Printer Friendly

Dividend tax cut and security prices: examining the effect of the Jobs and Growth Tax Relief Reconciliation Act of 2003.

Changes in dividend taxation for individuals under the Jobs and Growth Tax Relief Reconciliation Act of 2003 (JAGTRRA) allow us to test hypotheses motivated by the dividend valuation and policy theory. Using event studies, we find that, consistent with both the tax-clientele and short-term trader hypotheses, high-dividend stocks outperform low-dividend stocks with a reduction in dividend taxation. Consistent with the excess funds hypothesis, we find that firms that were currently paying no dividends and firms with high cash holdings, low debt ratios, and low Tobin's Q were winners under the 2003 dividend tax cut.

**********

In 2003, the taxation of dividends for individuals changed dramatically with the signing of the Jobs and Growth Tax Relief Reconciliation Act of 2003 (JAGTRRA) into law. For example, JAGTRRA lowered the federal individual tax rate on dividends from the top ordinary income tax rate of 38.6% at the time to the long-term capital gains tax rate of 15%. This was the lowest federal tax rate on dividends in almost 90 years. This dramatic change in individual dividend income taxation presents a unique opportunity in the history of dividend taxation to test hypotheses motivated by the dividend valuation and policy theory.

In this study, using key events associated with JAGTRRA, we present new research on some important hypotheses on dividend valuation and dividend policy. Using the abnormal returns for a large sample of firms for the original proposal by President Bush in January 2003 (hereafter referred to as the Proposal) and its final passage and signature into law in May of that year (JAGTRRA), we find new evidence that supports the following notions.

First, we find that firms with higher dividend yields earned higher returns around the time of the Proposal and around the date of the passage of JAGTRRA. We interpret this as evidence that individual investors cannot costlessly avoid dividend taxation, and that they are the marginal investors in dividendpaying stocks after these events. These inferences are consistent with both a conditional case of the tax-clientele hypothesis (Elton and Gruber, 1970) and the short-term trader hypothesis (Kalay, 1982 and Boyd and Jagannathan, 1994), which we call the post-marginal investor hypothesis.

Second, we find that dividend taxation affects the ability of managers to use dividend policy to reduce agency costs between managers and shareholders. We find that non-dividend-paying firms earned positive abnormal returns around the Proposal and passage of JAGTRRA relative to dividend-paying firms. This result implies that investors expect non-dividend paying firms to respond more vigorously to a dividend tax reduction than dividend-paying firms, and is consistent with the excess funds hypothesis (or free cash flow hypothesis) (Easterbrook, 1984, Jensen, 1986, and Lang and Litzenberger, 1989), the outcome model of dividend policy (LaPorta, Florencio, Shleifer, and Vishny, 2000), and the empirical analysis of corporate responses to JAGTRRA (Brown, Liang, and Weisbenner, 2004 and Chetty and Saez, 2005).

Third, we find that the abnormal returns associated with the Proposal and the passage of JAGTRRA are positively related to cash holdings and negatively related to debt ratios and Tobin's Q. We interpret these results as indications that investors expect firms with higher cash, lower debt ratios, and lower Tobin's Q to be more responsive to a dividend tax cut, given that the cost of distributing cash via dividends is lower. These results are also consistent with the excess funds hypothesis. Although other studies have examined the actual corporate responses to JAGTRRA, our study tests the relevant dividend policy hypotheses by using the market reactions to both the initial Proposal and the final passage of JAGTRRA.

The paper is organized as follows. In Section I, we discuss the motivation for the study and the predictors of event market returns associated with the Proposal and the passage of JAGTRRA. Section II describes the sample and variables. In Section III, we present the results of the tests of the post-marginal investor and excess funds hypotheses. Section IV summarizes and concludes.

I. Motivation

In a speech at the Friar's Club on January 7, 2003, President Bush called on Congress to eliminate the double taxation of dividends. Under his proposal, individuals would owe no taxes on dividends, provided that the issuing firm paid sufficient federal taxes at the corporate level. Because individual taxpayers' dividends were subject to ordinary income tax rates, the Proposal represented a potentially dramatic shift in dividends, firm values, and corporate dividend policy for both investors and firms.

On May 28, 2003, President Bush signed JAGTRRA into law. However, although JAGTRRA lowered the marginal tax rates on dividends, it did not include conditions for the federal taxes paid by dividend-issuing firms. In addition, JAGTRRA included a sunset provision under which the individual tax rate on dividends would revert to those of ordinary income, starting January 1, 2009.

A. The Proposal and JAGTRRA as News Events

For the market to react, investors need to be surprised by public events. The amount of press coverage, the headlines from the news articles, and the status of and actions on the bills in Congress that led to JAGTRRA strongly suggest that investors were surprised by both the Proposal and JAGTRRA.

Figure 1 shows the number of news articles per day published in The Wall Street Journal that contained the words "dividend", "tax", and "cut" in the citation or abstract per ABI/Inform, for the period starting December 1, 2002, about one month before the date of the Proposal, and ending June 30, 2003, about one month after Congress enacted JAGTRRA. We note that the daily article count peaks around the date of the Proposal (January 7, 2003), fades over the next few days, rises occasionally in the following months, and then reaches its highest and longest peak around the enactment of JAGTRRA in late May.

[FIGURE 1 OMITTED]

The market reaction to events such as the Proposal and JAGTRRA depends on two factors: whether investors care about the dividend tax rate of individuals, and if so, how they expect firm managers to respond to a dividend tax cut for individuals.

A review of other studies identifies two classes of hypotheses related to dividend taxation. These hypotheses make two predictions about market responses to the Proposal and JAGTRRA: dividend valuation that assumes that corporate dividend policy is static, and dividend valuation that allows dynamic corporate dividend policy. Absent a change in dividend payouts, the first class of hypotheses focuses on the marginal investor of stocks that already pay dividends. The second class of hypotheses focuses on corporate dividend policy. Because corporations are likely to change their dividends only if investors care about dividend taxation, we consider the first class of hypotheses first.

B. The Marginal Investor and Individual Dividend Taxation

The tax effect of dividend distribution is a well-researched subject in finance (see Kalay and Michaely, 2000). Two lines of theory provide motivation for the marginal investor to care about dividend taxation. According to the tax-clientele hypothesis (Elton and Gruber, 1970), investors select their stock holdings to minimize the tax bite of dividends. It follows that a high-dividend-tax-rate investor would avoid holding dividend-paying stocks, while a low/zero-dividend-tax-rate investor would prefer doing so. Koski and Scruggs (1998) find no empirical support for the tax-clientele hypothesis. However, Elton and Gruber (1970) and Litzenberger and Ramaswamy (1979), among others, report that returns on ex-dividend days are consistent with high-tax-rate individuals avoiding holding dividend-paying stocks.

A sufficient reduction in their dividend tax rates could make dividend stocks attractive enough to individuals such that they would become the new marginal investors in dividend stocks. Such a dividend tax cut, if unexpected, would cause dividend stocks to earn positive abnormal returns. The intuition behind this prediction is that the new marginal investors value dividend stocks more than the old marginal investors, resulting in an unexpected rise in the market prices of dividend stocks. Because it is conditioned on the tax clientele of dividend-paying stocks changing with the dividend tax cut, we call this case the "dynamic tax-clientele hypothesis."

According to the short-term trader hypothesis (Kalay, 1982 and Boyd and Jagannathan, 1994), individuals hold dividend stocks, but avoid the excess tax bite of dividends by trading with other market participants around the cure-dividend date. Lakonishok and Vermaelen (1986), Karpoff and Walkling (1988), and Michaely (1991), among others, provide evidence supporting the short-term trader hypothesis. If individual investors use short-term trading around ex-dividend days to manage their exposure to the excess taxation of dividends, they are likely to face transactions costs when they trade to avoid dividend taxation. Furthermore, if dividend tax rates are lowered to the capital gains tax rate, individuals no longer have the motivation to trade to avoid dividends. By saving the cost of trading around the ex-dividend day, individual investors will value the dividend-paying stock by the trading costs saved. Boyd and Jagannathan (1994) assert that when dividend taxation is higher than capital gains taxation, investors are more likely to trade high dividend-paying stocks than low dividend-paying stocks. It follows that a reduction in dividend taxation will benefit higher-dividend stocks more than low-dividend stocks.

When investors face costs for avoiding dividend taxation and become marginal investors after a dividend tax change, then news of a reduction in the dividend tax rate should be accompanied by a market response. These conditions are satisfied under both the dynamic tax-clientele hypothesis and the short-term trading hypothesis. Because these common conditions are not specific to any hypothesis, we call this commonality the "post-marginal investor hypothesis."

Since the post-marginal investor hypothesis implies that the market prices of high dividend-paying firms' stocks will increase because of the expected benefit of lower taxation on extant dividends, ceteris paribus, a decrease in the dividend tax rate has a larger percentage effect on value as the dividend-yield of a stock increases. Holding expected dividends constant, we call this positive dividend-yield effect due to a reduction in dividend taxation the "static tax-rate effect."

C. How Investors Might Expect Corporations to React to a Dividend Tax Cut

If individual investors care about dividend taxation, they would also be concerned with whether managers are likely to revise dividend payouts when the dividend tax rates fall. In fact, encouraging public corporate managers to pay out more cash to shareholders was a principal reason mentioned by the then chairman of the President's Council of Economic Advisors in promoting support for a dividend tax cut (Financial Times, 2003). It follows that investor expectations about corporate responses to a dividend tax cut would be reflected by market prices when the Proposal and the passage of JAGTRRA were announced. Because these market responses would be due to the dynamic tax-rate effect, they would be distinct from the static tax-rate effect.

The excess funds hypothesis posits that the retention of earnings by a company poses agency problems because it is easy for managers to waste these resources. High dividend taxation encourages firms to retain earnings and increases the chance of managerial waste. With a reduction in dividend taxation, investors would expect managers to feel pressured to initiate or increase dividends under the new tax regime, thus reducing the agency conflict. Investors would perceive that firms that already pay dividends have less agency conflict between shareholders and managers compared with non-dividend-paying firms. Therefore, non-dividend-paying firms are likely to be under greater pressure to start paying dividends than those firms already paying dividends are to increase their dividends when the dividend tax rate decreases.

To make predictions about investors' expected corporate responses to the Proposal and JAGTRRA, we rely primarily on the excess funds hypothesis. Several financial variables may be correlated with excess funds.

First, firms with high cash holdings are likely to have higher excess funds. Second, investors may perceive that firms with higher free cash flow are likely to have more financial capacity to increase dividend payouts. Third, firms with a lower debt ratio could have higher financial capacity to pay cash dividends because they are committed to pay less interest. Fourth, investors are likely to perceive firms with lower Tobin's Q as having higher agency costs of excess funds.

II. The Sample

Our sample consists of common stocks of domestic US firms. We obtain the stock returns data from the Center for Research in Security Price (CRSP) database and the financial statements data from the Compustat database. Following Fama and French (2001), we exclude utilities (SIC code 4900-4949) and financial firms (SIC code 6000-6999) from our sample. We exclude utilities because their dividend policies may be affected by regulation, and financial firms because their financial ratios are not comparable to those of industrial firms. To reduce the chance that missing data might affect our analysis, we also require that a firm has valid CRSP stock returns for each day in the estimation period (255 days starting 45 days prior to January 7, 2003), price and share data as of January 2, 2003, and relevant financial statement data for the most recent quarter or year prior to January 7, 2003. This results in our full sample of 2,948 firms.

Except as noted, our financial variables are from Compustat data for the third calendar quarter of 2002, and the market values are computed using CRSP data as of January 2, 2003. We define the variables as follows: MKTVAL is the market value of common equity in millions of dollars, using CRSP closing price and shares outstanding on January 2, 2003. DIVZERO is a dummy variable equal to one if Compustat reports that the firm did not pay an ordinary dividend, and zero otherwise. DIVYIELD, the dividend yield, is the annual dividend reported by Compustat, divided by stock price per CRSP as of January 2, 2003.

CASH/TA is the ratio of cash and short-term investments to total assets. FREE/TA is the ratio of free cash flow for the trailing 12 months divided by total assets. DEBT/TA is the ratio of long-term debt to total assets. TOBINQ is the ratio of market value of common equity plus total long-term debt plus preferred stock divided by total assets. LOWQ is a dummy variable equal to one for firms with TOB1NQ less than one, and zero otherwise.

Table I provides the descriptive statistics for the firm characteristics. For all variables except as noted, we use the full sample. With DIVYIELD, we use only firms that have DIVYIELD greater than zero (the dividend-stock sample, 660 firms).

The sample has features that are similar to other studies of US firms (e.g., Fama and French, 2001): the range of market equity value is highly skewed, and a majority of firms (77.6%) pay no ordinary dividends. Also, the distribution of cash holdings and Tobin's Q indicates that we can expect numerous firms in the sample to increase payouts due to the reduction in dividend tax rates. The mean (median) of CASH/TA is 0.22 (0.1), and 25% (10%) of the firms have at least 34% (62%) of total assets as cash. The sample is split about evenly between firms that have Tobin's Q roughly less than or greater than one; the median Tobin's Q is 0.963; the mean of LOWQ is 0.52.

To measure the market reactions to these events, we use both the cumulative abnormal returns (CARs) and standardized CARs (SCARs) for return windows that cover two significant events related to the dividend tax cut. For each of these events, we estimate the abnormal returns by using the market model with the value-weighted CRSP index as the market proxy. The market model estimation period is 255 days ending 45 days before January 7, 2003, and the model is estimated using the Scholes and Williams (1977) correction for nonsyncbronous trading. CAR[PROP] is the cumulative abnormal return for the four-trading-day window around the Proposal, January 3--January 8. CAR[PASS] is the CAR for the period in May covering the four-trading-day window that brackets when Congress finalized and passed JAGTRRA and when President Bush signed the bill into law two trading days later, May 22--May 28. SCAR[PROP] and SCAR[PASS] are CAR[PROP] and CAR[PASS] standardized for market model error and serial correlation per Mikkelson and Partch (1988).

III. Analysis of Event Returns

Here, we use the post-marginal investor and excess funds hypotheses to predict market returns for the Proposal and the passage of JAGTRRA.

A. Univariate Analysis

Table II shows the univariate analysis of the Proposal and JAGTRRA event cumulative abnormal returns (CAR[PROP] and CAR[PASS], respectively). Panel A uses correlation analysis and Panel B an analysis of the sorted groups. We perform both types of analysis with the full sample, except as follows: when DIVYIELD is greater than zero, we restrict observations to the dividend-stock sample described in Section II. To construct the groups, we use the firm characteristics to sort the entire sample into three equal-sized groups, except as follows: for the LOWQ dummy variable, the "High" ("Low") portfolio contains firms when the dummy variable is equal to one (zero).

Generally, the results are consistent across both univariate methods and provide considerable support for the hypotheses. Consistent with the post-marginal investor hypothesis, CAR[PROP] and CAR[PASS] are significant and positively correlated with dividend yield among dividendpaying firms (DIVIDEND is greater than zero) and CAR[PROP] (CAR[PASS]) is significantly higher for the high- compared to the low-dividend-yield groups by 1.14% (0.85%). Consistent with the excess funds hypothesis, both event returns are significant and positively correlated with cash holdings (CASH/TA), significant and negatively correlated with debt ratio (DEBT/ TA), and significant and positively correlated with the indicator variable for Tobin's Q less than one (LOWQ). Correspondingly, CAR[PROP] (CAR[PASS]) is significantly higher for the high- compared to the low-cash groups by 2.14% (2.76%), significantly lower for the high- compared to the low-debt-ratio groups by -1.59% (-1.19%), and significantly higher for the LOWQ group by 0.84%. The difference is not significant for CAR[PASS]. CAR[PROP] (CAR[PASS]) is significantly lower for the high- compared to the low-free-cash-flow groups by 1.22% (3.06%).

B. Multivariate Regression Analysis

For our multivariate analysis, we perform linear regressions of standardized CAR[PROP] and CAR[PASS] (SCAR[PROP] and SCAR[PASS], respectively), robust to heteroskedasticity following White (1980), with parametric explanatory variables winsorized at the 2.5% and 97.5% levels.

The results presented in Table III provide support for the post-marginal investor and excess funds hypotheses. We present our findings in five different models. In the first three basic models, we model CASH/TA, FREE/TA, and DEBT/TA with LOWQ as the key explanatory variable. In Model 4, to test if the effect of the CASH/TA variable on stock returns differs between low Tobin's Q and high Tobin's Q firms, we add the CASH/TA*LOWQ interaction variable to Model 1. In Model 5, we add the LMKTVAL*DIVZERO, CASH/TA*DIVZERO, and LOWQ*DIVZERO interaction variables to Model 1 to test if the effects of the LMKTVAL, CASH/TA, and LOWQ variables on stock returns differ between non-dividend-paying and dividend-paying firms.

To test the robustness of our regression models, we examined the Wall Street Journal news stories in the seven-month period around the Proposal and passage of JAGTRRA (December 1, 2002 to June 30, 2003), as well as the Congressional record on bills relevant to JAGTRRA. To test whether our variables have any significance in an out-of-sample period, we looked for a "quiet period" on the prospects of a dividend tax cut. It appears that the most quiet window happened somewhere between Thursday, March 20, 2003 through Tuesday, March 25, 2003. The returns for this trading window show no statistical significance at conventional levels for any of the variables.

Following Lie (2000), we use LMKTVAL as a control variable. The sign of the LMKTVAL variable is negative and significant at the 1% level in all five models for the Proposal, implying a strong Size-in-January effect. Keim (1983) finds that the negative Size-in-January effect occurs predominantly in the first week of trading in January.

Because the DIVYIELD variable is zero for 77.6% of the sample, we include the DIVZERO dummy variable to control for estimation bias for the DIVYIELD coefficient. The sign of the DIVZERO dummy variable is significant and positive for both the Proposal and the passage of JAGTRRA in the first four models and for the Proposal in Model 5. This finding is consistent with the dynamic tax-rate effect. Investors expected that non-dividend-paying firms are more likely to initiate dividend payments than are dividend-paying firms to increase their dividends. (1)

Consistent with the post-marginal investor hypothesis, the coefficient for the DIVYIELD variable is significant and positive for both the Proposal (SCAR[PROP]) and the passage of JAGTRRA (SCAR[PASS]) in all five models. This result indicates that, in both events, high-dividend-paying firms were relative winners compared with low-dividend-paying firms. Studying the effect of the Tax Reform Act of 1986, which also lowered dividend tax rates relative to capital gains tax rates, Jang (1994) also finds that high dividend stocks earned positive abnormal returns around that event.

Consistent with the excess funds hypothesis, the coefficient for CASH/TA is significant and positive for both the Proposal and JAGTRRA returns in the first four models and for the Proposal in Model 5. Lie (2005) finds that firms are likely to increase dividend payouts when they have high cash levels.

In Model 2, the sign of the coefficient for FREE/TA is negative and significant at the 10% level for the Proposal and at the 1% level for the passage of JAGTRRA. Lie (2000) shows that the sign of FREE/TA is significantly negative for firms that announce large special dividends. It is possible that investors expected firms with high free cash flow to respond to the dividend tax cut by issuing a large special dividend rather than by increasing or initiating regular dividends. Because of the sunset provision in JAGTRRA, a large special dividend can be a one-time event. However, if a firm pays regular dividends, since such dividends tend to stick, they are likely to continue after the expiration of the law in 2009.

In Model 3, the significantly negative sign of DEBT/TA at the 1% level for both the Proposal and the passage of JAGTRRA is consistent with the excess funds hypothesis. Because they are committed to pay less interest, firms with lower debt ratios are likely to have higher financial capacity to pay cash dividends. This result is in line with the previous findings by Lintner (1956) and Allen and Michaely (1995). Lie (2005) also finds that firms are likely to increase dividend payouts when they have low debt ratios.

Also consistent with the excess funds hypothesis, the coefficient for LOWQ is significant and positive in Models 1 and 4 for the passage of JAGTRRA, and in Models 4 and 5 for the Proposal. We note that LOWQ is statistically significant only when modeled with CASH/TA. Therefore, we add an interaction variable (CASIt/TA*LOWQ) to Model 1 and present our results in Model 4.

The coefficient of the CASH/TA*LOWQ interaction variable is not significant at the conventional 5% level either for the Proposal or for the passage of JAGTRRA. As noted earlier, the significantly positive sign of the DIVZERO variable indicates that non-dividend-paying firms received a more favorable market reaction after the dividend tax cut. Many of these firms are high Q growth firms. The market expected that growth firms with large cash holdings (e.g., high tech firms) would start paying cash dividends after the reduction in the dividend tax rate. Therefore, the cash holdings of LOWQ firms, many of which might be already paying dividends, were not as important for investors, hence the nonsignificant sign of the interaction variable.

In Model 5, the sign of the LMKTVAL*DIVZERO interaction variable is positive and significant at the 5% level for the passage of JAGTRRA. This finding indicates that large non-dividend-paying firms were winners relative to large dividend-paying firms, and that their stock prices reacted more favorably to the news that a new law might reduce dividend tax rates. This result is consistent with the excess funds hypothesis and with the widespread market expectation that large non-dividend-paying firms were likely to start paying dividends after the passage of the new law.

The sign of the CASH/TA*DIVZERO interaction variable is positive and significant at the 5% level for the passage of JAGTRRA. This result implies that the CASH/TA position of non-dividend-paying firms was more important to investors relative to the CASH/TA position of dividend-paying firms, which made the former relative winners with a more favorable market response to the expected tax rate reduction on dividends. Again, this result is consistent with the excess funds hypothesis and with the widespread market expectation that cash-rich non-dividend-paying firms were likely to start paying dividends after the passage of the new law.

The LOWQ*DIVZERO interaction variable has a positive sign and is significant at about the 5% level for the passage of JAGTRRA and at the 10% level for the Proposal. It indicates that the low Tobin's Q position of non-dividend-paying firms was more important for investors than the low Tobin's Q position of dividend-paying firms. Again, this finding is consistent with the excess funds hypothesis and with the widespread market expectation that non-dividend-paying firms with limited growth opportunities were likely to start paying dividends after the reduction in dividend tax rates.

C. Relative and Economic Significance of Explanatory Variables

To estimate the relative significance of the explanatory variables on the event returns, we multiply the regression coefficients of Models 1, 2, and 3 in Table III by one when they are dummy variables and by the variable range (3rd quartile minus 1st quartile) when they are parametric variables. To estimate the economic significance of these variables, we do similar factor multiplication using the regression coefficients from the same models using non-standardized CARs in place of standardized CARs.

Table IV shows the regression analysis results of the non-standardized CARs. In general, the regression analysis of non-standardized CARs is consistent with that of standardized CARs: the coefficients for DIVZERO, DIVYIELD, CASH/TA, and LOWQ are positive, and those for FREE/TA and DEBT/TA are negative. Additionally, the scale of the coefficients indicates the economic impact of marginal changes in the explanatory variables on the market returns for sample firms. DIVYIELD's Model 1 coefficients of 0.343 and 0.1742 for the Proposal and the passage of JAGTRRA imply that an increase in a firm's dividend yield by 10% resulted in 3.43% and 1.742% higher abnormal returns on these events, respectively. CASH/TA's Model 1 coefficients of 0.0241 and 0.0585 for the Proposal and the passage of JAGTRRA imply that an increase in a firm's cash ratio by 10% resulted in 0.241% and 0.585% higher abnormal returns on these events, respectively. DEBT/TA's Model 2 coefficient of-0.0159 for the Proposal and the passage of JAGTRRA implies that an increase in a firm's debt ratio by 10% resulted in 0.159% lower abnormal returns on these events.

Although these results are informative about the marginal impacts of firm characteristics, analysis using the range of firm characteristics can provide an idea about which firm characteristic had a larger relative or economic impact across the possible range of the variable. We use the results from Models 1 through 3 from Tables III and IV for this analysis, and report the results in Table V.

For the Proposal, results for both the standardized and non-standardized CARs suggest that investors believed that DIVZERO was the most important single attribute among these variables, followed by DIVYIELD and CASH/TA. For example, using Model 1, the relative impact of these three variables was 0.275, 0.168, and 0.102, respectively, while the economic impact was 1.11%, 0.75%, and 0.75%, respectively. By comparison, the relative and economic impact of FREE/TA (Model 2:-0.022 and -0.35%), DEBT/TA (Model 3:-0.06 and -0.43%), and LOWQ (Model 1: 0.056 and 0.4%) are considerably lower.

At the passage of JAGTRRA, the results suggest that investors considered CASH/TA as the most important attribute among these variables (in Model 1, 0.176 and 1.83%), followed by DIVZERO (in Model 1, 0.164 and 1.02%). We note that DIVYIELD was considerably lower at the passage of JAGTRRA (in Model 1, 0.125 and 0.38%) than at the Proposal (in Model 1, 0.168 and 0.75%), but LOWQ was somewhat more significant at the passage of JAGTRRA (in Model 1, 0.075 and 0.84%) than at the Proposal (in Model 1, 0.056 and 0.4%). These results suggest that investors considered the passage of JAGTRRA as more valuable for its potential to increase firm value due to changes in dividend policy than it was for the mere benefit of reduced taxation of dividends from dividend-paying firms.

IV. Conclusion

The dividend tax reform of 2003 provides us with a unique opportunity to test several hypotheses related to dividends. Using the returns of a large sample of US firms around President Bush's original proposal (the Proposal) for dividend tax reform in January 2003 and around the passage of dividend tax reduction per the Jobs and Growth Tax Relief Reconciliation Act of 2003 (JAGTRRA) in May 2003, we find convincing evidence that dividend taxation matters to investors, and that investors expect corporate management to be sensitive to investors' concerns about dividend taxation.

We first consider the tax-clientele hypothesis (Elton and Gruber, 1970) and the short-term trader hypothesis (Kalay, 1982 and Boyd and Jagannathan, 1994). We identify conditions that are sufficient for a dividend tax cut to generate a market response among dividend-paying stocks. Under the joint hypotheses that avoiding dividend taxation is costly for individuals and that individuals are the expected marginal investors in dividend stocks after a tax regime change, credible news about a significant dividend tax cut should generate positive market response for high dividend-yield stocks. Our results support this prediction.

The excess funds hypothesis (Easterbrook, 1984, Jensen, 1986, and Lang and Litzenberger, 1989) suggests that market responses to changes in dividend taxation are predictable, based on measures of financial capacity and agency costs. Consistent with these conjectures, firms with higher cash holdings, lower debt ratios, and lower Tobin's Q were relative winners with the initial Proposal and the final passage of JAGTRRA.

Our findings also indicate that the non-dividend-paying firms had a more positive market response to the expected decrease in the dividend tax rate compared to the dividend-paying firms. Regressions with interaction variables show that larger non-dividend-paying firms and those with high cash and low Tobin's Q levels had a significantly more favorable return response to the expected decrease in the dividend tax rate compared to dividend-paying firms.

Overall, we find strong evidence that dividend tax rates matter to both the market valuation of dividend payouts and to investors' expectations about corporate dividend policy.

We thank Bill Christie (the Editor) and an anonymous referee for their helpful comments and suggestions, and Sandra Sizer for copy editing. We also thank the participants of the EFA meeting in Norfolk and the FMA meeting in Chicago for helpful discussions.

References

Allen, F. and R. Michaely, 1995, "Dividend Policy," in R.A. Jarrow, V. Maksimovic, and W.T. Ziemba, Ed., Handbooks in Operations Research and Management Science: Finance, New York, NY, Elsevier Science.

Boyd, J.H. and R. Jagannathan, 1994, "Ex-Dividend Price Behavior of Common Stocks," Review of Financial Studies 7, 711-741.

Brown, J.R., N. Liang, and S. Weisbenner, 2004, "Executive Financial Incentives and Payout Policy: Firm Responses to the 2003 Dividend Tax Cut," University of Illinois at Urbana-Champaign Working Paper.

Chetty, R. and E. Saez, 2005, "Dividend Taxes and Corporate Behavior: Evidence from the 2003 Dividend Tax Cut," Quarterly Journal of Economics 120, 791-833.

Easterbrook, F.H., 1984, "Two Agency-Cost Explanations of Dividends," American Economic Review 74, 650-659.

Elton, E. and M. Gruber, 1970, "Marginal Shareholders Tax Rates and the Clientele Effect," Review of Economics and Statistics 52, 68-74.

Fama, E. and K. French, 2001, "Disappearing Dividends: Changing Firm Characteristics or Lower Propensity to Pay?" Journal of Financial Economics 60, 3-43.

Financial Times, 2003, "A Tax Cut that will Discipline Corporate America," by G. Hubbard, February 23, 19.

Jang, H.Y.J., 1994, "The Market Reaction to the 1986 Tax Overhaul: A Study of the Capital Gain Tax Change," Journal of Business Finance and Accounting 21, 1179-1193.

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

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

Kalay, A. and R. Michaely, 2000, "Dividends and Taxes: A Re-Examination," Financial Management 29, 55-75.

Karpoff, J. and R. Walkling, 1988, "Short-term Trading Around Ex-Dividend Days: Additional Evidence," Journal of Financial Economics 21, 291-298.

Keim, D.B., 1983, "Size-related Anomalies and Stock Return Seasonality, Further Empirical Evidence," Journal of Financial Economics 12, 13-32.

Koski, J.L. and J.T. Scruggs, 1998, "Who Trades Around the Ex-Dividend Day? Evidence from NYSE Audit File Data," Financial Management 27, 58-72.

Lakonishok, J. and T. Vermaelen, 1986, "Tax Induced Trading Around Ex-Dividend Dates," Journal of Financial Economics 16, 287-319.

Lang, L.H.P. and R. Litzenberger, 1989, "Dividend Announcements: Cash-Flow Signaling vs. Free-Cash Flow Hypothesis?" Journal of Financial Economics 24, 181-191.

LaPorta, R., L.S. Florencio, A. Shleifer, and R. Vishny, 2000, "Agency Problems and Dividend Policies around the World," Journal of Finance 55, 1-33.

Lie, E., 2000, "Excess Funds and Agency Problems: An Empirical Study of Incremental Cash Disbursements," Review of Financial Studies 13, 219-247.

Lie, E., 2005, "Financial Flexibility, Performance, and the Corporate Payout Choice," Journal of Business 78, 2179-2201.

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

Litzenberger, R. and K. Ramaswamy, 1979, "The Effects of Personal Taxes and Dividends on Capital Asset Prices: Theory and Empirical Evidence," Journal of Financial Economics 7, 163-195.

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

Mikkelson, W.H. and M.M. Partch, 1988, "Withdrawn Security Offerings," Journal of Financial and Quantitative Analysis 23, 119-134.

Mikkelson, W.H. and M.M. Partch, 1988, "Errata: Withdrawn Security Offerings," Journal of Financial and Quantitative Analysis 23, 487.

Scholes, M. and J.T. Williams, 1977, "Estimating Betas from Nonsynchronous Data," Journal of Financial Economics 5, 309-327.

White, H., 1980, "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity," Econometrica 48, 817-838.

(1) Chetty and Saez's (2005) ex post study of corporate responses to JAGTRRA finds that the percentage increase in the number of firms with a dividend initiation was greater than the percentage increase in the number of firms with dividend increases from the pre-JAGTRRA period to the post-JAGTRRA period. Brown, Liang, and Weisbenner (2004) report that many of the firms that increased dividends after JAGTRRA scaled back share repurchases, leaving total payouts little changed.

Christopher Gadarowski, Gulser Meric, Carol Welsh, and Ilhan Meric *

* Christopher Gadarowski is an Assistant Professor of Finance at Rowan University, Glassboro, NJ. Gulser Meric is a Professor of Finance at Rowan University, Glassboro, NJ. Carol Welsh is an Associate Professor of Accounting at Rowan University, Glassboro, NJ. Ilhan Meric is the Jesse ILL Harper Professor of Finance at Rider University, Lawrenceville, NJ.
Table I. Descriptive Statistics and Univariate Correlations of the
Sample

This table shows the descriptive statistics and Pearson correlation
coefficients for the firm characteristics of our full sample of
non-utility (SIC code 4900-4949), non-financial (SIC code 6000-6999)
U.S. firms with valid data found using CRSP and Compustat, and the
dividend-stock sample (660 full sample firms with positive dividend
yield). Except as noted, we compute all financial statement variables
from Compustat data for the third calendar quarter of 2002. DIVZERO
is a dummy variable equal to one if the firm pays no ordinary
dividend as of the third calendar quarter of 2002, and zero
otherwise. DIVYIELD is the annual dividend per share divided by the
closing stock price per CRSP as of January 2, 2003. MKTVAL is the
market value of common equity in millions of dollars. LMKTVAL is the
natural log of MKTVAL. CASH/TA is the ratio of cash and short-term
investments to total assets. FREE/TA is the ratio of free cash flow
for the trailing 12 months divided by total assets. DEBT/TA is the
ratio of long-term debt to total assets. TOBINQ is the ratio of
market value of common equity plus total long-term debt plus
preferred stock to total assets. LOWQ is a dummy variable equal to
one if TOBINQ is less than one, and zero otherwise. To reduce the
effect of outliers on correlation analysis, we first winsorize each
variable in the sample at the 2.5% and 97.5% levels within the full
or subsample. The p-values of the correlation coefficients appear
in parentheses.

Descriptive Statistics

 MKTVAL DIVZERO DIVYIELD>0 CASH/TA

N 2948 2948 660 2948
Mean 2,119.0 0.78 0.03 0.218
Std. Dev. 11,648.0 0.42 0.12 0.243
Skewness 13.59 -1.33 24.71 1.23
Kurtosis 230.80 -0.24 626.27 0.51
Maximum 253,512.5 1 3.17 0.991
97.5th Percentile 16,434.7 1 0.08 0.840
90th Percentile 2,940.3 1 0.04 0.623
3rd Quartile 723.2 1 0.03 0.340
Median 165.8 1 0.02 0.113
1st Quartile 40.4 1 0.008 0.027
10th Percentile 11.8 0 0.004 0.010
2.5th Percentile 4.4 0 0.001 0.002
Minimum 0.6 0 0 0

Univariate Correlations

 LMKTVAL DIVZERO DIVYIELD>0 CASH/TA

LMKTVAL
DIVZERO -0.399
 (<.0001)
DIVYIELD>0 0.252
 (<.0001)
CASH/TA -0.157 0.292 -0.235
 (<.0001) (<.0001) (<.0001)
FREE/TA 0.305 -0.204 0.139 -0.403
 (<.0001) (<.0001) (<.0001) (<.0001)
DEBT/TA 0.177 -0.135 0.136 -0.395
 (<.0001) (<.0001) (<.0001) (<.0001)
TOBINQ 0.283 0.031 -0.096 0.270
 (<.0001) (.0918) (<.0001) (<.0001)
LOWQ -0.350 0.041 0.242 -0.166
 (<.0001) (.0245) (<.0001) (<.0001)

Descriptive Statistics

 FREE/TA DEBT/TA TOBINQ LOWQ

N 2948 2948 2948 2948
Mean -0.038 0.164 1.437 0.520
Std. Dev. 0.308 0.213 2.261 0.500
Skewness -8.02 2.77 19.22 -0.08
Kurtosis 120.69 15.56 604.71 -1.99
Maximum 0.720 2.220 82.231 1
97.5th Percentile 0.228 0.645 5.163 1
90th Percentile 0.145 0.425 2.728 1
3rd Quartile 0.084 0.267 1.568 1
Median 0.025 0.094 0.963 1
1st Quartile -0.059 0 0.639 0
10th Percentile -0.277 0 0.426 0
2.5th Percentile -0.760 0 0.249 0
Minimum -6.691 0 0.042 0

Univariate Correlations

 FREE/TA DEBT/TA TOBINQ LOWQ

LMKTVAL
DIVZERO

DIVYIELD>0

CASH/TA

FREE/TA

DEBT/TA 0.076
 (<.0001)
TOBINQ -0.142 -0.104
 (<.0001) (<.0001)
LOWQ 0.009 0.073 -0.65836
 (.6087) (<.0001) (<.0001)

Table II. Univariate Analysis of Event Returns

This table shows the univariate analysis of the cumulative abnormal
returns for the Proposal (CAR[PROP]) and the passage of JAGTRRA
(CAR[PASS]). We use correlation analysis in Panel A and the average
returns of the sorted groups in Panel B. Except as noted, we compute
all financial statement variables from Compustat data for the third
calendar quarter of 2002. DIVYIELD is the annual dividend per share
divided by the closing stock price per CRSP as of January 2, 2003.
MKTVAL is the market value of common equity in millions of dollars.
LMKTVAL is the natural log of MKTVAL. CASH/TA is the ratio of cash
and short-term investments to total assets. FREE/TA is the ratio of
free cash flow for the trailing 12 months divided by total assets.
DEBT/TA is the ratio of long-term debt to total assets. LOWQ is a
dummy variable equal to one if TOBINQ, the ratio of market value of
common equity plus total long-term debt plus preferred stock to total
assets, is less than one, and zero otherwise. We perform our analysis
using the full sample (2,948 firms) except as follows: with DIVYIELD,
we use only firms with DIVYIELD greater than zero (660 firms). To
construct the groups, we sort the firms into three equal-sized groups
using the firm characteristics with the entire sample except as
follows: for dummy variable LOWQ, the "High" ("Low") portfolio
contains firms when the dummy variable is equal to one (zero). To
reduce the effect of outliers, we winsorize CAR[PROP] and CAR[PASS]
at the 2.5% and 97.5% levels within each sample used in the analysis.
The table also shows the differences in the means between the high
and low group averages (High-Low) along with the p-values (Panel A)
and the t-statistics (Panel B) for the two-tailed tests.

Panel A: Univariate Correlations

 DIVYIELD CASH/TA FREEITA

CAR[PROP] 0.180 0.096 -0.121
 (<.0001) (<.0001) (<.0001)
CAR[PASS] 0.110 0.201 -0.206
 (<.0001) (<.0001) (<.0001)

Panel B. Sorted Groups

CAR[PROP]

Groups DIVYIELD CASH/TA FREE/TA

High -0.0030 0.0134 -0.0011
Medium -0.0167 -0.0019 -0.0063
Low -0.0144 -0.0079 0.0111
High-Low 0.0114 *** 0.0214 *** -0.0122 ***
t-statistic 2.82 6.69 -3.64

CAR[PASS]

Groups CASH/TA FREE/TA

High 0.0091 0.0369 0.0080
Medium 0.0063 0.0125 0.0120
Low 0.0006 0.0093 0.0387
High-Low 0.0085 ** 0.0276 *** -0.0306 ***
t-statistic 2.21 8.91 -9.43

Panel A: Univariate Correlations

 DEBT/TA LOWQ

CAR[PROP] -0.061 0.054
 (<.0001) -0.0035
CAR[PASS] -0.055 0.040
 (0.0028) -0.0285

Panel B. Sorted Groups

CAR[PROP]

Groups DEBT/TA LOWQ FEDTAX/PI

High -0.0052 0.0052 -0.0049
Medium -0.0019 -0.0033
Low 0.0107 -0.0032 0.0087
High-Low -0.0159 *** 0.0084 *** -0.0136 ***
t-statistic -5.03 3.27 -5.43

CAR[PASS]

Groups DEBT/TA LOWQ FEDTAX/PI

High 0.0134 0.0221 0.0119
Medium 0.0200 0.0128
Low 0.0254 0.0168 ** 0.0296
High-Low -0.0119 *** 0.0053 ** -0.0177 ***
t-statistic -3.90 2.11 -3.66

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

Table III. Regression Analysis of Standardized Abnormal Returns

This table shows the regression results from the following model:

[SCAR[...].sub.i] = [c.sub.0] + [c.sub.1] [LMKTVAL.sub.i] +
[c.sub.2] [DIVZERO.sub.i] + [c.sub.3] [DIVYIELD.sub.i] + [c.sub.4]
[CASH/TA.sub.i] (or [FREE/TA.sub.i] or [DEBT/TA.sub.i]) + [c.sub.5]
[LOWQ.sub.i] + [c.sub.6] [CASH/TA.sub.i] * [LOWQ.sub.i] (or
[LMKTVAL.sub.i] * [DIVZERO.sub.i] or [CASH/TA.sub.i] *
[DIVZERO.sub.i] or [LOWQ.sub.i] * [DIVZERO.sub.i]) +
[[epsilon].sub.i],

where ... is either PROP or PASS. SCAR[PROP] is the standardized
cumulative abnormal return (CAR) for the four-day return period
associated with the Proposal: January 3-January 8, 2003. SCAR[PASS]
is the standardized CAR for the four-day return period associated
with the passage of JAGTRRA: May 22-May 28, 2003. Except as noted,
we compute all financial statement variables from Compustat data for
the third calendar quarter of 2002. DIVZERO is a dummy variable equal
to one if the firm pays no ordinary dividend as of the third calendar
quarter of 2002, and zero otherwise. DIVYIELD is the annual dividend
per share divided by the closing stock price per CRSP as of January
2, 2003. MKTVAL is the market value of common equity in millions of
dollars. LMKTVAL is the natural log of MKTVAL. CASH/TA is the ratio
of cash and short-term investments to total assets. FREE/TA is the
ratio of free cash flow for the trailing 12 months divided by total
assets. DEBTITA is the ratio of long-term debt to total assets. LOWQ
is a dummy variable equal to one if TOBINQ, the ratio of market value
of common equity plus total long-term debt plus preferred stock to
total assets, is less than one, and zero otherwise. For all models,
we use the full sample (2,948 firms). To reduce the effect of
outliers, we winsorize each variable at the 2.5% and 97.5% levels.
(T-statistics are reported in parentheses.)

 Proposal

Regressor Model 1 Model 2 Model 3

Intercept -0.1315 -0.0671 -0.0294
 (-0.13) (-0.64) (-0.28)
LMKTVAL -7.0000 *** -0.1506 *** -0.1518 ***
 (-5.24) (-6.95) (-7.02)
DIVZERO 0.3374 *** 0.3874 *** 0.3984 ***
 (4.54) (5.24) (5.43)
DIVYIELD 0.0964 *** 0.0970 *** 0.1023 ***
 (3.23) (3.23) (3.42)
CASH/TA 0.0954 ***
 -(4.96)

FREE/TA -0.0367 *
 (0.90)
DEBT/TA -0.0488 **
 (-2.54)
LOWQ 0.0686 * 0.0346 0.0393
 (1.74) (0.90) (1.01)
CASH/TA * LOWQ

LMKTVAL * DIVZERO

CASH/TA * DIVZERO

LOWQ * DIVZERO

Adj. [R.sup.2] 0.0590 0.0523 0.0533
F-value 37.95 *** 35.50 *** 34.20 ***

 Proposal

Regressor Model 4 Model 5

Intercept -0.1563 -0.1466
 (-1.42) (-1.42)
LMKTVAL -0.3606 *** -0.0610 ***
 (-7.00) (-3.60)
DIVZERO 0.3287 *** 0.3298 **
 (4.43) (2.03)
DIVYIELD 0.0935 *** 6.5023 ***
 (3.14) (2.74)
CASH/TA 0.1253 *** 0.5758 **
 (4.71) (2.37)
FREE/TA

DEBT/TA

LOWQ 0.1261 ** 0.1584 **
 (2.45) (2.46)
CASH/TA * LOWQ -0.0628 *
 (-1.74)
LMKTVAL * DIVZERO 0.0021
 (0.11)
CASH/TA * DIVZERO -0.2718
 (-1.08)
LOWQ * DIVZERO -0.1285 *
 (-1.73)
Adj. [R.sup.2] 0.0596 0.0592
F-value 32.15 *** 24.20 ***

 Passage of JAGTRRA

Regressor Model 1 Model 2 Model 3

Intercept -0.1228 -0.067 0.0617
 (-1.15) (-0.63) (0.58)
LMKTVAL -0.0021 0.0191 -0.0164
 (-0.10) (0.88) (-0.76)
DIVZERO 0.2173 *** 0.0154 *** 0.3303 ***
 (2.94) (3.96) (4.54)
DIVYIELD 0.0769 ** 0.0752 ** 0.0845 ***
 (2.48) (2.44) (2.74)
CASH/TA 0.1783 ***
 (8.80)
FREE/TA -0.1456 ***
 (-6.95)
DEBT/TA -0.0486 *
 (-2.54)
LOWQ 0.0999** 0.0559 0.0333
 (2.44) (1.39) (0.81)
CASH/TA * LOWQ

LMKTVAL * DIVZERO

CASH/TA * DIVZERO

LOWQ * DIVZERO

Adj. [R.sup.2] 0.0367 0.0278 0.0113
F-value 23.45 *** 17.87 *** 7.71 ***

 Passage of JAGTRRA

Regressor Model 4 Model 5

Intercept -0.1356 0.1848
 (-1.26) (1.27)
LMKTVAL -0.0020 -0.0290 *
 (-1.10) (-1.82)
DIVZERO 0.2128 *** -0.1749
 (2.87) (-1.11)
DIVYIELD 0.0745 ** 5.8084 **
 (2.42) (2.39)
CASH/TA 0.1935 *** 0.1265
 (7.30) (0.54)
FREE/TA

DEBT/TA

LOWQ 0.1294** -0.0352
 (2.45) (-0.54)
CASH/TA * LOWQ -0.0321
 (-0.08)
LMKTVAL * DIVZERO 0.0361 **
 (1.97)
CASH/TA * DIVZERO 0.4697 *
 (1.93)
LOWQ * DIVZERO 0.1407 *
 (1.90)
Adj. [R.sup.2] 0.0366 0.0380
F-value 19.67 *** 15.55 ***

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table IV. Regression Analysis of Nonstandardized Abnormal Returns

This table shows the regression results from the following model:

[CAR[...].sub.i] = [c.sub.0] + [c.sub.1] [LMKTVAL.sub.i] +
[c.sub.2] [DIVZERO.sub.i] + [c.sub.3] [DIVYIELD.sub.i] + [c.sub.4]
[CASH/TA.sub.i] (or [FREE/TA.sub.i] or [DEBT/TA.sub.i]) + [c.sub.5]
[LOWQ.sub.i] + [[epsilon].sub.i],

where ... is either PROP or PASS. CAR[PROP] is the cumulative
abnormal return (CAR) for the four-day return period associated
with the Proposal: January 3-January 8, 2003. CAR[PASS] is the CAR
for the four-day return period associated with the passage of
JAGTRRA: May 22-May 28, 2003. Except as noted, we compute all
financial statement variables from Compustat data for the third
calendar quarter of 2002. LMKTVAL is the natural log of the market
value of common equity in millions of dollars. DIVZERO is a dummy
variable equal to one if the firm pays no ordinary dividend as of
the third calendar quarter of 2002, and zero otherwise. DIVYIELD is
the annual dividend per share divided by the closing stock price
per CRSP as of January 2, 2003. CASH/TA is the ratio of cash and
short-term investments to total assets. FREE/TA is the ratio of
free cash flow for the trailing 12 months divided by total assets.
DEBT/TA is the ratio of long-term debt to total assets. LOWQ is a
dummy variable equal to one if TOBINQ, the ratio of market value of
common equity plus total long-term debt plus preferred stock to
total assets, is less than one, and zero otherwise. For all models,
we use the full sample (2,948 firms). To reduce the effect of
outliers, we winsorize each variable winsorize at the 2.5% and
97.5% levels. For each regressor, the table provides the estimated
coefficients, followed by the t-statistic in parentheses. For each
model, the table shows the sample size, adjusted R-squared, and the
F-statistic.

 Proposal

Regressor Model 1 Model 2 Model 3

Intercept 0.0044 0.0061 0.0105
 (0.66) (0.96) (1.64)
LMKTVAL -0.0040 *** -0.0036 *** -0.0041 ***
 (-5.24) (-4.80) (-5.29)
DIVZERO 0.0111 *** 0.0134 *** 0.0148 ***
 (3.13) (3.82) (4.25)
DIVYIELD 0.3430 ** 0.3372 ** 0.3770 ***
 (2.55) (2.48) (2.79)
CASH/TA 0.0241 ***
 (3.87)
FREE/TA -0.0245 ***
 (-2.86)
DEBT/TA -0.0159 **
 (-2.01)
LOWQ 0.0040 0.0026 0.0022
 (1.41) (0.95) (0.77)
Adj. [R.sup.2] 0.0297 0.0281 0.0254
F-Value 19.04 *** 18.03 *** 16.36 ***

 Passage of JAGTRRA

Regressor Model 1 Model 2 Model 3

Intercept -0.0022 0.0007 0.0121 *
 (-0.35) (0.12) (1.93)
LMKTVAL -0.0007 0.0003 -0.0014 *
 (-1.05) (0.49) (-1.92)
DIVZERO 0.0102 *** 0.0154 *** 0.0190 ***
 (3.02) (4.65) (5.79)
DIVYIELD 0.1742 0.1554 0.2240 *
 (1.32) (1.19) (1.71)
CASH/TA 0.0585 ***
 (9.31)
FREE/TA -0.0663 ***
 (-7.48)
DEBT/TA -0.0159 **
 (-2.09)
LOWQ 0.0084 *** 0.0054 ** 0.0029
 (3.10) (2.03) (1.04)
Adj. [R.sup.2] 0.0500 0.0464 0.0157
F-Value 32.00 *** 29.69 *** 10.41 ***

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table V. Analysis of Explanatory Variable Effects on Standardized
and Nonstandardized Abnormal Returns

This table shows estimates of the relative effects of explanatory
variables on event returns based on the regression analysis of
standardized and nonstandardized cumulative abnormal returns (CARs)
around the Proposal and the passage of JAGTRRA. We compute the
relative effects of each variable, except the intercept, by using
the product of the range of the variable times its regression
coefficient in Models 1 through 3 for standardized CARs (SCAR: Table
III) or CARs (CAR: Table IV). The range is one if a dummy variable,
otherwise it is the difference between the 3rd and 1st quartile
of the variable's distribution. Except as noted, we compute all
financial statement variables from Compustat data for the third
calendar quarter of 2002. LMKTVAL is the natural log of the market
value of common equity in millions of dollars. DIVZERO is a dummy
variable equal to one if the firm pays no ordinary dividend as of
the third calendar quarter of 2002, and zero otherwise. DIVYIELD is
the annual dividend per share divided by the closing stock price
per CRSP as of January 2, 2003. CASH/TA is the ratio of cash and
short-term investments to total assets. FREE/TA is the ratio of
free cash flow for the trailing 12 months divided by total assets.
DEBT/TA is the ratio of long-term debt to total assets. LOWQ is a
dummy variable equal to one if TOBINQ, the ratio of market value
of common equity plus total long-term debt plus preferred stock to
total assets, is less than one, and zero otherwise.

 Range Times Coefficient

 Passage
 Proposal of JAGTRRA

 Range SCAR CAR SCAR CAR

Model 1
LMKTVAL 1.25 -0.075 -0.50% -0.001 -0.09%
DIVZERO 1 0.275 1.11% 0.164 1.02%
DIVYIELD 0.022 0.168 0.75% 0.125 0.38%
CASH/TA 0.313 0.102 0.75% 0.176 1.83%
LOWQ 1 0.056 0.40% 0.075 0.84%

Model 2
LMKTVAL 1.25 0.076 -0.45% 0.009 0.04%
DIVZERO 1 0.315 1.34% 0.219 1.54%
DIVYIELD 0.022 0.169 0.74% 0.122 0.34%
FREE/TA 0.143 -0.022 -0.35% -0.079 -0.95%
LOWQ 1 0.028 0.26% 0.042 0.54%

Model 3
LMKTVAL 1.25 -0.076 -0.51% 0.008 -0.17%
DIVZERO 1 0.324 1.48% 0.161 1.90%
DIVYIELD 0.022 0.179 0.83% 0.122 0.49%
DEBT/TA 0.267 -0.060 -0.43% -0.055 -0.42%
LOWQ 1 0.032 0.22% 0.025 0.29%
COPYRIGHT 2007 Financial Management Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Gadarowski, Christopher; Meric, Gulser; Welsh, Carol; Meric, Ilhan
Publication:Financial Management
Article Type:Report
Geographic Code:1USA
Date:Dec 22, 2007
Words:8935
Previous Article:Information, incentive alignment, and company loan financing of insider trades.
Next Article:Determinants of idiosyncratic volatility for biotech IPO firms.
Topics:

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters