As tax expense reflects value lost to taxes paid, it should be negatively associated with value, provided nontax, value-relevant information is controlled for. However, valuation regressions estimated in prior research�using contemporaneous tax expense and nontax variables�document substantial variation in the coefficients on tax expense, ranging from significant negative to significant positive values. We show this variation is (a) caused by the omission of expected future profitability, and (b) explained by many factors that cause variation in the correlations among included variables and omitted future profitability. Unfortunately, difficulties associated with separating the impact of individual factors hampers tax research investigating determinants of the value relevance of tax expense.