Beta stability and the 1986 Tax Reform Act: some dividend policy implications.

Casey, K. Michael ; Duncan, John B. ; Watters, Michael P. 等

INTRODUCTION

The 1986 Tax Reform Act (TRA) represented the greatest change in United States tax laws in recent years, both for corporations and individuals. One of the major changes made by TRA was to eliminate much of the preferential treatment afforded capital gains. With the passage of TRA, the tax treatment of private investor income, whether from capital gains or dividends, became more balanced. Assuming that investors maximize after-tax, risk-adjusted portfolio returns, investment preferences should have changed.

Between June 1981 and December 1986, the federal government allowed taxpayers to exclude 60 percent of capital gains from taxation. During this time period, maximum marginal tax rates of 50 percent applied to dividend income and maximum marginal tax rates of 20 percent applied to capital gains. The 1986 TRA eliminated the capital gains exclusion, raising the maximum capital gains tax rate from 20 to 28 percent. The TRA also had the effect of lowering the maximum marginal tax rate on dividends to 38.5 percent. These changes became effective January 1, 1987. (1)

However, it is important to note, as Miller (1986) points out, that capital gains can be deferred indefinitely. Income does not have to be realized until the sale of the asset. Taxes on capital gains can also be offset by capital losses, reducing or possibly eliminating tax liabilities. Dividend income, on the other hand, is treated as ordinary income. Although capital gains and dividend income still differed due to the ability to defer capital gains and also offset capital gains with capital losses, the relative relationship between the two forms of income changed with the passage of the 1986 TRA. (2) A consequence of this change should have been a reallocation of investor preference for dividend income versus capital gains.

An additional change made by TRA was to repeal the investment tax credit (ITC). Eliminating the ITC brought the potential for a decrease in new investment. Without the ITC, potential investment projects would be required to generate higher cash flows before being considered feasible. As a result firms would have the ability to increase dividend payout by utilizing funds formerly designated for investment. Managers might recognize their decreased investment opportunities and pay out earnings in the form of dividends instead of retaining them for investment.

This line of reasoning delineates the two following reasons to expect a significant increase in dividend payout ratios subsequent to passage of the 1986 TRA: (1) changes in marginal tax rates, and (2) the elimination of the ITC.

The arguments presented above suggest an expected increase in dividend payout ratios following the passage of the 1986 TRA. The first argument addresses investor demand for higher dividends. The passage of the 1986 TRA could have had the effect of stimulating investor demand for higher dividends. Subsequent market pressures brought about by this increased demand should have forced firms to increase dividend payout ratios.

Opposing theories suggest that dividend payout ratios would remain the same, decrease, or increase insignificantly in response to the 1986 TRA. Miller and Scholes' (1978) classic study maintains taxes are irrelevant in determining dividend policy because various legal provisions exist that provide vehicles to avoid taxation on dividend income.

An alternative viewpoint relating to elimination of the ITC suggests that managers lower their dividend payout to provide additional funds for investment. Corporations might reduce dividend payouts, retaining a higher percentage of earnings to maintain current levels of investment, due to this change in tax laws. Cash flow has a strong positive influence on investment spending as documented by Vogt (1994) and Whited (1992).

The previous paragraphs delineate the following two reasons not to expect increases in dividend payout policy subsequent to the passage of the 1986 TRA: (1) the existence of tax avoidance vehicles, and (2) the elimination of the ITC.

Several studies attempted to address the question of whether firms adjusted dividend payout ratios subsequent to the passage of the 1986 TRA. Ben-Horim, Hochman, and Palmon (1987) and Abrutyn and Turner (1990) both focused on anticipated firm dividend policy response to the passage of the 1986 TRA. Other studies such as Bolster and Janjigian (1991), Means, Charoenwong, and Kang (1992), and Papaioannou and Savarese (1994) empirically tested for change in dividend payouts subsequent to passage of the 1986 TRA and found conflicting results.

A firm's choice of dividend policy is important if it affects shareholder wealth. Most studies concentrate on the shareholder wealth effects of tax law changes. A question remains as to whether firm management believed the changes caused by the 1986 TRA affected shareholder wealth and reacted to the 1986 TRA by changing dividend policy. This study focuses on whether management altered dividend policy as a result of the 1986 TRA. The choice of methodology allows for the possibility of these changes being spread over a several year period.

This paper is organized in the following manner. Sections two and three contain a review of the relevant literature. Section four includes a discussion of the data and methodology. Section five contains a presentation of the results, and the final section contains the conclusions.

LITERATURE REVIEW ON DIVIDENDS AND TAXES

Miller and Modigliani's (1961) proof of dividend irrelevance in the presence of perfect capital markets kindled a debate that continues today. If dividends have no impact on firm value, then the dividend decision is moot. If, however, dividends affect firm value, then management can establish an optimal payout policy. Much of the research, therefore, has focused on the question of dividend relevance in the presence of market imperfections such as taxes. Walter (1963) maintains that dividends matter because of the differential tax treatment of income. Elton and Gruber (1970), Pettit (1977), and Barclay (1987) all support the existence of tax clienteles.

Farrar and Selwyn (1967), Brennan (1970), Modigliani (1982), and Poterba and Summers (1984) maintain that payment of cash dividends should reduce firm value because of the tax disadvantage stockholders face when receiving dividends. Litzenberger and Ramaswamy (1979 and 1980) concluded that investors demand higher returns on dividend paying stocks to compensate for holding a tax disadvantaged stream of cash flows. Litzenberger and Ramaswamy (1982) found a positive relationship between stock returns and dividend yields, but concluded that it is indeterminate whether these results are attributable to taxes or some omitted variable. In contrast, Chen, Grundy, and Stambaugh (1990) found no tax penalty on cash dividends.

Other researchers such as Miller and Scholes (1978), Black and Scholes (1974), and Miller (1986) have contended that dividends are irrelevant because certain provisions in the U.S. tax code allow sheltering dividend income. DeAngelo and Masulis (1980) maintain that no dividends should be demanded or paid without the existence of these tax shelters. However, Feenberg (1981) and Feldstein and Green (1983) indicated that few people have the ability to shelter dividends, and Chaplinsky and Seyhun (1990) noted that many investors do not use the shelters that are available.

Long (1978), Poterba (1986), and Sterk and Vandenberg (1990) found modest support for a preference for cash dividends over capital gains consistent with Gordon's (1959) "bird-in-hand" theory. Most research contends that this "preference" results from the signaling value of dividends especially in relation to changes in cash flows [see Bhattacharya (1979) and Miller and Scholes (1982)]. Many studies, such as Kalay (1982), Lakonishok and Vermaelen (1983 and 1986), and Michaely (1991), have focused on ex-dividend day stock returns and credited all or a portion of the price reaction to short-term traders eliminating profits.

For dividend payout ratios to adjust to the passage of the 1986 TRA, management must treat dividend payouts as an active policy variable. Support for this argument comes from the findings of Baker and Farrelly (1988), Farrelly and Baker (1989), Baker, Farrelly, and Edelman (1985), and Baker (1989). Management appears to favor dividend stability and gradual changes consistent with Lintner's (1956) partial adjustment results.

A number of studies approach dividend relevance from the Miller and Rock (1985) perspective that describes dividends and investment as being "two sides of the same coin." Masulis and Trueman (1988), Chang and Rhee (1990), and Jensen, Solberg, and Zorn (1992) all document a link between the investment and dividend decisions. To date, no consensus exists regarding the relevance of dividends in the presence of taxes.

LITERATURE REVIEW ON DIVIDENDS AND THE 1986 TRA

Several studies have found that the passage of TRA affected firms in some manner. (3) Of particular relevance is a study by Jang (1994) that uses event study methodology and finds significantly positive (negative) abnormal returns for high-yield (low-yield) stocks as a result of TRA's enactment. Several of these studies indicated that reaction to the 1986 TRA differed by industry.

Some of the relevant research has attempted to address the question of whether firms adjusted dividend payout ratios subsequent to the passage of TRA. Ben-Horim, Hochman, and Palmon (1987) analyzed five major tax groups of security holders and their dividend policy preferences and concluded that "the reform is likely to induce firms to raise their dividend payout ratios." However, this finding focused on the tax reform's likely impact on corporate financial policy and based results on data from 1986 and earlier.

Abrutyn and Turner (1990), showed that 85 percent of CEOs surveyed expected no change in the dividend payout ratios as a result of TRA. Only 11 percent of CEOs indicated their firms would pay a higher percentage of after-tax profits in the form of dividends. However, the time frame of the survey, early in 1988, coincided with the first year that the tax changes were completed. Actual responses could be very different if investor demand appears to warrant a change in dividend payout ratios.

Bolster and Janjigian (1991) examined shareholder wealth effects and dividend policy changes. High dividend yield stocks significantly outperformed low dividend yield stocks immediately surrounding TRA's passage. Anticipation of passage seems to have caused much of the price adjustment to occur prior to formal passage. With respect to dividends, however, no evidence indicated that dividend payouts increased. Bolster and Janjigian defined dividend payout ratios as total dividends divided by total after tax earnings and examined aggregate dividend payouts, focusing only on the mean and median aggregate dividend payments for 883 firms. Their findings indicate that dividends increase monotonically throughout the time period, 1984-1989, with no significant change after passage. While their findings indicate that aggregate dividend payments remained constant in 1987, their analysis is not industry specific. They concluded that the existence of tax clienteles does not appear to impact significantly corporate dividend decisions.

In a study of dividend policy, Means, Charoenwong, and Kang (1992) focused on changes in dividend yield patterns over the period 1984-1988. Calculation of yearly dividend yield involved summation of monthly dividend yields divided by summation of monthly total yields, with all data being collected from the CRSP tapes. Total yields, from the CRSP tapes, include capital gains yields and dividend yields. The authors expected TRA to cause increasing dividend yields and decreasing capital gains yields. They divided the sample of firms into the following four groups based on percentage of dividend yield: (1) no dividend, (2) dividend yield less than 3 percent per year, (3) dividend yield between 3 percent and 5 percent, and (4) dividend yield greater than 5 percent. Their findings indicated that dividend yields had a downward trend over the period 1984-1986. Following TRA's passage, dividend yields started trending upward. (4) The authors did not notice any significant changes between 1987 and 1988 and concluded that firms had fully adjusted to the tax changes.

More recently, Papaioannou and Savarese (1994) conducted a study of 243 industrial firms and 40 utilities. The authors divided firms into quintiles according to their average dividend payouts during the fourteen quarters prior to TRA's passage. Matched pairs t-tests, conducted on each quintile, indicated that the three quintiles with the lowest dividend payout ratios increased dividend payouts significantly following the passage of TRA. The fourth quintile results were insignificant while the firms with the highest dividend payouts significantly decreased dividend payout ratios. The conflicting evidence presented in these sections indicates further research is necessary.

DATA AND METHODOLOGY

The sample used in the current study consists of 470 observations; one set of pre-1986 and one set of post-1986 observations for 235 firms. The number of firms with stable betas is 115 while the remaining 120 firms have unstable betas. The sample excludes firms in industries subject to significant regulation such as banking, insurance, air transportation, railroad, transportation, electric and gas utilities, financial services, and telecommunications. The sample also excludes: (1) foreign firms, (2) firms not paying dividends throughout the entire period (1982-1992), (3) firms experiencing negative cash flows during any one year of the study period, and (4) firms in the petroleum industry. (5) The study includes only firms that have positive earnings in every year and have data available in Value Line Investment Surveys during the entire time period of 1982 through 1992.

The previous research related to TRA extends, at best, back to the third quarter of 1983. The present study extends the analysis to include additional years of data, specifically 1982 through 1986 and 1988 through 1992. The last major change in the U.S. tax laws, prior to 1986, became effective in January 1982. At that time the highest marginal tax rate on dividends dropped from 70 percent to 50 percent. The fact that 1987 was a transition year for changes in the tax laws requires omission of data from 1987 . Other studies of dividend policy and the 1986 TRA, including Papaioannou and Savarese (1994), followed this procedure.

To test the impact of TRA on corporate dividend policy, the present research estimates a modification of Rozeff's (1982) model using data obtained from Value Line Investment Surveys. This revised model incorporates variables that allow testing for possible changes in dividend payout policy subsequent to the passage of the 1986 TRA. The model also introduces other variables to test the impact of investment spending and beta stability on dividend payout policy.

Beta measures the sensitivity of returns for securities compared to general market movements. A stock with a beta of one is expected to mirror the market. A stock with a beta of greater than one is expected to have returns that go up or down proportionately more than those of the market. Alternatively, a stock with a beta of less than one is expected to have its returns go up or down proportionately less than those of the market. According to Value Line's glossary, the beta coefficient is derived from a regression analysis of the relationship between weekly percentage changes in the price of a stock and weekly percentage changes in the New York Stock Exchange Index over a period of five years. In the case of shorter price histories, a smaller time period is used, but two years is the minimum (Lindahl and Wachowicz, 2001).

One basic premise of this study is that firms not changing dividend policy in response to the 1986 TRA exhibit significant changes in representative firm betas. Testing for this change requires regressing a series of twelve annual betas for each individual firm against time and dummy variables to detect for structural shifts. Estimation of the following model permits classification of firms as having either stable or unstable betas.

[Beta.sub.T] = [[beta].sub.0] + [[beta].sub.1]T + [[beta].sub.2]DV + [[beta].sub.3]DV*T

where:

[Beta.sub.T] = Individual firm beta in year T. T = years 1-12, 1981 through 1993 with 1987 omitted. DV = dummy variable coded 0 for years 1-6, and 1 for years 7-12.

Detection of either a change in slope ([[beta].sub.3] [not equal to] 0) or a discontinuity ([[beta].sub.2] [not equal to] 0) classifies a firm as having an unstable beta and it receives a dummy variable coding of zero for the STABLE variable in the following model. Stable beta firms receive a dummy variable coding of one for the variable STABLE.

After classification of firms into stable or unstable categories this study estimates the following modification of Rozeff's model:

Beta Stability Model

[PAY.sub.i] = [[beta].sub.0] - [[beta].sub.1]INS - [[beta].sub.2]GROW1 - [[beta].sub.3]GROW2 - [[beta].sub.4]STABLE + [[beta].sub.5]STOCK - [[beta].sub.6]CSP [+ or -] [[beta].sub.7]DV [+ or -] [[beta].sub.17]INS*DV [+ or -] [[beta].sub.27][GROW1.sup.*]DV [+ or -] [[beta].sub.37][GROW2.sup.*]DV [+ or -] [[beta].sub.47][STABLE.sup.*]DV [+ or -] [[beta].sub.57][STOCK.sup.*]DV [+ or -] [[beta].sub.67][CSP.sup.*]DV + [[epsilon].sub.i]

where:

[PAY.sub.i] = Average dividend payout ratio of firm i over a five year period calculated as a percentage of EPS. The two separate five year periods are 1982 through 1986 and 1988 through 1992.

INS = Percentage of common stock held by insiders at the end of each five year period, 1982 through 1986 and 1988 through 1992.

GROW1 = Realized five year average growth rate of revenues over each five year period, 1982 through 1986 and 1988 through 1992.

GROW2 = Forecasted future five year average growth rate of revenues using Value Line's estimates. Forecast period includes the years 1987 through 1991 for firms in pre-1986 sample and 1993 through 1997 for firms in post-1986 sample.

STABLE = Dummy variable coded 1 for firms classified as having stable betas over the period 1981 through 1993 with the year 1987 omitted, and 0 otherwise.

STOCK = Natural logarithm of number of common stockholders at end of each five year period, 1982 through 1986 and 1988 through 1992.

CSP = The ratio of total capital spending per share over each five year period (1982 through 1986 and 1988 through 1992) divided by total cash flow per share over each five year period. (can exceed 1.0)

DV = Dummy variable coded 1 for the years 1988 through 1992, 0 for the years 1982 through 1986.

[INS.sup.*]DV, [GROW1.sup.*]DV, [GROW2.sup.*]DV, [STABLE.sup.*]DV, [STOCK.sup.*]DV, and [CSP.sup.*]DV = interaction terms between the terms defined above and the dummy variable, DV.

[[epsilon].sub.i] = error term distributed as N(0,[[sigma].sup.2]).

Note that the operational signs indicate the hypothesized direction of impact each variable should have on dividend payout policy. The following discussion addresses Rozeff's justification for inclusion of each of the variables.

According to Rozeff (1982), the higher the percentage of stock held by insiders (INS), the lower the dividend payout ratio. Dividend payment functions as a bonding cost by decreasing the time and effort expended by outside ownership to monitor the corporation. If insiders hold a significant portion of the shares, the demand for higher dividend payout declines. Conversely, if insiders own very little of the firm's stock, then higher dividend payouts function to lower monitoring costs.

Rozeff also expects the previous five year average growth rate of revenues (GROW1) to have a negative impact on dividend payout policy. The rationale of this variable is that the higher the past revenue growth, the higher the past demand for investment funds to support that revenue growth. Rozeff considers the forecasted five year revenue growth (GROW2) to be a proxy for future investment capital needs. The expected relationship is also negative. The higher the forecasted revenue growth, the lower the dividend payout ratio, assuming maintenance of a target capital structure.

Rozeff used beta to proxy a firm's operating and financial leverage. According to Rozeff (1982), a firm with a higher use of total leverage will have a higher beta. Therefore, to lower the cost of external financing, these higher beta firms would choose lower dividend payout ratios. Dividend payout should then be negatively related to beta. This paper investigates whether beta stability is impacted by the dividend decision. A firm not increasing dividend payout ratio should see a greater instability in firm beta subsequent to TRA's passage. STABLE is a dummy variable included to address this issue.

Rozeff hypothesizes the dispersion of ownership, as measured by the total number of stockholders, should also affect dividend payout ratios. A smaller number of stockholders would be able to monitor the firm much more easily than a larger number. Therefore, as the number of stockholders increases, the likelihood of higher dividend payout ratios also increases. Again, the distribution of dividends functions to reduce monitoring costs. The natural log of the number of shareholders (STOCK) corrects for scale effects.

Rozeff (1982) finds all five original variables to be significant in explaining dividend payout. Later studies by Dempsey and Laber (1992) and Dempsey, Laber, and Rozeff (1993) replicate and extend Rozeff (1982) by examining another seven year period. These studies confirm the stability of Rozeff's original five variable model.

An additional variable suggested by the literature includes some measure of investment. Rozeff's two revenue variables, past and future revenue growth, are intended to proxy a firm's investment requirements, but more recent studies, specifically Dempsey, Laber, and Rozeff (1993), indicate that other variables may be more suitable for measuring investment. This study introduces a new variable, capital spending per share (CSP), as a proxy for investment. CSP's calculation sums capital spending per share over each five year period and divides by the sum of cash flows per share over the same five year period. This ratio represents the percentage of total cash flow required for investment needs during each five year time period and can exceed 1.0. As the ratio increases, firms might reduce dividend payout and satisfy investment requirements using internally generated funds first. The variable CSP should be a more accurate measurement of investment needs.

The current study includes a dummy variable representing the impact of the change in tax laws. A dummy variable DV, coded as zero, represents the time period 1982 through 1986, and coded as one, represents the second period, 1988 through 1992.

Calculation of Rozeff's original variable of PAY uses a five year period instead of a seven year period. Data for the variables INS and STOCK come from the end of the two five year periods, 1986 and 1992. Selecting the end of the period maintains consistency with all previous studies employing variations of Rozeff's model. Dempsey and Laber (1992) use alternative two-year and four-year mean payout definitions and found the model to be robust over these shorter time horizons.

RESULTS

Table l contains a presentation of the regression results. Calculation of the Durbin-Watson statistic indicates that serial correlation is not a problem.

Calculation of the correlation matrix and variance inflation factors indicates that this model suffers from multicollinearity. The use of a stepwise regression procedure can correct the problem of multicollinearity by retaining only one variable from a set of highly correlated regressors (Wonnacott and Wonnacott, 1979). Shalagan (1991) followed a similar procedure.

A stepwise regression procedure yields a more parsimonious model and determines which of the independent variables has the greatest impact on dividend payout policy. Stepwise regression is a common procedure in studies when models have large numbers of variables. Woodside and Trappey (1992), Malone, Fries, and Jones (1993), and Bagozzi (1978) all used a similar methodology.

Table 2 contains a presentation of the results using the SAS stepwise procedure with an [alpha] = 0.10. All the variables present in Rozeff's original study have the expected signs. Both models are overall significant at the .0001 level or better. The Durbin-Watson statistic indicated no serial correlation. Visual inspection of the residuals of the reduced model indicated no discernable patterns, but analysis of residuals consistent with White (1980) and Breusch and Pagan (1979) indicated heteroskedasticity. Consequently, column four in Table 2 also contains White's (1980) p-values after the correction for heteroskedasticity.

CONCLUSIONS

The results presented in Table 2 indicate both growth variables are significant in the overall model. In addition, firms with higher historical growth rates increased dividend payouts subsequent to TRA's passage. Possibly firms that had previously retained earnings for growth elected to distribute a higher percentage of earnings in a more favorable tax environment.

The results support the hypothesis that firms not increasing their dividend payouts subsequent to the passage of the 1986 TRA consistently had structurally unstable betas. The firms with structurally stable betas have significantly higher dividend payout ratios, with a coefficient of 0.075, than firms with structurally unstable betas.

The results of the study indicate that TRA's passage affected corporate dividend policy. This study provides additional evidence for the relevance of dividend policy in changing tax environments. Since firms alter dividend payout in response to tax changes, managers potentially should adopt a more active posture regarding dividend decisions. Assuming changes in dividend payout policy result from investor demand pressures, firms adjusting more rapidly to changes in tax treatment of capital gains versus ordinary income might conceivably benefit from rapid adoption of new policies. Temporary clientele imbalances might allow first-movers to increase shareholder wealth by providing a dividend policy more suitable to the new tax environment. The specifics of whether to increase or decrease dividend payout would vary with the direction of the tax change, the magnitude of the change and the growth prospects of the firm.

One of the most interesting findings is that investment (CSP) is positively related to dividend payout. This result supports the theory of dividend distributions functioning to reduce agency costs. Firms with higher levels of investment have higher payout ratios possibly forcing them into external capital markets more frequently to face more stringent monitoring pressures.

Possible public policy implications exist regarding the impact of tax changes on corporate America. These results indicate that lowering taxes on ordinary income relative to capital gains could change the aggregate level of dividends, thus affecting the income subjected to double taxation. It remains to be seen whether overall government receipts from corporate taxes and taxes on dividends increase as a result of this change.

REFERENCES

Abrutyn, S., and R.W. Turner (1990). Taxes and Firms' Dividend Policies: Survey Results. National Tax Journal, 43(4), 491-496.

Bagozzi, R.P. (1978). Salesforce Performance and Satisfaction as a Function of Individual Difference, Interpersonal, and Situational Factors. Journal of Marketing Research, 15(4), 517-531.

Baker, K.H. (1989). Why Companies Pay No Dividends. Akron Business and Economic Review, 20(2), 4861.

Baker, K.H. & G.E. Farrelly (1988). Dividend Achievers: A Behavioral Perspective. Akron Business and Economic Review, 19(1), 79-92.

Baker, K.H., G.E. Farrelly, & R.B. Edelman (1985). A Survey of Management Views on Dividend Policy. Financial Management, 14 (3), 78-84.

Barclay, M.J. (1987). Dividends, Taxes, and Common Stock Prices: The Ex-Dividend Day Behavior of Common Stock Prices Before the Income Tax. Journal of Financial Economics, 19(1), 31-44.

Ben-Horim, M., S. Hochman, & O. Palmon (1987). The Impact of the 1986 Tax Reform Act on Corporate Financial Policy. Financial Management, 16(3), 29-35.

Bhattacharya, S. (1979). Imperfect Information, Dividend Policy, and "the bird in the hand" Fallacy. The Bell Journal of Economics, 10(1), 259-270.

Black, F., & M. Scholes (1974). The Effects of Dividend Yield and Dividend Policy on Common Stock Prices and Returns. Journal of Financial Economics, 3(1), pp. 1-22.

Bolster, P. J., & V. Janjigian (1991). Dividend Policy and Valuation Effects of the Tax Reform Act of 1986. National Tax Journal, 44(4), 511-518.

Brennan, M. J. (1970). Taxes, Market Valuation and Corporate Financial Policy. National Tax Journal, December 23(4), 417-427.

Breusch, T. & A Pagan (1979). A Simple Test for Heteroskedasticity and Random Coefficient Variation. Econometrica, 47(5), 1287-1294.

Chaplinsky, S. & H.N. Seyhun (1990). Dividends and Taxes: Evidence on Tax-Reduction Strategies. The Journal of Business, 63(2), 239-261.

Chang, R.P. & G. Rhee (1990). The Impact of Personal Taxes on Corporate Dividend Policy and Capital Structure Decisions. Financial Management, 19(2), 21-31.

Chen, N., B. Grundy, & R.F. Stambaugh (1990). Changing Risk, Changing Risk Premiums, and Dividend Yield Effects. Journal of Business, January 63(1), S51-S70.

Cox, R.A.K., R.T. Kleiman, & R.G. Stout (1994). Tax Policy and Shareholder Wealth: Some Evidence from the Tax Reform Act of 1986. The Financial Review, 29(2), 249-273.

DeAngelo, H. & R.W. Masulis (1980). Leverage and Dividend Irrelevancy Under Corporate and Personal Taxation. The Journal of Finance, 35(2), 453-467.

Dempsey, S.J., & G. Laber (1992). Effects of Agency and Transaction Costs on Dividend Payout Ratios: Further Evidence of the Agency-Transaction Cost Hypothesis. The Journal of Financial Research, 15(4), 317-321.

Dempsey, S., G. Laber, & M. Rozeff (1993). Dividend Policies in Practice: Is There an Industry Effect? Quarterly Journal of Business and Economics, 32(4), 3-13.

Elton, E.J., & M.J. Gruber (1970). Marginal Stockholder Tax Rates and the Clientele Effect. The Review of Economics and Statistics, 52(1), 68-74.

Farrar, D.E., & L.L. Selwyn (1967). Taxes, Corporate Financial Policy and Return to Investors. National Tax Journal, December 20(4), 445-454.

Farrelly, G.E., & K.H. Baker (1989). Corporate Dividends: Views of Institutional Investors. Akron Business and Economic Review, 20(2), 89-100.

Feenberg, D., (1981). Does the Investment Interest Limitation Explain the Existence of Dividends? Journal of Financial Economics, 9(3), 265-269.

Feldstein, M., & J. Green (1983). Why Do Companies Pay Dividends? The American Economic Review, 73(1), 17-30. Gordon, M. J. (1959). Dividends, Earnings, and Stock Prices. The Review of Economics and Statistics, 41(2), 99-105.

Graddy, D.B., G. Homaifar, & K.W. Hollman (1992). Tax Reform's Impact on Insurance Industry Stock Returns. Journal of Risk and Insurance, 59(2), 284-290.

Grammatikos, T. & P. Yourougou (1990). Market Expectations of the Effects of the Tax Reform Act of 1986 on Banking Organizations. Journal of Banking and Finance, 14(6), 1171-1187.

Han, K.C. (1994). The Effect of the 1986 Tax Reform Act on Ex-Dividend Day Return Behavior. The Journal of Financial Research, 17(2), 175-186.

Hearth, D. & J.N. Rimbey (1993) The Dividend-Clientele Controversy and the Tax Reform Act of 1986. Quarterly Journal of Business and Economics, 32(1), 68-81.

Jang, H.J.(1994). The Market Reaction to the 1986 Tax Reform Overhaul: A Study of the Capital Gain Tax Change. Journal of Business Finance and Accounting, 22(6), 1179-1193.

Kalay, A. (1982). The Ex-Dividend Day Behavior of Stock Prices: A Re-Examination of the Clientele Effect. The Journal of Finance, 37(4), 1059-1070.

Lakonishok, J. & T. Vermaelen (1983). Tax Reform and Ex-Dividend Day Behavior. The Journal of Finance, September 1983, 38(4), 1157-1179.

Lakonishok, J. & T. Vermaelen (1986). Tax-Induced Trading Around Ex-Dividend Days. Journal of Financial Economics, July, 16(3), 287-319.

Lindahl, M. & J. Wachowicz, Jr. (2001). Judging Your Portfolio's Return, Given Its Risk. Review of Business, 22 (1-2), 59-61.

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

Litzenberger, R.H., & K. Ramaswamy (1979). The Effect of Personal Taxes and Dividends on Capital Asset Prices. Journal of Financial Economics, 7(2) 163-195.

Litzenberger, R.H., & K. Ramaswamy (1980). Dividends, Short Selling Restrictions, Tax-Induced Investor Clienteles and Market Equilibrium. The Journal of Finance, 35(2), 469-485.

Litzenberger, R.H., & K. Ramaswamy (1982). The Effects of Dividends on Common Stock Prices; Tax Effects or Information Effects? The Journal of Finance, 37(2), 429-443.

Long, J.B. (1978). The Market Valuation of Cash Dividends, A Case to Consider. Journal of Financial Economics, 6(2,3) 235-264.

Malone, D., C. Fries, & T. Jones (1993). An Empirical Investigation of the Extent of Corporate Financial Disclosure in the Oil and Gas Industry. Journal of Accounting, Auditing, and Finance, 8(3), 249-273.

Masulis, R.W., & B. Trueman (1988). Corporate Investment and Dividend Decisions under Differential Personal Taxation. Journal of Financial and Quantitative Analysis, 23(4), 369-385.

Means, D.B. Jr., C. Charoenwong, & Y. Kang (1992). Changing Dividend Policies Caused by the Tax Reform Act of 1986: An Empirical Analysis. Journal of Economics and Finance, 16(3), 153-160.

Michaely, R. (1991). Ex-Dividend Day Stock Price Behavior: The Case of the 1986 Tax Reform Act. The Journal of Finance, 46(3), 845-859.

Miller, M.H. & R.J. Shiller (1986). Behavioral Rationality in Finance: The Case of Dividends. The Journal of Business, 59(4), S451-S468.

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

Miller, M.H., & K. Rock (1985). Dividend Policy under Asymmetric Information. The Journal of Finance, 40(4), 10311051.

Miller, M.H., & M.S. Scholes (1978). Dividends and Taxes. Journal of Financial Economics, 6(4), 333364.

Miller, M.H., & M.S. Scholes (1982). Dividends and Taxes: Some Empirical Evidence. Journal of Political Economy, 90(6), 1118-1141.

Modigliani, F. (1982). Debt, Dividend Policy, Taxes, Inflation and Market Valuation. The Journal of Finance, 37(2), 255- 273.

Papaioannou, G. J., & C.M. Savarese (1994). Corporate Dividend Policy Response to the Tax Reform Act of 1986. Financial Management, 23(1), 56-63.

Pettit, R.R. (1977). Taxes, Transaction Costs and the Clientele Effect of Dividends. Journal of Financial Economics, 5, 419-436.

Poterba, J.M. (1986). The Market Valuation of Cash Dividends: The Citizens Utilities Case Reconsidered. Journal of Financial Economics, 15(3), 395-405.

Poterba, J.M., & L.H. Summers (1984). New Evidence That Taxes Affect the Valuation of Dividends. The Journal of Finance, 39(5), 1397-1415.

Robin, A.J. (1991). The Impact of the 1986 Tax Reform Act on Ex-Dividend Day Returns. Financial Management, 20(1), 60-70.

Rozeff, M.S. (1982). Growth, Beta, and Agency Costs as Determinants of Dividend Payout Ratios. The Journal of Financial Research, 5(3), 249-259.

Shalagan, S. (1991). In the Eye of the Beholder: Commercial Bank Priorities for Maximizing Shareholder Value. Journal of Business and Economic Studies, 1(1), 33-44.

Siddiqi, M.A. (1995). An Indirect Test For Dividend Relevance. The Journal of Financial Research, 18(1), 89-101.

Skinner, D.L. (1993). Twenty-Five Years of Tax Law Changes and Investor Response. The Journal of Financial Research, 18(1), 61-70.

Sterk, W.E. & P.A. Vandenberg (1990). The Market Valuation of Cash Dividends and the Tax Differential Theory of Dividend Policy: A Case Revisited. The Financial Review, 25(3), 441-455.

Value Line Investment Surveys, vols. 1-39, 1982-1992.

Vogt, S.C. (1994). The Cash Flow/Investment Relationship: Evidence from U.S. Manufacturing Firms. Financial Management, 23(2), 3-20.

Walter, J.E. (1963). Dividend Policy: Its Influence on the Value of the Enterprise. The Journal of Finance, 18, 280-291.

White, H. (1980). A Heteroskedasticity Consistent Covariance Matrix. Econometrica, 48(4), 817-838.

Whited, T. (1992). Debt, Liquidity Constraints, and Corporate Investment: Evidence from Panel Data. Journal of Finance, 47(4), 1425-1460.

Wonnacott, R.J. & T.H. Wonnacott (1979). Econometrics. John Wiley & Sons.

Woodside, A.G.& R.J. Trappey, III (1992). Finding Out Why Customers Shop Your Store and Buy Your Brand: Automatic Cognitive Processing Models of Primary Choice. Journal of Advertising Research, 32(6), 59-78.

ENDNOTES

(1) The Taxpayer Relief Act of 1997 reduced the capital gains tax rates that had been in place since the passage of 1986 Tax Reform Act. When fully phased in, most capital gains on assets held more than five years will be taxed at a maximum rate of 18 percent. Assets held between one and five years will be taxed at a maximum rate of 20 percent.

(2) Certain investors already prefer dividend income in spite of the preferential tax treatment afforded capital gains.

(3) Robin (1991), Cox, Kleiman, and Stout (1994), Graddy, Homaifar, and Hollman (1992), Han (1994), and Siddiqi (1995) all find market reactions subsequent to the passage of TRA. In contrast, Michaely (1991), Hearth and Rimbey (1993), Skinner (1993), and Grammatikos and Yourougou (1990) find no reaction to TRA's passage.

(4) The authors did not indicate why dividend yields trended upward. A decrease in price while maintaining a stable dividend could yield the same upward trend.

(5) Firms in the petroleum industry are subject to different accounting practices which makes EPS not strictly comparable to other industries.

K. Michael Casey, Henderson State University

John B. Duncan, The University of Louisiana at Monroe

Michael P. Watters, Henderson State University

TABLE 1: Regression Results for EPS Dividend Payout on Rozeff's
Original Variables, Capital Spending and Beta Stability

 Prob. >
Variable Parameter Estimate [absolute value of T]

INTERCEPT 0.486038 0.0171
INS -0.000393 0.8058
GROW1 -0.017087 0.0002
GROW2 -0.013358 0.0272
STOCK 0.011582 0.5760
CSP 0.131384 0.0978
STABLE 0.089331 0.0527
DV 0.349929 0.2099
[INS.sup.*] DV -0.000293 0.8938
[GROW1.sup.*] DV -0.014435 0.0258
[GROW2.sup.*] DV -0.000511 0.9603
[STOCK.sup.*] DV -0.011458 0.6797
[CSP.sup.*] DV -0.020087 0.8509
[STABLE.sup.*] DV -0.034759 0.5934

[R.sup.2] = 0.2247
F Value = 10.163 *
Adjusted [R.sup.2] = 0.2026
N = 470

* = significant at the 0.0001 level

TABLE 2
Stepwise Regression Results for EPS Dividend Payout on Rozeff's
Original Variables, Capital Spending and Beta Stability

 Parameter Prob. > White's Prob. >
Variable Estimate [absolute value of T] [absolute value of T]

INTERCEPT 0.603403 0.000100 0.000100
GROW1 -0.017455 0.000100 0.000100
GROW2 -0.013784 0.004500 0.007800
CSP 0.129039 0.011900 0.002600
STABLE 0.075115 0.011900 0.016200
DV 0.204855 0.000300 0.008900
[GROW1.sup.*]
 DV 0.014135 0.013400 0.033700

[R.sup.2] = 0.2225
Adjusted [R.sup.2] = 0.2124
F Value = 22.080 *
N = 470
Durbin-Watson = 1.801

* significant at the 0.0001 level