Print Reference This Reddit This

Strategic Ownership Dividend Policy And Price Drop Ratio Finance Essay

In this proposed research, we suppose that ex-dividend day price behavior is motivated by tax heterogeneity between dividend and capital gain and also size of dividend yield (dividend policy). Furthermore, it supposed that dividend policy is a function of tax heterogeneity between dividend and capital gain and strategic ownership. Based on literature review, it is assumed that price drop ratio on ex-dividend day motivated by dividend policy and while dividend policy caused by strategic ownership, so we can assume that ex-dividend day price behavior is motivated by strategic ownership. Cross-sectional data will be use to investigate the effect of strategic ownership on dividend policy. In addition, another regression model will predict the mediator effect of strategic ownership on price drop ratio on ex-dividend day.

Therefore, it is supposed that strategic ownership is capable of moderating the effect of tax heterogeneity and dividend policy on price drop ratio. By introducing strategic ownership, we try to increase the accuracy of regression model for predicting price drop ratio on ex-dividend day, which seems to be a puzzle yet.

Background of Study

1.1 Introduction

On a stock’s ex-dividend day, we should expect to see stock`s price drop by the amount of the dividend. If the stock were to be sold before that day, the seller would relinquish to the buyer the right to receive the dividend. After the stock goes ex-dividend, however, the right to the dividend remains with the seller and is not transferred to the buyer. Ignoring the time value of money in the short period between the ex-date and the dividend payment date (usually 2-4 weeks), the stock price must drop by the value of the dividend to prevent arbitrage.

However, much empirical research has documented that stock prices drop by significantly less than the dividend on the ex-day. The first explanation for this effect [Elton and Gruber (EG 1970)] cited preferential tax treatment of capital gains. Specifically, when dividends are taxed at a higher rate than capital gains, investors would not be willing to subject themselves to higher tax rates by holding a stock through the ex-dividend day unless they were compensated by a smaller ex-day price decline. However, there has not been unanimous agreement about whether ex-dividend stock price behavior is truly driven by differential taxation of dividends and capital gains.

Other arguments [e.g., Kalay (1982)] point out that short-term traders are taxed identically on dividends and capital gains, so any deviation from a one-for-one price drop-to-dividend relationship creates an apparent arbitrage opportunity if transaction costs are low enough. However, if higher transaction costs inhibit “dividend capture” activity, the ex-day effect hypothesized by EG may still persist.

More recent explanations for the ex-day effect point to different aspects of market microstructure. One of the interesting microstructure models is that of Bali and Hite (1998). They argue that when dividends may be paid in any amount but stock prices are constrained to discrete multiples of 12.5 cents, ex-day prices should drop by the largest tick multiple less than the dividend, regardless of tax considerations. Recent events have provided an excellent opportunity to test the discreteness hypothesis, as the major exchanges converted to “decimalized” pricing in 2001, thereby removing the supposed cause for the effect. Graham, Michaely, and Roberts (2003) and Jakob and Ma (2004) find that the ex-day effect continues to exist even in the absence of large discrete price increments, thereby casting serious doubt on the explanatory power of the discreteness model.

This study focuses almost exclusively on tax-based explanations for ex-dividend stock price behavior, while we consider strategic ownership as a moderating variable. This moderating variable provides an obvious opportunity to improve the predictability of price drop ratio on ex-dividend day. In this research, we will investigate the accuracy of this claim in practical world.

1.2. The importance of the study

The ex-day pricing problem has endured as a subject of interest in financial economics for several reasons. First, ex-day pricing ultimately suggests something about how stockholders value dividends versus capital gains, which has implications for corporate dividend policy (assuming managers maximize shareholder wealth).

Next, with improved knowledge of the determinants of ex-day pricing, all market participants can make better decisions, either by timing transactions or by engaging in other profitable trading activities around the ex-day.

Finally, the ex-day phenomenon is an anomaly with implications for market efficiency, and such anomalies naturally call for an explanation.

Research Questions and Objectives

This research will examine the moderating effect of strategic ownership on ex-day price behavior. Therefore, Main objective of this research is whether strategic ownership has a moderating effect on price drop ratio on ex-dividend day.

According to the literature, price drop ratio is a function of tax heterogeneity and amount of dividend yield (dividend policy). In addition, the second school of thought argues that dividends create a tax disadvantage for the investors who received them because they are taxed much more heavily than price appreciation (capital gain). However, it is the firm`s marginal stockholders who determine how dividends are viewed. Individual investors may dislike the tax on high dividends, but pension funds, institutional investors and other strategic owners may not share the same misgivings because of their lower tax statue or tax-exemption rules. So, it assumed that there is a relationship between dividend policy and strategic ownership. So, the main question of this research is as follow:

Does strategic ownership moderate the effects of tax-heterogeneity and dividend policy on price drop ratio on ex-dividend day?

This main question includes several sub questions as follow:

Is there any relationship between tax-heterogeneity and price drop ratio on ex-dividend day?

Is there any relationship between dividend policy and price drop ratio on ex-dividend day?

Is there any relationship between tax heterogeneity and dividend policy?

Does the relationship between tax-heterogeneity and dividend policy moderate by strategic ownership?

Does the relationship between tax-heterogeneity and price drop ratio moderate by strategic ownership?

Does the relationship between dividend policy and price drop ratio moderate by strategic ownership?

Literature Review

3.1 Tax heterogeneity and Price drop ratio

In one of the earliest published studies on the ex-day anomaly, Campbell and Beranek (1955) observed that the ex-dividend price drop was on average slightly less than the dividend (~90%) for a small sample of medium-sized to large cash distributions paid on NYSE stocks. While they do noted that “a tax-paying individual will do better to sell before an ex-dividend date but to buy after it,” and that “for a tax-exempt institution, the rule is exactly the reverse,” they did not attempt to give a formal explanation for why the price should not normally drop by the full dividend amount.

Using a one-year sample with 4,148 dividends, Elton and Gruber (EG 1970) confirmed that ex-day stock prices tend to fall by significantly less than the dividend and developed a model explaining the effect. They showed that when dividends are taxed at a higher rate than capital gains, the stock price must drop by less than the dividend for investors to be indifferent between (i) selling the stock cum-dividend (i.e. before the distribution) and (ii) holding the stock, receiving the dividend, and selling ex-dividend.

If selling the stock cum-dividend, the value will be:

Value1 = Pcum * (1 - tcg)

If holding the stock, receiving the dividend, and selling ex-dividend the value will be:

Value2 = D(1 – td) + Pex * (1 - tcg)

Where Pcum is the last cum-dividend price, Pex is the first ex-dividend price, D is the dividend amount, td is the tax rate applied to dividend income, and tcg is the tax rate applied to capital gains.

For rational investors, these two values should be the same and indifference:

Value1 = Value2 →

Pcum * (1 - tcg) = D(1 – td) + Pex * (1 - tcg) →

Pcum * (1 - tcg) - Pex * (1 - tcg) = D(1 – td) →

(Pcum - Pex) * (1 - tcg) = D(1 – td) →

(Pcum - Pex) / D = (1 – td) / (1 - tcg)

Specifically, the relationship is given by :

Price Drop Ratio

It follows that PDR < 1 when tcg < td, as the rates have frequently been for individual investors.

Critics of the tax theory have noted that preferential tax treatment of capital gains over dividends applies only to long-term individual investors. Corporations historically have been allowed to exclude 70 to 85 percent of dividends received from taxable income, thereby reducing their effective dividend tax rates below the capital gains rate. Other investors, such as short-term traders and tax-exempt funds, are taxed (or not taxed) identically on dividends and capital gains and thus should be indifferent between the two types of income. In fact, short-term arbitrageurs should exploit any difference between the dividend and the ex-day price drop until the two are approximately equal, with any remaining random discrepancies due to market imperfections. [See Brooks and Edwards (1980), Kalay (1982), Miller and Scholes (1982), and Lakonishok and Vermaelen (1986) for arguments advocating this position.] This reasoning is compelling, and it casts doubt on the validity of marginal tax rates inferred from the EG model.

Kalay (1982) revisits the original EG study using a slightly different methodology for the same sample period to correct for potential biases. Interestingly, even after the adjustments, he still affirms EG’s basic findings that ΔP/D is on average less than 1 and positively related to the dividend yield.

In view of the above arguments, the persistence of abnormal ex-day returns is at first puzzling. After all, it seems that anything but a one-for-one price drop creates a virtual arbitrage opportunity. On the other hand, there is little doubt that positive ex-day returns do happen, as documented by countless studies, many of which are cited later in this review. Of course, negative ex-day returns (where the price drops by more than the dividend) are well-documented, too – usually in stocks with the highest dividend yields.

High-yield stocks are often attractive targets for “dividend capture” by corporations who are generally taxed much less on dividends than on capital gains. Upon further reflection, it should also be puzzling if the tax preferences of long-term investors do not influence stock prices around the ex-dividend date.

It might seem, then, that the EG model of tax-based dividend clienteles is in competition with the short-term trading hypothesis, and that the focus of empirical tests should be to determine which one better describes observed price behavior. It would be a mistake, however, to view the two hypotheses only as separate, opposing stories. In fact, they can be complementary, as different classes of traders may coexist, and all may influence the supply of and demand for securities around the ex-date.

While most ex-day pricing research has focused on taxable cash dividends on common stocks, one can also make inferences about existing theories by observing the price behavior of different securities and around other distribution types. Eades, Hess, and Kim (1984) document negative excess returns for preferred dividends, as might be expected for high-yield dividend capture targets. However, Stickel (1991) obtains conflicting results. In his sample of nonconvertible preferred stocks, he finds positive abnormal returns and volume on the ex-day, with returns declining for more liquid stocks. So far, this is consistent with a synthesized model where both long-term investors and short-term arbitrageurs influence prices around preferred dividends. Inconsistent with this framework, however, is Stickel’s finding that trading volume increases with liquidity for low-yield but not high-yield preferreds. Preferred dividends are of course relevant to ex-day pricing theories, but it is perhaps more interesting to compare observations around non-taxable distributions against those around the usual taxable dividends.

Eades, Hess, and Kim (1984) and Grinblatt, Masulis, and Titman (1984) examine ex-day price behavior around stock dividends and splits, which are non-taxable. According to both the EG model and the short-term trading hypothesis, these studies should find ex-day price drops fully reflective of the dilution caused by additional shares. In fact, neither does. Eades et al. (1984) report that “non-taxable stock dividends and splits are priced on ex-dividend days as if they are fully taxable.” Oddly, Grinblatt et al. (1984) note higher positive ex-day returns for stock dividends than splits, possibly due to the added inconvenience investors face when dealing with odd lots.

Green and Rydqvist (1999) study a unique security –Swedish lottery bonds– to which special rules apply. Coupon payments on the bonds (distributed by lottery) are not subject to income tax, but capital gains are taxed at the ordinary rate. Furthermore, the regulatory environment is not conducive to short-term arbitrage using these securities. Consistent with the EG tax model, Green and Rydqvist find that the bond price drops by about 130 percent of the distribution on the “ex-coupon” day, and that the bonds frequently trade at negative pre-tax yields.

Elton, Gruber, and Blake (2005) examine two samples of closed-end funds. In one sample, distributions are not taxed (but capital gains are); in the other, distributions are taxed normally. As expected, market-adjusted price drop ratios are greater than one for the non-taxable distributions but less than one for the taxable sample. Price drop ratios for both samples also behave as predicted by the EG model following tax law changes in 1993 and 1997. Similarly, Milonas, Travlos, Xiao, and Tan (2002) examine taxable and non-taxable dividends in the Chinese stock market and find price behavior mostly consistent with the tax theory.

So, base on this literature review, tax heterogeneity can be introduced as the main explanatory variable, which can be modeled as follow:

Price Drop Ratio

3.1.1 Short-Term Trading and Transaction Costs

The tax heterogeneity framework suggests that positive ex-day returns may exist for some stocks, but those returns are likely to vanish (or even become negative) when the dividend yield is high enough, the risk small enough, and/or transaction costs low enough to attract arbitrageurs or even corporate dividend capturers. Consistent with this view, Karpoff and Walkling (1990) find for a sample of NASDAQ stocks that ex-day returns were positively related to transaction costs as measured by bid-ask spreads, especially for high-yield stocks.

Michaely and Vila (1995, 1996) find evidence that abnormal ex-day trading volume was greater for high-yield stocks, lower for riskier stocks, and higher when options could be used to hedge a risky dividend capture position. In May 1975, the New York Stock Exchange (NYSE) switched from a fixed commission system to negotiated commissions, a move that is widely agreed to have reduced transaction costs. Consistent with the integrated tax clientele/short-term trading framework, studies generally find lower ex-day returns and higher trading volume after the introduction of negotiated commissions, particularly for higher-yield stocks with lower measured transaction costs proxies. [See Lakonishok and Vermaelen (1986), Karpoff and Walkling (1988), Eades, Hess, and Kim (1984, 1994), and Naranjo, Nimalendran, and Ryngaert (2000).]

However, transaction cost is not introduced into this proposed research model, because it variant through different size (lots) of trading and it is difficult to be measured for each stock one by one. However, if the average transaction cost is considered for each stock, the effect is the same for all stocks. So, it can be eliminated from the model.

3.2 Dividend policy and Price drop ratio

EG sort their sample into deciles by dividend yield and compute the mean PDR for each group. Using the equation above, EG then impute investors’ marginal tax rates for each dividend yield deciles. They find that PDRs generally increase with dividend yield, suggesting that investors in lower tax brackets prefer stocks with higher dividend yields, while higher bracket investors prefer lower-yield stocks. Thus, EG confirm the existence of the ex-day effect, derive a mathematical model tying the effect and its magnitude to tax rates, and provide economic reasoning for a specific type of “dividend clientele” as suggested generally by Miller and Modigliani (1961).

The size of the ex-day effect documented by EG is actually somewhat larger than what had been previously documented by Campbell and Beranek (1955) and Durand and May (1960). However, this is exactly what the EG tax theory would predict since both earlier studies used samples of stocks with higher dividend yields.

Boyd and Jagannathan (1994) develop a theoretical model that explicitly models the tax heterogeneity brought about by interaction among different types of market participants. Their model predicts a nonlinear relationship (for which they find some evidence in their data) between the percentage price drop (ΔP/P) and the dividend yield. It also shows that when the dividend yield is high enough, corporate dividend capturers have a trading advantage over tax-neutral arbitrageurs, thus explaining why PDRs are often greater than one for high-yield stocks. In the same spirit, Michaely and Vila (1995, 1996) develop a model and provide empirical evidence suggesting that abnormally high ex-day trading volume occurs precisely because of the above-mentioned tax heterogeneity. They also note that “because of the risk involved [in dividend capture strategies], no traders will take an unlimited position, regardless of the price movement.” Thus, ex-day abnormal returns do not necessarily imply that arbitrageurs are not at work; instead, they may have taken the largest possible positions that their tolerances for risk will allow.

According to the literature till now, price drop ratio supposed to be a function of tax heterogeneity and size of dividend (dividend policy):

Price Drop Ratio =

Where is the average price and D is dividend.

3.3 Tax Heterogeneity and Dividend Policy

As shown by Miller and Modigliani (1961), dividend policy does not affect share prices under the assumption of a perfect and complete capital market. In their theoretical model, a firm is capable to achieve any desirable payout level by issuing or repurchasing common shares and every shareholder can replicate any desired stream of payment by selling or purchasing common shares at the capital market (Ibid). However, in reality capital markets are not perfect, and one reason for that is taxation.

As the market value of shares is determined by after-tax cash flows, any differential tax treatment of capital gains relative to dividends may influence corporate dividend policy. The changes in dividend policy may include altering the dividend payout ratio and replacing cash dividends with stock repurchases or stock dividends.

In the case when all investors are taxed similarly and the effective tax rate on dividends exceeds the effective tax rate on capital gain, it is optimal to pay no dividends at all. Even if the statutory tax rates are similar, the possibility to choose the moment to realize the capital gains (i.e., tax timing option) is valuable (see e.g., Constantinides 1984, Emery and Gehr 1988) and therefore from tax perspective, unrealized capital gains are superior to dividends. If, on the other hand, the effective tax rate on dividends is lower than on capital gains and taxes are the only source of market imperfection, it would be optimal to pay out all earnings as dividends and use external financing.

By the way, base on the second school of thought for dividend policy, which argue that dividends create a tax disadvantage for the investors who received them because they are taxed much more heavily than price appreciation (capital gain), so dividend policy assumes to be a function of Tax Heterogeneity.

Dividend Policy = ε0

3.4 Strategic Ownership and Dividend Policy

Dividend policy is unaffected if the effective tax rates on dividends and capital gains are the same, but different investors (domestic vs. foreign, individual vs. corporate) are taxed differently. However, such tax discrimination may lead to permanent changes in the ownership structure of companies.

In a world without transaction and agency costs and information asymmetry, but with different tax rates on dividends and capital gain, companies should choose the dividend policy which is the most tax effective. However, in reality different investors may prefer different dividend policies due to differential tax treatment of investors (e.g., non-flat tax rates, different tax rates for domestic and foreign investors, different tax rates for individual and institutional investors, etc). As dividend income is usually taxed more highly than capital gain, this suggests that investors in high tax brackets should hold shares with a low dividend yield and vice versa (so-called dividend clientele2). So, it can be assumed that the relationship between tax heterogeneity and dividend policy can be moderated by the ownership structure.

Dividend Policy =

ε0

Where, stow is strategic ownership. Strategic ownership is the percentage of total shares in issue held strategically by following institution and not available to ordinary shareholders.

Foreign held shares: The percentage of total shares in issue held by an institution domiciled in a country other than that of the company.

Employee held shares: The percentage of total shares in issue held by employees, or by those with a substantial position in a company that provides significant voting power at an AGM.

Investment’s company held shares: The percentage of total shares in issue held as long term strategic holdings by investment banks or institutions seeking a long term return.

Pension fund held shares: The percentage of total shares in issue held by pension funds or endowment funds.

Cross holdings: The percentage of total shares in issue held by one company in another.

Government held shares: The percentage of total shares in issue held by a government or government institution.

Other holding: The percentage of total shares in issue held strategically outside one of the following categories.

In their excellent survey of literature on payout policy, Allen and Michaely (2002) distinguish between two types of clientele models: static and dynamic. The most important difference between them is that in static models investors trade just once, while in dynamic models they are allowed to trade multiple times (Ibid). According to Allen and Michaely (2002), a static view is appropriate when transaction costs are extremely high.

If the static form of dividend clientele really exists, companies should not change their dividend policy too often, as it would cause shareholders to switch firms and burden them with brokerage costs, and in many cases also with capital gain taxes. Brav et al. (2005) document that managers hesitate about introducing extreme changes in their payout policy as it might cause changes in ownership structure, thereby negatively affecting the company’s stock price.

Changes in dividend policy are also viewed as signals about companies’ future prospects and therefore companies may be reluctant to lower their payout ratios even if it would be tax effective (see e.g., Miller and Rock (1985), John and Williams (1985)). Most of the earlier studies used static models. Elton and Gruber (1970) among others found some evidence to confirm the existence of static clientele. However, direct studies of stock ownership have found no significant tendency for high-income groups to prefer stocks with a low dividend yield (Kalay 1982).

Price Drop Ratio and Strategic Ownership

3.5.1 Stock Strategic Ownership and ex-Dividend Day Trading Volume

Michaely and Vila (1995) showed that Ex-day excess trading volume is determined in the following manner:

where D is the amount of dividend per share, Ki is the level of risk tolerance for investor i, αi, is the tax-induced preference for dividends versus capital gains for investor i, is the average preference for dividends versus capital gains in the economy, weighted by investors’ levels of risk tolerance Ki, σe2 is the total risk of a stock, and N is the total number of investors in the economy. This equation indicates that trading volume is a positive function of the dividend and of the degree of tax-induced heterogeneity. If there is no tax-induced investor heterogeneity, that is, if αi = for all i’s, then there is no extra trading around ex-dividend days, even in the presence of tax-disadvantaged dividends (e.g., even if < 1). Thus, one necessary condition for ex-day excess trading volume is the presence of tax-induced investor heterogeneity.

As, this research just examines the relationship between strategic ownership and trading volume around ex-dividend day and there is not another research in this area which investigates the effect of ownership structure on price drop ratio, it seems to be a gap in ex-dividend literature.

3.5.2 The moderating effect of ownership structure on price drop ratio

According to the literature review we know that price drop ratio is a function of tax heterogeneity and dividend policy as follow:

Price Drop Ratio =

Moreover, the second school of thought on dividend policy argued that dividend policy is a function of tax heterogeneity which moderates by strategic ownership as follow:

Dividend Policy =

ε

So, what seems to be a gap and we are going to investigate it, is the moderating effect of strategic ownership on ex-dividend day price drop ratio.

This research gap will be investigated through following model:

Price Drop Ratio =

Theoretical frame work and Hypothesises

As mentioned in literature review, the prick drop ratio on ex-dividend day is a function of tax heterogeneity and Dividend policy (amount of dividend yield to average stock price). Dividend policy is a function of tax heterogeneity which moderates by strategic ownership. So, our theoretical frame work could present as follow:

Price Drop Ratio

H2

H1

H6

H5

H2

H1

Strategic Ownership

H6

H4

Tax Heterogeneity

Dividend Policy

H3

H3

H5

H4

Base on this theoretical frame work, Research Hypothesizes could be mentioned as below:

Hypothesis 1: Holding other variables constant, there is a significant relationship between tax heterogeneity and price drop ratio.

Hypothesis 2: Holding other variables constant, there is a significant relationship between dividend policy and price drop ratio.

Hypothesis 3: Holding other variables constant, there is a significant relationship between tax heterogeneity and dividend policy.

Hypothesis 4: Holding other variables constant, the relationship between tax-heterogeneity and dividend policy moderates by strategic ownership.

Hypothesis 5: Holding other variables constant, the relationship between tax heterogeneity and price drop ratio moderates by strategic ownership.

Hypothesis 6: Holding other variables constant, the relationship between dividend policy and price drop ratio moderates by strategic ownership.

5. Research Methodology

5.1. Research Design

Research design involves decisions regarding the purpose for study, where the study will be conducted, the extent to which the researcher manipulates and control the study, the temporal aspects of the study and the level at which the data will be analyzed.

The purpose of study is hypothesis testing, the study setting is non-contrived, the researcher interference is minimal, the time horizon is cross-sectional, and the unit of analysis is individual (in the firm level).

Since the knowledge about the concepts is extensive, the hypothesis testing is suitable and as the time horizon covers a one years period in the past, the time horizon is cross-sectional, and since the research question is about each firm, the unit of analysis, individual, is suitable.

Moreover, since I will use accounting base data to calculate the strategic ownership and dividend policy and so we used exactly the financial statement, the validity is high. Also, I will use stock exchange data to calculate price drop ratio, Taxes and transaction costs, which have a high validity as well. We will use data-stream database for getting our required data.

Also, we know that the reliability of a measure indicates the extent to which the measure is without bias and hence offers consistent measurement across time and across the various items in the measurement. As the measures of the study are based on the theory of previous studies, we can be somewhat sure that the reliability is high too.

In order to examine the research hypotheses, the Price Drop Ratio for all companies existing in the sample for one year (2009) will be calculated by this formula:

Price Drop Ratio

Where Pcum is the last cum-dividend price, Pex is the first ex-dividend price and D is the dividend amount.

The ratio of number of stocks owned by strategic owners (n) to total number of stocks (N) will be calculated as strategic ownership.

In other words, Strategic ownership is the percentage of total shares in issue held strategically by following institution and not available to ordinary shareholders.

Foreign held shares: The percentage of total shares in issue held by an institution domiciled in a country other than that of the company.

Employee held shares: The percentage of total shares in issue held by employees, or by those with a substantial position in a company that provides significant voting power at an AGM.

Investment’s company held shares: The percentage of total shares in issue held as long term strategic holdings by investment banks or institutions seeking a long term return.

Pension fund held shares: The percentage of total shares in issue held by pension funds or endowment funds.

Cross holdings: The percentage of total shares in issue held by one company in another.

Government held shares: The percentage of total shares in issue held by a government or government institution.

Other holding: The percentage of total shares in issue held strategically outside one of the following categories.

Tax heterogeneity can be measured as follow:

Where, td is the tax rate applied to dividend income, tcg is the tax rate applied to capital gains.

Finally dividend policy can be calculated by dividing amount of dividend (D) to the average share price .

Dividend Policy =

In order to examine the first hypothesis (the relationship between tax heterogeneity and price drop ratio), the following regression model will be examined.

Price Drop Ratio

Second hypothesis which investigate the relationship between dividend policy and price drop ratio can estimated by following regression model:

Price Drop Ratio

To re-examine the overall effect of tax heterogeneity and dividend policy on price drop ratio the following model can be applied:

Price Drop Ratio

In the same way, the effect of tax heterogeneity on dividend policy, which is the third hypothesis, can be estimated as follow:

Dividend Policy = ε0

The moderating effect of strategic ownership on the relationship between tax heterogeneity and dividend policy (forth hypothesis) can be tested by following model:

Dividend Policy =

ε

Likewise, the moderating effect of strategic ownership on the relationship between tax heterogeneity and price drop ratio (fifth hypothesis) can be tested as follow:

Price Drop Ratio =

The moderating effect of strategic ownership on the relationship between dividend policy and price drop ratio (sixth hypothesis) can be tested as follow:

Price Drop Ratio =

To re-examine the overall moderating effect of strategic ownership on the relationship between Tax heterogeneity, dividend policy and price drop ratio, the following model is introduced:

Price Drop Ratio =

5.2 Data Collection

5.2.1 Research Population

The population of the research includes all major companies associated in Kuala Lumpur stock exchange and also NYSE during one year period (2009), excluded companies which their data are not accessible or complete during this period of time. The reason for choosing these two stock exchange market is that capital gain in Kuala Lumpur stock exchange is not subject to tax, while in NYSE the shareholders have to pay capital gain tax.

The taxes on capital gain and dividend in NYSE and KLSE are as follow for year 2009:

Country

Capital Gain Tax Rate

Dividend Income Tax Rate

USA

15%

35%

Malaysia

0%

26%

5.2.2 Sampling

All major companies associated in Kuala Lumpur stock exchange and also NYSE during one year period (2009), which possesses the required information (audited financial statements) has be chosen as a sample for calculating the model variables. So, the type of sampling design will be use in the study is probability. In the probability sampling, the elements in the population have same known chance of being selected as sample subject. As I explained above, because I will consider entire population and exclude firm that did not pass the criteria, all firms had the same chance to be selected. So, the type of sampling is probability.

Base on research proposals, the hypothesis testing will be done for both stock market separately or as a pool sample (whenever tax heterogeneity considered).

.

5.2.3 Data Gathering Tools

The required data will be gathered by referring to the existing data on data-stream data-base, which are secondary data.

5.2.4 Data Analyzing

In this research, the data will be analyzed by Eviews software. In order to examine the research hypothesis, first data will be check for unit root problem. After that, hypothesized regression model will be test and all other tests such as multi-collinearity, heteroskedasticity and auto-regression (which are competent with the specific characteristics of this research) will be used. For examining the meaningfulness of the parameters, t, F and Chi2 statistics will be used to interpreting the regression results.

6. The time schedule of the research

The entire time for doing this research is 1.5 years.

Gant chart (Project schedule):

Need help with your literature review?

Our qualified researchers are here to help. Click on the button below to find out more:

Literature Review Service

Related Content

In addition to the example literature review above we also have a range of free study materials to help you with your own dissertation: