Professional Documents
Culture Documents
Hosam en
Hosam en
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&M’s 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 firm’s
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
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 firm’s 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.
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
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.
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 χ 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:
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 STOCK: natural log of number of common Positive
costs stockholders
INSD: percentage held by insiders Negative
H2: Ownership FAML: family dummy = 1 if firm is family Negative
structure* owned, and 0 otherwise.
STATE: state dummy = 1 if firm is owned by the Positive
government or its agencies, and 0 otherwise.
INST: institution dummy = 1 if firm is owned by Positive
an institution, and 0 otherwise.
MULT: multiple owners dummy = 1 if firm has Positive
more than one type of owners who control at least
10% of the stock, and 0 otherwise.
53
Dr. Husam-Aldin Nizar Al-Malkawi December 2007
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):
yit = x'it β + α i + ε it ,
*
(2)
with the observed dependent variable being such that:
if yit ≤ 0 ,
*
yit = 0,
if yit > 0 .
* *
= yit ,
The random effects α i and the error term εit are assumed to be
iid N (0, σ α2 ) and iid N (0, σ ε2 ) , respectively, and independent of xi1, …
xiT. To estimate β 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 ( χ 2 ) statistics derived
from the Wald test statistic. The Wald test statistic follows a χ 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
See StataCorp (2003) and Verbeek (2000) for the likelihood function.
7
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. Min Max CV Obs.
DYLD 0.027 0.035 0 0.400 1.282 1316
STOCK 3.068 0.746 1.114 5.407 0.243 885
INSD 0.289 0.246 0 0.945 0.851 936
FAML 0.303 0.460 0 1 1.517 1181
STATE 0.210 0.408 0 1 1.941 1181
INST 0.539 0.499 0 1 0.926 1181
MULT 0.368 0.483 0 1 1.311 1182
AGE 15.738 13.533 0 70 0.860 1598
AGESQ 430.717 703.841 0 4900 1.6341 1598
MBR 1.290 1.025 -0.020 12.670 0.795 1382
DER 2.143 6.939 -177.242 69.270 3.238 1511
TURN 0.299 0.671 0 7.993 2.243 1491
DTAX 0.500 0.500 0 1 1.000 1592
MCAP 15.731 1.403 11.156 21.356 0.089 1320
EPS 0.282 1.774 -8.388 28.299 6.285 1514
NONFIN 0.738 0.440 0 1 0.597 1920
Note: Variables are defined in Table 1.
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 ( ρ ), 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 Chi-
square 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 χ2 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
For computing the marginal effects see, for example, Greene (1999).
9
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 Model 3
Marginal Marginal Marginal
Independ. Coefficient Effects Coefficient Effects Coefficient Effects
Variables Estimates (%) Estimates (%) Estimates (%)
Constant -0.266*** -0.278*** -0.300***
(-4.10) (-5.43) (-6.25)
STOCK -0.008 -0.241 — — — —
(-1.01)
-0.060*** -0.041** -0.043**
INSD -1.795 -1.237 -1.273
(-2.61) (-2.04) (-2.29)
-0.004
FAML (-0.43) -0.130 — — — —
0.024* 0.021* 0.019*
STATE (1.93) 0.770 (1.90) 0.672 (1.83) 0.610
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 author’s 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 firm’s 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 firm’s 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 firm’s financial
leverage and dividend yield. Table 3 shows that the coefficients on debt-to-
equity 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
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, 1199-
1227.
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
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
68
December 2007
STOCK INSD FAML STATE INST MULT AGE AGESQ MBR DER TURN DTAX MCAP EPS TANG
STOCK 1.000
INSD -.362 1.000
FAML -.166 .107 1.000
for the Explanatory Variables
69
AGESQ .163 .170 -.042 .365 -.084 .104 .946 1.000
Journal of Economic & Administrative Sciences
MBR -.148 .215 -.062 .117 .018 .031 .039 .017 1.000
DER -.007 .098 -.053 .040 .012 -.058 .121 .151 .434 1.00
TURN .085 -.216 -.053 -.160 -.090 -.160 -.095 -.097 -.102 .014 1.000
DTAX -.087 .047 .079 -.092 .129 .080 .007 -.004 -.294 .037 -.074 1.000
MCAP .178 .085 -.202 .244 -.030 .016 .315 .368 .306 -.009 -.160 -.151 1.000
EPS -.020 -.022 -.069 .058 -.055 -.056 .068 .042 .094 -.075 -.060 -.073 .178 1.000
TANG .010 .018 -.124 .003 -.037 -.100 -.113 -.023 .014 .094 .112 .038 .080 -.084 1.000
VIF 1.32 2.35 1.61 2.20 2.34 2.32 1.42 1.81 1.45 1.11 1.21 1.51 1.14 1.11
Dr. Husam-Aldin Nizar Al-Malkawi December 2007
ﻤﻠﺨﺹ ﺍﻟﺩﺭﺍﺴﺔ
ﺘﻬﺩﻑ ﻫﺫﻩ ﺍﻟﺩﺭﺍﺴﺔ ﺇﻟﻰ ﻤﻌﺭﻓﺔ ﻤﺤﺩﺩﺍﺕ ﺴﻴﺎﺴﺔ ﺘﻭﺯﻴﻊ ﺍﻹﺭﺒﺎﺡ ﻟﻠﺸﺭﻜﺎﺕ ﺍﻷﺭﺩﻨﻴﺔ ،ﻭﺍﻋﺘﻤﺩﺕ
ﺍﻟﺩﺭﺍﺴﺔ ﻋﻠﻰ ﺒﻴﺎﻨﺎﺕ ﺘﺠﻤﻊ ﺒﻴﻥ ﻗﻴﻡ ﻤﻘﻁﻌﻴﺔ ﻭﻗﻴﻡ ﺴﻠﺴﻠﺔ ﺯﻤﻨﻴﺔ ) ،(panel dataﻭ ﻗﺩ ﺸﻤﻠﺕ
ﺍﻟﺩﺭﺍﺴﺔ ﺠﻤﻴﻊ ﺍﻟﺸﺭﻜﺎﺕ ﺍﻟﻤﺩﺭﺠﺔ ﻓﻲ ﺴﻭﻕ ﻋﻤﺎﻥ ﺍﻟﻤﺎﻟﻲ ﺨﻼل ﺍﻟﻔﺘﺭﺓ 1989ـ ،2000
ﻭﺍﺴﺘﺨﺩﻤﺕ ﺍﻟﺩﺭﺍﺴﺔ ﻤﻨﻬﺞ ﺍﻻﻨﺘﻘﺎل ﻤﻥ ﺍﻟﻨﻤﻭﺫﺝ ﺍﻟﻌﺎﻡ ﺇﻟﻰ ﺍﻟﺨﺎﺹ ﻟﻼﺨﺘﻴﺎﺭ ﺒﻴﻥ ﺍﻟﻔﺭﻀﻴﺎﺕ
ﺍﻟﻤﺘﻨﺎﻓﺴﺔ ،ﻭﻟﺩﺭﺍﺴﺔ ﻤﺤﺩﺩﺍﺕ ﻜﻤﻴﺔ ﺃﻭ ﻤﺴﺘﻭﻯ ﺍﻹﺭﺒﺎﺡ ﺍﻟﻤﻭﺯﻋﺔ ﺘﻡ ﺍﺴﺘﺨﺩﺍﻡ ﻨﻤﻭﺫﺝ ﺘﻭﺒﺕ
) ،(Tobitﻭﺃﻅﻬﺭﺕ ﺍﻟﻨﺘﺎﺌﺞ ﺃﻥ ﻋﻭﺍﻤل ﻤﺜل ﻨﺴﺒﺔ ﻤﻠﻜﻴﺔ ﺍﻹﺩﺍﺭﺓ ﻭﻭﺠﻭﺩ ﺍﻟﺤﻜﻭﻤﺔ ﻓﻲ ﻫﻴﻜل ﻤﻠﻜﻴﺔ
ﺍﻟﺸﺭﻜﺔ ﻭﻜﺫﻟﻙ ﺤﺠﻡ ﻭﻋﻤﺭ ﻭﺭﺒﺤﻴﺔ ﺍﻟﺸﺭﻜﺔ ﺘﻠﻌﺏ ﺩﻭﺭﹰﺍ ﻤﻬﻤ ﹰﺎ ﻓﻲ ﺘﺤﺩﻴﺩ ﺴﻴﺎﺴﺔ ﺘﻭﺯﻴﻊ ﺃﺭﺒﺎﺡ
ﺍﻟﺸﺭﻜﺎﺕ ﺍﻷﺭﺩﻨﻴﺔ ،ﻭ ﺨﻠﺼﺕ ﺍﻟﺩﺭﺍﺴﺔ ﺇﻟﻰ ﺩﻋﻡ ﻨﻅﺭﻴﺔ ﺍﻟﻭﻜﺎﻟﺔ ﻭﺍﻻﺘﺴﺎﻕ ﻋﻤﻭﻤ ﹰﺎ ﻤﻊ ﻨﻅﺭﻴﺔ ﺍﻭﻟﻭﻴﺎﺕ
ﺍﻟﺘﻤﻭﻴل ،ﺒﻴﻨﻤﺎ ﻟﻡ ﺘﺩﻋﻡ ﻨﺘﺎﺌﺞ ﺍﻟﺩﺭﺍﺴﺔ ﻨﻅﺭﻴﺔ ﻭﺠﻭﺩ ﻤﺤﺘﻭﻱ ﻤﻌﻠﻭﻤﺎﺘﻲ ﻓﻲ ﺘﻭﺯﻴﻌﺎﺕ ﺍﻷﺭﺒﺎﺡ.
70