FIRST DRAFT

22 November 2003

Effect of Dividend Announcement on Shareholders’ Value:

Evidence from Dhaka Stock Exchange

Md. Hamid Uddin*

Abstract

Academic literature suggests that dividend payments should have no impact on shareholders
value in the absence of taxes and market imperfections. Hence, companies should invest excess
funds in the positive net present value projects instead of paying out them to the shareholders.
Literature also suggests that market valuation of stocks depends on the expected future dividends.
If company pays out all of the earnings, funds for future investment will decrease and dividend
may not increase in the future. Moreover, when dividend is taxable, paying out more cash would
increase the shareholders tax liability. Despite these theoretical arguments for not paying
dividends, companies often pay cash dividends to their shareholders possibly to signal
information about the future earnings prospects. Our empirical results based on 137 samples of
dividend paying companies listed on Dhaka Stock Exchange, showed that investors do not gain
value from dividend announcement. Indeed shareholders lost about 20 percent of value over a
period of 30 days prior to the dividend announcement through to 30 days after the announcement.
The lost value may be partially compensated because of the current dividend yield. Overall, the
evidence tends to support the dividend irrelevancy hypothesis. Evidence also indicates that
dividend payment does not signal any information to the investors, which needs to be further
investigated.

1. INTRODUCTION

The goal of corporate entities is to maximize the value of shareholders’ investment in the

firm. Managers pursue this goal through their investment and financing decisions.

Investment decisions involve with selection of positive net present value projects while

financing decisions involve with selection of a capital structure that would minimize the

cost of capital of firm. Apart from the investment and financing decisions, managers need

to decide on regular basis whether to payout the earning to shareholders, reducing the

agency problem (Jensen and Meckling, 1976). However, the question remains whether

paying out of earnings would essentially create value for the shareholders or not. A

*
Md Hamid Uddin, PhD, is an Assistant Professor of Finance at Institute of Business Administration,
University of Dhaka. All enquires and comments should be sent to the author at Tel: 880-2-8617815, Fax:
880-2-8615583, e-mail: huddin@hotmail.com Author wishes to thank Nousheen, Saunia, Ashna, Farhana,
Ashik, Tanveer, and Adnan for their excellent support in data processing.
dividend payment provides cash flow to the shareholders but reduces firm’s recourses for

investment; this dilemma is a myth in the finance literature.

A great deal of theoretical and empirical research on dividend effects has been

done over the last several decades. Theoretically, cash dividend means giving reward to

the shareholders that is something they already own in the company; hence this will be

offset by the decline in stock value (Porterfield, 1959 and 1965). In an ideal world

(without tax and any restrictions) therefore dividend payments would have no impact on

the shareholders’ value (Miller and Modigliani, 1961)1. In the real world, however a

change in the dividend policy is often followed by change in the market value of stocks.

The economic argument for investor’ preference to dividend income was offered by

Graham-Dodd (1951). Subsequently, Walter (1956) and Gordon (1959 and 1962)

forwarded the dividend relevancy idea, which has been formalized into a theory,

postulating that current stock price would reflect the present value of all expected

dividend payments in the future.

Other researchers made efforts to further understand the dividend controversy.

Among them, Brennan (1970 and 1973), Litzenberger and Ramaswamy (1979 and 1980)

showed that it is not optimal for the investors to receive dividends if their marginal tax

rate is greater than zero, and investors’ after-tax expected rate of return (discount rate)

depends on the dividend yield and systematic risk.2 This leads to an idea that at least

dividend might have some tax-induced effect on the share prices. Average investors,

subject to their personal tax rates, would prefer to have less cash dividend if it is taxable:

1
It was further showed that the irrelevancy of dividend policy holds even after dropping the assumption of
ideal economy.
2
Black and Scholes (1974) argued however that tax effect is not uniform for all investors, because different
investors are subject to different tax rates depending on the level of their wealth and income.

2
size of optimal dividend inversely related to personal income tax rates (Pye, 1972).

Hence, stocks prices tend to decline after announcement of dividend increase.

Empirical studies however showed mixed evidence, using the data from US,

Japan and Singapore markets. A number of studies found that stock price has a

significant positive relationship with the dividend payment [Gordon (1959), Ogden

(1994), Stevens and Jose (1989), Kato and Loewenstein (1995), Ariff and Finn (1986),

and Lee (1995)], while others found a negative relationship [Loughlin (1989) and Easton

and Sinclair (1989)]. A negative relationship between dividend announcement stock

returns is expected due to tax effect, but researchers tended to relate the positive

relationship between the stock returns and dividend announcement with the information

effect of dividend. The dividend information hypothesis postulates that cash dividend

carries information regarding the future cash flows of firm that is to be reflected in the

market price of stock after announcement of dividend, particularly when dividend

increases [Bhattacharya (1979) Bar-Yosef and Huffman (1986) and Yoon and Starks

(1995)].

The theoretical literature on dividend effects has been well developed.

Researchers largely accepted that dividend per-se has no impact on the shareholders’

value in an ideal economy. However, in a real world, dividend announcement is

important to the shareholders because of its tax effect and information content. In this

study, we have examined the dividend effect on shareholders’ value in Dhaka Stock

Market with a sample of 137 companies who announced dividend over a period from

October 2001 to September 2002. Our results showed that Cumulative Abnormal Return

(CAR) of 137 stocks portfolio increased shortly before the announcement of dividends

3
but this value increase did not sustain in the ex-dividend periods. Indeed, the

shareholders’ of dividend paying companies lost significant amount of value over a

period of 30 days after the dividend announcement. However, the lost value can be

partially compensated by the dividend yield.

As CARs are negative in the periods after dividend announcement, evidence

suggests dividend announcements do not carry information about the future earnings and

cash flows of the companies. Hence, our findings are not consistent with the dividend

information hypothesis. The negative CAR in the ex-dividend periods is apparently

consistent with the dividend tax-effect hypothesis. This could be possible because

dividend income generally needs to be included in the investors’ personal income for tax

purposes. Since the dividend yield partially compensates the value lost in the ex-dividend

periods, investors’ overall value remains largely unchanged after dividend payments.

Hence, evidence seems to be consistent with Miller and Modigliani hypothesis (1961).

The rest of the paper is organized as follows: in Section 2, we discuss the

methodology. Characteristics of samples are described in Section 3. The empirical

findings are presented and analyzed in Section 4. Finally, Section 6 concludes the paper.

2. METHODOLOGY

In order to study the impact of dividend announcement on shareholders’ value we use

two measures: (i) daily market-adjusted abnormal return (MAAR) and (ii) daily

cumulative abnormal return (CAR). MAAR indicates the relative daily percentage price

change in the dividend paying stocks compared to the change in average market price.

We use DSE all-share price index as the proxy of average market price. MAAR is

calculated as follows:

4
 (1)

Where,

MAARit is the market adjusted abnormal return for security i over time t
Rit is the time t return on security i, calculated as (Pit – Pit-1)/Pit-1.
Where, Pit is the market closing price of stock i on day t. Pit-1 is
the market closing price of stock i on day t-1.
Rmt is the time t return on the DSE all-share price index calculated as
(It –It-1)/It-1. Where, Iit is the market index on day t. It-1 is the
market index on day t-1.

The market adjusted abnormal return (MAAR) shows the change in individual stock’s

value due to the dividend announcement. As the percentage change in market index

(average market price) is deducted, the remainder gives us the unsystematic portion of the

value change, which is specific to that particular stock resulting from its dividend

announcement. MAAR is calculated over a period starting to –30 days to +30 days

relative to the dividend announcement day (0-day).

The second measure used is cumulative abnormal returns (CAR), which measures

the investors’ total return over a period starting from well before the announcement of

dividend to well after the dividend announcement day. We use a 61-day window period

staring from -30-day to +30-day relative to the dividend announcement day (0-day).

CAR is computed as follows:

t= j
CARt = ∑ MAARt (2)
t =1

5
Where, CARt is cumulative abnormal return, MAARt as defined above, j denotes the day

-30 through day +30.

Finally, we used parametric test to determine the statistical significance of market

adjusted average abnormal return of dividend paying stocks over the window period (-30

day to +30 day relative to dividend announcement). The t-statistics were calculated cross-

sectionally by using the standard deviation of abnormal returns of the portfolio of 137

dividend-paying stocks. Moreover, t-test suggested in Brown and Warner (1980, p. 251-

252) is also applied to test the statistical significance of the cumulative abnormal returns.

3. SAMPLE DESCRIPTIONS

The sample includes 137 companies listed on the Dhaka Stock Exchange (DSE) who

announced dividends between October 2001 and September 2002. We considered the

sample period as the stable year in recent periods, as this period followed immediately

after the change of political power in the country when political chaos was generally

absent and market run smoothly. A stable market period is essential for collecting

samples. Otherwise, the empirical results may be contaminated by the other factors such

as market volatility. A breakdown of the sample companies according to industrial

sectors is given below in Table 1 and Figure 1.

Table 1 shows that the highest average dividend was paid in the Fuel and Power

sector, followed by that in the investment sector. The highest dividend was announced in

the food sector, and lowest in the Jute and Services sectors. In Jute sector, only one

company announced dividend during the sample period. The average dividend was 19.5

percent with standard deviation of 12.9 percent. Overall, the table shows that our sample

includes stocks from all sectors, except the paper sector. The number of samples are also

6
fairly equally distributed with 10 to 20 stocks from each sector - except Paper, Jute and

Services sectors. This is also noted that out of 137 companies, 34 companies announced

dividend in 2001 and 103 in 2003. Sample also displays that 108 companies belong to A-

category, 17 belong to B-category and 12 belong to Z-category3. Therefore, the empirical

result based this sample is likely to be reliable.

Table 1: Distribution of Sample Companies Listed on DSE

Sector Number of Maximum Minimum Average
Companies Dividend Dividend Dividend
Bank 16 50.0% 16.5% 27.0%
Chemical 14 40.0% 5.0% 18.3%
Engineering 15 50.0% 10.0% 19.5%
Food 15 60.0% 8.0% 19.0%
Fuel and Power 3 50.0% 5.0% 33.0%
Insurance 17 43.0% 10% 20.1%
Investment 10 50.0% 7.5% 23.7%
Jute 1 - - 10.0%
Paper 1 - - -
Service 3 10.0% - 10.0%
Textile 22 40.0% 5.0% 10.3%
Misc. 20 50.0% 5.0% 21.1%
Total 137

3
DSE classify the listed companies into A, B and Z categories. A category companies are good stocks as
their operating performance are assessed to be good and pay regular dividends, B-category companies are
moderate companies whose operating performance are satisfactory and pay some dividends from time to
time, and Z-category companies are those whose operating performance are not good and normally pay no
dividend.

7
 Figure 1: Average dividend in different industrial sectors

35.00%

30.00%

25.00%
Dividend %

20.00%

15.00%

10.00%

5.00%

0.00%
el

te

.
e
g

od
al
nk

ve ce

isc
ile
t

ic
pe
in

en
Fu

Ju
ic

an
Fo
Ba

rv

xt
er

M
Pa
em

stm
ur

Se

Te
ne
Ch

s
gi

In
En

In

Induatrial Se ctors

4. Empirical Findings and Analyses

4.1. Market Adjusted Abnormal Return

Findings reported in Table 2 shows that average market adjusted abnormal return

(MAAR) on the day of dividend announcement was only 0.8 percent, which was not

statistically significant. This could be due to the fact that the information of dividend

payment often leaks out to the market a few days before the announcement made by the

company. Hence, the announcement of dividend normally carries no surprise to the

market. Therefore, evidence shows that MAARs of day –4 and –3 are about 2 percent and

2.9 percent respectively, which are significant at 5 percent level. This suggests that

market reacts earlier than the actual announcement of dividend.

8
Table 2: Market adjusted abnormal return (MAAR) of 137 dividend-paying DSE
stocks over a window period starting from day –30 to day +30 relative to dividend
announcement day (0-day).
Day relative to dividend Average MAAR t statistic
announcement
-30 -0.049 -1.091
-29 -0.012 -0.664
-28 -0.002 -1.037
-27 0.004 1.275
-26 -0.001 -0.239
-25 0.006 1.625
-24 0.005 1.451
-23 0.005 2.071
-22 0.002 0.738
-21 0.01 0.579
-20 0.004 -2.031**
-19 0.01 0.497
-18 -0.002 -0.884
-17 0.001 0.268
-16 0.009 1.526
-15 0.016 3.877***
-14 0.003 1.155
-13 0.02 0.653
-12 -0.004 -1.951*
-11 0.011 1.953*
-10 0.008 1.161
-9 -0.002 -0.958
-8 0.015 0.374
-7 0.004 0.867
-6 -0.005 -1.892*
-5 0.002 0.215
-4 0.020 2.038**
-3 0.029 2.172*
-2 -0.007 -1.064
-1 -0.003 -0.439
0 0.008 0.554
1 -0.004 -0.669
2 -0.007 -0.779
3 0.001 0.386
4 -0.014 -1.635
5 0.004 0.976
6 -0.011 -1.218
7 -0.039 -1.748*
8 -0.029 -1.474
9 -0.008 -0.370
10 0.004 0.458
11 -0.036 -1.590
12 -0.02 -1.086
Cont…next page

9
 Table 2…cont
Day relative to dividend Average MAAR t statistic
announcement
13 0.001 0.194
14 -0.015 -0.956
15 -0.016 -0.892
16 -0.019 -0.915
17 -0.045 -1.569
18 -0.04 -1.342
19 -0.0028 -1.168
20 0.000 0.051
21 0.001 0.122
22 -0.010 -0.958
23 -0.001 -1.707*
24 -0.005 -0.534
25 0.005 0.660
26 0.0012 0.695
27 0.004 0.634
28 -0.0046 -2.052**
29 -0.003 -0.196
30 0.008 1.093
Note: Asterisks in the last column denotes that the corresponding MAAR is statistically significant. The
asterisks ***, **, and * indicate the level of significance (based on the t values) at respectively the 1, 5 and
10 percent level.

Prior to dividend announcement, MAAR also found significant on the days –6,

-11, -12, and –20. However, the percentage returns on these days are less than that on the

day –4. On other days, MAAR is insignificant. Therefore, evidence tends to confirm that

market reacts a few days before the announcement of dividend is made. During the post-

announcement periods (day +1 to +30), all MAARs are insignificant except those on day

+7, +23, and +28. Overall, MAAR results suggest that the effect of dividend

announcement is not strong in Dhaka Stock Exchange. Shareholders gain only about 4

percent value about three/four days before the announcement of dividend but no

significant value gain on the announcement day.

10
4.2. Cumulative Abnormal Return

Results in Table 3 shows that investors do not gain value from dividend

announcement. Evidence depicts that CAR had risen from -4.9 percent on day -30 to a

level of 10.5 percent on the day of dividend announcement. But the gained value was lost

over the next 30 days after dividend announcement, as CAR dropped to –19.52 percent

on the day 30. Although results tends to suggest that investors may have overreacted to

the dividend announcement, the evidence generally consistent with the dividend

irrelevance hypothesis of Miller and Modigliani (1961). This is because investors seemed

to gain no value from the dividend announcements.

Findings also show that investors lost more value in the ex-dividend period than

the value gained in the pre-dividend period. This finding tends to suggest that dividend

announcement does not carry information about the future earnings and cash flow of the

companies. In Bangladesh, DSE and Security and Exchange Commissions (SEC)

generally rate the performance of the listed companies based on their regular dividend

payments. Hence, companies may like to retain their good standing by paying regular

dividends.4 In the presence of a kind of indirect pressures from the regulatory authorities,

the companies may not be able to effectively signal the future earning prospects through

their dividend announcement.

4
Payment of dividend is purely a corporate financial decision of a firm. Dividend per se should not have
any impact on the shareholders’ value but companies may like to pay dividend due to tax effect or
information effect. Since dividend income was tax free in Bangladesh, tax effect should not be found at
DSE. If companies pay dividend that should be for signaling any information to the investors, but when
regulatory authorities rate the companies based on their regular dividend payments, investors may become
confused about the purpose of dividend announcement by the companies.

11
Table 3: Cumulative abnormal return (CAR) of 137 dividend-paying DSE stocks
over a window period starting from day –30 to day +30 relative to dividend
announcement day (0-day).
Days relative to dividend CAR t-values
announcement
-30 -0.049 -1.091
-29 -0.061 -1.571
-28 -0.063 -1.141
-27 -0.059 -1.271
-26 -0.060 -1.091
-25 -0.054 -1.781
-24 -0.049 -2.052
-23 -0.044 -1.781
-22 -0.042 -1.111
-21 -0.032 -0.890
-20 -0.028 -0.111
-19 -0.018 -1.194
-18 -0.02 -1.123
-17 -0.019 -1.178
-16 -0.010 -1.152
-15 0.006 3.125
-14 0.009 1.962
-13 0.029 1.561
-12 0.025 1.333
-11 0.036 3.152
-10 0.044 2.150
-9 0.042 2.111
-8 0.057 4.591
-7 0.061 4.781
-6 0.056 3.589
-5 0.058 3.456
-4 0.078 7.985
-3 0.107 8.978
-2 0.100 9.781
-1 0.097 5.891
0 0.105 6.888
1 0.101 5.578
2 0.094 4.595
3 0.095 4.258
4 0.081 8.654
5 0.085 4.568
6 0.074 3.245
7 0.035 1.589
8 0.006 0.015
9 -0.002 1.065
Cont…next page

12
 Table 3….cont.
Days relative to dividend CAR t-values
announcement
10 0.002 0.860
11 -0.034 -3.656
12 -0.054 -7.891
13 -0.053 -2.456
14 -0.068 -5.892
15 -0.084 -5.654
16 -0.103 -4.444
17 -0.148 -10.652
18 -0.188 -5.456
19 -0.1908 -4.658
20 -0.1908 -3.222
21 -0.1898 -4.658
22 -0.1998 -7.698
23 -0.2008 -4.587
24 -0.2058 -7.159
25 -0.2008 -9.346
26 -0.1996 -11.986
27 -0.1956 -12.654
28 -0.2002 -10.258
29 -0.2032 -8.348
30 -0.1952 -8.147

Our results reported above support the dividend irrelevance proposition in the Dhaka

Stock Market. It is however important to examine further whether dividend carries any

information, e.g., future earnings. We leave it for future research because the earning data

for the year following the current dividend payment is not yet available.5 Hopefully,

future findings on the dividend information hypothesis would validate the current

evidence from DSE, which supports the dividend irrelevancy proposition in Bangladesh.

5
We have collected the dividend announcement data from October 2001 to September 2002, which have
been analyzed in this paper. Currently, we are keeping the track of information on earning announcements
by the same companies from October 2002 onward. Hopefully, we will get most of the earning
announcements by the end of year 2003, which will be analyzed in next paper on dividend information
hypothesis.

13
 Figure 2: Cumulative abnormal return (CAR) of 137
DSE listed companies over period from day -30 to day
+30 relative to dividend anncouncement
Cumulative abnormal
return

1
5
9
13
17
21
25
29
-27
-23
-19
-15
-11
-7
-3
Event days relative to dividend anncouncement

5. Conclusion

In academic literature, it was suggested that dividend payments have no impact on the

shareholders’ value (Miller and Modigliani, 1961) in the absence of taxes and other

market imperfections. A dividend payment provides cash flow to the shareholders but it

reduces firm’s recourses for investment. Hence, firms should not pay dividend if they

have any positive net present value project in hand. However, Walter (1956) and Gordon

(1959 and 1962) showed that valuation of stock depends on the expected future

dividends. If company pays out all the earnings to shareholders, funds for future

investment will decrease and dividend may not increase in the future. Therefore,

theoretical literature suggested that dividends payout should not be desirable provided

that companies can better invest their funds. Moreover, cash dividend is not desirable if

investors need to pay taxes on their dividend income. Given the valid reasons for not

14
paying dividends, an announcement of dividend payments may carry some information

for the market and stock prices may be adjusted accordingly.

Based on the 137 DSE listed companies declaring dividends during October 2001

and September 2002, we found that investors do not benefit from dividend

announcement. Over the period starting from 30 days prior to dividend announcement to

30 days after the announcement of dividend payment, investors incurred losses upto

19.52 percent of stock value. Although this loss of value is partially compensated by the

current dividend yield, investors in Bangladesh seemed to have no net gain due to

dividend payments. The evidence from DSE tends to support Miller and Modigliani

(1961) hypothesis of dividend irrelevancy. Apart from the academic significance of our

findings, the regulatory authorities (DSE and SEC) may wish to review their policy of

company evaluation, which emphasizes on dividend payments by the listed companies, in

the context of empirical evidence on the dividend effects.

15
REFERENCE

Ariff, M. and Finn, F. J., “Announcement Effects and Market Efficiency in a Thin Market: An
Empirical Application to the Singapore Equity Market”, Asia Pacific Journal of
Management, Vol. 6, 1986, p: 243-267.
Bar-Yosef, S. and Huffman, L., “The Information Content of Dividends: A Signalling Approach”,
Journal of Financial and Quantitative Analysis, Vol. 21, No. 1, 1986.
Bhattacharya, S., “Imperfect Information, Dividend Policy, and ‘the bird in the hand” Fallacy”,
The Bell Journal of Economics, Vol. 10, No. 1, 1979, p: 259-270.
Black, F. and Scholes, M., “The Effect of Dividend Yield and Dividend Policy on Common Stock
Prices and Returns”, Journal of Financial Economics, Vol. 1, 1974, p: 1-22.
Brennan, M. J., “Taxes Market Valuation and Corporate Financial Policy”, National Tax
Journal, Vol. 26, 1970, p: 1115-1121.
Brown, S.J. and Warner, J.B., Measuring Securities Price Performance, Journal of
Financial Economics, Vol. 8, 1980, p. 205-258.
Easton, S. A. and Sinclair, N. A., “The Impact of Unexpected Earnings and Dividends on
Abnormal Returns to Equity”, Accounting & Finance, Vol. 29, 1989, p: 1-19.
Gordon, M. J., “Dividend, Earning, and Stock Prices”, The Review of Economics and Statistics,
Vol. 41, 1959, p: 99-105.
----------------, “The Savings Investment and Valuation of a Corporation”, The Review of
Economics and Statistics, Vol. 44, 1962, p: 37-51.
Graham, R. E. and Dodd, P., “Security Analysis”, McGraw-Hill Book Co. 1951.
Jensen, M. and Meckling, W., “Theory of Firm: Managerial Behaviour, Agency Cost,
and Ownership Structure”, Journal of Financial Economics, Vol: 3, 1976, p: 306-
360.
Lee, B. S., “The Response of Stock Prices to Permanent and Temporary Shocks to
Dividends”, Journal of Financial and Quantitative Analysis, Vol. 30, 1995, p: 1-
22.
Litzenberger, R. H. and Ramaswamy, K., “The Effects of Personal Taxes and Dividends on
Capital Asset Prices”, Journal of Financial Economics, Vol. 7, 1979, p: 163-195.
----------------, -----------------------, “Dividend, Short Selling Restrictions, and Tax-
Induced Clientele and Market Equilibrium”, The Journal of Finance Vol. 35,
1980.
Loughlin P. H., “The Effect of Dividend Policy on Changes in Stockholders’ Wealth”, A PhD
Thesis, Graduate School of Saint Louis University, 1982, USA.
Kato, K. and Loewenstein, U. “The Ex-Dividend-Day Behavior of Stock Prices: The Case of
Japan”, The Review of Financial Studies, Vol: 8, 1995, p: 816-847.
Miller, M. H., and Modigliani, F., “Dividend Policy, Growth and The Valuation of
Shares”, The Journal of Business, Vol. 34, 1961, p: 411-433.

16
Porterfield, J. T. S., “Dividend, Dilution, and Delusion”, Harvard Business Review, Vol.
37, 1959, p: 56-61.
-----------------------, “Investment Decisions and Capital Costs”, Englewood Cliffs, N.J.,
Prentice-Hall, 1965.
Pye, G. “Preferential Tax Treatment of Capital Gains, Optimal Dividend Policy, and Capital
Budgeting”, The Quarterly Journal of Economics, Vol. 86, 1972, p: 226-242.
Ogden, J. P., “A Dividend Payment Effect in Stock Returns”, Financial Review, Vol. 29, 1994, p:
345-369.
Stevens, J. L. and Jose, M. L., “The Effect of Dividend Payout, Stability, and Smoothing on Firm
Value”, Journal of Accounting Auditing & Finance, Vol. 7, 1992, p: 195-216.
Walter, J. E. “Dividend Polices and Common Stock Prices”, The Journal of Finance, 16, 1956, p:
29-41.
Yoon, P. S. and Starks, L. T., “Signaling, Investment Opportunities, and Dividend
Announcements”, The Review of Financial Studies, Vol. 8, 1995, p: 995-1018

17