How U.S. GAAP Distorts the Effective Tax Rate As a Measure of Tax Avoidance

Abstract

Many studies use GAAP effective tax rates (ETR) as proxies for tax avoidance and rely on the maintained assumption that very low (high) ETRs represent the greatest (least) tax avoidance. We provide large-sample empirical evidence on how well ETRs capture cross-sectional differences in tax avoidance. Using a sample of 6,789 income tax footnote disclosures from 2008 through 2016, we document that factors largely unrelated to tax avoidance explain most of the deviation from the statutory tax rate for ETRs below 5% and above 40%. For example, factors associated with firm performance comprise over 50 percent of the deviation for ETRs below 5% and 20 percent for ETRs above 40%. We also examine alternative tax avoidance measures and find the measure proposed by Henry and Sansing (2018) best reflects expected changes in tax avoidance across its distribution. Our findings inform researchers about factors unrelated to tax avoidance that drive significant deviations in ETRs from the statutory tax rate, which is of increasing importance as the number of studies examining the consequences of tax avoidance grows.