On the Use of Residuals as Dependent Variables

Abstract

We discuss some pitfalls associated with using residuals as dependent variables in accounting research. We provide basic intuition for why the typical implementation of this procedure generates biased coefficients and standard errors that can lead to incorrect inferences with both type I and type II errors. We illustrate model specification issues associated with estimating regressions in two steps versus one and discuss possible over-specification and the use of fixed effects. We provide Stata code and sample data to help illustrate and understand this bias and conclude with concrete guidance for researchers and reviewers of archival research.