Technology transfer via FDI and absorptive capabilities

Johannes Stephan, Institut Bergakademie Freiberg fr Wirtschaftsforschung Halle und TU

Introduction and motivation of analysis

When discussing Foreign Direct Investment (FDI) as an engine for economic growth, we are concerned with the conditions that make foreign affiliates an active part in the advancement of technology in the host economy. A large body of literature has emerged on those conditions and typically distinguishes between direct technology transfer or spillovers between foreign investor and foreign affiliate on the one hand and indirect effects between the foreign affiliate and its host economy business environment. Thereby, direct technology transfer and spillover may be conceptualised as a precondition for the indirect effects on the host economy. Amongst the many conditions for the direct knowledge and technology effects of FDI discussed in the literature (e.g. investment motives, investment strategies in general, parent-foreign affiliate relationships, institutional environment, etc.), we here focus on absorptive capacities of the foreign affiliate. The central role of this condition for our overarching question becomes particularly apparent if we consider the example of a foreign investor from e.g. West Europe who seeks to invest into a firm in East Europe: only if this foreign affiliate has sufficient knowledge about and experience with the technology used by the foreign investor, will the foreign affiliate be able to make active use of the foreign knowledge and technology and thereby contribute to spur productivity growth in the host economy. In the adverse case, a foreign affiliate with little or no absorptive capacities will solely have an effect on employment and income growth (if at all, which obviously depends on the extent of the replacement effect of the foreign investment). Many more conditions are certainly important, but previous research does suggest that it is mainly this factor that proved to be a particular bottleneck in Central and East Europe (Dyker et al., 2006, pp. 79-82).1 In general, we may assume that the potential for direct technology and knowledge transfer and spillovers will increase with the ability of the foreign affiliate to absorb the knowledge and technology that the foreign investor has. Girma and Strobl (2005), using UK establishment data, find that absorptive capacity matters for productivity spillovers and suggest a U1

Needless to say that one further pre-condition pays an all important role here, and that is the strategy of the foreign investor with regard to the extent to which the investor allows its foreign affiliate to tap its technology or knowledge- related ownership advantage. However, we assume that even in the most extreme case of technological marginalisation of an affiliate by its investor, the foreign affiliate will be able to benefit from its links to the foreign firm with more a advanced or different knowledge stock.

shaped relationship between productivity growth and FDI interacted with absorptive capacity. In contrast Casttellani and Zanfei (2003) as well as Barrios and Strobl (2002) find no significant evidence for a positive relationship between absorptive capacities and spillovers. Hence our focus on the conditions for technology transfer and spillovers: we assume that absorptive capacities are amongst the set of necessary conditions without going as far as to hold that there is a pre-determinable and general relationship between capacities and effects. The dimension of absorbing foreign knowledge and technology originating from a foreign investor incorporates many meanings and specifications in the literature. At the most general level, absorptive capacity or capability (here: of a foreign affiliate) is typically defined as the firms ability to recognize valuable new knowledge and technology, to integrate it into the firm, and use it productively. Cohen and Levinthal (1989 and 1990) argue that Research and Development (R&D) stimulates not only innovation but also increases a firms absorptive capacity. R&D helps to develop firms ability to identify, assimilate, and exploit outside knowledge which is likely to increase the incident of technology diffusion. In a slightly changed version of the original idea of the two faces of R&D, Lane and Lubatkin (1998) conceptualise absorptive capacities as being a relative concept rather than an absolute firmspecific measure, in as much as capacities will depend on the similarity of organisations experiencing interorganisational learning. Next to the use of the traditional Cohen and Levinthal-characterisation of absorptive capacities (e.g. Barrios et al., 2003), research has focussed on absorptive capacity as a function (also) of human capital (e.g. Keller, 1996; Borensztein et al., 1998; Kinoshita, 2000; Kneller, 2002; Schoors and van d. Tool, 2002; Todo and Miyamato, 2002). Damijan and Knell (2003) use investments into intangible assets, and Blalock and Gertler (2004) apply an even more comprehensive approach controlling for human capital endowment as well as the gap between productivity levels of domestic and foreign firms. Finally, the firms organisational structure and combinative capabilities have been found to contribute to a firms absorptive capacity (Van den Bosch et al. 1999). In this empirical analysis, we use both the R&D and the human capital aspects to proxy the level of absorptive capacity. In this analysis, we make use of a unique firm-specific database on foreign investment affiliates in a selection of Central East European countries, generated by fieldwork in the framework of an EU 6th Framework Programme research project U-Know. We use this data to provide answers on two separate research questions: the first is concerned with the foreign investors interest in using existing and developing absorptive capacities of subsidiaries in our selection of post-socialist economies. This is analysed under particular consideration of the foreign investors strategic motive in the investment. The second question compares the potentials of each country in our selection of post-socialist economies to benefit from a dynamic technology transfer between parent and foreign affiliate that goes in both directions. The latter is 2

measured in terms of a combination of two firm-specific factors, the foreign affiliate-performance in autonomy in business functions and in absorptive capacities. Here, a theoretical taxonomy is generated by reviewing literature and plausibility assumptions and then tested for the subsidiaries in our database. The central idea of this analysis is to link absorptive capacities and levels of autonomy from parent investors to the potentials for technology transfer, here in particular for a dynamic, two-way knowledge transfer between foreign affiliate and parent. The article starts with a brief description of the data used in the analysis. Following a discussion of the strategic investment motives of and for foreign affiliates, chapter 3 develops a four-quadrant theoretical taxonomy, describing the assumed potentials for direct technology transfer as determined by autonomy in business functions and absorptive capacities. This is then applied empirically by use of the firm-specific data and discussed in terms of the resulting potentials for direct technology transfer for each of the host countries. The criteria we use here are the endowment of countries with the different performance of foreign affiliates along the two coordinates autonomy and absorptive capacities. The concluding chapter summarises the results and discusses the shortcomings and caveats of the analysis.

The data used in the analysis

For this kind of analysis, information at the level of the foreign affiliate is needed. In late 2002, early 2003, a European consortium, financially supported by the EU in its 5th Framework Programme, conducted coordinated field work projects interrogating manufacturing foreign affiliates in a selection of Central and East Europe: Estonia (with information on 73 foreign affiliates), Slovenia (69), Poland (153), Slovakia (78), Hungary (85). In 2007, a follow-up on the first wave was conducted, again financially supported by the EU in its 6th Framework Programme, this time focussing on East Germany (with information on 295 foreign affiliates), Slovenia (40), Poland (110), and Romania (220). Without financial support, Croatia also joined our field work and collected 144 filed-out questionnaires. The data used in this analysis is exclusively from the second wave the first wave served but as a pilot study for further research at a later stage. In both waves, the questionnaires were centrally designed, whereas the implementation of the surveys were decentrally organised. All firms from the population were approached in written form or by phone and invited to participate in the survey. Firms received the questionnaire by post, fax, or as an electronic version. In Romania, due to the large size of the population, a random sub-sample was drawn from the population. Here, all firms in the subsample received the questionnaire by post, and the interviews were realised face-to-face. In East Germany, all firms from the population were contacted by phone and invited to take part in the survey. Most interviews were realised directly by phone, only few foreign investors affiliates preferred to fill in the

questionnaire in its written form. The project deliberately allowed the country teams to choose the most appropriate method and timing individually. For East Germany, a complementary survey (also conducted by the IWH and using the same questionnaire) focussed upon West German multinationals that own at least 10 per cent of equity and/or are the ultimate owner of a legally independent firm located in East Germany. As a multinational enterprise, these firms are industrial firms that undertook FDI in at least one country outside Germany.2 The total population of our second wave field work consists of foreign affiliates located in Croatia, Slovenia, Poland, Romania and East Germany in 2006. The selection of these regions in economic transition tries to balance country size, geographic location, and level of economic development. This population includes different types of foreign investors apart from industrial companies such as mutual and pension funds, banks, foundations, individuals/families, or any combination of these different types of owners. The populations of foreign affiliates in Poland, Romania, and East Germany were drawn from the Amadeus database. The Polish and East German populations were supplemented with data from the respective foreign investment agencies (Invest in Germany IIG, Invest in Poland PAIZ). The East German population data was additionally supplemented with information from the European Investment Monitor (2006) and EUs industrial R&D investment scoreboard (2005).3 The Croatian population of foreign affiliates was compiled using information provided by the Institute for Business Intelligence (Zagreb/Croatia). In the Slovenian case, the population was drawn from statistics provided by the Bank of Slovenia. The cumulated total population across the five host economies of our analysis consist of 6.833 manufacturing firms with foreign investors employing about 1.1 million employees (see Table 1: Country composition of total population).

From our point of view, it is important to include West German investors in East Germany, because these investments were made in a transition region just as well and hence compare well to the investments made into the other transition economies. For a detailed description of the East German sample see also Gnther, Stephan, and Jindra (2008).

Table 1

Country composition of total population of manufacturing firms with foreign investors

Nr of FAs in % of total population 16.74 5.61 23.21 3.38 51.07 100.00 Employmen t 195,429 40,791 211,055 56,033 550,361 1,053,669 in % of total population 18.55 3.87 20.03 5.32 52.23 100.00

East Germany Slovenia Poland Croatia Romania Total

1,090 365 1,511 220 3,325 6,511

Notes: FA denotes foreign affiliates. The extraordinarily large share of Romanian FAs in the total population is probably distorted upward: even the latest editions of the Amadeus database include firms that have ceased to exist. This was particularly problematic in the Romanian case, the databases for the total populations of the other countries were cleaned more thoroughly. Source: The IWH FDI Micro database

The resulting total sample of foreign invested companies in the second wave holds information from 809 enterprises that account for an employment of over 214.000 employees across the five countries surveyed (see Table 2: The IWH FDI Micro database country composition and response rates). In terms of number of foreign affiliates, the sample constitutes 12.43 per cent of the total population, and in terms of employment 20.33 per cent respectively. The sample response rates vary across countries. In terms of number of firms, it ranges from 6.62 per cent in Romania to 65.45 per cent in Croatia. In terms of employment, it ranges from 16.22 per cent for the Romanian sample to 65.97 per cent in Croatia. Such differences in response rates are mainly explained by the differences in the size of the respective populations of foreign affiliates and the financial means available to collect filled-out questionnaires. The larger populations of Romania, Poland and East Germany hence tend to reduce response rates. In the cases of Romania and Poland, this resulted in a underrepresentation within the sample both in terms of numbers of foreign affiliates and employment. Particularly noteworthy is also that the response rates for Poland and Romania differ strongly between the response rates according to the number of foreign affiliates (much lower) and to the number of employment (much higher). This indicates that the foreign affiliates in the sample are in fact larger than in the respective countrypopulations and we have a bias towards large firms in those two countries. In fact, the total sample is underrepresented for micro (1-9) and small enterprise (10-49), and consequently overrepresents medium seized (50 - 249) and large (above 250) firms.

Table 2

The IWH FDI Micro database country composition and response rates
Nr of FAs in % 36.46 4.94 13.60 17.80 27.29 100.0 0 Employment 39,876 9,686 38,408 36,963 89,292 214,225 in % 18.61 4.52 17.93 17.25 41.68 100.0 0 Response rate for Nr of FAs 27.06% 10.96% 7.28% 65.45% 6.62% 12.43% Response rate for employment 20.40% 23.75% 18.20% 65.97% 16.22% 20.33%

East Germany Slovenia Poland Croatia Romania Total Source:

295 40 110 144 220 809

The IWH FDI Micro database.

Derived from the above said, the target of our analysis here is focussed upon affiliates of foreign investors in our selection of Central East European host economies. The database only includes manufacturing industries. No explicit distinction was made in any of the country-specific subsamples between affiliates of foreign investors with or without further affiliates in other foreign host economies. A foreign affiliate is defined as a legally independent enterprise with a foreign equity participation of at least 10 per cent and/or an ultimate owner located abroad. In principal, no restriction in terms of firms size was introduced, yet, the Croatian and Romanian participants decided to include only foreign owned firms with a minimum of 10 employees into their population. In general, the samples generated in the second wave are representative of the total population in terms of NCAE-2 industrial branches. In the East Germany case, representativeness was also positively tested for the regional aggregates of Raumordnungsregionen (ROR) (which is a functional aggregate of 97 regions of sizes between NUTS2 and NUTS3) and for the size of foreign affiliates. As formerly centrally planned economies, the countries and regions analysed here were virtually closed with respect to foreign investment from the West before 1989 (Dunning 1993, Hunya, 1997). Our dataset hence consists predominately of investments undertaken after 1989. In fact, only eight investments in our dataset were made before 1989 (of which are five into East Germany in the second half of the 1980s). Amongst our countries, Slovenia was probably the most protective of FDI and nevertheless had accumulated an FDI stock of nearly 9.5 per cent of GDP by 1995 and nearly 21 per cent by 2003. Romania started to receive noticeable amounts only after 1996 and accumulated a comparable stock by 2003. Poland experienced large FDI-inflows right after the start of the transition process whilst Croatia started to receive noticeable amounts only after 1995. Still, Croatias stock per GDP in 2003 already surpassed that of Poland with nearly 32 per cent and

nearly 24 per cent respectively.4 Of course, Poland and Romania are much larger economies than Croatia and Slovenia with Croatia and Slovenia around half the size of Romanias GDP and around one eighth of Polands GDP.

Investors strategic motives, subsidiaries absorptive capacities, and the local knowledge base

An investor who places particular weight on access to local knowledge, skills, and technology in his strategic motive may either hope to find such technological capacities in the foreign affiliate he chooses for his investment or may plan to develop the foreign affiliates capacities. In developing the foreign affiliate, the foreign investor may either use his own network or make use of the local knowledge base or innovation system. The analysis proceeds in three steps. In a first step, we want to find out what strategies are dominant for the foreign investments in our selection of post-transition economies. Of the five possible strategic motives at the time of the initial entry that we asked foreign affiliates, access to new markets or the increase of an existing market share turn out be the dominant strategic motive in all economies but in Slovenia (see Graph 1: Averages of strategic motives for foreign investment at the time of entry). Here, the motive to increase efficiency across the foreign owner network is slightly higher than the host market motive. Our main interest here is in the strategic motive to tap the local host economy knowledge base in terms of existing knowledge, skills of personnel, and available technology. This motive appears to have been important in East Germany and Romania (second only to the market motive) and rather low in the ranking of the five in Poland. This picture does not change much in the strategic motives that prevail today.

Whilst Slovenia and Croatia both belonged to former Yugoslavia and therefore had more elements of a market economy than in our two other countries, Croatias economic development and systemic transformation was retarded significantly during war time.

EDE
4

HRO

PL

RO
4

SLO

Access new markets or increase market share Increase efficiency in foreign owner network Access local knowledge, skills, and technology

Follow key clients Access local natural resources

Graph 1

Averages of strategic motives for foreign investment at the time of entry

Note: EDE...East Germany, HRO...Croatia, PL...Poland, RO...Romania, SLO...Slovenia. Source: The IWH FDI Micro database.

The second step asks whether the foreign affiliates that place particular weight on access to the local knowledge base are correspondingly also characterised by high absorptive capacities? In order to answer this, first, a proxy for absorptive capacities has to be developed. Then, a correlation analysis will provide some answers to the question. In the database, we have information about absorptive capacities which is derived from the literature: own R&D (Cohen and Levinthal, 1990). This follows the idea that R&D typically has two effects on the investing firm: first, it provides the basis for own technological development (e.g. in the form of inventions and innovations), and second, it raises the firms ability to adapt technology developed elsewhere to suit its own particular environment. The data we have generated in the field work does contain information that we may use to proxy absorptive capacities of local subsidiaries: R&D or innovation-related expenditure in % of total sales 2005 and Share of R&D personnel in 2005. This information was directly interrogated from the managers of the foreign affiliates and cross-checked with firm-financial databases where this was possible (using the Amadeus database). We construct a common indicator between R&D employment and R&D expenditure: this way, absorptive capacity is high, where both expenditure and personnel for R&D are high simultaneously, and vice versa. Where firms employ a large number of R&D personnel yet do not spend a lot in terms of share of R&D expenditure in total sales, or where large expenditure is not 8

matched by a large share of specialised R&D personnel, we assume absorptive capacities to be less developed. Furthermore, because both knowledge-related proxies can be expected to be quite industry-specific with large variations between industries as a typical feature, we here use firm-specific deviations of our composite indicator from the average of the indicator in the respective 2-digit NACE industry. This way, we get positive and negative firm-specific deviations, and their magnitude reflects the extent to which the foreign affiliate has a higher or lower absorptive capacity in comparison to what is typical in the firms own industry. At the country-level, negative and positive deviations of all foreign subsidiaries are averaged to provide a picture of the typical endowment of each country with absorptive capacities amongst firms (see Graph 2: Averages of deviations from industry-means of absorptive capacities).
15 -5 0 5 10

EDE

HRO

PL

RO

SLO

mean of adapt_dev

sd of adapt_dev

Graph 2:

Averages of deviations from industry-means of absorptive capacities

Note: EDE...East Germany, HRO...Croatia, PL...Poland, RO...Romania, SLO...Slovenia. Source: The IWH FDI Micro database.

It is nor a surprising result that the average of absorptive capacities amongst East German foreign affiliates is highest5: after all the proxy is measured in R&D related indicators of the firms, and here, East Germany can be expected to house firms that are technologically more active than in the other countries of our sample. At the very least, this reflects the typical association of levels of economic development, measured in GDP per capita
5

Noteworthy, however, is the observation that standard deviations are much higher in the case of East German foreign affiliates, suggesting that net of deviations from the average over all foreign affiliates hides significant heterogeneity.

(or average technology-level of the country, measured in aggregate productivity) and the intensity of technological activity amongst firms. From related research, however, we know that the East German National Innovation System is furthest developed amongst all transition economies or regions, and here in particular the state of transformation of research institutions (universities, public and private research institutes, like the Fraunhofer, the Max-Planck, the Leibnitz-list) (see e.g. Gnther et al., 2008). That the level is also comparatively high amongst Romanian foreign affiliates may be a statistical phenomenon: what counts as R&D expenditure and R&D personnel depends on each countrys definition in the law. The net deviation of absorptive capacities is negative for all other countries, the lowest for Polish and Slovenian foreign affiliates. To analyse the relationship between strategic motives and absorptive capacities, a correlation analysis amongst the firms with the strategic motive today of tapping the local knowledge base and the absorptive capacity of the foreign affiliate achieved today turns out to be weakly positive with a coefficient of 15.3 and significant at the 1 per cent level. Apparently, there is some association between the motive geared towards the local knowledge base and the absorptive ability of the foreign affiliate. But, because we can only use the association between the strategic motive and the absorptive capacity that the foreign affiliate displays today (there is no data on absorptive capacity at the time of entry of the foreign investor), we cannot tell whether the absorptive capacity was available right from the start or whether it was developed by the foreign investor after the direct investment took place. We are able, however, to test in a third and final step whether the most important sources for subsidiaries absorptive capacities either lie in own R&D by the foreign affiliate itself, or rather lie in local sources of knowledge and technology, or finally lie in technology-sources provided by the foreign investor and its network. In case of the latter, we can interpret that absorptive capacity is developed by the foreign investor, i.e. we diagnose direct technology transfer. Where we find a close association between absorptive capacity and local sources of technology, we have established some indication of indirect technology transfer via a functioning local and yet internationalised innovation system, i.e. a well-aligned network for technological development. A close association between absorptive capacity and own R&D as its source indicates that foreign investors appear to rely to some extent on the technological abilities of their foreign subsidiaries. Because we do not want to pre-assume a particular direction of causality, we refrain from using a regression analysis with absorptive ability as the dependent variable and the sources listed above as determinants. Rather, we test correlations between each source of technological knowledge for R&D and innovation at the foreign affiliate and the foreign affiliates absorptive capacity. The results we receive using the complete sample of foreign affiliates suggest that internal own sources of technological knowledge are in 10

fact positively correlated with absorptive capacities. The correlation coefficient is 37.7 and is significant at the 1 per cent level. Local host economy sources (R&D carried out with local suppliers and with local scientific institutions) both also turn out to be positively correlated with coefficients at 15.6 and 29.5 respectively and both significant at the 1 per cent level. No significant correlation could be established for local customers, though. Sources that lie within the network of the foreign investor (R&D carried out at the headquarter of the foreign investor network) are not significantly correlated with absorptive abilities. Across all subsidiaries of our sample, we conclude that where foreign affiliate-technological development takes place, this apparently has no bias on the foreign parent network but rather is sourced mainly from within the foreign affiliate and from the local innovation system. In terms of the requirements for absorptive capacities, we conclude that they appear to originate from the foreign affiliates internal and local sources. This underlines the importance of the local environment and innovation network for foreign investments in our post-transformation economies. Broken down by each post-socialist economy of our sample, the results vary slightly: internal foreign affiliate-own sources of technological knowledge and local scientific institutions are not relevant for Slovenian subsidiaries absorptive capacities (though the number of observations is low at 30 and 28 subsidiaries respectively and hence not very robust). The relevance of local suppliers for the development of absorptive capacities is also not significant for East Germany (here with a total number of 83 subsidiaries). In the case of Croatia, the parent headquarter as a source of technology for absorptive capacities is in fact significant at the 1 per cent level with a total number of 90 foreign affiliates in the analysis).

Comparative analysis of potentials for dynamic technology transfer

The overarching research question of this contribution is concerned with a cross-country comparative perspective: which of the countries appear to contain the largest potentials for a dynamic, two-way technology transfer via FDI, given their individual endowment with different kinds of FDI-subsidiaries? 3.1 Conceptual approach: the theoretical taxonomy

To find some answer to this question, we develop a conceptual, or theoretical taxonomy. This taxonomy is derived from organisational theory and the international business and management strategy literature and features two dimensions: (i) the management-relationship between the foreign affiliate and the parent investor, as well as (ii) the level of absorptive capacity of the foreign affiliate. With respect to the role that the foreign affiliate assumes within the network of the foreign investor (the management-relationship), the literature

11

assumes at the most general level that the stronger the competencies of the foreign affiliate in terms of its own management vis--vis control by the headquarter, the better it may be able to adapt foreign technology it may get access to from the foreign investor and the stronger will also be the positive technology-impact of the foreign affiliate on the host economy environment (e.g. Holm/Malmberg/Slvell 2002). However, this may not be linear: at early stages, parent companies can be adverse to technological incongruity (Dyker/Stolberg 2003, following Ozawa 1979 and Wells 1983) and may impose their own technological culture on their affiliates via direct control. Here, we assume that by installing the parents best practice without its adaptation to the particularities of the host country particularities, the process of direct technology transfer remains restricted to a static one-way transfer. The process may become dynamic with the foreign affiliate maturing in terms of its own expertise and gradually assuming a more active role in the adaptation of the parents technology (compare this to the short-term and long-term impacts as conceptualised by Tunzelmann, 2004, in his networkalignment approach). By enhancing its adaptive capabilities, the foreign affiliate establishes a process of technological interaction to the benefit of both partners, the parent and the foreign affiliate (Birkinshaw and Hood, 1998). This is what we refer to here as the dynamic, two-way knowledge transfer between foreign affiliate and parent investor.
typical mature FAs large potentials for static tech transfer large potentials for dynamic tech transfer unused benefits from parent small potentials for static tech transfer small potentials for dynamic tech transfer

s A F

s A F

,t n e h

,t n e h fA i F e r a s
Conceptual taxonomy
Note: FA denotes foreign affiliate

e f s n a r t

y m n o t u a

-r w b p h s n o i t a l e d r

fA i F o m n v i g e r a s

i h f l t o p g i s a e r c n

l a t n f o e l e
parent acts as a hindrance ? large potentials for static tech transfer small potentials for dynamic tech transfer typical young, immature FAs large potentials for static tech transfer small potentials for dynamic tech transfer high adaptive ability low adaptive ability Ability of FAs to adapt foreign technology received from parent

y b t m e g n a M u s e r p t a n i m o d
Figure 12

t n e r a p f h c r f s e t o p g n i l a

e f s n a r t

The theoretical taxonomy locates foreign affiliates into four stereotypical categories, each of which we assume to be linked over time: in the lower right quadrant, we assume foreign affiliates to be rather young and immature, they have low adaptive abilities and pretty much copy what they receive in terms of knowledge and technology. They also have a dominant parent and hence very little autonomy with respect to decisions over business functions in management. In such a scenario, we expect large potentials for a static, oneway direct technology transfer from parent to foreign affiliate. If such foreign affiliates do not develop their own adaptive abilities (or absorptive capacities) but at the same time gradually assume more independence from their parent companies and responsibilities for their own business functions (i.e. if they are left alone, see top right quadrant in the Figure: Conceptual taxonomy), then we assume the potentials for direct technology transfer to fall. If, however, foreign affiliates develop their own absorptive capacities yet are not granted more autonomy from their parents (bottom left quadrant), then we would characterise parent companies as a hindrance to the development of a dynamic technology-generating and technology-sharing network: potentials for direct technology transfer remain large but also static. Only in the case where the foreign affiliate simultaneously develops adaptive abilities AND is granted a more active role in the management-relationship in the investors network will this network be able to generate potentials for a dynamic, two-way knowledge transfer between foreign affiliate and parent. With Central East Europe being a region that has a long industrial history and is already in the process of catching up to the West in terms of technology and factor-prices, we assume that networks of parent investors and foreign affiliates here typically tend to develop in this direction over time (if they do not get stuck somewhere along the road, e.g. in the top right quadrant). 3.2 Empirical analysis and the empirical representation of the theoretical taxonomy When we use our FDI-database for all countries and all industries, we are able to locate all foreign affiliates into this taxonomy. This analysis at the firm level proxies absorptive capacities by the above explained interaction between R&D expenditure and R&D personnel. To account for industry-specific effects in the latter indicators, we again use the deviations of each foreign affiliate from the industry-means of all foreign affiliates (see above in the descriptive part). Business function-autonomy was interrogated directly by asking managers of foreign affiliates about their perception of their own autonomy in a set of seven individual business functions, namely production and operational management, market research and marketing, basic and applied research, product development, process engineering, strategic management and planning, investment projects and finance. Managers were asked to select on a rating scale between 1 and 4 with diminishing levels of autonomy (was rescaled to reverse in the analysis for a better match with intuition). In our analysis, we use averages over all business functions 13

and benefit from the seven sub-questions that managers perception has been interrogated more precisely than would have been the case with only one overall assessment of autonomy. Graph 3 (Averages of foreign affiliate autonomy levels) depicts the average levels of foreign affiliates levels of autonomy by host country: again, not surprisingly, the group of East German foreign affiliates have, on average, the feeling that they are in fact quite autonomous in deciding about business functions, and in particular higher than amongst most other countries foreign affiliates. Surprisingly, however, Romanian firms managers would rank their level of autonomy on average just as high as the East German ones do, and Polish and Slovenian managers feel the most dominated and controlled by their foreign parents. But: differences are not so pronounced between countries, and the answers will also depend to some degree on mentalities. Still, it makes sense to compare the managers perception for each firms with the same firms absorptive capacity, following the assumptions in our taxonomy above.
3 0 1 2

EDE

HRO

PL

RO sd of autonomy

SLO

mean of autonomy

Graph 3:

Averages of foreign affiliate autonomy levels

Note: EDE...East Germany, HRO...Croatia, PL...Poland, RO...Romania, SLO...Slovenia. Source: The IWH FDI Micro database.

The results are depicted in the empirical representation of the conceptual taxonomy above. Each dot represents one of the 510 foreign affiliates of the sample that answered the questions that were necessary to construct the indicators. The four-quadrant taxonomy is presented for the total population of our foreign affiliates (bottom right) and again for each of the host countries. The interpretation of results per host economy provides us with some indication as to how the countries compare in terms of endowment with different kinds of foreign affiliates. Form this, we attempt to infer the host 14

economies respective potentials of direct technology transfer between investor and affiliate. A graphical interpretation of the resulting empirical representations of this taxonomy already provides some interesting insights: if we take the picture for the total population as a benchmark for typical allocation patters of foreign affiliates in Central East Europe during this time, then FDI into the region appears to rather flock towards the right of the vertical centre of the taxonomy, i.e. a bias on low but negative deviations from the respective NACE 2-digit industry-averages. On the left hand side of the taxonomy, absorptive capacities are more dispersed with some firms already achieving higher levels.6 In terms of the horizontal allocation of firms in the twodimensional space, the totality of firms tend upwards suggesting high levels of autonomy. This is a result of the ranking scale used in the field work, where a level of 2 suggests that a business functions is currently undertaken mainly foreign investor network, whereas a level of 3 denotes mainly the foreign affiliate (both in the re-scaled version of the indicator). In terms of numbers (see Table 3 Shares of foreign affiliates allocated into the four quadrants of the taxonomy), the largest share of foreign affiliates (45.7%) are located in the top right quadrant, denoting left alone foreign affiliates with neither large potentials for a two-way direct technology transfer between parent and foreign affiliate, nor large potentials for a static one-way transfer from parent to foreign affiliate. 27.3 per cent are located in the lower right quadrant, i.e. immature foreign affiliates with only little potentials for a dynamic two-way technology transfer yet large potentials for the static one-way kind of technology transfer. In only 7.2 per cent of cases would we assume that the parent firm rather acts as a hindrance to the technological development of the foreign affiliate by not granting the kind of autonomy that the level of absorptive capacity of the firm would warrant. Nearly 20 per cent of all firms however are located in the region in which we characterise firms to contain large potentials for both the static and the dynamic kind of technology transfer. At the country level, using the geographical centre of all foreign affiliates in the space of the taxonomy, results suggest that East German foreign affiliates (and possibly Romanian) have probably the largest potentials for a dynamic technology transfer (top left). Croatian, Polish, and Slovenian Foreign affiliates probably the lowest potentials (top right). In terms of shares of foreign affiliates (Table 3), 33.3 per cent of all East German foreign affiliates are located in the top left quadrant. Whilst in Romania, an also large share (20 per cent) is located in this quadrant, all other countries assessed here have very low shares here and hence suggest very little potentials for a two-way dynamic technology transfer (Croatia: 8.2%, Slovenia: 7.5%, and Poland with
6

Negative and positive deviations even out in the picture for all foreign affiliates, this is per definition necessarily so, it is a result of the particular construction of the composite indicator with equal weights for all firms. At the same time, it is clear that this does not apply to country-subsets. Here, a right hand bias indicates comparatively lower absorptive capacities, a left hand bias higher capacities.

15

only 5.4%). The top right quadrant has high shares mainly in Croatia (63.3%), but also in Slovenia (50%) and Poland (47.8%). This is the location where we would expect rather low potentials for technology transfer. Yet Poland has a nearly equal share of firms located in the bottom right quadrant, where foreign affiliates are assumed to enjoy the possible prospect of being able to move up the learning curve.

EDE
4

HRO

PL

autonomy

RO
4

SLO

Total

1 20

10

-10

-20 20

10

-10

-20 20

10

-10

-20

adapt_dev
Graphs by host

Graph 4

Graphical representation of empirical results in the conceptual taxonomy

Note: EDE...East Germany, HRO...Croatia, PL...Poland, RO...Romania, SLO...Slovenia. Source: The IWH FDI Micro database.

16

Table 3

Shares of foreign affiliates allocated into the four quadrants of the taxonomy
Mature FAs Left-alone FAs East Germany Croatia Poland Romania Slovenia 39.5% 63.3% 47.8% 44.3% 50.0%

East Germany Croatia Poland Romania Slovenia

33.3% 8.2% 5.4% 21.0% 7.5%

Parent as hindrance East Germany Croatia Poland Romania Slovenia 8.6% 4.1% 3.3% 10.2% 2.5% East Germany Croatia Poland Romania Slovenia

Immature FAs 18.5% 24.5% 43.5% 24.6% 40.0%

Note: Percentages are calculated as shares of foreign affiliates of one country in the total number of firms of that country, not of the totality of foreign affiliates. Hence, all numbers for each country add up to 100 per cent (notwithstanding rounding errors).

The results hence suggest a ranking of countries in terms of potentials for technology transfer, based on their endowment with different kinds of foreign affiliates with respect to comparative levels of autonomy and comparative absorptive capacities. Such a ranking would probably be headed by East Germany and followed by Romania. If a location in the top right quadrant can be assumed to signify a half-way towards the top left quadrant where potentials for dynamic technology transfer is highest, then the ranking would proceed with Croatia and Slovenia. This assumes that those countries subsidiaries are in a position to better their potentials for technology transfer by increasing their adaptive capacities whilst already having achieved some autonomy from their parent investor companies. Poland would then assume a lower position in such a raking, a large share of Polish foreign affiliates may be characterised as being rather at the start of the typical foreign affiliatelearning curve that was characterised in the conceptual taxonomy.

Conclusions
This chapter set out to analyse FDI foreign affiliates in Central East Europe to shed some light on the developmental role of foreign affiliates in the 17

Central East European countries and regions of East Germany, Croatia, Poland, Romania, and Slovenia. This developmental role is assumed to depend on strategic motives of FDI projects, absorptive capacities of foreign affiliates and the source of such capacities (own sources, local host economy sources, or the foreign investor), and the management-relationship between the investor and its affiliate (i.e. the autonomy of the affiliate from the investor in terms of business functions). The analysis of strategic motives of foreign investment projects by use of a unique database constructed in a recent field work established that it is foremost the motive of access to new markets or the increase of an existing market share that appear to be the dominant across Central East Europe. This is particularly pronounced in East Germany and Poland. In Croatia, Slovenia and Romania, the strategic aim to increase efficiency in the foreign investors network also turned out to be of particular importance. In East Germany, access to localised knowledge, skills, and technology additionally plays a dominant role for foreign investors. In terms of the sources of absorptive capacities, we find that where foreign affiliate-technological development takes place, this apparently does not rely very much on the foreign parent network but rather is sourced mainly from within the foreign affiliate and from the local innovation system. In terms of the requirements for absorptive capacities, we conclude that they appear to originate from the foreign affiliates internal and local sources. This underlines the importance of the local environment and innovation networks for foreign investments in our post-transformation economies. The simultaneous characterisation of foreign affiliates along the two criteria of absorptive capacities and autonomy offers some important insights into the potentials of FDI-led economic development: the endowment of host economies with different kinds of foreign affiliates may be assumed to determine the potentials of a host economy to benefit from direct technology transfer between parent and affiliate. Here, the analysis develops a conceptual taxonomy that allows us to determine the potentials for technology transfer by assessing the endowment of each country with foreign affiliates, characterised along the two criteria of absorptive capacities and autonomy. If we can assume that the foreign affiliates in our sample are in fact representative of the total population for FDI projects in each of the countries, then we can tentatively make a ranking of countries according to the potentials of technology transfer. In such a ranking, East Germany would contain the largest potentials for dynamic, a two-way direct technology transfer between foreign parent and its affiliate. The ranking would proceed with Romania with a larger share also of affiliates where the parent investor may be considered a hindrance to the further development of the affiliate and a larger share of affiliates that still remain to increase their own technological abilities. The latter characterisation also applies to Croatian and Slovenian affiliates, hence those countries would come next in the ranking. Poland would come last in the ranking, mainly because of its foreign affiliates can 18

predominantly be characterised as technologically immature, both by having little absorptive capacities and by at the same time being rather dependent on the decisions of their foreign investor. This lowest rank, however, is also attached with the possible development of affiliates along the typical learning curve. Those results, however, have to be taken with due care: first, the answers of foreign affiliates with respect to autonomy are highly subjective and possibly also country-specific due to differences in mentalities. Furthermore, the deviations around the geographical centre need to be taken into account when interpreting the results where they present averages: for autonomy, standard deviations are particularly high in Romania, but the levels of standard deviations in per cent of the means here are nothing like the levels for absorptive capacity in all countries, and here in particular for Romania again and here also East Germany. Still, the analysis offers some clear and interesting results.

References
Barrios, S. , H. Goerg, and E. Strobl, 2003, Explaining Firms' Export Behaviour: R&D, Spillovers and the Destination Market, Oxford Bulletin of Economics and Statistics, Vol. 65(4), pp. 475-496. Birkinshaw, J. and N. Hood (eds), 1998, Multinational Corporate Evolution and Subsidiary Development, London: Macmillan. Blalock, G., and J. Gertler, 2004, Welfare gains from FDI through technology transfer to local suppliers, mimeo, University of California, Berkeley. Borensztein, E., J. De Gregorio, and J-W. Lee, 1998, How Does Foreign Direct Investment Affect Economic Growth?, Journal of International Economics, 45, pp. 115-35. Cohen, W. and D. Levinthal, 1989, Innovation and Learning: The Two Faces of R&D, The Economic Journal, Vol. 99, No. 397, pp. 569-596. Cohen, W. and D. Levinthal, 1990, Absorptive capacity: a new perspective on learning and innovation, Administrative Science Quarterly 35 (1), 128152. Damijan, J.P. and M. Knell, 2003, Impact of privatization methods on the accessibility of local firms to international knowledge spillovers through trade and foreign investment: Evidence from Estonia and Slovenia, University of Oslo, mimeo. Dunning, J.H., 1993, Multinational Enterprises and the Global Economy (Wokingham, England: Addison-Wesley Publishing Company). Dyker, D.A. and C. Stolberg, 2003, Productivity and capability in the transition economies - a historical and comparative perspective, University of Sussex, mimeo. Dyker, D.A. et al (ed.), 2006, Closing the EU East-West Productivity Gap, London: Imperial College Press.

19

Gnther, J., J. Stephan, and B. Jindra, 2008, Foreign Subsidiaries in the East German Innovation System Evidence from Manufacturing Industries, Applied Economics Quarterly Supplement, 59, pp. 137-165. Holm, U., Malmberg, A. and . Slvell, 2002, MNC Impact on Local Clusters, IIB Institute of International Business Working paper No. 02/11, IIB, Stockholm. Hunya, G., 1997, Large privatisation, restructuring and foreign direct investment. In: Zecchini S. (ed.) Lessons from the Economic Transition in Central and Eastern Europe in the 1990s. (Dordrecht: Kluwer Academic Publishers) 275300. Keller, W., 1996, Absorptive capacity: On the creation and acquisition of technology in development, Journal of Development Economics, 49/1, pp. 199-227. Kneller, R., 2002, Frontier Technology, Absorptive Capacity And Distance, Research Paper Series: Globalsation, Productivity and Technology, 2002 (24), Leverhulm Centre. Kinoshita, Y., 2000, R&D and technology spillovers via FDI: Innovation and absorptive capacity, Working paper no. 349, The William Davidson Institute, University of Michigan Business School. Lane, P.J. and M. Lubatkin, 1998, Relative absorptive capacity and interorganizational learning, Strategic Management Journal, 19/5, pp. 461-77. Ozawa, T., 1979, Multinationalism, Japanese Style, Princeton: Princeton University Press. Schoors, K. and B. van der Tool, 2002, Foreign Direct investment spillovers within and between sectors: Evidence from Hungarian data, Working Paper no. 157, University of Ghent. Todo, Y. and K. Miyamato, 2002, Knowledge Diffusion From Multinational Enterprises: The Role of Domestic And Foreign Knowledge-Enhancing Activities, OECD Development Centre, Technical Papers No. 196. Tunzelmann, N. von, 2004, Network alignment in the catching up economies of Europe. In: The Emerging Industrial Structure of the Wider Europe (Mc Gowan et al, ed.), Routledge Studies in Business Organisations and Networks, pp. 23-37. Van den Bosch, F.A.J., H.W. Volberda, and M. De Boer, 1999, Co-evolution of firm absorptive capacity and knowledge environment: Organizational forms and combinative capabilities, Organization Science, 10/5, pp. 55168. Wells, L.T., 1983, Third World Multinationals, Cambridge, MA: MIT Press.

20