Identifying Accounting Conservatism in the Presence of Skewness

Abstract

Previous studies have cast doubt on the construct validity of the asymmetric timeliness (AT) coefficient from the Basu (1997) model as a measure of conditional conservatism. The purpose of this paper is to clarify the exact conditions under which the AT coefficient fails to identify accounting conservatism. We show both analytically and by extensive simulations that the AT coefficient is not a reliable measure of accounting conservatism in the presence of skewness in returns and/or earnings unless returns are strictly exogenous. While earnings skewness is a predicted consequence of conditional conservatism, return skewness is arguably unrelated to conservative reporting. Return skewness therefore distorts the AT coefficient as a measure of conservatism. Simple skew reducing transformations or outlier-robust estimators do not overcome the distortion. Empirically, we provide evidence that cross-sectional variation in the AT coefficient, especially across groups sorted on firm size, is highly correlated with cross-sectional variation in skewness.

Keywords: conditional conservatism; asymmetric timeliness; piecewise linear regression; skewness

JEL Classification: M41, C15