth century model
of the Factories Acts, which set maximum working hours for women and young children working in industry, was
repealed in 1989 (Deakin 1990). More importantly, the decline of sectoral collective agreements has removed working
time protections for many employees. Sectoral agreements were an effective mechanism for the reduction of the basic
working week to thirty-nine from forty hours as recently as the early 1980s. However, working time is now increasingly
subject to arrangements made at company level which may or may not involve collective bargaining. In the absence of
a legal basic working week, and with the decline in collective bargaining coverage, it is possible for employers to avoid
paying overtime altogether.
Much of this is set to change since in 1999, after several years' vacillation, the provisions of the 1994 EC Directive on
Working Time were nally incorporated into UK labour law. However, maximum use was made of the exceptions and
derogations allowed by the Directive.
Unemployment Compensation and the Tax-Benet System

The income gap between those in employment and those relying on social security widened signicantly in the 1980s.
During this decade the extent to which unemployment benet compensated for lost earnings fell by 30 per cent for
single men on average earnings and more for married men with children (Deakin and Wilkinson 1991). This was
largely due to the abolition of earning-related supplements and child dependency additions, and the failure to uprate
short-term benets in line with earnings. Taxation of unemployment benet, introduced in the early 1980s, also
reduced its value for individuals with irregular working patterns.
Signicant changes were also made to the rules on eligibility for social security which have reduced entitlement to
benet and increased the stringency of the job search test. Firstly, the effective qualication period for contributory
benets was extended and in 1988 those under eighteen were excluded from receiving non-contributory
unemployment benets. Secondly, while social security provision in Britain has always contained an element of
work discipline, in the late 1980s the job search duties of unemployed claimants were increased. As a result, claimants
were required to show that they were actively seeking work, and the range of jobs which a claimant could be required
to accept was widened to include work with lower wages or hours.
Additional requirements were introduced with the jobseeker's allowance which replaced unemployment benet in
1995. Entitlement to contributory benets was limited to six monthsit was previously twelve monthsafter which
the claimant is entitled only to the meanstested
allowance. Further, the claimant is now required to enter a job-seeker's agreement, which can specify not only the
lowest wage for which the claimant should be willing to work, but which also imposes a duty on the claimant to
comply with any reasonable job-seeking direction. Failure to comply can result in a reduction in, or the total withdrawal
of, benet for up to twenty-six weeks.
Increasingly, benets have aimed to supplement individual and/or household income from low-paid employment.
Family credit (1985) introduced benets of this kind which were paid net of taxes and national insurance contributions
with a long taper by which benet is only gradually reduced as wages rise above a certain applicable amount. To
qualify, the claimant had to show that he or she was normally engaged in paid employment for initially twenty-four and
then sixteen or more hours per week. Many of the essential features of this system have been retained under the new
working family tax credit introduced by the new Labour government and implemented from October 1998.
Expenditure on family credit was 494 million in 1989; in 1996, it was 1.7 billion. In 1989, just after the benet was
introduced, there were 285,000 recipients; by August 1998, there were approximately 782,000 recipients of whom
393,000 had partners, and whose dependants totalled 1,577,000. At this time, the average weekly payment of family
credit was 59.20 and the average weekly income from wages of family credit recipients was 115 (DSS 1998).
The Outcomes: The Structure of Unemployment and Employment

Although there is general consensus that the labour market underwent deep structural reforms during the 1980s and
1990s, there is less consensus as to the relative importance of each of the regulatory changes or indeed of their
respective effects on levels of unemployment and employment. Nevertheless, many commentators both in Britain and
in other countries have pointed to the signicant decline in UK unemployment in the 1990s, which stood at 6.3 per
cent on the ILO denition in 1998; low by European standards. Furthermore, although during the rst two to three
years of the current recovery the UK encountered jobless growth, since 1993 it has experienced a substantial rise in
employment by comparison to other European systems. During the four years up to Spring 1997, the number of
employed rose by over 1 million, equivalent to an increase of 4.25 per cent on LFS gures, and employment
continued to rise, showing an increase of 306,000 in the year up to Spring 1998 (LFS). On closer examination,
however, it is clear that certain trends within the UK labour market are not all favourable. It is also the case that there
is only tenuous evidence linking deregulatory policies to changes in the structure of employment and unemployment.
Unemployment
Even though the UK economy underwent signicant exibilization in the 1980s and 1990s, the level of
unemployment did not consistently fall during this period, but rather became more volatile (Alogoskous et al. 1995).
Unemployment swung from 6.4 per cent in 1980 to 12.4 per cent in 1983, down to 6.8 per cent in 1990, up again to 10
per cent in 1993 and down to 4.6 per cent on the claimant count or 6.3 per cent on the ILO count in 1999.
Further, the proportion of working age population experiencing at least one spell of claimant unemployment did not
decline during the early to mid 1990s as compared to the mid to late 1980s (Robinson 1997; Table 5.4). This nding is
supported by evidence from the British Household Panel Survey which shows that between 1991 and 1994, 71 per
cent of all men were in continuous employment while 29 per cent experienced at least one spell out of work over all
four years (Gosling et al. 1997). Redundancy rates, however, were higher in 1996 than those witnessed in the late 1980s
(Taylor and Booth 1996), although they have declined considerably since the early 1990s (Table 5.5).
While the proportion of the working population experiencing unemployment has remained stable, the incidence of
unemployment has become more evenly spread both in terms of age, industries, occupations and earning deciles in the
early 1990s when compared with the mid-1980s (Robinson 1997). However, as noted earlier, unemployment in the
UK has become increasingly polarized by household, a trend which can be traced back to the late 1970s. By 1990 the
proportion of workless households in the population was double that of 1975; since then the gure has risen, sharply
peaking at 18.7 per cent in 1996 and declining slightly to 17.9 per cent in spring 1998 (Table 5.2). In 1998, 19 per cent
of children, or over 2.7 million, lived in workless households. In addition, nearly 60 per cent of workless couples with
children had been without work for three or more years, creating a signicant dependency on welfare provision.
There is an apparently close relationship between the level of unemployment and the incidence of long-term
unemployment (Machin and Manning 1998). In the UK, as the labour market has recovered towards the end of the
1990s, long term unemployment has fallen. The number
Table 5.3. Labour Market Status Data

All
Year Total in em- Employment ILO unem- ILO unem- Total eco- Economic in-
ployment rate (%) ployed (000s) ployment rate nomically inactivity rate
(000s) (%) active (000s) (%)
1984 22,904 69.0 3,084 11.9 7,183 21.7
1990 25,654 75.2 1,890 6.9 6,551 19.2
1991 25,120 73.5 2,321 8.5 6,737 19.7
1992 24,461 71.4 2,706 10.0 7,078 20.7
1993 24,205 70.6 2,871 10.6 7,218 21.0
1994 24,384 71.0 2,686 9.9 7,281 21.2
1995 24,634 71.5 2,415 8.9 7,397 21.5
1996 24,897 72.0 2,294 8.5 7,375 21.3
1997a 25,961 72.8 2,068 7.4 7,656 21.5
1998a 26,267 73.4 1,788 6.4 7,747 21.6
Men
1984 13,599 78.2 1,840 11.9 1,954 11.2
1990 14,745 82.7 1,106 7.0 1,983 11.1
1991 14,320 80.1 1,452 9.2 2,099 11.7
1992 13,744 76.8 1,812 11.6 2,345 13.1
1993 13,497 75.3 1,936 12.5 2,496 13.9
1994 13,627 75.9 1,780 11.6 2,557 14.2
1995 13,798 76.6 1,569 10.2 2,650 14.7
1996 13,890 76.8 1,503 9.8 2,689 14.9
1997a 14,503 77.7 1,318 8.3 2,845 15.2
1998a 14,695 78.4 1,104 7.0 2,945 15.7
Women
1984 9,305 59.0 1,244 11.8 5,229 33.1
1990 10,908 67.1 784 6.7 4,568 28.1
1991 10,800 66.2 869 7.4 4,638 28.4
1992 10,717 65.6 894 7.7 4,733 29.0
1993 10,708 65.4 935 8.0 4,722 28.9
1994 10,757 65.7 906 7.8 4,724 28.8
1995 10,835 66.0 846 7.2 4,746 28.9
1996 11,006 66.8 791 6.7 4,686 28.4
1997a 11,458 67.3 750 6.1 4,811 28.3
1998a 11,573 67.8 684 5.6 4,802 28.1
a
UK gures. Other gures are for GB.
Notes: LFS 199498 seasonally adjusted. For all persons of working age.
Source: See Table 5.3.
unemployed for more than twelve months fell from about 845,000 people, or 38 per cent of the unemployed, in
autumn 1996 to about 625,000, or 33 per cent, in autumn 1997 (Machin and Manning 1998). Although these gures
suggest an improvement in the level of long-term unemployment, the rates experienced in 1997 were similar to those
in the late 1980s, and the long-term unemployed still represent a majority of the total unemployed (Table 5.6).
Another reason identied by commentators for the persistence of long-term unemployment is the decline in outow
rates from unemployment, although it should not be assumed that this decline has primarily affected the long-term
unemployed (Machin and Manning 1998). For example, while the rate of outow from unemployment into
employment in the UK in the mid 1990s improved and was comparable to that observed at the end of the 1980s, it is
signicantly lower than that experienced at the end of the 1970s, before labour market reforms were instituted (Table
5.7).
Alongside long-term unemployment, there is also evidence in the UK of a high degree of recurrent unemployment
among the bottom-end of the labour market. While unemployed workers do not seem less likely to nd work again,
around 45 per cent of people making a fresh claim for jobseeker's allowance in 1997 had made a claim in the previous
six months (Gregg et al. 1998).
The nature of the relationship between the changing structure and composition of unemployment and the changes to
the social security system remains, at best, uncertain. At a theoretical level it is widely assumed that generous social
security benets raise the equilibrium unemployment rate or NAIRU (Layard et al. 1991). However,
Table 5.4. Number of People Making at Least One Claim for Unemployment-Related Benets
Five year period Millions % of working age population
198589 10.5 27.8
198690 10.2 27.1
198791 10.3 27.3
198892 10.3 27.4
198993 10.4 27.6
199094 10.5 27.8
199195 10.6 28.0
Source: JUVOS (Joint Unemployment and Vacancies Operating Statistics System); ESRC Data Archive, University of Essex, as analysed by
Robinson, 1997.
Table 5.5. Redundancies and Redundancy Rate

Year redundancies ('000s) Redundancy rate per thousand employees
1977 158 7.2
1978 173 7.7
1979 187 8.3
1980 494 22.0
1981 532 24.9
1982 398 19.0
1983 324 15.8
1984 245 11.8
1985 235 11.1
1986 238 11.4
1987 144 6.8
1988 108 5.0
1989 144 (4.7) 6.4
1990 181 (7.0) 8.1
1991 391 (11.1) 17.8
1992 324 15.1
1993 262 12.3
1994 205 9.6
1995 220 10.2
1996 207 9.4
Notes: Redundancy rate is based on statutory returns (ES955) to Department of Employment. This measures conrmed redundancies and
comprises a consistent series between 1977 and 1991. After 1990, published estimates are based on individual returns from the Labour
Force Survey which are on average 30% higher. The difference is due to the exclusion in the former series of redundancies of 10 or less
from any one establishment. The gures in parentheses for the years 19891991 refer to calculations made under the method used for the
former series.
Source: Employment Gazette, Labour Market Trends, various editions.
empirical investigation has failed to nd evidence of a strong relationship between unemployment related benets and
unemployment duration (Atkinson and Mogensen 1993). There are, moreover, few empirical studies which have
directly sought to assess the impact of changes in the rules relating to social security entitlement of the kind which took
place in Britain in the 1980s and 1990s. Some analyses have ascribed the size of the fall in unemployment in the
middle 1980s and its timing, to the tightening of social security administration brought about in 1986 by the Restart
programme, under which the long-term unemployed were subject to regular monitoring by employment service
ofcials to assess
Table 5.6. Epi Joblessness Indicators, 19841997

Year Season More than 6 Months Unemployed
('000s)
1984 Spring 2033
1985 Spring 1941
1986 Spring 1887
1987 Spring 1787
1988 Spring 1406
1989 Spring 1056
1990 Spring 904
1991 Spring 1051
1992 Spring 1539
Summer 1649
Autumn 1671
Winter 1731
1993 Spring 1796
Summer 1780
Autumn 1724
Winter 1704
1994 Spring 1677
Summer 1622
Autumn 1496
Winter 1430
1995 Spring 1441
Summer 1410
Autumn 1320
Winter 1284
1996 Spring 1302
Summer 1242
Autumn 1170
Winter 1124
1997 Spring 1076
Source: Labour Force Survey, 198497 (not seasonally adjusted).
their availability for work (Layard et al. 1991; Dicks and Hatch 1989). However, there are problems with the hypothesis
that incentives for the unemployed have been improved by these means. Other research suggests that a major
disincentive for the unemployed is the precariousness of work and income in most of the jobs which are on offer to
them. Entering employment often means giving up a regular benet income, in return for an irregular wage income
and insecure employment, without the guarantee that net income from employment will be higher than that from
benets (McLaughlin et al. 1989).
Further disincentives arise from the uneven geographical dispersion
Table 5.7. Probability of Movement to and from Different Employment States (%)
Year Employed Unemployed Inactive
From From Total From From Total From in- From in- Total
employed employed outow unem- unem- outow active to active to outow
to unem- to inac- rate ployed to ployed to rate employed unem- rate
ployed tive employed inactive ployed
All
1977 2.9 3.1 6.0 47.1 8.2 55.3 16.4 5.1 21.5
1979 2.5 3.2 5.7 43.4 13.4 56.8 15.6 4.2 19.8
1981 5.3 3.5 8.8 28.0 9.9 37.9 15.2 7.1 22.3
1983 4.3 4.0 8.3 29.3 13.8 43.1 13.0 5.8 18.8
1984 4.0 3.9 7.9 30.6 15.8 46.4 14.7 6.7 21.4
1985 3.7 3.0 6.7 30.1 17.1 47.2 16.8 7.5 24.3
1986 3.8 3.2 7.0 30.1 18.6 48.7 17.1 7.7 24.8
1987 3.9 3.1 7.0 33.3 16.4 49.7 17.7 7.2 72.9
1988 3.1 2.7 5.8 37.5 16.5 54.0 18.8 6.8 25.6
1989 2.9 2.7 5.6 38.7 16.0 54.7 19.4 6.1 25.5
1990 3.1 2.8 5.9 39.5 15.1 54.6 19.1 6.5 25.6
1991 4.4 4.2 8.6 34.8 15.8 50.6 14.3 6.3 20.6
1992 4.2 5.3 9.5 31.0 21.6 52.6 13.3 4.3 17.6
1993 3.7 5.0 8.7 30.4 21.2 51.6 12.2 4.3 16.5
1994 2.9 5.1 8.0 33.1 20.9 54.0 12.9 4.3 17.2
Men
1977 3.0 1.1 4.1 46.5 5.4 51.9 24.9 5.0 29.9
1979 2.5 1.4 3.9 43.8 9.0 52.8 22.3 5.1 27.4
1981 6.0 1.4 7.4 25.7 8.2 33.9 23.3 12.6 35.9
1983 4.8 2.5 7.3 27.4 12.7 41.1 16.4 7.1 23.5
1984 4.2 2.4 6.6 28.8 13.8 42.6 17.0 7.2 24.2
1985 3.9 1.6 5.5 27.4 15.3 42.7 20.4 8.5 28.9
1986 4.1 1.9 6.0 27.4 17.2 44.6 20.5 8.5 29.0
1987 4.2 1.9 6.1 30.5 14.7 45.2 20.2 8.0 28.2
1988 3.2 1.3 4.5 35.0 14.0 49.0 21.6 7.4 29.0
1989 3.0 1.4 4.4 36.1 14.0 50.1 21.6 6.1 27.7
1990 3.3 1.4 4.7 36.0 12.5 48.5 20.7 7.1 27.8
1991 5.2 2.6 7.8 30.6 13.9 44.5 14.2 6.9 21.1
1992 5.3 3.7 9.0 27.8 17.6 54.4 13.4 5.0 18.4
1993 4.7 3.6 8.3 28.0 17.8 45.8 12.0 5.1 17.1
1994 3.6 3.8 7.4 30.8 16.8 47.6 12.1 4.7 16.8
Women
1977 2.8 6.4 9.2 48.3 14.0 62.3 14.2 5.1 19.3
1979 2.5 6.2 8.7 42.7 21.4 64.1 13.8 3.9 17.7
1981 4.2 6.7 10.9 33.0 13.7 46.7 13.1 5.7 18.8
1983 3.6 6.3 9.9 33.9 16.6 50.5 11.9 5.4 17.3
1984 3.5 6.0 9.5 34.8 20.6 55.4 13.8 6.5 20.3
1985 3.3 5.0 8.3 36.0 20.9 56.9 15.5 7.1 22.6
1986 3.4 4.9 8.3 35.8 21.6 57.4 15.8 7.4 23.2
1987 3.5 4.7 8.2 39.5 20.1 59.6 16.7 6.9 24.3
1988 2.9 4.6 7.5 42.7 21.8 64.5 17.7 6.6 24.6
1989 2.7 4.5 7.2 43.8 20.1 63.9 18.5 6.1 24.7
1990 2.8 4.6 7.4 46.5 20.3 66.8 18.4 6.3 20.4
1991 3.3 6.3 9.6 43.3 19.5 62.8 14.4 6.0 20.4
1992 2.7 7.2 9.9 37.3 29.1 66.4 13.2 4.0 17.2
1993 2.3 6.7 9.0 35.4 28.4 63.8 12.3 3.9 16.2
1994 2.0 6.7 8.7 37.7 29.4 67.1 13.2 4.1 17.3
Source: Labour Force Survey, 197794, Spring quarter.
of employment opportunities. The incidence of unemployment appears to vary considerably geographically (EPI
1998). Some travel-to-work areas, in particular suburban areas and semi-urban areas, have tight labour markets, there
being as many job vacancies as job seekers. However, some travel-to-work areas, especially inner cities, traditional
industrial conurbations and remote rural and seaside tourist areas remain relatively depressed. In these areas job
vacancies are comparatively scarce, with up to four active job seekers per vacancy. Many of the more prosperous areas
and the more depressed areas are located in the same regions.
Employment
Although employment levels having been rising, during the 1990s levels of economic inactivity also expanded
signicantly in the UK, rising from 19.6 per cent in Spring 1990 to 21.6 per cent in Spring 1998 (Table 5.3). This rise
has been totally accounted for by men, with levels of male economic inactivity increasing by nearly 12 per cent between
1992 and 1996, equivalent to approximately 300,000 men (LFS). In contrast, economic inactivity fell amongst women
by 2.1 per cent or 100,000 women during the same period. Between 1992 and 1996 economic activity among women
increased by 2.28 per cent, equivalent to 264,000 women entering the labour market. This suggests that the proportion
of women wanting to work continues to rise as the economic recovery matures.
The relationship between overall employment levels and the numbers employed in exible or non-standard work is
also important. While in the early to mid-1990s only a small minority of jobs created were full-time and permanent, the
proportion has increased as the economy has recovered. In the year from winter 1996 to winter 1997, over 80 per cent
of all new jobs created were in this category. This was the case for over 86 per cent of all new jobs for men, although
only for 44 per cent of new jobs for women, with 27 per cent of new jobs for women being part-time and permanent,
and 17 per cent being part-time temporary jobs.
Overall, however, women have fared better than men in the labour market in the 1990s. By summer 1998, the
employment rate for women, 68.1 per cent, had risen above the pre-recession peak of 67.1 per cent in 1990, whereas
the employment rate for men, 78.7 per cent, was 4 per cent below its peak of 82.4 per cent (Table 5.3). The decline in
male employment rates between 1990 and 1996 was concentrated amongst young men and those of prime age (2549).
Since 1993, employment growth has been faster for men aged 5064 than those aged 2549. By contrast, the greatest
increase in employment for women has been among those aged 2534.
One of the main reasons why women have fared better in the labour market in the 1990s appears to be the effects of
the restructuring which took place in different sectors of the labour market. It is evident that the restructuring which
commenced in the 1980s has continued into the 1990s, with a decline in employment being concentrated in agriculture
and the utilities, such as energy and waterwhich are traditional employment areas for menand an expansion in the
service and business sectorswhich employ a relatively higher proportion of women.
Other, non-legal factors, such as the waning of cultural attitudes which had once underpinned the male breadwinner
family, and the pressure on low-income households to have multiple sources of incomes, may be important in
explaining why the employment rate for women in Britain has increased while it remained static in many other EU
countries. However, the regulatory framework does not seem to offer any obvious explanations for this aspect of the
British experience. Since many married women with incomplete records of social security contributions do not qualify
for unemployment-related benets, the cuts in these benets which took place after 1979 would be largely irrelevant to
their situation. As we explain further below, the system of social security taxation encouraged part-time employment at
very low rates of pay but discouraged movement into higher pay brackets and cannot be said to have increased overall
employment, as opposed to having an inuence on the form in which labour was contracted (Dickens 1992).
Job Tenure and Turnover Rates

The rise or fall of job tenure is one measure of mobility and therefore potentially of exibility within a labour market.
Over the twenty years since 1975 median job tenure has declined by 12 per cent (Gregg and Wadsworth 1996), with a
signicant decline taking place during the early 1980s as a result of rising unemployment and restructuring, especially in
the manufacturing sector. Since 1985, it appears that tenure has moved counter-cyclically. While overall changes in
tenure have been modest, signicant different trends emerge when the gures are broken down by gender, with men
suffering a decline (Robinson 1997), whereas median tenure for women since the mid-1980s has grown. It is likely that
this change can largely be explained by women taking advantage of statutory rights to return to work after pregnancy
or connement, although a decline in fertility or childbirth may also help to explain the trend.
Movements within the labour market play an important role in enabling the economy and rms to respond to shocks
or shifts in
demand. Indeed every year more than 6 million people either leave a job, nd a new job or do both. Those moving
from unemployment into work are generally being required to accept lower status and lower paid jobs than previously
experienced with anything up to a 10 per cent pay cut on their return to the labour market. The pay gap between
workers who retain their jobs and those who are displaced is around 15 per cent (Gregg et al. 1998). Job-to-job moves
within the labour market are generally resulting in longer hours, with a clear switch being made by movers from part-
time jobs into full-time jobs. However, there is a signicant increase in the use of temporary contracts. Indeed a pattern
of mobility appears to be emerging with workers moving in two stages, rst from part-time work or government
training schemes into full-time temporary work, followed by a second stage move into permanent full-time work
(Gregg and Wadsworth 1997).
Non-Standard Forms of Employment and Flexible Working Time

Over the last two decades, there has been a decline in full-time, permanent employment and a corresponding increase
in the use of atypical forms of work in the UK. The reasons why employers use exible working practices appear to
vary, but mainly include traditional reasons, such as the need to meet peaks in demand and the temporary unavailability
of full-time, permanent employees (Wood and Smith 1987; McGregor and Sproull 1991; Beatson 1995). Growth in
part-time work began in the 1960s and 1970s, prior to the general trend towards deregulation in the 1980s. The only
signicant rise in atypical work since 1979 has been the rapid increase in self-employment, which was concentrated in
the early 1980s. Since 1990 the growth in self-employment has almost ceased. During the mid-1990s there has also
been an increase in the use of xed-term contracts especially in the public administration, health and education.
Although temporary employment, on LFS gures, still constitutes only a small proportion of working practices in
Britain (Table 5.8), there is some evidence to suggest that the LFS under-estimates the numbers in xed-term
employment (Burchell et al. 1999).
Despite the relative decline in standard work, no clear pattern emerges to link the growth of non-standard
employment with changes in employment protection legislation. The timing of legal changes bears no obvious relation
to the emergence of non-standard work. In any case, given the comparatively weak level of regulation of full-time and
indeterminate-duration employment, it would be inappropriate to regard atypical forms of work as necessarily
exible in comparison to the rigidity of the standard employment relationship.
Table 5.8. The Changing Composition of All Employment, 19791997

1979 1984 1990 1997
Full-time employees 76.7 69.7 67.1 65.2
Full-time perma- 67.4 64.8 61.7
nent
Full-time tempo- 2.3 2.3 3.5
rary
Part-time employees 16.1 18.8 19.4 22.2
Part-time perma- 16.5 17.2 19.2
nent
Part-time tempo- 2.3 2.2 3.0
rary
Full-time self-em- 6.5 9.4 11.3 9.9
ployed
Part-time self-em- 0.7 1.9 2.1 2.6
ployed
Note: Excluding those on Government schemes and unpaid family workers.
Source: Labour Force Survey, Spring; EPI Employment Audit (various issues).
The limited signicance of employment protection law for hiring decisions is borne out by empirical studies, which
suggest that employers do not regard this legislation as imposing hiring costs which would amount to a signicant
disincentive to hiring (Daniel and Stilgoe 1978; Brown et al. 1981; Evans et al. 1985; Dickens et al. 1985). The same
studies found evidence that the introduction of unfair dismissal legislation increased the degree of screening of
applicants and, to that extent, led to fewer subsequent dismissals.
Studies of the growth of exible forms of work suggest that the tax-benet system has had a much more powerful
effect on the form of hiring than employment law has. This effect is the consequence of the differential tax treatment
of certain forms of non-standard employment. Hence, the growth of both self-employment and part-time work at
low rates of weekly pay in the 1980s can convincingly be linked to features of income tax (Harvey 1995) and social
security contributions which, in effect, subsidized these forms of work. The numbers employed for less than sixteen
hours per week went up from 6 per cent of the employed labour force in 1979 to 16 per cent in 1996; 30 per cent of
women employees with dependent children were in this category (EPI 1997). There is evidence of clustering of weekly
working time around the fteen-hours gure in the 1980s, which suggests that the social security contributions system
may have had an impact on the type of employment contract offered by certain employers (Dickens 1992).
However, these effects are waning thanks to recent regulatory changes. In the mid-1990s the Inland Revenue took
steps to regularize the position of many freelancers and self-employed construction workers, in the process eliminating
many of the tax advantages they formerly enjoyed (Harvey 1995). Changes made to the structure of social security
contributions in the 1998 and 1999 Budgets will also have the
effect of removing much of the incentive for employment at very low rates of weekly pay.
Although there is no evidence that changes in the legal rules relating to qualication for unfair dismissal led directly to
an increase in atypical forms of employment, they nevertheless had an impact on the proportion of the working
population protected by basic employment rights. In 1975, when the basic qualifying period for unfair dismissal
protection was six monthsone year for those working between eight and sixteen hours per week91 per cent of
employees of working age had sufcient tenure to qualify for protection. The actual level qualifying would, in practice,
be less than this because the statutory rules dening continuity of employment had the effect that continuity might be
forfeited by a break in employment, so requiring the employee to start again. 1.8 million self-employed workers were
also excluded. In 1990, when two years' service was requiredve years for those working between eight and sixteen
hours per weekonly 62 per cent of employees qualied; self-employment had risen to 3.4 million. By 1995 the
number protected had increased to 70 per cent of employees, in part because of the abolition of the separate
thresholds for part-time workers, and because of reduced voluntary quits during a period of recession (Gregg and
Wadsworth 1996).
Working Time Flexibility

Signicantly, the most widely used form of exible employment in the UK continues to be the variation of employees'
working hours from week to week (Casey et al. 1997). British working time patterns are highly varied, and the degree of
variation is increasing. In 1992 a study by the Department of Employment estimated that only 10 per cent of
employees in the UK normally worked forty hours per week, in contrast to 34 per cent of employees in the rest of the
European Community. Over 72 per cent of employees in the other then eleven member states worked between thirty-
ve and forty hours per week, compared to 36 per cent in the UK (Wareing 1992). The pattern of regulation in the UK
appears to be a major cause of this wide dispersion of actual hours worked. Although basic weekly hours in most UK
industries are close to norms set elsewhere in the EC, overtime and shift-working in the UK are only weakly regulated.
Both overtime and shift-working tend to uctuate with the economic cycle; gures from 1988, when overtime working
was at historically high levels, indicate that over 41 per cent of British male workers were employed for forty-six hours
or more per week, compared to a gure of 23 per cent for the EC as a whole.
The British economy also has a large number of workers employed for less than the standard working week. In
particular, part-time work for only a few hours per week is widespread, and the use of zero-hours contracts is also
growing. Longitudinal data indicate that the tendency for working time arrangements to become more varied has
increased since the mid-1970s (EPI 1997). The numbers working for less than sixteen hours a week amounted to 6 per
cent of the employed labour force in 1979; by 1996 this had increased to 16 per cent. In 1996 nearly 30 per cent of
women employees with dependent children worked less than sixteen hours per week. There has been a substantial
decline in the numbers working the standard week of between thirty-three to forty hours; this accounted for 48 per
cent in 1979 but only 34 per cent in 1996. Those working very long hoursmore than forty-eight per
weekincreased only slightly from 17 per cent of employees in 1979 to 20 per cent in 1996.
In short, working time in Britain is subject to
a relatively exible or permissive system, with high levels of shift, overtime and part-time working and relatively
weak tendencies towards working-time reduction for full-time workers, except as a function of sectoral and
occupational change. There is little evidence of a dramatic transformation of working-time practices in the 1980s;
what is more signicant is the picture of a relatively exible working-time system prior to the current vogue for
deregulation (Rubery et al. 1994).
Wage Levels, Structure and Differentials

There is a growing consensus amongst commentators that increased exibility and the comparatively low
unemployment rate in Britain have been achieved at the price of a deterioration in wages, growing insecurity and
increasing poverty (Marx 1999). After a long period of wage stability, wage inequality has grown sharply over the last
two decades, resulting in higher differentials than at any other time during this century (Machin 1996). There is also
evidence of a lack of income mobility, particularly among the lower paid (Dickens 1997a). The pay gap between the
lowest decile and the highest decile of workers has been growing with the real earnings of workers in the lowest decile
not rising in real terms during the 1990s.
When the gures are broken down by sex, it is clear that both women and men face wage immobility (Dickens 1997b).
Women, however face high levels of wage inequality, since they are paid on average only 79.6 per cent of male average
earnings. Further, the hourly rate of manual and non-manual women workers relative to men and of part-time workers
remain much lower than 79.6 per cent. This is primarily due to
the concentration of employment for women in low paying industries or occupations. Similarly substantial pay
inequality is experienced by members of ethnic minorities.
The impact of the decline in collective bargaining on pay and terms and conditions of employment is difcult to assess,
because of other relevant factors including changes in general labour market conditions, and increased competition in
product markets. The weakening of union power might have been expected to lower the union/ non-union pay
differential and to reduce inationary pressures. However, the signicance of the union mark-up is limited in the
context of the UK labour market, where, traditionally, unions negotiated both for their own members and for non-
members at both establishment and sector level. Studies have found that in the early 1980s, the union mark-up was no
more than 10 per cent for unskilled workers and was negligible for skilled workers. Only the presence of a preentry
closed shop signicantly raised the wages of unionized workers, by up to 23 per cent, in a situation of competitive
product markets (Stewart 1990).
The declining strength of trade unions and the removal of institutional support for the wages oor has a clearer
relationship to the growth in inequality in the UK in the 1980s and 1990s. There was a decline in the factor share of
income from employment in national and household income as a whole (Ryan 1996), and while the earnings of the
bottom decile of earners showed a rise in real terms over the period from 1980, the earnings distribution widened
considerably. Studies suggest that the decline in the coverage of collective bargaining was responsible for between one
eighth and one quarter of the increase in earnings inequality (Gosling and Machin 1995; Schmitt 1995). The reduction
in the power of the wages councils after 1986 is estimated to have led to an increased dispersion of pay, and to lower
relative pay for younger workers, who were removed from the scope of statutory wages orders at that time (Machin
and Manning 1996): [i]n that respect at least, legislation has contributed directly to the widening of pay inequalities in
contemporary Britain (Brown, et al. 1997: 7980).
The alternative explanation for inequality would place more emphasis upon falling demand for unskilled labour as a
result of changes in technology and in the structure of demand for goods and services. However, comparative studies
suggest that these factors can only account for a part of the polarization of earnings in systems such as the UK, the
USA and New Zealand which have engaged in policies of market liberalization. These more exible systems have seen
the pay of their relatively less skilled workers fall more rapidly than in systems which offer more systematic institutional
support for investment in
human capital, such as Germany and Sweden. This suggests that shifts in the demand for skills do not tell anything like
the whole story: there is no evidence that these skill shifts have made a substantial contribution to the rise in European
unemployment nor that labour market inexibility per se is associated in any simple way with such effects as have been
observed (Nickell 1997: 71).
Conclusion
This chapter has sought to evaluate the contribution of regulatory reform to changes in the structure of employment
and unemployment in Britain in the 1980s and 1990s. A widely held view of the impact of deregulation might run as
follows. High unemployment at the start of the 1980s was the legacy of over-regulation during the years of the post-
war, welfare state consensus. The Conservative administrations of the 1980s and 1990s responded by removing
regulatory rigidities, in the process freeing up market forces. While the impact of liberalization took some time to be
lter through, by the mid-1990s the UK labour market had achieved a high degree of exibility, the effects of which
included a higher employment rate and a superior record of job growth to most of its European neighbours.
As we have seen in this chapter, this popular view does not accord well with the evidence. However, many would still
hold to the view that the labour market has been rendered more exible, but that this has been at the price of greater
inequality. This is because, rstly, the effects of unemployment have not been evenly spread across the working
population but instead have been concentrated on a substantial minority of householdsaround 20 per cent of the
total populationin which none of the potential earners is in paid employment. Secondly, although the numbers of
young and long-term unemployed are lower than in most other EU countries, there is substantial churning between
unemployment and low-paid, irregular and short-term work. Thirdly, there is evidence to suggest that upwards
mobility is limited for those in employment, so that those in low-paying occupations and industries tend to stay in
them. Overall, then, we might conclude that Britain's recent experience is one of a growing employment rate, but that a
growing employment rate does not, in itself, cure the problem of social exclusion.
Even then, the proposition that the labour market reforms of the 1980s and 1990s are responsible for the recent shifts
in the structure of employment needs to be treated with some scepticism. Social scientists
144 References
prepared to make this assertionof whom there are, in fact, relatively feware on slippery ground, since there is little
or no methodological consensus on the best way even to address the question, let alone settle it. The manner in which
legal and regulatory changes interact with extra-legal norms in inuencing economic behaviour is, to say the least,
poorly understood. The very complexity of the regulatory framework and the inherent difculties in distinguishing the
effects of regulatory change from other factors militate against clear-cut ndings. As we have seen, rising female
participation in paid employment is largely responsible for the increase in the employment rate in the 1990s. However,
there is very little to link this to changes in the legal and regulatory system.
The most convincing studies are those which locate the British experience in both a comparative and a long-run
perspective. When this is done, it becomes apparent that the British labour market, while seemingly more exible, does
not have a better record on employment and unemployment than many other EU countries (Nickell 1997). The claim
that Britain has crossed the river owes a good deal to extrapolations of relatively recent trends dating from the mid-
1990s, or just to plain wishful thinking.
References
Addison, J., and Siebert, W. S. (1991), The Social Charter of the European Community: Evolution and Controversies
Industrial and Labor Relations Review, 44/July: 597625.
Alogoskous, G., Bean, C., Bertola, G., Cohen, D., Dolado, J., and Saint-Paul, G. (1995), Unemployment: Choices for
Europe. London: Centre for Economic Policy Research.
Atkinson, A., and Mogensen, G. V. (eds.) (1993), Welfare and Work Incentives. Oxford: Clarendon Press.
Atkinson, J. (1985), Flexibility, Uncertainty and Manpower Management. Brighton: Institute of Manpower Studies.
Beatson, M. (1995), Labour Market Flexibility, Employment Department Research Series No. 48. London: HMSO.
Brown, W. (ed.) (1981), The Changing Contours of British Industrial Relations: A Survey of Manufacturing Industry. Oxford:
Blackwell.
and Wadwhani, S. (1990), The economic effects of industrial relations legislation since 1979, National Institute
Economic Review, 131: 5770.
Deakin, S., and Ryan, P. (1997), The Effects of British Industrial Relations Legislation 197997 National Institute
Economic Review, 161: 6983.
Burchell, B., Deakin, S., and Honey, S. (1999), The Employment Status of Individuals in Non-standard Employment.
Employment Relations Research Series, Department of Trade and Industry.
References 145
Casey, B., Metcalf, H., and Millward, N. (1997), Employers' Use of Flexible Labour. London: Policy Studies Institute.
Daniel, W. (1985), The United Kingdom, in M. Cross (ed.) Managing Workforce Reduction. London: Croom Helm.
and Stilgoe, E. (1978), The Impact of Employment Protection Laws. London: Policy Studies Institute.
Deakin, S. (1990), Equality under a market order: the Employment Act 1989 Industrial Law Journal, 19: 119.
and Morris, G. (1998), Labour Law. 2nd. edn. London: Butterworths.
and Wilkinson, F. (1991), Labour law, social security and economic inequality. Cambridge Journal of Economics, 15:
12548.
Department of Employment (1985), Employment: the challenge for the nation, Cmnd. 9474, London: HMSO.
Department of Social Security (1998), Family Credit Quarterly Statistics, July.
Dicks, M., and Hatch, N. (1989), The Relationship between Employment and Unemployment. Bank of England Working Paper.
Dickens, L. (1992), Whose Flexibility? Discrimination and Equality Issues in Atypical Work. London: Institute of
Employment Rights.
Hart, M., Weekes, B., and Jones, M. (1985), Dismissed: A Study of Unfair Dismissal and the Industrial Tribunal System.
Oxford: Blackwell.
Dickens, R. (1997a), Male Wage Inequality in Great Britain: Permanent Divergence or Temporary Differences? in P.
Gregg (ed.) Jobs, wages and poverty: patterns of persistence and mobility in the exible labour market. London: Centre for
Economic Performance, LSE.
(1997b), Wage Mobility in Great Britain Employment Audit, Spring.
Emerson, M. (1988), Regulation or Deregulation of the Labour Market: Policy Regimes for the Recruitment and
Dismissal of Employees in Industrialised Countries European Economic Review, 32: 775817.
Employment Policy Institute (EPI) (1997), Employment Audit 5. London: EPI, 710.
(1998), Employment Audit 9. London: EPI, 1217.
Epstein, R. (1983), A Common Law for Labour Relations: a Critique of the New Deal Labor Legislation Yale Law
Journal, 92: 13571408.
(1984), In Defense of the Contract at Will University of Chicago Law Review, 51: 94782.
Evans, S., Goodman, J., and Hargreaves, L. (1985), Unfair Dismissal Law and Employment Practice in the 1980s. Research
Paper No. 53, Department of Employment. London: HMSO.
Giddens, A. (1998), The Third Way. The Renewal of Social Democracy. Cambridge: Polity Press.
Gosling, A., and Machin, S. (1995), Trade unions and the dispersion of earnings in British establishments, 198090
Oxford Bulletin of Economics and Statistics, 57: 16784.
Gosling, A., Johnson, P., McCrae, J., and Paull, G. (1997), The dynamics of low pay and unemployment in early 1990s Britain,
Working Paper Series 1997/3 London: Institute of Fiscal Studies.
146 References
Green, F. (1993), The impact of trade union membership on training in Britain Applied Economics, 25: 103343.
Gregg, P. (ed.) (1997), Jobs, wages and poverty: patterns of persistence and mobility in the exible labour market. London: Centre
for Economic Performance.
and Wadsworth, J. (1996), Feeling Insecure Employment Audit, Autumn.
(1997), A Year in the Labour Market, Employment Audit, Summer.
Knight, G., and Wadsworth, J. (1998), Down and Outs in Beverley, Yorks Employment Audit, 9.
Machin, S., and Metcalf, D. (1993), Signals and cycles: productivity growth and changes in trade union status in
British companies Economic Journal, 103: 894907.
Grubb, D., and Wells, W. (1993), Employment Regulation and Patterns of Work in EC Countries OECD Economic
Studies, 21: 758.
Hakim, C. (1990), Core and periphery in employers' workforce strategies: evidence from the 1987 ELUS survey
Work, Employment and Society, 4: 15788.
Harvey, M. (1995), Towards the Insecurity Society: The tax trap of selfemployment. London: Institute of Employment Rights.
Hayek, F. A. (1960), The Constitution of Liberty. London: Routledge & Kegan Paul.
(1973), Law, Legislation and Liberty. Volume 1: Rules and Order. London: Routledge.
Hills, J. (1998), Thatcherism, New Labour and the Welfare State, CASE paper CASE/13, Centre for Analysis of Social
Exclusion, LSE.
Hunter, L., McGregor, A., McInnes, J., and Sproull, A. (1993), The Flex-ible Firm: Strategy and Segmentation,
British Journal of Industrial Relations, 31/3: 383407.
Joseph Rowntree Foundation (1995), Inquiry into Income and Wealth. London: JRF.
Layard, R., Nickell, S., and Jackman, R. (1991), Unemployment, Macroeconomic Performance and Labour Markets. Oxford:
Oxford University Press.
Lindbeck, A., and Snower, D. (1989), Wage Setting, Unemployment and Insider-Outsider Relations American Economic
Review, 76: 2359.
McGregor, A., and Sproull, A. (1991), Employer Use Strategies: Analysis of a National Survey. Department of Employment
Research Paper No. 83. London: HMSO.
Machin, S. (1996), Wage Inequality in the UK Oxford Review of Economic Policy, 12/1: 4764.
and Manning, A. (1998), Long-Term Unemployment: Exploding Some Myths Employment Audit EPI 8.
McLaughlin, E., Millar, J., and Cooke, K. (1989), Work and Welfare Benets. Aldershot: Avebury.
Marx, I. (1999), Low pay and poverty in OECD countries Employment Audit 10: 1721.
Metcalf, D. (1989), Water notes dry up: the impact of the Donovan reform
References 147
proposals and Thatcherism at work on labour productivity in British manufacturing. , British Journal of Industrial
Relations, 27: 131.
Minford, P., Davies, D., Peel, M., and Sprage, A. (1985), Unemployment: Cause and Cure. (2nd Edn) Oxford: Basil
Blackwell.
Nickell, S. (1997), Unemployment and labour market rigidities: Europe versus North America Journal of Economic
Perspectives, 11: 5574.
Nolan, P. (1989), Walking on water? performance and industrial relations under Thatcher Industrial Relations Journal,
20: 8192.
and Marginson, P. (1990), Skating on thin ice? David Metcalf on trade unions and productivity British Journal of
Industrial Relations, 28: 22747.
OECD (Organisation for Economic Co-operation and Development) (1994), Collective Bargaining: Levels and
Coverage Employment Outlook.
Pollert, A. (ed.) (1988), Farewell to Flexibility. Oxford: Blackwell.
Robinson, P. (1997), Insecurity and the exible workforce: measuring the ill-dened Paper presented to the 12th
Annual Employment Research Unit Conference, Cardiff Business School, September 1997.
Rubery, J., and Wilkinson, F. (eds.) (1993), Employer Strategy and the Labour Market. Oxford: Oxford University Press.
Deakin, S., and Horrell, S. (1994), United Kingdom, in G. Bosch, P. Dawkins and F. Michon (eds.) Times are
Changing: Working Time in 14 Industrialised Countries. Geneva: International Institute for Labour Studies.
Ryan, P. (1996), Factor shares and inequality in the UK Oxford Review of Economic Policy, 12/1: 10626.
Schmitt, J. (1995), The changing structure of male earnings in Britain, 197486 in R. B. Freeman and L. F. Katz (eds .)
Differences and Changes in Wage Structures. Chicago, Ill.: University of Chicago Press.
Stewart, M. (1990), Union wage differentials, product-market inuences, and the division of cents, Economic Journal,
100: 112237.
Taylor, M., and Booth, A. (1996), The Changing Picture of Male Unemployment in Britain Working Papers of the
ESRC Research Centre on Micro-social Change, 96/18.
Wareing, M. (1992), Working arrangements and patterns of working hours in Britain Employment Gazette 100: 88100.
Wood, D., and Smith, P. (1987), Employers' Use StrategiesFirst Report of the 1987 Survey. Department of Employment
Research Paper No. 63. London: HMSO.
6 Going Different Ways: Labour Market Policy in
Denmark and Sweden
Anders Bjrklund
A Review of the National Debate in Denmark and Sweden
Sweden and Denmark Up Until 1990

Until around 1990, the Swedish labour market model was considered a success story with low unemployment rates,
negligible long-term unemployment, and a high degree of equality in labour market outcomes. Sweden was particularly
praised for its active labour market policies and its centralized wage bargaining system with responsible unions. In the
Swedish labour movement this successful outcome was attributed to the policy model developed by the trade union
economists Gsta Rehn and Rudolf Meidner in the late 1940s. Even though their ideas were not popular in all political
circles in Sweden, the favourable outcome gave the defenders of the Swedish model a strong position in the public
debate.
In the 1980s, the proponents of the Swedish labour market model received additional support from more academic
circles, when Richard Layard and his colleagues79 started to advocate the Swedish approach to labour market policy
and wage bargaining. Layard et al. were in favour of an active labour market policy and centralized wage bargaining,
two ingredients of the traditional Swedish model. They did, however, lay more emphasis on the design of the
unemployment compensation
79
The policy recommendations by Layard et al. were perhaps most forcefully stated in their book Unemployment (Layard, et al. 1991), but their proposals had already
received a great deal of publicity in the mid-1980s. In a public speech delivered to the Swedish Economic Association (published in Ekonomisk Debatt 1987: 5), Layard
presented his major ideas to a Swedish audience. His popular book How to Beat Unemployment (1986) was translated into Swedish and published by Tiden, the publisher
associated with the trade union of manufacturing workers (LO ).
Denmark and SwedenGoing Different Ways 149
system than was usually the case in Sweden. They proposed unemployment compensation with xed benet duration,
but combined with job and training programmes offered to unemployed job searchers whose benet periods were
about to expire. In contrast to many central European countries, Sweden had limited benet periodssixty weeks for
most workers; ninety weeks for those above 55 years of age. Layard et al. claimed that Sweden had been able to avoid
persistently high unemployment after the oil price shocks thanks to its labour market policy model.
Like Sweden, Denmark had high labour force participation among women and a quite narrow wage distribution; but in
the political circles that praised Sweden's labour market policy, Denmark was considered a severe failure. Ever since
the rst oil crisis in the mid-1970s, Denmark had suffered from high unemployment of around 10 per cent. In the late
1980s the consensus view on Denmark was that this level of unemployment had become structural in nature and could
not be reduced by conventional aggregate demand policy. At that time Denmark had a generous unemployment
compensation system with high replacement ratios, long benet periods, and eligibility rules that were favourable to
students and other entrants to the labour force. Further, Denmark did not have anything like Sweden's set of active
labour market policies for unemployed job searchers. Critics of Denmark laid emphasis on both these weaknesses,
whereas Swedish critics of Denmark tended to focus on the lack of active labour market policies.
It is interesting to reconsider a cross-country based regression equation published by Layard, Nickell and Jackman
(1991), which formed part of their evidence in favour of the Swedish model and of their critique of the policy pursued
in Denmark. A linear equation that explained average unemployment during 198388 using the following indicators:
unemployment benet duration, the replacement ratio, active labour market spending, coverage of collective
bargaining, union co-ordination, employer co-ordination and change in ination, accounted for over 90 per cent of the
variation in the dependent variable in a sample of twenty countries. Table 6A in the Appendix reports predicted
unemployment rates using this equation, the contributions of the independent variables, and actual unemployment
rates for Denmark and Sweden (see Appendix). The small deviations between predicted and actual unemployment
rates for both Denmark and Sweden lend credibility to the equation's ability to explain the outcomes for these two
countries. Both countries scored highly on collective bargaining and union and employer coverage, these factors
producing a negativethat is, unemployment reducingimpact. The difference between the two
150 Anders Bjrklund
countries could be explained by benet duration, replacement ratio, and in particular by the use of active labour market
policies.80
The economic development during the 1990s has completely changed the discussion about labour market policy in the
Nordic countries. The critics of the traditional Swedish labour market model have gained a much stronger position in
the debate. Unemployment in Sweden virtually skyrocketed from less than 2 per cent in 1990 to over 8 per cent in
1993 (see Fig. 6A in the Appendix). If cyclical labour market policy measures were added, the increase in total
unemployment would be from 4 to 12 per cent of the labour force. The gure also shows that unemployment in
Denmark rose sharply between 1987 and 1993.
The causes and cures of the high unemployment rate have been intensely discussed in Sweden. Even though there is a
considerable amount of consensus that a sharp fall in aggregate demand was the original cause of the rise in
unemployment, structural factors were increasingly emphasized in the debate on how to emerge from the recession
and how to reduce unemployment. Denmark had, in some respects, become used to its high level of unemployment,
but discussion on structural measures to combat unemployment continued. The next section of this chapter provides
an overview of the public discussion on structural reforms in the labour market, with the focus on unemployment
benets, the wage bargaining system, employment protection legislation, earnings dispersion and working hours. The
actual changes in policy and wage bargaining systems, and the outcomes of these changes, are discussed in the
following sections.
The Discussion in the 1990s

The issue of whether the present high unemployment in Sweden is cyclical, and can be reduced by demand
management, or structural, and therefore also requires structural reform if it is to be reduced, cannot yet be considered
settled. In the present public discussion both views have their supporters. Compared with the 1980s and earlier,
however, the new ingredient is discussion of structural labour market reform and the political conditions to implement
such reforms. The Danish debate, on the other hand, has almost entirely focused on structural problems in the labour
market. In what follows, I shall summarize the discussion on structural reforms and ignore the macro-oriented issues.
80
This equation has been widely discussed in the literature, and it has been been reestimated several times. See Bellman and Jackman (1996), Calmfors (1993), Forslund and
Krueger (1997), Jackman (1994) and Zetterberg (1993).
Unemployment Benets and Labour Market Policy

The Danish discussion on these issues can best be summed up by the word activation. During the 1980s and early
1990s, there was growing consensus that the Danish unemployment benet was quite generous, and in combination
with high negotiated minimum wages offered low incentives for especially low-skilled unemployed to take a job. Smith
(1998) reports results from a survey showing that over 35 per cent of unemployed women would not make monetary
gains by taking up a job. The equivalent number for men was around 25 per cent. Studies by Peder Pedersen and Nina
Smith suggest that a substantial proportion of the unemployed were not to be considered part of the effective labour
supply for other reasons.81 In particular, many young women with family responsibilities were unable to take on jobs
immediately despite the fact that they collected unemployment benets. In addition, Denmark had a quite generous
early retirement scheme.
It was also recognised in the Danish discussion that job search without help from active measures would have severe
economic consequences for unemployed job seekers and their families. Therefore, programmes that offered help in
the transition from unemployment to a job were considered to be necessary components of labour market reform. In
1992 and 1993 a public committeethe Socialkommissionenpublished various reports setting out quite detailed
proposals for such labour market reforms. It required another public committeethe Vlfrdskommissionento take
political decisions on the major labour market reforms, which are examined below.
The Swedish unemployment system has been intensively discussed in the 1990s. Most of the parameters of the system
have been scrutinized by a remarkably long series of public committees: the replacement rate, the duration of benets,
and the membership fees paid by members into their unemployment insurance (UI) fund. One common theme in this
discussion has been that the Swedish UI system has not in reality had a limited benet period, as argued by Richard
Layard, but in practice has been of indenite duration, lasting until retirement at the age of 65. The reason is that
Sweden, in the 1980s, introduced the formal right of unemployed persons whose benets are about to expire, to obtain
a place on a training programme or a temporary job subsidized by the National Labour Market Board. These training
courses or jobs, in turn, will last for ve to six months, or exactly the time needed to requalify for a new period of
unemployment benet. The effect of this right to training or
81
See Pedersen and Smith (1997) for several references. These results and the ones from Smith (1998) pertain to the situation around 1993, but the rules and wage structure
that gave rise to these results were in effect also during the 1980s.
152 Anders Bjrklund
to a temporary job has been that the duration of the Swedish UI system has become indenite. A frequent policy
proposal by economists, employers' representatives, and the non-socialist parties has been that the number of periods
with unemployment benet and programme participation should be limited.
Wage Bargaining System

The pros and cons of centralized wage bargaining have been widely discussed in both Denmark and Sweden. The
defenders of centralized bargaining have argued along the same lines as Richard Layard. At the central level, the
negative impact of too high wage increases on employment can be taken into account, or internalized. At intermediate
levels of bargaining, like the industry level, there are no such incentives. In Swedish discussion this hypothesis has been
labeled the Calmfors-curve due to the study by the Swedish economist Lars Calmfors and John Drifll (1988).
This argument has been criticised on three grounds. First, critics deny that centralization is an efcient means to
moderate growth in the real wage costs. The reason for this is that centralized agreements are not completely
determined at the central level, but are followed by more detailed agreements at both the industry and rm level. The
actors at each level of negotiations feel themselves under pressure to obtain wage increases from their bargaining in
order to justify their existence to their members. The representative of the Danish Employers Federation, Richard
Larsen (1994), claims that a centralized bargaining system may easily become a multi-level bargaining system. In a
system of this kind, the union leaders at each level are eager to show results to their members, and there is therefore an
upward pressure on total wage increases. Similar claims have been made concerning the Swedish wage bargaining
system.
A second argument against centralized wage bargaining has been that it has adverse effects on the wage structure.
Unions at the central level have generally been very eager to implement wage equalization clauses in agreements. In
addition, market forces at industry, occupational and regional levels are not allowed to affect wages in a centralized
system. According to both arguments, the wage structure will not conform to the market forces and thereby cause
unemployment.
A third line of reasoning, common among representatives of the employers' federations, is that the productivity of
single workers is determined at the level of the rm. Therefore, a rm must be able to use the mode of payment that
best ts the workers and the job tasks. To achieve this, the argument runs, wages must be set at the local level.

Employment security legislation has been a major issue in the Swedish discussion. The famous Lindbeck
Commissionappointed by the non-socialist government in late 1992 and headed by the leading Swedish economist
Assar Lindbeckrecommended several changes in this legislation.82 Among the Commission's 113 policy proposals,
several focused on labour market exibility issues. They recommended changes so that rms could hire more
workers on a temporary basis, but also changes in the rules governing dismissals and re-employment after dismissals
that would reduce employers' ring costs. In 1993 a public committee (SOU 1998: 12) presented concrete proposals
along the same lines, although the unions were strongly opposed to them.
In Denmark, employment security legislation has been less of an issue in public debate. There are strong reasons for
this being the case: rst, current legislation leaves it to the social partners to agree upon the desired level of protection;
second, employers can easily resort to lay-offs when, due to weather conditions or market uctuations, they need to
reduce their wage bill. Workers on lay-off, in turn, are offered quite generous unemployment benets, and can be
expected to be recalled when demand improves. The generosity of the system from the point of view of employers was
somewhat reduced by decisions in 1989 and 1991, after which the rst two days of unemployment had to be paid by
the employer. But the system still offers a exible device for rms facing short-term uctuations in demand.
Wage Dispersion
A common theme in debate in both countries has been that the unions have compressed wage and earnings
distributions, with high unemployment among low-skilled workers as a consequence. In Sweden, this has been the
recurrent message of the inuential annual reports issued by the economic council of the SNSthe Center for
Business and Policy Studies. A further argument in the discussion has been that low earnings differentials among
workers with low and high levels of education also imply a disincentive to further study. In Denmark, the Economic
Council has discussed these issues, especially in its 1994 report, in much the same spirit.
In both countries, the opponents in this discussion have not surprisingly been the unions. Although the unions in
general have recognized that factors like technological change, trade and European economic
82
The ofcial report of the Lindbeck Commission was published in March 1993 and received much attention in the Swedish media. An English version was published about
a year later, see Lindbeck et al. (1994).
154 Anders Bjrklund
integration may represent a threat to their low-skilled members, they have not endorsed the policy recommendations
of the interest groups which are dominated by academic economists. Their alternative policy recommendation has
instead been more training and education, and especially targeted programmes for adults with low basic education. In
both countries there is also a growing interest in various forms of tax relief for low-skilled workers.
Working Time and A-Typical Jobs

There are some notable differences between the two countries with respect to attitudes to work sharing and atypical
jobs as measures to promote employment. In Sweden, the majority of economists, trade unions, employers and
political parties have been strongly opposed to reducing working time as a means to promote employment. In the
1980s, early retirement for labour market reasons was a feasible and popular option for workers older than 58 whose
unemployment benet periods had expired. Ironically, this option was eliminated just before the deep recession in the
early 1990s started to raise unemployment. Further, two political partiesthe Left Partythe former Communist
Partyand the Green Party have been in favour of work sharing policies, among other things, for the purpose of
reducing unemployment. In the leading circles, however, work sharing has been considered a very expensive and
inefcient employment policy. In 1997 the Social Democratic government reintroduced an option for older
unemployed people to retire, but were heavily criticized for this decision.
The Danish discussion has been different, maybe because unemployment has been high for so long. At the peak of
unemployment in the early 1990s, there was an intensive discussion about various paid-leave schemes. One concrete
proposal became known as the garbage-collector model (skraldemandsmodellen). This model involved paid leave each
fourth week for garbage collectors if the absent person was replaced by an unemployed worker. It was temporarily
tried in one local community, but further experimentation was stopped by the Department of Labour. The intensive
discussion did, however, also generate a theoretical analysis in the journal of the Danish Economic Association
(Pedersen 1994). That work sharing had some appeal in Denmark is also illustrated by the fact that paid-leave schemes
are important ingredients of the great labour market reform in 1994 and 1996.
There is also a marked difference between Sweden and Denmark in political support for stimulating new jobs in the
home service sectorsuch as cleaning and home care. Denmark has implemented subsidy schemes to support such
jobs and Sweden, too, has given this kind of employment policy lengthy consideration. This issue has, to some extent,
split the Social Democratic Party in which the left wingers have opposed such schemes. The right wingers among the
social democrats and the non-socialist parties have been strongly in favour of such schemes and accused their
opponents for being too negative. Indeed, this discussion has not only been about the potential of such schemes in
reducing unemployment, but has also had an ideological component. The expression maid-jobs has frequently been
used by opponents of these home service schemes to illustrate what they perceive is the social status of the resulting
jobs.
Actual Changes in Policy and Wage Bargaining Systems
Unemployment Benets, Related Benet Systems, and Labour Market Policy

The most important changes in labour market policy to have taken place in the two countries during the 1990s are the
Danish labour market reforms introduced in 1994 and 1996. The three major components of these reforms are
presented in more detail in the Appendix.
First, after having been of indenite duration in practice, a formal limit was set on the duration of unemployment
benets. In 1994, it was decided that the maximum period should be seven years and participation in labour market
policy programmes was to be compulsory after four years of unemployment. In 1996, a further reduction in benet
period took place. The most marked reduction was made for people below 25 years of age. The maximum benet
period for this group was reduced to six months, after which placement in a programme ensued. For people over 25,
the maximum period was reduced to ve years. In addition to these reductions of benet period, the eligibility criteria
for benets were made stricter.
Second, a number of active measures were introduced to facilitate the transition from unemployment to a job. Within
the rst three months of unemployment, the employment ofce formulates an individual action plan for the
unemployed job-seeker. This is done in co-operation with the job-seeker, taking his/her situation as well as labour
market conditions into consideration. The 1994 reforms established that compulsory full-time activation through
education or job training must take place after four years of unemployment; this was reduced to two years in 1996.
Special efforts have also been made to activate uninsured unemployed job-searchers.
Third, quite generous paid leave arrangements were introduced. Paid
156 Anders Bjrklund
leave for education was made particularly attractive for unemployed workers. Child care leave for up to a year was also
introduced, but in this respect Denmark took a step closer to the rules that had existed for longer in the other
Scandinavian countries. It also became possible to take pure sabbatical leave with 80 per cent of UI-benets as
compensation. A requirement for this kind of leave was that an unemployed person should be hired as a replacement
for the person on sabbatical leave.
Unemployment benets and labour market policy have changed to a much lesser extent in Sweden during the 1990s,
even though the deep recession has generated the expected increase in the number of persons employed through
labour market policy programmes. The maximum replacement rate has been reduced twice: from 90 to 80 per cent in
1993 and down again to 75 per cent in 1996. However, since these reductions were made in most social insurance
schemes in order to reduce the federal budget decit, they were not specic to unemployment benets.
That the benet duration period has been a sensitive political issue is illustrated by the political decisions that have
been taken in Sweden. In the spring of 1994, the non-socialist government decided to set at limit of about two and a
half years on the benet period: UI benets for 60 weeks, followed by a six month period of programme participation,
followed by one nal period of UI benets for 60 weeks. The change did not become effective, however, because the
new Social Democratic government reverted to the original rules. None the less, in 1996 the Social Democrats decided
to place a limit on the total benet period, but with another practical solution and a somewhat longer total benet
period than that proposed by the former government. This change too did not come into effect, because protests by
the unions forced the government to revert to the original rules.
As regards labour market policy in general, job and training programmes have been expanded during the deep
recession of the 1990s. A large number of programme slots have also been needed in order to full the rights of those
approaching benet expiry to a training course or temporary job. At the beginning of the recession there was strong
emphasis on traditional labour market training, whereas the traditional temporary jobsrelief work or
beredskapsarbetenwere not as frequently used as in previous recessions. In 1993 it was realised that many labour
market trainees participated in order to renew their benet periods.83 To prevent this use of labour market training,
new temporary
83
See Regnr (1997) for an evaluation of labour market training in the early part of the recession.
job programmes were introduced to perform the function of renewing benet periods.
The total number of persons employed by means of temporary programmesexcluding those for the
handicappedhas amounted to around 5 per cent of the labour force. Since 1997, this number has been declining,
partly due to offers to the unemployed to participate in the regular educational system and still receive the equivalent of
their unemployment benets. This expansion of the general educational system has been in addition to a general
increase in this system throughout the 1990s; the duration of compulsory schooling has in practice been increased to
the age of 19, and the number of study places and students at universities has almost doubled since the late 1980s. This
expansion of the educational system has been strongly supported by the unions.
Overall, after its labour market reforms of 1996, and the lack of such reforms in Sweden, it is Denmark that probably
comes closer to the recommendations of Richard Layard. The equation used in Table 6.1 would most likely predict
lower unemployment in Denmark than in Sweden.

Political decisions on employment protection legislation taken in Sweden also demonstrate how politically sensitive an
issue it has been. In 1994, the non-socialist government made two changes to the existing rules governing lay-offs and
probationary periods of hiring. In cases of lay-offs, the employer must divide its staff with similar job tasks into
groups, and within each group lay-offs must take place in reverse order of workers' length of employment with the
company. In 1994 it was decided that the employer may make two exceptions to this rule per group in order to retain
those workers that it nds most valuable. The non-socialist government also extended the probationary period of
hiringbefore an employment contract becomes validfrom six to twelve months in 1994. Both changes were
abolished by the new Social Democratic government in 1995 and replaced with the old laws.
No major changes have taken place in Denmark as regards employment protection. It has, however, become more
costly to use temporary lay-offs to adjust to short-term uctuations in demand, since the employer must now pay for
the two rst days of unemployment benet. In general, Denmark's employers still rank high in cross-country
comparisons of exibility in adjusting labour force to uctuations in demand (OECD 1996a).
158 Anders Bjrklund
Wage Bargaining System

The Swedish wage bargaining system used to be famous for its high degree of centralization and for the low degree of
state intervention.84 From 1956 to 1981, bargaining at the national union level was coordinated through central
agreements between the Swedish Employers' Confederation (SAF) and the Swedish Trade Union Confederation (LO).
The LO mainly organised blue-collar workers in the private sector, but the agreements between the LO and the SAF
set the pattern for the other parties on the labour market. At the beginning, it was the SAF that took the initiative on
these central agreements, which later became the main instrument with which the LO equalized wages: the wage
solidarity policy, which had its heyday between 1966 and 1980.
Although the employers initiated central wage bargaining, they also came to question the value of it. The break with
central bargaining came in 1983 when the Swedish Engineering Employers' Association (VF, later VI) concluded a
separate agreement with Swedish Metal Workers' Union (Metall). The remaining parts of the SAF-LO area were
covered by a central agreement. The rest of the 1980s was characterized by a combination of some central co-
ordination with some free bargaining at the national industry level. The nalized central agreements except that of the
VF, did not in general contain the detailed regulations in the form of low-wage provisions that were introduced during
the 1970s. Public sector agreements also became less detailed in this respect, showing that public and private employers
were pursuing similar goals.
Around the 1990s there were marked signs of severe weaknesses in the Swedish economy. The Social Democratic
Government acted to achieve a low level of wage growth by trying to persuade the social partners to agree to low wage
increases. When this attempt failed, the government appointed an expert group consisting of a former
DirectorGeneral of the National Labour Market Board and previous chief negotiators from several organisations.
The purpose of this group was to assist the social partners to form agreements on a low level of wage increases. After
several attempts and with implicit threats of coercive wage legislation, all parties nally concluded similar so-called
stabilization agreements that expired in March 1993. Elvander (1997) describes the work of the group as a de facto
reinforced mediation institute, also containing elements of incomes policy and arbitration. Overall, the degree of
governmental involvement in this process was very high by historical Swedish standards.
84
Elvander and Holmlund (1997) provide a good introduction to the history of the Swedish wage bargaining system and to the changes made in the 1990s.
Since 1993, the trend of the 1980s towards more decentralized bargaining with negotiations at the industry level has
continued. Although there is a general consensus among the social partners that the wage bargaining process is not
working well, there is less consensus on what changes are needed. In the spring of 1997 the Social Democratic
government appointed the General Director of the National Institute of Economic Research to design a new
mediation institute. His proposal was presented in 1998 and will represent a new phase in Swedish wage bargaining if it
is implemented (SOU 1998: 41).
The Danish wage bargaining system was also characterized by a high degree of centralization in the 1970s (Ibsen and
Pedersen 1997; Larsen 1994; and Pedersen 1994). Settlements at the central level between the LO and the DA, the
employers' confederation, gave strong signals both to local negotiations and to other sectors of the labour market. In
the late 1970s, the parliament intervened in central bargaining on several occasions.
Since the early 1980s there has been a gradual process of decentral-ization. One dimension of this decentralization is
that the LO and DA no longer directly participate in the wage bargaining process. Instead, wage agreements are
reached by under-organizations at the industry level. Generally, the organizations for the manufacturing industries
have been the front-runners in the wage bargaining rounds. Therefore, a certain degree of co-ordination is taking place
in this new system of wage bargaining as well.
A further factor enhancing decentralization is that central agreements at the industry level have gradually become less
binding on individual rms. The Danish system of wage negotiations has traditionally distinguished between a normal
wage system and a minimum wage system. In the former system, wages are negotiated centrally and only minor
additions can be made at the local level. In the latter system, central bargaining sets a oor for nal wages, which are
otherwise determined at the local level. In the private sector, the minimum wage system has gradually become more
common. Pedersen (1994) reports that, in 1994, 85 per cent of workers in the DA/LO area were covered by this
system, an increase of some 35 percentage points since the late 1980s. The public sector, however, principally relies on
the normal wage system, although the scope for local deviations has increased to some extent during the 1990s.
Owing to these changes in the wage-bargaining system over time, one would expect industry- and rm-specic factors
to have become more important determinants of the wage structure than previously, although the magnitude of the
changes is not necessarily large.
160 Anders Bjrklund
The Outcomes
What have been the outcomes of these changes in policy and wage bargaining systems?
For several reasons it is hard to provide clear answers to this question. One difculty is that the most marked change
took place in Denmark at a late stage: the reforms of the unemployment benet system in 1996. Another problem is
that for Sweden it is difcult to distinguish between the effects of the small reforms that have been made and the
effects of the recession as such. None the less, I shall attempt to examine and discuss these effects as regards a number
of outcomes in the labour market.
Unemployment
Denmark
The most obvious indicator of labour market performance in Denmark since the labour market reforms in 1994 and
1996 is the unemployment rate. Figure 6B (Appendix) presents monthly data on the unemployment rates in the 1624
and 2559 age groups.
The gure shows that there has been a striking decline in youth unemployment in Denmark since the introduction of
the labour market reforms. There is also a noticeable decline among the prime-age. Other sources reveal that the fall of
long-term unemployment and unemployment of unskilled workers has been particularly strong. However, it is still too
early to determine the contribution to this favourable development made by the labour market reforms in general and
by the stricter UI-rules in particular. There was a general recovery in Denmark over the same period, and, to the extent
that the labour market reforms have been effective, it is not easy to establish how much of the decline is due to the
separate parts of the reforms.
The labour market reforms have been examined in a number of recent reports by the Danish Institute for Social
Research (Socialforskningsinstituttet). Even though these studies suffer from their short follow-up period, they offer
valuable information for those seeking to evaluate the reforms. Langager (1997) focuses on the consequences for
unemployment benet recipients, using information obtained from a large-scale survey. The respondents reported
general satisfaction with their individual action plans, so this seems to have been a useful part of the activation process.
Further, a large proportion of participants in job training and education programmes found these measures useful,
believing that their labour market prospects had improved.85 A considerable number of those receiving job training
from employers claimed that they had similar job tasks to other employees. At the time of the survey, in late 1995 and
early 1996, it emerged that regular employment with rms offering job training was more common than employment
with other rms.
Bjerregrd Bach (1997) has studied the labour market reform from the point of view of the rms using a mail
questionnaire survey of over 4,000 rms. He nds that the public employment exchange ofces are not greatly used by
rms for recruitment. However, in order to use the wage subsidies that are part of the job training schemes, the rms
must involve the employment ofces. When asked about the main restriction on not employing more persons on job
training schemes, rms most frequently replied that they did not expect the prospective applicants to have the relevant
qualications. There were also questions in the survey whether the rms would have hired the persons even without
the subsidies for job-training slots. Their replies indicate that such deadweight effects may be important in the private
sector, in which one out of four rms reported that the same person would have been hired even without subsidies,
and one out of two that recruitment would have taken place without subsidies. However, a signicant net employment
effect is consistent with deadweight effects of this magnitude.
Andersen, Appelborn and Weise (1996) have investigated paid-leave schemes by means of surveys covering both
participants and rms. Perhaps not surprisingly, most participants report that they were pleased with their spells of
paid leave. Nor is it surprising that women and public-sector employees were over-represented among those taking up
opportunities for paid leave. The replies to the survey also reveal that those on educational leave found the skills
acquired useful in their jobs, and around 70 per cent of rms agreed that workers returning from training had obtained
useful new skills. On the other hand, rms report problems with child care leave; small rms in particular say that they
may become more cautious in hiring women with young children due to the costs involved in nding substitutes for
them when they go on leave.
85
Langager also presents a non-experimental econometric study of the impact of training participation on the degree of unemployment during a period of time. He employs
panel data to estimate a so-called xed-effect model where the changes over time in participants' and non-participants' unemployment experiences are compared. He nds
positive effects of programme participation, i.e. the degree to which unemployment is reduced by participation. However, a model of this kind may be difcult to use if
participation is compulsory. The estimated parameters of programme participation could capture the difference between early and late participation; in that case the results
cannot be used to determine programme-participation in general.
162 Anders Bjrklund
One gains the general impression from these early evaluation studies that the activation policies of Danish labour
market reforms are serious attempts to improve the quality of the transition process from, particularly, long-term
unemployment to a job, and that they are not make-work labour market policies designed to make the unemployment
gures look better. None the less, it is often easier to nd positive effects of labour market programmes in interview
studies than in quantitative analyses of real outcomes like post-programme employment and earnings. Rosholm (1998:
ch. 5), for example, found mixed results in his study of the post-programme impact on unemployment and
employment durations of the employment subsidy schemes in effect during the 1980s. So, more thorough studies are
needed to evaluate the Danish labour market reform.
Sweden
The central issue concerning Sweden is the turnover from unemployment into programme participation and back to
unemployment again. What has been the impact of a system that allows this type of turnover? Because there have been
no changes in the system, there is no information that can identify the effects of the system, or the effects of eliminating
the right to move back and forth between unemployment and programme participation.
There are, however, two high quality studies which shed light on the structure of unemployment exit-rates in the
presence of the right to renewable benets periods through programme participation. (Carling et al. 1996; Thoursie
1998). Both studies have employed large samples of individual unemployment spells collected from the administrative
records of the National Labour Market Board (AMS). The time-periods covered by the two studies were 19913 and
19915 respectively. They estimate so-called hazard models from which it is possible to derive unemployment exit-
rates. Both studies distinguish between UIrecipients and non-recipients, and examine two destinations from
unemployment: to a regular job, and to a labour market programme.
The basic results are illustrated in Fig. 6C (Appendix). There are two striking patterns in exit-rates. First, after falling in
the rst part of the unemployment spell, the exit-rate to regular jobs for UI-receivers increases close to the sixty week
time limit when unemployment benets expire. This increase in exit-rates cannot be found among those who do not
receive UI-benets, which suggests that the limit on the duration of unemployment benets exerts incentive effects on
search behaviour. Second, exit-rates to labour market programmes rise markedly close to the date of benet expiry,
which suggests that the programmes full an important function of giving entitlement to further benet periods. An
interesting overall conclusion is that the limited benet period seems to exert incentive effects despite the fact that there
is guaranteed participation in a programme, which in turn grants the right to a new benet period. Thus, the guarantee
has not eliminated the incentive effects of a limited benet. It may be, though, that the effect of the limited benet
periodthe rise in exit-rates just prior to sixty weekswould have been even stronger without this guarantee.
Another conclusionand a negative one for proponents of labour market policyis that potential programme
participation seems to discourage unemployed job seekers whose benets are about to expire.
The defenders of the right to renewable benet periods by programme participation have advanced two main
arguments. First, a stricter limit on the benet duration period would have negative consequences on income
distribution, or at least it would force many unemployed persons to apply for means-tested social assistance benets
instead of collecting unemployment benets. A study by Ackum et al. (1995), which examined the characteristics of
persons with long and repeated unemployment spells, not surprisingly provides support for this view: such persons
were over-represented in the lower part of the income distribution and they were more liable to collect social assistance
benets than others. A second line of defence has been that the impact of introducing a limited benet period would
be relatively small, because only few unemployment spells last long enough for the right to the renewal of benets to
exert a marked impact on total unemployment. Ackum et al. also provide gures which shed light on this issue. They
looked at the number of unemployed who would have lost benets if the stricter rules about benet duration proposed
by the non-socialist government had been effective, and if there had been no behavioural changes by the unemployed.
The number amounted to about 0.6 per cent of the labour force. A further 1.1 per cent would have lost their benets
within six months. It therefore seems that a signicant number of unemployed persons would have received much
stronger incentives for job-nding if the proposed benet duration period had been in effect.
Another study has exploited the natural experiment created by the reduction of the replacement ratio from 90 to 80
per cent in July 1993 (Harkman 1997). This study used data on newly-registered unemployed persons at employment
exchange ofces in the period from July 1992 to July 1994. A standard type of hazard model, which explains
unemployment exit-rates, was estimated. The model allowed for different changes in exit-rates in July 1993 for UI-
receivers and non-receivers. The difference in changes in exit-rates can be attributed to the reduction of the
replacement rate. This approach to estimating the effect of
164 Anders Bjrklund
unemployment benets acknowledges that there may be general changes in opportunities to exit unemployment at the
time of the change in policy, but if these changes are the same for both UI-receivers and non-receivers, the difference
in the change will capture a true effect. The author of the study nds an effect in the direction predicted by standard
search theory, and its magnitude is about the same as that obtained by several previous cross-section studies: the
elasticity unemployment compensation with respect to the exit-rate to regular jobs is around -0.7. The author discusses
alternative explanations for the results and notes that the precision of the estimates is quite low. None the less, the
results suggest that economic incentives cannot be neglected when seeking explanations of job search outcomes.
Wage Dispersion
One may expect changes in policy and wage bargaining systems to exert several effects on the structure of wages. It is
natural to start by looking at the evolution of overall wage or earnings dispersion. Unfortunately, data on wage
dispersion are meagre, especially in Denmark. Regularly published ofcial data are not available for Denmark and
have only very recently become available in Sweden. The available evidence for the period examined here is
summarized in Table 6B in the Appendix. Two complementary series are given for Sweden. Sweden 1 covers 19927
and uses ofcial data from Statistics Sweden based on reports from employers. Sweden 2 covers selected years from
198191 and is based on survey responses of individuals about earnings and hours worked. That the latter series
reveals higher inequality suggests that measurement errors are higher in data based on responses from individuals. The
Danish data stem from reports from employers. In Table 6B, the overall wage dis-persion is shown as the 90/10
percentile ratio, where the percentiles refer to the overall wage distribution. This overall wage dispersion is broken
down into inequality in the lower half (the 50/10 percentile ratio), and inequality in the upper half (the 90/50 percentile
ratio) of the distribution.
In Denmark, wage dispersion was quite stable from 1980 to 1990, despite the gradual decentralization of the wage
bargaining system with more scope for rm- and industry-specic factors to inuence wages. Additional information
from Datta Gupta et al. (1998) using the standard deviation of log hourly wages as the measure of inequality, conrms
the stable pattern over this period: the standard deviation was 0.238 in 1983 and 0.245 in 1989. The same source
reveals a minor increase to 0.259 in 1994.
Stability is also the main impression gained of the Swedish experience. One can discern some increase in both series
and hence in both periods of time, but the increases are modest. A frame of reference to the experiences of Denmark
and Sweden is provided by the fact that the 90/10 percentile ratios for males in the United States increased from
around 3.25 in 1980 to around 4.3 in the mid-1990s.
This quite stable pattern of the overall wage dispersion does not rule out that the wage structure in some dimensions
may have been affected by the structural changes in the labour market in general and by the wage bargaining system in
particular. The impact of rm- and industry-specic factors on wages may have increased even if the overall wage
dispersion as measured by percentile ratios has been quite stable. Some studies have addressed the issue of the
importance of such factors as regards wages. Holmlund and Zetterberg (1991) estimated wage equations on pooled
cross-industry and time-series data, and found that internal industry-specic factors are more important in Germany
than in Sweden, Norway and Finland, and that they are even more important in the United States. Holmlund (1997) re-
estimated such equations on data covering twenty-eight industries and 30 years (196493). He focused in particular on
the question of whether there was a shift in the determinants of wages from 1983 onwards. The overall nding of his
estimates is that the impact of rms' performance on wages was small but increasing after 1983.
One reason why only small changes in the structure of wages are observed in the early 1990s may be that the rate of
ination fell sharply. If wages are sluggish downwards in nominal terms, the structure of relative wages may change
more easily, the higher the ination in the economy. With the present very low rates of ination in both countries, it
may therefore take some time before the full consequences of more decentralized wage-setting are observed.
Although the structure of wages has been quite stable, the distribution of disposable income may have changed
markedly, since it is also sensitive to working hours and unemployment. The evolution of inequality of disposable
income in the four Nordic countries in the early 1990s has been examined by Aaberge et al. (1999). They nd that in
Denmark this kind of inequality was not at all responsive to the uctuations in unemployment over the period. Their
results suggest that unemployment benets largely counteracted the impact of more unequally distributed market
income. Pedersen and Smith (1998), using standard relative poverty measures, conrm this nding. The share of the
Danish population aged 1875 with income below the median, or
166 Anders Bjrklund
alternative relative poverty measures, was stable up until 1995 when unemployment started to decline.
In Sweden, there has been a modest increase in the inequality of disposable income during the 1990s. The largest
changes seem to have affected young people, whose income fell considerably in the early 1990s. The rise in
unemployment combined with low coverage of unemployment compensation is most likely to have been one
contributing factor, while the rise in high school and college enrolment rates is another. The survey of income
inequality in the major OECD countries by Gottschalk and Smeeding (1997) shows that Denmark and Sweden
belonged among the four or ve countries with the highest income equality in 1992.
The Evolution of Employment ContractsPermanent Versus Temporary Jobs

The Swedish labour force surveys since 1987 yield useful information on types of employment. In addition to the
distinction between employed and self-employed workers, they also contain information on whether employed persons
have permanent or temporary employment contracts. There is a noticeable increase in the fraction of workers on
temporary contracts from 1990 to 1997, and a corresponding decline in the fraction with permanent jobs. The overall
increase in the fraction with temporary jobs was around 4 percentage points, with a much stronger increase among the
young.
This evolution has ignited the discussion about the nature of jobs in the future labour market. But it is still too early to
determine whether the rise in temporary employment contracts is structural or cyclical. The latter explanation seems
reasonable, since one would expect employers to be cautious with their hiring policies in a recession. Further, workers
may feel forced to accept temporary jobs in a period when few vacancies are available. On the other hand, the increase
of temporary jobs continued when the economy improved from 1994, so there might also be structural changes
involved. Among structural explanations, the temporarily more generous treatment of temporary contracts by
employers may account for some of the overall increase in 1994, but not all of it.
The Role of Unions

It is of particular interest to examine how union membership rates have evolved in Denmark and Sweden in the 1990s.
High unemployment and the decentralization of the wage bargaining process has reduced the bargaining
power of unions. But in both Denmark and Sweden there seem to have been counteracting forces involved.
In both countries the unemployment insurance system is organized as a so-called Gent system, by which is meant that
membership of an unemployment insurance fund is necessary to obtain eligibility for unemployment compensation.
Membership fees, however, cover only a small proportion of total costs, and the government's contribution is both the
largest and the sole marginal source of nancing.
The unemployment insurance funds are run by, or are closely related to, the unions. Often, membership of the union
and of its unemployment insurance fund is considered to be one and the same thing. Belgium and Finland also have
Gent systems. It is widely assumed that the Gent system has a positive impact on union membership rates.
In periods of rapidly rising unemployment, one may expect demand for unemployment insurance to increase. In a
country with unemployment insurance organised as a Gent system, union membership rates may as a consequence
also increase.86 Even though rising unemployment is likely to reduce the bargaining power of unions, there may be a
counteracting effect involved if union membership rates increase.
Table 6C reports the rates of union membership among all employed workers in Denmark and Sweden from 1980 to
1996. Owing to some statistical issues, two series are reported for Sweden. The evidence is consistent with the
hypothesis that rapidly rising unemployment has a positive impact on rates of union membership in a country with a
Gent system. During the period 1986 to 1990, when the labour market was overheated, with extremely low
unemployment, the rate of union membership fell in Sweden by about 4 percentage points. This was the rst period of
declining union membership ever recorded in Sweden. In the 1990s, however, the downward trend was reversed and
the entire decline in membership rates had been recovered by 1995. According to Kjellberg (1997), Scheuer (1992)
reports a marked increase in Danish rates of union membership in the 1970s when Denmark had rapidly rising
unemployment. Further, Finland also experienced rising rates of union membership in the 1990s when unemployment
rose dramatically. Finally, the level of union membership is high in Denmark, Finland and Sweden from an
international perspective, and the same is true of Belgium compared with its neighbour countries in central Europe.
There are, however, further explanations for the high overall rates of union membership in the four countries with
Gent systems. For
86
See also Rothstein (1989) for claims about such effects.
168 Anders Bjrklund
example, in the Swedish setting, unions perform important functions in rms contemplating layoffs, so there may be
additional reasons why union density rates have increased in Sweden in the 1990s. Scholars in the Swedish industrial
relations system have also found a trend towards more union involvement in rms' long-term planning (Nilsson 1997).
But there are some differences between the two countries with respect to the administration of labour market policy.
Several observers claim that the regional labour market councils in Denmark, which implement the active policies,
have given union representatives important tasks due to the labour market reforms (Boje and berg 1999; and Torng
1999). In Sweden, the corporatist ingredient of the administration of labour market policy has become less important
for the unions due to the retreat of employers from most boards of national agencies, and also because power has been
decentralized from the National Labour Market Board to its own civil servants (Bergstrm 1997).
To sum up, there is no doubt that the unions in Denmark and Sweden are still large and powerful in the mid-1990s.
They have been on the defensive in discussion of structural reform in the labour market, and high unemployment has
had an impact on their bargaining position. But membership rates are still high, and, particularly in Sweden, the
unionsespecially LOhave closely inuenced the policy of the Social Democratic government since 1994.
Concluding Discussion
Denmark and Sweden offer two most interesting casestudies in terms of labour market evolution and structural
policies. One would expect to nd numerous similarities between two countries with such strong geographic, linguistic,
political, social and academic links. There are, however, several signicant differences. The design of labour market
policy has differed for many decades, with a stronger emphasis on active measures for unemployed workers in Sweden
and more generous unemployment benets in Denmark. Not surprisingly, therefore, for several decades prior to the
early 1990s, Sweden experienced very low unemployment, whereas Danish unemployment came close to the double
digit level. None the less, both countries had compressed wage distributions by international standards, and as late as
1992 both Denmark and Sweden belonged among the OECD countries with the most equal distributions of
disposable income. At that time, the crisis in Sweden had already brought unemployment to around 8 per cent, so that
one lesson to be drawn from the experience of the two countries is
apparently that low unemployment is not a necessary condition for equally distributed disposable incomes.
The Swedish crisis in the 1990s generated intense discussion on structural labour market reform; many economists
active in public policy debate argued in favour of such reforms. This debate was in several respects similar to the one
that had been going on in Denmark for a long time, although some issues, like employment protection legislation, were
emphasized more in Sweden than in Denmark.
Discussion in Denmark led to the labour market reforms of 1994 and 1996. In terms of policy changes from which
lessons can be drawn, this is the most interesting development in the two countries. At the time of writing, the
proponents of these reforms have tended to prevail in the debate, owing to the rapid decline in unemployment in
general and among young people in particular. It is, however, too early to determine the extent to which the stricter
rules for unemployment benets or the active policies, or the overall recovery of the economy, can account for this
evolution. It is also too early to evaluate the potential negative effects of these reforms on wage and income inequality.
The development of labour market policies in Denmark and Sweden also raises the more political issue of why such
structural measures were taken in Denmark but not in Sweden. In point of fact, though, Swedish discussion has started
to focus on the political conditions for structural labour market reform. It is noteworthy that in the autumn of 1997,
the SNS, which has advocated such reforms for a long time, organized a conference on this topic with the political
economist Gilles St Paul as a keynote speaker. In Sweden, there is no doubt that the political struggle on such reforms
has taken place within the Social Democratic Party, and the union representatives in this party have so far successfully
fought against all proposals for such reforms. The unions have also had very strong incentives to take up this ght,
because the existing system of UI-benets based on unemployment insurance funds tied to the unions, facilitates the
recruitment of new unions members. Even though the Swedish unions have seen their bargaining power reduced by
high unemployment and more decentralized wage bargaining, the existing unemployment benet system has probably
enabled them to recruit new members.
How, therefore, could Denmark with a Social Democratic government and a similar unemployment benet system
decide on these labour market reforms without provoking a vehement reaction from the unions? One possible
explanation is that after twenty years of high and quite well compensated unemployment, it nally came to be regarded
as very costly and a nancial threat to basic public services like education, health and care of the elderly. Sweden made
budget cuts in these
170 Anders Bjrklund
public programmes to eliminate the huge budget decits caused by the crisis, and there is now growing concern about
the declining quality of Swedish schools, child care centres, and health care centres for elderly. If unemployment
remains high and costly for public budgets, and if the Danish experiment proves to be successful, the likelihood of
structural labour market reforms in Sweden will increase markedly in the next few years.
At the same time, an interesting experiment is at present being conducted in Sweden. As a substitute for structural
reforms of unemployment benets and employment protection legislation, the Swedish government has undertaken
the massive expansion of education among both young and adult workers. If this policy also turns out to be successful,
it may provoke constructive discussion on alternative models for combating structural problems in the labour market.
Appendix 171
Appendix
Main ingredients of the Danish labour market reforms 1994 and 1996
The 1994 (January 1 ) Measures St
Activation of the Unemployed

1. After three months of unemployment an individual action plan must be formulated.
2. All unemployed persons are entitled to one year of education or job training or to a grant to start their own
enterprise within two years of unemployment.
3. The unemployment period is divided into two (four plus three years). Full-time activation through job offers,
training and education is compulsory in the second period.
4. Benets during education and job training are not to fall below unemployment benets.
5. Activation efforts were extended to cover the uninsured unemployed.
Unemployment Benets
1. UI can be obtained for no longer than seven years. Job offers and activation cannot be used to restore benet
entitlements, but participation on paid-leave schemes is in addition to the UI period.
2. The unemployed are authorised to combine an annual income from ordinary work up to DKr 27,500 with 80
% of maximum benets.
Paid Leave Arrangements

1. Education: Unemployed UI-recipients and persons employed for three years could take educational leave for a
whole year and receive pay according their level of UI-benets. The benet period for unemployed persons
was prolonged by the time spent on educational leave.
2. Sabbatical: A person in employment for three years could take paid sabbatical leave and receive 80 % of UI-
benets if s/he was replaced by an unemployed job seeker.
3. Child care: Unemployed UI-recipients obtained the right to keep their benets and take child care leave for up
to a year. Employed workers obtained the same right for at most half a year.
The 1996 (January 1 ) Extensions St
Activation of the Unemployed

1. Compulsory full-time activation was advanced by two years to take effect after two years of unemployment.
2. Advancement of training (lasting 18 months) for low-skilled young people to take effect after six months of
unemployment.
172 Appendix
Unemployment Benets
1. UI can be obtained for a maximum of ve years. Paid leave for education can no longer prolong the benet
period.
2. Access to UI requires 52 weeks of non-subsidized work (formerly 26 weeks) over the last three years.
3. The maximum benet period reduced to 6 months for people below 25 years of age.
4. UB can be accessed from the age of 19, as the minimum age for membership of UI funds is increased by two
years.
5. A supervisory unit to oversee the UI funds was created to control enforcement of job availability criteria.
Paid Leave Arrangements

6. Education: The rules became less generous for unemployed persons, for example the benet period was no
longer prolonged by the time on leave.
7. Sabbatical: The rules were made somewhat less generous.
8. Child care: The maximum duration of paid leave for unemployed workers was reduced to 26 weeks.
Note: Some paid leave arrangements were already in force in 1994, but the labour market reform brought a major
extension to them.
Sources: This information is based on information in Andersen et al. (1996); OECD (1996a); Langager (1997); and
Rosdahl (1996).
Table 6A. Predicted Unemployment and Underlying Factors
Factors Denmark Sweden
Constant 0.24 0.24
Benet duration 2.30 1.10
Replacement ratio 15.30 13.60
Active labour market policy spend- -1.03 -4.50
ing
Collective bargaining, union and -9.75 -9.75
employer coverage
Change in ination 1.05 1.33
Predicted rate of unemployment, 8.1 2.0
average 198388.
Actual rate of unemployment, aver- 9.0 2.2
age 198388.
Source: Layard et al. (1991).
Appendix 173
Table 6B. Hourly Wage Inequality, Denmark and Sweden, 19801997

Year Denm. Denm. Denm. Swe. 1. Swe. 1. Swe. 1. Swe 2. Swe 2. Swe 2.
P50/P10 P90/P50 P90/P10 P50/P10 P90/P50 P90/P10 P50/P10 P90/P50 P90/P10
1980 1.41 1.52 2.14
1981 1.41 1.53 2.16 1.30 1.54 1.99
1982 1.40 1.55 2.17
1983 1.38 1.57 2.17
1984 1.39 1.58 2.20 1.31 1.48 1.93
1985 1.41 1.54 2.17
1986 1.42 1.55 2.20 1.35 1.48 1.99
1987 1.41 1.56 2.20
1988 1.40 1.55 2.17 1.32 1.52 2.01
1989 1.39 1.57 2.18
1990 1.38 1.57 2.17
1991 na na na 1.35 1.54 2.08
1992 na na na 1.23 1.45 1.78
1993 na na na 1.25 1.45 1.81
1994 1.25 1.46 1.81
1995 1.24 1.45 1.80
1996 1.24 1.47 1.83
1997 1.25 1.50 1.87
Notes: Figures are given in percentile ratios and relate to both sexes.
Sources: For Denmark see OECD, Employment Outlook, 1996, table 3.1. Sweden 1 is based on the new earnings statistics that Statistics
Sweden has produced since 1992. Source: Lnestatistisk rsbok various issues. Sweden 2 comes from Edin & Holmlund (1995) who in turn
have used data from the Level of Living Surveys in 1981 and 1991 and from HUS surveys (Household Market and Non-market Activities)
in 1984, 1986 and 1988. In both these surveys questions are asked about earnings and working hours at the present job and hourly earnings
is constructed out of this information.
174 Appendix
Table 6C. Union Membership Rates, 19801996

Year Denmark Sweden 1 Sweden 2
1980 75 79 78
1981 79 78
1982 80 79
1983 80 80
1984 81 80
1985 77 82 81
1986 I 77 83 82
1986 II 87 86
1987 86 85
1988 75 85 84
1989 76 84 83
1990 76 83 82
1991 76 85 83
1992 77 87 84
1993 77 89 85
1994 76 89 85
1995 75 88 85
1996 87 84
Source: Kjellberg (1997).
Appendix 175
Fig. 6A. Unemployment rates in Denmark and Sweden 19871997
Source: Statistics Denmark, register based data, Statistics Sweden, survey based data.
176 Appendix
Fig. 6B. The evolution of unemployment in Denmark since 1994. Monthly data. 1624 years, and 2559 years
Source: Statistics Denmark, register data.

Appendix 177
Fig. 6C. Stylized patterns of conditional exit-rates out of unemployment in Sweden during the recession in the 1990s
according to two recent studies.
Source: From Carling et al. (1996) and Thoursie (1998).

178 References
References
Aaberge, R., Bjrklund, A., Jntti, M., Pedersen, P., Smith, N., and Wennemo, T. (1999), Unemployment shocks and
income distributionHow did the Nordic countries fare during their crises?, Scandinavian Journal of Economics.
Ackum, A. S., Bjrklund A., and Harkman, A. (1995), Unemployment insurance, labour market programmes and
repeated unemployment spells, Swedish Economic Policy Review, 2: 10128.
Andersen, D., Appelborn, A., and Weise, H. (1996), Orlovevaluering af orlovsordningerne. Copenhagen: Socialforskningsi
Instituttet, 96: 11.
Bellman, L., and Jackman, R. (1996), The impact of labour market policy on wages, employment and labour market
mismatch, in Gnther Schmid et al. (eds.), International Handbook of Labour Market Policy and Evaluation. London:
Edward Elgar.
Bergstrm, J. (1997), Korporatismens ddSAFs avhopp frn Arbetsmarknadsverkets styrelser, Arbetsmarknad och
Arbetsliv, 3: 16584.
Bjerregard, Bach, H. (1997), Virk somhederme og erbejdsformidlingens tilbudevaluering at arbejdsmarkedsuefor-
men II: Socialforskningsi Instituttet, 97: 19.
Boje, T. P., and berg, R. (1999), Den nya danska arbetsmarknadspolitikenrttigheter svl som skyldigheter,
Arbetsmarknad och Arbetsliv, 5: 526.
Calmfors, L. (1993), Lessons from the macroeconomic experience of Sweden, European Journal of Political Economy, 9:
2572.
and Drifll, J. (1988), Bargaining Structure, Corporatism, and Macroeconomic Performance, Economic Policy, 6:
1361.
Carling, K., Edin, P. A., and Holmlund, B. (1996), Unemployment Duration, Unemployment Benets and Labor
Market Programs in Sweden, Journal of Public Economics, 59: 31334.
Datta Gupta, N., Oaxaca, R. L., and Smith, N. (1998), Wage Dispersion, Public Sector Wages and the Stagnating Gender Wage
Gap, Working Paper 98-18, Centre for Labour Market and Social Research, University of Aarhus.
Edin, P. A., and Holmlund, B. (1995), The rise and fall of solidaristic wage policy? in R. Freeman and L. Katz (eds.),
Differences and changes in wage structures. Chicago, Ill.: University of Chicago Press and the National Bureau of
Economic Research (NBER).
Ekonomisk Debatt (1987), Att bekmpa arbetslsheten (How to beat unemployment), discussion from a meeting with
Swedish Economic Association with, among others, Richard Layard and Ingemar Sthl.
Elvander, N., and Holmlund, B. (1997), The Swedish bargaining system in the melting potinstitutions, norms and outcomes in the
1990s. Stockholm: Arbetslivsinstitutet.
Forslund, A., and Krueger, A. (1997), An evaluation of the Swedish active labor market policynew and received
wisdom, in R. Freeman et al. (eds.), The Welfare State in Transition. Chicago, Ill.: University of Chicago Press and the
NBER.
References 179
Gottschalk, P., and Smeeding, T. (1997), Cross-national comparisons of earnings and income inequality, Journal of
Economic Literature, 35: 63387.
Harkman, A. (1997), Arbetslshetsersttning och arbetslshetstidvilken effekt hade snkningen frn 90 till 80
procents ersttningsniv? in A. Harkman et al. (eds.), Arbetslshetsersttningen och arbetsmarknadens funktionsstt.
Stockholm: Arbetsmarknadss-styrelsen, Ura 1997: 1.
Holmlund, B. (1997), Some new evidence on wage behaviour in Sweden in Elvander and B. Holmlund, The Swedish
bargaining system.
and Zetterberg (1991), Insider Effects in Wage DeterminationEvidence from 5 Countries, European Economic
Review, Vol. 35, No. 5, pp. 100934.
Ibsen, F., and Lisbeth, P., (1997), Lonforhandlinger p det danske arbejdsmarked, in Er der vejetilfuld beskaegelse,
Socialforskningsi Instituttet, 97: 13, Kpenhaun.
Jackman, R. (1994), What can active labour market policy do?, Swedish Economic Policy Review, 1: 22157.
Kjellberg, A. (1997), Fackliga organisationer och medlemmar i dagens Sverige. Stockholm: Arkiv frlag.
Langager, K. (1997), Indsatsen over for de forsikrede ledigeevaluering af arbejdsmarkedsreformen I. Copenhagen:
Socialforskningsi Instituttet, 97: 20.
Larsen, R. B. (1994), Decentraliseringsprocessen, in Arbejdsmarkedpolitisk rbog. Copenhagen: Arbejdsministeriet.
Layard, R., Nickell, S., and Jackman, R. (1991), Unemployment. Oxford: Oxford University Press.
Lindbeck, A., Molander, P., Persson, T., Petersson, O., Sandmo, A., Swedenborg, B., and Thygesen, N. (1994), Turning
Sweden Around. Cambridge, Mass.: MIT Press.
Nilsson, T. (1997), Fackets nya roll. Frn frhandling till partssamverkan i lokalt utvecklingsarbete, Arbetsmarknad och
Arbetsliv, 3: 18592.
OECD (1996a), Economic Survey: Denmark. Paris: OECD.
(1996b), Employment Outlook. Paris: OECD.
Pedersen, J. (1994), Londannelsen p LO/DA-omrdet, in Arbejdsmarkedpolitisk rbog. Copenhagen: Arbejdsminister-
iet.
Pedersen, L. H. (1994), En konomisk analyse av Skraldemandsmodellen , Nationalkonomisk Tidskrift, 132: 115.
Pedersen P. J., and Smith, N. (1997), Arbejdsutbud och incitamenter, in Lisbeth Pedersen (ed.), Er der veje til fuld
beskaeftigelse? Copenhagen: Socialforskningsinstituttet 97: 13.
(1998), Low incomes in Denmark, 19801995, Working Paper 199819, Centre for Labour Market and Social
Research, University of Aarhus.
Regnr, H. (1997), Training at the Job and Training for a New Job: Two Swedish Studies, Swedish Institute for Social Research,
Stockholm University, Dissertation series no. 29.
Rosdahl, A. (1996), Youth unemployment in Denmark, in Youth unemployment in the North, Copenhagen: TemaNord
96: 592.
Rosholm, M. (1998), Transitions in the Labour Market, Ph.D. Thesis, Department of Economics, University of
Aarhus.
180 References
Rothstein, B. (1989), Marxism, Institutional Analysis, and Working Class Power: The Swedish Case, Politics and Society,
18: 31745.
Scheuer, S. (1992), Denmark: return to decentralisation, in A. Ferner and R. Hyman (eds.), Industrial Relations in the
New Europe. Oxford: Blackwell.
Smith, N. (1998), Okonomiske incitamenter till at arbejde, in N. Smith (ed.), Arbejde, incitamenter og ledighed. Aarhus:
The Rockwool Fund and Aarhus University Press.
Sveriges Offentilga Utredningar (SOU) (1998), Medling och lnebildning. Slutbetnkade frn utredningen om ett frstrkt
frlikningsmannainstitut, 98: 141.
Thoursie, A. (1998), Studies on Unemployment Duration and on the Gender Wage Gap, Dissertation no. 35, Swedish
Institute for Social Research, Stockholm University.
Torng, J. (1999), Workfare with welfare: recent reforms of the Danish welfare state, Journal of European Social Policy, 9:
528.
Zetterberg, J. (1993), Arbetslshet, arbetsmarknadspolitik och lnefrhandlingssystem, SOU 1993: 43.
7 The Dutch Miracle?
Cees Gorter
Introduction
The Dutch economic experience of the last fteen years is intriguing for both policy-makers and researchers. Put
briey, this experience can be regarded as a blend of wage moderation and part-time job growth. This development
has coincided with a remarkable rise in female labour participation from 35 per cent to 53 per cent in the period since
1982, reaching a level, however, that is still very low in comparison with other European countries. This development
has led to an economy in which part-time jobs, largely lled by females, form a key feature (see Fig. 7.1). The
prominent role of atypical employment in the Dutch economy is quite different from other countries in the
worldFreeman (1998) refers to the Dutch case as the rst part-time economy in the world.
In the past fteen years, the Dutch government has implemented several policies to enhance labour market exibility
and to increase international competitiveness. Important measures were taken to reform the labour market-
exibilization and modication of the social security systemand to alter the process of wage determination. In
particular, the government decided to abstain from wage control measures after 1982 if the social partners would agree
on wage restraint; since then social partners negotiate on wages and employment in a consensus-seeking environment
in which the public interest plays a crucial role. In other words, the social partners were given the freedom to engage in
decentralized bargaining since 1982.
Since 1994, a liberal-labour governmentan exceptional combination in Dutch politicshas been in power. This new
government has introduced a number of new policy directions which have led, most notably, to product market
deregulation, such as anti-cartel measures in line with European regulations and less stringent rules on business hours
(shops) and working time; and to the stimulation of job creation
182 Cees Gorter
Fig. 7.1. Employment growth in The Netherlands, 19871997

Note: Part-time (PT) employment: less than 35 hours per week.
for disadvantaged groups such as the long-term unemployed by subsidizing public-sector jobs at the lower end of the
labour market. There has also been an interesting shift in the focus of labour market policy in recent years: whereas the
emphasis used to be on combating unemployment, the most important issue in the late 1990s is boosting participation
in general and increasing the participation of females, ethnic minorities and the elderly in particular.
What, then, is the overall situation of the Dutch labour market at present? The positive features of the Dutch labour
market, the bright side, are most evident in employment growth, which has been remarkably high 1.8 per cent average
annual growth of jobs in 198393, compared with the OECD average of 0.4 per cent for the same period. The
unfavourable features, the dark side, consist especially in the high level of inactivity in The Netherlands; the labour
market participation rate is lowthe net rate, that is, employment over the working population, was just 62 per cent in
1994despite the rapid rise in female labour market participation in recent decades. Disability enrolment has been
excessively high in the past. The absolute number of disabled persons came close to one million in 1989, and the
absolute number of disabled workers is about double the registered number of unemployed people. The rate of
unemployment, however, is not exceptionally high in The Netherlandsbetween 5 and 6 per cent in 1997when
viewed from an international perspective.
The Dutch Miracle? 183
The National Debate

The political debate on the functioning of the labour market is centred, in The Netherlands, around three major
themes. First, the social security system is criticized for being too generous, and thus too costly at the aggregated level,
and not able to provide the proper incentives for non-working people to enter the labour market. Related to this, the
economic policy debate is concerned with the low labour participation rate due to signicant inactivity among persons
of working age. Second, the trend towards labour market deregulation and labour market exibilization is clearly
present, but its effects on labour market performance are not undisputed. Third, questions arise whether the
development of continuous wage moderation will remain to be benecial for employment growth. To discuss each of
these issues somewhat more specically, it is rst useful to sketch the context in which the trends of revising the social
security system, the deregulation and exibilization of the labour market and the policy of wage restraint emerged. As
in many industrialized countries, the Dutch economy was hit by major crises at the beginning of the 1980's:
unemployment increased at an enormous rate, prots declined and investment funds shrunk, many rms faced mass
dismissals or even went bankrupt, and the public budget decit reached unusually high levels. The economic trend
called for major corrections to the institutional system in general, such as large cutbacks in governmental expenditures,
and institutional arrangements in the labour market in particular. Due to this bad economic condition, the managing
role of the government was heavily criticized, in particular the school of supply-side economists argued that due to
institutional rules the smooth functioning of economic, that is labour and product, markets was strongly impeded (Van
Sinderen 1990). The response of the national government during the 1980s, led by Prime-Minister Ruud Lubbers, was
to cut public expenditures and public employment; to reform the social security system, for example through the
introduction of more stringent eligibility rules; and to strive for lower taxes and also social charges for rms. This type
of market-oriented thinking became prevalent among Dutch politicians and academics and established the origins of
the three major issues mentioned above. What are the main questions related to these topics?
The rst issue concerns the revision of the social security system since it became evident in the 1980s from the large
numbers of unemployment and disability benet recipients that public expenditures had grown to an unsustainable
level. Although originally social insurance programmes were regarded as a stabilizing force in the economy and
184 Cees Gorter
associated with a highly developed welfare state, the welfare lossesin particular via the distortionary tax wedge
because employers and employees largely nance the insurance themselvesbecame too high. The government was
therefore forced to reform the system.
An additional problem in the Dutch system arouse due to its highly complex and fragmented organizational structure
in which incentives to restrain costs at the implementation stage are missing (Stigter 1997). Moreover, it was protable
for both the employer and the worker to use disability as an exit route, instead of dismissal. As a result, a large group of
workers had become disabled for economic reasons (Aarts et al. 1993) and the number of persons on disability
schemes had grown excessively in The Netherlands. In fact, this aw of the system lasted until 1993 when it was
decided to install a Parliamentary Inquiry Commission to investigate the organizational structure of social insurance.
This inquiry basically showed that it was easy for redundant workers to enter social insurance schemes, especially
disability enrolment, whereas re-entering the labour market was hard due to a lack of incentives. Despite strong
protests in society, it was therefore decided, in the beginning of the 1990s, to tighten the access to social insurance
schemes and to redene employment disability more strictly. The rethinking on the optimal design of the social
insurance schemes has not yet come to an end. Disability schemes in particular, are being reconsidered because, after
an intial decline in the use of disablity benets in the mid-1990s their use has started to rise since 1998. What has
happened so far is that sickness schemes were privatized in 1996 making employers fully responsible for sickness
payments. Additionally, employers have to pay disability premiums according to their disability enrollmentthe higher
the rm's ow into disability, the higher the premium.
A second, but less intensive, debate is on the effectiveness of labour market reforms. Public and academic opinions are
divided on this topic. Supporters of deregulation, found for example in the Ministry of Economic Affairs, stress the
expected positive effects on economic performance through employment and production, and claim these effects
could even be stronger when combined with more competitive product markets. (Van Sinderen et al. 1994). Opponents
of reforms put forward that more external exibility on the labour market is not necessarily good for the economic
performance of rms. For example, Kleinknecht et al. (1997) argue that external exibility is not likely to positively
affect the performance of knowledgeintensive rms since investments in human knowledge are easily lost in this way.
The institutional change towards decentralized wage bargainingwhich will be discussed in full detail in the next
sectionthat resulted
in a long-lasting period of wage-restraint, was believed to have led to positive employment effects (De Vries 1995; Den
Butter and Van Vuuren 1998). Interestingly, the government was able to maintain a guiding role in the wage
determination process through its entitlement to intervene in wage formationalthough in practice, this right has not
been not exercised since 1982and from its attempt to persuade the social partners to moderate wages and strive
after general socioeconomic goals as well: job creation, working time reduction, and schooling and training. Thus,
bipartite consensus-building in the wage determination process is encouraged by a government which does not
interfere directly but plays a facilitating role. This organized decentralized arrangement is considered to be partly
responsible for The Netherlands' ability to moderate aggregate wages during the 1980s and 1990s. The Netherlands'
successful performance has become known as the Dutch Miracle or as the Dutch Polder Model. To explain this
miracle, however, causes besides wage moderation should be cited, and in particular the favourable effects of
international economic recovery. An issue that came to the fore in the 1990s is whether the instrument of continuous
wage moderation would be even harmful, in the long-run, due to the lack of incentives to be innovative and to remain
competitive on the international market. Kleinknecht (1994) argues that major rises in wages would cause a
schumpeterian process of creative destruction which would lead to better macro-economic outcomes than continuous
wage moderation.
It is noteworthy that the public debate on abolishing the mandatory extension of the collective labour agreementthis
extension or wider coverage is carried out by the Minister of Social Affairs and Employmenthas largely come to a
standstill. In The Netherlands, the large majority of workers are bound by a collective agreement or an extension;
union coverage, the coverage ratio of collective agreements as a percentage of the dependent labour force, was 72 per
cent in 1990 (Van den Wijngaert 1994). Union densitydened as union members aged under 65 as a percentage of
the working population is, however, much lower at 29 per cent in 1990 (Van den Wijngaert 1994). Of this group, 11
per cent of employees are covered by mandatory extension. It was argued by some economists that the mandatory
extension of the collective labour agreements prevented low wage competition, thereby making the wage structure too
rigid (Oudshoorn and Vijlbrief 1995; Ministry of Economic Affairs 1995; Van Wijnbergen 1996). To date, however,
mandatory extension has not been abolished. Moreover, recent research has shown that the mandatory extension
exerts little upward pressure on the wage level (Lever and Marquering 1995; Van Praag and Hop 1995; Freeman,
Hartog, and Teulings 1996).
186 Cees Gorter
Dynamics of Rigidities/exibility: The Dutch System of Flexicurity

The general trend to be observed isas in many European countriesone of deregulation or exibility. This makes
the labour market more exible, reduces adjustment costs, and goes hand-in-hand with the provision of more legal
protection, or security for exible workers. This typical Dutch combination is referred to as exicurity.
Employment Protection (Firing Rules) 87
An important institutional feature of the functioning of the Dutch labour market is, as in many Western European
countries, the high lay-off costs due to stringent regulations protecting workers against unreasonable lay-offs. In The
Netherlands, the law demands that an employer must have acceptable reasons for laying off workers (for more details
see Hassink 1996). In some European countries, the workforce can be temporarily decreased by resorting to short-
time work arrangements. The regulation of short-time work is, however, strict in The Netherlands as compared with
countries such as Belgium and Germany. It is only permitted when a rm faces abnormal difculties such as a high,
temporary, reduction of business activity due to structural factors, and approval must be obtained from the Minister of
Social Affairs. Not surprisingly, therefore, this method of temporary workforce reduction is not much used in The
Netherlands.
Dismissals may occur either at the individual or the collective level. In the latter case, Dutch law requires that at least
twenty workers must be redundant within a period of 3 months. For each individual dismissal, the employer is obliged
to gain permission from the Director of the regional employment ofce, according to the Labour Relations Act
(Buitengewoon Besluit Arbeidsverhoudingen) of 1945. This rule of prior admission is a phenomenon exclusive to The
Netherlands (Den Broeder 1996). After the decision has been taken by the regional employment ofce, both workers
and employers may appeal in court if they consider the decision to be unfair or incorrect. The ring procedure
becomes more complex when the employer wants to make collective dismissals. First, the warning period of
notice46 weeks in case of an individual dismissalmust be extended by one month. Second, the employer must
87
The discussion of employment protection rules will be based to a large extent on information systematically summarized in Den Broeder (1996), where the Dutch system is
compared with the German system.
consult the work council88 and the union, while the regional employment ofce must be informed, although the
employer no longer needs to gain permission from this ofce. Third, a social plan must be developed in which
severance payments are specied for the workers involved. Fourth, the indemnity to be paid to workers varies from
one to two month's pay, but is absent for individual dismissals. To summarize, the ring procedure is lengthy, and
therefore costly for the employer. An alternative approach for an employer to re a worker is to apply directly to the
court, where the judge decides relatively rapidly on the amount of nancial compensation to be paid to that worker.
For the worker, this procedure implies that his/her chances of being red may be higher, but, if this happens,
substantial nancial compensation for the job loss is likely. In practice, this second approach to ring has become
increasingly popular in the 1990s among Dutch employers; in fact, one out of three cases is dealt with in this new way
(Van der Heyden et al. 1995).
Given the complex, costly, and lengthy ring procedures for employers (see also, OECD 1996), it makes sense for
them to search for other methods of work force reduction.89 During the 1980s, employers extensively exploited the
disability enrolment option by encouraging workers who were not strictly disabled to be classied as disabled as an
alternative to ring. In addition to this abnormal exploitation of disability enrolment, rms can also use voluntary
quits to reduce their workforce (Hassink 1996). Finally, Dutch employers may hire workers on xed-term contracts or
employ them through temporary work agencies, as is discussed in the next section.
Regulation of Work EntryFlexible Labour Contracts

With respect to exible employment, a striking feature of the Dutch situation is the importance of temporary work
obtained through temporary work agencies (TWA). About 3 per cent of total employees hired per year are engaged
through TWAs, whereas the share of total exible employment is approximately 10 per cent. The popularity of exible
contracts can be largely explained by the demand for short-term employment exibility to accomplish downward
adjustments of employment. However, it has recently been emphasized by Kleinknecht et al. (1997) that employers also
have a strong interest in long-term
88
An organization governed by the employees themselves which protects the workers' interests in the rm.
89
However, it is worth mentioning that requests for dismissals made to the employment ofce are usually approved: for example, 76.8 per cent of the requests received
permission in 1992 (Mayes and Soteri 1994).
188 Cees Gorter
relationships based on commitment, because of their investments in rm-specic human capital.

As for the regulations on exible employment, these have three main features. First, workers on exible contracts, so-
called ex-workers, have less favourable access to unemployment benets than workers on permanent contracts. This
is because of the minimum threshold values set for the number of days worked before becoming unemployed;
moreover, pension funds may exclude workers on xed-term contracts. Second, employers may use exible contracts
for various purposes, such as to replace sick workers, but one of the main ones is to screen new workers. In addition,
exible contracts are not restricted to particular cases or to a maximum duration. Third, Dutch employers are free to
hire temporary workers through the TWAs. In 1996, a new collective agreement was agreed that makes labour
relations for agency workers more secure. This agreement between the employers' association of agencies and the
unions, sets minimum pension rights and gives continuation of employment after two years of temporary work. Since
January 1999, this agreement has been put into practice in legalization: the so-called Flex-Act. Hence, a worker's legal
position becomes stronger, the longer s/he works on a temporary basis, moving from contract to contract. The same
agreement undertakes to give greater exibility to the use of agency workers: for example, employers will be allowed to
hire such workers for longer durations. This combination of greater security for the (agency) worker with more
exibility for the employer neatly illustrates the concept of exicurity in the Dutch labour market.
It can be argued that part-time work has become a key feature of the Dutch labour market (OECD 1995). The share
of part-time work has grown rapidly from 15 per cent in 1975 to 35 per cent in 1994 (Hemerijck and Van den Toren
1996), mainly because of the rise in female labour market participation. In November 1996, the Dutch government
implemented a new rule providing equal legal status for part-time workers and full-time workers. Equal access to
unemployment benets has also been secured in recent years by removing the minimum thresh-olds for part-time
workers. In the near future, the government intends to give full-time workers the right to working hours reduced to 80
per cent of a full-time working week. This may be especially benecial to workers with children who do not wish to
interrupt their careers but also want to work fewer hours so that they can attend to their home duties. In the rst years
following 1982, potential increases in real wages were not used to reduce working time from 40 to 38 hours per week.
Further cuts in working time were desired by the unions, but they were impeded by the emergence of bottlenecks in
particular segments of the labour market.
Wage Determination
In the past, the Dutch government was closely involved in the formation of wagesfor an overview see Van de
Wijngaert 1994but since 1982 it has abstained from interference in wage determination. Prior to 1982, however,
direct intervention in wage formation by the government, under the Wage Determination Act of 1970, was a frequent
occurrence. The Wage Determination Act was modied in 1986 so that governmental measures on wages were
restricted to emergency situations and severe shocks. Although in practice this ended governmental interference in the
wage determination process, the Minister of Social Affairs and Employment threatened to intervene in 1992 and 1993
in the interest of reducing unemployment growth and easing the burden on the social security system.
More generally since 1982, the government still participates in tripartite wage debates with employers and employees'
associations to reach consensus on sensible wage growththat is, the government seeks to exert its inuence
through policy statements and tripartite consultations. Moreover, the government is also represented, with the social
partners, in the Social Economic Council (Sociaal Economische Raad) which comments on economic developments in the
country.90
Hence an important change in wage policy took place after the period 196382, during which the government had
directly intervened in wage determination on frequent occasions. As mentioned before, the Dutch economy was hit by
a major crisis at the beginning of the 1980s. 1982 saw the signing of a now famous agreement on collective wage
bargaining in Wassenaar, which stated that the trade unions and employers' confederations were primarily responsible
for wage bargaining. A second important outcome of the agreement was that employers acknowledged the unions'
interest in the development of employment, a goal which was to be achieved through the reallocation of labour (by
means of a general reduction of working time).91 The third key feature of the so-called Wassenaar Accord was that it
put an end to the application of private wage growth rates to the public sector. In essence, employers and employees
agreed on wage restraints in return for employment growth. In other words, the unions relinquished the right to wage
indexation while the employers recognised that job sharing was one of the means to stem the rise in unemployment
(van Veen 1987).
From 1982, a decentralization process was set in motion whereby
90
Since 1995, however, consulting the Social Economic Council is no longer mandatory.
91
However, two years later in 1984 the employers opposed further working time reductions (Albeda 1985).
190 Cees Gorter
collective labour agreements were reached at the industry level and also, at an increasing rate, at the rm level.92
Nevertheless, central governmental controls did not entirely disappear, since the government maintained its right to
intervenealthough only in exceptional circumstances, as mentioned earlierand it was also able to inuence the
consensus-building process at the national level through the Foundation of Labour (Stichting van de Arbeid), a joint body
comprising the employers' and employees' federations.93 This kind of governmental persuasion exists because the
outcomes at the national level serve as guidelines for industries. In this sense, one can classify this new system of
collective wage bargaining as a mixture of decentralized and centralized attributes.
A decentralized trend in collective bargaining is clearly present in the Dutch economy. This tendency towards
decentralization has been accompanied by the central government's role as the guiding agent in the wage formation
process. Institutionalized consensus-building has been an important feature of the Dutch system in the past, and it is
still an important ingredient of the wage formation process. This feature is also evident in the follow-up to the
Wassenaar agreement: the Memorandum A New Course signed in December 1993 which provides for wage
moderation together with measures for rms' investment, job creation, working time reduction, and training. In short,
the wage formation process is not simply passing through a phase of decentralization, it is moving towards what has
been called centralized decentralization (Due et al. 1994).
To conclude, the institutional setting of wage formation is currently moving towards increasing decentralization on the
one hand, with more rm-level agreements, especially in large companies like PTT, Philips, Shell, and Unilever; while
maintaining centralized consensus-forming negotiations on the other, thereby providing a guideline for industries.
Hemerijck and Van den Toren (1996) describe the new-found bipartite consensus as follows: . . . The 1990s accords
reinstated a policy priority for voluntary wage restraint in the pursuit of business protability, investment and
employment growth and job redistribution. All the 1990s agreements afrmed the freedom of sectoral collective
bargaining and the primacy of self-regulation in industrial relations. In practice, the central employers' and workers'
organizations frequently meet at the national level to formulate joint declarations on bargaining and employment
conditions at decentralized levels, before actual collective bargaining takes place. This Dutch habit of consensus-
building may be
92
Firm-level agreements were initially used by employers to try to de-standardize labour contracts and the working hours regime.
93
Employers and employees have equal representation in the Foundation of Labour.
due to what Axelrod (1984) has called the shadow of the future, which refers to the idea of we meet again.
How long this state of consensus will last, and whether it is capable of coping with substantial uctuations in the
business cycle, remains to be seen. If it really was the experience of a profound economic crisis that made unions and
employers' associations decide to join forceswhich can be regarded as a case of punctuated change (Krasner
1988)then one can doubt whether wage moderation will continue in periods of lasting economic prosperity.
So far, a description has been given of the institutional changes in the process of wage determination. But how did real
wages develop throughout this period? How much are real wages, set in collective agreements, reduced and if so, which
groups contributed to this development most strongly? The outcomes on real wage rates resulting from bargaining are
as follows (gures are based on Salverda, 1999). Wage inequality is low, relative to other countries, and hardly changes
over the period 197994 because wage declines occured both at the top and the bottom of the wage distribution. Real
wage rates, as agreed upon through bargaining, are reduced considerably. For the total group of wage earners involved
in collective agreements real wages drop about 10 per cent during the period 197985, and 5 per cent overall between
1979 and 1994. Wage earners in the public sector face a much stronger reductiondropping 26 per cent in the period
197985than those working in the private sector who drop only 8 per cent in the same period. Finally, minimum
wages have fallen signicantly for both adultsdropping 20 per cent in the period 198096and youthdropping
35 per cent in the same period.
Working-Time Regulations
A further important step toward less regulation on working hours was taken in 1996, when new rules were gradually
implemented to facilitate working at irregular hours and to allow for individual variations in the length of the working
week. It is possible to achieve even more exibility when negotiations between employers' and workers' associations
take place through collective bargaining agreements or through co-determination at the rm level, that is, the rm's
management bargains with the work council. In this way, packages of working-time arrangements may be produced in
which working time is reducedsuch as a shortening of the working week as desired by the unionsand the
employer's exibility is enhanced, facilitating work at irregular hours and variable working time. This has happened, for
example, in the banking sector, PTT Netherlands and the civil service.
192 Cees Gorter
Social Security System

The Dutch social security system is quite extensive and, according to the 1994 OECD survey (and Visser 1997), it can
be described thus: . . . Labour market policy remains characterised by generous income support measures.
Replacements rates ( . . .) for all types of family situations and for all unemployment spells are among the highest in the
OECD area.
Since the mid-1990s, a number of changes have been made to the design of the social security system (Vijlbrief 1992;
Visser 1997). A major change that occurred due to the Wassenaar accord was that the link (koppeling) between the
development of contractual wages in the private sector and benet incomes has been abandoned since 1982. The
Labour Party managed to reinstate this link in 1990, although it was then made conditional on the inactivity ratiothat
is, the number of benets recipients per 100 employed personswhich should be lower than or equal to 82.6.
Important modications of the system have been:
(1) the reduction of unemployment benets in 1985 from 80 per cent to 70 per cent of the pre-unemployment
wage;
(2) the passing of the new Unemployment Insurance Act in 1987, which made the maximum duration of
unemployment benets more closely dependent on the period for which insurance contributions have been
paid;
(3) the decline in statutory real minimum wages during the 1980s of about 15 per cent on average and even more
among young people (Mot and Teulings 1990), which also has direct consequences on the level of social
benets, given that the minimum social benet level amounts to 70 per cent of the minimum wage; and
(4) the abolition of the Sickness Act in 1996, thereby transferring a social right into an industrial right for workers
(cf. Hemerijck and Van den Toren 1996).
In addition to this major change in the Unemployment Insurance Act, the assessment procedure of the degree of
disability was also modied since the labour market consideration in this assessment was eliminated. This
modication was motivated by the fact that the Dutch disability programmes exhibit a large hidden unemployment
component. Further measures were taken in 1992, 1993, and 1994 in which the access to disability and sickness
benets was limited and nancial incentives for employers and employees were put into use. As a result, decreasing
volumes of sickness and disability beneciaries were observed (Stigter 1997).
It is well known that in welfare states the tax wedge positively inuences the use of social insurance schemes due to the
downward spiral
of social security and labour participation; and this also holds for the Dutch case (Den Butter and Vijlbrief 1992). By
downward spiral it is meant that a rise in the number of insurance claimants leads to higher governmental
expenditures on social security which has to be nanced through higher taxes. The latter results in less activity in the
labour market due to higher, distorted wages and this, in turn, gives rise to more payments for social security. The
Dutch level of the tax wedge reached a peak in the major recession of 1983 at 48 per cent for the so-called modal
incomes, and did not decrease much afterwards, except for the minimum incomes for which it did fall. Due to the
price and volume measures taken in the social security system during the 1980s and 1990s respectively, the use of social
insurance schemes has been reduced and hence, the impact of the downward spiral has become weaker (Den Butter
and Kock 1997). In 1996, the Dutch government introduced a wage-subsidy to lower the employer's wedge for low-
paid workers. This measure of tax-exemption affects about one out of six workers and consequently, it appears to
form a major cost component: more than 2 billion guilders in 199798 (Salverda 1999). According to Salverda, the
employment effects of the wage subsidy are small if there are any at all.
Conclusion
External exibility to adjust working hours and the level of employment to output uctuations seems to be obtained by
the extensive use of part-time and exible contracts, respectively (see also Delsen and De Jong 1997). The use of
exible contracts to adjust the demand for employment is supported by permissive regulations on xed-term contracts
and the hiring of temporary workers through TWAs. This strategytogether with measures designed to reduce the
costs of the social security systemforms an essential part of the Dutch policy of labour market deregulation. Table
7A in the Appendix gives an overview of the major changes in institutional arrangements. Hence, ring regulations do
not necessarily impede downward employment adjustment. Nevertheless, it is true that the need to gain permission
before an individual dismissal prolongs the ring procedure. The permissive regulation of exible work may, in fact,
result from an increase in the complexity of ring rules (Schmann et al. 1995).
To conclude, exible working arrangements exert a positive effect on employment dynamics and push up temporary
employment. Not surprisingly, Dutch employers and their main representatives are in favour of exibilization and
deregulation. This is not to imply, however, that employers do not recognize the importance of improving the quality
of
194 Cees Gorter
the production process, as has become clear in the recent qualitative expansion of collective bargaining agreements.
The other side of the coin may be, however, that deregulation and exibilization create a dual labour market with
insiders retaining permanent positions while semi-outsiders occupy inferior labour market positions in which
temporary work is accepted because no permanent job is available. The trade unions, therefore, are essentially against
exible labour contracts,94 but by adopting this stance they fail to attract ex-workers as members. Since the union
does not then bargain for these workers, this may result in an even weaker labour market position for semi-outsiders
or ex-workers. This is clearly a dilemma for unions now confronted with a decline in their memberships from 37 per
cent in 1979 to 25 per cent in 1987 (Van den Wijngaert 1994).
Labour Market Performance in the Nethrlands
The Structure of Unemployment

It is necessary to return the 1980s in order to trace the development of unemployment, since in that period the Dutch
economy was confronted by qualitative and quantitative mismatches of demand and supply. Figure 7.2 shows that the
unemployment rate peaked in The Netherlands in 1983 at 11.7 per cent. This peak was followed by a gradual decline
to 5.3 per cent in 1992. After an increase in the unemployment rate between 1992 and 1994 due to a short recession,
unemployment declined again subsequently.
Unemployment is unevenly distributed among individuals in The Netherlands, as in many European countries. This
can be seen from Fig. 7.3, in which a signicant rise in the relative share of long-term unemployment can be observed
during the 1980; in fact, the duration of unemployment tripled in the period 197790 (De Beer 1996). The
development, in 199095, in the absolute value of long-term unemployment makes clear (see Fig. 7.4) that the
proportion of long-term unemployment has been persistently high over time: in 1995, the share of long-term
unemployment once again rose to 53 per cent, whilst it had been just below 50 per cent in 1992 and 1993. Another
sign of the unequal distribution of unemployment over socio-economic groups is that the proportion of low-skilled
unemployed is high: in 1995, 22 per cent of the
94
Further analysis of part-time and exible jobs will consider whether this dualism occurs in The Netherlands, and if so, which categories of worker it affects.
Fig. 7.2. Unemployment rate in the Netherlands, 19801997 (% of the total labour force)
Fig. 7.3. Long-term unemployment, 19801990

unemployed had only nished primary school. Ethnic minorities are clearly over-represented in the pool of
unemployed, accounting for 25 per cent of total unemployment in 1995, and it should be emphasized that this group is
mainly resident in the cities of Amsterdam and Rotterdam, where the unemployment rate and especially
unemployment duration is generally high, and extremely high for this group in particular. Finally, the unemployment
rate among school-leavers in 1995 was 11 per cent compared to the average rate of about 6 per cent.
196 Cees Gorter
Fig. 7.4. Development of the absolute number of long-term unemployed, 19901995
Note: Based on the new denition of unemployment.

To sum up, the unemployment problem in The Netherlands is less a matter of becoming unemployed than of
remaining unemployed for too long (see Gorter 1991). It is therefore of utmost importance to identify the
determinants of unemployment duration, so that the socio-economic groups at risk can be identied. Several studies
on unemployment duration have been carried out in the last decade (Van den Berg 1990; Theeuwes et al. 1990; Gorter
and Gorter 1993; Gorter et al. 1993) in which the emphasis is usually also on the impact of unemployment benets on
duration. Most of these studies, however, show that the level of unemployment benets does not inuence
unemployment duration positively, indicating that there is no push effect of unemployment benets on
unemployment durationfor an overview see Groot and Jehoel-Gijsbers (1990). It is however found (Lindeboom
and Theeuwes 1993) that coming to the end of the benet entitlement period affects the leaving rate of unemployment
positively. As regards the role of personal and labour-market related features, these studies of unemployment duration
reveal the following pattern: leaving unemployment is relatively more difcult for the elderly, for those who have
already been unemployed for a long period, those living in regions or cities with unfavourable labour market
conditionsespecially ethnic minorities living in big citiesand the poorly educated. Some studies also
indicate that the group of highly-educated persons take more time to obtain a job.
The Dynamics in Employment

Investigating labour market exibility by means of unemployment outows is not straightforward when the exits from
unemployment include transitions into non-participation as well. A more appropriate way to look at exibility in the
labour market might therefore be to investigate employment owsows into and out of employment, and job-to-job
owsin order to assess the adjustment capabilities of workers to constantly changing economic circumstances. For
this purpose, the information on employment ows in The Netherlands (197191) gathered by Gautier (1997) and
based on administrative data from the Social Security Council, can be used. In Gautier's study, employment outow is
computed by summing the ow from employment into unemployment and into non-participation and employment
inow is obtained by using this outow gure together with information on the change in the stock of employment.
The pattern of employment inow and outow over the period 197191 is shown in Fig. 7.5.
During the crises at the beginning of the 1980s, employment inow
Fig. 7.5. Employment ows in the Netherlands, 19711991
Note: EIR = Employment Inow Rate, EOR = Employment Outow Rate, SUM = EIR + EOR, (shaded areas are
recession periods).
Source: Gautier, 1997.
198 Cees Gorter
collapsed whereas employment outow increased. In the years that followed, 198291, employment inow increased
steadily from 6 per cent to about 10 per cent in 1990 and dropped again in 1991, whilst employment outow was more
stable during this period at around 8 per cent. Gautier (1997) also shows that total worker turnover, dened as the sum
of employment inow, employment outow and job-to-job movements, was 33.5 per cent on average in the period
197191, which was much lower than the levels in Italy (67 per cent), France (59.6 per cent), Germany (43.8 per cent),
and Denmark (57.3 per cent), but much higher than the level in Spain (15.9 per cent) (Appendix Table 7B). This result
suggests that Dutch workers adjust relatively slowly to changing economic conditions.
At the individual level, however, the picture of job mobility is not particularly unfavourable: average job tenure in the
late 1990s is about seven years (Ministry of Social Affairs and Employment 1996), and even more importantly, the
proportion of jobs with a duration of less than one year is the highest among the countries of Western Europe (OECD
1995). The existence of a large share of completed job spells of short duration was also reported by Lindeboom and
Theeuwes (1991), who studied the determinants of job duration in The Netherlands. They found that the Dutch
pattern of job mobility is one of a labour market in which a substantial number of workers with life-time jobs co-exists
with parts of the labour market where very short job duration are prevalent. High turnover co-exists with permanent
job attachment. Workers with high turnover rates appear to be mainly women with part-time jobs, whereas middle-
aged and older men are usually employed in full-time jobs and thus experience long job duration (Lindeboom and
Theeuwes 1991). In this respect, women can be regarded as losers in the long run, since they are not likely to receive
much support in developing specic human capital and their job careers in general when employed in these kinds of
short-duration jobs.
Brief discussion is required of a second issue concerning employment dynamics at the micro-level: the inow of
employment. In particular, the hiring strategies of employers will be highlighted with respect to their choices between
employed and unemployed candidates when lling vacancies. A recent study by Russo et al. (1997a) has shown that
educational requirements are of decisive importance for the outcome of the recruitment process. From their empirical
analysis of the Dutch labour market in the period 199194, Russo et al. conclude that when higher educational
standards are required, employers are more inclined to hire employed job-seekers; conversely hiring the unemployed
when seeking lower educational standards. The preference of employers to hire non-unemployed people is made plain
by the fact that, of the 700,000 newly-created
jobs in 198390, fewer than 100,000 were lled by unemployed workers (Visser 1997). Again, important reasons for
hiring newcomers like students leaving school or women re-entering the labour market, instead of the long-term
unemployed are an actual or perceived lack of training, work-experience and educational skills among the unemployed.
Part-Time and Flexible Jobs

A crucial feature of the Dutch labour market is the large share of atypical employment: in 1996 part-time workers
represented 36.7 per cent of total employment, while temporary workers were 13.3 per cent. The presence of atypical
employment calls for further analysis.
Three out of four part-time jobs are held by women. Russo et al. (1997b) analyse the main features of newly-created
jobs by using data on lled vacancies during the period 199194, and nd that part-time jobs are more likely to be
posted for positions requiring a lower level of general education and no experience. This suggests that part-time jobs
may be concentrated in the secondary segment of the labour market.
The focus now shifts to temporary employment in order to establish the winners and the losers in this increasingly
important segment of the Dutch labour market. The trend towards more exible work in the Dutch labour market is
obviously desirable from the employers' point of view. However, although some workers may want more exibility in
the form of exible working hours, number of hours, and temporary leave, it is debatable whether less regular work is
used to employ the weaker groups in the labour markethence creating a segmented labour market. Russo et al.
(1997b) address this issue by looking at temporary contracts in The Netherlands.95 This investigation of exible work
in the Dutch labour market has found that it does not seem to be persistently allocated to less favoured groups.
Moreover, it becomes clear that the over-representation of women in temporary jobs is not at variance with their
preferences. Thus, as far as temporary work in the late 1990s is concerned, there appears little risk of a dual Dutch
labour market in which winners and losers are clearly identied. Russo et al. (1997b) also nd evidence of a recent
structural shift towards the use of temporary jobs by both rms and workers, which raises the question as to whether
segmentation will become stronger if this exibilization trend continues.
95
As already mentioned, temporary contracts in The Netherlands also include contracts stipulated with an agency which serves as a mediator between the employer and the
employee.
200 Cees Gorter
Fig. 7.6. Unemployment, vacancies, and matches, The Netherlands, 19801993

Source: Gorter, Nijkamp and Pels, 1997.
Development of Labour Market Efciency Over Time

As outlined in earlier, the Dutch labour market was faced by enormous discrepancies between supply and demand at
the beginning of the 1980s. To illustrate this in more detail, Fig. 7.6 shows the opposite movements of unemployment
and vacancies in the period 198093.96 Moreover, it can be seen that the ow of lled vacancies increased considerably
during the 1980s, reaching its maximum level in 198788 and, after a decline in 198990 showing a second peak in
1991.
There is a debate in The Netherlands over the extent to which a worsening of the functioning of the labour market has
contributed to high unemployment. Gorter et al. (1997) have applied a matching approach to the Dutch labour
market in order to investigate the development of labour market efciency over time allowing for regional differences
in performance. To this end, they estimate a matching function that species the relationship between the ow of lled
vacancies and the stocks of unemployment and vacancies. Their estimation results show that, for the period 198093,
regional differences in labour market efciency
96
The sharp decrease of unemployment in 1989 was mainly due to a change in denition, but at the same time there was an improvement in the measurement of
unemployment which also led to a drop.
are not responsible for regional differences in the ow of lled vacancies.

As regards the general pattern of labour market efciency over 198093, Gorter et al. (1997) have found an interesting
cyclical movement of labour market performance (see Fig. 7.7). The outcomes presented in Fig. 7.7 show that
efciency improved during the rst recession (198183), continued to be high during the recovery period (198488),
dropped signicantly in the boom (198990), and nally returned in the second major recession (1993) to the level
reached during the previous recession. Three important conclusions can be drawn from these ndings. First, a
relatively high level of efciency during a period of rapid employment growth indicates that a large pool of potential
employees is available. Second, the substantial improvement of efciency in a major recession period may at rst sight
be somewhat surprising. However, this nding of an increased ability to produce matches between unemployed job-
seekers and vacancies is remarkably consistent with the observation, at the aggregate level, that in many industrialized
countriesthe United States, Germany, and The Netherlandsthe ow from
Fig. 7.7. Labour market efciency over the business cycle in The Netherlands, 19801993
Notes: (k) labour market efciency.
k adjusted refers to efciency (k) adjusted for the change in the denition of unemployment from 1989 onwards.
202 Cees Gorter
unemployment to employment increases during a recession (Burda and Wyplosz 1994; Schettkat 1996).97 Third, the
functioning of the labour market in 1993 was at least as efcient as it had been more than a decade beforea period
covering about two cycles. This last conclusion means that during a period that can characterized, from a
policyoriented perspective, as one of deregulation and exibilization, labour performance moved basically counter-
cyclically but did not change to a level in 199193 that differed from its initial level at the beginning of the 1980s.
The Impact of Wage Restraint: Empirical Evidence

As regards to the impact of wage moderation on economic performance in The Netherlands, the evidence so far is not
conclusive. The organized-decentralized arrangement of the 1980s and 1990s in The Netherlands may be held partly
responsible for reigning in wage growth at the macro-level. The huge rise in employment that began in the second part
of the 1980s was to a large extent stimulated by this strategy of wage restraint. The Ministry of Social Affairs and
Employment (1990) reported that wage growth in the period 19839 was 1.2 per cent per annum lower than it would
have been without the new institutional arrangements. However, besides wage moderation, the favourable effects of
international economic recovery should also be cited to explain this ourishing performance. The Dutch Central
Planning Bureau (CPB 1991) has estimated that during the period 198590 wage moderation was twice as important
for job growth as the international economic upswing during the same period. In absolute terms, an additional
employment growth of 415,000 persons is credited to the wage moderation policies through which real wage growth
was kept continuously lower than productivity growth in that period (CPB 1991). This observation is also made by
Den Butter and Van Vuuren (1998) who analysed the role of wage moderation in the trends for economic growth,
productivity and employment in The Netherlands by using a stylized reduced-form macro-economic model.
Simulation experiments with this model supported the importance of wage restraint in the Dutch Polder Model.
More specically they conrmed that the interest of Dutch unions in the aggregate employment and participation
levels contributed to the acceptance of the real wage moderation strategy. They also found that employment effects are
much more effectively achieved by wage moderation than by a reduction in working time.
97
In addition, recent research on the lling of vacancies at the micro-level (Russo et al., 1997a) has also conrmed that when there is an excess supply on the labour market,
employers are more likely to hire unemployed candidates than employed workers.
Some other studies on the causes of the Dutch miracle have also pointed at other factors responsible for the long-
lasting period of employment growth. For example, Delsen and De Jong (1997) have argued that in the extended
employment growth period, the appreciation of the Dutch Guilderlinked to the value of the German Markdid
not lead to a decline in international competitiveness and protability, as occurred in Germany, because wage
moderation, supported by the unions who recognized the importance of employment, was strong enough to
counteract the appreciation effect. This role of the real depreciation of the Dutch Guilder is also empirically supported
by the simulation exercises of Den Butter and Van Vuuren (1998). The extended duration of the growth period is also
partly attributed by Delsen and De Jong (1997) to macro-economic luck. By this they mean that the Dutch economy
experienced a sequence of favourable, temporary macro-economic events that boosted effective demand: rst, the
international economic recovery in the late 1980s; second, the positive impact of the German unication on the Dutch
economy; third, the dominance of the agricultural and chemical sectors in The Netherlands which remained almost
unaffected during the recession of 199394; and fourth, the historically low interest rate since 1996 that caused a hugh
rise in durable consumption, such as housing. Further doubts on the succesfulness of the Dutch model have been put
forward by Salverda (1999). He concludes that the period of wage moderationthat actually started a few years before
the famous Wassenaar accord was madehas not resulted in employment growth per capita. Salverda argues therefore
that it appears that the 1982-accord deserved less credit for starting wage moderation than for continuing it.
International competitiveness as reected by modest export growth gures per capita, relative to other industrialized
countries, also did not improve during the period 197996 (Salverda 1999). A development in favour of the Polder
Model's effectiveness was the deliberate choice of social partner to moderate wages in the 199293 when
unemployment began to rise again.
Conclusions
The institutional setting of the Dutch labour market can be described as harmonious as regards industrial relations.
The Dutch system of wage formation, in which a decentralization tendency is accompanied by centralized consensus-
forming negotiations, is at present generally supported by employers, unions and the government. As far as labour
204 Cees Gorter
market regulation is concerned, the Dutch economy is typied by a concept of exicurity in which security for
workers is matched by exibility for rms. In particular, external exibility in the Dutch labour market is achieved by
the extensive use of part-time and exible contracts, a practice sustained by permissive regulations on xed-term
contracts and the hiring of temporary workers through TWAs. On the other hand, protective regulations apply to
workers on permanent contracts via complex and lengthy ring procedures, and to an increasing extent also to ex-
workersespecially those employed through TWAs. Finally, the eligibility conditions for unemployment benet have
been made more stringent in the last ten years. To sum up, the institutional representation of The Netherlands may be
viewed as relatively stable over the late 1980s to 1990s.
However, the crucial question is this: how does this system work in terms of labour market performance? Not too
badly, seems to be the answer. First, employment has been growing rapidly in The Netherlands since the mid-1980s,
although it should be immediately pointed out that this growth is particularly high for part-time jobs and, to a lesser
extent, temporary jobs. The darker side of Dutch labour market performance is less visible in the unemployment rate
which is relatively low when seen from a European perspective; the labour market participation rate, however, is
remarkably low compared with other European countries, and particularly among the elderly and women. Despite a
low unemployment rate of about 56 per cent, long-term unemployment accounts for around 50 per cent of total
unemployment in the Dutch labour market and appears to be concentrated among the elderly and ethnic minorities,
who suffer most markedly from their lack of appropriate skills. Second, the dynamic pattern of employment ows
shows that the Dutch labour market is, on average, characterised by a relatively low turnover rate of about 3035 per
cent of total employment per year. However, the distribution of job durations reveals that this relatively low average
value for turnovercompared with other European countriesis a combination of low turnover rates for full-time
workers, who are mostly middle-aged and elderly males, and high turnover rates for part-time working females. As
regards the ow into employment, a striking nding is that during the period of job growth, it was mostly non-
unemployed persons, in particular school-leavers and females, who lled the newly created part-time jobs. Third, the
functioning of the Dutch labour marketmeasured by its efciency in matching unemployment and vacanciesat the
beginning of the 1990s appears to have been as good or as bad as it had been over a decade before. Since then, several
labour market deregulation and decentralization schemes have been implemented, but these apparently have not
coincided with a structural
change in labour market efciency. Fourth, the Dutch labour market does not show evident signs of a segmented
labour market in which exible work is persistently allocated to less favoured groups in the labour market.
To conclude, deregulation and exibilization measures in the Dutch labour market since 1980 have not inuenced the
degree of employment protection or wage exibility, or else they have only done so to a limited extent. However, they
have clearly promoted the extensive use of temporary and exible work, and also brought an increase in working hours
exibility. In the light of these observations, it seems unlikely that this trend of deregulation and exibilization has
played a major role in the favourable Dutch experience of job growth. This Dutch miracle can be ascribed principally,
but maybe not fully, to a continuous course of wage moderation based on an organized-decentralized wage bargaining
system in which social partners reach agreements whereby wage restraint is exchanged for socio-economic goals like
employment creation and working time reduction. The most recent example of this manner of consensus-building
between the Dutch social partnersa new central accord signed in November 1997shows an exchange of wage
moderation by the unions for the employers' responsibility to organize and nance schooling activities, together with
the introduction of exible wage components such as a wage premium in the case of good performance by the
individual worker. If this new agreementintended to serve as a guideline for the new collective agreements at the
sectional and rm levelagain proves to be successful in terms of employment growth, then it may increase the
labour participation of the long-term unemployed, the partially disabled, and welfare recipients who have not yet
beneted from the favourable economic development of the late 1980s and 1990s.
Acknowledgements
The author wishes to thank Jacques Poot and Wiemer Salverda for their useful comments on an earlier versions of this
report.
206 Cees Gorter
Appendix
Table 7A. Institutions and Deregulation, the Netherlands
Institutions/regulations with respect Main feature(s) in 1997 Major change(s) since 1980
to
Employment protection costly ring procedures (prior ad- None
mission)
Flexible work use of temporary work agencies Central agreement 1996: easier use
of ex-workers, more legal rights for
ex-workers exicurity
permissive regulation towards exi- Flex-Act 1999: legalizing the central
ble/temporary contracts agreement of 1996
Wage determination (bargaining) organized-decentralization Wassenaar accord 1982: trade
unions and employers associations
responsible for wage bargaining
Working time None Wassenaar accord 1982: reduction
of working time [to 38 hours per
week]
1996: increase in working hours
exibility
Social security system generous income support measures 1980s: decline in real minimum
wages (-15%)
1985: reduction of unemployment
benets (-10% previous wage)
1987: more stringent eligibility con-
ditions for maximum duration of
unemployment
199294: Tightening of eligibility for
sickness and disability payments
1996: abolition of Sickness Act
(privatization)
References 207
Table 7B. Selected International Studies on Employment Flows

Author(s)/year Country Period In (%) Out (%) Sum (%)
Albk and Sor- Denmark 8091 28.9 28.4 57.3
ensen (1994)a
Antolin (1995) Spain 7791 7.9 7.4 15.9
Blanchard and USA 6886 41.2 34.7 75.9
Diamond (1990)
Burda and Wy- France 87 28.9 30.7 59.6
plosz (1994)
Germany 87 22.3 21.5 43.8
Japan 87 9.3 8.7 18.0
Gautier and Holland 7191 17.2 16.3 33.5
Broersma (1994)
Leonard and van Belgium 21.0 20.5 41.5
Audenrode
(1995)
Mumford and Australia 8087 + -
Smith (1996)
Contini and Italy 8591 33.0 34.0 67.0
Revelli (1994)
a
Manufacturing sector only.
Notes: in = inow, out = outow, sum = turnover; + indicates pro-cyclical movements; - denotes counter-cyclical movements; Yearly ows
were calculated as: (12 monthly ows).
Source: Gautier, 1997.
References
Aarts, L. J. M., Dercksen, W. J., and Jong, P. R. (1993), Arbeidsongeschiktheid, een Internationale Vergelijking, Sociaal
Maandblad Arbeid. 48: 75569.
Albk, K., and Sorensen, B. (1994), Worker ows and job ows in Danish Manufacturing. Copenhagen: University of
Copenhagen (mimeo).
Albeda, W. (1985), Recent Trends in Collective Bargaining in The Netherlands, International Labour Review, 124.
Antolin, P. (1995), Gross worker ows: how does the Spanish evidence t the stylized facts. Paper presented at the Centre for
Economic Policy Research workshop, Berlin.
Axelrod, R. (1984), The Evolution of Co-operation. New York: Basic Books.
Beer, P. de (1996), Het onderste kwart. Werk en werkloosheid aan de onderkant van de arbeidsmarkt. Rijswijk: Sociaal en
Cultureel Planbureau.
Berg, G. J. van den (1990), Search behaviour, transitions to nonparticipation and the duration of unemployment,
Economic Journal, 100: 84265.
Blanchard, O. J., and Diamond, P. (1990), The cyclical behavior of gross ows of U.S. workers, in Brooking Papers on
Economic Activity. Cambridge, Mass.: MIT Press.
Broeder, C. den (1996), Institutions at work. Commitment and exibility on the German and Dutch labour markets. Research
Memorandum 131, The Hague: Central Planning Bureau.
208 References
Burda, M., and Wyplosz, C. (1994), Gross Worker and Job Flows in Europe, European Economic Review, 38: 1287315.
Butter, F. A. G. den, and Vijlbrief, J. A. (1992), Sociale Zekerheid en Arbeidsparticipatie, Tijdschrijft voor Politieke
Economie, 14: 3749.
and Kock, U. (1997), Sociale Zekerheid en Arbeidsmarktdynamiek, in Nederland, Research Memorandum
199757, Free University Amsterdam.
and Vuuren, A. P. van (1998), Het poldermodel en de keuze tussen werk, inkomen en vrije tijd, Tijdschrift voor
Politieke Economie, 20/4: 3962.
Central Planning Bureau (1991), De werkgelegenheid in de jaren tachtig, Working Document 41, The Hague: Central
Planning Bureau.
Contini, B., and Revelli, R. (1994), Job creation and destruction in Italy. Working Paper, Rome: Bank of Italy.
Delsen, L., and Jong, E. de (1997), Het wankele mirakel, EconomischeStatistische Berichten, 82: 3247.
Due, J., Madsen, J. S., Jensen, C. S., and Pedersen, L. K. (1994), The Survival of the Danish Model. Copenhagen: DJOF
Publishing.
Freeman, R. B. (1998), War of the models: Which labour market institutions for the 21st century, Labour Economics, 5:
124.
Hartog, J., and Teulings, C. N. (1996), Pulling the plug. OSA-report W144, The Hague: SDU/OSA.
Gautier, P. A. (1997), The ow approach to the labour market, Dissertation, Amsterdam: Tinbergen Institute/Thesis
Publishers.
and Broersma, L. (1994), The timing of labor reallocation and the Business cycle. Amsterdam: Tinbergen Institute
(mimeo).
Gorter, C. (1991), The dynamics of unemployment and vacancies on regional labour markets, Dissertation,
Amsterdam: Tinbergen Institute/Thesis Publishers.
Nijkamp, P., and Rietveld, P. (1993), Barriers to Employment: Entry and Re-entry Possibilities of Unemployed
Job Seekers in the Netherlands, De Economist, 141/1: 7095.
Nijkamp, P., and Pels, E. (1997), Vacancy Dynamics and Labor Market Efciency in the Dutch Labor Market,
Growth and Change, 28: 173200.
Gorter, D., and Gorter, C. (1993), The Relation between Unemployment Benets, the Reservation Wage and Search
Duration, Oxford Bulletin of Economics and Statistics, 55/2: 199214.
Groot, W., and Jehoel-Gijsbers, G. (1990), De invloed van de uitkeringshoogte op de werkloosheidsduur, Economische
Statistische Berichten, February: 17781.
Hassink, W. H. J. (1996), Worker ows and the employment adjustment of rms, Dissertation, Amsterdam:
Tinbergen Institute/Thesis Publisher.
Hemerijck, A., and Toren, J. P. van den (1996), The resurgence of corporatist policy coordination in an age of globalization. The case
of the Netherlands. Rotterdam: Erasmus University/Department of Public Administration (mimeo).
Heyden, P. F. van der, Kaar, R. H. van het, Knegt, R., Roebroek, J., Sol E., and Wilthagen, A. C. J. M. (1995), Instituties
van de arbeidsmarkt: een retrospectieve studie. OSA-Report V43, The Hague: SDU/OSA.
References 209
Kleinknecht, A. (1994), Heeft Nederland een loongolf nodig, Tijdschrift voor Politieke Economie, 17/2: 524.
Oostendorp, R., and Pradhan, M. (1997), Flexibiliteit en prestaties, Economische Statistische Berichten, November:
86970.
Krasner, S. (1988), Sovereignty. An international perspective, Comparative Political Studies, 21: 6694.
Leonard, J., and Audenrode, M. van (1995), Corporatism run amok: job stability and industrial policy in Belgium and the United
States. Berkeley: Institute of Industrial Relations (mimeo).
Lever, M. H. C., and Marquering, W. A. (1995), Effect van AVV op werkgelegenheid blijkt beperkt, Economische
Statistische Berichten, May: 4848.
Lindeboom, M., and Theeuwes, J. J. M. (1991), Job duration in the Netherlands: The co-existence of high turnover
and permanent job attachment, Oxford Bulletin of Economics and Statistics, 53: 24364.
(1993), Search, Benets and Entitlement, Economica, 60: 32746.
Mayes, D. G., and Soteri, S. (1994), The right of dismissal and labour exibility. OSA-report W121, The Hague: SDU/OSA.
Ministry of Economic Affairs (1995), Toets op het concurrentievermogen. The Hague: SDU/Ministry of Economic Affairs.
Ministry of Social Affairs and Employment (1990), Rapportage Arbeidsmarktbeleid 1990. The Hague: SDU/Ministry of
Social Affairs and Employment.
(1996), De Nederlandse verzorgingsstaat in internationaal en economisch perspectief. The Hague: SDU/Ministry of Social
Affairs and Employment.
Mot, E., and Teulings, C. N. (1990), Minimumjeugdloon en werkgelegenheid. OSA-report W75, The Hague: SDU/OSA.
Mumford, K., and Smith, P. (1996), Labour market ows, cycles and hiring in Australia. York: University of York (mimeo).
OECD (1994), Economic Surveys 19931994: Netherlands. Paris: OECD.
(1995), Job study. Paris: OECD.
(1996), Employment Outlook 1996. Paris: OECD.
Oudshoorn, C., and Vijlbrief, J. A. (1995). Het concurrentievermogen van de Nederlandse economie, in L. van der
Geest and J. van Sinderen, (eds,), Kracht en zwakte van de Nederlandse economie. Rotterdam: ESB, 3146.
Praag, B. van, and Hop, J. P. (1995), De matigende invloed van de Nederlandse vakbeweging, Economische Statistische
Berichten, November: 9969.
Russo, G., Gorter, C., and Schettkat, R. (1997a), Searching, Ranking and Hiring, Research Memorandum 199717.
Amsterdam: Free University Amsterdam.
and Moolenaar, D. (1997b), Temporary Jobs and Temporary Workers: The Dutch Experience (Lavori
temporanei e Lavoratori temporanei: l'esperienza olandese), Lavoro e Relazioni Industriali, 2: 61102.
Salverda, W. (1999), Is there more to the Dutch Miracle than a lot of part-time jobs? Groningen, University of
Groningen (mimeo).
Schettkat, R. (ed.) (1996), The Flow Analysis of Labour Markets. London: Routledge.
Schmann, K., Rogowski, R., and Kruppe, T. (1995), Fixed-term contracts in the
210 References
European Union. Paper presented at the European Association of Labour Economists conference 1995, Lyon, France.
Sinderen, van J. (1990), Belastinghefng en economische groei. Groningen: Wolters Noordhoff.
Bergeijk, P. A. G. van, Haffner, R. C. G., and Waasdorp, P. M. (1994), De kosten van economische verstarring
op macro-niveau, Economische Statistische Berichten, March: 2749.
Stigter, A. I. (1997), Dutch experiences in reforming social insurance and implementing active labour market policy in
19871995, Research Memorandum 199722, Amsterdam: Free University.
Theeuwes, J., Kerkhofs, M., and Lindeboom, M. (1990), Transition intensities in the Dutch labour market
19801985, Applied Economics, 22: 104361.
Veen, C. van (1987), Meer markt, minder overheid; terug naar de onderneming, in A. Knoester (ed.), Lessen uit het
verleden. 125 jaar Vereniging voor de Staathuishoudkunde. Leiden/Antwerp: Stenfert Kroese.
Vijlbrief, J. A. (1992), Unemployment Insurance and the Dutch labour market, Dissertation, Amsterdam: Tinbergen
Institute/Thesis Publishers.
Visser, J. (1997), Two cheers for corporatism, one for the market. Industrial relations, unions and labour markets in
The Netherlands, British Journal of Industrial Relations.
Vries, B. de (1995), Een halve eeuw werk, werk en de werking van de arbeidsmarkt, Inaugural lecture, Erasmus
University, Rotterdam.
Wijnbergen, S. van (1996), Institutionele aspecten van de arbeidsmarktproblematiek in Nederland, OSA-report W139,
The Hague: SDU/OSA.
Wijngaert, R. van de (1994), Trade Unions and Collective Bargaining in The Netherlands, Dissertation, Amsterdam:
Tinbergen Institute/Thesis Publisher.
8 Germany: A Regulated Flexibility
Susanne Fuchs and Ronald Schettkat
Introduction
To many non-Germans, the German economy appears as an efcient, high-skilled, and high-income economy
producing top-quality products for world markets. Its welfare state and trade unionism are well developed, and income
inequalities are modest. Without too many tensions, furthermore, the West-German economy has managed to
integrate East Germany. To many Germans, however, the German economy is an outdated model of social consensus,
unable to pass necessary reforms, and is therefore stuck with persistently high unemployment, sluggish employment
growth, an underdeveloped service sector, inadequate wage dispersion, excessively powerful unions, insufcient labour
mobility, and too high taxes and social security contributions.
Certainly, the German labour market is highly regulated both legally and through collective negotiations. German
unionization is not especially high at about 30 per cent, but what really stands out is the very high degree of employer
organizationabout 90 per cent membership. Collective bargaining is guaranteed in the Grundgesetz (constitution)
but the state is not permitted to intervene, as is possible in other European countries. There are laws guaranteeing the
right to establish works councils, legal protection against dismissal, a Codetermination Act. In general, however, these
laws provide a regulatory framework for the organization of interests rather than for direct government intervention.
Regulations limit the scope for discretionary decisions by employers, but they also create opportunities. Works
councils, for example, give workers a voice and thus improve decisions (Hirschman 1970; Freeman and Medoff 1984)
and facilitate adjustments such as the reduced working hours subsidy which provides a short-term alternative to
dismissals. Most criticism focuses on legislation concerning dismissal
212 Susanne Fuchs and Ronald Schettkat
protection, co-determination, labour standards, non-wage labour costs and wage determination via collective
bargaining.
Labour Market Trends

The West-German economy is facing substantially different labour markets trends from those in East Germany.
Certainly the West has an unemployment problem, but the unemployment rate was only about 10 per cent in 1997,
compared to the East-German rate of about 18 per cent. Even so, labour force participation has declined substantially
and many workers are in public work and training programmes.
The Labour Market in West Germany
Unemployment
Unemployment in West Germany has been high since the early 1980s (Table 8.1). However, unemployment is difcult
to measure (Schettkat and Verhagen 1999) because some unemployed people may not appear in the unemployment
statistics. A corrected unemployment rate including hidden unemployment shows even higher rates and, most
importantly, the difference between the two series is much greater in the late 1980s and 1990s than in the early 1980s.
It is debatable whether or not hidden unemployment should be included in the unemployment rate, but it certainly
gives an impression of the importance of labour market measures. Table 8.1 also gives the German unemployment
rate with unemployment and the labour force measured in a similar way to those in the USA. It may come as a surprise
that, calculated in this way, the West-German unemployment rate was, until recently, only about level with, or even
below the US rate. In the mid-1990s the US rate declines whereas German unemployment rises.
The most important feature is the increasing proportion of long-term unemployment. In 1975 only 10 per cent of the
unemployed were out of work for more than one year, compared with almost one-third in 1996 (Table 8.1).
Furthermore, most of this increase represents people who had been unemployed for more than two years.
Nevertheless, there is substantial movement in the unemployment pool and the average duration fell in the early 1990s
as the West-German economy boomed following unication. Why is long-term unemployment increasing so much
and why does it remain at such high levels? Its growth seems consistent with the Eurosclerosis thesis (Giersch 1985),
with an increasing skills
Germany 213
Table 8.1. Unemployment Trends in West Germany

Years Unemployment rates in West-Germany Share of un- USA unem-
Ofcial rate Open and Using US Below 20 employed lon- ployment rate
Hidden concepts ger than 1 year
1970 0.7 NA a
0.5 NA 6.0 4.9
1975 4.7 NA 3.4 NA 9.0 8.5
1980 3.8 3.9 2.8 3.2 11.9 7.1
1981 5.5 5.5 4.0 4.9 12.8 7.6
1982 7.5 7.8 5.6 7.7 17.8 9.7
1983 9.1 9.3 6.9 9.1 23.6 9.6
1984 9.1 9.2 7.1 7.9 27.2 7.5
1985 9.3 9.7 7.2 8.1 28.9 7.2
1986 9.0 9.5 6.6 7.4 29.4 7.0
1987 8.9 9.8 6.3 6.6 30.1 6.2
1988 8.7 9.9 6.3 6.1 30.6 5.5
1989 7.9 9.2 5.7 NA 29.0 5.3
1990 7.2 8.5 5.0 5 27.2 5.6
1991 6.3 7.9 4.3 4.5 26.9 6.8
1992 6.6 8.4 4.6 5 26.2 7.5
1993 8.2 10.3 5.7 6.4 26.2 6.9
1994 9.2 11 6.5 7.3 31.2 6.1
1995 9.3 11.2 6.5 8 32.3 5.6
1996 10.1 12.3 7.2 9 32.2 5.4
1997 11 12.9 NA 9.2 35.0 4.9
a
NA = Not available.
Source: BLS 1998 for unemployment rates according to US concepts; other data from Sachverstndigenrat 1998.
mismatch and unemployment benets causing unemployment. But such a pattern may also occur in periods featuring
excess labour supply.
In Germany, unemployment is concentrated among the low-skilled. The ratio of unemployment among less-skilled
workers to that among skilled workers is 2.5 in Germany and 2.2 in the USA even though wage dispersion in Germany
is much narrower (OECD Employment Outlook 1998). This is consistent with the Nickell and Bell (1996) argument
that the exibility of wages seems not to affect relative unemployment rates. Given the very different underlying skill
distributions in the USA and Germany, it is not clear which economy actually has the wider skill-controlled wage
dispersion. The International Adult Literacy Survey suggests that Germans with relatively low educational atlaintment
score as high as average Americans in literary skills. This
Table 8.2. Unemployment Labour Participation, and Employment-Population Rates by Skill Level, Germany and the
USA
Skill level Germany USA
Unemploy- Labour force Employment Unemploy- Labour force Employment
ment rate participation population ra- ment rate participation population ra-
rate tio rate tio
I 13.1 50.3 43.7 8.0 70.0 64.4
Highschool
graduates in
USA, Real-
schule in Ger-
many
II+ 5.2 70.5 66.8 3.6 84.0 81.0
Some college
or more in
USA Appren-
ticeship or
higher educa-
tion in Ger-
many
Source: Freeman and Schettkat (2000).
Germany 215
lack of low-skilled workers might explain why, in Germany, low-wage jobs are rare. It is the difference in participation
by skill that differs, not the relative unemployment rates. This may be where high reservation wages affect the job
distribution.
Employment
Germany's employment ratio was in 1970 higher than the American, but is now 10 percentage points lower. This is a
dramatic change, mainly caused by a sharper decline of German industry and sluggish growth of services. Indeed, it
was not until the 1990s that service sector employment in the German economy equalled that in the USA in 1970
(Freeman and Schettkat 1998).
The higher proportion of industrial employment and the lower share of service employment in Germany are often
explained by a lack of specialization among German rms. It is claimed that German companies do not outsource,
which implies that the manufacturing sector includes many service activities which are outsourced in other countries.
However, Schettkat and Russo (1998) showed that German rms actually outsource more services than American
rms. This result is supported by industry-occupation matrices (Freeman and Schettkat 1998). The difference between
Germany and other countries lies in consumer services (Appelbaum and Schettkat 1996; Freeman and Schettkat 1998,
1999a). This may well be related to the wage structure, that is, the narrow German wage dispersion and high social
security contributionsissues which are discussed in The German Debate later in this chapter.
Labour Mobility
It is generally argued that employment protection leads to a reluctance to hire. The job-nding ratehirings from
unemployment divided by unemploymenthas declined. However, a declining job-nding rate does not in itself
demonstrate a malfunctioning labour market, that is, a declining hiring intensity by rms, a decreased search intensity
by the unemployed, or mismatching. The job-nding rate would also decline if unemployment were caused solely by a
negative aggregate demand shock.
The standard tool for analysing the functioning of labour markets is the Beveridge curve; the vacancy-unemployment
curve. Variations of aggregate demand are usually identied with movements along a stable Beveridge curve,98 whereas
variations in the pace of structural change, of skills mismatch, and of the functioning of the labour market are
98
The matching function, although related to the Beveridge curve, is a more general method of analysing the functioning of labour markets (Blanchard and Diamond 1989;
Schettkat 1996).
Table 8.3. Employment-Population Ratios for West-Germany and the USA, 19701996.
Year Germany United States
Total Agricul- Manufac- Services Total Agricul- Manufac- Services
ture turing ture turing
1970 67.6 5.7 32.9 29.0 61.9 2.8 20.5 38.6
1975 64.5 4.4 29.3 31.9 61.8 2.5 18.2 41.1
1980 64.9 3.4 27.9 33.7 65.9 2.3 19.3 44.2
1981 63.8 3.2 26.9 33.7 65.8 2.3 19.0 44.5
1982 62.3 3.1 25.7 33.5 64.6 2.3 17.6 44.7
1983 60.8 2.9 24.7 33.1 64.9 2.3 17.4 45.2
1984 60.6 2.8 24.4 33.3 66.9 2.2 18.2 46.4
1985 60.9 2.7 24.4 33.7 67.6 2.1 18.2 47.3
1986 61.6 2.7 24.6 34.4 68.5 2.1 18.2 48.2
1987 62.1 2.6 24.5 35.0 69.7 2.1 18.1 49.5
1988 62.4 2.4 24.4 35.6 70.8 2.0 18.3 50.5
1989 62.9 2.3 24.5 36.1 71.8 2.1 18.4 51.3
1990 63.6 2.2 24.7 36.7 72.2 2.1 18.1 52.0
1991 65.5 2.1 25.1 38.2 71.0 2.1 17.2 51.8
1992 65.9 2.0 24.7 39.1 70.9 2.0 16.7 52.2
1993 64.8 1.9 23.2 39.6 71.3 2.0 16.4 53.0
1994 64.3 1.8 22.4 40.1 72.3 2.1 16.6 53.6
1995 64.1 1.8 23.4 40.4 72.7 2.1 16.8 53.9
1996 63.1 1.7 22.5 40.5 73.1 2.1 16.7 54.3
Notes: Manufacturing comprises manufacturing, mining, and construction; Services comprises transportation, communication, public
utilities, trade, nance, public administration, private household services, and miscellaneous services.
Source: Employment data from BLS 1998; Population 1565 from OECD (1998), Economic Outlook Database, CD-Rom.
traditionally identied by shifts of the curve.99 Figure 8.1 displays the West-German Beveridge curve for the period
1960 to 1997. This suggests that there are breaks in the function in the mid-1970s and again in the early 1980s.100
Wolfgang Franz (1987, 1991; Franz and Siebeck 1991) nds a shift of the curve in 1975, and in 1988, an outward shift
and a very steep slope (Franz and Siebeck 1991: 32). A very steep Beveridge curve means that a major increase in
vacancies is needed to reduce unemployment. However, controlling for long-term unskilled unemployment, which
may be seen as endogenous, the outward shift of the Beveridge curve turns into an inward shift (estimation period
1974 to 1988).101 The mismatch
99
However, recent analyses have shown that shifts of the Beveridge curve may have many causes (Brsch-Supan 1991; Schettkat 1992). In particular, the curve shifts
outwards if mobility in the labour market increases. If an indicator for labour mobility between unemployment and employment is included in the Franz regressions, the
outward shift of 1983 becomes an inward shift (Schettkat 1995). In other words, the extended regression model suggests that the outward shift of the German Beveridge
curve was caused by higher rather than lower mobility between unemployment and employment.
100
Shifts of the curve are more debatable (Schettkat 1993).
101
Other detailed occupational analyses have found a highly exible West-German labour market.
Germany 217
Fig. 8.1. The West-German Beveridge curve
problem was thus mainly one of unskilled, long-term unemployment; as Franz and Siebeck (1991) point out, long-term
unemployment is endogenous. Furthermore, adding the 1990s to the Beveridge curve in Fig. 8.1 shows that the very
steep slope of the curve was a short-term phenomenon.
The Beveridge curve is based on the stock of vacancies and the stock of unemployed workers. Flow-duration curves
show the dynamics of job vacancies, distinguishing the ow from the duration component (Schettkat 1992). In a pure
business cycle, where the aggregate shock does not affect the skill structure in the economy, the ow of new vacancies
will increase if the economy moves out of recessionpoint A in the vacancy diagram in Fig. 8.2. Increased labour
demand will be easily satised from the large unemployment pool and the duration of vacancies will remain low. At
some point, however, hiring will become more difcult. This will occur at zero unemployment in an ideal labour
markethomogeneous labour, no frictionsbut at some positive level of unemployment in real labour markets.
Vacancy duration will increase until labour demand actually diminishes after point B. At the end of the recession, the
ow-duration curve will arrive at point A again. The illustration of a pure business cycle for unemployment resembles
that for vacancies, but a recession is characterized by a long duration of unemploymentpoint A in the
unemployment diagram of Fig. 8.2. Unemployment duration will shrink in the expansionary period and bottom out in
the boompoint
Fig. 8.2. Flow-duration curve of unemployment and vacancies in West Germany

Notes: A = recesson, B = boom.
Germany 219
B. Short vacancy duration coincides with long unemployment duration and vice versa.
The vacancy and unemployment curves can shift in opposite directions. Obviously the ow-duration curve for
unemployment has shifted outwards with every business cycle. The business cycles show the described loops, but each
business cycle involves a longer duration of unemployment. In the case of an increasing malfunctioning of the labour
market, the ow-duration curve for vacancies would be expected to respond in a similar way. It should become more
and more difcult to ll vacancies, and the vacancy curve should shift outwards. Again, the actual ow-duration curve
for vacancies shows the business cycle loops, but these occur at an increasingly shorter average duration. This hardly
supports the view that recruiting has become more difcult and that a skills mismatch has constrained growth in West
Germany.
Finally we investigate the composition of hirings. Firms can hire from different pools: the unemployed, the employed,
anda feature of particular importance in the German economyfrom their own apprentices. If the unemployed
have adequate skills, rms will prefer to hire from among the unemployed because attracting workers from other rms
usually entails a wage premium (Schettkat 1996). The rate of hiring other rms' employees declined substantially
between the 1970s and the 1980s; with overall hires declining as a result of reduced job-to-job mobility. Job mobility is,
however, endogenous to the labour market; a function of the macro-economy is high when the pool of non-employed
is low but low when the pool is large. The reason for this effect is that hiring from among the employed creates a hiring
or vacancy chain engendering an additional hiring effort by rms who lose workers, whereas hiring from among the
unemployed shortens the chain by reducing job-to-job mobility (Akerlof et al. 1988; Schettkat 1992, 1994, 1996). Low
job-to-job mobility in a situation of high unemployment is therefore a sign of a well-functioning, rather than
malfunctioning, labour market, because it arises when rms can hire from among the unemployed.
The Labour Market in East Germany After Unication

After unication in 1990, the East-German labour market virtually collapsed. However, this is only partly reected in
the unemployment rates, which were held down for example by public work measures and training programmes. The
rate of overt plus hidden unemployment outlined in Table 8.5 illustrates this point. Whereas overt unemployment was
up to almost 20 per cent in 1997, hidden unemployment added another 8 per cent. The data shows that there was a
shift from massive use of a
Table 8.4. Labour Market Mobility in West Germany

Years Job-nding rate Total Hiring rate Hires Job-to-job Mobility
from unemploy-
ment/Not in labor
force
1970 25.2 33.8 18.3 15.5
1971 20.5 30.7 18.0 12.7
1972 16.1 30.5 18.6 11.9
1973 14.7 35.7 18.6 17.1
1974 7.1 27.2 19.0 8.2
1975 4.2 25.6 21.1 4.5
1976 4.4 28.6 22.4 6.2
1977 4.6 28.7 22.2 6.5
1978 4.8 29.2 22.1 7.0
1979 5.6 29.6 22.7 6.9
1980 5.5 29.8 22.1 7.7
1981 3.9 29.0 22.2 6.8
1982 2.7 26.0 22.6 3.5
1983 2.3 26.6 23.7 2.9
1984 2.3 26.4 24.2 2.2
1985 2.3 27.1 25.1 2.0
1986 2.5 27.5 25.2 2.3
1987 2.4 27.3 23.7 3.5
1988 2.4 27.4 23.7 3.7
1989 2.7 29.7 23.7 6.0
1990 2.8 32.4 22.5 10.0
1991 3.0 30.6 20.2 10.4
Notes: Job-nding rate = unemployed who nd a job / unemployed; Hiring rate = hires / employment.
Source: Schettkat 1995, computations are based on IAB Arbeitskrftegesamtrechnung.
so-called active labour market policy to a passive labour market policy. In addition, the employment-population ratio
fell by about ve points.102
Economic unication followed the policy of giving the East-German Mark equal value to the Deutschemark (DM),
and East Germans were promised that their pensions and wages would quickly rise to West-German levels. It was
believed that the DM, together with freed markets, would raise the East-German economy to world standards without
any cost to the West-German economy. This idea informed the climate in which the early rounds of collective
bargaining took place, leading to wage agreements which promised Western wage standards within a few years, and
inicting an additional cost shock on the East-German economy. This is discussed in Collective Bargaining later in
the chapter.
102
For a comprehensive discussion of the economics of German unication see Sinn and Sinn (1994).
Germany 221
Table 8.5. East German Labour Market Trends

Years Unemployment Rates Share of Employment-population ratios
Ofcial Open + Below 20 Unem- Total Agricul- Manufac- Services
Hidden years ployed ture turing
longer
than 1 year
1991 10.3 29.5 n.a. n.a. 56.4 3.5 23.0 29.9
1992 14.8 32.8 n.a. 23.2 52.2 2.3 18.6 31.2
1993 15.8 30.5 7.1 31.0 52.3 2.0 18.3 32.1
1994 16.0 28.1 6.7 31.6 53.2 1.9 18.5 32.7
1995 14.9 26.2 7.4 28.4 53.7 1.9 18.8 33.1
1996 16.7 26.3 9.0 25.3 53.1 1.8 18.2 33.1
1997 19.5 27.1 10.9 30.1 51.4 1.8 17.4 32.2
Source: Sachverstndigenrat 1998.
The currency union was a cost shock to East-German industry. Not surprisingly, therefore, East-German industry
broke down and attempts to restructure turned into a privatization programme which was only partially successful. As
a result, unication turned out to be extremely costly, with annual transfers amounting to about 5 per cent of West
Germany's GDP (Jacobi et al. 1998).
Statutory unication meant that West-German institutions were introduced in the East. West-German unions feared
the establishment of a low-wage labour market in the East and pushed for rapid equalization of wage levels. West-
German employers had no interest in new upcoming competitors, however, East-German labour market conditions
created pressures on institutions like collective bargaining, leading to the introduction of so-called opening clauses in
collective agreements. A decade later, it is still the case that only a small proportion of East-German employers are
organized; extremely high unemployment has created pressure for wage exibility, which has also spilled over into
West Germany.
The German Debate

Extreme positions on deregulation are rarely voiced in the German debate. The mainstream tends to support the
modication of existing legal and institutional features. The same is true of actual policy which has been generally
pragmatic; for example, dismissal protection has been deregulated by a law permitting xed-term contracts. Although
actual deregulation policies have been pragmatic, some more extreme positions can be found in academic debates and
in politics where, for
example, the Kronberger Kreis, a group of neo-liberal economists, has pushed for free markets.
A recent example of the controversy is the demand by the red-green government for lower interest rates to stimulate
economic growth and employment and the reply of the European Central Bank (ECB) to the effect that labour market
reform is needed to improve employment. Labour market regulations and welfare state provisions are claimed to
distort the functioning of labour markets, which have become sclerotic (Giersch 1985). According to Siebert,103
. . .Most importantly, a whole set of measures raised the reservation wage: the duration of unemployment benets
was partly increased; it was made easier to obtain such benets; the conditions under which the unemployed were
expected to accept jobs were interpreted more generously; governmental schemes for the unemployed were
extended; the relative distance between the lowest wage in the labour market and non-working income in welfare
programs became more narrow; and the minimum wage, which is applied in some countries, was raised. (Siebert
1997)
The OECD seems to follow the same line of argument: the generosity of unemployment payments and related
welfare benets increases the natural rate of unemployment and reduces the speed of readjustment (Martin 1996).
Consequently, the OECD suggests that Germany should lower replacement rates, cut the duration of benets, and
strengthen work availability and eligibility conditions in order to encourage low-paid employment (OECD 1997).
Several reports have been written on the potential for deregulation; the most prominent being the Deregulierungsbericht
prepared by a government-appointed committee of economists (Deregulierungskommission 1991; Donges 1992) and the
report of the Monopolkommission (1994). Many economists who argue in favour of substantial deregulation use the
perfect labour market as their point of reference.104 This approach was pushed in the 1980s by the neo-liberal
Kronberger Kreis. It was also advocated by the Monopolkommission, whose chairman at that time, Carl-Christian von
Weiszcker, was also a member of the Kronberger Kreis. The Monopolkommission was supposed to report on trends
endangering free competition. It interpreted its task as including labour market coalitions, and classied collective
bargaining as a bilateral monopoly where the unionthe monopolist on the supply sidehad three possible ways of
setting wages.105 If the union set wages at the
103
Siebert's paper is written with reference to Europe but he draws many of his arguments from the German context.
104
For a more elaborate discussion of this issue, see Buttler et al. 1996.
105
See also the recommended measures of the Zunkunftskommission of Saxony and Bavaria which include a signicant reduction in wages and social security benets; the
encouragement of part-time work and entrepreneurial activity in the area of low-paid personal services, such as house-cleaning; a basic low income model for pensioners;
and the incentives for individuals to accumulate savings in order to replace social security. The report aims at a complete revision of social attitudes, partly abolishing the
institutional and mental framework of post-war Germany, including corporatist patterns of negotiation, the social security system, legal protection of labour interests, and,
nally, deeply rooted intuitions concerning social justice.
Germany 223
equilibrium level, then collective bargaining was unnecessary because this would also, according to the
Monopolkommission,106 be the market result. If the union set wages below the equilibrium level, rms would pay
higher wages leading to wage drift. In these two cases, the union would do no harm but would be unnecessary.
However, the Monopolkommission argued that union ofcials needed to create a premium for their members and that
wages, for this reason, were set above the equilibrium level, reducing demand and creating unemployment. The
Monopolkommission provided no empirical evidence on actual union wage-setting.
Collective bargaining is important, but the view of Germany as a country with exclusively centralized wage
negotiations is overstated. Although the metal workers' union (IGM) may be identied as a wage-leader there is
substantial variation in wage levels and wage increases across industries.
Wage dispersion is much narrower in Germany than, for example, in the USA. In the USA the top decile earns 4.1
times as much as the lowest, whereas in Germany it is only 2.3 times as much (OECD, Employment Outlook 1998).
But the skill distribution is also much wider in the USA than in Germany and it is unclear how the two wage
distributions compare if measured in efciency units. Detailed comparative analysis shows that the difference in wage
dispersion between the USA and Germany is less between industries but rather within each industry (Freeman and
Schettkat 1998).
Wage compression is claimed to cause unemployment because low-skilled labour in the industrialized competes with
the South (Wood 1997; Freeman 1997). Technological change is another arguement along this line. Excessive wages
for unskilled labour in Germany cause substitution of labour by capital and insufcient wage dispersion leads to
substitution of low-skilled by skilled labour. Although there is no legal minimum wage in Germany, the social security
system denes a reservation wage below which nobody is willing to work (Sinn 1998; Scharpf 1993; McKinsey Global
Institute 1997). Rather than guaranteeing a high standard of living, high minimum wages actually keep low-skilled
workers out of work. (McKinsey Global Institute 1997). In this way, the
106
A recent study shows that the wage residual not explained by human capital variables tends to be higher in decentralized wage bargaining systems than in co-ordinated
bargaining systems (Hartog and Teulings 1996).
social security system drives away low-skilled jobs, while public-sector jobs are not being created (Zukunftskommision
der Friedrich-Ebert Stiftung 1998). This reasoning has produced suggestions that wage subsidies should be introduced
for low-skilled, low-paid labour, either directly or through a negative income tax (Scharpf 1993).
However, recent analysis shows that the reservation wage dened by the social security system is not as high as is
commonly believedabout 30 per cent of the mean wage (Freeman and Schettkat 1998). It may therefore be that
labour supply constraints are over-emphasized in the German discussion. Nevertheless, non-wage labour costs and
taxes may create product demand constraints (Freeman and Schettkat 1999a). A study by Krueger and Pischke (1998)
rejects the hypothesis that wage structure and employment growth are correlated. They nd no evidence of any
signicant difference in growth patterns between low-wage or high-wage economies. For similar results with a different
data set, see Freeman and Schettkat (1998).
A major issue in the wage debate is the degree of wage differentiation permitted in collective bargaining. Many
proposals have been madeand actually implementedto permit rm-specic adjustments to the Flchentarifvertrag or
collective agreement for an industry-region combination. More exibility, it is argued, would allow the survival of
rms whose existence may be threatened by standard wages. Some employers' representatives take an extreme
position and argue for maximum exibility at the rm level. Consequently, they even claim that the right to strike has
no place in the modern world. However, there are also employers who fear that rm-level wage negotiations will bring
conict into the workplace and shift competition from product markets to factor markets. In fact, of course, rm-level
wage negotiations also require the right to strike at the level of the individual company. Certainly, if negotiations take
place at this level, works councils are more likely to agree to wage restraint, or even to wage reductions, when their
company is in nancial difculties. On the other hand, many works councils are likely also to use their power to push
up wages whenever a particular rm is doing well. Probably the most important consequence of rm-level
negotiations is that conict will no longer be delegated to employers' associations and unions but will be conned
within the company, possibly creating severe friction both between employers and employees and between different
groups of workers (Franz 1992).
Meanwhile, there has been wide-ranging discussion on the future of the Flchentarifvertrag. The key argument is that
wages should adjust to rm-specic conditions because wages dened in a Flchentarifvertrag cannot t rm-specic
requirements and may therefore eliminate jobs in less productive rms. These management claims are often supported
Germany 225
by their works councils, and wages below the standards dened in collective agreements are often agreed upon
unofcially, especially in East Germany. Meanwhile, various collective agreements allow for substandard wages if rms
are under economic pressure and cannot survive otherwise (Bispinck 1997). In addition, sub-standard wages can be
paid to new hirees in the chemical industry.
Two developments seem to have undermined the strength of collective agreements. The rst is the call for a reduction
of the standard working week to 35 hours. A variety of working-time regimes have been established since 1984 when
this was phased in (Bispinck 1997: 553). The second is connected to the imposition of the West-German industrial
relations framework on the East-German economy. In East Germany, collective agreements were concluded on the
assumption that the government was committed to the one-to-one exchange rate policy. In 1991 it was agreed for the
metal industry that wages in the East would be increased in four stages to reach the West-German level in 1994. This
agreement was mainly negotiated by representatives from the West because the employers were poorly organised in the
East and the old union was no longer trusted. However, the nal increase to 100 per cent wage parity was postponed
in 1993 following a conict between the metal union (IGM) and the metal employers' association in East Germany.
The employers' association unilaterally cancelled the wage agreement and IGM called a strike in East Germany.
The result, however, was not only that wage increases were postponed, but that so-called ffnungsklauseln (opening
clauses) were introduced into collective agreements.107 These permit wage reductions in cases where companies nd
themselves in severe economic difculties. Since February 1997, opening clauses have also been used in the West-
German metal industry.
Another major deregulation issue is dismissal protection, where some revisions of the law have actually been
implementeddiscussed later in this chapter. The relaxation of dismissal protection under the Kohl government was
prompted by the argument that dismissal protection and severance payments increase labour costs and thus induce
rms to hire less labour in upswings, because employers foresee difculties in shedding labour in a downturn
(Schellhaass 1989). However in that case, relaxing dismissal protection would only affect the variation in employment,
not the average level. It has also been argued that dismissal protection raises labour costs, and therefore reduces the
level of employment. But, of course, this argument only holds for a given labour
107
They are opening clauses in the sense that they open up branch-level collective agreements to company-specic arrangements concerning wages. To save face, the unions
prefer the term Hrteklauseln, which implies a highly selective and conditional use of these clauses, which were implemented in 1993 in the face of resistance by the IGM.
demand function if dismissal costs are additional to other labour costs. This type of argument is often applied to all
non-wage labour costs, since unions understand employers' contributions to the social security of their employees as a
masked wage.
The report of the Deregulierungskommission proposes modifying the legal framework of collective agreements by, for
example, permitting sub-tariff wages for the long-term unemployed, reducing the inclusiveness of collective
agreements, or substantially reducing dismissal protection. Some of these measures were adopted by the Kohl
government despite opposition from unions and social-democratic politicians. The confederation of employers'
associations, which represents domestic trade interests, feared substantial wage increases and both the unions and
employers' confederation expected ruinous competition. However, the report aimsdespite the severity of some of its
proposalsto modify the legal and institutional framework of industrial relations, rather than abolish it (Keller and
Seifert 1997).
In 1998 the Schrder government restored the earlier levels of dismissal protection, sickness benets, and pensions.
This echoes its position that macro-economic conditions are the main reason for unemployment. This seems to be
correct so far as the business cycle in the late 1990s is concerned. However, unemployment has risen in Germany over
a 25 year period and employment has declined at a time when it has increased elsewhere.
The Dynamics of Rigidity and Flexibility: Specic Regulations and

Labour Market Performance
Collective Bargaining
The Industrial Relations Framework

The government cannot directly inuence wage bargaining, only the framework within which bargaining takes place
(Collective Agreement Act, TVG 1949). This is an important difference between Germany and other European
countries. Industrial relations in Germany are unusual in other ways too. Firstly, there is an elaborate legal basis for
labour relations both in the workplaceworks councils as dened in the Works Constitution Act, BetrVG. 1952,
revised in 1972and at company board levelCodetermination Act, MitbG. 1976. Workers therefore have many
means of being obstructive and severely damaging the economy if they wish to do so. But this has not happened and
actual industrial relations have continued very much in the spirit of the legal framework,
Germany 227
which is co-operative rather than confrontational. Members of the supervisory board are expected to act in the
interests of the specic company. Likewise, works councils are expected to co-operate with managementas they
often do, even against their unions' advice. Strike activity is infrequent in Germany, although a few conicts have been
resolved only after substantial action, such as the dispute over shorter working hours in the 1980s.108 However, as long
as the labour market parties are still engaged in negotiations, no strike action is allowed, and any decision on strike
action must be reached via specic decision procedures.
Employers' organisations are divided into three types: Employers' Associations (Confederation of German Employers'
Associations (BDA)), Chambers of Industry and Commerce (German Association of Chambers of Industry and
Commerce (DIHT)) and Business Associations (Federation of German Industries (BDI)). Only the subsections of the
BDAlike Gesamtmetal, the association of employers in the metal industriesare involved in collective bargaining at
the regional level. Both Chambers of Industry, Commerce and Business Associations are concerned with the regulation
of apprenticeships, licensing and the regulation of trade practices, and public relations. The DIHT and the BDI have
proposed signicant labour market deregulation, for example in the sphere of collective bargaining, suggesting single-
employer negotiations at company level and enabling works councils to negotiate collective agreements. The BDA,
however, tends to adopt the approach of improving regulations rather than seeking to introduce radical change. This
may be related to its direct responsibility for wage negotiations.
Unions are organised by industry.109 The German Trade Union Confederation (DGB) encompasses eleven industry
unions. Following unication, the DGB temporarily doubled its membership, but this subsequently declined as
employment fell in both East and West. Like the BDA, the DGB is not directly involved in collective bargaining. Both
confederations act as platforms for the formulation of common policies, but wage goals are formulated and negotiated
at the branch level. Some companies have not joined the employers' association and negotiate their own agreements;
VW, for example. All in all, there are currently several thousand collective agreements in Germany (Bispinck 1997),
108
One of the longest strikes in West-German history occurred in 1956 and 1957 in Schleswig-Holstein's metal industry and concerned paid sick leave. Other severe labour
conicts were the 1978 strike on the introduction of new technology, employment protection and training, the 1984 strike over shorter working hours in the metal and
printing industries, and the 1992 strike in banking and public services.
109
The only major exception is the white collar union (DAG) which was in some opposition to the DGB but is now negotiating a merger with the DGB.
although most of them are just regional variations on general agreements.

In 1995, the Kohl government tried to re-establish concerted action (the so-called Kanzlerrunde or chancelor's
roundtable) but did not succeed because the unions felt they were being ignored in the formulation of public policy,
and because employment trends were not encouraging. Nevertheless, tripartite consultations inspired by the Dutch
model once again play a major role in the Social-Democratic/Green government's policies to boost employment, now
known as the Bndnis fr Arbeit or alliance for jobs. One of the proposals in the alliance for jobs is a new early
retirement programme developed by the Minister for Employment, Walter Riesterthe former vice-president of the
IGMwhich will enable workers to take early retirement without actuarial adjustments at the age of 60. Because the
government promised to lower social security contributions, Minister Riester proposed tariff-funds, to be nanced
out of negotiated wage increases. This measure may be seen as an additional pension insurance contribution under a
different label. The alliance for jobs came under pressure because the unions wanted to exclude wage issues from the
discussions, whereas the employers insisted that the inclusion of wages was essential. In addition, the unions asked for
substantial wage rises in the spring of 1999 and the resulting wage conict hardly provided an encouraging background
to the new concerted action initiative.
The unication of Germany led to the introduction of West-German industrial relations in East Germany, although
there was no appropriate institutional structure for it in the new Lnder. For example, there were hardly any private
companies, but only Kombinatehuge, vertically integrated industrial organizations which needed to be restructured.
Therefore, the degree of organization of employers in the new Lnder was, and still is, weak. Negotiated wages and
low productivity force many of them to pay salaries lower than the tariff wage in the new Lnder. Their works councils
agree to these arrangements, so the aftermath of unication has established a trend towards decentralization in wage
bargaining which is spilling over into West Germany. Ironically the introduction of shorter working hours in the 1980s
also shifted some aspects of bargaining to the level of the individual company.
Changes in Collective Agreements (Flchentarifvertrag)

About 84 per cent of all German workers are covered by collective bargaining arrangements (Bispinck 1997).
Formally, these involve bargaining between a particular industry union and an employers' association for a specic
region, such as the metal industry in the northern parts of Baden-Wrtemberg. The resultant collective agreement is
binding on
Germany 229
all employers who are members of the employers' association.110 Collective bargaining has come under pressure for
several reasons. Despite an obvious problem of low productivity in the East-German economy, high wages were
negotiated for East-German employees after unication; small companies complained about the dominance of large
companies in the employers' associations, and bargaining at rm and establishment-level became increasingly
important.
The acceptance of opening clauses in the East was a signal for employers to abandon collective bargaining or to ignore
agreementseven though the unions were prepared to accept a moderate, exible policy on wage levels and working
hours. This trend rst emerged in the new Lnder, with employers in West Germany following suit (Bispinck 1997:
553f.). The company and industry-specic variations on agreements include for example: exible working hours,
limited contracts, suspension of negotiated wage increases, lowering of wages, lower wage levels for specic groups of
employees such as apprentices, and the suspension or reduction of special payments.
Effects of Wage Compression

It is sometimes argued that, because collective bargaining sets minimum wage standards, it cuts off part of the labour
demand function, especially at the bottom of the wage scale. The result is excessive wages for less-skilled workers, and
this causes the substitution of less-skilled by skilled labour. Indeed, the wage differential between less-skilled and skilled
labour has narrowed in Germany. This may explain why unemployment is concentrated among the less-skilled.
The elasticity of substitution is a measure indicating the amount of labour substitution related to relative wages. An
elasticity of substitution of 1.0 would indicate that a 10 per cent increase in the relative wage of less-skilled workers will
reduce relative employment in this group by 10 per cent. Estimated values for this elasticity in Germany vary
substantially: from 0.3 to 3.0 (Fitzenberger and Franz 1997; Steiner 1997; Beissinger and Mller 1998). In a
comparative analysis of the USA and Germany, Freeman and Schettkat (1999a) found a low elasticity of substitution
and conclude that changes in relative wages can explain only a small part of the US-German difference in employment-
population ratios for the less-skilled. Thus, a higher degree of wage exibility may slow substitution processes but will
hardly reverse them.
Some collective bargaining contracts permit sub-tariff wages for new hirees. Keller and Seifert (1997) report that 14
per cent of all new hirees
110
Employers who are not members of the employers' association will have their own individual contract with the union or may be free to negotiate with their employees if the
Ministry of Economics has not extended the collective agreement to the whole industry.
are on sub-tariff wages, that is, pay equivalent to 9095 per cent of the tariff wage in the chemical industries. This
measure is designed to compensate for the productivity differential, especially among the long-term unemployed.
Does Unemployment Insurance Create Unemployment?

Protection against income loss during unemployment is based on a two-tier system. Everybody with a work history
meeting certain minimum criteria qualies either for unemployment benetcurrently 65 per cent of the previous net
income or 60 per cent for singlesor for means-tested unemployment assistance, which provides roughly 50 per cent
of previous net income. The latter is a safety net provision for those whose period of eligibility for unemployment
benet has expiredusually after one yearor who are not eligible for insurance benets. The scheme is administered
by the German Labour Agency, Bundesanstalt fr Arbeit (BA), and nanced through contributions from employers and
employees on an equal basis.111
One frequently-cited explanation of high German unemployment rates is the assumed relationship between the
generosity of unemployment benets (Siebert 1997) and workers' willingness to risk unemployment. Generous
unemployment benets may raise the NAIRU. It is true that unemployment benets are much more generous in
Germany than in the USA, but they have certainly not become more generous since the 1970s. In the 1980s and 1990s
replacement rates declined and eligibility requirements were tightened. This should have lowered rather than raised the
German NAIRU. The composition of unemployment benets has changed in favour of unemployment assistance. In
the early 1970s, fewer than 10 per cent of the unemployed received unemployment assistance; in 1995 it stood at
around 25 per cent. At the same time, the proportion of the unemployed receiving the higher unemployment insurance
benets decreased from 65 to less than 40 per cent in the mid-1980s (Cramer 1986; Ermann 1988), but rose again to
50 per cent in 1995.
Detailed econometric analysis (Hunt 1995; Hujer and Schneider 1995; Steiner 1997; Zimmermann 1993) reveals no
signicant impact of unemployment benets on the duration of unemployment. The duration of benet eligibility
seems, however, to have an effect. Among those who argue that benets are too generous, Giersch et al. (1992: 204)
argue that unemployed workers have a disincentive to accept jobs which pay less
111
Part of the cost of service provision is met through general taxation rather than insurance contributions (Schmid et al. 1992).
Germany 231
than 50 per cent of their previous wage because the social security system provides an alternative income at around
that level.
Dismissal Protection and Working Time
The Changing Legal Situation

Dismissal protection is one of the prime factors held responsible for the alleged distortion of the labour market. Once
a regular contract of employment has been concluded, it is assumed to be for an indenite period,112 except in the case
of apprenticeship agreements, which are generally time-limited. The Dismissal Protection Act of 1969 states that any
dismissal must be justied either by economic reasons, such as demand conditions for the specic rm, or by the
personal performance of the employee concerned. Unjustied dismissals are illegal. As in other areas, the law denes
minimum standards and the collective bargaining parties are free to conclude more generous agreements.
Since 1993, workers to be dismissed must be given notice of at least four weeks but collective bargaining may lead to
longer periods, usually increasing with seniority.113 Employers are obliged to investigate alternatives to dismissal, such
as a different job in the same rm. The works council must be consulted in advance, and is involved in negotiating
severance payments, which are usually highequivalent to about ve to ten months' wages (Hase et al. 1992;
Sengenberger 1987; Hemmer 1988).
Several steps have been taken to deregulate dismissal protection. The major step was the introduction in 1985 of the
Employment Promotion Act Beschftigungsfrderungsgesetz, which relaxed the conditions concerning xed-term
employment contracts. Fixed-term contracts had been possible before, but the employer had to show good reason
for them. The new law provided unconditional freedom for the conclusion of xed-term contracts up to eighteen
months in duration. Highly controversial when rst introduced, the law was initially limited to a period of four and a
half years (Bchtemann and Hland 1989). Since January 1996, when the law was once again extended for a limited
period, this time to 31 December 2000, employers have been allowed to conclude xed-term contracts for periods of
up to two years. Contracts
112
Bielenski (1997) argues that, under the German Civil Code (art. 620 BGB), contracts of employment were originally assumed to be time-limited, but that court decisions
created the exception to the rule (unlimited contracts).
113
Before 1993, the period of notice for blue collar workers was 2 weeks, and for white collar workers, 6 weeks. The distinction between blue and white collar workers has
become blurred over time and is nowadays mainly related to membership of one or other compulsory pension insurance scheme.
may be extended up to three times within the two-year period.114 The hope was that relaxing the constraints on xed-
term contracts would reduce dismissal costs and thus increase employment.
Small rms with a maximum of ve employees were exempted from dismissal protection laws and, in October 1996,
the limited dismissal protection was extended to include companies with fewer than ten employeesroughly 80 per
cent of all companies. Again, the hope was that this would enable small companies, especially those in the crafts and
service sectors, to respond more exibly to demand uctuations. However, this legislation was reversed by the Social-
Democratic/ Green government in 1999.
The Impact of Dismissal Protection on Employment

The controversial Employment Promotion Act was accompanied by research to evaluate its effects. Bchtemann and
Hland's study took place from May 1985 to April 1987, a period in which the economy was recovering from
recession, and therefore exactly the kind of period in which such a law would be expected to have the strongest impact.
They found that xed-term contracts were used mainly in small and medium-sized companies, which typically have a
large share of low-skilled labour. This shows that the relevance of short-term contracts is limited to that segment of the
labour market in which neither workers nor employers invest much in the relationship. 40 per cent of rms adopting
xed-term contracts use them as an extended probationary period.115
The Bchtemann-Hland analysis concluded that the job-creation effect of the Employment Protection Act was very
modestabout 25,000 persons, or roughly 0.1 per cent of German employment at that time. Furthermore, it found
that about two-thirds of employers refrained from using xed-term contracts because they were interested in long-
term rather than xed-term employment relationships. Using different data, Abraham and Houseman (1993) and
Kraft (1994) found no evidence that the Employment Promotion Act had changed the speed of employment
adjustments.
A second evaluation (Bielenski, et al. 1994) comes to conclusions similar to those of Bchtemann and Hland.
Between 1985 and 1994 the share of xed-term contracts remained at 56 per cent of all new contracts and, most
surprisingly, it hardly varied over the business cycle (Bielenski 1997; Kraft 1994). The empirical evidence gives
deregulators no reason to be euphoric, but neither does it support those critics who
114
For employees aged over 60 there is no limitation on xed-term contracts.
115
The evidence presented is based on Bchtemann and Hland (1989).
Germany 233
predicted the collapse of industrial relations in Germany if employers were allowed to conclude xed-term contracts.
Employment Stability But High Working Volume Flexibility

If the institutional framework affects company behaviour, one would expect to see German rms behaving very
differently from US rms when faced with demand shocks. Abraham and Houseman (1993, 1995) and Kraft (1994)
investigated whether employment protection in Germany actually reduced adjustments in the volume of work. Given
the differences between German and US dismissal protection laws, one would expect US companies to adjust their
work force more rapidly than German rms in response to demand shocks. Starting from the observation that overall
employment measured in hours, or working volume, can be inuenced either by variations in numbers of persons
employed or by changes in average hours worked, Abraham and Houseman conducted an in-depth comparison of
adjustment processes in German and US rms in response to a negative demand shock (for a similar study, see Kraft
1994). Their results are summarized in Fig. 8.3.
The Abraham-Houseman analysis shows that rms in both economies reduce volume in a very similar way and at a
roughly similar speed (top panel of Fig. 8.3). However, the means whereby this is achieved differ substantially (lower
panels of Fig. 8.3). In the USA, rms respond in the short run mainly by dismissing workers and to a lesser extent by
reducing average hours worked. In Germany, by contrast, rms reduce average working hours and are less inclined to
cut the number of staff. In the long run, however, rms in both economies reduce their work force. The divergent
short-run behaviour of US and German rms may well be explained by their different cost structures. In the USA,
overtime premiums are high, at 50 per cent, dismissals are easy and there is no subsidy for short-time working. To
German rms, on the other hand, overtime premiums are low at about 20 per cent, dismissals are more difcult, and
short-time work is subsidized. Given that rm-specic human capital is at risk in a lay-off system, Abraham and
Houseman conclude that it may be more efcient to vary working hours than reduce employees. The Abraham-
Houseman analysis also shows that German rms are well able to reduce their work force. Kraft (1994) conrms the
Abraham-Houseman analysis.
An example of the German strategy is the short-time working agreement at Volkswagen (VW) in 1993. High severance
payments were one reason why the VW management proposed shorter working hours instead of mass redundancies in
1993 (Meinhardt et al. 1993). The VW employees agreed to a working time reduction of 20 per cent and a pay
Fig. 8.3. Employment elasticity in German and US rms

Notes: Reaction to a Permanent Negative Demand Shock, Manufacturing.
Source: data based on Abraham/Houseman (1993).
cut of between 12 and 15 per cent in exchange for a guarantee of employment. However, VW workers were at the
top end of the German wage scale and the workers' acceptance of pay cuts must be viewed in the light of the
uncertainty of the situation for individual workers, since management did not specify in advance who would be
affected by dismissals. With an envisaged 20 per cent reduction of the workforce, it was certain that the impact would
not be limited to specic groups. The VW
Germany 235
agreement is seen as a pioneering advance in working-time policies. It encouraged the metal union's agreement to
heavy pay cuts in exchange for shorter working hours and employment protection.
Until the Working Time Act, which came into effect in 1994, working hours in Germany were regulated by a law
dating from 1938. The Act enables employers to react more exibly and allows for an extension of the standard 40-
hour week to 60 hours per week for a period of up to six months or even longer if the employees concerned agree.
Regarding work on Sundays and public holidays, the new law creates some exibility through generous exceptions. In
addition, night and shift work is heavily deregulated although the proportion of night and shift work has remained
stable.
Skill Flexibility: Dual Education in Interaction With Retirement
Dual Education and Youth Unemployment

Germany's dual education system is admired around the world because it is believed to provide highly skilled and
highly exible workers and also low youth unemployment. In 1990 the unemployment rate for the under-20s was only
5 per cent, substantially below the European average.
The International Adult Literacy Survey (OECD 1997) clearly shows that the strength of the German education
system lies at the lower end of the range of formal educational attainment: the average score of Germans without
secondary education qualications was as high as the mean American score in the Survey.
It is well known that a substantial part of labour mobility is caused by younger workers searching for the right job. This
mobility component is smaller in Germany because apprenticeships in the dual education system have a selection
function as well as an educational function (Franz and Soskice 1995). Firms training young workers for about three
years have a chance to evaluate them as employees and this reduces the information problem inherent in the hiring
process; rms tend to retain only the most satisfactory.116 This is another element which has reduced labour market
mobility in Germany.
Although apprentices have up-to-date skills, stagnation of employment and strong employment protection based on
seniority rules mean that rms encounter difculties in hiring them. For young workers to enter the system, older ones
must exit at the same rate. The apprenticeship system turns young workers into insiders and it is commonly
116
The apprenticeship agreement is xed for the apprenticeship period. Prolongation of the relationship requires a new contract of employment.
accepted in Germany that widespread premature retirement is preferable to massive youth unemployment.
Insider-Outsider Relations and Early Retirement

Early retirement including through disability pensions, often preceded by a period of unemployment, has helped not
only to modernize the skill structure of individual companies but also to reduce the overall labour force (Schusser
1987). German pension laws guarantee entitlement to a pension from the age of 60 following a year's unemployment.
Workers can leave the rm at the age of 59, often receiving additional payments from the former employer.117
Early retirement is an effective means of reducing and/or modifying the workforce, but it is expensive. Inefciencies
arise because rms and workers externalize at least some of their costs to the pension system and require an increasing
share of GDP to be spent on pensions. Consequently, efforts have been made to burden employers and early retirees
with at least a proportion of the costs. In 1984 it was decided that, under certain conditions, employers had to cover
unemployment benets for up to one year if they dismissed older workersoften the rst step towards early
retirement. After complex and difcult negotiations between the government and the then opposition during the
1980s, a radical reform of the pension system was approved in 1989. This included new forms of nancing, new
contributions, and a change in the retirement age.
Later, in 1990, a heavy burden was placed on the pension system by the inclusion of East-German pensioners in the
West-German pension system. This raised pension contributions by almost 3 percentage points and provoked a new
reform bill. Under the new law, early retirement pensions are calculated with actuarial reductions. For example, a
worker retiring in March 1997 at the age of 60, ve years before the standard retirement age, now has to accept a 0.3
per cent reduction in his/her pension. The reduction level will gradually increase to 18 per cent for unemployed people
taking early retirement at the age of 60 in or after January 2007.118
Since August 1996, part-time work for elderly workers has, to some extent, taken the place of early retirement.
Employees have the option of working part-time from the age of 55, on 70 per cent of their
117
Distinguishing the unemployed by one-year age groups shows that numbers leaving unemployment for non-participation peak at the age of 60 and account for more than
93\
118
For women, reductions for those retiring at the age of 60 start in March 2000 (0.3\
Germany 237
full-time wage. The part-time arrangement must be agreed with the employer. Provided the employer meets the 20 per
cent difference between nominal and net part-time income, pays pension contributions corresponding to 90 per cent
of the employee's full-time income, and hires a former apprentice or an unemployed person to replace the hours lost,
the German labour agency will reimburse the costs.
In autumn 1997, a pioneering collective agreement was reached on the basis of this new law following mediation in the
metal industry. Workers in the metal industry may now shift to part-time work from the age of 55, provided there is a
company-level agreement. Income compensation for older part-time workers has been raised to 82 per cent and
workers are in any case entitled to work on a part-time basis from the age of 61.119
The German Social-Democratic/Green government of the late 1990s has also initiated consultations and negotiations
between unions, employers' associations and government (the so-called alliance for employment). These will include a
debate about the possibility of re-establishing a right to early retirement at the age of 60 without any reduction in
pension. According to the Minister for Employment this should be nanced through a new fund (Tariffond), to be
established jointly by employers and employees. He suggested that 1 per cent of the increase in wages should be used
to nance this fund. Reactions to this idea have been mixed. Employers' organisations state that the costs of this
measure will explode in the near future, despite the presumption that early retirement will be used to decrease the size
of companies rather than to create new jobs. The IGM welcomed the idea in the rst instance, although a leading
member of the union's executive board opposed the fund as being too expensive for employees and a devaluation of
the existing pension system. It is hard to avoid the suspicion that this fund is a way of achieving a hidden increase in
non-wage labour costs.
Other Deregulation
Statutory Sickness Benets

In 1995 the Bundestag approved a new law on sickness benets. This reduced benets from 100 to 80 per cent of the
net income during the rst six weeks of sickness.120 This law was viewed as a major change which
119
The IGM suggests that employees should work full-time between the ages of 55 and 60 on 82\
120
Another change was also made, imposing on new employees a four-week waiting period to become eligible to receive statutory sickness benets.
would force unions to accept labour standards and to reduce non-wage labour costs. In 1996, again after a strike, the
IGM negotiated a collective agreement on the restoration of sickness benets to 100 per cent. In practice, therefore,
the new regulation applies only to roughly 20 per cent of employees. Shortly after the elections, the Schrder
government re-introduced more generous statutory sickness benets. The law now once again guarantees 100 per cent
benets for all employees, although it retains the four-week waiting period for new employees.
Private Placement Services

Until 1994, public employment agencies had a monopoly of placement services.121 The new law passed in that year
allowed private agencies to offer their services. However, they had to accept stringent limitations on fees, which can
only be charged to employers (Walwei 1998). The number of licenses increased steadily to 4,500 in mid-1997, after
which there was a decline (Walwei 1998: 15). A 1996 evaluation showed that the new regulation had little effect.
Between August 1994 and December 1996, roughly 100,000 placements were made, about 75 per cent of which
resulted in jobs lasting only seven days or less. The market share of private placement services accounts for only 2 per
cent of all placements.
Conclusions
The German labour market suffers from persistently high and predominantly long-term unemployment, its
employment-population ratio is low and the service sector is underdeveloped. Is this the outcome of an overly
regulated labour market? In this chapter we have summarized the main lines of the argument and reviewed the
empirical evidence for the over-regulation case. As previous studies have done for other countriesfor example,
those in this book and by Blank and Freeman (1994) and Gregg and Manning (1997) we must conclude that the
empirical evidence in support of this explanation of Germany's unemployment is extremely weak.122
Our review of the evidence does not suggest that labour laws have substantially reduced the exibility of the West-
German labour market. Indicators often claimed to provide clear evidence of the malfunctioning labour market, such
as the rise in long-term unemployment and
121
Except in the case of the placement of top managers.
122
From an evaluation of individual incentives to participate in the labour market Klaus Zimmerman (1993: 240) concludes similarly: . . .the disincentive effects of benets
and taxation are not strong enough to justify large revisions of the welfare state.
References 239
reduced job-to-job mobility prove on closer examination to be less than convincing.

The reason why most analyses have failed to detect negative effects of regulation may be that they are usually based on
aggregated data, which may not reveal the distorting effects of regulation. However, even detailed analysis of the
impact of specic regulation measures, such as dismissal protection, has failed to identify the exibility-disturbing. This
may be because these measures are not isolated but embedded in a multitude of other measures, which may partly
affect the effects of deregulation. For example, xed-term contracts failed in practice to produce the hoped-for
employment effects. This conclusion means that the performance of the actual labour market is not well understood
and that there are offsetting mechanisms in operation through other formal or informal institutional arrangements,
which are difcult to conceptualize in theoretical models.
Another possible reason why it has proved impossible to nd the promised effects of deregulation measures is that
actual deregulation has not gone far enough; the measures implemented may have been half-hearted compromises. It
is true that changes made in the German labour laws have fallen far short of radical, free-market deregulation
proposals. However, even in the UK, where more radical institutional changes have been implemented, the impact of
these measures on employment seems to be weak (Gregg and Manning 1997).
Even if the bad news is that deregulation measures have not achieved positive employment effects, the good news is
that they have not had the destructive effects claimed by their opponents. The relaxation of the requirements for xed-
term contracts has not boosted employment, but neither has it destroyed social cohesion. In a world of imperfect
markets, long-term, stable employment relationships are in the interests of both sides.
References
Abraham, K., and Houseman, S. (1993), Job Security in America: Lessons from Germany. Washington: Brookings Institution.
(1995), Labor Adjustment Under Different Institutional Structures, in Buttler, F. et al. (eds.), 285315.
Akerlof, G. A., Rose, A., and Yellen, J. L. (1988), Job Switching and Job Satisfaction in the U.S. Labor Market,
Brookings Papers on Economic Activity, 2: 495582.
and Hessenius, H. (1991), East Germany in from the Cold: The Economic Aftermath of Currency
Union, Brookings Papers on Economic Activity 1.
240 References
Appelbaum, E., and Schettkat, R. (1996), The Increasing Importance of Institutions for Employment Performance,
in G. Schmid et al., International Handbook of Labor Market Policy and Policy Evaluation. London: Edward Elgar,
791810.
Beissinger, T., and Mller, J. (1998), Wage Flexibility and Employment Performance: A Microdata Analysis of Different Age-
Education Groups in German Industries, Regensburg: University of Regensburg (mimeo).
Bell, L. A. (1986), Wage Rigidity in West Germany: A Comparison with the U.S. Experience, Federal Reserve Bank N.Y.
Quarterly Review, Autumn, 1121.
Bellmann, L., and Mller, J. (1996), Institutional Inuences on Interindustry Wage Differentials, in F. Buttler, et al.,
Institutional Frameworks and Labor Market Performance. London: Routledge, 13267.
Bentolila, S., and Bertola, G. (1990), Firing Costs and Labour Demand: How Bad is Eurosclerosis? Review of Economic
Studies, 57: 381402.
Bertola, G. (1992), Labor Turnover Costs and Average Labor Demand, Journal of Labor Economics, 4.
Bielenski, H. (1997), Deregulierung des Rechts befristeter Arbeitsvertrge. Enttuschte Hoffnungen, unbegrndete
Befrchtungen, WSI-Mitteilungen, 50/8: 5327.
Kohler, B., and Schreiber-Kittl, M. (1994), Befristete Beschftigung und Arbeitsmarkt. Empirische Untersuchung ber
befristete Arbeitsvertrge nach dem BeschFG. Forschungsbericht 242, Bonn: Bundesministeriums fr Arbeit und
Sozialordnung.
Bispinck, R. (1997), Deregulierung, Differenzierung und Dezentralisierung des Flchentarifvertrags. Eine
Bestandsaufnahme neuer Entwicklungstendenzen der Tarifpolitik, WSI Mitteilungen. 50/8: 55161.
Blanchard, O. J., and Diamond, P. (1989), The Beveridge Curve, Brookings Papers on Economic Activity, 1: 160.
(1994), Ranking, Unemployment Duration, and Wages, Review of Economic Studies, 61: 41734.
Blank, R., and Freeman, R. B. (1994), Evaluating the Connection between Social Protection and Economic Flexibility,
in R. Blank (ed.) Social Protection versus Economic Flexibility. Chicago: University of Chicago Press, 2142.
BLS (Bureau of Labor Statistics) (1997), Comparative Civilian Labor Force Statistics, Ten Countries. Washington: BLS.
Brsch-Supan, A. (1991), Panel Data Analysis of the Beveridge Curve: Is There a Macroeconomic Relation between
the Rate of Unemployment and the Vacancy Rate?, Economica, 58: 27997.
Bruno, M., and Sachs, J. D. (1985), Economics of Worldwide Stagation. Cambridge, Mass: Harvard University Press.
Bchtemann, C. F., and Hland, A. (1989), Befristete Arbeitsvertrge nach dem Beschftigungsfrderungsgesetz 1985,
Forschungsbericht 183, Bonn: Bundesminister fr Arbeit und Sozialordnung.
Burtless, G. (1983), Why is Insured Unemployment so low?, Brookings Papers on Economic Activity, 1: 22549.
Buttler, F., Brandes, W., Dorndorf, E., Gaum, W., and Walwei, U. (1992), Flexibility and Job Security in the Federal
Republic of Germany, SAMF Arbeitspapier, 1.
References 241
Franz, W., Schettkat, R., and Soskice, D. (1996), Institutional Frameworks and Labor Market Performance. London:
Routledge.
Calmfors, L., and Drill, J. (1988), Bargaining structure, corporatism and macroeconomic performance, Economic
Policy, 3: 1361.
Cramer, U. (1986), Wieviel Arbeitslose erhalten noch Lohnersatzleistungen?, Mitteilungen aus der Arbeitsmarkt und
Berufsforschung, 2: 2038.
Deregulierungskommission (1991), Marktffnung und Wettbewerb. Deregulierung als Programm? Stuttgart: Pschel.
Donges, J. B. (1992), Deregulierung am Arbeitsmarkt und Beschftigung. Tbingen: Mohr.
Engels, W., Gutowski, A., and Hamm, W. (1986), Mehr Markt im Arbeitsrecht. Bad Homburg: Frankfurter Institut fr
wirt.-pol. Forschung.
Ermann, K. (1988), Arbeitsmarktstatischtische Zahlen in Zeitreihenform, Betrge zur Arbeitsmarkt- und Berufsforschung,
31.
Fitzenberger, B., and Franz, W. (1997), Flexibilitaet der qualikatorischen Lohnstruktur und Lastenverteilung der Arbeitslosigkeit:
Eine konometrische Analyse fuer Westdeutschland, ZEW Discussion Paper No. 97-32, Mannheim.
Franz, W. (1987), Strukturelle und Friktionelle Arbeitslosigkeit in der Bundesrepublik Deutschland, in G. Bombach,
B. Gahlen and A. E. Ott (eds.), Arbeitsmrkte und Beschftigung, Fakten, Analysen, Perspektiven. Tbingen: Mohr,
30127.
(1991), Match and Mismatch in the German Labor Market, in F. Padoa-Schioppa (ed.), Mismatch and Labor
Mobility. Cambridge: Cambridge University Press, 10535.
(ed.) (1992), Structural Unemployment. Heidelberg: Physica.
and Siebeck, K. (1991), A Theoretical and Empirical Analysis of Structural Unemployment in Germany:
Reections on the Beveridge Curve, in W. Franz (ed.), Structural Unemployment. Heidelberg: Physica, 158.
and Soskice, D. (1999), The German Appienticeship System, in Buttler, F. et al. (eds.) 20834.
Freeman, R. B. (1988), Evaluating the European View that the United States Has No Unemployment Problem,
American Economic Review, 78: 2949.
(1995), Are Your Wages set in Beijing?, Journal of Economic Perspectives, 9/3: 1532.
and Medoff, J. M. (1984), What do Unions do? New York: Basic Books.
and Schettkat, R. (1998), From McDonald's to McKinsey: Comparing German and US Employment and Wage
Structures, Paper presented at the Leverhulme II conference, Labour Market: Stocks and Flows, Institute of
Economics and Statistics Oxford NBER-Paper.
(1999a), Zwischen Fastfood und Excellence; Die Beschftigungslcke in Deutschland im Vergleich zu den USA,
Hamburger Jahrbuch fr Wirtschafts- und Gesellschaftspolitik. Hamburg: Weltwirtschaftsarchiv.
(1999b), The Role of Wage and Skill Differences in US-German Employment Differences, Jahrbcher fr
Nationalkonomie und Statistik: 4966.
(2000), Low Wage Services: Interpreting the US-German difference, in M. Gregory (ed.), Labour market
Inequalities. Oxford: Oxford University Press.
Geppert, M., and Kachel, P. (1995), Die Treuhandanstalt am Ende. Historischer Abri und kritische Beurteilung aus
volkswirtschaftlicher und organisationstheo
242 References
retischer Perspektive, in R. Schmidt and B. Lutz (1995), Chancen und Risikeb der industriellen Restrukturierung in
Ostdeutschland. Berlin: Akademie Verlag.
Giersch, H. (1985), Eurosclerosis, Kiel Discussion Paper No. 112. Kiel: Institut fr Weltwirtschaft.
Paqu, K. H., and Schmieding, H. (1992), The fading miracle. Four decades of market economy in Germany. Cambridge:
Cambridge University Press.
Gregg, P., and Manning, A. (1997), Labour market regulation and unemployment, in D. J. Snower and G. De la
Dehesa (eds.), Unemployment Policy. Government Options for the Labour Market. Cambridge: Cambridge University Press.
Hartog, J., and Teulings, C. (1996), Dutch Income Inequality in International Perspective, Paper presented at the
Conference on Globalization and the New Inequality, Utrecht.
Hase, D., Neumann-Cosel, R., and Rupp, R. (1992), Handbuch Interessenausgleich und Sozialplan. Kln: Bund Verlag.
Hemmer, E. (1988), Sozialpraxis in der Bundesrepublik: eine empirische Untersuchung. Kln: Deutscher Institutsverlag.
Hirschman, A. O. (1970), Exit, Voice, and Loyalty. Responses to Decline in Firms, Organizations, and States. Cambridge Mass:
Harvard University Press.
Hujer, R., and Schneider, H. (1995), Institutionelle und strukturelle Determinanten der Arbeitslosigkeit in
Westdeutschland: Eine mikrokonomische Analyse mit Paneldaten, in B. Gahlen, H. Hesse and H-J. Ramser
(eds.), Arbeitslosigkeit und Mglichkeiten ihrer berwindung, Wirtschafswissenschaftliches Seminar Ottobeuren, Band
25, J.C.B. Mohr: Tbingen.
Hunt, J. (1995), The effect of unemployment compensation on unemployment duration in Germany, Journal of Labor
Economics, 13: 88120.
Jacobi, O., Kller, B., and Mller-Jentsch, W. (1998), Germany: Facing New Challenges, in A. Ferner and R. Hyman
(eds.) (1998), Changing Industrial Relations in Europe. Oxford: Blackwell, 190238.
Keller, B. (1991), Arbeitspolitik. Mnchen und Wien: Oldenbourg.
and Seifert, H. (1997), Eine Zwischenbilanz der Deregulierung, WSIMitteilungen. 50/8: 52231.
Kommission fr Zukunftsfragen der Freistaaten Bayern und Sachsen Erwerbsttigkeit und Arbeitslosigkeit in
Deutschland, Entwicklung, Ursachen und Massnahmen, (1997) Teil II: Ursachen steigender Arbeitslosigkeit in
Deutschland und anderen frhindustrialisierten Bonn: Lndern.
Kraft, K. (1994), A comparison of employment adjustment in France, Germany, Great Britain and Italy,
Wissenschaftszentrum Berlin, FS I 92207.
Krueger, A., and Pischke, S. (1999) Observations and Conjectures on the U.S. Employment Miracle, Paper presented
at the AEA meetings 1999, New York.
Layard, R., Nickell, S., and Jackman, R. (1991), Unemployment. Oxford: Oxford University Press.
Lehmbruch, G. (1984), Concertation and the Structure of Corporatist Networks, in J. H. Goldthorpe (ed.) Order and
Conict in Contemporary Capitalism. Oxford: Oxford University Press.
McKinsey Global Institute (1997), Removing Barriers to Growth and Employment in France and Germany. Washington:
McKinsey \&Company.
Martin, J. P. (1996), Measures of replacement rates for the purpose of international comparisons: a note, OECD
Economic Studies, 1/26: 10015.
References 243
Meinhardt, V., Stille, F., and Zwiener, R. (1993), Weitere Arbeitszeitverkrzungen erforderlich Zum Stellenwert des
VW-Modells, Wirtschaftsdienst, January.
Mller, J., (1998), Die Entwicklung der Lohnungleichheit in Deutschland, in Statistisches Bundesamt (ed.), Einkommen
und Vermgen in DeutschlandMessung und Analyse. Stuttgart: Metzel-Pschel, 16993.
Monopolkommision (1994), Mehr Wettbewerb auf allen Mrkten. Baden-Baden: Nomos Verlagsgesellschaft.
Nickell, S. (1998), Unemployment: Questions and some answers, The Economic Journal, 108: 80216.
and Bell, B. (1996), Changes in the Distribution of Wages and Unemployment in the OECD countries, American
Economic Review, 86/5: Papers and Proceedings, 3028.
Nullmeier, F., and Rueb, F. W. (1993), Die Transformation der Sozialpolitik: Vom Sozialstaat zum Sicherungsstaat. Frankfurt,
Main, and New York: Campus.
OECD (1997), International Adult Literacy Survey. Paris: OECD.
(1997), Implementing the OECD Jobs Strategy: Lessons from Member Countries' Experience, Council at
Ministerial Level, 2627 May, Paris.
(1997), Literacy Skills for the Knowledge Society. Paris: OECD.
(1998), Historical Statistics, Diskettes. Paris: OECD.
Rowthorn, R. E. (1992), Centralisation, employment and wage dispersion, The Economic Journal, 102.
Rudolph, H. (1984), Die Entwicklung der Vermittlung in Arbeit, Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, 2:
16882.
Sachverstndigenrat (1998), Jahresgutachten 1998. Deutscher Bundestag, Drucksache 14/73.
Scharpf, F. W. (1993), Von der Finanzierung der Arbeitslosigkeit zur Subventionierung niedriger Erwerbseinkommen,
Gewerkschaftliche Monatshefte, 44: 43343.
Schellhaass, H. (1989), Sozialplne aus konomischer Sicht, Zeitschrift fr Arbeitsrecht, 167207.
(1990), Das Arbeitsrecht als Beschftigungshemmis?, in C. F. Bchtemann and H. Neumann (eds.), Mehr Recht
durch weniger Arbeit?. Berlin: Edition Sigma, 87104.
Schettkat, R. (1992), The Labor Market Dynamics of Economic Restructuring. New York: Praeger.
(1992a), The Labor Market Dynamics of Economic Restructuring; The United States and Germany in Transition. New York:
Praeger.
(1993), On Vacancies in Germany, Jahrbcher fr Nationalkonomie und Statistik, 213/2: 20922.
Schettkat, R. (1994), Flexibility through Mobility: A Function of the Macroeconomy, Structural Change and Economic
Dynamics, December.
(1995), The Macroperformance of the German Labor Market, in Buttler, F. et al. (eds.), 31642.
(1996), Labor Market Flows over the Business Cycle: An Asymmetric Hiring Cost Explanation, Journal of
Theoretical and Institutional Economics, 64153.
(1997), Employment Protection and Labor Mobility in Europe: An Empirical Analysis using the UE's Labor
Force Survey, International Review of Applied Economics, 11/1: 10518.
244 References
and Russo, G. (1998), Are Structural Economic Dynamics a Myth? Changing Industrial Structure in the Final
Product Concept, Working Papers Onderzoekschool Arbeid. Welzijn. Sociaal Economisch Bestuur, 4.
and Verhagen M. (1999), International Unemployment Trends: Measurement and Stylized Facts, in Wagner, H.
(ed.) Globalization and Unemployment, Berlin, Heidelberg 1999: Springer.
Schmid, G., Reissert, B., and Bruche, G. (1992), Unemployment Insurance and Active Labor Market Policy. Detroit: Wayne
State University Press.
Schusser, W. H. (1987), Stufenweiser bergang in den Ruhestand aus der Sicht der betrieblichen praxis, in Schmhl,
W. (ed.), Verkuerzung oder Verlaengerung der Erwerbsphase, Tbingen.
(1988), Stufenweiser bergang in den Ruhestand aus der Sicht der betrieblichen Praxis, in W. Schmhl (ed.)
Verkrzung oder Verlngerung der Erwerbsphase, Verkuerzung oder Verlaengerung der Erwerbsphase. Mohr: Tbingen,
21324.
Sengenberger, W. (1987), Struktur und Funktionsweise von Arbeitsmrkten. New York: Campus.
Siebert, H. (1997), Labor Market Rigidities: At the Root of Unemployment in Europe, Journal of Economic Perspectives,
11/3: 3754.
Sinn, G., and Sinn, H-W. (1994), Jumpstart: The Economic Unication of Germany. Cambridge, Mass.: MIT Press (reprint
edition).
Sinn, H-W. (1998), Der Sozialstaat in der Zwickmhle, Handelsblatt, November 24: 53.
Soskice, D. (1990), Wage Determination: the Changing Role of Institutions in Advanced Industrialized Countries,
Oxford Review of Economic Policy, 6/4.
Steiner, V. (1997), Extended Benet-Entitlement Periods and the Duration of Unemployment in West German,
Discussion paper, Centre for European Economic Research, 14.
and Wagner, K. (1997), Relative Earnings and the Demand for Unskilled Labor in West German
Manufacturing, ZEW Discussion Paper No. 97-17.
Teulings, C., and Hartog, J. (1998), Corporatism of competition? Cambridge: Cambridge University Press.
Tobin, J. (1972), Ination and Unemployment American Economic Review, 62.
Wagner, J. (1985), Mangelnde Faktormobilitt eine Ursache fr Arbeitslosigkeit?, Wirtschaftsdienst, 65/6: 297303.
Walwei, U. (1998), Job Placement in Germany: Developments Before After Deregulation, IABTopics, Nuremberg.
Wood, A. (1995), How Trade Hurt Unskilled Workers, Journal of Economic Perspectives, 9/3: 5780.
Zimmermann, K. F. (1993), Labour responses to taxes and benets in Germany in A. B. Atkinson and G. V.
Mogensen (eds.), Welfare and Work Incentives. A North European Perspective. Oxford: Clarendon Press.
Zukunftskommission der Friedrich-Ebert-Stiftung (1998) Wirtschaftliche Leistungsfhigkeit, sozialer Zusammenhalt,kologische
Nachhaltigkeit, Drei Zieleein Weg. Bonn: Dietz.
9 France: The Deregulation That Never Existed
Miguel A. Malo, Luis Toharia, and Jerme Gauti
Introduction
France is not generally considered to be a case where labour market exibilities or rigidities have played a substantial
role in the debate on unemployment. Although it is generally regarded as one of the countries which follow a
Mediterranean pattern (Siebert 1997), rigidity has not been a central focus of public attention, and it can be argued
that the labour market policies adopted by French governments have not attempted to introduce signicant changes in
terms of exibility, unlike some of the other European countries covered by this study. This is not to imply that
unemployment is not a major problem in France. Indeed, it was one of the key issues in the recent Parliamentary
elections, and France is one of the countries most in favour of including unemployment and social problems as
fundamental elements in the reform of the Maastricht Treaty.
A Review of the National Debate and Literature

Figure 9.1 plots the evolution of the labour force and employment, as well as the unemployment rate, in France over
the past twenty years.123 Starting from an unemployment rate below 5 per cent in 1977, employment stagnation until
1983 and the minor fall in 198385, together with continued growth of the labour force throughout the period, pushed
the
123
The data in the gure are drawn from various OECD publications. The most recent gures are from Enqute Emploi, the yearly labour force survey. Since there are minor
differences among the various sources, we have linked the series ourselves, both here and throughout the chapter.
246 M. Malo, L. Toharia, and J. Gauti
Fig. 9.1. Labour force, employment, and unemployment in France

Source: OECD, Economic Outlook, June 1997; updated by authors on the basis of INSEE Premire, n530, June 1997.
France 247
rate to double digit levels by 1985. This was the period in which the Socialist Party formed the government (1981) and
launched a series of measuresamong them the so-called Lois Auroux, more specically geared towards the labour
market, though by no means the most signicant ones among the measures adoptedto tackle France's economic
problems. The solutions were in keeping with the tradition of the left and the labour movement, although they were
conventionally Keynesian in their economic principles: increasing the minimum wages and social welfare benets,
creation of jobs in the public sector, and nationalisation of the largest banks and rms. However, this single-country
experiment failed, and the French government revised its policies, adopting a more markedly European strategy. As in
many other European countries, the second half of the 1980s was a period of economic expansion. Employment
increases eventually outweighed the growth of the labour force and the unemployment rate declined, although by 1990,
when the recovery was at its peak, it remained just below 9 per cent. However, the employment gains were relatively
modest, averaging 0.8 per cent per year during the 198590 period. The crisis of the early 1990s fed unemployment
and the rate exceeded 12 per cent in 1994, with more than 3 million people out of work for the rst time in recorded
history. The brief recovery of 199596 improved matters somewhat, but unemployment still remains at the same
record levels reached in the mid-1990s.
As already mentioned, labour market rigidities, in their various forms, seem to have played a relatively minor role in
this evolution. It is interesting to inspect the main elements of current French labour market debate as summarized by
a recent study, commissioned by the European Commission (DG-V) from one of the leading French consultancy
rms (Bernard Brunhes Consultants):
1. Working time management and reduction: France is one of the leading European countries in terms of the
debate on economy-wide, statutory, working time reductions, or work sharing, as a means to boost
employment. This issue is to some extent related to labour market exibilityinasmuch as it concerns the
ability to cope with the vagaries of the business cycle through working time adjustmentsand, interestingly
enough, it has recently tended to gain importance in the French debates over exibility. Since the Socialists
regained government in 1997, working time reduction is increasingly being considered an important element in
aiming at substituting internal exibilityincreased hirings and reduction in precarious forms of
employment, in exchange for higher working time resilience and even reductionfor external, less desirable,
exibilitydismissals and redundancies, and the use of atypical contracts.
2. The inability of the French economy to create employment in the service sector. Compared with other
European countries, France's service sector is underdeveloped. Cultural and social factors are cited as the
causes of this specically French feature.
3. Reducing the costs of unskilled labour. Although this appears to be generally accepted in France as one
possible remedy for unemployment, there is no general agreement on how it can be achievedthrough wage
reduction or, and this is more widely accepted, through a reduction of payroll tax and the general tax wedge.
There is also debate on the size of the cuts required, on their expected effects, or the problems which this
measure might generate in terms of social security nancing. In this case, as in many other European countries,
minimum wages and especially tax policy are seen as inhibiting unskilled labour employment. However, the
minimum wagethe salaire minimum interprofessionnel de croissance or SMICappears to be a sacred,
unchallenged institution. Since 1980, it covers 910 per cent of the working population, which gives a measure
of its signicance as a wage oor the unskilled.
4. The effectiveness of labour market policies, which are hindered by the dispersion of the institutions managing
the various measures. Their effectiveness is debated in terms of the need to reform the state's current role in
the labour market, but fewer questrions are raised as to the need for such government intervention.
The study makes no mention of labour market exibilities or rigidities as key variables in explaining unemployment, or
as issues in the debate on its possible solution. Labour market exibility, therefore, does not appear to be one of the
central issues in current debate on unemployment in France. This is not to say, however, that it has never been a
signicant factor. One of the main policies adopted by the right-wing government in 1986 was the elimination of prior
authorization for dismissals and the introduction of new grounds for the use of atypical contracts. However, it is also
true that many of these measures seeking to enhance labour market exibility were later off-set or reversed. Moreover,
measures designed to allow the more exible deployment of human resources, and most notably the better
management of working hours, have always been a signicant component of labour market policies. The French
experience may be understood as one of ambivalence between efforts to meet employers' demands for higher
exibility and the continuous introduction of successive regulations and legislative measures, the efcacy of which tend
to remain unassessed (Goetschy 1998).
Although some labour exibility measures have been adopted, as we
France 249
shall see, in France the main labour market institutions have remained largely unchanged over the 1980s and 1990s.
Indeed, this lack of signicant institutional change is sometimes blamed for the upsurge in unemployment. It is clear
from Fig. 9.1 that the French economy has been unable to create enough jobseven including those vacated by
retireesto meet the continuously rising ow of entrants; a ow which has shown no signicant response to the
evolution of employment opportunities.
The Dynamics of Rigidity and Flexibility

Analysed in this section are the institutional changes introduced over the 1980s and 1990s. The analysis will be
combined with discussion of the effects of these changes on labour market performance. Table 9A in the Appendix
(adapted from Bertola and Ichino 1995) provides a useful starting point. It presents the various institutional changes
introduced, as well as a subjective evaluation of their impact on labour market exibility. It will be seen that the
government's political stance appears to have been a key factor in such changes. Thus, after the Socialist government
had enacted the Lois Auroux in 198283a move intended to strengthen workers' rights; increasing rigidity from the
employers' point of viewmore exibilizing measures were adopted by the right-wing government in the second half
of the 1980s. In any case, it should be pointed out that, according to one leading expert in labour law (Lyon-Caen
1993), the French way towards exibility has not followed the path of deregulation, but rather that of more regulation.
Regulatory Changes in Firings and Hirings

We focus rst on the changes considered to be the main exibility-favouring measures introduced during the 1980s:
the removal of mandatory prior authorization for collective dismissals on economic grounds; and the easier use of
temporaryxed-term and otherwiseand part-time contracts.
The employers' organizations had long pressed for the removal of prior administrative authorization for dismissals.
The impact of this measure is difcult to assess, however, because the administrative statistics on collective dismissals
disappeared jointly with the requirement for authorization. The data on entry into the ANPE (the French Public
Employment Service) show a substantial increase in these sorts of dismissals: by 17 per cent between the end of 1985
and the end of 1986.
In addition, the Enqute Emploi supports the suggestion that dismissed permanent workers were replaced by temporary
ones (Maurau 1993). The removal of mandatory prior authorization also strengthened the role of work councils, and
gave rise to a rapid judicial procedure regulating the causes and the form of economic dismissals. In 1989, a new law
extended the role of the public administration in collective dismissals by introducing a new requirement, the social
plan.
In 1993, this requirement was reinforced by a further law which established that the social plan must contain detailed
provisions, as opposed to merely general declarations, concerning the workers to be dismissed; otherwise the collective
dismissal would be considered void. Another law passed in 1992, following the EC Directive of 24 June 1992,
increased the procedures and rights of workers facing economic dismissal; the law considering economic dismissals to
include every sort of job loss or separation stemming from a rm's economic problems, although these separations
might be quits, for instance. Finally, the ve-year law passed in 1993 sought to promote employment stability by
allowing greater exibility in working hours. This is discussed in Working Time below.
Some sort of incomplete legal round trip thus appears to have characterized the evolution of the French regulation of
economic dismissals. The initial impetus towards exibility provided by the removal of prior authorization was partly
off-set by the more stringent constraints subsequently imposed on rms ring workers, larger rms in particular,
whose obligations became more binding. In addition, incentives were introduced to encourage the use of adjustment
mechanisms as an alternative to dismissals. It can be argued that the initial changes were an effort to facilitate
workforce adjustment. However, when it seemed that these easier adjustments were giving rise to the excessive
substitution of xed-term workers for permanent ones, palliative measures were introduced in order both to protect
dismissed workers and to provide rms with alternative adjustment mechanisms. It is interesting to note in this
respect, that although the 1993 law only affects in practice 15 per cent of dismissals, some voices in the employers'
organization are expressing a desire to return to prior authorization, considered less burdensome and probably less
uncertain than the current web of legal and union controls.
As regards atypical work, the main changes introduced were the new regulations enacted by the Chirac government in
1986, which opened the door to the widespread use of new xed-term, temporary or agency work, and part-time
contracts. Although these forms of contract had existed before, the new measures, especially the ordinance of August
1986, enabled rms to hire xed-term or temporary workers for their
France 251
normal, non-temporary, activities, while the maximum period for such arrangements was extended to two years.
Although later reforms tended to restrict the use of these atypical jobs, it can be safely argued that the 1986
regulations marked a signicant break in the way that French rms could resort to temporary arrangements to solve
their labour adjustment needs. These so-called aided jobs, involving the various training contracts as well as the jobs
of community service in the non-market sector, however entail worse working conditions, among them wages in
practice smaller than the SMIC. They primarily affect, and increasingly so, youngsters under 25: in 1992, 20 per cent of
them held one of these aided jobs; by 1996, the percentage had grown to 35 per cent.
What was the labour market effect of these changes to the ring and hiring regulations? Although it is tempting to
relate them to the employment increases of the second half of the 1980s, many other factors were probably involved in
this growth, which was experienced throughout Europe. At any rate, although Maurau (1993) is not particularly in
favour of exibility, it does recognize that the changes regarding hirings and rings exerted a positive, albeit minor,
inuence on employment. Employment, however, did not increase to any marked extent, compared with other
European countries, since it averaged 0.8 per cent over the growth period of 198590; which implies that the changes
probably added little to the average employment stock on a longer-term basis (recall Figure 9.1).
On the other hand, it is clear is that these changes did increase the proportion of workers in atypical forms of
employment. For example, according to Eurostat (OECD 1996: Fig. 1.1), on the basis of the European Labour Force
Survey, the proportion of wage and salary workers with temporary statuswhich covers training, xed-term and
agency work contractswas below 4 per cent in 1984. The proportion consistently increased throughout the rest of
the decade to reach around 1011 per cent by 1990, and remained more or less stable thereafter. If this increase is
related to the 198586 regulationsand there appear to be sufcient grounds to argue as muchone may conclude
that the exibility at the margin, inasmuch as the basic rights of workers hired on permanent contracts were left
unchanged, introduced in France in 1985 enlarged the segment of the economy in which job insecurity prevails. This
dual labour market interpretation is supported by recent analyses of the French labour market (most notably, Saint-
Paul 1996).
Nevertheless, these various changes have not altered the core of French labour law (Lyon-Caen 1993)Bertola and
Ichino (1995) also argue that these institutional changes were minor and relatively inconsequentialbecause the
regulatory changes have never pursued global
change, if this is dened as change in the employment security of all workers. Saint-Paul (1996) attempts to explain this
sort of reform from a political economy perspective as a realistic way to increase labour market exibility in the long
term. The cornerstone of his argument is that goverments will only implement labour market reforms if these do not
damage the majority of people. Hence xed-term contracts are a compromise which enables the government to
introduce some exibility into the labour market without endangering the employment security of most permanent
workers. Overall, the results of the changes can be summarized as a relatively minor and short-lived increase in
employment together with an expansion of the secondary, unstable segment of the labour market.
The Wage Determination System

It is widely known that France has a system of wage determination which differs considerably from those of the other
European countries. The traditional interest of trade unions in political, as opposed to industrial, goals, together with
the government's incisive role in enforcing incomes policies, has given rise to a relatively weak system of collective
bargaining. The main changes to have occurred since 1980 are only indirectly linked with legislative or institutional
innovations. The most signicant of these changes have been those introduced in the early 1980s by the Lois Auroux,
the aim of which was to increase the unions' presence at the plant level, and for which purpose a more signicant role
was given to plant-level collective agreements. Probably the most important single measure has been the introduction
of obligatory rm-level bargaining. In addition, the sanctioning of plant-level agreements which may conict with the
sectoral agreementeven though they are limited to working time arrangements, the so-called accords drogatoireshas
been a further step towards decentralization. It should be pointed out, however, that these changes have not been
made in order to increase the responsiveness of wages to the economic situation of rms. Rather, this appears to have
been a side-effect of the changes.
As a consequence, the number of rm-level agreements has signicantly increased since the early 1980s, while the
number of sectoral agreements has decreased. However, the percentage of employees covered by sectoral agreements
has grown, and national sectoral bargaining has increased at the expense of regional agreements. This feature should be
linked with the tendency for sectoral agreements to act less as safety nets in wage determination and occupational
classication, and more as general frameworks regulating numerous aspects
France 253
of labour relations and providing a clearer structure for rm-level bargaining between employers and workers
representatives. On the whole, therefore, sectoral agreements have diminished in importance as constraints on wage
determination, since their formal application in many rms, especially small and medium-sized enterprises (SMEs), is
hindered by the trade unions' lack of inuence in small rms, which is one of the unregulated areas pointed out by
Segrestin (1990). The only exception has been the campaign conducted by the trade unions to strengthen sectoral
agreements, as part of their wider strategy to increase lower wages.
Overall, as Brunhes Consultants (1997) stress, the main trend in wage determination has been the increasing
individualization of wage formation, as measured by the increasing proportion of employees receiving payments
linked either to their personal characteristics or to their performance. This appears to have been the principal device
used by French rms to increase wage exibility. These extra payments affect 80 per cent of wage- and salary-earners
in the private sector (Colin 1995), and they amount to 18 per cent of the total wage bill for these workers. Such extra
payments are related to rm size, and in 1994 they were less frequent in rms with fewer than 200 employees (DARES
1995). Moreover, they are cyclical: according to DARES (1994), the share of rms paying extra payments closely
follows the business cycle, increasing in booms and diminishing in recessions.
In terms of wage growth, French wages have been much more restrained since 1978 and especially since 1983,
compared with the previous period: the real-wage productivity ratio followed an increasing trend over the 1960s and
early 1970s, boosted in 197576, slightly decreasing until 1982 and then plummeting until the late 1980s, reaching its
lowest level in 30 years (INSEE 1998). It can thus be argued that the economic and collective bargaining changes of
the early 1980s, although not initiating the period of wage restraint, were instrumental in the consolidation and
exacerbation of the declining trend of wages vis--vis productivity.
As for wage dispersion, despite the trade unions' endeavour to increase the wages of the worst paid, it appears that this
variableas measured by the ratio of the ninth to the rst decile of earnings (OECD 1996)has not varied
signicantly, certainly to a lesser extent than in other Western countries. Despite these relatively minor overall
movements in wage distribution, however, the statutory minimum wage has increased at a faster rate than average
hourly wages1.8 as opposed to 1.0 in the post-78 periodso that the ratio of minimum wage to average wage is the
second highest in Europe. In the highest, The Netherlands, it should be remembered that the minimum wage is
calculated as a
function of age, which is not the case in France. Consequently, the minimum wage may have some effect on the
employment of low-skilled workers. Indeed, a study by Sneessens and Shadman-Mehta (1995) shows that the relative
wage of less-skilled workerswho are presumably those at the bottom of the wage distribution scalehas
substantially increased in France since the 1970s. In close relation to this, Sneessens and Shadman-Mehta nd
empirical evidence that non-neutral technological progress favours skilled labour. They estimate the trend rate of
substitution of skilled for unskilled workers at around 4 to 5 per cent per year before 1974 and at around 2 to 3 per
cent thereafter.
In this regard, there seems to be a general agreement in the French debate on labour market reforms that the labour
taxes on low-skilled labour should be reduced. The problem is choosing the right instrument and solving the problem
thus created of nancing social security. Some studies (reviewed in OECD 1995) emphasise the effects of decreasing
the minimum wage on the employment of specic groups, such as young people and low-skilled workers. This does
not appear to be a good political strategy, however, because in France the minimum wage is considered to be a
cornerstone of social peace. Beginning in 1993, mainly in the ve-year law, and reinforced in 1995 and 1996, various
measures designed to reduce the labour costs of the unskilled have been implemented by the progressive introduction
of exemptions from employer social contributions for such workers. Thus, the ratio of the labour cost at wages equal
to the SMIC over the labour cost at wages equal to median wages, which had signicantly increased over the 1970s and
through the early 1980s, had returned by 1996 to its 1970 level (CSERC 1996). These measures are expected to
alleviate the problem of unemployment among unskilled workers, but only in the medium term.
On the whole, the French system of wage determination has followed a long-term trend towards more decentralized
collective bargaining and less government intervention. However, no dramatic changes have taken place in the
structure of the system. Concern about the possible negative effects of a too compressed wage structure in terms of
high unemployment differentials against the least skilled, has led to the adoption of measures aimed at reducing the
labour costs of this category of worker. Their effects are not expected to be apparent until some time has elapsed.
Working Time
As already mentioned, working time is probably the area in which French regulations have made the most signicant
effort to provide
France 255
exible forms of operation for employers. The most important change in terms of labour market exibility has been
the ve-year law of 1993.124
In France, working time is determined by employers, who set a so-called collective schedule which must abide by
certain limits imposed by the law as well as, in some cases, industry-wide agreements. The general conditions are as
follows:
1. The maximum working week is 39 hours; any work above that level is considered overtime, whereas work
below it may be considered, for full-time workers, to be partial unemployment, which carries unemployment
benets.
2. In some instances, this 39-hour limit may be exceeded provided that certain other limits are respected: 48
hours in the same week, 46 hours per week in any 12-week period, 10 hours per day, with other limits set for
special categories of workers.
3. Overtime may be resorted to within certain limits: in addition to the global limit established in the preceding
point, a total of 130 overtime hours per year may be worked, with an administrative authorizationissued by
the labour inspector, as the representative of the Ministry of Labourbeing required to exceed that ceiling.
The rst 8 hours of overtime, computed on a weekly basis, are paid at a premium of 25 per cent, and any
additional hours at a 50 per cent premium. Employers can avoid these extra payments by compensating
overtime with free time. According to a study by Cette and Taddei (1994), ofcial overtime work in 1991 was
equivalent to 680,000 full-time jobs, and amounted to around 3 per cent of total employment in that year and
to 30 per cent of unemployment.
4. Short-time working, or partial unemployment as it is known in France, is the reverse of overtime, and is used
mostly by rms trying to avoid lay-offs in times of underuse of capacity. Short-time working may result from a
total close-down of plants or from a reduction in the number of hours worked per day or week. Its use
obviously varies with the business cycle: in 1993, a very bad year in terms of business conditions, the number
of short-time working days allowed by the Ministry of Labour was twice that of the previous year, reaching 24
million (Braud et al. 1994), the equivalent of 140,000 jobs. This provides a comparison with overtime which
suggests that shorttime working is much less used. One of the reasons for its relatively infrequent use is the
way in which the system is nanceddirectly
124
Leaving aside the recent debate on reducing the working week to 35 hours. However potentially important this measure may be, the debate over its implementation and its
employment effects has not been framed in exibility/rigidity terms, even though, in the late 1990s the French government has argued for working time reduction as a way
to increase internal exibility.
through a at-rate subsidy paid by the state which employers must completeas well as the relatively low benet paid
to workers, which is on average 50 per cent of their wage, with a minimum benet.125 There is a yearly maximum of
700 hours per worker.
5. Part-time work is dened as amounting to at least one-fth less than the legal working schedule of 32 hours per
week or less, and it may be computed on a weekly, monthly or yearly basis. There are no restrictions on its use,
and its costs are proportional to the time worked. Since 1992, various incentives have been introduced to
encourage the use of part-time work, for instance a reduction in the social security contributions paid by
employers when the part-time contract is open-ended and its duration is more than 16 hours per week. The
incidence of part-time work has steadily increased in France, especially among women. With a level around the
European average, in 1997 it affected 16.6 per cent of total employment and 31 per cent of the female labour
force. By comparison, the level was 12 per cent in 199091 and 11 per cent in 1985, which implies that the
growth in the importance of part-time work is a relatively recent phenomenon.126
The ve-year law of 1993 sought to lay the basis for a more exible use of working time in France, although its actual
application was left to the social partners through collective bargaining. The most signicant change was the
introduction of new types of working time modulationthe calculation of working hours on a yearly basis and the
distribution of these hours through out the year. Three types of modulation are currently available to French rms:
1. Type 1, introduced in 1982, is used to vary weekly hours around the legal maximum of 39, which must be the
average amount of hours worked during the modulation period. Overtime premiums or time-off must be
given for hours exceeding 39 but they may not be counted against the yearly maximum overtime hours of 130,
and the absolute weekly maximums of 48/46 hours must be respected.
2. Type 2, introduced in 1987, is similar to type 1, except that the maximum weekly hours must be 44, or 48 if
agreed at the sector level, and overtime is not to be paid at a premium or otherwise compensated. However,
this modulation entails a quid pro quo in terms of working time reduction, nancial compensation, training
schemes or job creation.
125
In exceptional cases, the benet may reach 100 per cent through a special subsidy designed to prevent economic dismissals. As suggested by Mosley and Speckesser (1994),
the French system is substantially less generous than either the German or the Italian ones.
126
See The Structure of Employment and unemployment below for the detailed gures, based on Gissot and Mercier (1997) and Brunhes Consultants (1997).
France 257
3. Type 3, introduced in 1993: in this case, the reference period is a year and the only limits are the absolute weekly
maximums. Overtime hours must be remunerated, and rms must undertake to reduce weekly working
hoursthe newly-agreed working week being the one used to dene overtime.
The latter type of modulation is generally regarded as the most exible for rms, given its use of the year as the
reference period. However, it appears the this new type of working time exibility has not been extensively used in
collective agreements (OECD 1995).
The ve-year law also changed the regulation of short-time working. A new special scheme was introduced which
extended the maximum number of short-time hours to 1200 over a period of 18 months. While no change was made
to the benet level, the at-rate subsidy for hours in excess of 700 was raised, although part of this was to be paid out
of unemployment insurance funds (UNEDIC). The reform was intended to be temporary, and in fact it was
discontinued at the beginning of 1996.
Overall, working time exibility in France appears to have increased in recent years, and French rms apparently do
not face signicant problems in their deployment of human resources over time.
Unemployment Compensation
As in other European countries, unemployment compensation in France consists of two levels of benet: insurance
benets and assistance benets. The specic feature of the French situation is that the system is managed by agencies
created by the social partners (UNEDIC and the ASSEDICs), although this is not a unique arrangement in Europe.
The French system of unemployment compensation has traditionally been regarded as one of the most generous in the
OECD area (OECD 1991). The picture grows less clear, however, after the reforms introduced in 1992 in order to cut
both the expense of the system and its coverage. As usually happens, these cuts were not prompted by economic
analysis of the disincentive effect of unemployment benets, but rather by the need to curb expenditure.
The changes introduced can be summarised as follows:
1. The minimum entitlement conditions have been tightened: whereas in the previous system the minimum
requirement was 3 months of contributions in the last 12 months, it is now 4 months over the past 8 months.
Although the reference period has been shortened, increasing the obligatory contribution period means that
people in very short-term employment are excluded from any compensation.
2. For the unemployed with a contribution record amounting to 6 months over the last 12, the previous system
provided a maximum entitlement period of 21 or 15 months depending on whether the person was older or
younger than 50; under the present regulations, these people are entitled to only 7 months of benets, unless
they have paid contributions for 8 months, in which case the same entitlement periods apply as before.
3. For those with contribution periods of 12 months over the last 24, the previous system provided a maximum
period of 30 or 45 months, depending as before on the age of the person; people with 6 months of
contributions over the previous 12 but with an overall contribution record of 10 years in the previous 15
received the same benet. Under the present rules, 14 months of contributions in the past 2 years are required
to qualify for these benets, and the second possibility has been eliminated.
4. For workers older than 50 with contribution periods amounting to 24 months over the last 36, the previous
system xed an entitlement period of 45 months, or 60 months for those over 55; under the new system, 27
months of contribution are required.
5. Whereas the old system envisaged various types of insurance benet, such as basic, extended, end-of-
entitlement, extension of end-of entitlement, the new system comprises a single basic benet which, after a
certain period, is reduced by a xed amount every four months. The interval between these reductions was
extended to six months in 1996. The new system apparently implies an overall lower substitution rate (OECD
1995).
6. The previous system also provided unemployment assistance benet for workers with an employment record
of 5 years over the preceding 10, and a smaller benet paid for an indenite period renewed every six months.
The new system has preserved this so-called solidarist payment.
The OECD considers these changes to be a deep modication of the unemployment compensation system (OECD
1995). According to the Ministry of Labour sources quoted by the OECD study, the beneciaries of the insurance
system decreased by more than 8 per cent in 1994, from 70 per cent of the unemployed to less than 60 per cent. The
average substitution rate also fell, from 54.3 per cent to 51.7, and the average monthly payment decreased by 10 per
cent in real terms between 1990 and 1994. These reductions meant, of course, a substantial decline in the level of
expenditures by the system.
These changes were partly off-set by increased take-up in solidarist systems: not only the unemployment assistance
benets but also the
France 259
Revenu minimum d'insertion (RMI) created in 1989 as a minimum guaranteed income programme. Many of those
excluded from unemployment compensation applied to the RMI instead: with the consequence that although these
solidarist systems did not change their regulations, the number of people beneting from them increased.
What, then, have been the effects of these changes? The presumed negative effects of unemployment benets are well-
known: they reduce the motivation of the unemployed to search actively for jobs. Moreover, studies carried out in
various countries suggest that the exit rates from unemployment are higher in the rst months of receiving
unemployment benets than when these benets approach exhaustion. Interestingly, there are relatively few studies
specic to the French situation. Two of the most recent ones, both published in 1996, refer to data for the late 1980s
(Cases 1996; Joutard and Ruggiero 1996), and they largely conrm the above presumptions. Consequently, the
199293 cuts in the unemployment system should have implied higher wage exibility, and presumably have acted
favourably on employment. As far as we can tell, however, no studies exist with which to assess the effects of the
reform.127
Overall, therefore, although the unemployment compensation system has been reformed in a manner presumably
favourable to exibility, the results of these changes remain to be seen.
The Structure of Employment and Unemployment

Figure 9.1 has already shown the general evolution of employment and unemployment in France. On the whole, two
distinct periods can be identifed since 1985: rst, an increase in employment and a decrease in unemployment between
1985 and 1990, followed by a relatively stable level of employmentmore precisely, a decline until 1993 and then a
certain amount of recoverytogether with a continuous rise in unemployment, with the sole exception of 1995,
thereafter.
The increase in employment during the rst period may be linked to the rise in xed-term work, although this was a
period of economic expansion throughout Europe. As so often happens in the social sciences, the impossibility of
carrying out controlled experiments prevents more solid conclusions from being reached.
127
If they are to be rigorous, these evaluation studies must be based on panel data, which take a long time to gather; this probably explains the lack of studies on the effects of
the reform.
What has been the evolution of employment in terms of its composition? The rst dimension to inspect is the
structure of employment by economic sector. Table 9B in the Appendix shows the evolution for the period 19851995
and provides evidence that the French economy has continued to lose jobs in agriculture and industry, including
construction, in favour of services. However, as mentioned in the Introduction, this growth of the service sector is
considered in France to be anomalous with respect to other European countries. Indeed, the most recent OECD
gures show that the weight of services in employment is larger in France than in most European countries. This is an
interesting nding in view of the idea, mentioned in the review of the national debate, that the French service sector is
underdeveloped.
Table 9C, which sets the rates of employment variation by industry using a more detailed breakdown, may provide a
clue to this apparent paradox. In the table, the sectors are ranked from greater to lesser employment creators. The
gures cover the period 199296, the most recent period after the most recent changes described in the previous
section. Non-market services were the only sector to create a signicant amount of employment: above 7 per cent.
This suggests that the public sector probably acted as a buffer in these years to prevent what might otherwise have
been constant employment decline. This sector is by far the largest employer in the French economy. Miscellaneous
Market Services, mostly personal services, is the other sector that displays employment increase, and especially so in
1994 and 1995 when its growth rate was above 3 per cent. Two other service industries, Distribution and Transport
and Telecommunications, show a constant level on average over the period. All the remaining sectors, including Real
Estate/Insurance/Finance, Manufacturing, Construction and Agriculture, experienced employment losses over the
period.
The above-mentioned paradox concerning the weight of services in French employment can thus be interpreted in
terms of the inability of the service sector, especially market services, to outweigh continuing losses of employment in
manufacturing and agriculture. Construction is always a more pro-cyclical sector and in fact registered a small
employment increase in 1995. According to a recent study which has enjoyed wide publicity in France, the number of
jobs per capita in retail and hotels-restaurants was in the United States 80 and 130 per cent respectively higher than in
France (Piketty 1997). The author relates these differences to the much higher relative cost of unskilled workers in
France, a point already mentioned in the section on the wage-determination system.
Table 9D presents various other features of employment. The proportion of women in total employment has
continued to grow, which is
France 261
probably linked with the sectoral evolution analysed above. As for self-employment, its weight in total employment has
followed a downward sliding trend, and is now approaching a proportion of 10 per cent. This gure is clearly below
the EU average of 15 per cent (European Commission 1996), a nding probably due to the relatively low weight of
agriculture in total employment, although it may also indicate that the French do not see self-employment as a real
employment alternative.
A similar picture emerges for part-time work. The proportion of part-timers has followed a rising trend, as has been
the case elsewhere in the European Union. The current level of 16 per cent exactly matches the EU average,128
although the female rate is somewhat lower29.5 as compared to 32 per cent in 1996, the latest gure available for
the EU.
Regarding xed-term work, as dened in Eurostat terms,129 there was a constant rise between 1985 and 1990,
coinciding with employment growth, which lends some weight to the notion that the measures in favour of xed-term
contracts passed in 1985 were probably benecial to employment.130 Since 1990, the proportion of temporary workers
has remained largely constant, with a slowly rising trend, although some acceleration is evident in 1995 and 1996. As in
the case of part-time work, this gure is close to the EU average.
Figure 9.2 shows the evolution of unemployment broken down by gender and age, as well as the proportion of the
long-term unemployed. Gender differentials appear to have remained relatively stable over time, with the female rate
some 4 points higher than that for males. The rate of unemployment among young people under 25 follows a pattern
similar to the overall rate, but its variations are more marked. Thus, whilst young people beneted from the
employment recovery of 198590 more than did their older counterparts, they were the more signicant losers in the
recession of the early 1990s. The slight recovery of 1994 led to only a minor improvement in the youth unemployment
rate, which continued to rise in 1996 and 1997. In total, whereas the overall rate increased from 8 per cent in 1990 to
more than 12 per cent in 1997, the rate for young people increased from 20 to almost 30 per cent.131 Of
128
It is well known that there are wide differences within this average. France thus just happens to be in the middle. See Eurostat (1997b).
129
The Eurostat denition stems from a question regarding the temporary nature of employment. The gures from the Enqute Emploi yield much smaller percentages of
xed-term workers, i.e. workers with a contrat dure dtermine ).
130
Of course, it may be that jobs created under this special scheme would have been created in any case. This deadweight effect is very difcult to estimate, however.
131
The precise gure varies depending on the source. According to the Enqute-Emploi, the youth unemployment rate was 27.9% in 1997. However, according to Eurostat's
homogeneous gures, it stood at 30.0%.
Fig. 9.2. Unemployment rates: gender, youth and long-term unemployed, France, 19851997
Sources: Eurostat and Enqute Emploi, linked by authors.
France 263
course, the more marked evolution of the youth unemployment rate only reects the fact that recessionary and
expansionary periods are usually associated with the blocking and opening of labour market entry, which obviously
affects young people to a greater extent. However, it is worth mentioning that while the overall French unemployment
rate is close to the EU average12.4 per cent in France as opposed to 10.9 for the EU in 1996the youth
unemployment rate in France is signicantly higher: 28.9 as opposed to 21.8 per cent in the EU as a whole (Eurostat
1997a). This suggests that France may have a special problem with youth unemployment compared with the rest of
Europe.
Figure 9.2 also shows the evolution of the proportion of the unemployed in search of work for more than one year:
the long-term unemployed. The path followed by long-term unemployment is quite understandable. First, during the
long unemployment recovery of the late 1980s, the decrease in unemployment was matched by a decrease in the
proportion of those in search of a job for more than one year. This means that during that period both the short-term
and the long-term unemployed beneted from employment creation, and there was consequently no special long-term
unemployment problem, or at least it did not worsen during that period. The rise of unemployment related to
employment losses meant a continuous ow of entrants into unemployment, thereby increasing their share of total
unemployment and reducing the proportion of the long-term unemployed. Finally, when unemployment continued to
rise but employment stabilized, long-term unemployment started to accumulate again as the ow of entries into
employment diminished and the ow of exits remained relatively stagnant. Despite the recent build-up, however, long-
term unemployment accounts for a much smaller proportion of overall unemployment in France than in many other
European countries: the EU average in 1996 stood at 48 per cent, 10 points above the gure for France, and only
Sweden, Austria, Denmark, Finland and Luxembourg had smaller proportions (Eurostat 1997a).
Figure 9.3 provides a further interesting breakdown of unemployment, this time by level of education. These gures
are taken from the Enqute-Emploi and refer to the period 199197. It will be seen, in general terms, that all the
unemployment rates shown display similar trends in the 1990s. There appear to be three levels: people without
qualications display a consistently higher unemployment rate, which has remained more or less stable at 17 per cent
since 1994. Those with some sort of secondary education show rates somewhat lower than the overall average with
some advantages for those who have completed their upper secondary educations, the Baccalaurat. Finally, the
Fig. 9.3 Unemployment rate, by level of education, France, 19911997

Source: Enqute Emploi.
France 265
unemployed with university degrees show distinctly lower rates: around 8 per cent for those who have completed some
university-level study but do not have a degree and around 7 per cent for those with a university qualication.
This pattern prompts two considerations. Firstly, it is clear that France suffers from a specic problem of
unemployment among the less-educated, or ess-skilled, if education is taken as a proxy for that variable. In this sense, it
could be argued that, as already pointed out, the problem is related to the relatively higher labour costs of these
unskilled workers. If this approach is correct, then the 1993 measures aimed at reducing this relatively higher cost have
yet to exert their expected effects on the differential unemployment of the unskilled. The problem, however, may be
that relative costs are only a part of the story. If there is a shortage of jobs for macro-economic reasons, due to
restrictive macro-policy for example, and the labour market functions on the basis of internal markets, then a job
queue will develop in which the less-skilled, that is, those with higher training costs, will be those with a lower
probability to get hired. As a consequence, the high relative unemployment rate of the unskilled can be analysed as a
crowding-out or ladder' effect, the more skilled being hired on low skilled jobs, which, in return, induces an
occupational downgrading for them. Thus, lowering relative labour costs may be expected to favour the position of the
unskilled in the job queue but the effect of such a measure will not be in any way decisive (see Gauti 1997).
Secondly, the relatively small differences between the two pairs of levels corresponding to secondary and university
education should also be stressed. It seems that the French labour market has underestimated the value, in terms of
unemployment, of completing these levels of education. The explanation may be related to the fact that the majority of
students drop out of university before nishing their degree courses, which suggests that more middle-level degrees or
diplomas are required. Alternatively, generation effects may be at work: if young people show a greater tendency to
nish their secondary/university educations, this may explain their relatively higher unemployment rates. In addition, if
nding rst employment is a similar experience for all young people in terms of the time that it takes them, regardless
of their level of education, then it may be that the relatively high, or less-than-expected low, unemployment rate among
the highly educated masks this rst job search process. The published data do not permit more in-depth analysis which
could discriminate among these possible explanations.
Concluding Comments
On the whole, the situation in France in the late 1990s reveals quite substantial stability. After the 199093 crisis, the
economy levelled out, implying a more or less stable employment level and a constantly rising unemployment rate, fed
by the constant growth of the labour force. The various changes made to the labour market institutions have not been
sufcient to ease France's unemployment problem. Of course, the changes introduced are not particularly dramatic
from a comparative point of view; indeed, they may be called meagre.
Although the French debate does not appear to have viewed the lack of labour market exibility as a cause of the
problem, other authors and institutions, most notably the OECD (1997c) argue that the French labour market requires
more forceful action to be taken in this area. However, there is no evidence either for or against this position. Those
measures that have been introduced, such as the easing of xed-term contracts and the elimination of the
administrative authorization for dismissals in 1986, have been followed by a period of employment creation; but it
cannot be argued in any way that they were instrumental, not even partially, in this employment recovery. And,
although it could be said that these measures were partly reversed in 1989 and 1993, when collective dismissals were
required to provide more protection for the dismissed workers, the main substance of the changes persisted. Yet they
were not sufcient to prevent the employment losses of the early 1990s, nor were they instrumental in provoking the
stronger employment stance of the mid-1990s. Of course, the other measures adopted, as modest as they arethe
reduction of social contributions by the low-paid, reform of the unemployment insurance systemhave not produced
signicant changes in the level and composition of employment and unemployment. The French case probably
provides evidence which the defenders of deregulation can use to argue that more is required to solve unemployment
problems, but it equally yields arguments for those who maintain that institutional changes in the labour market
framework are not conducive to better employment and unemployment conditions.
France 267
Appendix
Table 9A. Institutional Changes in France
Institutional changes Year Effects on exibility
Enhanced role of works councils 1982 -
Fixed-term contracts for short du- 1982 -
ration in specic areas
Incentives for temporary contracts 1985 +
in job creation programmes
Removal of limitations on xed- 1986 +
term contracts
Abolition of prior authorization for 1986 +
collective dismissals
Lower entry wage for young work- 1986 +
ers
Provisions for the prevention of 1989 -
dismissals
Reform of the unemployment in- 199293 +
surance system
New working time schedules 1993 +
Source: Adapted from Table 6 in Bertola and Ichino (1995).
Table 9B. Distribution of Employment by Major Economic Sector, France, 19851995

Sector/Year 1985 1990 1991 1992 1993 1994 1995
Agriculture 8.2 6.4 6.0 5.9 5.5 5.2 4.9
Industry 32.4 30.4 30.0 29.6 27.5 26.9 27.0
Services 59.4 63.2 64.0 64.5 67.0 67.9 68.1
268 References
Table 9C. Evolution of Employment by Industry, France, 19921996

Industry Employment in 1996 (1992 = 100) Share of each industry in total em-
ployment, 1996 (%)
Non-market services 107.5 30.0
Miscellaneous market services 105.5 22.7
Distributive services 100.3 12.9
Transport and telecommunications 100.2 6.2
Fuel and power products 97.6 1.1
Renting, insurance, nance 96.1 2.8
Agro- and food industries 95.1 2.6
Land transport equipment manufactur- 93.9 1.7
ing
Intermediate goods manufacturing 91.4 5.2
Industrial capital goods manufacturing 91.3 4.6
Construction 89.3 6.8
Non-durable consumer goods manu- 87.9 4.7
facturing
Household durables manufacturing 87.7 0.2
Agriculture 86.1 4.8
(Average) 99.7 100.0
Source: National accounts gures, as reported in INSEE, L'conomie franaisedition 19971998, Paris, 1997.
Table 9D. Atypical Forms of Work, France, 19851997 (% Of Total Employment)

1985 1990 1991 1992 1993 1994 1995 1996 1997
Self-em- 12.6 12.9 12.6 12.5 11.8 11.8 11.6 NAa NA
ployed
Part-time 10.9 11.9 12.1 12.7 13.9 14.9 15.5 15.8 16.6
work
Both
genders
Females 21.8 23.6 23.5 24.5 26.3 27.8 28.9 29.5 30.9
Males 3.2 3.3 3.4 3.6 4.1 4.5 5.0 5.2 5.2
Fixed-ter- 4.7 10.4 10.1 10.5 10.9 11.0 12.2 12.5 NA
m work-
ers
a
NA = Not available.
Source: EUROSTAT and Enqute-Emploi.
References
Braud, M., Lallement, M., and Lefvre, G. (1994), Job protection and work sharing in France: from short-time
working to compensated reduced working of long duration, Paper presented at the at the Conference on Short-
time Working and Structural Change, organized by the Wissenschaftzentrum-Berlin WZB, Berlin, December.
References 269
Bertola, G., and Ichino, A. (1995), Crossing the river: a comparative perspective on Italian employment dynamics,
Boyer, R. (1995), The Future of Unions: Is the Anglo-Saxon Model a Fatality, or Will Contrasting National
Trajectories Persist?, British Journal of Industrial Relations, 33/4: 54556.
Brunhes Consultants (1997), tude sur le march du travail: France. Labour Market Studies Series, Brussels: European
Comission.
Cette, G., and Taddei, D. (1994), Temps de travail, mode d'emploi. Paris: La dcouverte.
Charraud, A. (1993), The Impact of Temporary Work Legislation in France, 19801989, in C. Buechtemann, (ed.)
Employment Security and Labor Market Behavior. New York, NY.: Ithaca, ILR Press, 37482.
Colin, C. (1995), Les primes et complments de salaire dans le secteur priv INSEE Premire, 406, October.
CSERC (Conseil Suprieur de L'Emploi, des Revenus et des Cots) (1996), L'allgement des charges sociales sur les bas
salaires. Paris: La Documentation Franaise.
DARES (Direction Gnrale de L'Animation, de la Recherche, des Etudes et des Statistiques) (1994), Dix ans
dvolution des salaires dans le secteur priv', Premires Synthses, 96.
(1995), Lindividualisation des salaires recule dans les PME en 1994, Premires Synthses, 102.
Drze, J. H., and Sneessens, H. (1996), Technological development, competition from low-wage economies and low-
skilled unemployment, in D. Snower and G. De la Dehesa (eds.), Unemployment Policy. Government Options for the
Labour Market. Cambridge: Cambridge University Press.
European Commission (1996), Employment in Europe1996. Luxembourg: Ofce for Ofcial Publications of the
European Communities.
EUROSTAT (Statistical Ofce of the European Communities) (1997a), Labour force survey, Principal results 1996, Eurostat
Statistics in focus, Population and social conditions. Luxembourg: Statistical Ofce of the European Communities, 8/97.
(1997b), Part-time work in the European Union, Eurostat Statistics in focus, Population and social conditions. Luxembourg:
Statistical Ofce of the European Communities, 13/97.
Flanagan, R. J., Soskice, D., and Ulman, L. (1983), Unionism, Economic Stabilization and Incomes Policies. European
Experience, Washington: The Brookings Institution.
Fougre, D., and Kamionka, T. (1992), Mobilit et prcarisation sur le march franais du travail: une analyse
longitudinale pour les annes 19861988, Economie et Prvision, 102/103: 15778.
Gauti, J. (1997), Insertion professionnelle et chmage des jeunes en France, Regards sur l'actualit, 233 JulyAugust.
Gissot, C., and Mercier, M. A. (1997), Chmage et emploi en 1997: premier rsultat de l'enqute sur l'emploi, INSEE
Premire, June: 530.
Goetschy, J. (1998), France: the limits of reform, in A. Ferner and R. Hyman, (eds.), Changing industrial relations in
Europe. Oxford: Blackwell, 35794.
270 References
INSEE (Institute National de la Statistique et les Etudes Economiques) (1997), L'conomie franaisedition 199798.
Paris: INSEE.
(1998), L'conomie franaisedition 199899. Paris: INSEE.
Lyon-Caen, A. (1993), Workers' Protection and the Regulation of Labor Relations in France during the 1980s, in C.
Buechtemann, (ed.) Employment Security and Labor Market Behavior. New York, NY.: ILR Press, Ithaca, 34757.
Maurau, G. (1993), Regulation, Deregulation, and Labor Market Dynamics: The Case of France, in C. Buechtemann,
(ed.) Employment Security and Labor Market Behavior. New York, NY.: ILR Press, Ithaca: 35873.
Maurin, E., and Torelli, C. (1992), RMI et comportements sur le march du travail, Economie et Statistique, 252: 3742.
Mosley, H., and Speckesser, S. (1994), Short-time working in France, Germany and Italy; incentive structures, uptake
and labour market impacts, Paper presented at the Conference on Short-time Working and Structural Change,
organized by the WZB, Berlin, December.
OECD (1991), Employment Outlook. Paris: OECD.
(1995), Economic SurveysFrance, 1995. Paris: OECD.
(1997a), Employment Outlook. Paris: OECD.
(1997b), Labour Force Statistics, 1997 Edition. Paris: OECD.
(1997c), Implementing the OECD Jobs StrategyCountry Reccomendations. Paris: OECD.
Piketty, T. (1997), Les crations d'emploi en France et aux Etats-Unis: services de proximit contre petits boulots, Note de la
Fondation Saint-Simon, December.
Redor, D. (1997), Does the Public Sector Lead the Wage Evolution in the French Economy, Labour, 11/2: 35172.
Rogers, J., and Streeck, W. (1994), Workplace Representation Overseas: The Works Councils Story, in R. Freeman
(ed.), Working under Different Rules. New York: Russell Sage Foundation, 97156.
Rojot, J. (1995), Proli giuridici, in La gestione delle Eccedenze di Personale in Europa, Documenti CNEL (Consiglio
Nazionale dell'Economia e del Lavoro) 58/2: 6498.
Saint-Paul, G. (1996), Dual Labor Markets. Cambridge, Mass.: MIT Press.
Segrestin, D. (1993), Recent changes in France, in G. Baglioni and C. Crouch (eds.), European industrial relations. The
challenge of exibility. London: Sage Publications.
Siebert, H. (1997), Labor Market Rigidities: At the Root of Unemployment in Europe, The Journal of Economic
Perspectives, 11/3: 3754.
Sneessens, H., and Shadman-Mehta, F. (1995), Real wages, skill mismatch and unemployment persistence, Annales
d'Economie et Statistique, 37/38: 25592.
10 Italy: The Long Times of Consensual
Re-Regulation
Manuela Samek Lodovici
Introduction
International comparisons point to Italy as one of the most rigidly regulated labour markets in the industrialized world,
especially in terms of hiring and ring rules and of limits on the use of atypical contracts (Lazear 1990; Grubb and
Wells 1993). However, this is offset by a complex mix of decentralized small rms, which are subject to lesser
constraints on employment relationships; a large share of the black economy and non-dependent labour, and
widespread use of social shock absorbers. These peculiar dimensions of labour market exibility have produced
distortive effects on the structure and performance of the Italian labour market which, only in the late 1990s, are
starting to be tackled by structural reforms.
The Italian labour market is characterized by a low level of regular employment and a greater segmentation than other
OECD countries. Segmentation cuts across gender, age prole, skills and territory, and reects large disparities in the
labour market protection of different components of the labour force (Table 10.1).
The employment and participation rates are signicantly lower than the EU average, especially for women and young
people, and in Southern Italy. The employment rate has been declining since the early 1980s, especially for the least
educated. Employment composition is also different from other European countries: regular dependent employment is
relatively limited, as is the incidence of regular part-time and xed term employment, which are respectively 7 per cent
and 8.8 per cent of total employment in 1997. On the other hand the proportion of self-employment is high, at about
32 per cent of total employment in 1997, and non-regular work accounts for 22.5 per cent of total work (ISTAT 1997).
Table 10.1. Italy: Main Labour Market Indicators, 1997

Italy Centre-North South
Employment rates (%)
young peoplea 24.7 32.3 13.4
men (total) 55.4 58.6 49.5
women (total) 29.0 34.1 19.2
Total 41.7 45.9 33.9
Participation rates
(% of working age pop-
ulation)
young people 37.6 41.5 30.7
men (total) 61.2 61.9 60.2
women (total) 34.8 38.5 27.9
Total 47.5 49.7 43.5
Unemployment rates
(%)
men (total) 9.5 5.1 17.9
men (adults) 6.8 3.6 13.1
women (total) 16.8 11.4 31.0
Total 12.3 7.6 22.2
Long-term unemploy-
ment (12 months and
over)
(% of total unemployment)
men (total) 68.2 56.9 74.4
women (total) 68.0 59.5 76
Total 68.1 58.5 75.2
a
1524 years.
Source: Istat, Labour Force Survey, 1997.
Unemployment rates are higher than in most OECD countries and display very marked territorial and gender/age
disparities. In 1997 the gap between the Southern region with the highest unemployment rate, Campania, and the one
with the lowest, Trentino-Alto Adige, was 22 percentage points. This gap has been widening since the late 1980s. The
gender gap is the second highest in Europe, after Spain, while the youth unemployment rate is the highest, at 33.5 per
cent and the gap with the adult unemployment rate the widest: 24.5 percentage points. Italy has the highest percentage
of long-term unemployment among the industrialized countries, at 68%. In contrast with other European countries, 60
per cent of the long-term unemployed are young people in search of rst employment.
The nineties have been characterised by wage moderation and declining unit labour costs, thanks to high productivity
gains. The scal and contributive wedge on dependent labour is, however, still high relative
Italy 273
to other European countries, and the dispersion of earnings by profession, and labour costs differentials by industry
and territory, are low compared with productivity differentials, even if they have increased slightly in the 1990s.
A Review of National Debate and Literature

Most commentators regard the characteristics of the Italian labour market to be an expression of its regulatory
framework, especially the regulation of employment relationships, that is, wage-setting and employment rigidity.
Debate on labour market exibility has been ongoing in Italy since the early 1970s, but it has intensied since the mid-
1980s, when important measures were taken to reduce numerical rigidity. The debate and research has focused less on
functional exibility issues and the relationship between labour market exibility, labour quality and investment in
human capital.
Until the mid-1980s, the Italian approach to labour market regulation was generally accepted by the social partners.
The protection of core dependent workers and family heads by the strict regulation of employment relations, and the
use of public resources to support labour hoarding, reected the interests of both the trade unionswhose median
voter was the core industrial workerand rms, which did not have to bear the costs of labour hoarding and could
rely on high internal exibility. Nor were governments particularly keen on exibility: the state's monopoly of
placement, and the practice of discussing and agreeing every decision on collective dismissals in large rms with the
social partners, were important political instruments.
The effects of the Italian regulatory system during the 1970s and 1980s are widely acknowledged. In the rst place, it
reduced adjustment costs during recessions and the costs of industrial restructuring for both workers and rms. Until
the mid-1980s, the costs of employment adjustment in medium and large rms were largely borne collectively through
the use of temporary lay-offs and other measures. According to some authors (Giavazzi and Spaventa 1989; Mariani
and Zeli 1995), CIG (the Italian Wages Guarantee Fund) and other income support schemes, were a way to transfer
resources to prots. They thus nanced industrial reorganization and helped to improve the long-term performance of
the Italian industrial sector. The costs to the public nances were relatively low in relation to other European countries,
due to the low extension
of benets coverage. These measures also greatly helped in reducing industrial and social conict, especially in
Southern Italy, where income support may last for many years, and during phases of indus-trial restructuring. In
addition, they encouraged the decentralization of rm-level collective bargaining on more exible working time and
working conditions. The negative aspects of the systemhigh segmentation of the labour market, with difculties of
access to employment for youth and women, and the large-scale growth of an underground economywere
considered necessary costs for socially desirable distributive goals.
Growing pressures from international competition and constraints on public spending, together with declining union
power, have changed the social partners' attitude towards labour market exibility in the early 1990s. Trade unions and
workers have accepted exibility as necessary to ght high unemployment, but call for rules to prevent a totally
unregulated market. Their position is well evidenced in a document written by the former socialist Labour Minister
Giugni, the father of the 1970 Workers' Statute, which urges the introduction of regulated exibility (Ministero del
Lavoro 1994). Employers and their organizations, for their part, stress the need to increase exibility and mobility in
the labour market through the deregulation of labour contracts, fewer constraints on hiring and ring decisions, closer
links between wage rates and company performance, and more local or plant-level decision making.
In the important income agreement of July 1993 the unions accepted reduced constraints on the use of xed-term and
part-time contracts and the introduction of temporary or agency work, which entailed the abolition of the public
monopoly of placements. In addition, the agreement abolished wage indexation and increased the scope of local/ rm-
level bargaining. On the other hand, the difculties and lengthiness of the legislative process, which only in 1998 led to
the introduction of agency work, is indicative of the division in public opinion concerning these issues and the
impracticability of strong and rapid deregulation in Italy.
Most authors (Lunghini 1994; Rapiti 1994; Salvati 1988) argue that greater labour market exibility is not in itself
enough to solve Italy's labour market problems. Moreover, the incentivation of atypical contracts, without revising the
protection given to permanent workers, may produce negative substitution effects, as the Spanish experience suggests.
Italian unemployment and the low use of human resources result from structural problems that must be addressed
with medium-term measures. Greater exibility should be accompanied by more efcient employment services and
training opportunities, and by renewed
Italy 275
investments in the South, where exibility measures are unable to overcome structural unemployment by their own
(SVIMEZ 1996).
More recent research is concerned with the real degree of exibility achieved in recent years, and its effects. While it is
widely acknowledged that wage exibility has increased since the income agreements of 1992 and 1993, there is less
agreement on the degree of external exibility and labour mobility produced by legislation on hiring and ring
regulations and on atypical contracts. Christodes (1996) detects a structural break in the Italian labour market
conditions since the early 1990s which he associates with the 1991 legislation on collective dismissals; other authors
(Boeri 1997; Ganoulis 1997; Boragine and Fabi 1997; Demekas 1995) argue that this legislation, by regulating collective
dismissals, only eliminated one anomaly in the Italian system, but did not bring comprehensive change to the general
regulatory framework. According to these authors, segmentation across rms has been reinforced, and small rms are
still penalized by the high costs associated with collective dismissals and the lengthiness and complexity of the related
procedures. In the early 1990s layoffs were facilitated by the widespread use of early retirement, but later on budget
constraints have forced a reduction of these measures and ring and hiring procedures remain relatively rigid.
Bertola and Ichino (1995) provide a good summary of the ongoing debate. In their view, Italy is now going through a
transitional stage between extreme rigidity and exibility. The 199193 recession coincided with a greater degree of
external exibility engendered by legislation on collective dismissals and xed-term contracts. This gave rise to more
rapid and deeper employment reduction than did earlier recessions. The move towards exibility should be rapid in
order to ensure that recovery brings the employment and efciency improvements that one would expect. Transition
to exibility should therefore be accelerated by greater credibility and by paying closer attention to its redistributive
effects. However, the case-by-case approach followed to date has not improved credibility: rms and workers expect
that a specic solutionusually in terms of income support and employment guaranteewill be found, with the costs
externalized to the taxpayer.
As regards wage exibility, debate in the late 1990s centres on the effectiveness of the 1993 income agreement and its
evolution in a situation of low ination. According to most commentators, there is now a need for a bargaining pattern
more closely linked to local and plant conditions, since wage rates should increasingly reect productivity changes
(CNEL-IRS 1997a). The existence of legally binding wage oors, negotiated for each sector and category at the
national level for every part of Italy, in the presence of wide productivity differentials,
is argued to be one cause of the high level of Southern unemployment, and for unemployment among women and
young people (Casavola et al. 1995).
The debate also concerns pensions and income support policies. The lack of a universal scheme of income support
during unemployment, together with the inefciency of the public placement services, are considered by most
commentators to be major obstacles against labour mobility (Boeri 1997; Dell'Aringa 1995; Franco and Sestito 1995).
Reform of the income support system has been mooted since the late 1980s, but the sensitivity of the issue and the
many interests involved have prevented any radical innovation. The debate has recently been fuelled by the revision of
the pension system and by the growing proportion of precarious jobs. The nancial sustainability of the current
pension system and the need for intra-generational equity are being discussed, together with the necessity to reduce the
other negative effects of the current social protection system, which fails to provide an effective safety net for those
with weak or absent links with the labour market (Rostagno and Utili 1998)
The Dynamics of Rigidity/exibility: Trends in Labour Market

Regulation
The Italian approach to regulation is based on social concertation. Legislative regulation covers employment relations,
but not wage and working hours, which are left to bargaining arrangements. Regulatory legislation has usually followed
collective or tripartite agreements between the social partners, or else it has been an ad hoc response to particular
problems. The result is a hodgepodge of case-by-case measures, with no clear, comprehensive framework.
It is only with the 1990s that there was some progress towards increased exibility at both the legislative and the
contractual agreement levels. Table 10.2, which updates Bertola and Ichino (1995), presents the development of
institutional regulation in Italy. The main stages are the 1983, 1992, 1993, and 1996 national agreements providing for
greater exibility in wages and labour contracts; the 1984, 1987, and 1997 legislation on employment contracts, and the
1991 legislation on collective dismissals.
The 1997 so-called Treu Package, the most recent step towards greater exibility, is indicative of the Italian approach
to deregulation: some easing of regulation on hirings, rings and atypical contracts while
Italy 277
Table 10.2. Institutional Labour Market Flexibility Developments in Italy

Institutional change Year Internal External Labour cost
Prohibition of interme- 1960 -
diation in hiring
Restriction on temporary 1962 -
labour contracts
Regulation of ring pro- 1966 -
cedures
Cassa Integrazione Guadagni 1969 + -
Straordinaria
Charter of workers' 1970 - -
rights (Statuto dei Lavora-
tori)
Removal of temporal 197278 +
limits to CIG
benets and extensions
in the use of CIG
Reform of procedures 1974 - -
for labour litigations
Egalitarian wage index- 1975 -
ation clause (Scala Mobile)
Reduction of wage in- 1977 +
dexation coverage
Youth employment 1977 +
measures
Industrial restructuring 1977 + -
Possibility of temporary 1977 +
contracts in tourism and
trade
Restrictions on workers' 1978 +
mobility across rms
Payroll tax reductions 1980 +
(Fiscalizz. Oneri Soc.)
Reduction of contractual 1983 +
indexation
Solidarity contracts and 1984 +
work sharing
Temporary training con- 1984 + +
tracts (Contratti di Forma-
zione e Lavoro)
Part-time and exitime 1984 +
contracts
Initial liberalization in 1984 +
hiring procedures
GEPI and extension of 1985 + +
CIG benets
Reduction of payroll 1986 +
taxes in the south
Reform of indexation 1986 +
system
Renewal of temporary 1986 + +

training contracts
Extensions in the use of 1987 + +
xed-term contracts
Flexible compensation 1988 +
system at FIAT
Reform of ring regula- 1990 -
tions (L.108)
Reform of CIG and lay- 1991 - +
off procedures (L.223)
Elimination of indexation 1992 +
(end of Scala Mobile)
Counter-reform of CIG 1993 + -
and layoff procedures
The Giugni income 1993 + +
agreement
Extension of CIG 1993 +
ad hoc rescue measure 1993 +
for specic plant cri-
ses
End of payroll tax re- 1994 -
ductions in Southern
Italy
Labour Agreement 1996 + +
New measures in the 1996
eld of social shock ab-
sorbers:
Temporary jobs +
schemes (LSU)
Extension of mobility +
benets to small rms
Realignment contracts +
Help to business start- +
up for the young
Increase of hiring -
quotas (from 6% to 12%)
in the areas with low
unemployment rate
Treu Package (applica- 1997
tion of the 1996 Labour
Agreement):
Introduction of +
Agency contracts (appli-
cation since Jan. 1998)
Incentives for working + +
time reductions
Incentives for part- + +
time contracts
Liberalization of xed- + +
term contracts
Reform of apprentice- + +
ship and work-training
contracts
Italy 279
Exension of monetary +
resources for temporary
jobs schemes (LSU)
Incentives to regional + + +
labour mobility
Extension of territorial + +
pacts and area contracts
Realignment contracts +
Privatization of em- 199798 + + +
ployment conditions in
public sector
Tax reform 199798 +
End of public place- 199798 + +
ment monopoly and de-
centralization of
employment services and
active labour market
policies
Draft legislation on 35 1998 - -
hours working week
Social Pact and Budget 1998
Law 1999: government
mandate to:
reform social shock 1999 + +
absorbers
reform education and 1999
training systems
reform employment 1999 +
incentives
reduce contributive 1999 +
and tax wedge on labour
costs
reinforce concertation 1999
Notes: Relevant measures are shaded.
+ indicates enhancement of exibility.
- indicates reduction of exibility.
blank cells indicate no effect.
Source: Bertola and Ichino's Table 3 (1995: 385) integrated with 199498 measures.
maintaining a strong regulative approach, all premised on concertation with the social partners. We may thus talk of
re-regulation instead of deregulation in the Italian case.
The December 1998 Social Pact conrms and reinforces the concervative approach to industrial and social relations by
involving local authorities, and by giving mandate to the executive for reforming the education and training system and
the system of social shock absorbers and labour incentives. The Social Pact also supports the bottom-up approach to
local development and regulation adopted since 1996 through the use of local pacts and area contractsthe so-called
Programmazione negoziata.
Employment Protection
The most important step toward external exibility occurred in 1991 with the introduction of legislation regulating
collective dismissals. Prior to 1991, collective redundancies were governed by collective agreements rather than by
legislation. Collective agreements strictly regulated the periods of notice and required consultation with the trade
unions, joint discussion of planned redundancies and examination of ways to avoid them. In 1991, following the EU
Directive, legislation on collective dismissals in some predened sectors132 was introduced in Italy. According to the
new rules, on consultation with the unions, redundant workers are laid off and placed on a mobility list which entitles
them to mobility benets, except in the case of workers from small rms, and to re-employment facilities.
The regulation of individual dismissals is, on the other hand, still extremely restrictive. Compared to other countries,
sanctions against unfair dismissals are particularly severe and widely applied. Legislation was passed in 1966, on the
basis of the 1965 national collective agreement. The 1970 Worker Statute strengthened protection against unfair
dismissals by extending its coverage to rms with at least sixteen employees and by introducing more severe sanctions.
In 1990 coverage was extended to all rms, although smaller ones are subject to less stringent sanctions in cases of
unfair dismissal. The rigidity of these norms is heightened by the lengthiness and complexity of procedures and by
uncertainty over the nal outcome, which largely depends on subjective interpretations by the labour courts of the
existence of justied grounds for dismissal. The entire process from notication of dismissal to decision by the court
may take up to ten years.133
132
Originally, the law only covered the industrial sector. It was then extended to large commercial rms and other small sectors in crisis.
133
Individual dismissals must be on justied grounds: either misconduct by the employee or an objective reason relating to the enterprise. An employee may challenge a
dismissal by ling a written statement within 60 days. The case is heard by a specialist judge (pretore del lavoro ) who decides whether the employer has demonstrated justied
motive. If the judge considers this not to be the case, the dismissal is void and the employer must compensate the worker for damage and re-admit him/her to the
workforce. Reinstatement as a remedy for unfair dismissal is obligatory in establishments with more than fteen employees. Should the worker not wish to be reinstated, he/
she receives additional compensation amounting to fteen months' pay. Smaller rms may choose between re-hiring or paying compensation, which depends on length of
service and size of the rm.
Italy 281
The stringency of dismissal regulations in Italy has been limited by two important measures widely used during
restructuring periods: the special wages guarantee fund (CIGS) and early retirement introduced in the late 1960s, to
ease industrial restructuring by providing generous benets for industrial workers. CIGS, which was absorbed by the
mobility allowance in 1991, supplemented income for a period that could last several years. Workers were not laid off;
instead, they maintained their employment contract with the rm. Early retirement schemes provide generous
incentives to labour force exit by industrial workers employed in rms receiving CIGS benets and with at least fteen
years of social security contributions. During the 1980s, CIGS and early retirement measures helped to restrict the
growth of open unemployment.134 CIGS became a de facto unemployment benet for workers who had been denitively
laid off even if they were still formally employed.
Hiring Regulation
The regulation of hirings was formally very rigid until 1991, when the obligation to hire in sequence from a waiting list
compiled by the Public Employment Ofces (collocamento obbligatorio) was abolished for the private sector. In 1994, the
nal constraints were removed, and now rms may hire workers directly, being only obliged to notify their hirings
within ten days. Hiring regulation remains, however, relatively restrictive by international comparisons: Italy still has
mandatory quotas for the long-term unemployed and workers on the mobility list, at 12 per cent of total hirings, and
for the disabled: 15 per cent of the work force for rms with over thirty-ve employees. Italian legislation also requires
rms to give preference to workers dismissed during the previous twelve months.
Atypical contracts for dependent employment, such as xed term, part-time and agency work have been liberalized
only in 199697. Legislation on atypical forms of employment denes the basic regulatory
134
During the 1990s, a reduction of 532,000 units in industrial employment translated into an increase of only 253,000 unemployed workers with previous employment
experience. In that period, the use of CIGS was equivalent to about 220,000 workers at zero hours on average per year. A further 32,000 workers on average per year were
involved in early retirement schemes.
framework, leaving the denition of possible extensions and implementation to sectoral collective bargaining.
Part-time work was legalized in 1984, but its opportunities have been only recently improved by adjustments of social
security contributions and pension provisions. Part-time work may be horizontal or vertical, with working hours clearly
specied in a written contract. The use and extent of part-time work is dened by collective bargaining at the national
and rm level; collective agreements may also envisage the possibility of overtime by part-time workers. Changes from
full-time to part-time contracts must be approved by the local Public Employment Ofce. These rules, combined with
the high level of protection afforded to part-time workers, which is substantially the same as full-timers; organizational
difculties in small manufacturing rms and the scarce willingness of Italian workers to accept part-time jobs,135 help to
explain why part-time work is little used in Italy, representing only 7.3 per cent of total employment in 1997.
In Italy xed-term contracts have been traditionally considered to be solutions for specic production problems, never
as a measure to combat unemployment. For this reason they were strictly limited by Italian labour legislation until
1987, when the denition of cases and forms of use was left mainly to collective bargaining. The 1997 Treu Package
provides for greater exibility in their use by increasing the number of times these contracts are renewable, and by
reducing sanctions in cases of violation of contract regulations. Monetary sanctions are instituted instead of the
automatic transformation of a part-time into a permanent contract. Since 1997 xed-term contracts have been
introduced for public managers.
Only a special form of xed-term contractwork-training contracts for young people, or contratto di formazione lavoro
(CFL)has been encouraged since 1984 through reductions in social security contributions and some simplication of
hiring procedures. These contracts are designed to improve the chances of permanent employment for young people
in need of training. For this reason, compared with the less regulated Spanish experience, their introduction did not
radically increase hiring probabilities and did not reduce the average duration of unemployment among young people,
although it did improve their likelihood of nding stable employment (Adam and Canziani 1997). The age limit,
however, produced negative substitution effects in the hiring of adult low-skilled workers. Substitution effects have
also operated with respect to the younger and less-educated
135
In Italy 37 per cent of part-timers and 49.5 per cent of xed-term workers would prefer to work full time or on permanent contracts. Availability for these jobs is however
increasing: from 31 per cent of labour supply in 1989 to 39 per cent in 1994 (Olini 1997).
Italy 283
apprentice workers, who have declined in numbers in the 1980s, due to overlaps and competition between the two
forms of incentive for young people (Brunetta and Turatto 1988). The lack of clear targeting, besides age, increases the
probability of deadweight effects: recent monitoring of CFL shows a pro-cyclical pattern with sharp declines during
downturns and the greater use of these contracts in the low unemployment regions of the North (Croce 1997). In the
early 1990s, restrictions on the use of the facilities linked with these contracts reduced the number of low-skilled young
people hired on them, numbers dropping 33 per cent between 1987 and 1995, while increasing by 11 per cent the
number of highly educated young people hired on them (ISFOL 1996).
The 1997 Treu Package has devised a new model of work training and apprenticeship contracts and new temporary
contracts to ease the school-to-work transition. Reforms of work training and apprenticeship contracts provide for
longer duration and special training provisions which should increase and improve their training content. The age of
potential users is increased from 29 to 32 years. In the case of apprenticeship contracts it is also possible to reduce
entry wages.
Temporary or agency work was legalized only in 1997, after fteen years of discussion. The regulatory framework
provided by the Treu Package is rather restrictive, its intention being to provide protection for temporary workers.
Interim agencies must be authorized by the Ministry of Labour, they must operate in at least four Italian regions, and
they must deposit a large cautionary sum for their rst two years of activity. Moreover, interim agencies must pay 5 per
cent of the workers' gross wages into a special Vocational Training Fund. The law allows only a limited amount of
temporary employment, which must be agreed upon by collective agreements,136 and forbids its use for unskilled
workers. In addition, agency workers must be guaranteed the minimum contractual wage of their sector of
employment. According to recent estimates (Banca d'Italia 1997), the more protected agency workers could susbstitute
for currently less protected free lance workers and collaborators.
Wage Determination
Italian wage formation has been, until very recently, relatively unresponsive to rising unemployment. Such wage rigidity
was the result of the multi-tiered system of wage bargaining that operated until the 1993 income agreement. Bargaining
used to take place at national, sectoral,
136
A collective agreement in April 1998 xed the maximum incidence of agency work at 8 per cent of a rm's employees.
Fig. 10.1. Inter-sectoral wage dispersion, Italian industry, 19701995
Source: CNEL-IRS (1997a: 67, Graph 1.8).

and rm levels which might overlap and cover the same ground. Wage oors bargained at the national level for each
industry were legally binding on all rms in all regions. In addition, prior to 1992 a wage-indexation mechanism
operated which was based on a at-rate backward looking system that greatly reduced wage differentials and prolonged
over time the effects of inationary shocks. Subsequent agreements at the regional and/or rm level could re-open
bargaining on wages and working conditions to ameliorate any of the aspects considered at the national level. The
results were: a decline in interindustry wage differentials (Fig. 10.1), an inationary bias due to the effect of automatic
wage indexation, and a lack of co-ordination between different bargaining levels which led to a wage-spiral in the 1970s
and 1980s.
The low level of wage dispersion by qualication and skill levels, together with the sectoral and dimensional
specialization of the Italian economy, helps also to explain the low level of investments in human capital in Italy
compared with other OECD countries.
Wage rigidity has been progressively reduced since the 1983 agreement between the social partners on wage
indexation, and especially since the 1992 and 1993 income agreements. Wage bargaining currently occurs at a national
sectoral level and at the rm level. At the national level each sector has a four-year national labour contract which
governs labour relations, and a two-year settlement regulates nominal pay increases in line with forward looking
ination targets set by the government in agreement with the social partners. Firm-level bargaining concerns internal
labour conditions and the component of wage increases linked to the rm's economic situation and productivity gains.
Italy 285
The rigidities imposed by the wage indexation mechanism have been entirely eliminated.
In the early 1990s new collective bargaining rules and a new pay-determination system were introduced in the public
sector as well. Since 1994 negotiations have been conducted by a special independent agency (ARAN) on behalf of the
government, subject to the deadlines imposed by budgetary targets. Recent agreements in the public sector give more
weight to pay awards based upon individual performance at the expense of seniority-based pay increases.
Other important forms of regulated exibility in the late 1990s, which may lead to the introduction of territorial
dispersion in wage oors and working conditions, are area contracts and territorial pacts introduced by the 1996 Pact
for Employment. Local authorities and the social partners in economically depressed areas may agree upon special
working and wage conditions and simplied administrative procedures. According to employers associations'
estimates, the use of area contracts and territorial pacts in Southern Italy could lead to labour cost reductions of
between 25 and 30 per cent.
While progress has been made towards greater real wage exibility, the contributive wedge over labour costs, or
employers and workers social security contributions, is still among the highest in the OECD. It has increased up to
1997, due to the reduction in payroll based tax credits for employers, and the increased social security contributions
paid by employees. This is particularly marked in Southern Italy where it increased 10 per cent between 1992 and
1995.137
The 199798 tax reform and the provisions of the 1998 Social Pact should reduce the tax and contributive wedge on
dependent labour by transferring general social provisions such as health, maternity and family contributions onto
general taxation and by reducing scal pressure. When fully operational the new regional tax on business activity
(IRAP), is expected to reduce average labour costs by about 1 per cent, with labour intensive rms beneting more
from the new system. Special deductions for new hirings in Southern Italy, the hiring of apprentices, and work training
contracts are also envisaged.
Working Time Regulation

Functional exibility, particularly exibility in working time, in part compensates for strict hiring and ring regulations.
By international standards, Italian rms have greater exibility in their use of working time, especially overtime, shift
work and work on Saturdays. Working
137
According to Svimez (1996) this is likely to produce, all other things being equal, a 10 per cent decline in employment in the period 199499.
time is regulated only by collective agreements at the rm level and this has traditionally permitted a high level of
controlled exibility in terms of hours and working time. Overtime is widely resorted to by rms during upturns, and
it is de facto accepted by trade unions. In fact, the cost of overtime is lower than an additional normal hour for the rm,
given the wage base and the conventional working hours used to compute it (Origo 1997).138
Atypical hours, especially shift work and work on Saturdays, are also more widespread than in other EU countries.
Eurostat data show that in 1995 almost 18 per cent of Italian dependent workers were usually employed on shift work,
compared to an EU average of 12 per cent, and 35 per cent were usually employed on Saturday work relative to an EU
average of 23 per cent. Collective agreements at the plant level now increasingly envisage different forms of exibility
including the greater use of xed-term contracts, vertical part-time, or work only on Saturdays and Sundays, and
exible hours.
Working time exibility in Italy has traditionally been supported by two other important instruments used by rms
during downturns: Cassa Integrazione Guadagni Ordinaria (CIGO) and solidarity contracts. Both measures provide for the
partial reduction of working time during downturns, with income support supplements for workers.
Within this framework, the draft legislation presented by the government in 1998 for a reduction of the standard
working week to 35 hours by the year 2001 would introduce additional rigidites and increase hourly labour costs. For
this reason it is considered with scepticism among both employers and unions, which have signed an agreement in
1997 conrming the 40-hour working week, instead of the previous 48, and leaving any decision relative to further
reductions in working time to collective bargaining. The Treu package, on the other hand, provides incentives to
reduce working time through graduated reductions in social security contributions.
Social Security Policies

Income support during temporary or permanent loss of work is another key element in proper understanding of the
underlying exibility/ rigidity of labour markets. As long as income support during unemployment is high, there is less
opposition to dismissals and external exibility may be improved; but on the other hand there is a greater
138
Estimates by Origo (1997) show that the cost of overtime hours in 1995 was only 75 per cent of ordinary working hours in the chemicals sector, 90 per cent in the
manufacturing sector and 91 per cent in the textiles sector.
Italy 287
pressure on wages and higher de facto wage oors (Dell'Aringa and Samek Lodovici 1997).
Italy's social security system is consistent with the general regulatory approach of its labour market: the entire system is
based upon the over protection of insider workers and pensioners. Recent estimates (Rostagno and Utili 1998) show
that over 61 per cent of total social security outlays are accounted for by old age and survivors'pensionssome 20
percentage points above the EU average. Accordingly, spending for other forms of assistance, such as housing,
unemployment, and family benets, is much lower than the EU averageunder 10 per cent of GDP against a EU
average of almost 17 per cent of GDPand contemplates a comparatively narrow range of contingencies. The same
authors show that this system, and its ill-designed targeting mechanism, is extremely inefcient: it has generated severe
gaps in coverage, discouraged mobility and risk-taking in the labour market, and distorted intragenerational equity.
Only less than a quarter of total spending on welfare measures accrues to families below the poverty line.
Attempts to reform the social security system and especially to reform the pension system, due to its nancial
insustainability in the medium run, have been clashing against the interests of the strongest segments of the employed
population. Only partial and gradual reforms of the pension system have been carried out during the Amato (1992),
Dini (1995), and Prodi (1997) governments,139 while reform of social shock absorbers has been on the governments'
agenda since the late 1980s. The only steps taken up to now have been the introduction of a mobility allowance for
workers laid off through collective dismissal in 1991, and the experimentation of minimum guaranteed income
launched by the Ministry of Social affairs in 1998. Reform should tackle a number of shortcomings of the current
system. In particular:
1. The fragmentation of benets available and the differences in the treatment accorded to unemployed workers,
which depend on the characteristics of the job that s/he has lostsector, size of rm, type of employment
relationand on the reason for the job loss, either collective or individual dismissal. Workers from medium or
large industrial and commercial rms, dismissed temporarily or denitively due to restructuring or crisis, and
building sector workers, are those most
139
CIG and mobility allowances for collective dismissals are equivalent to 80 per cent of previous wages and may last up to ve years according to location and age. Initially
these benets were restricted to industrial rms with more than fteen employees. Large commercial rms were only included in 1991. During the 199294 recession,
benets were temporarily extended to workers in some other service sectors and to commercial rms with at least fty employees.
favoured by the Italian system; whereas workers in small rms and in most service rms are only eligible for the much
less generous ordinary unemployment benets.140 Those with no employment experience and the self-employed are not
eligible for any benet at all.
2. The absence of assistance provisions for the long-term unemployed once they have exhausted their
unemployment insurance entitlements, and of special arrangements in terms of family or housing benets.
Disability pensions have been used as assistance benets, thanks to their relatively lax eligibility requirements.
Social benets to the needy, usually the elderly, are provided by local authorities, with the benet level and
duration varying widely across regions.
3. The long duration of of CIGS and mobility benets, and their high replacement rates. The length of the benet
period can be increased by the sequential use of different provisions. Older workers on mobility lists, could,
before 1993, receive benets until early retirement; prime-age workers may combine CIGS and mobility
benets for a maximum of ve years, or six years in Southern Italy.
4. The absence of a link between the payment of benets and re-training or job search requirements and the lack
of incentives to accept regular but poorly paid jobs. Only since 1993 has the government considered the use of
temporary job schemes, or lavori socialmente utili, as a screening device to detect moral hazard. In fact, to date
these schemes have been organized on a voluntary basis, and they are not used to screen effective
unemployment. It is no coincidence that these measures have failed in Northern Italy, where many workers
nd work in the black economy and wait until the subsidy expires before regularizing their position. On the
other hand, in Southern Italy these schemes are used to extend income support for the long-term unemployed,
and they tend to create pressures for permanent jobs in the public sector.
5. The little use of active policies to help the unemployed nd work. Job creation measures and employment
services are conspicuously lacking, owing to the ineffectiveness of public employment services and the legal
prohibition on private agencies, which was only repealed in 1997. The reform of the CIG in 1991 introduced
some measures to help those on CIGS and mobility lists, but they are little used.
140
Ordinary unemployment benets may be obtained by workers dismissed individually or collectively, but not eligible for mobility benets. It is necessary to have at least 2
years of contribution or 78 days for seasonal workers. The benet level is equivalent to 30 per cent of last wage. For a detailed description of the evolution of income
support measures in the 1980s see Franco (1993), Tronti (1993), and de Luca-Bruni (1993). For this reason Italy's position in international comparisons depends largely on
which benet is considered. Average benets indicators are also misleading because of the wide differences in benet levels and duration among different segments of the
labour force.
Italy 289
In 1999, the Ministry of Labour set up a commission of experts to produce a proposal for a general reform of the
system of social shock absorbers. Based upon an earlier proposal (Onofri Commission 1996), current discussion calls
for a simplication of the existing system, reducing the number of benets now available, the extension of insurance
benets to all those who have lost their jobs, and the provision of welfare benets for the long-term unemployed and
low-income groups. Insurance benets should provide both for job losses and reductions in working hours.141
The Actual Coverage of Regulation

One measure of the stringency of labour market regulations is their effective coverage. In Italy, regulation mainly
concerns dependent workers in rms with more than fteen employees. This area of the labour market is relatively
small in relation to the other Western economies. Ichino (1996) estimates that such workers account for only 40 per
cent of total employment, and they are concentrated in the industrial and public sectors. A large share of the Italian
labour market thus falls outside binding regulations and is characterized by exible employment arrangements and
little, if any, employment protection. Self-employment, the small rms sector and non-regular work constitute large
areas of exibility in the labour market with numerous overlaps and no clear-cut boundaries. These forms of
employment have been increasing their share of total employment in the 1990s.
In 1997 self-employed workers accounted for 28 per cent of total employment, relative to the 17 per cent EU average,
and their share has been quite stable since the early 1980s. These workers are mainly own-account workers and
employers. In recent years, occupational positions midway between dependent and independent, or freelance, work
have increased rapidly, especially in northern Italy and in the service sector. This category comprised some 1.3 million
workers in 1997, equivalent to 6.5 per cent of total employment. In Southern Italy self-employment is mainly a way to
absorb excess labour supply and to circumvent scal and employment regulations: it is concentrated in trade services,
which represents 61 per cent of employment in this sector, and in agriculture, and displays a marked cyclical trend (Fig.
10.2). In Northern Italy, instead, it is a result of a tradition of craft work and small family rms, of exemption from
heavy employment regulation, and of the expansion
141
The benet may be used as a re-employment bonus if the unemployed person starts a new job before his/her benet expires. In addition, the benet may be used as a
start-up sum by those opting for self-employment or as a subsidy, reducing labour costs for rms employing workers on mobility. People on mobility benets who accept
part-time or xed-term jobs are only temporarily suspended from the benet.
Fig. 10.2. Employment evolution in Italy by geographic area, 19811997

Notes: Index of Employment levels: employment in 1981 = 100.
Source: Banca d'Italia and ISTAT.
Italy 291
of professional groups in business and personal services. It is concentrated in the industrial and service sectors and
follows a much more stable pattern than in Central and Southern Italy.
8590 per cent of Italian rms have fewer than fteen employees: the size threshold of the rigid regulation of
employment conditions induces small rms to contain their employment expansion, and especially large rms to
decentralize production phases to small rms. Moreover, a large part of the small rm sector has been induced to
merge with the underground economyequivalent to 2026 per cent of GDP according to Istat estimatesin order
to eliminate adjustment costs.
In small rms, labour turnover is rather high. Recent estimates based on administrative datasets (Contini et al. 1994;
Rapiti 1994; OECD 1996) show that job turnover in the private sector is around 23 per cent. Average separation rates
range around 34 per cent, compared to 30 per cent in Germany and 62 per cent in the USA. They are particularly high
in the commercial, construction and retailing sectors, all of which are characterized by the large incidence of small
rms.142
Non-regular work is another way to circumvent strict regulation,143 and it is another buffer against the scarcity of jobs
opportunities in the ofcial labour market, particularly for the weakest segments of the labour force, women and
youths (de Luca and Bruni 1993: 148). 1996 estimates from Istat National Accounts (Table 10.3) measure almost 5
million standard full-time labour units in non-regular work, which is equivalent to 22.3 per cent of the total standard
labour units in Italy. Non-regular work in Italy relates mainly to tax evasion, evasion of social security contributions,
and avoidance of labour legislation relating to minimum wages, health and safety measures, hiring and ring legislation,
working time, working conditions, and so on. Non-regular work is more widespread in Southern Italy, where it
represents 34 per cent of total labour as compared with only 18 per cent in the Centre-North (ISTAT 1997).
In the late 1990s, actions against non-regular work have been reinforced through simplication of labour regulation,
more exible working conditions, tax deductions on certain expensesmainly to prevent tax evasion by
professionalsthe reduction of non-wage labour costs, and the introduction of realignment contracts. Realignement
contracts allow preferential tax treatment and the use of lower labour
142
Administrative data as a base for job turnover estimates are criticized because they probably overestimate rms and job turnover rates. In fact, the turnover rates resulting
from labour force surveys are much lower (Boeri 1997; ISTAT 1997).
143
The underground economy includes: criminal activities; activities undeclared for economic reasons such as tax evasion, and non-compliance with contractual and legislative
norms; activities not surveyed due to non-visibility; and informal activities. Estimates of the black economy usually include the last three types of activity.
Table 10.3. Incidence of Regular and Non-Regular Labour Units on Total Standard Full-Time Labour Units
Labour units 1980 1985 1990 1996
Dependent labour 69.8 68.1 68.7 68.8
Regular 59.2 56.7 56.7 56.9
Non-regular: 10.7 11.4 12 11.9
of which irregular 6.7 6.2 6.5 6.3
of which undeclared em- 1.2 1.3 1.1 0.5
ployed workers
of which non resident 1.2 2.2 2.5 3.1
immigrants
of which second job- 1.5 1.7 1.9 2.1
holders
Independent labour 30.2 31.9 31.3 31.2
Regular 19.7 20.8 20.7 20.8
Non-regular: 10.4 11.1 10.6 10.4
of which undeclared em- 1.4 1 0.7 0.7
ployed workers
of which non resident 0 0 0 0
immigrants
of which second job- 4.9 5.9 6 5.9
holders
Total labour units 100 100 100 100
Regular 78.9 77.5 77.5 77.7
Non-regular: 21.1 22.5 22.5 22.3
of which undeclared em- 2.7 2.3 1.8 1.2
ployed workers
of which non resident 1.2 2.2 2.5 3.1
immigrants
of which second job- 6.4 7.7 7.9 8
holders
Source: Istat (1997).
standards, in terms of wages and working conditions, for a period of some years to rms willing to gradually emerge
into the regular market. These contractual arrangements have been introduced in some sectors, such as textiles and
agriculture, by collective agreements since 1991, and in 1996 a special legislative Act has provided scal deductions for
rms quitting the black economy, while the Treu Package includes specic funds to support emerging rms.
The Outcome: The Structure of Employment and Unemployment
The Effects of Regulation on the Structure of the Italian Labour Market

According to a large part of the Italian literature (Del Boca and Rota 1993; Tronti 1993; Liso 1994; Negri and Saraceno
1995), the rigidity of hiring and ring regulations and constraints on employment contracts,
Italy 293
together with the characteristics of the social protection system and social shock absorbers, have given rise to major
distortions in the Italian labour market.
The most negative effect has been the freezing of regular employment (Tronti 1993). According to Tronti's estimates,
a 1 percentage point increase in the number of hours covered by the CIG resulted, in the 197288 period, in a 0.75 per
cent slowdown in industrial hirings. This worsened the segmentation of the Italian labour market, discouraging labour
mobility for protected workers, and increased the difculties in nding regular long-term jobs for workers with little or
no labour experienceyoung people and women in search of rst employment in particular. The nancing of these
measures, largely out of public funds and without disincentive mechanisms, also favoured collusion among workers
and rms.
The long duration of CIG and mobility benets has increased both workers' dependence on the subsidy and long-term
unemployment. These benets, together with the inefcient system of controls, have also helped to expand the
underground economy. Workers on CIG or on mobility and early retirement schemes have no interest in giving up
their generous benets by nding a regular job. Finally, the greater generosity of mobility benetslinked to collective
dismissalsin relation to unemployment benetslinked to individual dismissalstogether with the lower costs and
risks of collective dismissals, have stimulated rms, in agreement with the trade unions, to use adjustment patterns
based on mass dismissals rather than small, individual adjustments144 (del Boca and Rota, 1993).
Some indicators of the level of sclerosis and segmentation in the Italian labour market in comparison with other
OECD countries in the late 1980s and early 1990s are given in Table 10.4 which presents the monthly ows into and
out of unemployment and long term unemployment based on OECD estimates. Together with Spain, Italy has the
lowest inow rates into unemployment, but also the lowest outow rates from unemployment. The inow into
unemployment is much higher for young people, at 51 per cent of total inows compared with only 11 per cent for
workers aged at least 45, and long-term unemployment is concentrated among the young. 60 per cent of the long-term
unemployed are in the 1524 age bracket. Employment protection and the extensive use of early retirement to support
rms' restructuring help to explain, on the other hand, the low rate of unemployment of people with
144
Individual dismissals are in fact relatively more expensive and riskier than collective dismissals, both for rms which have to pay an exit bonus directly to the worker and risk
being made to reinstate the worker by a court decision, and for workers, who receive a much lower and shorter unemployment benet (Del Boca and Rota 1993).
Table 10.4. Unemployment Flows and Long Term Unemployment, 19791991

Unemployment (% of Long-term unemployment (stock) (% of total unemployment)
source population)
monthly outows 1524 2544 >45 men women
inows
Germany 1979 0.2 n.a.a na na na 50 50
1989 0.3 7.6 12 42 46 51 49
1991 0.2 8.0 8 43 48 51 49
France 1979 0.3 6.1 28 46 26 37 63
1989 0.4 6.7 15 63 22 40 60
1991 0.3 5.5 13 63 23 40 60
Italy 1979 0.3 na na na na 50 50
1989 0.2 2.8 59 37 5 40 60
1991 0.2 3.6 52 43 5 39 61
United 1979 0.6 na na na na 75 25
Kingdom 1989 0.6 13.7 22 39 39 71 29
1991 0.6 13.4 18 43 39 74 26
Spain 1979 0.3 3.7 57 26 17 60 40
1989 0.2 2.2 40 39 22 43 57
1991 0.2 2.0 34 38 28 40 60
Japan 1979 0.3 19.5 9 52 39 70 30
1989 0.4 22.0 11 32 57 71 29
1991 0.3 23.6 13 42 46 75 25
United 1979 2.1 47.4 28 39 33 60 40
States 1989 2.0 48.2 16 55 29 70 30
1991 2.1 37.3 14 53 33 68 32
a
na = not available.
Source: OECD, Employment Outlook 1993: 8889, table 3.3.
previous employment, and the characteristically low level of unemployment among older and low-skilled workers in
Italy.
Most of the unemployed, being secondary workers with family support, that is, children and housewives, prefer to
work in the black economy while waiting for a job in the primary sector, rather than compete through wages.145 Wait
unemployment, wage rigidity with respect to unemployment, and low labour mobility have thus been further
consequences of segmentation (Bertola and Ichino 1995; Casavola et al. 1995; Faini et al. 1996; Meldolesi 1996). Micro-
labour force data show that almost 80 per cent of all the unemployed are living
145
For this same reason in Italy, contrary to other European countries, unemployment is not strongly linked to poverty.
Italy 295
with their families as dependents, while only 20 per cent are heads of families. A large share of the unemployed are not
willing to accept jobs out of their township40 per cent in 1997and reservation wages appear at the level of current
wages. Reservation wages, especially those of unemployed women, appear higher in Southern than in Northern Italy.
This has been explained with a wait unemployment theory, by which Southern unemployed youth, especially the
highly educated, wait for a good job in the public sector or in large private rms, instead of accepting less secure jobs
in the private sector (Bodo and Sestito 1991). Waiting is possible thanks to family support and irregular jobs.
The low territorial mobility of the Italian labour force is thus linked, on the one hand, to discouragement effects, and
on the other to widespread family welfare systems for secondary components of the labour force, and to the presence
of the irregular economy. Low territorial mobility is therefore related to regulatory patterns: employment protection
norms, the practice of national wage determination; and to the high costs of mobility in a situation where geographical
differences in living costs are particularly high, the housing market is not competitive, and transportation is not well
developed.146
Effects of Recent Changes in the Regulatory Framework on Labour Market

Developments
Empirical ndings show that the 1990s reforms have had effects on labour market performance, even if it is too early
for some of the most important changes to have displayed their full effects. Labour market data reveal increases in
external exibility, especially on the exit side, wage exibility, and in the use of part-time and xed-term contracts, with
dependent employment dynamics more responsive to the economic cycle, especially in the manufacturing sector in
Northern Italy.
Changes in regulations and in labour market conditions are also reected in the opinions of employers and workers.
EU employer surveys (EC 1995) report a sharp decline in the proportion of rms indicating restrictions in hiring and
ring as important obstacles against taking on more labour: the percentage fell from 61 per cent in 1989 to 30 and
then 25 per cent respectively in 1994. A survey on workers (International Survey Research 1996), on the other hand,
reveals a dramatic reduction in the employment security perceived by Italian workers. The
146
The role of increased housing and living costs differentials in constraining labour territorial mobility is well documented by Cannari et al. (1991) who estimate that, all other
things being equal, if North-South housing price differentials were the same as in the mid-1970s, the South-North mobility of labour would increase by 17 per cent.
fear of losing one's current job is also the main reason for the increase recorded in the number of employed workers
searching for new jobs in Italy, from 2.9 per cent in 1993 to 6.7 per cent in 1996. Worsening labour market conditions
and mobility incentives explain the different attitudes of Southern unemployed toward geographical mobility and
atypical work contracts: the number of unemployed people in Southern Italy willing to move to other areas in order to
nd a job has been rapidly increasing from 54.5 per cent in 1995 to 60.6 per cent in 1997; the number of those not
willing to accept atypical forms of working contracts has declined from 64.4 to 58.6 per cent.
A more detailed analysis shows that new modes of marginalization are emerging in the Italian labour market which
require a more comprehensive reform, including specic and well targeted labour market and training measures, and a
rapid revision of the social protection system.
The responsivity of employment to economic conditions appears greater during downturns, than during recoveries
(Fig. 10.2).147 The recession of the early 1990s reduced dependent employment by 1,271,000 units over the period April
1992January 1995, while the 199598 recovery has produced an increase of only 216,000 units. Contrary to what
occurred in previous recessions, the increase in separation rates augmented the number of adult unemployed with
previous employment (Fig. 10.3).
The recovery of 199698 produced mainly increases in hours worked by the employed and in part-time and temporary
jobs which increased respectively by 17.3 per cent and 21.2 per cent while permanent full time jobs declined (Fig. 10.4).
The use of atypical working hours also increased in 199396, especially in rms with more than 200 workers and in the
textiles sector (Istat 1997), involving mainly women with low educational qualications.
Panel data relative to the 199596 period show that women, the very young, and older workers were most affected by
these changes (Table 10.5). Women have higher probabilities of losing their jobs than men, and they register fewer
direct passages to another job. A higher percentage of laid-off women leave the labour force. Regarding age, the
probability of losing the previous job displays a U-shaped pattern, with
147
The stability of dependent employment is, however, still high in Italy compared with other countries, even if slightly less so than in the late 1980s. In 1995, the median length
of dependent employment was 8.3 years and the average was 11.1 years. Almost 45 per cent of dependent workers in 1995 had been employed for more than 10 years, and
20 per cent for more than 20 years, while only 8 per cent had been employed for less than one year. The probability of employed workers losing their jobs is 5 percentage
points lower than the EU average (ISTAT 1997), but entry ows into unemployment have increased from 0.14 in 1985 to 0.41 in 1994.
Italy 297
Fig. 10.3. Unemployed with previous jobs and equivalent workers on CIG benets in industrial rms, 19701998
(quarterly moving averages)
Notes: DSS = Unemployed with a previous job; CIGDSS = Equivalent workers in CIG.
Source: own estimates on Banca d'Italia, different years; Istat, Indagine sulle forze di lavoro, different years.
young people and older people presenting the highest probabilities. The incoming restrictions to early retirement
provisions and seniority pensions for older workers, all measures which, up to now, have largely contained
unemployment among older low-skilled workers, is going to rapidly increase their number, making the Italian labour
market more similar to that of other EU countries.
Young people were hit particularly hard by the recession of the early 1990s: it was youth employment that ensured
external exibility in that period, while adult employment remained relatively stable (Croce 1997). Young people
represent, since the mid-1980s, the highest inows into dependent employment, but this happens mainly through
temporary contracts and is related to measures facilitating rms in substituting younger for older workers through
temporary contracts and early retirement. During 1998 the use of temporary contracts to employ young people has
increased, especially in Southern Italy, where most of the temporary employment incentives apply. At the end of 1998
temporary employment represented 18.4 per cent of total dependent employment for people between 1529 years of
age and 14.5 per cent of dependent employment in Southern Italy, relative to a national average of 9.4 per cent. The
recovery in youth employment since 1995 is still too weak to compensate for the decline which occurred in the early
1990s: in 1998 the employment of people aged under 25 years was still 13.8 percentage points below the level of 1993.
The employment cyclicity of young workers appears thus to be largely asymmetric, with large falls in downturns and
little recovery in upturns.
Flexibilization trends do not affect long-term unemployment, which
Fig. 10.4. Part-time and xed-term contracts by geographical areas, July 1993April 1997 (quarterly moving averages).
Notes: July 1993 = 100.
Source: ISTAT.
Italy 299
Table 10.5. Quarterly and Annual Exit Routes from Dependent Employment 19951996 (%)
After 3 months After 1 year
Other em- Unem- Out of la- Total Other em- Unem- Out of la- Total
ployment ployment bour force ployment ployment bour force
Sex
Male 1.7 1.5 1.7 4.9 5.0 2.0 4.9 11.9
Female 1.1 1.4 2.9 5.4 4.7 2.4 5.6 12.7
Age
1524 3.6 4.0 3.4 10.9 9.8 5.9 7.5 23.2
2529 1.9 2.2 1.0 5.1 6.4 2.5 2.2 11.1
3039 1.3 1.6 1.3 4.3 5.2 2.1 2.1 9.4
4049 0.7 0.9 0.9 2.5 2.6 1.5 2.6 6.8
5054 0.1 0.4 2.4 2.9 1.9 2.1 13.4 17.4
55+ 0.7 0.9 7.3 8.9 1.5 1.5 23.5 26.5
Education
a
University 1.0 0.8 1.8 3.7 1.3 2.9 8.0
degree
Secondary 1.5 1.2 1.3 4.0 4.9 1.9 3.2 10.0
school
Compulso- 1.7 2.2 2.1 6.0 5.8 3.2 5.0 14.0
ry school
Without 1.1 2.0 3.2 6.3 3.7 2.7 10.2 16.5
education
Territorial area
North-We- 1.2 0.9 2.0 4.1 4.4 1.7 5.6 11.8
st
North-Ea- 2.0 1.9 1.8 5.7 7.0 2.3 6.0 15.4
st
Centre 1.3 1.1 1.8 4.2 4.3 2.1 5.3 11.7
South 1.6 3.7 2.2 7.5 4.4 4.9 4.5 13.8
Total 1.5 1.8 2.0 5.2 5.0 2.6 5.4 13.0
a
Figures unavailable.
Source: ISTAT (1997).
has increased with respect to total employment from 59 per cent in 1993 to 68 per cent in 1997. In Southern Italy, the
incidence has risen from 68 to 75 per cent. Most of the long-term unemployed are still young people searching for
their rst job. Recent estimates (ISTAT 1997) based on panel data from April 1993April 1996 show that rst-job
seekers have only a 20 per cent chance of nding employment compared to 40 per cent for the unemployed with
previous work experience; and that their unemployment duration is signicantly higher (Table 10.6). Gender and
education are important discriminatory factors in the chances of young people nding permanent jobs: males have a
higher probability of nding permanent jobs than females, and young people with a university degree are twice as likely
to do so compared with young people with only compulsory education. In addition, weaker segments of the labour
supply have higher probabilities of nding temporary or precarious jobs.
As regards wage exibility, changes in the wage formation process contributed to real wage declines in the 199296
period and to the decline of ination (Fabiani et al. 1997), but labour cost dispersion and earning differentials increased
only slightly in the 1990s across sectors and qualications, while they remained stable across rms' size and location
(CNEL-IRS 1997a). The diffusion of territorial pacts and area contracts, together with recent employment subsidies
aimed at supporting the school-to-work transition in Southern Italy, should have larger effects on age-related wage
differentials. According to the Bank of Italy (Bouced' Italie 1998) entry wages declined relative to average earnings in
1997 and non-wage labour costs for newly hired young people are going to be reduced by 30 per cent over a three year
period.
Conclusions
The Italian experience highlights the risks associated with excessive regulation of employment relations: the
development of an extremely segmented market, the exclusion of rst-job seekers from regular employment, the low
elasticity of wages to unemployment, the inefciencies deriving from low inter-regional and inter-sectoral labour
mobility and the presence of a large underground economy.
The strategies adopted to reduce labour market rigidities have been based on the extensive use of collective bargaining
and the involvement of the social partners in the denition of new areas of intervention. While these procedures entail
long adjustment processes, they explain
Italy 301
Table 10.6. Labour Market Conditions for Young First Job-Seekersa After One Year from First Survey 19951996 (%)
Permanent Temporary Job-seeker Potential Out of la- Total by row
employment employment job-seeker bour force
Sex Male 13.9 6.0 56.6 11.8 11.8 100
Female 11.4 8.9 55.8 14.2 9.7 100
Education University 20.1 9.4 52.4 11.6 6.4 100
degree
Secondary 14.2 7.4 54.9 12.6 10.9 100
school
Professional 11.6 9.2 47.6 18.5 13.1 100
qualication
Middle 11.0 6.6 58.3 12.6 11.5 100
school
Primary 6.6 6.6 64.8 13.1 8.9 100
school
Territorial North 26.0 12.4 38.9 10.6 12.1 100
area Centre 18.8 9.7 46.8 14.4 10.2 100
South-West 5.1 3.9 64.9 16.0 10.1 100
South-East 9.5 6.9 58.7 12.1 12.8 100
Islands 6.9 5.0 68.4 9.9 9.8 100
Desired Only full- 13.6 7.5 53.8 14.0 11.2 100
time
worktime Only part- 15.6 13.5 47.0 12.6 11.3 100
time
Any time 9.8 5.6 65.1 9.8 9.6 100
a
young: 1524 years.
Source: ISTAT, 1997.
the relatively low rate of conict in Italy, and have fostered a cooperative attitude in the social partners.
Since the early 1990s, a process of re-regulation in Italy has improved some dimensions of exibility relating to wage
dynamics, working time, ring procedures and atypical work, but much less so in terms of hirings, social security
schemes and non-wage labour costs. The initial effects of this process are visible in the labour market, especially in
Northern Italy and as regards exit exibility. They are particularly detrimental to young people and women, who
appear to have been hardest hit by the recession of the early 1990s, while older and low-skilled workers up to now have
been protected by early retirement provisions. The labour market position of the latter will probably worsen as budget
constraints are likely to impose a suspension of such schemes.
The length of the transition from greater exibility to labour market performance and the asymmetry of the
relationship which is stronger in downturns and much weaker in upturns, is probably due to the slow pace and the
asymmetry of the re-regulation process. The Italian case seems to bear out those who have recently argued that a
piecemeal approach to labour market reform does not help to reduce unemployment, but rather increases the burden
on specic sectors of labour supply (Bertola and Ichino 1995).
Other forces are however inuential in preventing the full effects of labour market re-regulation from making
themselves felt in Italy, and they are linked to the structural weaknesses of the Italian economy: namely under-
investment in both human and physical capital, and the underdevelopment of Southern Italy. The recovery of
investment requires political stability, improved macro-economic conditions and a more efcient public administration.
A wide ranging reform of education and vocational training is needed to improve labour force skills and competencies.
Southern Italy requires adequate development policies, which will be more effective the more closely they are
accompanied by increased exibility. Moreover, if exibilization of the labour market is to be more effective, it must be
matched by greater exibility and enhanced competition in the product and nancial markets, and by a signicative
effort to reduce the extent of the underground economy.
Finally, increased exibility in the labour market should be accompanied by a different approach to labour policies. The
aim should be to support job search and labour mobility, to increase the probabilities of the unemployed and the
discouraged nding employment, and to reduce unemployment duration by means of well targeted active and training
policies. A prerequisite for the improved effectiveness of labour market intervention is the more efcient
administration of labour policies and a transparent structure of intervention.
References 303
Since 1996 important steps in these directions have been taken, but the pace of structural reform need to be more
rapid and comprehensive in order to fully exploit all policy complementarities.
References
Adam, P., and Canziani, P. (1997), I contratti a tempo determinato in Italia e Spagna, Lavoro e relazioni industriali, 2:
952.
Ajello, V., Feliziani, D., and Samek Lodovici, M. (1994), Flessibilit ed efcienza del mercato del lavoro, Quale Impresa
Documenti, 13.
Banca d'Italia (1997), Bollettino Economico. 29 Rome: Banca d'Italia.
(1998), Relazione annuale, 1997. Rome: Banca d'Italia.
Bertola, G., and Ichino, A. (1995), Crossing the river: a comparative perspective on Italian employment dynamics,
Economic Policy, 21.
Bodo, G., and Sectito, P. (1991), Levie dello Suiluppo. Bologna: Il Mulino.
Boeri, T. (1997), Le istituzioni del mercato del lavoro: un confronto internazionale, in S. de Nardis and G. Galli (eds.),
La disoccupazione italiana. Bologne: Il Mulino, 91128.
Boragine, I., and Fabi, M. (1997), Rigidit contrattuali e collocamento, in S. de Nardis and G. Galli (eds.), La
disoccupazione italiana. Bologna: Il Mulino, 30323.
Brunetta, R., and Turatto, R. (1988), Gli effetti delle politiche del lavoro. Prime veriche, in Ministero del Lavoro e
della Previdenza Sociale, Rapporto 1988. Lavoro e Politiche dell'Occupazione in Italia. Rome: Poligraco dello Stato.
Bulgarelli, A., and Giovine, M. (eds.) (1988), Il lavoro e i giovani. Prima indagine nazionale sui contratti di formazione e lavoro.
Milan: F. Angeli.
Cannari, L., Nucci, F., and Sestitlo, P. (1997), Mobilit Territopia le ecortodille abitotioui: un aualisi eu.
Casavola, P., Gavosto, A., and Sestito, P. (1995), Salari e mercato del lavoro locale, Lavoro e Relazioni Industriali, 4:
3983.
Centro Studi Investimenti Sociali (CENSIS) (1996), 30mo Rapporto sulla situazione sociale del Paese. Milan: F. Angeli.
CLES (1995), Analysis of factors favouring a more effective employment policy in Italy. (mimeo) Rome: CLES, June 1995.
CNEL-IRS (1997a), Rapporto sulle retribuzioni e sul costo del lavoro. Milan: Giuffr Editore.
CNEL (1997b), Rapporto sul mercato del lavoro. (mimeo) Rome: CNEL, April 1997.
Coe, D., and Snower, D. (1997), Policy Complementarities: The Case for Fundamental Labour Market Reform, IMF
Staff Papers, 44/1: 135.
Contini, B., and Revelli, R. (1992), Imprese, occupazione e retribuzioni al microscopio. Bologne: Il Mulino.
304 References
and Rapiti, F. (1994), Young in, Old out: nuovi pattern di mobilit nella economia italiana, Lavoro e relazioni
industriali, 3: 5176.
Christodes (1996), Italian unemployment 197595: an analysis of macroeconomic shocks and policies using evidence
from a structural vector autoregression, IMF Working Papers, 36.
Croce, G. (1997), Flessibilit e occupazione giovanile nella prima met degli anni novanta, in L. Frey (ed.), Le
informazioni sul lavoro in Italia: signicato e limiti delle informazioni provenienti da indagini sulle famiglie. Quaderni di
Economia del Lavoro, 59: 77123.
Del Boca, A., and Rota, P. (1993), Hiring and Firing Rules in Italy: What Practices do Companies Follow? Discussion Paper,
University of Brescia, Department of Economic Science.
Dell'Aringa, C. (1995), L'accordo di Luglio del 1993 e gli effetti sul costo del lavoro: le prime veriche, Rivista
Internazionale di Scienze Sociali, 3: 50124.
and Samek Lodovici, M. (1997), Policies for the unemployed and social shock absorbers: the Italian experience,
South European Society & Politics, Special issue on Southern European Welfare States, 1/3: 17297.
de Luca, L., and Bruni, M. (1993), Unemployment and labour market exibility: Italy. Geneva: International Labour Ofce.
Demekas, D. G. (1995), Labour Market Institutions and Flexibility in Italy, Labour, 9/1: 343.
De Nardis, S., and Galli, G. (eds.) (1997), La disoccupazione italiana. Bologne: Il Mulino.
Di Nicola, P. (1995), Quale essibilit? Lavoro atipico e part-time in Italia e in Europa, Economia & Lavoro, 3: 2138.
EC (European Commission) (1995), Performance of the European Union Labour Market. Brussels: EC.
Fabiani, S., Locarno, A., Oneto, G. P., and Sestito, P. (1997), Nairu: Incomes Policy and Ination, OECD Working
Papers, 187.
Faini, R., Galli, G., and Rossi, F. (1996), Mobilit e disoccupazione in Italia: un'analisi dell'offerta di lavoro, CSC
Working Paper, 6.
Fondazione Giacomo Brodolini (FGB) (1996), Labour Market Studies. Italy. Luxemburg: FGB.
Franco, D. (1994), L'indennit di disoccupazione, Lavoro e Relazioni Industriali, 2: 191232.
and Sestito, P. (1995), Il sistema di protezione sociale dei disoccupati: alcune riessioni su una possibile riforma,
Politica Economica, 11/3: 389412.
Ganoulis, I. (1997), Flussi del mercato del lavoro e vincoli ai licenziamenti, in S. de Nardis and G. Galli (eds.), La
disoccupazione italiana. Bologne: Il Mulino, 275301.
Gavosto, A., and Sestito, P. (1994), Costi di aggiustamento e essibilit dell'occupazione: l'eterogeneit tra piccole e
grandi imprese, Lavoro e Relazioni Industriali, 2: 71106.
Giavazzi, F., and Spaventa, L. (1989), Italia: gli effetti reali dell'inazione e della disinazione, Rivista di Politica
Economica, JulyAugust.
Grubb, D., and Wells, W. (1993), Employment regulation and patterns of work in EC Countries, OECD Economic
Studies, 21: 758.
References 305
Ichino, P. (1996), Il lavoro e il mercato. Milan: Mondadori.

ISFOL (1996), Rapporto 1996. Milan: F. Angeli.
(1998), Il lavoro in Italia: proli, percorsi, politiche. Milan: F. Angeli (mimeo).
ISTAT (Instituto Statistico Italiano) (1997), Rapporto Annuale. La situazione del Paese nel 1996, Rome: ISTAT.
Lazear, E. P. (1990), Job security provisions and employment, Quarterly Journal of Economics, 105: 699726.
Liso, F. (1994), L'indennit di disoccupazione, Lavoro e Relazioni Industriali, 2: 191232.
Lunghini, G. (1994), Tre alternative sulla disoccupazione, Economia e Politica Industriale, 83: 619.
Mariani, P., and Zeli, A. (1995), Un'esplorazione sulle relazioni tra la CIG, il ciclo economico industriale italiano e le
sue componenti, Economia & Lavoro, 3: 6984.
Meldolesi, L. (1996), L'elevata mobilit del lavoro nel Mezzogiorno della speranza, in G. Galli (ed.), La mobilit della
societ italiana. Rome: Sipi, 35999.
Ministero del lavoro e della previdenza sociale (1994), Obiettivo occupazione: una strategia di medio periodo per il mercato del
lavoro italiano. (mimeo) Rome: MLPS.
Negri, N., and Saraceno, C. (1995), Le politiche contro la povert in Italia. Bologne: Il Mulino.
OECD (1994), The OECD Jobs Study. Paris: OECD.
(1996) Employment outlook, Paris: OECD.
(1997), OECD Economic Surveys 19961997. Italy. Paris: OECD.
(1999), OECD Economic Surveys 19981999. Italy. Paris: OECD.
Olini, G. (1997), Pi salario o meno orario? Le preferenze dei lavoratori, in CNEL-IRS, Rapporto sulle retribuzioni e sul
costo del lavoro, Milan: Giuffr Editore, 20311.
Origo, F. (1997), La convenienza relativa del ricorso al lavoro straordinario, CNEL-IRS, Rapporto sulle retribuzioni e sul
costo del lavoro, Milan: Giuffr Editore, 23745.
Rapiti, F. M. (1994), Ma di quale essibilit ha bisogno l'economia italiana? Alcune note da un'indagine empirica,
Prospettiva Sindacale, 86: 6790.
Rossi, F. (1997), Disoccupazione, offerta di lavoro e lavoro sommerso, in S. de Nardis and G. Galli (eds.), La
disoccupazione italiana. Bologne: Il Mulino, 3189.
Rostagno, M. V., and Utili, F. (1998), The Italian Social Protection System: The Poverty of Welfare, IMF Working
Paper, 74.
Salvati, M. (1988), Mercato del lavoro e relazioni industriali. Note in margine al dibattito su essibilit ed
occcupazione, Politica Economica, 1: 360.
SVIMEZ (1996), Rapporto 1996 sull'economia del Mezzogiorno. Bologne: Il Mulino.
306 References
Treu, T. (1993), Employment protection and labour relations in Italy, in C. F. Buechtemann (ed.) Employment Security
and Labour Market Behaviour, Ithaca, NY: ILR Press.
Tronti, L. (1993), Employment Security and Labour Market Segmentation: Economic Implications of the Italian Cassa
Integrazione Guadagni, in C. F. Buchtemann (ed.), Employment Security and Labour Market Behaviour. Ithaca, NY:
ILR Press.
11 The Spanish Experiment: Pros and Cons of
Flexibility at the Margin
Luis Toharia and Miguel A. Malo
Introduction
Spain is usually considered to be an economy with a highly rigid labour market. Most aggregate indices of exibility
tend to place it at the bottom of the list together with Italy, and until 1984 this was also the opinion of both experts and
employers in Spain. The country's unemployment situation, with a rate of over 20 per cent of the labour force, only
compounded the belief. What may be called the Spanish experiment started in 1984, when the government
introduced what was perceived as a signicant change in labour market operations by allowing new entries with more
exible, less costly contracts, although it left the situation of those already in employment untouched. The situation was
judged adequate until 1994, when a new wave of reforms was introduced in order to put some limits on the use of
temporary contracts and to change dismissals procedures. Finally, in 1997, a further set of changescreating new
incentives for permanent contractshave been made to the Spanish labour market. Evaluating this experiment of
exibility at the margin in terms of its inuence on the level and structure of employment and unemployment is the
purpose of this chapter.
A Review of the National Debate and Literature

Figure 11.1 plots the evolution of employment, unemployment and the unemployment rate in Spain over the period
19701996. Several features of this gure are worth stressing:
1. The rise in unemployment in the 197585 period was due to the strong employment adjustment undergone by
the economy. The
308 Luis Toharia and Miguel A. Malo
Fig. 11.1. Labour force, employment, and unemployment, Spain, 19701996

Source: Labour Force Survey, homogeneous series.
The Spanish Experiment 309
behaviour of the labour force was relatively moderate, especially prior to 1981.
2. During the late 1980s, the Spanish economy was able to create jobs at a rate unknown in its previous history.
However, due to the arrival into the labour market of the more numerous and healthierhence with lower
infant mortality ratescohorts born in the 1960s, as well as the rapid increase in female participation in the
labour market, the labour force grew much faster in the second half of the decade. The result was that the
unemployment rate only declined to 16 per cent by 1991, despite intense job creation.
3. The recession of the 1990s meant a signicant reduction in employment, which, together with the continuing
growth of the labour force, resulted in a sharp rise in unemployment, which stood close to 25 per cent by 1994.
4. The recent recovery has brought moderate advances in employment and a small reduction in the
unemployment rate, which, at the end of 1997, is slightly above 20 per cent.
The evolution just described has prompted debate on the exibility, or otherwise, of the Spanish labour market. The
1984 changes can be interpreted as a response by the recently elected Socialist government to the pressures applied by
employers, who would in fact have preferred a more sweeping reform.148 The exibility at the margin thus introduced
was seen as a positive development, and it was grudgingly accepted by the trade unions, although at the beginning the
Unin General de Trabajadores (UGT) which was very close to the Socialist Party gave some support to this reform.
The employment expansion that accompanied the post-1985 recovery of the economy pushed the notion of exibility
into the background. There was almost no debate on the issue, and whatever discussion that did take place can be
considered as a residue from the past.149
The only instance of debate occurred in 1988, when the Socialist party tried to launch a youth employment
programme designed to introduce cheaper contracts for people under 25 without past job experience. This
programme triggered a general strike in December 1988 which expressed the trade unions' frustration with a situation
in which an economic boom was attended only by slow increases in wages and social expenditures. The strike was an
enormous success, and partly as a result, the stillborn youth employment programme was forgotten. At the same
148
This was also the position taken up by Jos Luis Malo de Molina (1985), an expert at the Bank of Spain, who thought it a mild and inadequate solution. The Bank of Spain
has generally taken a tough line in favour of labour market deregulation.
149
The volume edited by Malo de Molina (1988) contains various views for and against labour market exibility.
time, the government selectively expanded unemployment benets as part of an agreement with the unions to pay the
social debt that had accumulated during the boom years.
It was not until the early 1990s that the debate on exibility was resumed. However, the central issue now was not so
much the lack of exibility as its possibly excessive use at the margin introduced in 1984. In 1990 the government
commissioned a study by independent experts on reform of xed-term contracts (Segura et al. 1991). Although the
proposals contained in the report were not adopted, the idea that xed-term employment was creating distortions in
the labour market, mostly in terms of growing labour market segmentation, gained ground among experts.150 Around
this time the temporary worker rate became stationary at one-third of salaried workers.
In early 1992, the government introduced a decree-law which reformed the unemployment protection system as well
as various aspects of the labour-contract system. However, the debate over the problems of the labour market
persisted, now concentrating on the notion of labour market reform as opposed to exibilization or deregulation.
The debate centred not only on unemployment compensation but also on the cost of ring permanent employees,
considered to be the cornerstone of labour market dualism. There were also those who argued that the Spanish
collective bargaining system was not working efciently because it stood midway between the two supposedly best
worlds of complete centralization and complete decentralization.
In this situation, in early 1993 the government presented a green paper on labour market reform which was to be
discussed with the social partners at the Economic and Social Council, a corporatist institution in which representatives
of employers and trade unions sit together with experts appointed by the government and representatives of other
social organisations such as farmers and housewives. However, political stalemate led to the calling of general elections,
and this aborted the debate.
Although re-elected in June 1993, but without an absolute majoritywhich obliged it to seek parliamentary support
from the Catalonian Nationalists, more liberal in their economic beliefsthe Socialist government did not revive its
earlier plans for reform. Rather, under pressure from a political climate which considered labour market reforms to be
essential, but which also emphasized that they should be the result of a tripartite social pact, it began negotiations with
the social partners.
150
See for example, Jimeno and Toharia (1993a, 1994), both originally written in 1991; and Bentolila and Dolado (1994), originally written in 1992.
However these negotiations failed, and the government decided to press ahead with reform on its own, approving a
bill and a decree-law in early December 1993 which were validated by Parliament in May 1994, at which time the
reform achieved full legal strength.151
The thrust of the 1994 reform, which left only a few aspects of the labour code untouched, was to shift the balance of
power in the workplace towards employers. And so it was perceived by the trade unions, who called a new general
strike in January 1994 against the reforms, although the unions clearly enjoyed less popular support than had been the
case in 1988. However, to many commentators, both political and academic, the reforms were actually much milder
than the above general statement might lead one to believe.152
Not surprisingly, therefore, when the right-wing Popular party won the general elections in 1996, labour market reform
again returned to the political agenda.153 Once again the government felt it necessary to reach agreements with and
between the social partners to implement such a reform. In 1996 and 1997, a series of agreements have been signed on
various aspects of industrial relations. The most signicant of them were reached between the social partners in April
1997, and were translated into a decree-law by the government in May. The thrust of these agreements is twofold: on
the one hand, to encourage the use of permanent contracts; on the other, to enhance the role of collective bargaining.
In this section we have provided a cursory description of the evolution of the debate on labour market exibility and
reform in Spain since the early 1970s. Debate in the late 1990s still centres on the need for further reform.154 However,
it does not seem that the government is planning any further changes in current legislation, at least not until the results
of the 1997 decrees have been thoroughly evaluated.
151
1993 saw several publications specially devoted to reform of the labour market. See, for example, the special issue of Economistas (1993), the review of the professional
association of economists in the Madrid region. The special issue of the bulletin of the employers think-tank, the Crculo de Empresarios (1993), is a balanced collection of
papers on labour market reform by authors representing the views both of the social partners and the academic world.
152
The 1994 reform generated an extensive literature in the labour law eld, because it changed the wording of most of the articles in the LET (1980 Workers' Statute). From
an economic perspective, however, analyses have been much more limited. See, for example, Malo and Toharia (1997).
153
As an indication, see the special issue of Revista Econmica de Catalunya (1997), the review of the professional assocation of economists in Catalonia, which contains a set of
articles on labour market reform, written in early 1997.
154
For a review of various positions on the situation of the labour market in the face of the integration into the Economic and Monetary Union, see Papeles de Economa Espaola
(1997).
The Dynamics of Rigidity/exibility

This section explores in detail the changes that have taken place in the 1980s and 1990s in the rigidity/exibility of the
labour market. It concentrates on four aspects in particular: rings and hirings, wage determination, working time, and
unemployment compensation. To provide a general background, Table 11A in the Appendix presents a synopsis of the
various institutional changes introduced in these areas since the early 1970s.
Firings and Hirings

Spain's fundamental labour law, the 1980 Workers' Statute (Ley del Estatuto de los Trabajadores, henceforth LET)
established three types of dismissals: disciplinary dismissals, objective dismissals, and economic dismissals. The
innovative feature of the LET was that it sought to facilitate duly justied dismissals by enhancing and clarifying the
notion of objective dismissals which was introduced in 1977. Objective dismissal meant justied rings for lack of
adjustment to the job, recurrent justied absence from work, or technological changes; but not for economic reasons,
except in the case of rms with fewer than fty employees. These types of dismissals too had to be justied but, unlike
disciplinary dismissals, they carried much lower severance pay: 20 days' wages per year of seniority, with a maximum of
12 months' wages; as opposed to 45 days' wages per year of seniority with a maximum of 42 month's wages in the case
of unfair disciplinary dismissals, these amounts now being clearly specied in the law. However, these objective
dismissals could be challenged by the workers affected, and this made employers reluctant to resort to them, for if the
dismissal was declared unfair, the severance pay was the same as that paid in the case of disciplinary dismissals (Malo
1998).
As for economic, usually collective, dismissals, the LET provided for a procedure which required negotiations between
the employer and the workers' representatives, and an agreement between the parties was a guarantee, although not a
requirement, of the procedure's acceptance by the administrative authority. For this case the LET established severance
pay equal to that provided for objective dismissals, although it is likely that in many instances the actual amounts
agreed, openly or otherwise, were signicantly higher. The very few data available conrm this belief (Malo and
Toharia 1999).
A signicant change to the law was introduced in 1984, which, although it did not alter the procedures for permanent
workers, did have
an impact for employers taken as a whole: the LET was reformed to make it signicantly easier to hire non-permanent
workers, even for the normal activities of companies. Although temporary contracts had always been allowed in Spain,
their almost unconstrained use was truly innovative. In addition to various specic contracts such as training and
practice contracts for young people, and xed-term contracts for launching new activities, the key component of the
reform was the so-called employment promotion xed-term contract. This contract was to last for a minimum of 6
months and a maximum of 3 years and could be renewed by 6-month periods up to the maximum. Upon expiry, the
rm had to disburse severance pay of 12 days' wages per year of seniority, but the worker could not sue the employer.
Once the maximum period had been reached, the worker had to be given a permanent contract, or the rm could not
hire another worker for the same job. It is obvious how difcult is was to monitor fullment of this requirement, and
how easy it was to redene the job. An additional feature of these contracts was that they entitled their holders to
unemployment compensation at a rate of 3 months of benet for each 6 months of contract. This was a deep change
in the Spanish Labour Law, because it allowed employers to break the so-called causality principle: that temporary
contracts could only be used for temporary activities of the rm. Therefore, a rm could perform permanent activities
hiring succesive temporary workers.
The second half of the 1980s may be termed, from the labour market exibility viewpoint, the era of temporary
contracts. Spanish societythat is, not only employers and employees, but also society at largecame to realize that
traditional permanent labour contracts were less important but also that, although xed-term contracts provided
employment, they were much more precarious, and could not guarantee a minimum, dependable standard of living.
More importantly, the appearance of temporary contracts on a sizeable scale did not prevent rms from ring
permanent workers: the number of rings remained more or less constant over this period (Malo and Toharia 1994).
This was probably due to the fact that these dismissals involved older workers and rms, many of whom needed to be
replaced, for productivity reasons, by younger and newer ones. For example, in the winter of 199192, when the
economy experienced its rst signicant employment decline after the buoyant late 1980swhen employment grew at
rates above 4 per cent per annumtwo-thirds of employment losses corresponded to permanent employees and only
one-third to xed-term workers (Toharia 1996). What happened was that, in relative terms, the weight of dismissals in
total separations became negligible: as measured by the entries into the unemployment benets system, the proportion
corresponding to permanent employees dropped to around 10 per cent by the end of the 1980s (Jimeno and Toharia
1993b).
The situation remained unchanged until 1992, when, led by the reform of the unemployment compensation system,
which raised the requirement to gain access to the insurance-based system to 12 months of contributions, the
employment-promotion contracts were also reformed. Their minimum length was to be 12 months, which was also
the renewal period, and the maximum remained at 3 years. At the same time, some tax exemptions were removed in
the case of practice and training contracts. These may be considered relatively minor changes, although their potential
impact in hindering hirings should not be overlooked.
The 1994 reform was much more profound. The employment promotion contracts were virtually eliminated,
persisting only as a remnant to be used for workers older than 45 or the long-term unemployed. It was an attempt to
restore the causality principle in the Spanish Labour Law. In addition, and presumably as a quid pro quo, the procedures
for ring were restructured in an attempt to reduce costs. A clearer distinction between individual and collective
dismissals was introduced, so that the law envisaged what was already happening in practice, but with quantitative
criteria xed so that the two types could be distinguished. Economic reasons were more clearly included among the
reasons justifying objective individual dismissals for all rms, independently of their size. The procedures for
individual dismissals were altered at various points considered to be crucial, in an effort to reduce the uncertainty
surrounding them. As for collective dismissals, an attempt was made to facilitate them in principle by reducing the time
allowed for the labour market authority to respond the proposal. However, other requirements more cumbersome for
rms, such as the presentation of a social plan for rms larger than fty employees, were also introduced. Finally, the
need to notify the authority that a collective dismissal procedure was under way remained.155
The most recent change in the regulation of ring costs and hirings has resulted from the 1997 agreement between the
social partners. Given the persistent duality between temporary and permanent workers, despite the 1994 reforms, the
means for agreement was once again a trade-off between more exibility for permanent workers and less exibility for
temporary ones. The result has been the creation of a new type of permanent contract, with lower severance payment
in cases of unfair dismissal33 days' wages per year of seniority with a maximum of 24 months' wagesrestricted to
workers under 30 or over
155
In the late 1990s, Spain remains the only European country with such a requirement for collective dismissals. France removed it in 1986.
45, the long-term unemployed, or the disabled. The trade-off also resulted in the imposition of stricter conditions on
the use of temporary contracts, together with elimination of the employment-promotion xed-term contracts. To this
arrangement, the government has added tax incentives in the form of a reduction of employers' social security
contributions over the rst two years of the contract: the reduction is set at 40 per cent of the current rate of 24 per
cent of the wage, except for employees aged over 45 and the disabled, for whom the reduction is 60 per cent.
On the whole, the Spanish pattern as regards the exibility of hirings and rings has been a search for more exibility
while trying to protect the positions of incumbent workers in the labour market. Whereas in the early 1980s workforce
adjustment for economic reasons could be considered rigidalthough under the law such adjustment required the
intervention of the administrative authorities, in practice rms used disciplinary dismissals in lieu of economic
onesthe situation changed after 1984, when the entry of massive numbers of xed-term workers enabled rms to
use them when economically necessary. The very fact, however, that dismissals of permanent workers remained at the
same level as they had been previously can be interpreted as indicating than this was not a sufcient exibility measure
for rms. On the other hand, it might imply that dismissals costs were not a signicant factor in hindering hirings;
rather, rms still needed to adjust their workforces by getting rid of less productive workers. If the latter interpretation
is correct, a reduction of ring costs in 1984 might have produced a socially undesirable faster adjustment of
permanent workers without guaranteeing their replacement by younger ones.
Wage Determination
The basic regulatory elements of the collective bargaining system were established in the LET, and they have remained
more or less constant since then. Trade unions were declared legal in 1977, and employers' organizations were created
at around the same time. Although the social partners were not signatories to the so-called Moncloa Pacts reached at
the end of 1977 as a concerted effort to curb ination and labour unrest, they clearly endorsed its contents. Indeed, it
can be argued that these Pacts started a new period in Spanish industrial relations marked by the existence of
economy-wide agreements, sometimes signed by different groups of agentsfor example one trade union
confederation would sign with the Confederacin Espaola de Organizaciones Empresariales (CEOE) with or without
the government, on other occasions all social partners and the government would sign. These
agreements established, amongst other things but most importantly, a wage range with which collective agreements
had to comply. Indeed, most agreements, no matter who the signatory parties had been, were widely followed at
individual bargaining tables. With the sole exception of 1984, agreements were signed between 1980 and 1986. No
other such nation-wide agreements have been reached since then, not even in 1997, when the Moncloa Pacts were
signed. Whatever their other merits or faults, these agreements were clearly instrumental in the disinationary process
that marked the Spanish economy in the rst half of the 1980s.
The interesting point is that these agreements, like the rest of the collective bargaining structure, were not regulated by
the LET. Instead, the law was very exible, leaving the bargaining parties to decide which precise structure to follow,
that is, to decide what were to be the precise boundaries, or coverage, of the agreements to be signed. These could be
of two broad types: rm-level and above, that is, sector level, national level, etc. Once signed, these agreements were
legally binding on the signatory parties and their represented members, although agreements could regulate the
conditions under which rms under economic strain need not follow its clauses. The law established minimum
requirements which the bargaining parties had to meet in order to gain the right to bargain. Regarding workers, the
representativeness criteria were to be based on union elections, in which all workers were entitled to vote, and which
were to be held every four years. This was important both at the rm levelwhere works councils consisting of
elected workers representatives could be set up, the number of members depending on the size of the rm, and with
representatives entitled to exemption from work in order to carry out their union dutiesbut also at the sectoral level,
where the right to sit at a certain bargaining table depended on the number of votes received.
In practice, the structure that emerged tended to mirror the structure of collective agreements and administrative
regulations that existed during the dictatorship. True, these were not really collective agreements in the democratic
sense of the expression, but they had existed since the 1960s, after the enactment of a law on surrogate collective
bargainingsince free unionism was not allowedin 1958. In addition, for sectors not covered by collective
bargaining, administrative regulations had been established, some of them dated from the 1940s. Although the LET
stated that these regulations were to be replaced by collective agreements or else abolished, some of them in fact
survived until 1994, when the government nally repealed them. Nevertheless, as the incorporation process to the
collective bargaining of the issues and sectors covered by administrative regulation has been slow and incomplete, the
social partners have enacted an agreement to cover in a systematic way these holes, avoiding any possibility of
deregulated sectors.
Thus the structure of collective bargaining tended to be dominated by provincial sector agreements which were the de
facto determinants of actual minimum wages, well above the statutory minimum wage, the relevance of which is more
closely linked to other non-wage matters, such as social security contributions for both employees and the self-
employed. This predominance is the result of the great fragmentation of collective bargaining: there are about 5,000
bargaining units. The provincial sector agreements are the most important in terms of the number of workers
involved, although rm-level agreements, covering mostly large rms, were also signicant. The law did not establish a
hierarchy between the different types of agreement, especially in terms of the wages agreed and the matters to be
negotiated at each level, but it was generally accepted that the sector agreements established minimum conditions to be
improved at the rm level; no detailed studies exist on this point, however. This lack of hierarchy has been sometimes
pointed to as one of the disorderly elements in the Spanish system (Lorences and Felgueroso 1996). A negative effect
of this disorder is the succesive bargaining of the same topics, mainly wages increases, at different levels of the
collective bargaining affecting the same rms.
In the late 1990s, there has been some effort to give more coordination to the collective bargaining structure.
Employees organizations and the main trade unions signed the Acuerdo Interconfederal sobre Negociacin Colectiva or
General Agreement about Collective Bargaining in 1997. The main objective is to give advice and orientation to the
bargaining parties at local and sectoral levels. The framework of the proposed structure is the national collective
agreement promoting the specialization of contents by bargaining levels. The issues reserved for national sectoral
collective bargaining are: wage structure, not wage increase!; maximum working time and criteria for its distribution;
improving safety and health requirements provided by the law; and adaptation of union rights in the rm. Sectoral level
agreements, independent of their geographical scope, should be specialized in changes in occupations and vocational
training. Implicitly, the wage increases seem to be attributed to the lower level: province-sectoral, rm- or plant-level
agreements. The time elapsed since this general agreement is still too short to allow a full appreciation of the real
importance and the actual implementation of this attempt to introduce a more rational structure in collective
bargaining.
What, therefore, can one conclude, in terms of labour market exibility, regarding the Spanish system of wage
determination? There are two aspects to consider: the responsiveness of wages to the
unemployment situation, and the existence of sufcient wage dispersion corresponding to the different situations of
rms.
As regards wage responsiveness to unemployment, it could be argued that, given the consistently high levels of
unemployment in Spain since the early 1980s, the very fact that wages have undergone positive growth is prima facie
evidence of a lack thereof: real wages were 15 per cent higher in 1995 than in 1980. However, this simple argument is
misleading when one considers the complexities of the political economy. In the early transition period, 197577, the
free trade unions were still illegal and their demands were clearly out of order. There is a general agreement (Dolado
and Jimeno 1997; Garca Serrano and Toharia 1994; Zabalza 1996) that a real wage gap developed at that time, and
some commentators have argued that it tended to become permanent. The following period was one of disination,
and only in one year, in the wake of the failed coup of 1981, did the unions accept negative real wage growth in
exchange for the promise of job creation. The expansion of the second half of the 1980s induced the unions to adopt a
more aggressive stance, and some authors (most notably Bentolila and Dolado 1994) argue that this was reinforced by
the duality between permanent and temporary workers. Wages only receded in 199495, partly as a result of the efforts
to prevent the application of the 1994 reform, but also as a response to the employment crisis of 1993. It could thus be
argued that aggregate wages, though not fully responsive to the vagaries of the business cycle, have been more sensitive
to the evolution of employment than to the level of unemployment. This should be related to the peculiar structure of
Spanish unemployment, the weight of which is spread across families, thereby greatly reducing its social impact. This is
discussed in The structure of Employment and Unemployment below.
Until 1986, when nation-wide agreements tended to determine wage growth, it was claimed that wage differentials
entailed signicant rigidity because they prevented rms from bargaining on the basis of their specic conditions. It
can thus be argued that, up until 1986, the Spanish collective bargaining system was rather centralized. Since then the
structure has grown much more mixed; indeed, collectively bargained wage dispersion has tended to increase. Jimeno
(1992) is one of the few to have studied the efciency of the Spanish collective bargaining system. His conclusions tend
to support the idea that Spain, especially since 1987, occupies the middle ground between centralization and
decentralization, the worst situation for labour market performance. Overall, collective bargaining in Spain has tended
to favour uniform wage increases throughout the economy, thus preventing wage dispersion from increasing.
Although no reliable information exists on this matter,
the general impression is that wage differentials, for example by sector or occupation, tended to compress in the early
years of the transitional period towards democracy. They then stabilized as a result of the across-the-board application
of the nation-wide agreements, but more recently have tended to widen again as the lack of such agreements has given
a certain leeway to higher wage dispersion.
Working Time
The most signicant changes in working time regulation in Spain were introduced as part of the 1994 reform.
Previously, the only measures adopted were the establishment, in 1983, of a maximum working week of 40
hourswhich affected very few workers because most of them were already working fewer hoursand the regulation
of part-time work in 1984, establishing a situation where part-time workersdened as those working less than two-
thirds of the normal working week in their sector of activityenjoyed proportionally the same rights as their full-time
counterparts. The LET regulated other aspects such as overtime, which was to be limited to 80 hours per year and paid
at a 50 per cent premium. It was also possible for employers to resort to temporary lay-offs or to short-time working,
in either case drawing funds from their accrued unemployment benets.
The 1994 reform maintained the 40-hour limit but introduced three important changes:156
1. The working time limit was to be interpreted on an average yearly basis, with a daily limit of 9 hours unless
otherwise agreed in collective bargaining or otherwise. Whilst this was already agreed by collective bargaining,
provided the daily limits were respected (UGT 1994), it now became a legal provision and the daily limit could
be set by agreement.
2. The period of rest between two consecutive working days was still xed at 12 hours; however, the reform
allowed agreements to be reached on irregular working schedules.
3. The limit of 80 hours overtime per year was maintained, with one important qualication: those hours
compensated with free time in the following 4 months would not be counted within that limit. At the same
time, overtime hours were no longer to be paid at the obligatory 50 per cent premium; overtime pay was now
to be determined by collective bargaining. The provisions regulating temporary lay-offs and short-time
working remained unchanged.
156
Interestingly enough, these changes were in line with the proposals of two of the leading Spanish Labour Law experts, who helped to draft the LET and whose positions are
usually thought to be close to those of the employers' confederations (Sagardoy and Rivero 1993).
The 1997 agreements do not include any provisions relating to working time, the only exception being the declaration
that the parties concerned urge examination of every aspect relating to working time management, the duration and
the redistribution of the working day, in order to enhance the performance of rms. Although there is some talk of
work-sharing and reducing the working week, this is not generally thought to be a solution to Spain's unemployment
problem, and the social partners did not include the item on their bargaining agenda.157
Generally speaking, Spanish employers do not regard working time restrictions to be a problem. The 1994 reform
introduced changes designed to make the use of working time more exible, but again these were not considered to be
a priority by employers; nor has there been any study evaluating its consequences in terms of the performance of the
labour market.
Unemployment Compensation
Although unemployment compensation has existed in Spain since the 1960s, the modern system was established in
1980 by the Ley Bsica de Empleo. Sub-title II of this law was converted into the Ley de Proteccin del Desempleo
(Unemployment Protection Law, or LPD) in 1984. The enacting of this law coincided with the reform of the LET
designed to ease the use of xed-term contracts, which established that a worker with 6 months of contributionsthe
minimum duration of a xed-term contractwas entitled to 3 months of unemployment insurance benet (UI) at a
rate of 80 per cent of the previous wage, tax-free. The general rule was that for each successive 6 months of
contributions, 3 more months of benets would accrue to the worker. The maximum duration was 24 months, to be
granted to those with 4 years of contributions over the previous 4 years, which was the reference period in general.
The substitution rate decreased to 70 per cent after 6 months and to 60 per cent after 12 months.158 The system was
supplemented by an unemployment assistance benet (UA) to be paid to workers who had not acquired the right to UI
but had contributed for at least 3 months, and also to those who exhausted their UI period but had family
responsibilities' and had passed a family income test. A special provision for workers aged over 55 established that UA
would be paid until
157
Probably because they thought they would never reach an agreement, as the employers were clearly against these measures, even though the trade unions have a more
favourable position towards them.
158
It should also be noted that benets were subject to a oor equal to the statutory minimum wage (SMW) and to a ceiling of 170 per cent of the SMW, to be increased to up
to 210 per cent for those with two or more dependent children.
retirement to those meeting all the requirements for entitlement to a retirement pension except ageestablished at 65
by Spanish Social Security law. UA was paid at a at rate of 75 per cent of the statutory minimum wage. There was a
special provision where workers could receive all their unemployment benets in a lump sum if they intended to set up
their own businesses or participate in a workers' co-operative or worker-owned rm.
This system underwent some change in 1989 when, in the aftermath of the 1988 general strike, the government
extended the duration of UA and reduced the age for indenite, that is until retirement, assistance benet to 52. The
system became such that a person dismissed at age 47, having worked for the previous 4 years, with family
responsibilities and with at least 15 years' contributions to the pension system, could retire from the labour market.159
This was certainly the purpose of the change, at a time when, despite huge employment growth, unemployment
declined much less than proportionately (recall Figure 11.1).
The system was thoroughly revised in 1992 when the government decided to implement a signicant reform of the
unemployment protection system, linked to change in the hiring regulations. The government was convinced that the
possibility of linking short employment contracts with unemployment benet periods was one reason for the nancial
strains experienced by the unemployment protection system, together with the idea that the system was too generous,160
and that this explained the resistance of the unemployment rate against any downturn despite employment growth
although there were no studies that substantiated such a position. Henceforth, the minimum contribution period was
to be 12 months over the previous 6 yearsthe new reference period. Each successive period of 6 months'
contributions was to provide 2 additional months of benet instead of 3. The maximum entitlement period remained
at 2 years. The replacement rate was reduced to 70 per cent for the rst 6 months but was otherwise left unchanged.
Those contributing for less than 12 months could obtain unemployment assistance.
The reform was supplemented with two further measures introduced
159
The worker would have 2 years of UI, 3 of UA and then, at 52, would qualify for the special scheme until reaching the age of 65. Of course, the amount of the retirement
benet would not be as high as that for a normal retirement, although it would at least provide a minimum guaranteed income. In fact, the number and proportion of these
workers does not appear to be very high, although it is difcult to know with certainty, and there are no studies on the matter.
160
The government's own calculations indicated coverage rates of the system amounting to over 100 per cent of the registered unemployed in early 1993. This, of course, is
logically impossible, and it was due to the fact that some unemployment beneciaries were not counted as registered unemployed. However, it provides an indication that the
coverage of the system was high, certainly higher than it had been a few years earlier.
in 1993. First, unemployment benets were made liable to income tax. Second, the notion of family responsibility was
tightened and more restrictive instructions were given to unemployment benet ofcers. The result of these changes
was a severe drop in the system's coverage rate. By 1995, the rate was back to its 1990 level, although it appears to have
stabilized thereafter.
What have been the labour market effects of these reforms? In principle, one would expect them to have increased the
outow from unemployment or, more precisely, to have increased the exit probability of unemployment beneciaries
compared with those not enjoying such benets. There are very few studies, however, on the inuence of the
unemployment system on the exit rate, and none on the inuence of the reform. Two main lines of analysis may be
mentioned. First, in a study based on the unemployment benets records, Cebrin et al. (1995) found that, contrary to
what one might expect on the basis of international experience, the exit rates of unemployment recipients did not
increase as they approached exhaustion. In their study of the probability of leaving the system before exhaustion, they
found that the period of entitlement was the most signicant variable. However, the effect was not continuous but
instead substantially increased for those with the maximum entitlement of 24 months. This suggested that the relevant
variable was not necessarily duration; rather, it was the specic characteristics of this group that accounted for their
particular behaviour. In addition, the replacement rate was not a signicant variable.
The weakness of this study, which referred to the pre-1992 period only, is that it was only able to observe
unemployment beneciaries, and it was thus unable to compare them with the rest of the unemployed. Using the
matched LFS records, Bover et al. (1996) have found a negative effect of unemployment benets on the rate of exit
from unemployment; an effect which was particularly signicant at short durations of unemploymentthe maximum
was observed at 3 months and faded away after 10 monthsPublished as a Bank of Spain study, this paper has been
very inuential, and it is probably responsible for the view that a reduction in unemployment benets would ease the
unemployment problem. However, it contained no specic evaluation of the 199293 reforms; nor is it without its
weaknesses,161 and neither the social
161
One weakness is the way in which the unemployment benet variable was constructed (see Alba 1996). Another is that half of the workers observed in the sample did not
receive benets on their rst observation, which implies that their previous job duration had been very shortpeople without past job experience do not qualify for
unemployment benets and were excluded from the sample. Thus, as suggested by Toharia (1997a), it may be that those without unemployment benets have a higher
probability of exiting unemployment but also a higher probability of re-entering. In addition, when considering the entitlement period, the evidence becomes less conclusive.
partners nor government are at present envisaging any further reform to the system.
The Structure of Employment and Unemployment

In previous sections we have described the various institutional changes to have taken place in Spain since the early
1970s. What have been the results of these developments in terms of employment and unemployment?
The basic gures on employment and unemployment have already been set out in Fig. 11.1. The general evolution
indicates that the employment boom of the second half of the 1980s followed the labour market reform of 1984. It is
highly likely that the changes introduced in 1984 exerted a positive, albeit small,162 effect on employment. The post-
1990 pattern raises doubts, however, as to whether this increase was merely an intensication of the response of
employment to GDP, as economic theory would lead us to expect, rather than a permanent shift in the level of
employment. An analysis of employment stability may help to resolve the issue.
Employment stability may be understood, following Garrido (1996), in two possible ways: as contractual stability or as
empirical stability. The notion of contractual stability refers to the proportion of employees holding xed-terms
contract in terms of the Labour Force Survey;163 that of empirical stability refers to the proportion of employees with
tenure in their jobs for less than one year. Another useful indicator is the proportion of employees with less than a year
of tenure still in employment a year later.164 Analysis of these three types of indicator (Fig. 11.2) suggests that there have
been two phases in Spain:
162
This was the wording of the Segura report (Segura et al. 1991) and it has generally been accepted as a valid statement by most experts.
163
The Labour Force Survey asks whether the contract is permanent or temporary, the latter including all categories, such as practice and training contracts for youngsters.
However, the fact that most answers indicate that the temporary contract belongs to the category others makes one suspicious of the accuracy of the answers, thus leading
to the consideration of temporary or xed-term contracts as a whole.
164
The way in which this indicator is calculated requires some explanation: it is the proportion of workers who in one year declare 1 to 2 years of tenure in their jobs at time t
over the number of workers who declare less than 1 year of tenure at time t-1. The way in which tenure was established in the LFS was changed in 1992 from answering by
intervals to the precise date of beginning the job. However, the coding of the LFS has maintained comparable categories and does not suggest that there was a break in the
series.
Fig. 11.2. Contractual and empirical employment stability, Spain, 19871997

Source: Labour Force Surrey.
1. Between 1985 and 1992, coinciding with expansionary business conditions, contractual stability declined; that
is, the proportion of employees with temporary contracts sharply increased.165 However, empirical stability
increased at a much slower rate, and the same can be said of the proportion of low-tenure workers still holding
their jobs after 1 year.
2. Since 1992, contractual stability has tended to remain relatively stable, even though this period has been one of
rst sharp recession and then mild recovery. However, and this is the most important point, empirical stability
has undergone substantial decline, as shown by the higher proportion of workers with less than 1 year of
tenure, as well as by the plummeting percentage of low-tenured employees who manage to retain their jobs.
This evolution can be interpreted as the probably unintended result of the various regulatory changes. The 1984
reformaimed at fostering temporary employment in order to increase total employmentwas probably effective
insofar as the response of employment to GDP increased, but it also set the scene for the development of a segment of
jobs characterized by high turnover.166 The existence of a regulation on xed-term contracts which established a
minimum duration of 6 months and a maximum of 3 years, after which workers were entitled to unemployment
benet, probably acted as a brake on such turnover. But when the minimum duration was increased to 1 year in 1992,
together with the more stringent entitlement conditions for unemployment benet, and later when the regulated xed-
term contracts were abolished altogether in 1994, in a move designed to yield more stability, exactly the opposite
happened. Initially, these further restrictions probably reduced the employers' ability to hire. Subsequently, however,
when employers were no longer constrained by the minimum limits but needed, or had become used to, short-term
contracts, they resorted to ordinary temporary contracts,167 with the paradoxical result that employment stability
sharply declined. This shift is also reected in the number of contracts registered with the employment ofces: from an
average of
165
As can be seen from Figure 11.2, it could be argued that the increase was more pronounced in the pre-1990 period.
166
Recent studies on turnover in Spain have found that worker turnover, as opposed to job turnover, is much higher in Spain than elsewhere in Europe. See Garca Serrano and
Malo (1997) and Garca Serrano (1998).
167
These are the so-called per task or service contracts, which have always existed but which were designed to satisfy the temporary needs of rms. By denition, their
duration is that of the task or service to be performed, which may be days, weeks or months. Casual information indicates that it was the employment ofces themselves
who encouraged employers to resort to this type of contract. Recent information from the Public Employment Service indicates that the average length of contracts in 1995
was around 30 days.
around 5 million in 198993, the numbers rose to 6 million in 1994, 7.3 in 1995, and 8.6 in 1996.168
It thus seems that the Spanish labour market has evolved into a dual structure, with two-thirds of employees retaining
permanent status and the rest in a highly mobile market. However, the important point in terms of social policy is
whether the workers who cater to this secondary market become trapped in their situation, and thus nd themselves
in circumstances resembling what can be termed social exclusion. This can be tested in three ways: rst, the
characteristics of temporary jobs at various points in time can be considered, in order to determine the extent to which
they are lled by the same people; second, the proportion of temporary workers who manage to obtain permanent
jobs, thereby escaping from the secondary market, can be examined; third, the evolution of the various age cohorts can
be followed in terms of their labour market situation.
As for the rst issue, Table 11B in the Appendix presents the characteristics of the stock of temporary workers in the
private sector at two points in time, 1992 and 1997. The reason for choosing these points is that they correspond to
what we have termed the second phase, that is, the phase when the labour market segmentation tended to get
encroached in the Spanish economy. There are two signicant features in Table 11B. First, the mean age of temporary
workers is more stable than that of permanent workers. Given that there is a lag of 5 years between the two points of
observation, this means that there has necessarily been a turnover within the group of temporary workers. Secondly, in
terms of the characteristics of the jobs involvedas represented by level of education, sector of activity, occupational
group: all of them proxies for the level of skill; as well as the composition by size of the enterprisethere is a
substantial degree of stability. This means that temporary workers are hired in rather specic jobs. One may therefore
conclude that the secondary sector probably responds to specic product market conditions which require more
exible labour market conditions, probably because of their inherent higher instability.
The second item of evidence concerns the proportion of temporary workers who obtain permanent jobs. The LFS
longitudinal data enable analysis of the dynamics of temporary and permanent employment. Given that each person is
interviewed for six successive quarters, it is possible to follow the situation of temporary workers for 15 months.
Figure 11.3 presents the prole of integration into permanent status of
168
See Ministerio de Trabajo y Asuntos Sociales, Boletn de estadsticas laborales, various issues. To put these gures in perspective, it should be pointed out that the number of
employees in 1996 was around 9 million.
Fig. 11.3. The employment situation of temporary workers (1992 and 1995)
Source: Labour force survey, matched les.
the workers interviewed for the rst time in 1992 and 1995.169 According to this information, between 10 and 15 per
cent of temporary workers observed at one point become permanent within just over a year. At this rate of exit, the
stock of temporary workers would be totally renewed in a period of 8 to 10 years. However, if account is taken of the
fact that other workers either lose their jobs or leave the labour force altogether, this replacement period is certainly
shorter. Moreover, it may also happen that some of those exiting from temporary employment to unemployment or
inactivity return to another temporary job, this being the paradigm of the permanently unstable worker. Such data,
however, do not provide a sufciently long follow-up on subjects for any rm conclusion to be reached on this point.
The nal item of evidence may help to resolve the preceding point. Instead of a longitudinal analysis proper, the length
of which is necessarily limited to 15 months, the LFS also allows for indirect longitudinal analysis based on the
following-up of cohorts of subjects. If the LFS yields a representative picture of the different cohorts of the
population, dened in terms of year of birth, it is then possible to follow groups of subjects over time. Table 11C in
the Appendix shows an exercise of this kind for the cohorts aged 20 to 35 in 1992, giving their situation ve years later.
It will be seen that despite the continuing prevalence of temporary work in Spain, people have, to a certain extent, been
able to integrate into more permanent jobs, albeit at a slower pace than used to be the case.170
On the whole, these three items of evidence suggest that the view of the demand-based, that is product market-led,
dual labour market adequately accounts for the Spanish situation. If this is so, one may conclude that the regulatory
changes introduced in 1984, probably together with the expansionary wave of the late 1980s, made the enlargement of
the secondary sector possible, and that once the latter had become established, the elimination of some of the
regulations could not reverse the situation.
But what about unemployment? Have the changes introduced in the regulatory framework of the labour market
affected its structure in any signicant way? Analysis of unemployment differentials by age and sex, as well as by level
of education, reveals that these differentials have tended to remain relatively stable over time, moving at the same rate
in times of expansion and recession (Garrido and Toharia 1996). A more
169
The analysis is restricted to workers for whom information is available at the six points of observation. In order to increase the sample size, four cohorts are included in
each prole: those entering in the four quarters of 1992 and those entering in the four quarters of 1995. The latter extend until the rst quarter of 1997.
170
On the issue of the labour market integration of young people, on which this assertion rests, see Garrido and Requena (1997).
signicant indicator is the proportion of unemployed without past job experience. This proportion tended to decrease
sharply with the expansion of temporary work in the late 1980s and since then has remained at a much lower level,
despite the increased unemployment of the early 1990s (Garrido 1996). This is, of course, the other side of increased
turnover.
A further indicator refers to long-term unemployment. If our dual labour market interpretation is correct, there should
be no signicant variations in the proportion of the unemployed out of work for more than a year. As it turns out, this
proportion has tended to follow the evolution of the global unemployment rate. Although duration dependence is
observed at the micro-levelthose who have been unemployed longer have a lower probability of exiting their
situationthere does not seem to be a ratchet effect like that observed in other OECD countries (Toharia, dir.,
1997).
Proper understanding of the Spanish unemployment problem, however, requires consideration of the role played by
the family. A substantial proportion of the unemployed are seeking a second income for the household, either the
spouses, usually female, or the children of the breadwinner. These categories of the unemployed show a signicantly
less active labour market search, a more choosy attitude towards employment and, on the whole, less likelihood of
leaving unemployment. But their situation is sheltered by the situation of their households. Two out of every three
unemployed persons live in a household in which someone else is employed. Moreover, nine out of ten of them live in
households in which someone else receives some kind of income from employment, unemployment benets or
pensions.171 Needless to say, in the case of young people, this situation is the other side of the secondary labour market.
What is happening, on the whole, is that young people are taking more time to nd stable employment and thereby
obtain the resources with which to establish themselves as independent households. The low levels of fertility in
Spanish society in recent years are but the reection of this situation.
Conclusion
Overall, one may argue that the exibility at the margin introduced in 1984 probably affected the Spanish economy's
capacity to create jobs to some extent. More importantly, however, it released the forces
171
For details see Toharia (1997b), who conducts thorough analysis of the income situation of the unemployed.
necessary for the creation of a wider dual labour market with a primary sector in which workers enjoy a substantial
stability and a secondary sector in which such employment stability is lacking. However, to the extent that this
secondary sector employs people who are not subsequently trapped in it, we cannot speak of distinct social
segmentation.
The subsequent reforms of 1994 and 1997 also show important lessons. The 1994 reform did not create signicant
economic effects notwithstanding its many legal changes. The 1997 reform has been intended to promote permanent
contracts, because the social partners and the Spanish society as a whole judged as excessive the wide use of temporary
contracts, especially in terms of the effects on families and young workers. The family supports a great part of the
weight of unemploymentshown in the relatively long time before the emancipation of womenbecause of the high
proportion of secondary unemployment in those directly related to the breadwinner. Linking this problem to the
debate on the reduction of unemployment benets, we cast some doubt on the increase of unemployment exits
through diminishing benets or entitlement periods. It is possible that a decrease in benets only will increase
monetary and non-monetary intra-family transfers without a signicant effect on unemployment.
To sum up, the experiment has probably been benecial to the economy as a whole because it allowed the more
permanent existence of this secondary sector, one of whose roles is precisely that of providing the economy with a
exible fringe. A long-term consequence of the 1984 changes, of course, has been the delayed social integration of
young people and the consequent decline in fertility rates, but these features are probably due just as much to the
unemployment situation that already existed in Spain in 1984. One could argue that the social integration of young
people would have been even more difcult if the economy had remained with a smaller secondary sector and lower
worker turnover.
Appendix 331
Appendix
Table 11A. Institutional Changes in Spain's Labour Market: Impact on Flexibility
Year Description of change Effects on exibility
1983 Establishment of a legal maximum working ? Few rms affected by this measure
week of 40 hours
1984 Elimination of the principle of causality for +
temporary contracts: xed-term contracts al-
lowed almost without restrictions; creation of a
myriad of other temporary contracts, such as
training and practice contracts, contracts for new
lines of business, etc.
Unemployment protection lawminimum pe- + Positive effect due to the jopint use of xed-
riod of contribubtion established at 6 month- term contracts and unemployment benets
smaximum entitlement period 2
yearsSubstitution rate 80/70/60
- Possible negative effects on search incentives
1989 Unemployment assistance extended in the case ?
of certain groups (older and long-term unem-
ployed)
1992 Reform of the unemployment protection ?
lawminimum contribution period extended to
1 year for unemployment insurance
Minimum contract period for xed-term work- -
ers increased to 1 year-incentives to the use of
training and practice contracts eliminated
199394 Minor changes in the regulations of unemploy-
ment assistance, reducing workers' entitlements.
1994 Fixed term contract eliminated as a general case;
refurbishing of labour law aimed at giving more
weight to collective bargaining; some changes in
the regulations of dismissal costs; elimination of
training contract and creation of apprenticeship
contract
- Elimination of xed-term contracts, although
compensated by the wider use of per task or
service contracts+ More incentives to decen-
tralized bargaining; increased power of employ-
ers? Changes in dismissal costs for permanent
employees
1997 Creation of a new permanent contract with a ? (Doubtful effects on dual structure of labour
lower dismissal cost, after agreement of social markets)
partners; incentives to increase permanent
contracts
332 Appendix
Table 11B. Characteristics of permanent and temporary workers, private sector, Spain, 1992 and 1997
Popul Fe- Mage Youth Old Base- Unie- Agric Manfg Constr Serv Bluec Un- Sma- Low- Low-
male du du skill lent ten dur
Both genders
Permanent
1992 4,265- 26.8 40.1 8.5 35.4 70.8 9.9 3.7 38.4 8.4 49.5 46.0 4.5 32.4 6.4
.3
1997 4,483- 32.1 40.3 7.0 36.6 61.2 12.9 3.1 39.6 3.0 53.9 48.2 0.0 33.6 7.5
.4
Temporary
1992 2,709- 36.0 29.9 41.1 12.6 74.7 5.5 8.1 25.5 19.7 46.7 49.5 11.9 42.2 76.1 20.7
.5
1997 2,886- 36.5 31.1 34.1 13.4 65.8 9.1 7.6 42.8 2.2 47.3 61.6 0.0 40.4 89.2 22.7
.2
Males Permanent
1992 3,123- 41.1 6.6 39.0 72.0 9.4 4.7 43.0 10.9 41.4 55.8 5.5 29.3 5.6
.8
1997 3,045- 41.3 5.7 40.3 62.7 12.1 4.0 48.6 4.0 42.9 53.9 0.0 29.1 6.3
.5
Temporary
1992 1,733- 30.3 38.7 13.6 77.6 3.9 9.7 26.7 29.7 33.8 65.4 17.0 39.3 76.8 21.5
.6
1997 1,832- 31.7 32.0 15.3 71.4 5.9 9.1 55.8 2.9 32.1 73.8 0.0 37.7 88.1 23.2
.4
Females Permanent
1992 1,141- 37.1 13.7 25.4 67.4 11.4 1.0 25.7 1.6 71.7 19.1 1.9 41.0 8.6
.5
1997 1,437- 38.2 9.8 28.8 58.0 14.7 1.2 20.5 1.0 77.2 36.1 0.0. 43.1 10.1
.9
Temporary
1992 975.9 29.0 45.4 10.7 69.4 8.4 5.4 23.4 1.8 69.5 21.4 3.0 47.3 74.7 19.3
1997 1,053- 29.9 37.7 10.1 56.2 14.6 5.1 20.2 0.9 73.7 40.4 0.0 45.0 91.1 21.9
.8
Meaning of indicators:
Popul: number of people in each group (thousands).
Female: % women.
Mage: mean age (years).
Youth: % under 25.
Old: % 45 or older.
Basedu: % with at best basic education.
Uniedu: % with some university degree.
Agric: % working in agriculture.
Manfg: % working in manufacturing.
Constr: % working in construction.
Serv: % working in services.
Bluec: % holding a blue-collar occupation.
Unskill: % holding an unskilled occupation.
Smalent: % working in rms with 10 employees or less.
Lowten: % with a tenure of less than 1 year.
Lowdur: % with a contract of 3 months or less duration (only applicable to temporary workers).
Source: LFS (Labour Force Survey).
Appendix 333
Table 11C. Cohorts aged 2530: the situation ve years later, Spain, 1992
Labour Males Females
Market Born 196872 Born 196367 Born 195862 Born 196872 Born 196367 Born 195862
Status (2024 in 1992) (2529 in 1992) (3034 in 1992) (2024 in 1992) (2529 in 1992) (3034 in 1992)
Per- 9.9 25.7 29.2 38.8 45.1 47.9 8.4 19.7 19.1 25.4 25.1 28.0
man-
ent
em-
ploy-
ees
Tem- 27.0 28.8 28.3 22.6 19.0 15.5 21.7 22.7 20.5 14.0 11.1 10.2
porary
em-
ploy-
ees
Other 7.5 11.5 13.7 17.3 18.7 20.7 3.6 6.3 6.7 7.7 7.9 8.6
em-
ployed
Unem- 5.0 4.8 2.8 1.4 0.7 0.1 8.5 8.5 6.8 3.5 3.0 1.8
ployed,
no
past
experi-
ence
Unem- 12.4 14.9 14.1 14.2 11.1 11.6 12.9 17.0 14.7 16.1 11.9 13.6
ployed,
past
experi-
ence
Stu- 27.0 9.9 7.2 1.4 1.2 0.4 34.5 10.6 6.9 1.9 1.4 1.1
dents
Other 11.2 4.6 4.8 4.2 4.4 3.7 10.5 15.2 25.3 31.4 39.5 36.8
inac-
tive
Source: LFS (Labour Force Survey).
334 References
References
Alba, A. (1996), Explaining the transitions out of unemployment in Spain: the effect of unemployment insurance,
Working paper 96-71, Department of Economics, University of Carlos III, Madrid.
Bentolila, S., and Dolado, J. J. (1994), Labour exibility and wages: lessons from Spain, Economic Policy, 5499.
Blanchard, O., and Jimeno, J. F. (eds.) (1994), Unemployment in Spain: Is there a Solution? London: Centre for Economic
Policy Research.
Bover, O., Arellano, M., and Bentolila, S. (1996), Unemployment duration, benet duration and the business cycle,
Estudios econmicos, Banco de Espaa, Servicio de Estudios, 57.
Cebrin, I., Garcia, C., Muro, J., Toharia, L., and Villagmez, E. (1995), Prestaciones por desempleo, duracin y
recurrencia del paro in J. Dolado and J. Jimeno, Estudios sobre el funcionamiento del mercado de trabajo espaol. Madrid:
Fundacin de Etudios de Economa Aplicada, 15593.
Crculo de Empresarios (1993), El mercado de trabajo espaol: reforma y creacin de empleo, Boletn, 57.
Economistas (1993), Special issue on Reforma del mercado de trabajo [Labour market reform], 57.
Garca Serrano, C. (1998), Worker turnover and job reallocation: the role of xed-term contracts, Oxford Economic
Papers, 50: 70925.
and Malo, M. A. (1997), Movilidad de trabasadores y de puestos de trabajo en empresas espaolas grandes.
and Toharia, L. (1994), La participacin de las rentas del trabajo y de los benecios en la renta nacional, Revista
de economa y sociologa del trabajo, 2526: 7384.
Garrido, L. (1996), La temporalidad, pacto intergeneracional o imposicin?, in Consejo Econmico y Social de la
Comunidad de Madrid, La duracin del contrato de trabajo. Madrid: CES Madrid, 4774.
and Requena, M. (1997), La emancipacin de los jvenes en Espaa Madrid: Instituto de la Juventud.
and Toharia, L. (1996), Paro y desigualdad, in Pobreza, necesidad y discriminacin, II Simposio sobre Distribucin de
la Renta y la Riqueza, Fundacin Argentaria, Visor Distribuciones, 963.
Jimeno, J. F. (1992), Las implicaciones macroeconmicas de la negociacin colectiva: el caso espaol, Moneda y Crdito,
195: 22381.
and Toharia, L. (1993a), The effect of xed-term employment on wages: evidence from Spain, Investigaciones
econmicas, 17/3: 47594.
(1993b), El despido y sus costes: qu reformas?, Economistas, 57: 1321.
(1994), Unemployment and Labour Market Flexibility: Spain. Geneva: ILO.
Lorences, J., and Felgueroso, F. (1996), Negociacin colectiva y empleo: experiencias recientes, Economistas, 70: 8693.
References 335
Malo, M. A. (1998), Las indemnizaciones por despido, un problemo de negociacin, modrid: Edi torial ACARL.
and Toharia, L. (1994), Los costes del despido en Espaa, Revista de economa y sociologa del trabajo, 2526: 18092.
(1997), Las indemnizaciones por despido: su origen, sus determinantes y las enseanzas de la reforma de
1994 Documentacin Laboral, 51: 1332.
(1999), Costes de despido y creacin de empleo, Madrid: Ministerio de Trabajo y Asuntos Sociales, Coleccin Estudios.
Malo de Molina, J. L. (1985), Coherencia del sistema de relaciones industriales y eciencia del mercado de trabajo,
Papeles de economa espaola, 22: 24464.
(ed.) (1988), El debate sobre la exibilidad del mercado de trabajo. Madrid: Fundacin F.I.E.S.
Ministerio de Trabajo y Asuntos Sociales, Boletn de estadsticas laborales, various issues.
Papeles de Economa Espaola (1997), Special issue on El mercado de trabajo en perspectiva europea, [The labour
market in a European perspective] 72.
Revista Econmica de Catalunya (1997), Special issue on La reforma del mercat de treball [Labour market reform] 32.
Sagardoy, J. A., and Rivero, J. (1993), La regulacin legal del mercado de trabajo, Boletn del Crculo de Empresarios, 57.
Segura, J., Durn, F., Toharia, L., and Bentolila, S. (1991), Anlisis de la contratacin temporal en Espaa. Madrid: Ministerio
de Trabajo y Seguridad Social.
Toharia, L. (1985), En torno a la supuesta rigidez del mercado de trabajo en Espaa, Boletn del Crculo de Empresarios,
30: 4763.
(1996), Empleo y paro en Espaa: evolucin, situacin y perspectivas, Ekonomiaz, 35: 3667.
(dir.) (1997), The Labour Market in Spain. Series on Labour Market Studies Brussels: EC.
(1997a), El sistema espaol de proteccin por desempleo, Papeles de economa espaola, 72: 192213.
(1997b), El paro en Espaa: puede ser tan alto?, Paper presented at the I Foro de reexin sobre el empleo en
Espaa, Valencia, October.
Unon General de Trabajadores (UGT) (1994), El Estatuto de los Trabajadores y la reforma laboralEdicin comentada.
Madrid: UGT.
Zabalza, A. (1996), La recesin de los noventa bajo la perspectiva de los ltimos treinta aos, Moneda y Crdito, 203.
Conclusions
GSta Esping-Andersen and Marino Regini
This volume has presented the main ndings stemming from detailed examination of eight European countries, as
well as from broader, cross-national comparative research. What, at this point, can we conclude? In highly stylized
form, the principal lessons that emerge and that have policy relevance, directly or indirectly, are the following.
Europe Is not Homogeneous

As far as labour markets are concerned, it is seriously misleading to consider Europe as a whole; and policy discussions
premised on a simple USA-Europe contrast should be avoided. There is simply too much heterogeneity within Europe
as regards unemployment levels, trends, duration and composition. A quick look at the tables and gures in each
country chapter is sufcient to appreciate the extent of such diversity.
Similarly, regulatory practices and deregulatory reforms diverge tremendously from one country to another. To be
sure, in no country of the European Union, with the possible exception of France, has recent legislation or collective
bargaining added further rigidities in the labour market. In all countries, labour market reforms have tended to inject a
higher degree of exibility. But in some caseslike Britain and Denmark, and to a lesser extent Sweden and The
Netherlandsexibility has acquired the role of a general principle on which the working of labour markets is based;
and of the guiding principle that informs all new legislation and social partners' strategies. In otherssuch as France,
Germany, Spain and Italylabour market deregulation has often been accompanied by other, apparently
contradictory, measures of re-regulation. This is because deregulation has been conceived as a controlled experiment,
aimed at inserting some degree of exibility in one or another segment of the labour market, but not at questioning its
general rules.
Conclusion 337
Regulation Affects Structure, not Levels of Unemployment

It makes a great difference whether we are concerned with levels or structure of unemployment. One of the clearest
results from our study is that the various forms of regulation do, or at least may, have some impact on who is
unemployed, but they cannot explain overall unemployment levels. Generous social benets, highly egalitarian wage
structures with high xed labour costs, and strong job protection can be shown to negatively affect the employment
chances, and thus increase the risks of unemployment, of three groups in particular: youth, women workers, and the
low-skilled work force. That is, such regulatory practices or rigidities have a selective effect on unemployment
probabilities. The main reason this is so, is that various kinds of employment protection affects ows, especially in and
out of jobs. But the same regulatory practices, contrary to widespread opinion, have no visible effect on aggregate
unemployment rates. That is, rigidities cannot explain why some nations, say Spain or France suffer higher
unemployment levels than others, say the Netherlands.
The lack of any visible direct relationship between regulatory practice and unemployment levels suggests that
deregulation reforms will probably not help reduce mass unemployment much. Since our study was explicitly limited
to the impact of regulation, we have little to say on what other factors may inuence levels in general, and the
permanency of mass unemployment in particular. An oft-cited argument is that its principal cause lies in overly
restrictive macro-economic policies, an issue not dealt with in this volume. Another, not necessarily conicting,
argument is that much of Europe's, especially Southern Europe's, unemployment problem stems from a labour supply
explosion coupled to heavy out-migration from agriculture and industry. A third, and more directly policy-relevant,
suggestion is that employment growth is vastly helped by long-term wage behavior. This is the lesson epitomized by
the so-called Dutch miracle, namely the importance of consensually bargained, sustained real-wage moderation across
the board.
As far as structure of unemployment is concerned, our study has paid special attention to the impact of deregulatory
policies over the 1990s, a period in which almost all European countries have introduced a substantial amount of
potentially meaningful reforms. A common characteristic of European deregulation, in fact, is its partial and targeted
character. Deregulation strategies have typically concentrated on the following fronts. A large number of countries
have reduced social benets
at the margin. Perhaps more importantly, in a few cases, like Denmark, duration of unemployment benets has been
reduced. Many countries have liberalized various forms of atypical contracts, such as part-time work, especially in the
Netherlands, or xed-term, temporary contracts: Germany and Spain stand out. There are also widespread efforts to
improve the employability of youth via subsidies, especially training contracts, and to liberalize dismissals. In no
country, except Britain to some extent, has there been any attempt at deregulation tout-court.
There is substantial evidence that these forms of partial deregulation, or adjustment at the margin, affect the
unemployment structure positively. The Danish duration reform has led to a rapid decline in youth unemployment
rates; subsidized training contracts do appear to augment youth employability; and The Netherlands' liberalization of
part-time jobs has stimulated women's employment tremendously.
Regulation and Flexibility Have Complex and Contradictory Effects

It is misleading to examine European rigidities or American exibilities without also considering offsetting
mechanisms. Rigidities are often coupled to informal practices such as self-employment or irregular, undeclared
contracts. On the other hand, even highly exible labour markets, such as the American, exhibit informal rigidities
regarding wages, such as efciency wages, or redundancies.
More generally, the regulatory regime of any advanced economy is a highly complex system, since it involves different
institutions, levels and mechanisms, and has multiple, and often contradictory, economic effects. Hence, outright
deregulation of labour markets may be neither feasible nor benecial in terms of greater exibility. Attempts to
dismantle some institutions or rules may simply lead to their replacement by other regulatory modes and mechanisms,
or they may generate hidden rigidities. Since the institutions that set constraints may also provide resources and
competitive advantages, the cumulative effect of such deregulation may be negative. Flexibility and deregulation are,
therefore, far from being simple recipes with unequivocalwhether positive or negativeeffects on labour market
performance.
The selective impact of regulation on unemployment incidence appears related to structural characteristics of labour
markets. The impact seems stronger in segmented, or insider-outsider labour
Conclusion 339
markets. Yet, in general we should avoid exaggerating the role of rigidities. Simply, the evidence is not powerful
because there are many alternative and plausible explanations for why some groups are particularly hard hit by
unemployment, or why some nations suffer from heavier unemployment levels than others.
Among the more important other reasons identied in this report, a few stand out. The rst is bargaining structure.
Highly co-ordinated systems of collective bargaining are better able to minimize the adverse effects of, say, overly
generous social protection for the unemployed, high wage costs or job rights. The crucial importance of collective
bargaining comes out especially in the oft-cited Dutch case where coordinated, long term wage restraint successfully
offsets other, possibly negative, effects. The second refers to the link between education systems and labour market
training. Where there are strong, institutionalized educational bridges, as in the German dual system, youth are less
likely to be unemployed or to suffer prolonged unemployment. The third reason is simply national idiosyncrasies:
Finland's recent burst of mass unemployment is related to the collapse of the ex-Soviet markets; Spain's massive
unemployment problem is to a large extent related to an unfortunate temporal coincidence of de-ruralization and
deindustrialization.
Policy Lessons
What lessons can we draw from the European experience with deregulatory reforms? Perhaps the single most
important lesson has to do with how to implement reforms. Our survey suggests very clearly that any meaningful
reform is most likely to pass and have an impact if it is based on consensus with and among the social partners. It
follows that countries blessed with a strong capacity for consensual policy-making and co-ordinated bargaining are, all
else considered, better placed. A second important lesson is that the kind of targeted and partial deregulation that has
characterized European reforms in the 1990s can bring positive, albeit generally selective, results. Since regulations
inuence composition rather than levels of unemployment, it is logical that deregulation strategies will mainly affect
composition.
More generally, one policy-relevant lesson that comes from this study is that some form of publicly subsidized
employment promotion for problem groups is a potential win-win strategy. Direct or indirectsay, via tax
concessionssubsidies help lower the employer cost of hiring
and training but do not disturb existing wage-setting practices or trade union prerogatives. In other words, exibilizing
via subsidization is more likely to nd broad acceptance. Even rather dramatic changes in the labour market regulatory
regime appear to be acceptable as long as they are targeted at specic groups, such as young people; focused on
regions of a country, particularly the more underdeveloped areas; and implemented for xed periods of time; as long as
these changes are the outcome of bread concertation.
Nevertheless, the positive effects of deregulation that we have been able to identify should be evaluated with
considerable caution. Flexibilization or liberalization of, say, hiring or ring practices may produce unwanted
substitution effects. Moreover, as especially our Italian study notes, partial deregulation risks producing asymmetries.
In the Italian case, easing collective dismissals has boosted lay-offs during downswings, but has not been accompanied
with hirings during upswings. In Spain, liberalizing xed-term contracts may have strengthened insiders' privileges and
thus labour market segmentation because the reform did not provide incentives to training or to re-hiring temporary
workers on regular employment contracts.
What, nally, seems to be the balance sheet? On the credit side, there is little doubt that some forms of deregulation
are desirable and that they may ease particular bottlenecks for particular labour force groupsin Europe, mainly
youth. There is no patent formula for deregulation that can or should be applied to Europe as a whole. In some
economies, a seemingly small intervention can bring about massive results, as in Denmark; in others not. The national
context is overdeterminant and this means we should be exceedingly sceptical about importing or copying models
from one country to the next. On the debit side, it should be clear that deregulation of the kind that is realistically
possible in Europe will not provide a panacea: it will probably do little to alleviate mass unemployment; it might shift
the incidence and thus affect the structure of unemployment. Of course, diminishing the extreme concentration of
unemployment among youth in some European countries would amount to a substantial welfare and efciency gain.
To wrap up, if we are permitted a more sweeping conclusion it would appear that in some countries at least, there are
two targets of reform that seem at least as urgent as simple deregulation. The rst has to do with collective bargaining
practice. Managing unemployment is greatly facilitated when, and if, the social partners are capable of strong
coordination and consensus-building. Reforming industrial relations systems may, in some cases, be a pre-condition for
successful labour market management. The second has to do with the interaction of
Conclusion 341
formal education and work-related training. Resolving the unemployment problem is much more difcult in countries
with a potential low-skill equilibrium, that is, where there is simultaneous market and government failure in the
production of the kinds of skills that are required.
Europe 45, 50, 52
Index Audenrode, M. van 207
Australia 689n., 207
Austria ; employment structure 56, 58, 679, 263; regulation
57, 100; changes 47, 62; employment protection 34, 55; tax
A New Course (1993) 190 wedge 601; wage determination 34, 41, 59Axelrod, R.
Aaberge, R. 165 191
Aarts, L. J. M. 184 BA (Bundesanstalt fr Arbeit) 230
berg, R. 168 Baccalaurat 263
Abraham, K. 33, 86, 2324 Bach, B. 161
accords drogatoires 252 Baden-Wrtemberg 228
Accornero, A. 15 Banca d'Italia 283, 290, 295, 300
Ackum, A. S. 163 Bank of Spain 309n., 322
activation programme (Denmark, 1996) 49 Barbieri, P. 66n.
Acuerdo Interconfederal sobre Negociacin Colectiva (General Agree- bargaining structure, see centralized bargaining; coordination;
ment about Collective Bargaining) 317 decentralized wage bargaining
Adam, P. 17, 77, 86, 282 Bavaria 222n.
Addison, J. T. 37, 118 BDA (Confederation of German Employers' Associations)
agency work ; France 251; Italy 274, 281, 302; Netherlands 2267
1878, 193, 2045 BDI (Federation of German Industries) 2267
aggregate demand 2, 150 Bean, C. 74
Agricultural Wages Board 124 Beatson, M. 11920, 124, 138
agriculture 137, 332, 337; France 260, 2678; Germany 216,
221 see also de-ruralization
Aid to Families with Dependent Children (AFDC) 73
Akerlof, G. A. 219
Alba, A. 322n.
Albk, K. 207
Albeda, W. 189n.
Albert, M. 26
ALMP (active labour market policies), see policy, active labour
market
Alogoskous, G. 712, 75, 778, 857, 89, 128
Amsterdam 195
Amsterdam Treaty 54
ANBA 236n.
Andersen, D. 161, 172
Anglo-Saxon countries ; dynamics of reform 44, 5562;
regulation effects 66, 74, 86ANPE (French Public
Employment Service) 249
Antolin, P. 207
Appelbaum, E. 215
Appelborn, A. 161
apprenticeships 80n., 235
ARAN 284
ASSEDICs 257
asymmetries 53, 340
Atkinson, A. B. 43, 75, 132
Atkinson, J. 16, 119
atypical work contracts 20, 33, 56, 78, 86, 338; Britain 13840;
Denmark and Sweden 1545; France 247, 2501, 268;
Italy 271, 274, 276, 281, 295, 300; regulation across
344 Index
Netherlands 191, 201

Beer, P. de 194 Buti, M. 334, 40n., 41, 44, 54, 59
Beissinger, T. 229 Butter, F. A. G. den 185, 193, 2023
Belgium 57, 60, 167; employment: patterns 56, 58, 68, 207; Buttler, F. 222n.
protection 34, 389, 55, 86, 186; regulation changes 48, Calmfors, L. 52n., 54, 82, 100, 150n., 152
51, 62; unemployment benet 44, 61, 74, 76, 101n.; wage Campania 272
determination 34, 40, 42, 59 Campbell, D. 12
Bell, L. A. 213 Canada 689n., 78n., 82
Bellman, L. 150n. Canziani, P. 17, 77, 86, 282
Bentolila, S. 85, 88, 310n., 318 Card, D. 82
Braud, M. 255 Carling, K. 162, 177
Berg, G. J. van den 196 Casavola, P. 276, 293
Berger, S. 26 Casey, B. 39n., 140
Bergstrm, J. 168 Catalonia 31011n.
Bernard Brunhes Consultants 247, 253, 256n. Cebrin, I. 322
Bertola, G. 6, 45, 72, 82, 859; on France 249, 251, 267; on Center for Business and Policy Studies (Sweden) 153
Italy 2759, 293, 300 centralized bargaining 41, 79, 889, 101, 106
Beveridge curve 21516; West Germany 217 CEOE (Confederacin Espaola de Organizaciones Empre-
Beveridge Report 121 sariales) 315
Bielenski, H. 231n., 232 Cette, G. 255
Bispinck, R. 225, 2279 CFL (contratto di formazione lavoro) 282
Bjrklund, A. 82, 14877 child-care< ; benets 48, 1045, 156, 16970; leave 161,
Blair, T. 49n., 120 1712Chirac, J. 250
Blanchard, O. J. 2, 99100, 108, 207, 215n.; regulatory effects Christodes 275
71, 80, 88, 92 CIG (Cassa Integrazione Guadagni) 273, 278, 288, 292, 295
Blank, R. 75, 85, 238 CIGO (CIG Ordinairia) 39n., 285
BLS (German Bureau of Labour Statistics) 213, 216 CIGS (CIG Straordinaria) 399n., 277, 2801, 2878, 302
Boeri, T. 2756, 303 Crculo de Empresarios 311n.
Boje, T. P. 168 CNEL 275, 300
Booth, A. 71, 128 Codetermination Act, German (MitbG, 1976) 21112, 226
Boragine, I. 275 Coe, D. 45
Brsch-Supan, A. 216n. Colin, C. 253
Bover, O. 322 Collective Agreement Act, German (TVG, 1949) 226
Boyer, R. 1213, 18 collective bargaining 5, 12, 142, 194, 33940; and deregulation
breadwinner model, see family 24,
Britain, see United Kingdom
British Household Panel Survey (BHPS) 128
Brittan, A. 54
den Broeder, C. 186
Broersma, L. 207
Brown, W. 73, 801; and Britain 116, 120, 125, 139, 142
Brunetta, R. 282
Brunhes Consultants 247, 253, 256n.
Bruni, M. 289, 303
Bruno, M. 100
Bchtemann, C. F. 32, 856, 2312
Budget Law (Italy, 1999) 279
Bundestag 237
Bndnis fr Arbeit (alliance for jobs) 228
Burchell, B. 139
Burda, M. 202, 207
business cycle 13, 33, 94, 255, 318; Germany 21819;
Index 345
149; France 252, 256; Germany 21112, 223, 22630; Department of Employment (UK) 116, 132, 140
Italy 274, 280, 282, 284; and regulatory reform 402, 45, Department of Labour (Denmark) 154
501; Spain 311, 31617, 319 see also centralized deregulation 14, 67, 14, 338; Britain 11544; Germany
bargaining; decentralized wage bargaining 2379; Netherlands 1834, 194, 2056; and regulatory
Communist Party (Sweden) 154 regimes in advanced economies 215
Conservative Party (UK) 11516, 1202, 143 Deregulierungsbericht 222
conservative regimes 335 Deregulierungskommission 222, 226
construction sector 260, 268, 284, 289, 332 de-ruralization 75, 81, 92, 1012, 1057, 339
Contini, B. 207, 289 Deutschmark 203, 220
convergence 534 DGB (German Trade Union Confederation) 227
co-ordination 26, 412, 59, 1001, 106 Diamond, P. 207, 215n.
coverage< ; benet 31, 38, 61, 185; Italy 28891; collective Dickens, L. 137, 139
bargaining 41, 62Cramer, U. 230 Dickens, R. 141
Croce, G. 282, 298 Dicks, M. 132
Crouch, C. 12, 26 DIHT (German Association of Chambers of Industry and
CSERC (Conseil Suprieur de L'Emploi, des Revenus et des Commerce) 227
Cots) 254 disability programmes 117, 315; Denmark and Sweden 157;
DA (employers' confederation, Denmark) 159 Italy 281, 287; Netherlands 182, 184, 192, 205; reform
DAG (white collar union, Germany) 227n. 389, 434, 47, 612
Dahrendorf, R. 14 dismissal 86; collective 267, 275, 314, 340; France 247, 250,
Daniel, W. 123, 139 2667; Italy 275, 280,
Danish Economic Association 154
Danish Employers Federation 152
Danish Institute for Social Research (Socialforskningsinstituttet)
160
Danthine, J. 54
DARES (Direction Gnrale de L'Animation, de la Recherche,
desertication Etudes et desertication Statistiques) 253
Datta Gupta, N. 164
Daveri, F. 42n.
De Luca, L. 289, 303
Deakin, S. 20, 49; on Britain 11544
decentralized wage bargaining 26, 456, 50, 82, 86; Denmark
and Sweden 159, 166, 1689; France 252; Italy 274;
Netherlands 1845, 18990, 205
de-industrialization 50, 75, 81, 92, 339; regulation and
unemployment 101, 105, 107
Del Boca, A. 291, 303
Dell'Aringa, C. 276, 286
Delors White Paper 54
Delsen, L. 193, 203
Demekas, D. G. 275
demographics 1056
Denmark 5, 100, 105, 14877, 198, 263, 336; employment and
unemployment patterns 678, 70, 747, 856, 945, 207;
national debate 14855; reform 31, 1712; labour
policies 57; policy lessons 340; regulation 36, 39, 412,
479, 512, employment protection 34, 55, and structure
338; wage setting 34, 59; welfare 602; working patterns
56, 58; unemployment 1602; unions, role of 1668, 174;
wage bargaining systems and policy changes 1559; wage
dispersion 1646, 173
346 Index
3023; protection 44; Germany 21112, 221, 2256, EPL (employment protection legislation) 34, 55, 62
2315 see also employment, protection Epstein, R. 116
Dismissal Protection Act (Germany, 1969) 231 Ermann, K. 230
distributive services 260, 268 Esping-Andersen, G. 17, 12, 19, 33, 33641; regulation:
Dolado, J. 803, 88, 310n., 318 effects 6695; and unemployment 99112
Donges, J. B. 222 Essex University 131
Dore, R. 26 ethnic minorities 142, 182, 196, 204
DQP (diversied quality production) 18 European Commission 53n., 86, 247, 261
Drifll, J. 54, 82, 100, 152 European Community (EC) 1, 46, 546, 92, 140; directives
dual education 2357 49; Collective Redundancies (1975, 1992) 123; on
Due, J. 190 dismissals (1992) 36, 250, 280; Equal Treatment in
Dutch Central Planning Bureau (CPB) 202 Employment (1976) 122; Working Time (1994) 117, 126
East Germany 212, 229; labour market trends 21921; and European Labour Force Survey 251
West German industrial relations framework 225, 228, European Union (EU) 92, 336; and Britain 115, 120, 137,
236 1434; dynamics of reform 30, 41, 43, 46; and France
ECB (European Central Bank) 222 261, 263; and Italy 271, 2856, 288, 294, 298, 303
Economic and Social Council 310
Economic Council (Denmark) 153
Economic Outlook 69n.
Economistas 311n.
Edin, P. A. 173
education 5, 78, 1047, 339, 341; Britain 120; Denmark and
Sweden 154, 1567, 1702; France 2635; Germany
214; Italy 2979; Netherlands 1968; Spain 326, 3323;
United States 214
efciency, labour market 2002, 204 see also matching
EIRR (European Industrial Relations Conference) 14, 16
elderly, see older workers
eligibility criteria 47, 61
Elmeskov, J. 69, 87
Elvander, N. 158
Emerson, M. 119
Employers' Associations; Germany BDA 227
employers' associations 250; Germany 211, 2289, 237;
Netherlands 189, 191; Spain 315, 317
employment 246, 268, 308; contracts 1234, 166; ows 197,
207; growth 512, 107, 1823; performance and
rigidities 703; protection 1112, 301, 45, 73, 8590;
Britain 1213; Denmark and Sweden 153, 157; and
external exibility 359; Germany 235; Italy 2801;
Netherlands 1867, 205; and unemployment 8490, 101,
1069; and wage setting 34; see also dismissal protection;
stability 94, 2335, 303, 3236, 330; structure 68; Britain
127, 13643; France 25965; Germany 21516, 234; Italy
267, 272, 290300; movement between 1345; Nether-
lands 1979; Spain 3239
Employment Gazette 132
Employment Promotion Act (Germany, 1985) 2312
Employment Relations Bill (UK, 1999) 120, 122, 125
EMU (European Monetary Union) 53, 311n.
Enqute Emploi 245n., 250, 261n., 2624, 268
EPI Employment Audit 11819, 133, 136, 139, 141
Index 347
103, 105; ows 95, 207, 293

Eurosclerosis thesis 100, 212; and regulation 6695 Franco, D. 276, 303
Eurostat 251, 2613, 268, 285 Franz, W. 21617, 224, 229, 235
Evans, S. 139 Freeman, R. B. 32, 72, 823; and Germany 211, 21415,
Fabi, M. 275 2234, 229, 238; and Netherlands 181, 185
Fabiani, S. 298 FS (exible specialization) 1820
Factories Acts (UK) 126 Fuchs, S. 21139
Faini, R. 293 full employment goal 1213, 67
Fair Wages Resolution (UK) 124 Ganoulis, I. 275
family credit 127 garbage-collector model (skraldemandsmodellen) 154
family solidarity 57, 102, 104, 137, 273; and reform 35, 44; in Garca Serrano, C. 318, 325n.
Spain 322, 32930 Garrido, L. 323, 3288n., 329
Felgueroso, F. 317 Gauti, J. 24568
female labour, see women Gautier, P. A. 1978, 207
fertility 1045, 32930 Gent system 167
Fiat 39, 278 German Civil Code 231n.
Finland 69, 165, 263, 339; dynamics of reform 48, 57, 59; Germany 165, 21139, 256n., 289, 336, 338; and Britain 124,
employment patterns 56, 58, 68; employment protection 143; education system 214, 339; employment patterns
34, 55, 62; welfare system 601, 167 56, 68, 234; ows 945, 207, 293; labour market trends
ring decisions and regulations 31, 378, 52, 856, 340; 21221; exibility 16, 32, 579, 22638; national debate
France 24952; Italy 271, 274, 2767, 294, 300;
Netherlands 1867, 193, 205; Spain 31215 see also
dismissal protection; employment, protection
rst job seekers 299
scal policy 99100
Fitzenberger, B. 229
ve-year law (1993) 250, 2547
xed-term contracts 338, 340; dynamics of reform 33n., 368,
456, 56; France 24952, 261, 2668; Germany 2313,
239; Italy 271, 2745, 279, 2812, 294, 296, 302;
Netherlands 188; Spain 310, 313, 315, 320, 325, 331
Flchentarifvertrag, see collective bargaining
Flex-Act (1999) 188, 206
exibility 4, 31, 58; contradictory effects 1421, 3389;
dynamics: Britain 116, 1207; France 24859; Germany
22638; Italy 27591, 301; Netherlands 18694; Spain
310, 31223; external 359; exibilization 181, 183, 194,
205; functional 16, 1921; hidden 23; numerical 16,
201; and service economy 1025; skill 2357; temporal
17, 1921; Britain 13841; Germany 2335; Netherlands
1878, 199; wage 1617, 1921
exicurity 18694, 204, 206
ows 6970, 10910
FMP (exible mass production) 1819
Fordism 13, 18
Forslund, A. 44, 150n.
Foundation of Labour (Netherlands) 190
France 198, 314n., 3367; employment structure 48, 56, 679,
94, 25965; exibility dynamics 32, 589, 24959;
national debate and literature 2459; regulation: dynamics
of reform 3651, 57, 602; effects 748, 816; employ-
ment protection legislation 34, 55; unemployment 101,
348 Index
2216; and the Netherlands 186, 198, 201, 203; home-ownership 99n.
regulation: dynamics of reform 33, 3841, 44, 468, Hop, J. P. 185
501, 602; effects 69, 78n., 79, 81, 86, 105; employment household effect 103
protection 34, 55; unication 50, 2201, 228 see also West Houseman, S. 33, 86, 2324
Germany Hujer, R. 230
Gesamtmetal 227 hump-shaped curve 82, 84, 8991, 100
Giavazzi, F. 273 Hunt, J. 54, 230
Gibbons, R. S. 32 Hunter, L. 120
Giddens, A. 1201 HUS surveys (Household Market and Non-market activities,
Giersch, H. 212, 222, 230 1984, 1986, 1988) 173
Gissot, C. 256n. hysteresis 80
Giugni income agreement 278 IAB Arbeitsfrftgesamtrechnung 220
globalization 31, 120; and competition 83, 93 Ichino, A. 6, 38n., 45; and France 249, 251, 267; and Italy
Glyn, A. 76, 81 2759, 293, 300; regulatory effects 72, 82, 85, 889
Goetschy, J. 248 Ichino, P. 288
Gorter, C. 181207 IGM (German metal workers' union) 223, 225, 228, 2378
Gorter, D. 196 ILO (International Labour Organization) 118, 12730
Gosling, A. 128, 142 immigration 83
Gottschalk, P. 166 income support 33, 35, 276, 286
Granovetter, M. 11 incomes policy 252
Great Britain, see United Kingdom industry 267, 302, 337; employment 215; workers 2801 see
Greece 40, 46, 50, 86; employment protection legislation 34, also construction; manufacturing
55, 62; exibility 568; welfare system 601 ination 32, 165, 172, 298
Green Party (Sweden) 154 informal employment 103
Gregg, P. 71, 11819, 131, 138, 140, 2389 Inland Revenue (UK) 139
Groot, W. 196 INSEE (Institute National de la Statistique et les Etudes
Grubb, D. 58, 86, 119, 271 Economiques) 246, 253, 268
guilder, Dutch 203 insideroutsider theory of unemployment 23, 116, 194,
Gutchess, J. F. 20 3389; effects of regulation 67, 75, 85, 912; Germany
Haagen Pedersen, L. 77n. 2357 see also segmentation
Hakim, C. 120 institutional change 206, 267, 2779, 331
Harkman, A. 163 interest associations 22
Harteklauseln 225n.
Hartog, J. 185, 223n.
Harvey, M. 139
Hase, D. 231
Hassink, W. H. J. 1867
Hatch, N. 132
Hayek, F. A. 116
health care 16970, 285
Hemerijck, A. 188, 190, 192
Hemmer, E. 231
Heyden, P. F. van der 187
Hills, J. 120
hiring decisions and regulations 31, 38, 52, 85, 340; Britain
1234; France 24952; Germany 215, 21920, 235;
Italy 271, 2747, 2813, 2924, 300; Netherlands 198;
Spain 31215
Hirsch, B. T. 37
Hirschman, A. O. 211
Hland, A. 2312
Holmlund, B. 75, 158n., 165, 173
Index 349
Langager, K. 1601n., 172

International Adult Literacy Survey 213, 235 Lange, P. 21
international competitiveness 312, 102, 274 Larsen, R. B. 152, 159
International Organization of Employers study (1985) 86 Layard, R. 67, 716, 87, 91, 99; and Britain 119, 1312;
International Survey Research (1996) 294 Denmark and Sweden 1489, 1512, 157, 172
investment 26, 301 Lazear, E. P. 867, 271
Ireland< ; employment patterns 56, 68; exibility 589; Left Party (Sweden) 154
regulation 69n.; dynamics of reform 33, 368, 41, 468, Leonard, J. 207
51, 57; employment protection legislation 34, 55; and LET (Ley del Estatuto de los Trabajadores) 311n., 31213, 31516,
welfare system 602ISFOL (Italy, 1996, 1998) 282 31920
ISTAT (Instituto Statistico Italiano) 78n., 2712, 28991, Level of Living Surveys (1981 & 1991) 173
2959, 303 Lever, M. H. C. 185
Italy 198, 256n., 271303, 307, 336, 340; employment patterns Ley Bsica de Empleo 320
68, 945, 101, 105, 207; employment structure 291300; Ley de Proteccin del Desempleo (Unemployment Protection Law,
exibility 32, 56, 589, 27691; regulation: dynamics of LPD) 320
reform 389, 423, 458, 501, 57, 602; effects 67, 69, LFS (Labour Force Survey)< ; and Britain 119, 1278, 130,
748, 813, 856; employment protection legislation 34, 1326, 139; and Italy 272; and Spain 308, 3224, 326,
55; trends 27691; and unemployment 101, 105 see also 328, 3323liberal welfare regimes 334
Northern Italy; Southern Italy
Jackman, R. 44, 14950n.
Jacobi, O. 221
Japan 124; dynamics of reform 35, 37, 39, 41, 55, 5960;
employment ows 207, 293; exibility 16, 21, 567
Jehoel-Gijsbers, G. 196
Jensen, P. 76n.
Jimeno, J. F. 76, 81, 310n., 314, 318
job< ; creation 69, 156, 309; mobility 53, 198; Germany
21520, 239; security 31, 33, 297; tenure 94, 112n.,
1378, 198, 3235, 332; training 156, 161; turnover rates
69, 856, 162; Britain 1378; Italy 289, 303; Spain 325,
32930Jong, E. de 193, 203
Joseph Rowntree Foundation 118
JUVOS (Joint Unemployment and Vacancies Operating
Statistics System) 131
Kalisch, D. W. 57, 601
Kalleberg, A. 16, 20n., 21
Keller, B. 226, 229
Kjellberg, A. 167, 174
Kleinknecht, A. 1845, 187
Kock, U. 193
Kohl, H. 2256, 228
Kraft, K. 2323
Krasner, S. 191
Kreis, K. 222
Krueger, A. B. 37, 44, 82, 150n., 224
Kruppe, T. 389n.
labour< ; demand 7780; market 129; Germany 21221;
Netherlands 194203; policy 57, 1512, 251; supply 103,
1056Labour Party (Netherlands) 192
Labour Party (UK) 120, 124, 127
Labour Relations Act (Netherlands, 1945) 186
Lnder, Eastern 46, 50, 2289
350 Index
matching 201, 204, 215, 21718

liberalization 33, 46, 143 Maurau, G. 2501
liberal-labour government (Netherlands) 181 Mayes, D. G. 187n.
Lindbeck, A. 2, 73, 116, 153 Medoff, J. M. 211
Lindbeck Commission 153 Meidner, R. 148
Lindblom, C. 12 Meinhardt, V. 233
Lindeboom, M. 196, 198 Meldolesi, L. 293
Liso, F. 291 Mercier, M. A. 256n.
LO (Denmark) 159 Metcalf, D. 120
LO (Sweden) 148n., 158, 168 Micklewright, J. 43
Lockwood, B. 77n. Minford, P. 116, 118
Lois Auroux (France, 19823) 247, 249, 252 minimum wage 77, 88; Britain 11617, 11920; France 248,
Lnestatistik rsbok 173 2534; Germany 223, 229; Netherlands 1912; reform
long-term unemployment 51, 68, 70, 74, 76; Britain 120, 128, 412, 51, 59, 62; and unemployment 99n., 101, 10810;
131, 143; Denmark and Sweden 160; France 2623; and wage structure 804
Germany 212, 21617, 226, 230, 2389; Italy 272, Minimum Wage Act (1998) 124
2878, 2923, 298; Netherlands 182, 1956, 199, 2045; Ministerio de Trabajo y Asuntos Sociales 326n.
Spain 31415, 329, 331 Ministry for Employment (Germany) 228, 237
Lorences, J. 317 Ministry of Economic Affairs (Netherlands) 1845
low-skilled 3, 337, 341; Britain 142; Denmark and Sweden Ministry of Economics (Germany) 229n.
160; France 248, 254, 260, 265; Germany 213, 21617, Ministry of Labour (France) 255, 258
2234; Italy 271, 282, 292, 298, 300; regulation: effects Ministry of Labour (Italy) 274, 283, 288
on 68, 7783, 85, 8791; and unemployment 99n., 1012, Ministry of Social Affairs and Employment (Netherlands)
105, 10712; Spain 332 1856, 189, 198, 202
LSU (temporary jobs scheme) 2789 Ministry of Social Affairs (Italy) 287
Lubbers, R. 183 mobility, labour 53, 198; Germany 21520, 239
Lunghini, G. 274
Luxembourg 263
Lyon-Caen, A. 249, 251
Maastricht Treaty (1991) 100, 245
McGregor, A. 138
Machin, S. 118, 128, 131, 1412
McKinsey Global Institute 223
McLaughlin, E. 120, 133
Madrid 66, 311n.
Malcolmson, J. 77n.
male employment patterns ; Britain 129, 1347, 141;
Denmark and Sweden 151, 165, 175; France 268; Italy
272, 297; Netherlands 182; Spain 3323Malo, M. A. 74n.
; on France 24568; on Spain 30733
Malo de Molina, J. L. 309n.
Manning, A. 71, 77n., 128, 131, 142, 2389
manufacturing sector 45, 92, 105; France 260, 268;
Germany 216, 221; Italy 284, 294, 302; Spain 332
Marginson, P. 120
Mariani, P. 273
Marimon, R. 76, 81
Marquering, W. A. 185
Marsden, D. 20, 24
Marshall, T. H. 12
Martin, J. P. 99n., 222
Marx, I. 118, 141
Index 351
France 245n., 246, 2534, 2578, 260, 266; and

Moftt, R. 73 Germany 216, 2223; and Italy 2712, 283, 285, 289,
Mogensen, G. V. 75, 132 2923; Jobs Study (1994, 1997) 35, 47, 77; Literacy and the
Moghadam, R. 35 Economy 78n.; and the Netherlands 182, 1878, 192, 198;
Mller, J. 229 regulation effects 689, 74, 7882, 858, 94, 104; and
Moncloa Pacts 31516 Spain 329 see also Eurostat; International Adult Literacy
monetary policy 99100 Survey
Monopolkommission 2223 ffnungsklauseln 221, 225
Morris, G. 122 older workers< ; Denmark and Sweden 16970; Germany
Mosley, H. 389n., 256n. 2367; Italy 287, 292, 295, 297, 300; Netherlands 182,
Mot, E. 192 196, 204; Spain 31314, 3312 see also pension systems;
multi-national rms 54 seniorityOlini, G. 302
Mumford, K. 207 Olson, M. 15
Myles, J. 75n. Onofri Commission 288
NAIRU (non-accelerating ination rate of unemployment) OPEC oil crises 71n.
69n., 131, 230 opening clauses 221, 225
national debate< ; Denmark and Sweden 14855; France Origo, F. 302
2459; Germany 2216; Italy 2736; Netherlands Oudshoorn, C. 185
1835; Spain 30711National Institute of Economic Pact for Employment (1996) 284
Research (Sweden) 159 paid leave 1556, 161, 1712
National Labour Market Board (AMS) (Sweden) 151, 158, 162, Papeles de Economa Espaola (1997) 311n.
168 Parliamentary Inquiry Commission (Netherlands, 1993) 184
NAWRU (non-accelerating wage rate of unemployment) 69, participation rate 183, 272
84n., 87
Negri, N. 291
Netherlands 3368; Dutch Miracle 83n., 99101, 181207,
337; Dutch model 203, 228, 339; employment patterns
68, 945, 207; performance 5, 183, 194203; exibility
17, 32, 56, 589; exicurity 18694; national debate
1835; regulation: dynamics of reform 31, 3844, 478,
502, 602; effects 6970, 74, 83; employment protection
34, 55, 57; minimum wage 2534; and unemployment
99100
New Zealand 68, 142
Nickell, S. 59, 61, 99, 1434, 149, 213; regulatory effects 67,
716, 82, 87, 91
Nijkamp, P. 200
Nilsson, T. 168
Nolan, P. 120
non-standard employment, see atypical work
non-wage labour costs 48
Nordic countries, see Scandinavia
North America 66, 70, 101, 111
Northern Europe 75, 101
Northern Italy 272, 303; employment structure 2934, 2967,
299300; regulation trends 282, 28790
Norway 46, 679, 165
O'Connor, J. 12
OECD (Organization for Economic Co-operation and Devel-
opment); and Britain 119; Denmark and Sweden 157,
166, 168, 173; dynamics of reform 334, 3840, 43, 49,
52n., 5561; Employment Outlook (1998) 104n., 213; and
352 Index
and exibility 1114, 1217, 1878; contradictory effects

part-time employment< ; Britain 117, 1367, 141; France 3389; temporal 3940, Italy 2856, Netherlands 191;
24950, 256, 261, 268; Germany 2367; Italy 271, 274, Germany 22638; hiring and ring: exible labour
2812, 2946, 302; Netherlands 1812, 188, 193, 199, contracts 1878; France 24952; Italy 2813; institu-
204; and reform 368, 45, 48, 50, 56, 58, 62; Spain 319; tional 1112; and structure 3378; Italy 2914; trends
and unemployment 99n., 1045, 338Pedersen, L. H. 76, 468; Italy 27691; and rigidity/exibility dynamics 1217;
154, 159, 165 and unemployment 99112Rehn, G. 148
Pedersen, P. 734n., 83, 151 related benet systems 1557
Pels, E. 200 relief work (beredskapsarbeten) 156
pension systems 121, 167, 188, 236, 2867 Rentenreformgesetz (1992) 236
performance, labour market 31; comparative 6770; Ger- replacement ratios 43, 478, 612, 79n.; Denmark and
many 22638; Netherlands 5, 183, 194203; and Sweden 14950, 156, 163, 172; Germany 230; Italy
rigidities 703 287; Netherlands 192; Spain 321
permanent work 1367, 166, 3323 Requena, M. 328n.
Philips 190 reservation wage 222, 224, 293
Piketty, T. 260 Restart programme 132
Piore, M. 13, 1516, 18 retailing sector 289
Pischke, S. 224 retirement, early< ; Denmark and Sweden 151, 154; Germany
Pissarides, C. 35, 77n.
Pizzorno, A. 12
placement schemes 37
Polder Model, see Netherlands, Dutch Miracle
policy< ; active labour market (ALMP) 76, 101, 11112;
Britain 121; Denmark and Sweden 14850, 172; dynamics
of reform 34, 389, 447, 52n., 57, 62; Germany 220; Italy
288; failure 31; lessons from deregulatory reforms
33941; social security, Italy 2868; wage bargaining
systems 1559Pollert, A. 119
Polyani, K. 22
Poot, J. 205
Popular party (Spain) 311
Portugal< ; exibility 56, 589; regulation 689, 78n., 86;
dynamics of reform 38n., 40, 42n., 4650, 62; employment
protection 34, 55; and unemployment 1001; welfare
system 57, 601post-Fordism 15, 17
poverty 48, 303
Praag, B. van 185
productivity 152, 272
Programmazione negoziata 280
PTT Netherlands 1901
Public Employment Ofces (collocamento obbligatorio) 281
Public Employment Service (Spain) 325n.
Rapiti, F. M. 274, 289
redgreen coalition government (Germany) 222, 228, 232, 237
redundancy 37, 128, 132, 247, 280
Reed, H. 49, 11544
reform dynamics 3062
Regini, M. 17, 66n., 71, 33641; regulation dilemmas 1127
Regnr, H. 156n.
regulation< ; across Europe 1127, 3244, 4852; Euro-
sclerosis thesis 6695; changes 257, 62; in EU 4552;
Italy 294300; coverage 28891; and deregulation 215;
Index 353
2357; Italy 281, 292, 295, 298, 300; reform, dynamics of Schneider, H. 230
35, 379, 434, 478, 57, 62; regulation: and unemploy- Schmann, K. 193
ment 101, 106, 10912; winners and losers 76, 78, 84, schools 16970
91retraining 37 Schrder, G. 226, 238
Revelli, R. 207 Schusser, W. H. 236
Revisa Econmica de Catalunya 311n. seasonal work 36
Riester, W. 228 segmentation 21, 35, 338, 340; Italy 271, 2745, 292, 298;
rigidity 23, 14, 106, 337; dynamics 1217; France 24959; Netherlands 205; Spain 310
Germany 22638; Italy 273, 27591, 300; Netherlands Segrestin, D. 253
18694; Spain 307, 31223; and employment perform- Segura, J. 310, 323n.
ance 703; hidden 23, 25 Seifert, H. 226, 229
Rivero, J. 319n. self-employment 78, 338; Britain 136; France 261, 268; Italy
RMI (Revenu minimum d'insertion) 259 288; and reform 35, 38, 46, 56; and unemployment 101n.
Robinson, P. 128, 131, 138 , 103
Rogerson, R. 87, 89 Sengenberger, W. 12, 231
Rosdahl, A. 172 seniority 36n., 37, 86, 295
Rosholm, M. 162 SERPS (State Earnings-Related Pension Scheme, UK) 117, 121
Rostagno, M. V. 276, 286 service sector 4, 15, 19, 32, 923; Britain 137; and exibility
Rota, P. 291, 303 1025; France 248; Germany 21516, 221; home sector
Rothstein, B. 167n. 1545; Italy 2889; miscellaneous market 260, 2678;
Rotterdam 195 non-market 268; private placement 238; Spain 332; and
Rowthorne, R. E. 82 unemployment 1056, 110
Rubery, J. 120, 141 Sestito, P. 276
Russo, G. 1989, 202n., 215 Shadman-Mehta, F. 254
Ryan, P. 142 Shell 190
sabbaticals 1712 shocks 2, 40, 71n., 99102, 138; asymmetric 53, 340;
Sabel, C. 13, 15, 18 exogenous
Sachs, J. 100
Sachverstndigenrat 213, 221
SAF (Swedish employers' central associates) 50
Sagardoy, J. A. 319n.
Saint-Paul, G. 2, 46, 51, 169, 2512
Salvati, M. 274
Salverda, W. 191, 193, 203, 205
Samek Lodovici, M. 83n.; Italy 271303; reform, dynamics of
3062
Saraceno, C. 291
Sartor, N. 77n.
Saxony 222n.
Scala Mobile (wage indexation) 277
Scandinavia< ; employment performance 31, 70, 106;
exibility 56, 589; labour policies 39, 57; minimum
wage 834; regulatory approach 3644, 4950, 55, 62;
social democratic/corporatist regimes 334, 601Scar-
petta, S. 712, 745n., 76, 87
Scharpf, F. W. 13, 2234
Schellhaass, H. 225
Schettkat, R. 86, 202; on Germany 21139
Schleswig-Holstein 227n.
Schmid, G. 52n., 230n.
Schmitt, J. 142
Schmitter, P. 22
354 Index
92; ination 283; oil price 149; social 47; supply-side 30 3239; unemployment ows 945, 207, 293; exibility 33,
short-term contracts 325 56, 589, 31223; and Italy 272, 282, 292; national
short-time schemes 35, 38, 255, 257 debate 30711; regulation: effects 6770, 758, 816, 88;
Sickness Act, abolished (Netherlands, 1996) 192, 206 employment protection 34, 55, 31314, 319n.; labour
sickness benets 43, 49, 205 see also disability programmes policy 57; reform 3840, 42n., 458, 501, 62; and
Siebeck, K. 21617 unemployment 101, 103, 105, 198; social security 601,
Siebert, H. 72, 75, 85, 222, 230, 245 321
Siebert, W. S. 118 Spaventa, L. 273
Sinderen, J. van 1834 SPD (Social Democratic Party, Germany) 222, 226, 228, 232,
Sinn, G. 220n. 237
Sinn, H.-W. 220n., 223 Speckesser, S. 256n.
skill distribution 214, 223 Sproull, A. 138
skraldemandsmodellen (garbage-collector model) 154 stability, job 94, 2335, 303, 3236, 330
small and medium-sized rms (SMEs) 35, 38, 48, 103, 253; Statistics Denmark 1756
Italy 271, 289 Statistics Sweden 164, 173, 175
Smeeding, T. 166 statutory minimum wage, Spain (SMW) 320n.
SMIC (salaire minimum interprofessionel de croissance) 248, 251, 254 statutory sickness benet, Germany 2378
Smith, N. 74n., 83, 151, 165 Steiner, V. 22930
Smith, P. 138, 207 Stigler, G. J. 15
Sneessens, H. 254 Stigter, A. I. 184, 192
Snower, D. 23, 45, 78, 116 Stilgoe, E. 139
Social Democratic Party (Sweden) 1549, 1689 Stockholm 66
Social Economic Council (Netherlands) 189 Streeck, W. 1516, 18, 212, 256, 54
Social Pact (Italy, 1998) 27980, 285 strike action 11617, 2245, 227, 309, 311, 321
social partners 336, 33940; Italy 276, 300; Spain 31011, structure 150, 274; regulation affects 91,
314, 330
social plan (France) 250
Social Security Council (Netherlands) 197
social security system 60; Britain 127, 1312; Italy 2868;
Netherlands 183, 1923, 205; and unemployment
benets 424
Social-Democratic/Green government (Germany) 222, 228,
232, 237
Socialist Party (France) 247, 249
Socialist party (Spain) 30910
Socialkommissionen (Denmark) 151
solidarity contracts 2856
Solow, R. 11
Sorensen, B. 207
Soskice, D. 3, 78, 82, 235; dilemmas of regulation 17, 26;
dynamics of reform 32, 41
Soteri, S. 187n.
SOU (Sveriges Offentilga Utredningar, 1998) 153, 159
Southern Europe 84, 337; dynamics of reform 3640, 42, 44,
46, 48, 503; conservative regimes 33, 35; employment
protection 55, 62; exibility 569; tax and contributive
wedge 601; single breadwinner model 56
Southern Italy 2712, 301, 303; employment structure 2934,
296300; national debate 2745; regulation trends 278,
2845, 28790
Soviet Union 339
Spain 30733, 33640; employment patterns 68, 94; structure
Index 355
3378; wage levels: differentials 1413; and minimum tripartite concertation 234, 310
wage 804 Tronti, L. 2912, 303
students 199 Tuelings, C. N. 192
Summers, L. H. 88 Turatto, R. 282
sunk costs 24 turnover rates 69, 856, 162; Britain 1378; Italy 289, 303;
supply-side economists 183 Spain 325, 32930
SVIMEZ (Rapporto 1996 sull'economia del Mezzogiorno) TWAs, see temporary, work agencies
274, 302 UB, see unemployment, benets
Sweden 143, 14877, 263, 336; employment patterns 679, UGT (Unin General de Trabajadores) 309, 319
945, 174, 177; wage setting 59, 173; exibility 32, 56, 58; UI (unemployment insurance), see unemployment, benets
labour market policy 13, 14877; dynamics of reform underground economy 48, 101n., 274, 303
36n., 39, 467, 49, 57, 62; effects 78n., 80, 82, 86, 101n.; UNEDIC (unemployment insurance funds) 257
employment protection 34, 55; national debate 14855; unemployment< ; benets/insurance: Britain 1267, 131;
unemployment 1624; unions, role of 1668; wage Denmark 14952, 1557, 169, 1712, recipients and non-
bargaining systems 1559; wage dispersion 1646; social recipients 160, 1624, 177; duration 79, 230, 338,
security system 601 Denmark and Sweden 14950, 1556, 172, and unemploy-
Swedish Economic Association 148n. ment 100, 105, 10810, 2301; France 2579, 266; Italy
Swedish Employers' Confederation (SAF) 158 295, 303; Netherlands 188, 192; reform 33, 39, 424, 47,
Swedish Engineering Employers' Association (VF, later VI) 612; regulation effects 7380; Spain 312, 314, 3203, 330;
158 Sweden 14952, 1557, 169, 1712, recipients and non-
Swedish Metal Workers' Union (Metall) 158 recipients 160, 1624, 177; ows 90, 95, 109, 177, 218;
Swedish Trade Union Confederation (LO) 148n., 158, 168 hidden 192, 219,
Switzerland 69, 86
Taddei, D. 255
Tariffond 237
tax and contributive wedge 42, 48, 60, 7780; France 248;
Italy 285; Netherlands 184, 1923; and unemployment
101, 104, 10611
Taylor, M. 128
technological innovation 312, 71, 80, 153
temporary< ; contracts 20, 378, 446, 50, 86, 338; Britain
136, 138; Denmark and Sweden 1567; France 24951,
267; Italy 274, 295, 2989; Netherlands 193, 199, 2045;
Spain 310, 31314, 318, 32533; dismissals 38; work
agencies (TWA): France 251; Italy 274, 281, 302;
Netherlands 1878, 193, 2045tenure, job 94, 112n.,
1378, 198, 3235, 332
Teulings, C. N. 185, 223n.
Thatcher, M. 489, 115
Theeuwes, J. J. M. 196, 198
Thoursie, A. 162, 177
Tiden 148n.
Toharia, L. 66n., 74n., 76, 81; on France 24568; on Spain
30733
Toren, J. P. van den 188, 190, 192
Torng, J. 168
trade unions, see unions
trade-offs 267, 31
training 5, 26, 154, 251, 341
transport and telecommunications 260, 268
Trentino-Alto Adige 272
Treu Package (1997) 276, 278, 2823, 286, 291
356 Index
221; and regulation 8490, 99112, 3379; structural 69, 2023, 205; Spain 312, 31519; differentials 47, 101,
85, 912, 99, 149; structure 23, 67, 8990, 101, 10812, 1413, 1534, 31819; dispersion 173, 223, 253, 284;
337; Britain 12732; Denmark 1602, 172, 1756; France distribution 834, 149, 169; exibility 229, 253, 2756,
2457, 25966; hysteresis 13; Italy 272, 291300; Nether- 298300; growth 141, 253; moderation 51, 107, 181,
lands 183, 1947; Spain 3079, 318, 3239, 333; Sweden 183, 272 see also minimum wageWage Determination Act
1624, 172, 1756; West Germany 21215 see also insider- (1970) 189
outsider theory; long-term; low-skilled; youth unemploy- Wagner, J. 223
mentUnemployment Insurance Act (Netherlands, 1987) wait unemployment 2923
192 Wall Street Journal 20
Unilever 190 Walwei, U. 32, 238
unions 57, 24, 402, 82; bargaining power 32, 49, 274; Wareing, M. 140
Britain 116, 142; centralization 901; coordination 149; Wassenaar accord (1982) 51, 18990, 203, 206
coverage 59; Denmark 5, 1524, 156, 1669; density 41, Weise, H. 161
59; France 250, 2523; Germany 2213, 2268, 237; welfare schemes 312, 35, 42, 73, 184; and unemployment
Italy 273, 280, 2856, 292; membership rates 167, 174; 7380
Netherlands 5, 1889, 191, 194, 203, 205; recognition
120; Spain 30911, 31518, 320; Sweden 1524, 156,
1669; unionism 12, 83, 316; unionization 32, 412, 211
United Kingdom 33, 57, 11544, 239, 3368; employment
patterns 68, 94; ows 95, 293; structure 12743;
exibility 16, 56, 589, 117, 1217; regulation: effects
6770, 75n., 7780, 823, 85; employment protection 34,
55; reform 313, 36, 3849, 512, 62; and unemployment
99101n., 104n., 105; tax and contributive wedge 601
United States 165, 201, 260, 289, 336, 338; and Britain 115,
125; employment patterns 11, 68, 94, 21316, 234; ows
95, 207, 293; exibility 1, 3, 16, 23, 38, 56; and Germany
212, 21415, 22930, 233, 235; regulation: effects 6670,
736, 7881, 83, 86; employment protection 34, 55;
reform dynamics 3540, 437, 52, 57, 62; and unemploy-
ment 1024n.; tax and contributive wedge 601; wages
59, 69, 82, 142, 223
Utili, F. 276, 286
utilities 137
vacancies 200, 218
Vlfrdskommissionen (Denmark) 151
van Veen, C. 189
VAT (value-added tax) 48
Vijlbrief, J. A. 185, 1923
Visser, J. 192, 199
Vocational Training Fund 283
Volkswagen (VW) 227, 2335
von Weiszcker, C-C. 222
de Vries, B. 185
van Vuuren, A. P. 185, 2023
Wadsworth, J. 118, 138, 140
Wadwhani, S. 120
wage< ; bargaining 31, 33, 82, 148; compression 76, 82, 87,
223, 22930, 254; councils 49, 116, 142; determination
59, 62, 73; Britain 1245; Denmark and Sweden 152,
1559; France 2524; Germany 212; and industrial
relations systems 402; Italy 2835; Netherlands 18991,
Index 357
see also social security

Wells, W. 58, 86, 119, 271
West Germany 229; industrial relations framework 225, 228,
236; labour market 21220, 238
Westergaard-Nielsen, N. 73
Wijnbergen, S. van 185
Wijngaert, R. van de 185, 189, 194
Wilkinson, F. 20, 120, 126
Williamson, O. 12
women 3, 46, 815, 92, 1018, 3378; in Britain 117, 130,
1357, 1412, 144; in Denmark and Sweden 149, 151,
161, 175; in France 256, 2601, 268; in Germany 236n.;
in Italy 2712, 274, 276, 289, 293, 295, 2978, 300; in the
Netherlands 1812, 1989, 204; in Spain 329, 3323
Wood, A. 80, 82, 223
Wood, D. 138
work entry, see hiring
work incentives 43, 737
Workers' Statute (Italy, 1970) 274, 280
Workers' Statute (Spain, 1980) 311n., 312
working time 12, 33, 512, 58, 62; Britain 117, 119, 1256,
13841; Denmark and Sweden 1545; France 247, 250,
2547, 267; Germany 225, 2315; Italy 3940, 2856,
300; Netherlands 191, 205; Spain 312, 31920
Working Time Act (Germany, 1994) 235
workless households 11819, 128
Works Constitution Act, German (BetrVG, 1952, 1972) 226
works councils 24, 224, 227, 250, 267, 316
Wyplosz, C. 202, 207
youth unemployment 3, 51, 33740; Britain 120, 143; causal
mechanisms 1012; Denmark and Sweden 169; employ-
ment protection 85, 879; France 251, 2613, 267;
Germany 2356; Italy 2716, 282, 289, 292, 2958, 300;
minimum wage 803; Netherlands 1912; regulatory
effects 68, 779, 901, 95, 99n., 1056; Spain 309, 313,
32930, 332; unemployment structure 10812; work
incentives 74
Zeli, A. 273
Zetterberg, J. 150n., 165
Zimmermann, K. F. 230, 238n.
Zukunftskommision der Friedrich-Ebert Stiftung 2234
Zunkunfskommission 222n.

Gosta Esping-Andersen

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Gosta Esping-Andersen

Uploaded by

Copyright:

Available Formats

Why Deregulate Labour Markets?

This page intentionally left blank

8 Germany: A Regulated Flexibility Susanne Fuchs and Ronald Schettkat 211

2A Employment protection legislation: evolution in the 1990s 55

5.5 Redundancies and redundancy rate 132

Gsta Esping-Andersen and Marino Regini

high-paying manufacturing jobs; a sectorally divided, two-class labour market?

Labour Markets Between Regulation and Flexibility

Types of Labour Flexibility: Trade-Offs and Contradictions

The Varying Role of Flexibility in Companies' Strategies

Contradictions Between the Different Types of Flexibility

Regulatory Regimes and Deregulation in Advanced Economies

Changes in Regulatory Regimes and Labour Market Policies

Manuela Samek Lodovici

Regulatory Approaches in European Countries in the Late 1980s

Employment Protection Regulation and External Flexibility

Working Time Regulation

Wage Determination and Industrial Relations Systems

Social Security and Unemployment Benets

thereby attenuating the pressure applied by outsiders for wage moderation.

Recent Changes in Labour Market Regulation in Eu Countries

Common Trends in the Regulatory Framework

Differences Across European Countries

Some Preliminary Conclusions

Table 2A. Employment Protection Legislation: Evolution in the 1990s

Table 2C. Incidence of Accompanying Labour Policies, 19961997

Table 2E. Wage-Setting: Main Trends Since the Early 1980s

Table 2F. Tax and Contributive Wedge, 1985 and 1994

Table 2G. Unemployment Benets

Table 2H. Summary of Action Taken to Change Regulation in the 1990s

Comparative Labour Market Performance

The Links Between Employment Performance and Rigidities

The Welfare State and Unemployment

Table 3.2. Unemployment and Social Benets

Wage Structure and the Minimum Wage

Table 3.3. Unemployment and Wage Structure

Employment Protection and Unemployment

Identiable Effects of Employment Protection

Table 3.4. Unemployment, Wages, and Labour Market Regulation

Table 3B. Indicators of Employment Stability

Table 3C. Unemployment Flows

Re-Thinking the Causal Mechanisms

Flexibility and the Service Economy

Table 4.1. Causes of Women's Employment

negative trend in personal service jobs, as indicated in the following formula:71

Demographics and Labour Supply

A Quantitative Reassessment of the Issue

Table 4.2. Explaining Cross-Sectional Employment Growth, 19791994

Table 4.3. Explaining Unemployment Levels, 1996

Explaining Unemployment Structure

Table 4.4. Explaining Unemployment Outows, 1994

Table 4.5. Explaining Unemployment Structure

Simon Deakin and Hannah Reed

Review of the National Debate and Literature

Table 5.1. Legislative Developments Affecting Labour Market Flexibility, UK

Table 5.2. Workless Households 19771998

The Dynamics of Rigidity/exibility: Trends in Labour Market

Employment Protection Legislation

Hiring and the Form of the Employment Contract

Unemployment Compensation and the Tax-Benet System

The Outcomes: The Structure of Unemployment and Employment

Table 5.3. Labour Market Status Data

Table 5.5. Redundancies and Redundancy Rate

Table 5.6. Epi Joblessness Indicators, 19841997