Classification Shifting or Model Bias‘ A Reply to Barua and Cready (2008)

Abstract

In McVay (2006), I present evidence of managers classifying core operating expenses as special, in the year of the special item, to increase core earnings (which is the focus of both investors and analysts). The analysis requires a model of expected core earnings, and I discuss the limitations of this model at length in the original paper. Barua and Cready (2008) build upon these stated limitations and claim that the evidence of classification shifting is a byproduct of the expectations model. In this reply, I argue against this claim. I first present evidence of classification shifting for a subset of firms without using a model of expected core earnings. I next critique each of the authors' empirical tests, and then replicate their main analysis and show that their results are not consistent with model bias. I conclude my reply by summarizing the improvements to the classification shifting model made by subsequent researchers.