Data Truncation Bias and the Mismeasurement of Corporate Tax Avoidance

Abstract

This study examines the use of the cash effective tax rate (Cash ETR) as a measure of corporate tax avoidance. The Cash ETR normalizes cash taxes paid by pre-tax book income. Empirical studies of the Cash ETR delete or winsorize many observations for which the measure is not meaningful. In addition, high pre-tax book income in the denominator biases the Cash ETR towards the statutory rate. We propose an alternative measure that is meaningful for all observations. Whereas Cash ETR studies suggest that the average corporation is tax-favored, our measure indicates that the opposite is true. In addition, our study suggests an explanation for inconsistencies in the extant literature regarding the relation between pre-tax profitability and tax avoidance. Our measure shows that that more profitable firms engage in greater levels of tax avoidance.