The tax implications of dividend policy on a firm's equity value continues to be a controversial issue in corporate finance. Although some studies indicate that taxes only weakly affect corporate managers' dividend policy decisions, other evidence suggests that investors react to dividend payouts in line with their effective taxation on dividend income and capital gains. The Tax Reform Act (TRA) of 1986, which lowered the top marginal tax rates for ordinary income and equalized the statutory tax rates for dividends and capital gains, provided an unique opportunity to investigate whether and how managers change dividend policy in response to significant tax changes: If the taxation of dividends and capital gains matters, the changes in the TRA suggest that corporations should increase their dividend payout ratios, except possibly in the case of those firms with already high payout ratios that attract corporate investors whose taxation of dividends increased in the post-TRA period. On the contrary, no systematic changes of dividend payout policies should be expected if corporate managers disregard tax factors, are unsure of the effective (as opposed to die statutory) marginal tax rates of their shareholders, or consider other factors like the information or signaling effects of dividend payouts. Examining, therefore, whether corporations changed their payout policies in accordance with tax-based predictions of dividend policy can provide evidence on the influence of taxes on corporate dividend policy decisions. We examined changes in dividend payouts for a sample of 283 firms drawn from the FORTUNE500 and FORTUNE50 lists of industrial and utility firms for which the relevant data from the 3rd quarter of 1983 to the 1st quarter of 1991 were available. For the aggregate samples of the industrial and utility firms, we found positive changes in payout ratios, but tests of statistical significance on the mean differences of payout ratios failed to reject the null hypothesis of zero mean difference in the dividend payout ratios before and after the TRA. However, when we classified the industrial firms into five quintiles according to their payout ratios in the pre-TRA period, we found strong evidence of significant dividend payout ratio changes. Specifically, the mean differences of payout ratios were positive and significantly different from zero for firms in the first three quintiles (low to medium payout ratio) and negative and significantly different from zero for firms in quintile 5 (very high payout ratio). Of the 49 firms in each of the quintiles 1, 2 and 3, 73, 71 and 61 percent, respectively, had positive payout ratio changes, whereas only 21 percent of the 48 firms in quintile 5 had positive changes. These findings suggest that firms with low to medium payout ratios, which attract ordinary investors with high effective marginal tax rates, increased their dividend payout ratios since dividends would be taxed less in the post-TRA period. On the contrary, firms with very high payout ratios, which attract corporate investors with low effective marginal tax rates for dividends, decreased their payout ratios since the taxation of dividends for such investors would be slightly heavier in the post-TRA period. For additional evidence on the impact of tax changes on dividend policy, we examined whether the relationship of dividends to current earnings shifted in the post-TRA period. The Lintner model suggests that current dividends are determined by past dividends and current earnings. If the TRA increased (decreased) the preference for dividends, then we should find a positive (negative) incremental adjustment of dividends to current earnings. Using industry groups, we found this prediction to be supported in seven of the 19 industry groups. Using firm-specific data, we found evidence of significant incremental adjustment of dividends to current earnings in 23.2 percent of the 276 firms analyzed. We also found that the percentage of firms with positive and significant incremental responses was higher for quintiles 1, 2 and 3 (low to medium payout ratios) and lower for quintiles 4, 5 and the utilities sample (high to very high payout ratios). These findings are indicative of tax-induced changes in dividend payout policies. The evidence from this study is suggestive of several important conclusions concerning corporate dividend policy decisions. First, corporate dividend policies were changed following the tax reform of 1986. Second, firms changed their dividend policies in line with the preferences of their respective dividend clienteles. Both of these conclusions are important for managers who set dividend policy according to tax-clientele considerations. Third, not all firms within each quintile experienced uniform changes in dividend policy. This implies that factors other than taxes may have also influenced dividend policy.
