Binary Controls for the Corporate Marginal Tax Rate

Abstract

This paper examines the properties of simulated corporate marginal tax rates to assess their incremental benefit over other measures of corporate tax status. While binary variables have generally been replaced by simulated rates when marginal tax rate controls are needed, the results of this paper suggest a variety of advantages to using binary variables (or combinations of binary variables) in empirical specifications. In both cross-sectional and panel data settings, the results demonstrate that simple and easily constructed binary variables provide statistically equivalent results. Further analysis shows that binary variables are also statistically superior to alternatives to simulated rates that have been suggested in their absence. Additional tests show that using binary variables to control for firms’ tax incentives in empirical studies has several advantages over other available tax rate measures, including dramatically larger sample sizes, the ability to capture greater cross-sectional variation in tax incentives, and coefficient estimates that represent readily observable and discrete tax states.

Current version not for citation.

Keywords: corporate tax rate, marginal tax rate, taxes, measurement error

JEL Classification: G3, H25, M4, C200