Assessing the Relation Between Taxes and Stock Returns: The Critical Role of Choosing the Tax Variable

Abstract

Firms invest non-trivial resources to avoid paying taxes. One of the presumed incentives for doing so is that it should increase the value of the firm. Surprisingly, a large number of studies find that tax expense is positively related to stock returns, suggesting that paying more taxes is good news. We re-evaluate this puzzling relation and provide new evidence that an increase in tax expense is indeed bad news. We begin by showing that the tax expense variable in a regression of returns on pretax income and tax expense captures the market’s pricing of book-tax differences, and importantly, the different tax variables used in the literature capture different components of book-tax differences. Using a regression that clearly distinguishes between temporary and permanent book-tax differences, we show that both temporary and permanent differences (after controlling for influential observations) are negatively related with contemporaneous returns. We also demonstrate that the positive association between contemporaneous returns and current and total expense documented in prior research is explained by influential observations and the intricacies of the accounting for temporary differences. Finally, we provide evidence that the positive relation between future returns and total tax expense, or tax expense momentum, can be attributed to the same market inefficiency captured by earnings momentum.