Journal of Economic & Administrative Sciences

Vol. 23, No. 2, December 2007 (44-70)

Determinants of Corporate Dividend Policy in Jordan:

An Application of the Tobit Model
Husam-Aldin Nizar Al-Malkawi
Department of Finance and Banking
Faculty of Administrative and Financial Sciences
Al-Ahliyya Amman University, Jordan
Abstract
This paper examines the determinants of corporate dividend policy in
Jordan. The study uses a firm-level panel data set of all publicly traded firms
on the Amman Stock Exchange between 1989 and 2000. The study develops
eight research hypotheses, which are used to represent the main theories of
corporate dividends. A general-to-specific modeling approach is used to
choose between the competing hypotheses. The study examines the
determinants of the amount of dividends using Tobit specifications. The
results suggest that the proportion of stocks held by insiders and state
ownership significantly affect the amount of dividends paid. Size, age, and
profitability of the firm seem to be determinant factors of corporate dividend
policy in Jordan. The findings provide strong support for the agency costs
hypothesis and are broadly consistent with the pecking order hypothesis. The
results provide no support for the signaling hypothesis.
1. Introduction
The topic of dividend policy is one of the most enduring issues in modern
corporate finance. This has led to the emergence of a number of competing
theoretical explanations for dividend policy. No consensus has emerged
about the rival theoretical approaches to dividend policy despite several
decades of research. A range of firm and market characteristics have been
proposed as potentially important in determining dividend policy. The
attempt to test these competing models and refine them has in turn spawned
a vast empirical literature. The empirical work on dividend policy has,

Tel. (+962) 799 666 527. E-mail address: halmalkawi@ammanu.edu.jo.

I am thankful and indebted to Michael Rafferty, Stephane Mahuteau, and Roger
Ham of the University of Western Sydney who supervised the dissertation from
which much of this paper is derived.

44

Journal of Economic & Administrative Sciences

December 2007

however, generally been focused on developed stock markets such as the UK

and US.
The examination of dividend policy in emerging stock markets has, until
recently, been much more limited. Yet the sorts of firm and market
characteristics that may influence dividend policy may in fact be more likely
to be present in developing markets in an exaggerated fashion than in
developed markets. This provided a central motivation for the present study.
This study seeks to add to that literature by providing a detailed analysis of
dividend policy in Jordan, an emerging market that has been particularly
poorly analyzed to date. By and large, emerging stock markets have several
similar characteristics so, to some extent, corporate dividend behavior in
Jordan may share some important similarities with other emerging equity
markets. Consequently, the findings of such a detailed country case study
could form the basis of future comparative research into other emerging
markets. Such findings may also provide the basis for reflection on empirical
research in developed markets.
The paper uses a firm-level panel data set of all publicly traded firms on
the Amman Stock Exchange (ASE) between 1989 and 2000. The study
develops eight research hypotheses, which are used to represent the main
theories of corporate dividends. A general-to-specific modeling approach is
used to choose between the competing hypotheses. The study examines the
determinants of the amount of dividends using Tobit specifications. The
factors that affect dividend policy in developed stock markets seem to apply
for this emerging market, but often in different ways and on a different scale.
The results suggest that the proportion of stocks held by insiders and state
ownership significantly affect the amount of dividends paid. Size, age, and
profitability of the firm seem to be determinant factors of corporate dividend
policy in Jordan. These factors are found to positively affect the level of
dividends. The findings provide strong support for the agency costs
hypothesis and are broadly consistent with the pecking order hypothesis. The
results provide no support for the signaling hypothesis.
The rest of the paper is organized as follows. Section 2 gives a short
discussion of dividend policy theories. Section 3 formulates the hypotheses.
Section 4 describes the data. Section 5 explains the methodology. Section 6
reports the results. The final section summarizes and concludes the paper.

45

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007
1

2. Theoretical Consideration and Prior Research

Financial and business historians have shown that dividend policy has
been bound up with the historical development of the corporation. In its
modern form, however, dividend policy theory is closely tied to the work of
Miller and Modigliani (1961, hereafter M&M) and their dividend policy
irrelevance thesis. M&M demonstrate that under certain assumptions
including rational investors and a perfect capital market, the market value of
a firm is independent of its dividend policy. In actual market practices
however, it has been found that dividend policy does seem to matter, and
relaxing one or more of M&Ms perfect capital market assumptions has
often formed the basis for the emergence of rival theories of dividend policy.
The bird-in-hand theory (a pre-Miller-Modigliani theory) asserts that in a
world of uncertainty and information asymmetry dividends are valued
differently to retained earnings (capital gains). Because of uncertainty of
future cash flow, investors will often tend to prefer dividends to retained
earnings. As a result, a higher payout ratio will reduce the required rate of
return (cost of capital), and hence increase the value of the firm (see, for
example Gordon, 1959). This argument has been widely criticized and has
not received strong empirical support.
The tax-preference theory posits that low dividend payout ratios lower
the required rate of return and increase the market valuation of a firms
stocks. Because of the relative tax disadvantage of dividends compared to
capital gains investors require a higher before-tax risk adjusted return on
stocks with higher dividend yields (Brennan, 1970). Several studies
including Litzenberger and Ramaswamy (1979), Poterba and Summers,
(1984), and Barclay (1987) have presented empirical evidence in support of
the tax effect argument. Others, including Black and Scholes (1974), Miller
and Scholes (1982), and Morgan and Thomas (1998) have opposed such
findings or provided different explanations.
Another closely related theory is the clientele effects hypothesis.
According to this argument, investors may be attracted to the types of stocks
that match their consumption/savings preferences. That is, if dividend
income is taxed at a higher rate than capital gains, investors (or clienteles) in
high tax brackets may prefer non-dividend or low-dividend paying stocks,
and vice versa. Also, the presence of transaction costs may create certain
clienteles. There are numerous empirical studies on the clientele effects
hypothesis but the findings are mixed. Taking different paths, Pettit (1977),
Scholz (1992), and Dhaliwal, Erickson and Trezevant (1999) presented
1

See Lease et al. (2000) for a comprehensive review of the literature.

46

Journal of Economic & Administrative Sciences

December 2007

evidence consistent with the existence of clientele effects hypothesis. Studies

finding weak or contrary evidence include Lewellen et al. (1978),
Richardson, Sefcik and Thomason (1986), Abrutyn and Turner (1990),
among others.
Despite the tax penalty on dividends relative to capital gains, firms may
pay dividends to signal their future prospects. This explanation is known as
the information content of dividends or signaling hypothesis. The intuition
underlying this argument is based on the information asymmetry between
managers (insiders) and outside investors, where managers have private
information about the current performance and future fortunes of the firm
that is not available to outsiders. Here, managers are thought to have the
incentive to communicate this information to the market. According to
signaling models (Bhattacharya, 1979, John and Williams, 1985, and Miller
and Rock, 1985) dividends contain this private information and therefore can
be used as a signaling device to influence share price. An announcement of
dividend increase is taken as good news and accordingly the share price
reacts favorably, and vice versa. Only good-quality firms can send signals to
the market through dividends and poor-quality firms cannot mimic these
because of the dissipative signaling costs (for example, transaction costs of
external financing, or tax penalties on dividends, or distortion of investment
decisions). Support for the signaling hypothesis can be found, for example,
in Pettit (1972), Michaely, Thaler and Womack (1995), Nissim and Ziv
(2001), and Bali (2003). Other researchers, however, found limited support
or rejected the hypothesis (see Watts, 1973, Gonedes, 1978, and Conroy,
Eades and Harris, 2000, among others). Note that the aforesaid studies
followed different approaches.
The information asymmetry between managers and shareholders, along
with the separation of ownership and control, formed the base for another
explanation for why dividend policy may matter; that is, the agency costs
thesis. This argument is based on the assumption that managers may conduct
actions in accordance with their own self-interest which may not always be
beneficial for shareholders. For example, they may spend lavishly on
perquisites or overinvest to enlarge the size of their firms beyond the optimal
size since executives compensation is often related to firm size (see Jensen,
1986). The agency costs thesis predicts that dividend payments can reduce
the problems associated with information asymmetry. Dividends may also
serve as a mechanism to reduce cash flow under management control, and
thus help to mitigate the agency problems. Reducing funds under
management discretion may result in forcing them into the capital markets
more frequently, thus putting them under the scrutiny of capital suppliers

47

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

(Rozeff, 1982, and Easterbrook, 1984). Many researchers have offered

empirical support for agency explanations for why firms pay dividends
including Rozeff (1982), Lloyd, Jahera and Page (1985), Jensen, Solberg and
Zorn (1992), and Holder, Langrehr and Hexter (1998), among others. Others
including Denis, Denis and Sarin (1994), Yoon and Starks (1995), and Lie
(2000) provided little support or reject the hypothesis.
Theories discussed above presented differing explanations for the
determinants of corporate dividend policy, and provide a partial solution to
the dividend puzzle (the debate between these explanations remains
unresolved).
3. Research Hypotheses and Selection of Proxy Variables
Section 2 showed the main theoretical arguments for dividend policy
along with the empirical evidence. This in turn is now used as a guide to
formulate research hypotheses given the key characteristics of Jordanian
capital market. This section also presents the selection of proxy variables
that are often suggested in the literature and their expected relationship to
dividend policy as measured by dividend yield.
Hypothesis 1: Dividends serve as a bonding mechanism to reduce agency
problems
The agency hypothesis of dividends predicts that dividend payments can
be used as a mechanism to alleviate agency problems. To test this hypothesis
two proxy variables are used. The first proxy for agency costs is the natural
logarithm of number of shareholders (STOCK), which is used to measure
ownership dispersion (see, for instance, Rozeff, 1982, and Deshmukh, 2003).
The hypothesized relation between dividend payouts and STOCK is
expected to be positive. That is, in order to alleviate the agency costs
associated with an increasing number of shareholders firms should pay more
dividends, other things being equal.
The second proxy for agency costs is the percentage of a firms common
stock held by insiders (INSD) (see, for example, Rozeff, 1982, Jensen et al.,
1992, and Holder et al., 1998). It has been argued that agency costs may be
reduced if insiders (managers, directors, and other executive officers)
increase their ownership in the firm, because this can help to align the
interests of both managers and shareholders (Jensen and Meckling, 1976).
Therefore, the higher the proportion of managers in firm ownership, the less
is the need for using dividends as a device to mitigate agency costs. Hence,
INSD variable is expected to bear a negative relation to dividend payouts.
48

Journal of Economic & Administrative Sciences

December 2007

Hypothesis 2: Ownership and control structures affect corporate

dividend policy
Although the relation between dividend payouts and a firms ownership
structure can be examined within the agency theory framework, in the case
of Jordan it is important to address the relationship separately because of the
variety of owners of Jordanian firms (for instance, families, institutions,
government, and foreign investors). To test the link between dividend policy
and ownership structure, a set of four dummy variables were included to
describe the ownership structure of the firm: family (FAML), state (STATE),
institution (INST), and multiple (MULT). To identify the ultimate owner of
the firm we employed the 10-percent threshold level of ownership, which is
the criterion used by the ASE throughout the study period. We propose that
dividend payouts are negatively related to FAML and positively to STATE,
INST and MULT, ceteris paribus.
Hypothesis 3: Dividends as a signaling device
The information content of dividends (signaling) hypothesis predicts that
dividends can be used to signal firms future prospects and only good-quality
firms can use such a device. The hypothesis can be examined by identifying
the relationship between information asymmetry and dividend payouts. A
potential proxy for the degree of information asymmetry is the trading
volume of a firms shares. In general, investors tend to invest in securities
that are better known in the market, that is, with less information
asymmetries. Therefore, other things being equal, the higher the information
asymmetry of a security the lower its trading volume. This analysis uses
annual share turnover (TURN) as a proxy measure for information
asymmetry (see, for example, Bartov and Bodnar, 1996, and Leuz and
Verrecchia, 2000). Using this proxy for information asymmetry, the
signaling hypothesis predicts an inverse relationship between share turnover
and dividend payouts.
Hypothesis 4: Firm growth and investment opportunities are negatively
associated with dividend payouts
Firms with high growth and investment opportunities will need the
internally generated funds to finance those investments, and thus tend to pay
little or no dividends. This prediction is consistent with the pecking order
hypothesis 2 proposed by Myers and Majluf (1984). Accordingly, we expect
2

The pecking order hypothesis suggests that firms finance investments first with the
internal finance, and if external financing is necessary, firms prefer to issue debt

49

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

the firms growth and investment opportunities, as measured by market-tobook ratio (MBR), to be negatively related to dividends payouts (see, for
instance, Deshmukh, 2003, and Aivazian et al., 2003).
Moreover, mature companies are likely to be in their low-growth phase
with less investment opportunities (see Barclay et al., 1995, and Grullon et
al., 2002). These companies do not have the incentives to build-up reserves
as a result of low growth and few capital expenditures, which enable them to
follow a liberal dividend policy. On the contrary, new or young companies
need to build-up reserves to face their rapid growth and financing
requirements. Hence, they retain most of their earnings and pay low or no
dividends. Therefore, the age of the firm (AGE) is used as a second proxy
for the firms growth opportunities. Several studies have related firm growth
to age (Farinas and Moreno, 2000, and Huergo and Jaumandreu, 2004,
among others). Other things held constant, as a firm gets older its investment
opportunities decline leading to lower growth rates, consequently reducing
the firms funds requirements for capital expenditures. Therefore, dividend
payout should be positively related to the firms age. Yet, we do not expect
the impact of age to always be linear. Thus, we allow the effect of age to be
non-linear by including the age squared (AGESQ). If the coefficient on
AGESQ appears to be negative, then our assumption of a quadratic
relationship between age and dividends is true.
Hypothesis 5: The firm size is positively associated with dividend
payouts
A large firm typically has better access to capital markets and finds it
easier to raise funds with lower cost and fewer constraints compared to a
small firm. This suggests that the dependence on internal funding decreases
as firm size increases. Therefore, ceteris paribus, large firms are more likely
to afford paying higher dividends to shareholders. In this study the firms
market capitalization of common equity (MCAP) is used as a measure for
size (see, for example, Deshmukh, 2003). Based on the above discussion and
consistent with previous research the size variable is expected to have a
positive relationship with dividend payouts.
Hypothesis 6: The firm debt is negatively associated with dividend
payouts

before issuing equity to reduce the costs of information asymmetry and other
transactions costs. (Myers 1984, and Myers and Majluf, 1984).

50

Journal of Economic & Administrative Sciences

December 2007

When a firm acquires debt financing it commits itself to fixed financial

charges embodied in interest payments and the principal amount, and failure
to meet these obligations may lead the firm into liquidation. The risk
associated with high degrees of financial leverage may therefore result in
low dividend payments because, ceteris paribus, firms need to maintain their
internal cash flow to pay their obligations rather than distributing the cash to
shareholders. Moreover, Rozeff (1982) points out that firms with high
financial leverage tend to have low payouts ratios to reduce the transaction
costs associated with external financing. Therefore, other things being equal,
an inverse relationship between financial leverage ratio, defined as the ratio
of total short-term and long-term debt to total shareholders equity (DER),
and dividends is expected.
Hypothesis 7: There is a positive relationship between a firms
profitability and dividend payouts
The decision to pay dividends starts with profits. Therefore, it is logical
to consider profitability as a threshold factor, and the level of profitability as
one of the most important factors that may influence firms dividend
decisions. In his classic study, Lintner (1956) found that a firms net
earnings are the critical determinant of dividend changes.
The pecking order hypothesis may provide an explanation for the
relationship between profitability and dividends. That is, taking into account
the costs of issuing debt and equity financing, less profitable firms will not
find it optimal to pay dividends, ceteris paribus. On the other hand, highly
profitable firms are more able to pay dividends and to generate internal funds
(retained earnings) to finance investments. Fama and French (2002) used the
expected profitability of assets in place for testing the pecking order
hypothesis. In another study, Fama and French (2001) interpreted their
results of the positive relationship between profitability and dividends as
consistent with the pecking order hypothesis.
Based on the above discussion, profitability is expected to be a key
determinant of corporate dividend policy in Jordan. To test this hypothesis,
the after tax earnings per share (EPS) is used as a measure of a firms
profitability. The hypothesized relationship between EPS and dividends is
positive.
Hypothesis 8: The relative tax disadvantage of dividends induces lower
dividend payouts
The tax-preference theory proposed that companies should retain rather
than distribute their income because of the preferential tax treatment of

51

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

capital gains versus dividends. Prior to 1996, investors paid no taxes on

income received from Jordanian companies in both the form of capital gains
and dividends. However, in 1996 the Jordanian Income Tax Authority
imposed a 10 percent tax rate on dividends3, resulting in making dividend
payments costly for investors. Since this study covers the period between
1989 and 2000, which includes five years following the implementation of
taxes on dividends, it is important to examine if the resulting relative tax
disadvantage of dividends induces Jordanian firms to reduce their payout
ratios. In order to examine any change in dividend payouts following the
change in tax law we introduced a dummy variable (DTAX), which divides
the study into two periods, post-tax (1996-2000) and pre-tax (1989-1995).
Based on the tax-preference argument, the association between DTAX and
dividends is anticipated to be negative.
4. The Data
The data employed in this study is derived from the annual publications
of the ASE. Based on a 12-year period (1989-2000) and 160 companies
listed on the ASE, a panel dataset was constructed4. This dataset consist of
all companies listed on the ASE covering four sectors: industrial, service,
insurance, and banks. These companies ranged from old to newly established
ones, and some companies were de-listed during the study period. Therefore,
the number of observations for each company is different. In order to gain
the maximum possible observations, pooled cross-section and time-series
data is used. Because the number of observations for each company is not
identical, this results in an unbalanced panel. The analysis is based on annual
data, because the data set is annual, and the ultimate sample consists of 759
firm-year observations. The present study includes both dividend-paying as
well as non-dividend-paying firms. The exclusion of non-dividend-paying
firms results in a well-known selection bias problem (see for instance, Kim
and Maddala, 1992, and Deshmukh, 2003).
5. Methodology
5.1 Choosing between Hypotheses: The General to Specific Method
In Section 3 eight hypotheses were developed and a set of related proxy
variables were identified. The selected variables constitute the general model
to be tested in order to determine the factors that may affect dividend policy
3

This law was again amended in 2001 and implemented as of January 2002
removing the 10 percent tax on dividends.
4
The list of companies used in the analysis is available upon request.

52

Journal of Economic & Administrative Sciences

December 2007

in the case of Jordan. To choose between the competing hypotheses and to

arrive at the best model that fits the data the general-to-specific method is
used. The general-to-specific method is generally referred to as the London
School of Economics approach to econometric modeling (see Hendry, 1995).
Based on this approach, the analysis starts with a general unrestricted
model, which includes all the variables that were identified and supported by
theories of dividend policy. Next, nested procedures are followed to arrive at
the best-fitted model. Further, in testing the competing models the
likelihood-ratio (LR) test is carried out. The statistic LR = -2[Log(LR)Log(LUR)] follows a 2 ( k ) , where k is the number of restrictions and the
null model is the restricted model. This test enables us to see whether the
additional parameters in the unrestricted model significantly increase the
likelihood. In other words, the test is used to confirm whether or not the
unrestricted model is statistically different from the restricted model.
The general model to be estimated using the Tobit specifications, for firm
i in period t [mathematical signs indicate the hypothesized impact on
dividend policy as measured by dividend yield (DYLD)] can be written as:
DYLD = 0 + 1 STOCK - 2 INSD - 3 FAML + 4 STATE
+ 5 INST + 6 MULT + 7 AGE - 8 AGESQ 9 MBR
- 10 DER - 11 TURN - 12 DTAX + 13 MCAP + 14 EPS
15 NONFIN + ,

(1)

where the variables are defined in Table 1 below. The table also provides a
summary of the research hypotheses and identifies the dependent variable
used in the regressions.
Table 1
Summary of Research Hypotheses and Proxy Variables
H1: Agency
costs
H2: Ownership
structure*

STOCK: natural log of number of common

stockholders
INSD: percentage held by insiders
FAML: family dummy = 1 if firm is family
owned, and 0 otherwise.
STATE: state dummy = 1 if firm is owned by the
government or its agencies, and 0 otherwise.
INST: institution dummy = 1 if firm is owned by
an institution, and 0 otherwise.
MULT: multiple owners dummy = 1 if firm has
more than one type of owners who control at least
10% of the stock, and 0 otherwise.

53

Positive
Negative
Negative
Positive
Positive
Positive

Dr. Husam-Aldin Nizar Al-Malkawi

H3: Signaling
H4: Investment
opportunities
H5: Size
H6: Financial
leverage
H7: Profitability
H8: Taxes
Control variable

December 2007

TURN: share turnover (proxy for information

asymmetry)
MBR: market-to-book ratio
AGE: age of the firm
AGESQ: the square of AGE
MCAP: natural log of market capitalization
DER: debt-to-equity ratio
EPS: the after-tax earnings per share
DTAX: tax dummy = 1 for the years 1996-2000,
and 0 otherwise.
NONFIN: is a dummy variable to control for
industry effects = 1 if a firm belongs to nonfinancial sector, and zero otherwise
DYLD: dividend yield (dividend-to-price ratio)

Negative
Negative
Positive
Negative
Positive
Negative
Positive
Negative
Negative/
Positive

Dependent
variable
* The 10% threshold level of ownership is used to identify the ultimate owner of the
firm.

5.2. Tobit Estimation

In considering the dividend decision firms have only two options, either
to pay or to not pay dividends. In Jordan, many companies do not pay
dividends at all, and even those who pay dividends do not pay them
continuously5. This gives the dependent variable (dividends) a special
feature in that it takes two outcomes. It is either equal to zero or positive.
Dividends can never be negative. Therefore, OLS is not an appropriate
method to analyze the payment of dividends, because of the nature of the
dependent variable. Indeed there is what one calls a mass point in 0
because the dividends paid by firms can only be positive or nil. The
appropriate technique in this case is to apply Tobit estimation. Kim and
Maddala (1992) explicitly supported this claim (see also Anderson, 1986,
and Huang, 2001).
The evaluation of the determinants of the amount of dividends is carried
out using the general specification of the censored data estimation, namely
the Tobit model. Indeed, the observed dependent variable, the amount of
dividend paid by each firm, may either be zero or positive. The data are then
censored in the lower tail of the distribution.
5
Of the 1511 cash dividend observations in our sample, 853 observations (56.45
percent) are zero dividends.

54

Journal of Economic & Administrative Sciences

December 2007

The estimation involves the following general structure: the latent

underlying regression for the amount of dividend paid by the firm is
defined as (see Verbeek, 2000):
*
(2)
yit = x'it + i + it ,
with the observed dependent variable being such that:
yit = 0,
*

= yit ,

if yit 0 ,
*

if yit > 0 .
*

The random effects i and the error term it are assumed to be

iid N (0, 2 ) and iid N (0, 2 ) , respectively, and independent of xi1,

xiT. To estimate the maximum likelihood estimation (MLE) method is

used6. The software package STATA (version 8) is used to perform the
estimation.
The main emphasis in the analysis of the results from MLE will be on the
significance of each estimated coefficient as well as on the overall
significance of the model as judged by the Chi-square ( 2 ) statistics derived
from the Wald test statistic. The Wald test statistic follows a 2 distribution
with degrees of freedom equal to the number of coefficient restrictions.
6. Results
Table 2 presents summary statistics of all variables used in the analysis.
The table reports the mean, standard deviation, minimum, maximum,
coefficient of variation (CV)7, and the number of observations for each of the
dependent and independent variables. The CV indicates that there is a
significant variation among the explanatory variables.
In order to ensure that our results are robust, several diagnostic tests were
performed. In attempting to detect multicollinearity, we computed the
variance inflation factors (VIF) for the independent variables. The estimated
VIF values were small (much less than 10, the rule of thump) with an
average of 1.64 indicating an absence of multicollinearity between the
variables. The correlation matrix also confirmed the absence of
multicollinearity among the explanatory variables used in the regressions
(see Appendix 1).

6
7

See StataCorp (2003) and Verbeek (2000) for the likelihood function.
The coefficient of variation CV is defined as the standard deviation over the mean.

55

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

Table 2
Descriptive Statistics for the Dependent and Independent Variables
Variable
Mean
Std. Dev.
0.027
0.035
DYLD
3.068
0.746
STOCK
0.289
0.246
INSD
0.303
0.460
FAML
0.210
0.408
STATE
0.539
0.499
INST
0.368
0.483
MULT
15.738
13.533
AGE
430.717
703.841
AGESQ
1.290
1.025
MBR
2.143
6.939
DER
0.299
0.671
TURN
0.500
0.500
DTAX
15.731
1.403
MCAP
0.282
1.774
EPS
0.738
0.440
NONFIN
Note: Variables are defined in Table 1.

Min
0
1.114
0
0
0
0
0
0
0
-0.020
-177.242
0
0
11.156
-8.388
0

Max
0.400
5.407
0.945
1
1
1
1
70
4900
12.670
69.270
7.993
1
21.356
28.299
1

CV
1.282
0.243
0.851
1.517
1.941
0.926
1.311
0.860
1.6341
0.795
3.238
2.243
1.000
0.089
6.285
0.597

Obs.
1316
885
936
1181
1181
1181
1182
1598
1598
1382
1511
1491
1592
1320
1514
1920

The likelihood ratio (LR) test reported at the bottom of table of the results
(presented below) provides a formal test for the pooled (Tobit) estimator
against the random effects panel estimator. For all estimated regressions, the
results of the LR test indicate that the panel-level variance component is
important and, therefore, the pooled estimation is different from the panel
estimation. The reported coefficient of Rho ( ), which is the panel-level
variance component, provides a similar test.
In attempting to test for heteroskedasticity, a two-stage test procedure is
employed. An important test for heteroskedasticity is the Lagrange
Multiplier test devised independently by Breusch and Pagan (1979) and
Godfrey (1978). The results of the test show that our models do not suffer
from a heteroskedasticity problem. For example, in the general model
(Model 1 of Table 3) the observed Chi-square value of 2.581 is not
significant at the 5 percent level (the 5 percent critical value from a Chisquare distribution with 1 degree of freedom is 3.841). Therefore the null
hypothesis of homoskedasticity (or no heteroskedasticity) is not rejected.
Table 3 gives the results of the maximum likelihood estimation (MLE) of
the random effects Tobit model. The table shows three models in which
56

Journal of Economic & Administrative Sciences

December 2007

dividend policy is measured by the dividend yield (DYLD). The second

column of each model reports the estimated marginal effect, which is the
probability that the dependent variable is uncensored8. That is, the
probability of a non-dividend-paying firm to pay dividends conditional on a
change in each of the explanatory variables. The Wald test statistics reject
the null hypothesis that the parameters in the regression equations are jointly
equal to zero (models 1-3).
The general model (Model 1) includes fifteen variables and encompasses
all of the models, with 759 firm-year observations. Of the fifteen variables
used in the model, twelve have the hypothesized signs with the exception of
STOCK and MBR. Seven variables are statistically different from zero. To
control for industry effects, a dummy variable (NONFIN) is included taking
a value of one if a firm belongs to industrial or services sectors, and zero
otherwise (for firms in the insurance or banking sectors).
In the process of moving from the general to the specific model, six
variables (STOCK, FAML, INST, MULT, MBR, and TURN) were dropped
from Model 1 generating Model 2. The likelihood ratio test (LRT) is
performed to test Model 2 (restricted) against Model 1 (unrestricted) and see
whether this process statistically provides additional explanatory power to
the model. In this case, the LRT statistic is LR= 2 [270.47 - 272.33] = 3.72,
and the critical value from a 2 distribution, with 6 degrees of freedom, is
12.59 at the 5 percent level of significance. Since the computed value is less
than the critical value, the null hypothesis is not rejected. That is, the null
model (Model 2), which is the restricted model, cannot be rejected.
Therefore, Model 1 does not provide a statistically significant increase in
likelihood over Model 2, which supports our exclusion of the aforesaid
variables. These procedures are repeated until we arrived at the best model
that fits our data, that is Model 3. The variables included in Model 3 possess
the anticipated sign and are statistically significant.
From models 2 and 3, the coefficient of NONFIN was not statistically
different from zero, suggesting that industry effects may not be important. At
the aggregate level, the regressions (models 1 3) are highly significant at
the 1 percent level or better.
The regression results of Table 3 show that the variable STOCK has a
negative sign but is not significantly different from zero9, indicating that
ownership dispersion does not appear to influence dividend policy in Jordan.

8
9

For computing the marginal effects see, for example, Greene (1999).
Using US data, Alli et al. (1993) obtained a similar result.

57

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

Table 3
MLE Results for Random Effects Tobit Model
Dependent Variable = DYLD
Model 1
Model 2

Independ.
Variables

Coefficient
Estimates

Constant

-0.266***
(-4.10)
-0.008
(-1.01)
-0.060***
(-2.61)
-0.004
(-0.43)
0.024*
(1.93)
0.008
(0.76)
0.004
(0.37)
0.004***
(4.52)
-0.00006***
(-3.67)
0.00015
(0.04)
-0.003**
(-2.59)
-0.0002
(-0.02)
-0.007
(-1.52)
0.016***
(3.62)
0.005**
(2.52)
-0.017
(-1.30)
0.597***
(10.46)

STOCK
INSD
FAML
STATE
INST
MULT
AGE
AGESQ
MBR
DER
TURN
DTAX
MCAP
EPS
NONFIN
Rho ()a

No. of obs.
Log LKHD
Wald test

Marginal
Effects
(%)

Model 3

Marginal
Marginal
Coefficient
Effects Coefficient Effects
Estimates
Estimates
(%)
(%)
-0.278***

-0.300***

(-5.43)

(-6.25)

-0.241

-1.795

-0.041**

-1.237

-0.043**

-1.273

-0.130

0.770

0.021*
(1.90)

0.672

0.019*
(1.83)

0.610

0.237

0.118

0.131

0.004***

0.128

0.004***

0.117

-0.002

(-2.04)

(3.93)

-0.00006***

(-3.64)

0.005
-0.090

0.479
0.136
-1.686

759
272.33

-0.003***

(-2.95)

-0.006
-0.224

-0.002

-0.090

(-2.29)

(3.83)

-0.00006***

(-3.83)

-0.002

-0.002***

(-2.61)

-0.066

-0.006
(-1.40)
0.015***
(4.72)

-0.194
0.464

0.005***

0.149

0.006***

0.167

(2.84)
-0.017
(-0.97)
0.588***
(10.46)

-0.532

759
270.47

2 (15)=85.75

2 (9)=78.26

0.016***
(5.34)

0.480

(3.36)

0.578***
(11.77)

759
269.15

2 (7)=78.03

P-value
0.000
0.000
0.000
LR testb
152.83
175.15
202.85
P-value
0.000
0.000
0.000
Notes: See Table 1 for definitions. t-statistics are in parentheses. *, ** and ***
denote significance at the 10, 5 and 1 percent levels, respectively. a proportion of
total variance contributed by panel level variance component. b provides a test for
pooled (Tobit) estimator against the random effects panel estimator.

58

Journal of Economic & Administrative Sciences

December 2007

This result may be attributed to the special characteristic of the Jordanian

capital market at which firms tend to have high levels of ownership
concentration. In such a case, the concentration of ownership may lead to a
concentration in control. Therefore, minority shareholders will not be able to
exert much influence on dividend policy.
As predicted, the coefficient of the second proxy of agency costs (INSD)
is negative and statistically significant at the 1 and 5 percent levels. This
variable is also economically significant. From Model 3, other things being
equal, a 1 percent change in the proportion held by insiders expected to
account for 1.273 percent change in dividend yield. This indicates that
insider ownership is an important determinant of corporate dividend policy
in Jordan, in particular the level of dividends. The negative and significant
relationship between dividend yield and insider ownership lends support to
Rozeff (1982), who proposed that the use of dividends as a bonding
mechanism to reduce agency costs is less important when there is higher
insider ownership.
From Hypothesis 2, the variables representing the firms ownership
structure are expected to influence its dividend policy. Four dummy
variables are used, FAML, STATE, INST, and MULT. Table 3 shows that in
Model 1 the coefficients of those variables have the expected signs but are
not significantly different from zero, with the exception of STATE. The
existence of the government or its agencies as a controlling shareholder
seems to influence the level of dividends paid. The positive significant
relationship between dividend yield and STATE is consistent with the
assertion that, ceteris paribus, state-controlled firms tend to pay more
dividends. This prediction is based on the argument that state-controlled
firms are often subject to a double set of agency costs since the ultimate
owners of these firms are the citizens. In addition, government entities may
have tax privileges so they prefer to invest in dividend-paying firms. The
marginal effects indicate that, a 1 percent increase in government ownership
in a firm would lead to an increase in the dividend yield by 0.610 percent.
These results are in line with those obtained by Gul (1999) for China, and
Gugler (2003) for Austria.
In testing the information content of dividends hypothesis (Hypothesis 3),
the firm share turnover ratio (TURN) is used as a proxy for information
asymmetry. As predicted, the result from Model 1 in Table 3 show that the
coefficient on TURN is negative, but the null hypothesis of zero coefficient
for the variable could not be rejected (t-statistic = -0.02). The evidence here,
therefore, does not provide support for the signaling hypothesis. One

59

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

possible explanation for this result is that the proxy for information
asymmetry is weak10.
Hypothesis 4 predicts that firms with high growth and investment
opportunities tend to retain their income to finance those investments, thus
paying less or no dividends. The variables MBR and AGE are used as
proxies for growth and investment opportunities. From the regression results
(Model 1) in Table 3, contrary to expectations, the coefficient of MBR is
positive and insignificant, whereas, as predicted, the coefficient of AGE is
positive and significantly different from zero. These findings indicate that
the market-to-book value ratio is not related to dividend yield, while firm
age is positively related to dividend yield. The positive coefficient on MBR
is analogous to that reported by Aivazian et al (2003) for Jordanian firms;
however, they obtain a significant coefficient in their results.
However, the hypothesized relationship between dividends and growth
and investment opportunities could not be rejected since the other proxy
variable for growth (AGE) is highly significant with t-statistics of 4.52, 3.93
and 3.83 in models 1, 2 and 3, respectively. The age of the firm is
consistently significant at the 1 percent level, suggesting that mature firms
tend to pay higher levels of dividends. From Model 3, as a firm becomes one
year older, the estimated increase in DYLD is 0.117 percent, ceteris paribus.
These results, reported in Table 3 , are consistent with prior research of
Manos and Green (2001) who found that the age of the firm and the payout
level are positively related for independent (non-group affiliated) Indian
firms. The result also provides empirical support for the maturity hypothesis
proposed by Grullon et al. (2002). That is, as firms become mature their
growth opportunities tend to decline resulting in lower capital expenditure
needs, and thus more cash flows are available for dividend payments. When
a mature firm has little or no investment opportunities, a high level of
dividends will reduce the discretionary resources available to managers that
could be wasted in perquisites or unprofitable projects. In other words,
dividends reduce the agency costs associated with free cash flow. The results
hence also provide support for the free cash flow hypothesis (Easterbrook,
1984, and Jensen, 1986).
To the best of the authors knowledge, this paper provides the first
attempt to document that the age of the firm is quadratically negatively
related to dividend yield as shown by the significant positive coefficients on
AGE and the significant negative coefficients on AGESQ (age squared) in
10
El-Khouri and Almwalla (1997) provided weak or no support for the signaling
hypothesis for Jordanian firms.

60

Journal of Economic & Administrative Sciences

December 2007

all regressions. This non-linear relationship between the age of the firm and
dividend yield (negative sign of AGESQ) may imply changes in a firms life
cycle, that is, movement from a lower growth phase to a higher growth
phase. In other words, when a firm finds new investment opportunities
(growth phase) it will pay less or no dividends because paying dividends is
no longer optimum. Again, this evidence reinforces the interpretation above
that the positive association between the age of the firm and dividend yield is
consistent with Grullon et al.s (2002) maturity hypothesis, and accordingly
with the free cash flow hypothesis of Easterbrook (1984) and Jensen (1986).
Another variable found to be a determinant of corporate dividend policy
in Jordan is the firm size (Hypothesis 5). As expected, Table 3 reports that
the coefficients on size (MCAP) are robustly positively correlated with
dividend yield. The t-statistics of the coefficients on MCAP for models 1, 2
and 3 are 3.62, 4.72 and 5.34, respectively, which are highly significant at
the 1 percent level. Firm size as measured by market capitalization is
positively related to DYLD with marginal effects of 0.480 percent. If for
example a firms market capitalization increases from JD 10 million to JD 15
million, the expected increase in its DYLD would be 0.195 percent11.
The positive and significant correlation between dividend yield and size
suggests that large firms are more able to pay dividends. This finding lends
support to the agency costs explanation. The intuition here is that the larger
the firm, the more difficult (costly) is the monitoring (i.e. the greatest the
agency problem). Thus, dividends could play a role in helping to alleviate
the agency problem. Also, the positive relation between dividend yield and
size supports the generally accepted view proposed by many finance scholars
that larger firms have easier access to capital markets (see, among others,
Lloyd et al., 1985, Holder et al., 1998, and Fama and French, 2002), and
have lower transaction costs associated with acquiring new financing as
compared to small firms (Alli et al., 1993).
Hypothesis 6 predicts a negative relationship between a firms financial
leverage and dividend yield. Table 3 shows that the coefficients on debt-toequity ratio (DER) are negative and statistically significant at the 5 and 1
percent levels. The t-statistics of the coefficients on DER for models 1, 2 and
3 are 2.59, 2.95 and 2.61, respectively. This suggests that firms with high
debt ratios tend to pay fewer dividends, and the level of dividend payments

11

[(LN 15,000,000 LN 10,000,000) * 0.480], see Washer and Casey (2004) for a
similar calculation.

61

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

thus seems to be negatively correlated with the level of financial leverage12.

From Model 3, if a firms financial leverage increases by 1 percent, the
estimated decrease in dividend yield will be 0.066. In relation to emerging
equity markets including Jordan, Aivazian et al. (2003) provided evidence
consistent with findings presented here. However, a word of caution should
be added. The sample used in this study consists of all companies listed on
the ASE including financial companies, which are highly levered.
Turning to the firms profitability (Hypothesis 7), the estimates of
earnings per share (EPS) are all positive and significant at the 1 percent
level, with the exception of Model 1 which is significant at the 5 percent
level. This suggests that profitability is a critical determinant of the level of
dividends paid by Jordanian firms. However, contrary to expectation, the
profitability variable does not seem to have large economic importance. An
increase in earnings per share of 1 percent would lead to an increase in
dividend yield of only 0.167 percent. The positive relationship between
profitability and dividends is well documented in the literature (see, for
example, Jensen et al, 1992, Han et al., 1999, and Fama and French, 2001,
2002).
The results from Table 3 show that the implementation of a 10 percent
tax rate on dividends in 1996 (Hypothesis 8) seems to have had no
significant influence on corporate dividend policy in Jordan. Consistent with
the tax-preference theory, the coefficients on DTAX are negative in models
1 and 2, but the hypothesis of zero coefficients for tax variable could not be
rejected. This result lends support for the findings of Omet (2004) that the
implementation of the tax had no impact on corporate dividend behavior in
Jordan13.
6. Summary and Conclusion
This paper has examined the main determinants of corporate dividend
policy in Jordan. Tobit specifications were used to examine the determinants
of the level or the amount of dividends paid. The results were obtained using
the maximum likelihood estimations of the random effects Tobit regressions.
12

Another variable (for non-financial firms) that was found to be significantly and
negatively related to dividend policy in Jordan is the asset tangibility. The results are
not reported here in order to save space but are available upon request.
13
It is worth mentioning here that when we performed the Wilcoxon test to see
whether dividend payout ratios differ between pre-tax (1989-1995) and post-tax
(1996-2000) periods, the null hypothesis that payout ratios for both periods have the
same distribution is rejected. The Wilcoxon test statistic yields a Z-statistic of 7.84
(P-value = 0.000).

62

Journal of Economic & Administrative Sciences

December 2007

Nested procedures were followed to arrive at the best-fitted model. In order

to ensure that our results are robust, several diagnostic tests were performed.
The data showed that ownership dispersion as measured by the natural
log of the number of stockholders (STOCK) seems to not be related to
dividend policy in Jordan. The fraction held by insiders (INSD), the second
proxy for the agency costs hypothesis, has negative impact on the level of
dividends paid. Similarly, the existence of government or its agencies
(STATE) in a firms ownership structure (controlling shareholder) affects
the amount of dividends (positively). Other variables of ownership structure
seem to have no influence on dividend policy. By and large, the evidence is
consistent with the agency costs explanation.
The firms age, size, and profitability positively and significantly affect
its dividend policy. The use of age, and especially age squared as proxies for
growth has not been used in empirical testing of dividend policy to the best
of the authors knowledge, and these results therefore suggest interesting
avenues for future research. Again, these findings are generally consistent
with the agency costs hypothesis and lend support to the pecking order
hypothesis. The analysis also found that a firms financial leverage is
significantly and negatively related to its dividend policy.
The study demonstrated that much of the existing theoretical literature on
dividend policy can be applied to an emerging capital market such as Jordan.
Many of the factors that were found to be significant in the determination of
dividend policy are the same as those found in developed capital markets.

63

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

References
Abrutyn, Stephanie, and Robert W. Turner (1990), Taxes and Firm's
Dividend Policies: Survey Results, National Tax Journal 43, 491-496.
Aivazian, Varouj, Laurence Booth, and Sean Cleary (2003), Do Emerging
Market Firms Follow Different Dividend Policies From U.S. Firms?,
The Journal of Financial Research 26, 371-387.
Alli, Kasim L., A.Qayyum Khan, and Gabriel G. Ramirez (1993),
Determinants of Corporate Dividend Policy: A Factorial Analysis, The
Financial Review 28, 523-547.
Anderson, G. J. (1986), An Application of the Tobit Model to Panel Data:
Modelling Dividend Behaviour in Canada, Working Paper, MacMaster
University.
Bali, Rakesh (2003), An Empirical Analysis of Stock Returns Around
Dividend Changes, Applied Economics 35, 51-61.
Barclay, Michael J. (1987), Dividends, Taxes, and Common Stock Prices:
The Ex-Dividend Day Behavior of Common Stock Prices Before the
Income Tax, Journal of Financial Economics 19, 31-44.
Barclay, Michael J., Clifford W. Smith, and Ross L. Watts (1995), The
Determinants of Corporate Leverage and Dividend Policies, Journal of
Applied Corporate Finance 7, 4-19.
Bartov, Eli, and Gordon M. Bodnar (1996), Alternative Accounting
Methods, Information Asymmetry and Liquidity: Theory and Evidence,
The Accounting Review 71, 397-418.
Bhattacharya, Sudipto (1979), Imperfect Information, Dividend Policy, and
"the Bird in the Hand" Fallacy, Bell Journal of Economics 10, 259-270.
Black, Fischer, and Myron S. Scholes (1974), The Effects of Dividend
Yield and Dividend Policy on Common Stock Prices and Returns,
Journal of Financial Economics 1, 1-22.
Booth, Laurence, Varouj Aivazian, Asli Demirguc-Kunt, and Vojislav
Maksimovic (2001), Capital Structure in Developing Countries, Journal
of Finance 56, 87-130.
Brennan, Michael J. (1970), Taxes, Market Valuation and Corporate
Financial Policy, National Tax Journal 23, 417-427.
Breusch, T. S., and A. R. Pagan (1979), A Simple Test for
Heteroscedasticity and Random Coefficient Variation, Econometrica 47,
1287-1294.
Conroy, Robert M., Kenneth M. Eades, and Robert S. Harris (2000), A Test
of the Relative Pricing Effects of Dividends and Earnings: Evidence from
Simultaneous Announcements in Japan, Journal of Finance 55, 11991227.
64

Journal of Economic & Administrative Sciences

December 2007

Denis, David J., Diane K. Denis, and Atulya Sarin (1994), The Information
Content of Dividend Changes: Cash Flow Signaling, Overinvestment,
and Dividend Clienteles, Journal of Financial and Quantitative Analysis
29, 567-587.
Deshmukh, Sanjay (2003), Dividend Initiations and Asymmetric
Information: A Hazard Model, Financial Review 38, 351-368.
Dhaliwal, Dan S., Merle Erickson, and Robert Trezevant (1999), A Test of
the Theory of Tax Clienteles for Dividend Policies, National Tax
Journal 52, 179-194.
Easterbrook, Frank H. (1984), Two Agency Costs Explanations of
Dividends, American Economic Review 74, 650-659.
El-Khouri, Ritab, and Mona Almwalla (1997), The Effect of Dividend
Changes on Security Prices: The case of Jordanian Companies, Abhath
Al-Yarmouk 13, 87-94.
Fama, Eugene F., and Kenneth R. French (2001), Disappearing Dividends:
Changing Firm Characteristics or Lower Propensity to Pay?, Journal of
Financial Economics 60, 3-43.
Fama, Eugene F., and Kenneth R. French (2002), Testing Trade-Off and
Pecking Order Predictions About Dividends and Debt, The Review of
Financial Studies 15, 1-33.
Farinas, Jose C., and Lourdes Moreno (2000), Firms' Growth, Size and
Age: A Nonparametric Approach, Review of Industrial Organization 17,
249-265.
Godfrey, Leslie G. (1978), Testing for Multiplication Heteroskedasticity,
Journal of Econometrics 8, 227-236.
Gonedes, Nicholas J. (1978), corporate Signalling, External Accounting,
and Capital Market Equilibrium: Evidence on Dividends, Income, and
Extraordinary Items, Journal of Accounting Research 16, 26-79.
Gordon, Myron J. (1959), Dividends, Earnings, and Stock Prices, Review
of Economics and Statistics 41, 99-105.
Greene, William H. (1999), Marginal Effects in the Censored Regression
Model, Economics Letters 64, 43-49.
Grullon, Gustavo, Roni Michaely, and Bhaskaran Swaminathan (2002), Are
Dividend Changes a Sign of Firm Maturity?, Journal of Business 75,
387-424.
Gugler, Klaus (2003), Corporate Governance, Dividend Payout Policy, and
the Interrelation Between Dividends, R&D, and Capital Investment,
Journal of Banking and Finance 27, 1297-1321.

65

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

Gul, Ferdinand A. (1999), Government Share Ownership, Investment

Opportunity Set and Corporate Choices in China, Pacific-Basin Finance
Journal 7, 157-172.
Han, Ki C., Suk Hun Lee, and David Y. Suk (1999), Institutional
Shareholders and Dividends, Journal of financial and Strategic
Decisions 12, 53-62.
Hendry, David F. (1995). Dynamic Econometrics (Oxford University
Press, New York).
Holder, Mark E., Frederick W. Langrehr, and J. Lawrence Hexter (1998),
Dividend Policy Determinants: An Investigation of the Influences of
Stakeholder Theory, Financial Management 27, 73-82.
Huang, Ho-Chuan River (2001), Bayesian Analysis of the Dividend
Behaviour, Applied Financial Economics 11, 333-339.
Huergo, Elena, and Jordi Jaumandreu (2004), Firms' Age, Process
Innovation and Productivity Growth, International journal of Industrial
Organization 22, 541-559.
Jensen, Gerald R., Donald P. Solberg, and Thomas S. Zorn (1992),
Simultaneous Determination of Insider Ownership, Debt, and Dividend
Policies, Journal of Financial and Quantitative Analysis 27, 274-263.
Jensen, Michael C. (1986), Agency Costs of Free Cash Flow, Corporate
Finance, and Takeovers, American Economic Review 76, 323-329.
Jensen, Michael C., and William H. Meckling (1976), Theory of the Firm:
Managerial Behavior, Agency Costs and Ownership Structure, Journal
of Financial Economics 3, 305-360.
John, Kose, and Joseph Williams (1985), Dividends, Dillution, and Taxes:
A Signalling Equilibrim, Journal of Finance 40, 1053-1070.
Kim, Byeong Soo, and G. S. Maddala (1992), Estimation and Specification
Analysis of Models of Dividend Behavior Based on Censored Panel
Data, Empirical Economics 17, 111-124.
Lease, Ronald C., Kose John, Avner Kalay, Uri Loewenstein, and Oded H.
Sarig (2000). Dividend Policy: Its Impact on Firm Value (Harvard
Business School Press, Boston, Massachusttes).
Leuz, Christian, and Robert E. Verrecchia (2000), The Economic
Consequences of Increased Disclosure, Journal of Accounting Research
38, Supplement, 91-124.
Lewellen, Wilbur G., Kenneth L. Stanley, Ronald C. Lease, and Gary G.
Schlarbaum (1978), Some Direct Evidence on the Dividend Clientele
Phenomenon, Journal of Finance 33, 1385-1399.

66

Journal of Economic & Administrative Sciences

December 2007

Lie, Erik (2000), Excess Funds and Agency Problems: An Empirical Study
of Incremental Cash Disbursements, Review of Financial Studies 13,
219-248.
Lintner, John (1956), Distribution of Incomes of Corporations Among
Dividends, Retained Earnings, and Taxes, American Economic Review
46, 97-113.
Litzenberger, Robert H., and Krishna Ramaswamy (1979), The Effect of
Personal Taxes and Dividends on Capital Asset Prices, Journal of
Financial Economics 7, 163-195.
Lloyd, William P., John S. Jahera, and Daniel E. Page (1985), Agency
Costs and Dividend Payout Ratios, Quarterly Journal of Business and
Economics 24, 19-29.
Manos, Ronny, and Christopher Green (2001), Business Group and
Dividend Policy: Evidence on Indian Firms, Working Paper No.33,
(Institute for Development Policy and Management, University of
Manchester).
Michaely, Roni, Richard H. Thaler, and Kent L. Womack (1995), Price
Reactions to Dividend Initiations and Omissions: Overreaction or Drift?,
Journal of Finance 50, 573-608.
Miller, Merton H., and Franco Modigliani (1961), Dividend Policy,
Growth, and the Valuation of Shares, Journal of Business 34, 411-433.
Miller, Merton H., and Kevin Rock (1985), Dividend Policy Under
Asymmetric Information, Journal of Finance 40, 1031-1051.
Miller, Merton H., and Myron S. Scholes (1982), Dividend and Taxes:
Some Empirical Evidence, Journal of Political Economy 90, 1118-1141.
Morgan, Gareth, and Stephen Thomas (1998), Taxes, Dividend Yields and
Returns in the UK Equity Market, Journal of Banking and Finance 22,
405-423.
Myers, Stewart C. (1984), The Capital Structure Puzzle, Journal of
Finance 39, 575-592.
Myers, Stewart C., and Nicholas S. Majluf (1984), Corporate Financing and
Investment Decisions When Firms Have Information That Investors Do
Not Have, Journal of Financial Economics 13, 187-221.
Nissim, Doron, and Amir Ziv (2001), Dividend Changes and Future
Profitability, Journal of Finance 56, 2111-2133.
Omet, Ghassan (2004), Dividend Policy Behaviour in the Jordanian Capital
Market, International Journal of Business 9, 287-299.
Pettit, R. Richardson (1972), Dividend Announcements, Security
Performance, and Capital Market Efficiency, Journal of Finance 27,
993-1007.

67

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

Pettit, R. Richardson (1977), Taxes, Transactions costs and the Clientele

Effect of Dividends, Journal of Financial Economics 5, 419-436.
Poterba, James M., and Lawrence H. Summers (1984), New Evidence That
Taxes Affect the Valuation of Dividends, Journal of Finance 39, 13971415.
Rodriguez, Alicia C., Migule A. Molina, Ana L. G. Perez, and Urbano M.
Hernandez (2003), Size, Age and Activity Sector on the Growth of the
Small and Medium Firm Size, Small Business Economics 21, 289-307.
Rozeff, Michael S. (1982), Growth, Beta and Agency Costs as
Determinants of Dividend payout Ratios, The Journal of Financial
Research 5, 249-259.
Scholz, John Karl (1992), A Direct Examination of the Dividend Clientele
Hypothesis, Journal of Public Economics 49, 261-285.
StataCorp. (2003), Stata Statistical Software: Release 8.0, (Stata
Corporation, College Station, Texas).
Verbeek, Marno (2000). A Guide to Modern Econometrics, (John Wiley &
Sons,Ltd, Brisbane).
Watts, Ross (1973), The Information Content of Dividends, Journal of
Business 46, 191-211.
Washer, Kenneth M., and K. Michael Casey (2004), A Note on the
Determinants of Equity Payouts, Journal of Applied Business Research
20, 83-90.
Yoon, Pyung Sig, and Laura T. Starks (1995), Signalling, Investment
Opportunities, and Dividend Announcements, Review of Financial
Studies 8, 995-1018.

68

INSD

STOCK
-.166

-.362

1.000

.484

.107

1.000

.013

1.000
-.161

-.144 1.000
1.000

STOCK INSD FAML STATE INST MULT AGE AGESQ MBR DER TURN DTAX MCAP EPS TANG

FAML
.405

1.000
.178 1.000

1.000

-.145

1.00
1.000
-.151

-.156

.434

1.000

-.074

-.073

STATE

.017

.014

-.160

INST

.151

.037

-.060

.545 1.000

.039

-.102

-.009

-.171

.121

-.294

-.075

.388

.031

-.095 -.097

.306

.480

-.058

.007 -.004

.094

-.236

.018

.368

MULT

.012

.042

1.000

-.090 -.160

.315

.157

.117

.080

.068

-.017

.040

.129

.016

.402

-.062

-.160

-.030

-.024

-.053

-.092

-.055 -.056

.080 -.084 1.000

.241
.215

-.053

.244

.038

1.51

-.090

.098

.079

.058

.112

1.21

AGE
-.148
-.216

-.202

.094

1.11

.946 1.000

-.007
.047

-.069

.014

1.45

.104

MBR
.085

.085

-.113 -.023

1.81

-.084

DER
-.087

-.022

-.037 -.100

1.42

.365

TURN
.178

.003

2.34 2.32

-.042

DTAX
-.020

-.124

2.20

.170

MCAP

.018

1.61

.163

EPS

.010

2.35

AGESQ

TANG

1.32

1.14 1.11
VIF

69

December 2007
Journal of Economic & Administrative Sciences

Appendix 1
Correlation Matrix and Variance Inflation Factors (VIF)
for the Explanatory Variables

Dr. Husam-Aldin Nizar Al-Malkawi

December 2007

:
.

-

) (panel data
1989 2000


) (Tobit


.

70

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.