The relationship between dividends and taxes continues to be an open empirical question. While some researchers find that taxes have a first-order negative effect on dividends, others find the effect is only minor. In this paper we argue that these conflicting results arise because dividends are also determined by corporate governance, which moderates the effect of taxes. Specifically, lower dividends do not just reduce taxes, but may also increase agency costs by making the free cash flow problem more acute.

Therefore, when regulators increase dividend taxes, firms with serious agency problems are reluctant to cut dividends despite the potential tax savings. We find strong support for the idea that the sensitivity of payout to taxes depends on potential agency conflicts.

We identify the causal effect of taxes on dividends by exploiting a large and clean regulatory shock in Norway in 2006 that increased the dividend tax rate for individuals from 0% to 28%. Because the tax shock is large, any change in dividend policy around the time of the tax reform is likely to be driven by taxes. Because the tax shock is clean, with a flat tax rate both before and after the tax reform, we avoid complications due to multiple tax brackets. Because dividends and capital gains are taxed identically and share repurchases are negligible, we can focus on just cash dividends. 

We examine the population of majority-controlled firms, using proprietary microdata on all family relationships and a million individual tax returns and salary receipts. We focus on the main agency problem for most firms in any economy, which is the conflict of interest between majority and minority shareholders. Consistent with our prediction, we find that the dividend payout ratio drops less after the tax increase the higher the potential conflict between the firm’s shareholders. For instance, the average payout ratio falls by 30 percentage points when the conflict potential is low, but only by 18 percentage points when the conflict potential is high. This result suggests that, because controlling shareholders trade off the tax effect against the conflict effect of dividends, the relationship between taxes and dividends depends on the severity of agency costs.

Firms with severe shareholder conflicts are reluctant to cut dividends. Therefore, such firms may look for ways to mitigate the increased tax burden after the tax shock. While the tax reform raises the tax on dividends paid to individuals, dividends paid to firms continues to be tax-free. This difference in dividend tax rates creates incentives to own shares indirectly through holding companies rather than directly, as such indirect ownership ensures that free cash flow is taken away from the majority shareholder’s control without triggering immediate tax payments. Therefore, we predict that higher dividend taxation for individuals increases the use of indirect ownership, particularly in firms where potential shareholder conflicts are high.

We find strong support for this prediction: the number of holding companies quadruples after the tax shock; the ratio of holding companies to all companies grows from 2% to 12%; this sharp growth in indirect ownership is unique to Norway; and firms with higher potential shareholder conflicts are particularly likely to be indirectly owned. Moreover, controlling for self-selection into indirect ownership does not alter our main result that firms with higher potential shareholder conflicts cut dividends less when dividend taxation increases. This evidence also suggests that indirect ownership can have positive effects. While a system of taxing intercorporate dividends makes it costlier to reduce agency costs by paying out free cash flow, a system of tax-free intercorporate dividends faced by the firms we analyze avoids this problem.

Overall, our evidence suggests that both taxes and agency costs are important for dividend policy, that the costly effect of dividends on taxes is actively traded off against the beneficial effect on agency conflicts, and that investors organize their ownership in ways that allow them to capture the beneficial effect of dividends on agency conflicts at the lowest possible tax cost.

These results shed new light on how the effect of taxes on dividends interacts with the main agency problem for most firms in any economy, which is the conflict of interest between majority and minority shareholders.

Janis Berzins is Associate Professor at BI (the Norwegian Business School).

Øyvind Bøhren is Professor of Finance at BI (the Norwegian Business School).

Bogdan Stacescu is Associate Professor at BI (the Norwegian Business School).