نرخ مالیات موثر در داخل سهامداران و سود تقسیمی

Using information collected from the Swedish tax authorities, we calculate insiders’ actual effective tax rates on dividends. With this unique dataset, we find a significant negative cross-sectional relationship between insiders’ effective tax rates and dividend payout. This result is consistent with a tax-induced clientele effect for dividends. We also look at the impact of large block trades on dividends. We find that when insiders with zero effective taxes sell blocks, subsequent dividend payments are significantly more likely to decrease. This provides evidence that large shareholders are adjusting dividends for their individual tax situations.

Researchers have extensively examined whether shareholders form clienteles based upon taxed-induced preferences for dividends. Ceteris paribus, the higher the tax rate paid on dividends, the lower the preferred dividend payout. Furthermore, when a controlling shareholder is in a high tax bracket, he may set a relatively low dividend payout. Perez-Gonzales (2003) finds that for firms with large shareholders, changes in tax rates affect firm dividend payout policy. However, Barclay et al. (2007) (henceforth BHS) look at block trades from individuals (who are taxed on dividend income) to corporations (who are not taxed on dividend income) and do not find evidence that corporate blockholders use their power to increase dividends. They conclude that taxes are not an important factor in a corporation’s decision to receive dividend income. We have collected a unique dataset from the Swedish tax authorities of the actual taxes paid by individual large shareholders. Although the tax code in Sweden is not identical to the US tax code, the average tax rate on dividends is lower for corporations than for individuals in both Sweden and the US. Also, like the US tax code, the Swedish tax code allows deductions against dividend income. As a result, for the insiders in our study, the effective tax rates paid on their dividend income varies from 0% to a maximum rate of 30%, with a mean and median of 17% and 30%, respectively. A significant drawback of previous studies examining the relationship between dividend policy and marginal tax rates is that the actual personal tax rate of individual shareholders is generally unknown. As a result, researchers need to make assumptions about individuals’ personal tax rates. For example, since corporations are not taxed on dividend income in the US, BHS assume that all corporations are in a lower tax bracket than individuals. However, individuals are allowed to reduce taxable dividend income through various deductions. Thus there are instances in which the effective tax rate on dividends of individuals is also zero. 1 In the BHS sample, any individual shareholder with an effective tax rate on dividends of zero would be miss-classified. This would make it difficult to find a tax-induced effect on dividends even if one actually existed. Using the actual personal tax rates of large individual investors, we avoid the potential miss-classification problem discussed above. As a result, we are able to conduct more powerful tests of the relationship between dividend payout and taxes. We examine two questions: First, do shareholders form clienteles based upon their tax-induced preferences for dividends?2 Second, do large shareholders receive private benefits of control by using their power to directly influence dividend payout for tax reasons (Shleifer and Vishny, 1986)? We find a significant negative cross-sectional relationship between insiders’ effective tax rates and dividend payout. This result is consistent with a tax-induced clientele effect for dividends. Although the average (and median) insider coalition has effective control with approximately 40% of the votes, this result cannot be used to indicate whether insiders use their power to actively manage dividends. We would find a negative relationship between insiders’ effective tax rates and dividend payment, whether insiders lowered dividend payout to reduce their effective tax rates or instead first set dividend policy for reasons unrelated to taxes and then adjusted their deductions to reduce their effective tax rate.3 To address the question of causality, we perform a second set of tests of the impact on dividends of large block trades. If insiders set dividend policy for personal tax reasons, we expect that when blocks are traded to lower effective tax rate investors, firm dividends should increase in the years following the block trade. However, we find that when block sellers with the highest effective tax rate sell their blocks, there is no significant difference in the subsequent dividend yields, payout or probability of dividend cuts. These results imply that being able to protect dividends from taxes is not sufficient to induce large shareholders to increase dividends. Furthermore, these results are consistent with BHS, who found that when individuals sold blocks to corporations, which they assumed had lower marginal tax rates, dividends were not affected. Alternatively, when blocks are traded to investors who have higher effective tax rates, dividends should decrease. We find that when insiders with effective taxes of zero sell blocks, subsequent dividend payments are significantly more likely to decrease. This provides evidence that large (controlling) shareholders are adjusting dividends to optimize their individual tax situations. Although BHS examine block trades from individuals to non-taxed corporations, they do not examine the reverse: block trades from non-taxed corporations to individuals. In summary, highly taxed controlling shareholders reduce dividends to avoid paying taxes. On the other hand, controlling shareholders with low marginal tax rates do not automatically increase dividends just because they will not pay taxes on their dividends.

Numerous theoretical and empirical papers have examined the relationship between effective tax rates of shareholders and firm dividend policy. The empirical studies have used various proxies for the effective tax rate of investors because personal tax rate information is generally not available. We have collected the actual effective tax rate of shareholders, which allows us to directly test the influence of the effective tax rates of investors on the dividend policy of the firm. We conclude that insiders, who have sufficient power, set firm dividends to reduce their personal tax burden.