Monetary policy restriction and dividend behaviour of Pakistani firms: an empirical analysis.
Ashraf, Muhammad Shahzad ; Mohsin, Hasan M.
1. INTRODUCTION
Dividend behaviour has extensively been reviewed by many
researchers from time to time across different countries. Empirical
evidences observed in most of the studies reveal equivocal results about
dividend theories [Bhattacharyya (2007)]. Since, in absence of any
unanimous findings, need for future research has not been restricted,
theoretically. In developing countries like Pakistan, where limited
research is available on corporate dividend policy, need for future
research is more looked for. Most of the available research papers,
address only firm specific determinants of dividend policy. Do
macroeconomic variables influence corporate financing decisions? The
need to address this question is the prime motive of this research
paper. Major objective of this paper is to observe dividend behaviour of
listed firms in Pakistan under monetary policy restrictions and this is
the first attempt of its kind in Pakistan to the best of my Knowledge.
This study is very relevant in present scenario since State Bank of
Pakistan (SBP) has been persistently pursuing restricted monetary policy
since 2005 to control inflation.
Miller and Modigliani are the focal names when we start thinking
about dividend theories. MM theory of irrelevance, as quoted by Van Home
(1998), based upon assumption of perfect capital market, states that
dividend policy has no affect upon value of the firm. Nonetheless, when
markets are not perfect, as it is, dividend policy does matter and
affect value of the firm as both managers and investor favour dividend
payments as validated by many researchers.
2. MONETARY POLICY IN PAKISTAN
Pakistan is an emerging economy. After deregulation and
privatisation, in 1990s, studying macro variables is of paramount
importance and interest. Pakistan started liberalisation of the economy
and also adopted market based monetary policy system. Main motive of
monetary policy is to ensure low inflation along with sustainable
economic growth. It regulates cost and allocates money and credit in the
economy. Before liberalisation, interest rates were used to be fixed by
the regulatory bodies whereas after liberalisation, State Bank of
Pakistan's (SBP) open market operation is announced to be the major
instrument of monetary policy in 1995.
In year 2001, although Pakistan put efforts to bring macroeconomic
fundamentals back on track its monetary policy had to be tempered due to
conflicting economic goals. In overall terms monetary policy remained
tight in year 2001 [SBP (2001)]. Macroeconomic discipline achieved in
year 2001 led to easing of monetary policy in year 2002. Trade deficit
was much lower than year 2001 and inflation was down to 3.5 percent [SBP
(2002)]. Year 2003 again witnessed strong boost rising real GDP growth
to 5.1 percent level. The scale and depth of improvement in year 2003 is
much higher than year 2002. SBP increased market liquidity by lowering
discount rate substantially [SBP (2003)].
Year 2004 again witnessed loose monetary stance being adopted by
SBP since couple of years. It not only led to an immense increase in
aggregate demand along with increase in real GDP growth to over 6
percent but also contributed to growing inflationary pressures in the
country [SBP(2004)]. In year 2005 there is an important transition in
monetary policy i.e. from accommodative to aggressive tightening,
although SBP had started raising benchmark interest rates early in year
2004. Inflation was the main driving force behind this move [SBP(2005)].
This move continued in year 2006 although the chief policy variable,
i.e. discount rate remained same. However, State bank focused on
draining excess liquidity from the market [SBP(2006)].
In order to temperate demand pressures in the country, SBP
sustained tight monetary policy in year 2007 [SBP(2007)]. Increased
inflationary pressures led SBP to continue this policy in year 2008 and
2009 also [SBP(2008) and SBP (2009)].
3. LITERATURE REVIEW
Starting from John Lintner (1956), noticeable work upon dividend
behaviour and policy has been carried out in different parts of the
world. Lintner, in his research to know how firms decide to distribute
their earnings revealed that current earnings and lagged dividends the
foremost factors to be considered in dividend decisions. He surveyed 600
firms and on basis of interviews of officials developed a model and
tested further. Results also reveal that firms tend towards their target
payout ratios by partial adjustments reflecting soothing behaviour.
Following Lintner, many researchers explored other dividend
determinants by extending/modifying Lintner's model. Dividend
policies of individual firms were studied by Fama and Babiak (1968) by
modifying Lintner's model. They deleted constant and added lagged
profits in the model. Al-Najjar (2009) studied dividend behaviour of
Jordanian firms and found that factors affecting dividend policy in
developed countries are same as in case of Jordan. Results of his study
also validated Lintner's Model. Author used Pooled and Panel logit
and tobit models on 86 non-financial listed companies. Ahmed and Javaid
(2009) observed determinants of dividend policy in Pakistan along with
testing of Lintner's model of dividend soothing using panel data of
320 non-financial firms. Results reveal that firms rely, mainly, on
current earnings and past dividends for dividend decisions along with
instability towards dividend soothing.
Do foreign affiliates of a multinational firm depict same dividend
behaviour like of a parent company to its common shareholders?
Interesting work completed by Desai, et al (2001), reveals that
majority-owned foreign affiliates of American companies portray same
dividend policy as of domestic companies paying dividends to diffused
common shareholders. Musa and Fodio (2009) by using a model developed by
Musa, studied dividend behaviour of Nigerian firms revealing that
previous dividend, current earnings, cash flow, investment and net
current assets have significant impact on dividend policy. Dividend
stability has been observed by Al-Yahyaee, et al. (2010) in Oman by
working on a selected sample firms using Lintner's model.
Eriotis (2005) examined, in Greece, the effects of distributed
earnings, size of the firm and changes in dividend and distributed
earnings from the last year. Data comprises of a sample of 149 firms for
a period of 5 years. Firms prefer to distribute each year a rather
constant dividend, by adjusting to distributed earnings and size.
Abor and Bopkin (2010) observed effects of investment opportunities
and some other financial variables including some macro variables
(inflation rate and GDP) as control variables. Study is based upon a
sample of 34 emerging market countries, including Pakistan, for a period
of 17 years from 1990-2006. Authors observed significant relationship
between potential investment opportunity and dividend policy. Rozeff
(1982) studied impact of agency costs, Beta (a proxy for financial and
operating leverage) and growth of a firm, upon dividend policy. He
observed significant results for all these variables.
Dividend behaviour similarity between US firms and developing
countries (eight emerging markets including Pakistan), observed by
Aivazian, et al. (2003). However, sensitivity of variables differs as
country specific situations may effect. Interesting result is that in
emerging markets, firms found to give higher dividend payments than US
firms, although these face more financial constraints, relatively.
Garrett and Priestly (2000) worked on aggregate stock market data of US
firms with extended Lintner model and claimed that target dividends are
a function of permanent earnings and lagged prices. They introduced new
model which assumes that managers tend to minimise costs while pursuing
for target dividends. Regarding Signalling theory, authors concluded
that dividends signal about positive shocks to current permanent
earnings and not to future pennanent earnings.
Bhattacharyya, et al. (2008) worked in a different dimension on a
hypothesis that high quality agents (managers) have access to more
positive NPV projects rather than low quality agents. High quality agent
demands higher compensation. Model based upon this hypothesis, had been
tested for Canadian firms over the period from 1993-95 using tobit
regression analysis. Canadian firms found to support this hypothesis.
Some authors have worked, specifically, on dividend determinants
related to ownership of firms. In Pakistan, ownership structure has
significant impact upon dividend payout policy where as cash flows have
insignificant impact. It is finding of a study by Afza and Mirza (2010),
upon 100 companies listed at KSE. Board of directors act as a tool to
monitor management and hence helps to resolve agency problems. However,
composition of board does matter and have influence on dividend policy
accordingly. In same way ownership structure also dominates corporate
decisions involving voting requirements. Higher the concentration of
ownership, higher will be chances of exploitation of minority
shareholder's rights. Abdelsalam, et al. (2008) examined above both
elements in Egypt for a pooled data of 50 firms for three years using
logit and tobit models. He found significant association between
institutional ownership and dividend policy and insignificant for board
composition.
In family controlled firms, independent directors have significant
impact on dividend policy. Atmaja (2010) observed this finding in his
study upon Australian firms over the period from 2002-2005 using panel
(random effects) regression. Pandy (2001) observed sensitivity of
dividend behaviour of Malaysian firms, using multi-logit analysis, to
changes in earnings. In addition to observe sensitivity, application of
Lintner's framework depicted less stable dividend policies. Four
possible behaviours i.e.: (a) omission; (b) decrease; (c) increase; and
(d) no change, observed to three possible changes in earnings i.e.
(increase, decrease and negative earnings).
Not only internal but external factors, like monetary policy, do
affect financial decisions of the firms. Pandey and Bhat (2007)
observed, in India, that monetary policies have significant influence
upon dividend behaviour and 5 percent to 6 percent reduction observed in
payout. Authors tested extended Lintner's model using GMM estimator
for data of 571 firms over a period of 8 years. Ameer (2008) worked out
upon determinants of dividend policy of Malaysian Banks. He used ordered
probit modelling technique, in addition to check speed of adjustment
through Lintner model, to check flexibility of dividend policy to
certain variables. In addition to firm specific, author observed
monetary policy effects on dividend payout.
Goddard, et al. (2006) tested smoothing and signalling hypothesis
upon 137 UK firms, over the period from 1970 to 2003. He observed
contemporary relationship between dividends, prices and earnings. Some
evidence in favour of both hypotheses has been revealed by causality
tests.
Hussainey and Eisa (2009) in addition to work on dividend
signalling hypothesis also included signalling behaviour of voluntary
disclosure statements incorporated in annual reports. By using event
study methodology, they observed behaviour of 33 UK non-financial firms
after a decline in their sustained earnings growth. Findings do not
support dividend signalling hypothesis however support disclosure
signalling behaviour. Nissim and Ziv (2001) examined signalling
hypothesis and revealed, empirically, that dividend changes signal
profitability level in subsequent years.
Bhattacharyya (2007) argues continuity of search for more
elucidations as he observed equivocal empirical results of dividend
theories. He collected empirical work done based upon clientele,
signalling and agency hypothesis and extracted stylised facts also.
Dividend policies are affected by legal corporate framework of a
particular country. Countries having better legal protection for
minority shareholders, observe higher payouts. Porta, et al. quoted
their findings by doing empirical work over a cross section of 4000
firms of 33 countries.
Baker and Wurgler (2004) introduced catering theory of dividend.
Authors proposed that when investors pay premium on stock price, they,
infact, anticipate dividends and managers cater to them by paying
dividends and vice versa. Empirical findings confirm to their theory.
In addition to explicit claims, there are implicit claims, upon an
organisation, by non-investor stakeholders (e.g. employees, customers,
vendors etc.). These stakeholders may suffer costs if a firm runs out of
business i.e. cost of jobs search by employees, increased maintenance
costs for customers etc. Firms offering more implicit guarantees are
more valued. These have to maintain higher liquidity levels to pay off
potential implicit claims. Hence being more conservative, trying to
avoid from financial distress, use more equity. Dividend payout is less
in these firms. Although this stakeholder's theory is not very
persuasive as firms maintaining this level of excellence earn higher
profits and hence higher payouts. Holder, et al. (1998) tested this
theory and validated existence of this relationship.
Michel (1979) observed industry impact upon dividend policy in
United States. There are similarities in structural characteristics of
firms of an industry. Hence, different industries would have varying
influences upon dividend policies as would have different investment
opportunities. Empirical results, concluded by Michel (1979), confirm
the assumption.
4. MODEL AND METHODOLOGY
The Lintner dividend model can be assumed as the mother of all
dividend behaviour models. Almost all researches on dividend behaviour
are based upon this model, modified model or its enhanced versions and
the same practice would be followed by us. However, our study focuses on
dividend payment behaviour of Pakistani firms in tight monetary policy
regime. In perfect capital market, as Miller and Modigliani proposed,
cost of internally generated and external funds would not be different.
But we are living in imperfect world and hence above proposition would
not stand valid. There would be an information asymmetry between
borrowers and lenders. A moral hazard of default would prevail.
Investors have to incur project monitoring costs and also demand risk
premium, hence cost of external funds will be greater than internal
funds. At times of restricted monetary policy, cost of external funds
increases and firms prefer to utilise internal funds provided that firms
have investment opportunities. To maintain internal reserves, for
internal financing, dividend payout decreases. Although firms may go for
external financing (debt), in case of monetary policy restriction, if it
has yet to attain optimum level of capital structure and want to gain
tax benefits of interest expense.
Below mentioned is our proposed replicated model of Pandey and Bhat
(2007).
[D.sub.it] = [varies] + [[beta].sub.1][E.sub.it] +
[[beta].sub.2][D.sub.it-1] + [[beta].sub.3][D.sub.it-2] +
[[beta].sub.4]M[R.sub.t]{[D.sub.it-1]) +
[[beta].sub.5]M[R.sub.t]{[D.sub.it-2]) + [[theta].sub.i] +
[[phi].sub.t] + [[mu].sub.it]
[D.sub.it] = Dividend for firm i in time t
[E.sub.it] = Earnings (net profit) of firm i in time t
[D.sub.it-1] = Dividend in lagged year 1
[D.sub.it-2] = Dividend in lagged year 2
M[R.sub.t] = Monetary restriction in year t - A dummy variable.
[[theta].sub.i] accounts for individual firm effect while
[[phi].sub.t] measures time-based effect. Earning is a major and
dominant dividend determinant for every firm. Lagged dividends, do have
impact upon dividend payout as firms tend to move gradually to target
dividends i.e. dividend soothing. Pandey and Bhat (2007) used two lagged
periods rather one. Monetary restriction is a dummy variable with value
1 in case of tight monetary policy and zero (0) otherwise. Identifying
monetary policy with a only one variable, like discount rate, lending
rates or money supply may not be very explanatory. Furthermore, in
Pakistan monetary policy announcements are twice and thrice times a year
from 2006 onwards and concluding a policy for whole year may be
difficult. Hence, rather using, discount rate, lending rates or money
supply etc. we use State Bank of Pakistan's Annual reports for
identification of restricted monetary policies in respective years. In
annual reports, a single line sentence, describing overall monetary
policy stance in that particular year, is available. From year 2001 to
2009, monetary policy is loose only in three years from 2002 to 2004.
Balanced panel data of 900 observations (100 cross section firms
for 9 years) is being used in estimation of above model. Unlike cross
section or time series, panel data encompass certain advantages. Gujrati
(2003) has cited these advantages quoted by Baltagi (1995). Panel data
takes heterogeneity into account through individual firm effect. A
combination of cross section and time series observations give more rich
information, more variability, less collinearity among variables, more
degrees of freedom and more efficiency. Panel data better detects and
measures effects that are not observable in pure cross section or time
series. The dynamics of change are better observed through panel data as
repeated cross sections of observation are studied. In our estimation
model, panel data would also serve best to study effect of monetary
policy restrictions over the years and the dynamics of change in
dividend payments.
4.1. Hypothesis
[H.sub.0] : [[beta].sub.1] = 0, [[beta].sub.2] = 0, [[beta].sub.3]
= 0, [[beta].sub.4] = 0, [[beta].sub.5] = 0 [H.sub.1] : [[beta].sub.1]
> 0, [[beta].sub.2] > 0, [[beta].sub.3] > 0, [[beta].sub.4]
< 0, [[beta].sub.5] < 0
Above proposed is a dynamic model with lagged dependent variable as
explanatory variable. Dynamic models are bit difficult to estimate.
Dynamic models estimation is recommended through usage of GMM estimator
as literature enforces it.
5. SAMPLE AND DATA
A sample of 100 firms listed at Karachi Stock Exchange has been
selected. To ensure equal participation of each industry, in sample,
equal sample size (proportionately to respective population size) from
each group has been selected. Source of panel data for the period of
2001-2009, is State Bank of Pakistan.
For industry classification, State bank's classification,
based upon economic grouping, has been used. State Bank of Pakistan has
classified firms in nine economic groups based upon logical similarity
in nature of business. Only non-financial firms are
being analysed like did by [Porta, et al. (2000); Rozeff (1982);
Ahmed and Javaid (2009); Al-Najjar (2009); Musa and Fodio (2009);
Hussainey and Eisa (2009)]. Financial structure of financial firms is
considerably different from non-financial firms. Regulatory restrictions
on financial firms influence their financing decisions and these
restrictions affect financial firms' more than non-financial firms.
Like in case of banks, these are bound to maintain a minimum capital
adequacy ratio at all times, under prudential regulations, and it
influence their financing decision. Ogler and Taggart (1983), cited in
Ameer (2008), p. 1], empirically observed this later mentioned finding.
Exclusion of Firms owned by State (wholly or partially, as best we
can identify) as their financing decisions may have been affected due to
government influence. This practice also adopted by other researchers
like Porta, et al. (2000); Afza and Mirza (2010).
In order to be more pragmatic, factors, which may create biasness
in research findings and hamper explanatory power of our explanatory
variables, have been considered while sample selection. Very small firms
having net sales less than PKR 100 million, firms having negative net
worth in more than one year, with unavailable data for one or more
consecutive years, in losses for more than one consecutive year and
those without dividend information are excluded. Pandey and Bhat (2007)
also applied few of these criteria while sample selection. Consideration
of losses and negative net worth is due to the fact that dividends are
basically a primary function of an organisation's profitability and
net worth. Firms, with better dividend payment record, have been
preferred in sample. Musa and Fodio (2009) also quoted Kumar and Lee
(2001) in favour of above point that reason for dropping zero dividend
payout firms is that relative performance evaluation of dividend model
is meaningless for such firms. Exclusion of negative worth firms also
supports this logic as firms facing losses will definitely not be able
to pay dividend and to check these firms in model will be meaningless.
Afza and Mirza (2010) also have qualification that firms should not have
missed dividend payment in more than one year and firms should not be in
losses. Care has been taken to take into account those firms which are
also part of KSE 30 or 100 index so that sample should represent maximum
of the market capitalisation.
6. RESULTS
Table 2, below, provides summary of descriptive statistics of
earnings and dividends. There is an increasing trend in profits and
dividends over the period of time as evident from their mean values.
There is more variability in earnings as compared to dividends. Mean
payout ratio prevails around 50 percent with less variability (standard
deviation about 13 percent).
Table 3 depicts estimation results. Model 1 is fixed effects model
(cross section fixed). Significant results for earnings and lagged
dividend (1) have been observed at p-value of less than 1 percent. A
coefficient of One lagged year dividend has greater influence upon
dividend payment rather current earnings where as interactive variable
of monetary restriction and lagged dividends have mixed and
insignificant results. Coefficient of determination has significant
value of 0.71. In model 2, random effects approach has been used. Here
lagged dividend has significant results with coefficient of 0.72 at
p-value of less than 1 percent. Current earnings have, comparatively,
less coefficient value of 0.06 but also insignificant at 10 percent.
Monetary restriction interactive variable 1, like in fixed effect model
has also negative coefficient but insignificant as p-value is higher,
even than 10 percent. Surprisingly, second monetary restriction
interactive variable has positive coefficient along with significant
results at 5 percent.
Model 3 comprises of GMM estimation, which is urged, in dynamic
panels. Our model is also a dynamic one. Lagged dependent variable may
create biasness and GMM can manage it well. GMM estimation requires
instruments and we have used explanatory variables as instruments.
Results reveal almost similar trends like in models 1 and 2. Both
earnings and one year lagged dividend have significant coefficient
values at p-value of 1 percent. Lagged dividend coefficient has higher
value than of earnings. Monetary restriction interactive variable, again
in this model, has negative coefficient supporting the hypothesis but is
insignificant even at p-value of higher than 10 percent. Second lagged
dividend is appeared with negative coefficient although insignificant.
MR1 has negative coefficient but insignificant and surprisingly MR2 has
positive coefficient and significant at p-value of 1 percent. Results of
all four models portray a very similar and significant finding that
first lagged dividend has a significant and highest positive impact upon
dividend decision of firms in Pakistan. Current year earnings do have a
positive and significant impact but follow the last year dividend in
dividend decision. Due to mixed results and insignificant value for
second lagged dividend variable, we can claim that monetary restriction
does not have any significant bearing on dividend decisions of Pakistani
firms although theory is opposite to the results. Coefficients of lagged
dividend in all models range from 0.3 to 0.7. In model 4 value is 0.626
with adjustment parameter (1-0.62) = 0.38. Target payout ratio
(0.20/0.38) is 53 percent. Firms seem to observe stable dividend
policies.
7. CONCLUSION
Observing effect of monetary policy on dividend behaviour is of
paramount importance and to best of our knowledge, it is first study of
its kind in Pakistan. Lintner's model has been used to test
dividend stability. For dynamic model estimation, GMM is strongly
recommended method of estimation and same has been used in addition to
fixed effects and random effect models. Pakistani firms have been
observed to follow relatively stable dividend policies. Firms have
moderate target payout ratios and adjustment factors. One year lagged
dividends have strongest influence upon dividend decisions followed by
current earnings. Insignificant results of monetary restriction variable
do not claim any effect on dividend decisions of Pakistani firms.
.Although second monetary interactive variable has positive coefficient
in GMM estimation but results of first MR interactive variable and
second lagged dividend variable lead to the above conclusion.
REFERENCES
Abdelsalam, O., A. Al-Masry, and S. Elsegini (2008) Board
Composition, Ownership Structure and Dividend Policies in an Emerging
Market: Further Evidence from CASE 50. Managerial Finance 34:12,
953-964.
Abor, J. and A. G. Bokpin (2010) Investment Opportunities,
Corporate Finance and Dividend Payout Policy: Evidence from Emerging
Markets. Studies in Economics and Finance 27:3, 180-194.
Afza, T. and H. H. Mirza (2010) Ownership Structure and Cash Flows
as Determinants of Corporate Dividend Policy in Pakistan. International
Business Research 3:3.
Ahmed, H. and A. Javid (2009) Dynamics and Determinants of Dividend
Policy in Pakistan: Evidence from Karachi Stock Exchange Non-financial
Firms. International Research Journal of Finance and Economics 25.
Aivazian, V., L. Booth, and S. Cleary (2003) Do Emerging Market
Firms Follow Different Dividend Policies from US Firms. The Journal of
Financial Research 26:3, 371-387.
Al-Najjar, B. (2009) Dividend Behaviour and Smoothing New Evidence
from Jordanian Panel Data. Studies in Economics and Finance 26:3,
182-197.
Al-Yahyaee, H. K., T. Pham, and T. Walter (2010) Dividend Stability
in a Unique Environment. Managerial Finance 36:10, 903-916.
Ameer, R. (2008) Product Market Competition, Regulation and
Dividend Payout Policy of Malaysian Banks. Journal of Financial
Regulation and Compliance 16:4, 318-334.
Atmaja, S. L. (2010) Dividend and Debt Policies of Family
Controlled Firms: The Impact of Board Independence. International
Journal of Managerial Finance 6:2, 128-142.
Baker, M. and J. Wurgler (2004) A Catering Theory of Dividends. The
Journal of Finance 59:3, 1125-1165.
Bhattacharyya, N. (2007) Dividend Policy: A Review. Managerial
Finance 33:1,4-13.
Bhattacharyya, N., A. Mawani, and C. Morrill (2008) Dividend Payout
and Executive Compensation: Theory and Canadian Evidence. Managerial
Finance 34:8, 585-601.
Desai, A. M., F. C. Foley, and R. J. Hines Jr. (2001) Dividend
Policy Inside the Firm. Journal of Economic Literature.
Eriotis, N. (2005) The Effect of the Distributed Earnings and Size
of the Firm to Its Dividend Policy: Some Greek Data. International
Business and Economics Journal 4:1.
Fama, E.and H. Babiak (1968) Dividend Policy of Individual Firms:
An Empirical Analysis. Journal of the American Statistical Association
63, 1 132-1 161.
Garret, I. and R. Priestley (2000) Dividend Behaviour and Dividend
Signalling. The Journal of Financial and Quantitative Analysis 35:2,
173-189.
Goddard, J., G. D. McMillan, and O. J. Wilson (2006) Dividend
Smoothing vs. Dividend Signalling: Evidence from UK Firms. Managerial
Finance 32:6, 493-504.
Gujarati, D. (2003) Basic Econometrics. (4th ed). New York: McGraw
Hill.
Holder, E.M., W. F. Langrehr, and L. J. Hexter (1998) Dividend
Policy Determinants: An Investigation of the Influences of Stakeholder
Theory. Financial Management 27:3, 73-82.
Hussainey, K. and A. J. Eisa (2009) Disclosure and Dividend
Signalling when Sustained Earnings Growth Declines. Managerial Auditing
Journal 24:5, 445-454.
Lintner, J. (1956) Distribution of Incomes of Corporations among
Dividends, Retained Earnings and Taxes.
Michel, A. (1979) Industry Influence on Dividend Policy. Financial
Management 8:3, 22-26.
Musa and I. Fodio (2009) The Dividend Policy of Firms Quoted on the
Nigerain Stock Exchange: An Empirical Analysis. African Journal of
Business Management 3:10, 555-566.
Nissim, D. and A. Ziv (2001) Dividend Changes and Future
Profitability. The Journal of Finance 56:6, 2111-2133.
Pandey, M. I. and R. Bhat (2007) Dividend Behaviour of Indian
Companies under Monetary Policy Restrictions. Managerial Finance 33:1,
14-25.
Pandey, M. I. (2001) Corporate Dividend Policy and Behaviour: The
Malaysian Experience.
Porta, I. R., et al. (2000) Agency Problems and Dividend Policies
Around the World. The Journal of Finance 55:1, 1-33.
Rozeff, S. M. (1982) Growth, Beta and Agency Costs as Determinants
of Dividend Payout Ratios. The Journal of Financial Research 3.
State Bank of Pakistan (2001-2009) State Bank Annual Reports.
Karachi: State Bank of Pakistan.
Van Horne, J. (1998) Financial Management and Policy. (10th ed).
New Jersey: Prentice Hall.
Muhammad Shahzad Ashraf <
[email protected]> is MS
Student at the Shaheed Zulfikar Ali Bhutto Institute of Science and
Technology, Islamabad. Hasan M. Mohsin <
[email protected]>
is Head, Department of Econometrics and Statistics, Pakistan Institute
of Development Economics, Islamabad.
Table 1
Sample Selection Criteria
Sample Size 100 firms. Equal sample size from each industry
(proportionate to their respective population size)
Study Period 2001-2009
Criteria Non-financial firms
Excluding very small firms (having net sales less
than 100 million)
Excluding firms having negative net worth in more
than one year
Excluding firms with unavailable data for one or
more consecutive years
Excluding firms in losses for more than one consecutive
Selection of firms with preferably better dividend
payment record
Table 2
Earnings (PKR Millions)
Year Mean Stdev Max Min
2001 299 1,162 10,859 -1.357
2002 314 900 7,287 -2,649
2003 397 786 6,102 -125
2004 499 894 5,588 -22
2005 647 1,161 7,855 -2
2006 750 1,359 7,558 -321
2007 794 1,686 10,597 -570
2008 620 3,027 19,655 -14,745
2009 644 1,701 9,415 -5,587
All Years 552 690,98 19,655 -14,745
Dividends (PKR Million) Payout (%)
Year Mean Stdev Max Min
2001 165 527 4,513 0 55%
2002 226 930 8,794 0 72%
2003 239 719 6,249 0 60%
2004 225 602 4,282 0 45%
2005 228 626 4,751 0 35%
2006 351 1,006 6,927 0 47%
2007 287 688 4,441 0 36%
2008 334 907 6,785 0 54%
2009 450 1,178 8,923 0 70%
All Years 278 217,027 8,923 0 50%
Table 3
Model 1 Model 2 Model 3
Variable FEM REM GMM
Eit 0.021 * 0.066 *** 0.203 *
t-value 6.593 1.717 6.590
p-value 0.000 0.086 0.000
Dit-1 0.353 * 0.725 * 0.626 *
t-value 4.846 7.043 9.092
p-value 0.000 0.000 0.000
Dit-2 0.059 0.014 -0.040
t-value 1.030 -0.135 -0.543
p-value 0.303 0.893 0.586
MRtxDit-1 -0.052 -0.179 -0.212
t-value -0.675 -0.700 -1.177
p-value 0.499 0.484 0.239
MRtxDit-2 0.056 0.395 ** 0.350 *
t-value 0.765 1.925 2.366
p-value 0.444 0.055 0.018
[R.sup.2] 0.714 0.721 0.672
Adjusted [R.sup.2] 0.664 0.719 0.670
Durbin-Watson 2.288 2.212
Prob (F-Statistics) 0.000
J-Statistic 5.28E
* Significant at 1 percent or less. ** Significant
at 5 percent. *** Significant at 10 percent.
