The Impact of Dividend Announcements on Shareholders Wealth

The Impact of Dividend Announcements on Shareholders Wealth

Abstract

This paper examines whether dividend announcements affect shareholders wealth in the short-term. The empirical results show that changes in the level of dividend payment generate abnormal returns, which is in accordance with the information theory of Lintner (1956). However, since the results are relative weak this paper supports the theory of Amihud & Li (2006) who argue that the information content of dividends is declining, for the reason that these days more information is readily available and shareholders are better informed, the asymmetry between management and shareholders is fading away. Furthermore, this paper argues that the amount of increases (decreases) in dividend payments is affected by the economic climate. Managers tend to increase dividends in bull markets for the reason that they become too overconfident in the economy and their own future company performance.

Key words: Dividend Changes, Dividend-signaling theory, short-term performance

JEL Classification: G35

Roelof Romp S1597892

July, 2008

I. Introduction

This paper studies the effects of cash dividend announcements on shareholders wealth in the short-term. In particular, this study examines whether a relation exists between changes in dividend announcements and short-term stock performance. In addition, this study investigates whether the total amount of increases (decreases) in dividend payments is affected by economic conditions. The study uses data of companies listed on the FTSE 1001 index for the period of January 1st 2000 till December 31st 2008.

In principle, a cash dividend is the money which will be paid to the shareholders. The decision to distribute cash dividends determines what funds flow to investors and what funds are retained by the company for reinvestment. Ross, Westerfield & Jaffe (2005) state that shareholders prefer a high-dividend policy for a couple of reasons: the desire for fixed income; the behavioral finance principle of self control to keep a stable investment portfolio; the information content that high dividends signal future earnings; and it reduces the agency costs between shareholders and management.

In the literature, dividend payout policy has been examined massively. Still, further examination in this field is valuable since the empirical results and theories disagree in time. In one of the first studies on dividend policies, Lintner (1956) describes the information content of dividends. He argues that dividend payments are not only based on past and current earnings, but as well on future earnings, which provide valuable information on future company performance. In this way, companies can use dividend announcements to signal shareholders whether they expect good or bad future performance. This implies that an increase (decrease) in dividend payment does affect shareholder value positively (negatively).

On the other hand, according to Miller & Modigliani (1961) it is irrelevant for shareholder value whether companies distribute cash dividends, stock dividends or repurchase shares. They assume that with well-functioning markets, neutral taxes and rational investors, the market value only depends on the income stream generated by its assets.

Since the release of Lintner’s theory that dividends contain information, the bulk of the empirical literature found empirical effects of changes in dividend announcements on shareholder value.

However, since a couple of years, the literature starts to debate whether this theoretical framework is still valid. Amihud & Li (2006) are very clear about this; they state that the effect of information content is declining. The cause for this decline are the institutional holdings, which reflect more sophisticated and better informed shareholders. These days so much company data are readily available that dividend announcements, which should contain valuable information about future earnings of the company, actually do no longer contain critical information for the market. This implies that the effect of changes in dividend announcements on shareholder value is fading away. Furthermore, the use of dividends as a signaling device is costly, since dividends are generally taxed as ordinary income. This remarkable change of thoughts in the field of dividend payout policy makes further research into this topic valuable for extending the current literature.

stock market and the amount of increases (decreases) in dividend payments to test whether this assumption hold.

This paper tries to contribute to the existing literature in exploring the recent effects of dividend policies on shareholder value. Since the literature disagrees in time, a study using recent data will bring up to date the current thought in dividend policies. Furthermore, this paper relates the percentage change in dividend payments with abnormal performance. Using these figures, it is possible to determine whether a relation exists between higher percentage change in dividend payments and higher abnormal performance. Additionally, a side issue in this paper examines whether a relation exists between stock market performance and the amount of increases (decreases) in dividend payments. The robustness of the empirical results will be checked with a two-factor regression model and various event periods, which is included in the appendix.

II. Literature review

In this section, it will be discussed which alternatives companies have for distributing cash dividends. Furthermore, an overview will be given of authors who contributed to the theory on dividend policies. After that, a couple of studies which uses UK market data will be discussed since this study also use data from the UK. Finally, the behavioral aspects of dividend policies will be discussed.

2.2 Different Payout policies

This paper examines cash dividends, which is in principle the money which will be paid to the shareholders. However, companies do have alternatives to compensate shareholders. These are stock splits, stock dividends and share repurchases. Shortly, these alternatives will be discussed.

Stock splits and stock dividends increase the number of outstanding shares, which reduce the stock price since each share now represents a smaller percentage of the total company value. In principle, the distribution of stock splits and stock dividends are just finer slicing of a given cake without effecting the total company value. The fundamental question with stock splits and stock dividends is: If these cosmetic changes are just financial manipulations, particularly when real costs are incurred in the process, why should companies continue to distribute stock splits and stock dividends? Lakonishok & Lev (1987) suggest that stock splits are mainly used to restore stock prices, which increase considerably during an unusual growth period, to a normal range, defined in terms of market and industry-wide prices. On the other hand, stock dividends are usually not used to adjust stock prices to normal levels. Lakonishok & Lev (1987) suggest that managers believe that they can signal to naïve shareholders that stock dividends will be temporarily higher regarded as a substitute for cash dividends.

(3) offset of dilution; the exercise of stock options by managers increases the number of stocks outstanding. In this case, companies buy back shares to offset dilution. (4) repurchase as investment; when the management of a company believes that their shares are undervalued or there are few profitable investment opportunities, then repurchase shares can be the best investment opportunity.

2.3 Overview of dividend announcements

In this section, the main contributions to the theory of dividend policies will be discussed. Lintner (1956) discussed the information content of dividends. He argues that dividend policy depends on past, current as well as on future earnings. When dividends also depend on expected future earnings, shareholders can use this as an information signal for future company performance. The majority of the literature accepted the theory of the information content of dividends discussed by Lintner (1956), which is nowadays a very common and widespread theory in finance.

Somewhat contrary to Lintner is the framework for dividend payments built by Miller & Modigliani (1961). They agree that dividends can convey information about future cash flows when markets are incomplete. However, their framework assumes that when capital markets are perfect, investors behave rational and taxes are equal to zero. According to this model investors are indifferent to dividend policy and are therefore irrelevant for shareholder value. In this case, it is not relevant whether a company pays out cash dividends, stock dividends or repurchases shares.

However, this still leaves the question whether the theoretical framework mentioned by Lintner exists. Pettit (1972) and Watts (1973) are pioneers in the field of testing this theory empirically. However, the results of these studies are not similar. Watts (1973) suggests that on average there is a very small positive relationship between unexpected dividend changes and future earnings, which is consistent with the information theory. However, these empirical results are so small that he concludes that his own results are trivial. Contrary to this, without a doubt Pettit (1972) clearly suggests that his empirical results support the information theory. On the other hand, both authors argue that the proposition of Miller & Modigliani (1961) that dividend policy is irrelevant for shareholder value does not hold.

rates more desirable, since dividends are taxed as ordinary income, and thus a company can increase value by reducing dividend payments. Black & Scholes (1978) studied this theory. They state that if companies are aware of the fact that they can influence the share price by their dividend policy, then they will adjust their dividend policy to supply levels which are most in demand. In this case, equilibrium will be reached and no company is able to affect share prices by changing their dividend policy. Black & Scholes (1978) call this the supply effect.

A study by Ahoromy & Swary (1980) supports the theory that dividend policy influences shareholders wealth. They examined dividend and earnings announcements, which are made public on different dates. Their results strongly support the theory that changes in dividend announcements provide useful information for shareholders. In addition, the results support the semi-strong form of the efficient capital market hypothesis, which state that stock markets adjust in an efficient manner to dividend announcements. Ogden (1994) argues that a positive dividend payment effect occurs to companies which have dividend reinvestment plans. When shareholders reinvest the dividends into shares of the company, the confidence of the stock market in the company increases, which will cause the demand for the stock to increase and raise its price.

A study of Easterbrook (1984) focuses on why companies pay dividends. Since dividends are taxed as ordinary income, it should be rational to reinvestment excess cash flows in the company instead of paying dividends. According to Easterbrook (1984) dividends reduce excess cash flows and may reduce agency costs that arise from conflicts between management and shareholders. These conflicts have to do with monitoring of managers and adjusting the level of risk taking by the managers. Examples of agency conflicts are: perquisite consumption, empire building or other value destroying activities. They conclude that companies pay dividends to reduce agency problems. A study of Miller & Rock (1985) supports this agency conflict theory. They assume that, contrary to the theory of Miller & Modigliani (1961), there is information asymmetry between managers and outside shareholders about the true state of the current earnings.

Lang & Litzenberger (1989) empirically examined the free cash flow theory using Tobin’s Q. Their results support the theory of Jensen (1986).

costly since it will be taxed as ordinary income. Therefore it is not a proper method to distribute excess cash flows.

In 2002 Grullon, Michaely & Swaminathan came up with the maturity theory. This theory relates changes in dividend policy with the company’s life cycle. As companies become more mature, the less investment opportunities arise and the more dividends will be paid. This process causes that the systematic risk of the cost of capital will significantly decline. They state that increases in dividends are important signs of the long maturity process, which will be positive news for the market. The market will react positive on increasing dividends for two reasons. The first is that before the dividend announcement the market was not aware of the declining risk, since this is a positive surprise for the market, stock prices will increase. Second, the market reacts positive to dividend increases since the company can no longer overinvest excess cash flows; this is according to the free cash flow theory of Jensen (1986).

Table 1: literature review

Author (year) Index Period Sample (time frame) positive effect on shareholders wealth

1970: First empirical studies support the theory of Lintner (1956) that dividends contain information

Pettit (1972) NYSE 01/01/64 - 30/06/68 625 (-3 till 3) yes

Watts (1973) NYSE 01/06/45 - 30/06/68 310 (-11 till 12) limited

Black & NYSE 01/01/26 – 30/03/66 1050 no

Scholes (1978)

Ahoromy & NYSE 01/01/63 - 12/31/76 149 (-10 till 10) yes

Swary (1980)

Lang & NYSE 01/01/79 - 31/12/84 429 yes

Litzenberger (1989)

Abeyratna, FTSE 01/01/86 - 31/12/91 617 yes

Lonie, Power & Sinclair (1996)

Ogden (1994) NYSE/AMEX 01/07/62 - 31/12/89 3505 (-5 till 10) yes

Benartzi, NYSE/AMEX 01/01/79 - 31/12/91 7186 very limited

Michaely & Thaler (1997)

Balachandran FTA all share index 01/01/86 - 31/21/93 299 (-12 till 20) yes

(1998)

2000: Authors start to debate whether the theory that information contains information still exists.

Conroy, Eades TSE 01/01/88 - 31/12/93 Unknown no

& Harris (2000)

Fama & NYSE/ AMEX 01/01/26 – 31-12-99 Unknown no

French (2001) NASDAQ

Gunasekarage & FTSE 01/01/89 - 31/12/93 1787 (-1 till 1) yes

Power (2002)

Chen, Firth & SZSE/SHSE 01/01/94 - 31/12/97 1499 (-7 till 7) no

Gao (2002)

Grullon, Michaely NYSE /AMEX 01/01/67 - 31/12/93 7642 (-1 till 1) yes

& Swaminathan (2002)

Amihud & NYSE/AMEX 01/07/62 - 31/12/00 444 (-15 till 15) no

Li (2006)

2.4 UK evidence on dividend announcements

Because this study uses data from the UK market, a couple of dividend studies to the UK market will be discussed. A note should be made before interpretation of the UK results. Till July 1997 the UK had a different tax system for dividends than most countries. Tax-exempt investors, like pension funds and charities, could claim cash refund of tax credits from the tax authorities. This created an incentive for paying out cash dividends instead of retain earnings in the company (Bond, Chennels & Devereux, 1996). In 1997 the British tax system is reformed and since this reformation shareholders are indifferent between capital gains and cash dividends, under the assumption of Miller & Modigliani (1961). Consequently, studies which uses data before 1997 can be influenced by this different tax system.

Abeyratna, Lonie, Power & Sinclair (1996) discuss the relationship between earnings and dividend announcements. Their results depend on the character of change in earnings and dividends. Positive news of both factors generate positive abnormal returns, while with bad news both factors generate negative abnormal returns. However, when companies announce positive as well negative about earnings and dividends at the same time, the bad news seem to offset the good news and no abnormal returns are noticed.

Balachandran (1998) examined dividend reductions between January 1986 and December 1993. His results supports the theory that dividend reductions decrease short-term performance.

In a study Gunasekarage & Power (2002) examine the post-announcement performance of UK companies in the short run and in the long run. Evidence in the short run, three trading days from t -1 to t +1, provides valuable information that supports the dividend-signaling theory. However, in the subsequent five-year period, companies which announce dividend decreases, outperform companies which announce dividend increases. The conclusion of the authors is that companies use dividend decreases to adept the corporate finance policy to turn company performance around. This is contrary to the view that dividends reductions convey negative information about the future performance of the company.

2.5 Behavioral finance aspects on dividend policies

biases. Kahneman & Tversky (1979) developed the prospect theory. This descriptive theory of choice under uncertainty argues that investors frame their decisions in terms of losses and gains, which cause that investors view each investment as separate investment instead of portfolio, which limit investor’s ability to minimize risk and maximize return. In this case, investors tend to prefer high dividend paying stocks since these are sure gains. Thaler & Shefrin (1981) developed the theory of self-control. Some investors tend to spend too much of their money. In this case, investors prefer cash dividends since they view this as consumable income, so they cannot spend from their investment portfolio.

III. Methodology

This section is subdivided into four parts. These are: positive dividend announcement changes, negative dividend announcement changes, constant dividend announcements and the relationship between the amount of increases (decreases) with stock market performance.

The assumption in this paper is that dividend announcements provide information on future cash flows and future performance of the company. Therefore, the expectation is that an increase in the level of dividend payment will cause short-term performance to increase. Consequently, the hypothesis will be:

0

H : A positive dividend change does lead to positive short-term abnormal performance.

1

H : A positive dividend change does not lead to positive short-term abnormal performance.

For a decrease in the level of dividend announcements the opposite will be expected:

2

H : A negative dividend change does lead to negative short-term abnormal performance.

3

H : A negative dividend change does not lead to negative short-term abnormal performance.

For constant dividend announcements no changes in short-term performance will be expected:

4

H : Constant dividend announcements does not lead to short-term abnormal performance.

5

H : Constant dividend announcements does lead to short-term abnormal performance.

The second issue discussed in this paper is whether there is correlation between the economic climate and the amount of increases (decreases) in dividend payments. The expectation is that when the economy moves upwards the amount of increases in dividend payments will increase. When the economy moves downwards the opposite will be expected.

6

H : A positive relation exists between an increase (decrease) in the stock market and the amount of increases (decreases) in dividend payments.

7

3.2 Empirical design

To examine empirically the adjustment between stock prices and dividend announcements, the market & risk adjusted return method of Brown & Warner (1985) will be used. This model compares the excess returns of share i on day t with the returns of the market index on day t with the OLS values of alpha and beta from the estimation period. In addition, this one-factor model is extended to a two-factor model where the 10-year government bond rate is included. This extension gives the regression model a better r-squared fit.

First the daily returns of the stock prices, FTSE index, and the 10-year government bond rate are calculated: ) / ( , , 1 ,t = it it− i Ln P P R (1) t i

R, is the actual return on share i on day t and Pi,t is the price of share i on day i. The Ln is the

logarithm function and is used to prevent that outliers influence the data negatively. This increases the reliability of outcomes of the dataset.

Assuming that the stock returns are distributed normal, a one-factor OLS market model will be used: t i t m i i R e R =

α

β

So, that the abnormal return is:

t m i t i t i R R AR, = , −

α

β

So, that the abnormal return is: t g g t m m i t i t i R R R AR_, = _, −

α

β

β

AR_, is the abnormal return.

α

β

is the error term. R_m,tis the return on the FTSE index on day t, while R_g,tis the return on the 10-year government bond on day t. In both regressions beta is calculated with the formula (Ross, Westerfield & Jaffe 2005):

) ( ) , ( ), , ( ), , ( , t m g k t m g k t i R Var R R Cov = = =

β

Finally, to assess whether the abnormal returns statistically differ from zero the t-test will be used.

) , ( ~ )) , ( var( ) , ( 2 , 2 1 2 1 2 1 i i t i r t t AR t t AR T = Ν

µ

σ

µ

σ

r Ν implies that the t-test is lognormal distributed (Campbell, Lo, & Mackinlay, 1997). When all abnormal returns in the event period are calculated, the cumulative abnormal return (CAR) can be calculated. This CAR is used to spot a possible upward (downward) trend in the abnormal returns:

t t t t AR t t CAR 2 2 2 1, ) ( =Σ = (8)

The variance of the CAR will be calculated with:

The t-test for the CAR is: ) , ( ~ )) , ( var( ) , ( 2 , 2 1 2 1 2 1 i i t i r t t CAR t t CAR T = Ν

µ

σ

Furthermore, to test the relationship between the amount of increases (decreases) in dividend payments and market performance, a correlation test will be applied:

) )( ( ) ( ) ( y y x x y y x x r − − − − =

∑

∑

Where r is the correlation, x the yearly FTSE 100 change and y the yearly change in amount of positive, negative or constant dividends.

To test whether the correlation is significant the next formula will be used:

) 2 /( ) 1 ( ₋ 2 ₋ = N r r T (12)

IV. Data

The sample uses data of companies listed on the FTSE 100 index on January 1st 2008. Daily closing stock prices, daily closing FTSE 100 index data and daily 10-year government bond rates are obtained from DataStream. In addition, dividend announcements dates are obtained from Extel Financial Ltd. Dividend payments in the UK are usually paid twice a year. One mid-year interim payment and one final-year payment. Interim payment tends to be a lower amount than the final-year payment (Balachandran 2003). Therefore, this study compares interim payments with last years interim payments and final-year payments with last years final-year payments.

The time frame for this study is the period January 1st 2000 till December 31st 2007. A condition for an announcement to be included in the sample is that it should not exceed a 100% change in dividend payment. This limitation is to prevent that outliers influence the results too much. For the final sample, data of 92 companies is used and this results in a sample of 1284 announcements. The event window is t -3 till t +3 where the announcement day is t +0. The estimation window is the period t -200 till t -4. This is based on an empirical study on dividend announcements by Pettit (1972). These observations are then subdivided into three categories, depending on the direction of change in dividend payment; (1) increase in the level of dividend payment, (2) decrease in the level of dividend payment, (3) no change in the level of dividend payment. Some descriptive statistics of these observations are provided in table 2 and 3.

Table 2: sample information

in sample:

companies 92

total dividend announcements 1284

dividends per year:

2000 125 2001 143 2002 142 2003 148 2004 162 2005 157 2006 198 2007 209 not in sample:

no company data available 4

no dividend announcements available 3

no dividends paid 1

Table 3: descriptive statistics

Sample Average Median Min Max Standard Skewness Kurtosis

deviation

increase in dividend announcements 1002 -0.0000 0.0000 -0.0048 0.0044 0.0013 -0.1818 0.2204

decrease in dividend announcements 152 0.0000 -0.0001 -0.0043 0.0048 0.0019 0.2331 -0.3620

no change in dividend announcements 130 0.0000 0.0002 -0.0057 0.0067 0.0020 -0.0990 0.1668

beta1 92 0.0618 0.0488 -0.19202 0.06533 0.1238 2.4924 14.850

alpa4 92 0.0003 0.0003 -0.00235 0.00246 0.0007 -0.2253 2.7175

1

average Beta of the companies divided by all companies in sample

2

lowest average company beta is Rolls-Royce Group

3

highest average company beta is Man Group plc

4

average Alpha of the companies divided by all companies in sample

5

lowest average company alpha is British Energy

6

highest average company alpha is Admiral Group

V. Empirical results and interpretation

The abnormal return of the companies in the sample are reported in a seven day event window (three days before and three days after the announcement). The sample data is subdivided into three categories. Table 4 provides information of increases in dividend payments, table 6 of decreases in dividend payments and table 7 of constant dividend payments. The T-statistic presented in the tables indicate whether the abnormal returns and the cumulative abnormal returns in the event window significantly differ from zero. In addition, the abnormal returns and cumulative abnormal returns, which are plotted in figure 1 and 2 present a graphical overview of the results. The results in table 4 indicate that companies which increase dividend payments, on average realize positive abnormal returns in the event period. This implies that the H0, that

positive dividend announcements affect shareholder value positively, can be accepted. Additionally, the abnormal return for the announcement day is 1.93135 and significant at a 10% level. This result supports the theory described by Lintner (1956) that when dividend payments increase, shareholder value also increases, since these dividend payments should convey positive information about expected future performance. However, since the abnormal return is only significant on a 10% level, the result is not very strong and thus similar to the empirical findings of Watts (1973), who also found weak significant evidence.

Table 4: the effect of positive dividend changes on abnormal return

days AR1 T (AR) CAR2 T (CAR) -3 0.05% 0.38104 0.05% 0.38104 -2 -0.02% -0.14427 0.03% 0.11839 -1 0.14% 1.08677 0.18% 0.44118 0 0.26% 1.93135* 0.43% 0.81372 1 0.03% 0.22073 0.46% 0.69512 2 0.07% 0.56329 0.54% 0.67315 3 0.03% 0.19145 0.56% 0.60434

This sample contains 1002 announcements * Denote significance at the 10% level. ** Denote significance at the 5% level. *** Denote significance at the 1% level.

1

Abnormal returns

2

Figure 1: abnormal returns of positive dividend changes Daily abnormal returns (increase)

-0,70% -0,50% -0,30% -0,10% 0,10% 0,30% 0,50% 0,70% -3 -2 -1 0 1 2 3

Days relative to dividend announcement

A R (% ) CAR (%) AR %

The increases in positive dividend payments are further subdivided in three categories; (1) increase with 0% till 10%, (2) 10% till 25% and (3) 25% till 100%. With this information it is possible to see whether a relationship exist between higher dividend payments and share performance.

Table 5: percentage change in increase in dividends

0% - 10%1 10% - 25%2 25% - 100%3 days AR T (AR) AR T (AR) AR T (AR)

-3 0.03% 0.25630 0.06% 0.42501 0.06% 0.48392 -2 0.03% 0.20482 -0.07% -0.50362 -0.02% -0.12009 -1 0.18% 1.34882 0.20% 1.52358 0.05% 0.39011 0 0.17% 1.28562 0.27% 2.01771** 0.34% 2.53280** 1 0.09% 0.65290 0.02% 0.11845 -0.01% -0.08466 2 0.10% 0.78802 0.08% 0.57555 0.04% 0.31823 3 0.02% 0.16911 0.02% 0.13294 0.04% 0.30661

* Denote significance at the 10% level. ** Denote significance at the 5% level. *** Denote significance at the 1% level.

1

All abnormal returns with an increase in dividend payment between 0% and 10%

2

All abnormal returns with an increase in dividend payment between 10% and 25%

3

All abnormal returns with an increase in dividend payment of 25% and 100%

prices. An explanation for the absence of significance in category 1 could be that effects are smaller when the percentage increase is lower.

5.2 Short-term performance of companies which decrease dividends

Table 6 provides information of decreases in the level of dividend payments. On average, the cumulative abnormal return decreases in the event period. This is according to the free cash flow theory of Jensen (1986), which indicates that decreases in dividend payments generate excess cash flows which leads to overinvestment an thus decreases in future performance. According to this theory, it should be expected that day 0 presents significant abnormal performance. Since the result on day 0 is only -1.38878 the result is negative as expected, nevertheless not significant. Only on day 3 the abnormal return is -1.6750 and significant at a 10% level, which is a remarkable result. This result indicates that stock markets react slowly to negative dividend changes or that negative dividend changes will be followed up by more negative company information, which leads to even larger decreases in the share price. With these results the H2,

that negative dividend payments affect shareholder value negatively, can be accepted. However, since the results are very weak it supports the results of Chen, Firth & Gao (2002), who argue that dividend announcements affect shareholder value only marginally.

Since the sample size is relative small, for example there are only 3 samples with a negative change in dividends above 50%, no further sub-dividing of the results will take place, as was done by positive dividend changes.

Table 6: The effect of negative dividend change on abnormal return

days AR T (AR) CAR T (CAR) -3 0.07% 0.37305 0.07% 0.37305 -2 0.02% 0.10001 0.09% 0.23653 -1 0.00% 0.01006 0.09% 0.15434 0 -0.27% -1.38878 -0.18% -0.23144 1 0.13% 0.65028 -0.05% -0.05510 2 0.01% 0.06309 -0.04% -0.03540 3 -0.32% -1.67850* -0.36% -0.27013

Figure 2: abnormal returns of negative dividend changes

Daily abnormal returns (negative)

-0,40% -0,30% -0,20% -0,10% 0,00% 0,10% 0,20% 0,30% 0,40% -3 -2 -1 0 1 2 3

Days relative to dividend announcement

A R ( % ) CAR (%) AR %

5.3 Short-term performance of companies which announce constant dividends It is expected that constant dividends would not lead to abnormal performance, since these announcements do not provide any information. The results in table 7 confirm the expectation that constant dividends do not provide any company information since there are not significant abnormal returns. Furthermore, the cumulative abnormal return in the event period decreases only slightly, which indicates that stock markets do not react on constant dividend announcements. This implies that the H₄, that constant dividend do not affect shareholder value, can be accepted.

Table 7: the effect of constant dividends on abnormal return

days AR T (AR) CAR T (CAR) -3 0.10% 0.48894 0.10% 0.48894 -2 -0.28% -1.39374 -0.18% 0.45240 -1 0.21% 1.04507 0.03% 0.04676 0 0.01% 0.04944 0.04% 0.04743 1 -0.08% -0.38269 -0.04% -0.03860 2 -0.02% -0.07846 -0.05% -0.04524 3 0.00% -0.02166 -0.06% -0.04187

5.4 Percentage change of dividend payments

This section examines whether the percentage change in dividend payments affect the abnormal return. Since the abnormal return on day 0 of positive dividend changes is significant and the abnormal return on negative dividend changes is almost significant, a positive relationship is expected between the percentage change in dividend payment and the abnormal return on day 0. Figure 3 plots the abnormal returns on day 0 and the percentage changes in dividend payments. Results in this paper show that constant dividend payments affect shareholder value very marginally. Therefore, these results are not included in the scatter diagram. The trend line with a R2 of 0.1859 shows indeed a positive relationship. This result shows that the larger the change in dividend payment is, on average the increase in abnormal return will be larger. However, the results are not very strong and this is in accordance with most of the current literature, who on average do not find such strong results as earlier studies. Since this paper also does not find very strong results, it strongly supports the study of Amihud & Li (2006) who argue that the information content of dividends is declining since these days more information is readily available and shareholders are better informed, the asymmetry between management and shareholders is fading away.

Figure 3: percentage change in dividends plotted against the abnormal returns on day 0

Percentage change in dividends versus abnormal return

-15,00% -10,00% -5,00% 0,00% 5,00% 10,00% 15,00% 20,00% -100,00% -50,00% 0,00% 50,00% 100,00% % dividend change A b n o rm a l r e tu rn

5.5 Economic climate and the amount of changes in dividend payments

A second issue in this paper discusses whether economic performance has effect on the amount of increases (decreases) in dividend payments. Therefore, a correlation test is conducted which combines the yearly return of the FTSE 100 with the amount of increases (decreases) in dividend payments. Table 8 provides an overview of the sample data used for this correlation test. The table is subdivided in four categories; (1) amount of constant dividend payments, (2) amount of increases in dividend payments, (3) amount of decreases in dividend payments and (4) the yearly changes in return on the FTSE 100. In this test, the correlation number can be between -1 and +1, where -1 implies perfect negative correlation and +1 implies perfect positive correlation.

Table 8: percentage change in amount of dividend payments and return of the FTSE 100

year no change %change increase %change decrease %change return FTSE

2000 16 n/a. 99 n/a. 10 n/a. -16%

2001 18 13% 106 7% 19 90% -16% 2002 21 17% 100 -6% 21 11% -14% 2003 25 19% 99 -1% 24 14% 31% 2004 18 -28% 116 17% 28 17% 18% 2005 10 -44% 135 16% 12 -57% 9% 2006 11 10% 176 30% 11 -8% 30% 2007 11 10% 171 -3% 27 145% 9%

Table 9: correlation between economic performance and amount of changes in dividend payments

correlation T (correlation)

Positive dividend payments -0,11 1.163

Negative dividend payments 0,46 -0.716

Constant dividend payments -0,30 -0.248

It is obvious that there is a relationship between the development of the FTSE 100 index and the direction of change in dividend payments, which implies that the H₆ can be accepted. The results from table 9 show, as expected, a negative correlation between the stock market and the amount of decreases in dividend payments of -0,30 and a positive correlation between the stock market and the amount of increases in dividend payment of 0,46. Nevertheless, none of these correlation tests is significant.

These results also supports the assumption of the behavioral finance aspect that managers act irrational when markets are bullish. I assume that managers increase dividends because they are too confident about future earnings. Although the theory that managers act irrational is supported by the results, it is mainly theoretical. A more specific study to this subject is necessary to check whether this theory hold.

On the other hand, empirical results of this study shows that markets also act irrational in the short-term since increases (decreases) in dividend announcements affect abnormal return positively (negatively). However, since the expectation about positive future earnings is based on the reliance that markets stay bullish, it is highly uncertain if these future earnings will be realized. So, in the long term these positive effects should disappear since no underlying economic value is added to the company.

VI. Conclusion

This paper examined whether dividend announcements affect shareholders wealth. To examine this empirically, the change in dividend payment is used to compute abnormal returns. Furthermore, it is examined whether economic conditions affect the amount of increases (decreases) in dividend payments.

The empirical results show that when the direction of change of dividend payment is positive, abnormal return on day 0 is significant on a 10% level and the cumulative abnormal return increases in the event period. This is in accordance with Lintner’s theoretical framework of the information content and in accordance with the empirical results of Watts (1973) who found relative weak results. Looking at the percentage change of positive dividend payments shows that a small increase between 0% and 10%, on average does not have a significant effect on the share price. Although an increase of 10% and higher, on average have a significant effect on day 0 on a 5% level. These results clearly indicate that a positive relation exists between the level of change in positive dividend payments and shareholders wealth. However, the empirical results are much weaker than found in the study of Pettit (1972).

The negative change in dividend payment shows somewhat different results. The abnormal return is only significant on day 3 and not, as expected, on day 0 which result is almost significant. Since the cumulative abnormal return in the event period is negative the H₂ hypothesis, that negative dividend announcements affect shareholder value negatively, can be accepted. However, the results are very weak and this is in accordance with Chen, Firth & Gao (2002), who argued that dividend announcements affect shareholder value only marginally.

The results of constant dividend payments are, as expected, not significant. This supports the H₄

hypothesis that constant dividend payments do not affect shareholder value.

of Amihud & Li (2006) who argue that the information content of dividends is declining since these days more information is readily available and shareholders are better informed, the asymmetry between management and shareholders is fading away.

The second issue in this paper is whether the performance of the economy is related to the amount of increases (decreases) in dividend payments. The correlation test performed to examine this empirically shows that indeed the performance of the FTSE 100 index affects the amount of increases (decreases) in dividend payments. When the market index increases, the more dividends increases and when the market index decreases, the more dividends decreases. This paper argues that these results are due to the behavioral finance aspect that managers act irrational in bull markets. In bull markets managers become too overconfident in the economy and the future earnings of their own company. This causes that they will increase dividends, which is not based on rational economic figures, but on irrational predictions about the future. However, this conclusion is mostly theoretical and further empirical research to this subject would be needed.

References

Abeyratna, G., Lonie, A.A., Power, M. and Sinclair, C.D., 1996. The Influence of Company Financial Performance on the Interpretation of Dividend and Earning Signals: A Study of Accounting- and Market Based Data, British Accounting Review, 28(3), 229-247

Ahoromy, J. and Swary, I., 1980. Quarterly Dividend and Earning Announcements and Stockholders’ Returns: An Empirical Analysis, The Journal of Finance, 35(1), 1-12

Amihud, Y. and Li, K., 2006. The Declining Information Content of Dividend Announcements and the Effect of Institutional Holdings, Journal of Financial and Quantitative Analysis, 41(3), 637-660

Baker, M. and Wurgler, J., 2004. A Catering Theory of Dividends, The Journal of Finance, 59(3), 1125-1165

Balachandran, B., 1998. Dividend Reductions and Interim Effect: UK evidence, working paper, School of Accounting & Commercial Law, Victoria University of Wellington

Balachandran, B., 2003. UK Interim and Fiscal Dividend Reductions: A Note on Price Reaction, The European Journal of Finance, 9(4), 379-390

Benartzi, S., Michaely, R. and Thaler, R., 1997. Do Changes in Dividends signal the Future of the Past, The Journal of Finance, 52(3), 1007-1034

Black, F. and Scholes M., 1978, The Effects of Dividend Yield and Dividend Policy on Common Stock Prices and Returns, Journal of Financial Economics, 1(1), 1-22

Bond, S.R., Chennells, L. and Devereux. M.P., 1996, Taxes and Company Dividends: A Microeconometric Investigation Exploiting Cross-Section Variation in Taxes, Economic Journal, 106, 320-333

Campbell, J.Y., Lo, A.W. and Mackinlay, A.C., 1997. The Econometrics of Financial Markets, 1st edition. Princeton University Press. Princeton: New Jersey

Chen, G., Firth, M. and Gao, N., 2002. The Information Content of Concurrently Announced Earnings, Cash Dividends and Stock Dividends: An Investigation of the Chinese Market, Journal of International Financial Management & Accounting, 13(2), 101-124

Conroy, R.M., Eades, K.M. and Harris, R.S., 2000. A Test of the Relative Pricing effects of Dividends and Earnings: Evidence from Simultaneous Announcements in Japan, The Journal of Finance, 55(3), 1199-1227

Easterbrook. H., 1984. Two Agency-Costs Explanations of Dividends, The American Economic Review, 74(4), 650-659

Fama, E. F. and French, K.R., 2001. Disappearing Dividends: Changing Firm

Characteristics or Lower Propensity to Pay? Journal of Financial Economics, 60(1), 3-44

Gunasekarage, A. and Power, D.M., 2002. The Post-announcement Performance of Dividend-changing Companies: The Dividend-signaling Hypothesis Revisited, Accounting & Finance, 42(2), 131-151

Grullon, G., Michaely, R. and Swaminathan, B., 2002. Are Dividend Changes a sign of Firm Maturity? Journal of Business, 57(4) 387-424.

Jensen, M.C., 1986. Agency Costs of Free Cash Flow, Corporate Finance and Takeovers, American Economic Review, 76(2), 323-329

Kahneman, D. and Tversky, A., 1979. Prospect Theory: Analysis of Decisions under uncertainty, Econometrica, 47(2), 263-291

Lang, L.H.P. and Litzenberger, R.H., 1989. Dividend Announcements: Cash Flow Signaling vs. Free Cash Flow Hypothesis, Journal of Financial Economics, 24(1), 137-154

Lintner, J., 1956. Distribution of Incomes of Corporation among Dividends, Retained Earnings and Taxes, American Economic Review, 46(2), 97-113

Miller, M.H. and Modigliani, F., and 1961. Dividend Policy, Growth and the Valuation of shares, Journal of Business, 34(4), 411-433

Miller, M. and Rock, K., 1985. Dividend Policy, Under asymmetric information, The Journal of Finance, 40(4), 1031-1051

Nofsinger, J.R., 2005. The Psychology of Investing, 3th edition. New Jersey: Pearson Education inc.

Ogden, J.P., 1994. A Dividend Payment Effect in Stock Returns, The Financial Review, 29(3), 345-369

Pettit, P.R., 1972. Dividend Announcements, Security Performance, and Capital Market Efficiency, The Journal of Finance, 27(5), 993-1007

Renneboog, L. and Trojanowski, G., 2006. Patterns in Payout Policy and Payout Channel Choice of UK Firms, European Corporate Governance InstituteECGI Finance Discussion Paper, 70

Ross, S.A., Westerfield, R.W. and Jaffe, J., 2005. Corporate Finance, international edition. Boston: McGraw - Hill

Thaler, R.H. and Shefrin, H.M., 1981. An Economic Theory of self-control, Journal of Political Economy, 39(2), 392-406

von Eije, J. H. and Megginson, W.L., 2008. Dividend Policy in the European Union, Journal of Financial Economics, forthcoming

Appendix

Table 10: the effect of dividend changes on abnormal returns using a two-factor model1

Increase in dividends Decrease in dividends No change in dividends days T (AR) T (CAR) T (AR) T (CAR) T (AR) T (CAR) -3 -0,07536 -0,07536 0,36330 0,36330 0,44812 0,44812 -2 -0,37679 -0,22607 -0,10380 0,12975 -1,09541 -0,32364 -1 0,67822 0,07536 0,20760 0,01730 0,59750 -0,01660 0 1,88395* 0,52751 -1,55702 -0,37628 0,54771 0,12448 1 0,37679 0,49736 0,31140 -0,23874 -0,04979 0,08962 2 0,75358 0,54006 0,15570 -0,17300 0,19917 0,10788 3 0,15072 0,48444 -1,08991 -0,30399 -0,59750 0,00711 1

The two-factor model uses the R_m,t(FTSE 100 index) and theR_g,t(10-year UK government bond rate)

Table 11: the effect of dividend changes on abnormal returns using an event window of -10 till +10

Increase in dividends Decrease in dividends No change in dividends days T (AR) T (CAR) T (AR) T (CAR) T (AR) T (CAR) -10 0.08597 0.08597 0.82470 -0.82470 1.57383 1.57383 -9 0.34696 0.21646 0.32005 -0.25232 0.51379 1.04381 -8 -0.35577 0.02572 -0.22375 -0.24280 -1.59475 0.16429 -7 0.31704 0.09855 1.25596 0.13188 1.38532 0.46954 -6 -0.31786 0.01527 0.47111 0.19973 0.64040 0.50373 -5 -0.16544 -0.01484 0.01534 0.16900 -0.84271 0.27932 -4 0.04587 -0.00617 0.33458 0.19265 -1.83571 * -0.02282 -3 0.37805 0.04185 0.36900 0.21470 0.54922 0.04868 -2 -0.14313 0.02130 0.09892 0.20183 -1.56556 -0.13068 -1 1.07826 0.12699 -0.00994 0.18065 1.17390 -0.00022 0 1.91622 * 0.28965 -1.37371 0.03935 0.05553 0.00484 1 0.21900 0.28376 0.64322 0.08967 -0.42986 -0.03137 2 0.55887 0.30492 0.06240 0.00140 -0.08813 -0.03574 3 0.18995 * 0.29671 -1.66029 * -0.03727 -0.02432 -0.03492 4 -0.44924 0.24698 1.08758 0.03772 -0.51680 -0.06705 5 -1.35306 0.14698 -1.64167 -0.06724 -0.66440 -0.10438 6 -0.65123 0.10002 -0.32350 -0.08231 0.55519 -0.06558 7 -0.19385 0.08370 -1.36621 -0.15364 -1.66087* -0.15421 8 -1.44186 0.00340 -1.01070 -0.19875 0.54848 -0.11723 9 -0.84102 -0.03881 -1.33363 -0.25549 -1.53110 -0.18792 10 -0.23296 -0.04805 0.47374 -0.22077 0.62359 -0.14928

Table 12: the effect of dividend changes on abnormal returns using an event window of -5 till +5

Increase in dividends Decrease in dividends No change in dividends days T (AR) T (CAR) T (AR) T (CAR) T (AR) T (CAR) -5 -0.16628 -0.16628 0.01544 0.01544 -0.75267 -0.75267 -4 0.04611 0.02305 0.33663 0.17604 -1.63955 -0.81977 -3 0.37998 0.12666 0.37126 0.24111 0.49053 0.16351 -2 -0.14386 0.05903 0.09953 0.20571 -1.39827 -0.22693 -1 1.08377 0.26397 -0.01001 0.16257 1.04847 0.02814 0 1.92602* 0.54098 -1.38212 -0.09487 0.04959 0.03172 1 0.22012 0.49514 0.64716 0.01113 -0.38393 -0.02765 2 0.56173 0.50347 0.06279 0.01758 -0.07872 -0.03404 3 0.19092 0.46874 -1.67044* -0.16997 -0.02172 -0.03267 4 -0.45154 0.37671 1.09424 -0.04355 -0.46157 -0.07556 5 -1.35998 0.21883 -1.65171* -0.18974 -0.59341 -0.12264

* Denote significance at the 10% level. ** Denote significance at the 5% level. *** Denote significance at the 1% level.

Table 13: the effect of dividend changes on abnormal returns using an event window of -1 till +1

Increase in dividends Decrease in dividends No change in dividends days T (AR) T (CAR) T (AR) T (CAR) T (AR) T (CAR) -1 1.09160 1.09160 -0.01011 -0.01011 1.04781 1.04781 0 1.93992* 1.51576 -1.39623 -0.70316 0.04956 0.54869 1 0.22171 1.08441 0.65377 -0.25085 -0.38369 0.23789

* Denote significance at the 10% level. ** Denote significance at the 5% level. *** Denote significance at the 1% level.

Table 14: Companies incl. in the sample

1. Admiral Group

2. Alliance & Leicester

29. Experian Group Ltd 30. FirstGroup 31. Friends Provident 32. G4S Plc 33. GlaxoSmithKline 34. Hammerson 35. HBOS Plc 36. Home Retail Group 37. HSBC Hldg 38. ICAP 39. Imperial Tobacco 40. Intercontinental Hotels 41. International Powers 42. ITV Plc 43. Johnson Matthey 44. Kazakhmys Plc 45. Kingfisher 46. Land SecuritiesGroup 47. Legal & General 48. Liberty Intl. 49. Lloyds-TSB Group 50. Lonmin