Relative versus incremental information content

Abstract

This study distinguishes between incremental and relative information content. Incremental comparisons ask whether one accounting measure provides information content beyond that provided by another, and apply when one measure is viewed as given and an assessment is desired regarding the incremental contribution of another (e.g., a supplemental disclosure). Relative comparisons ask which measure has greater information content, and apply when making mutually exclusive choices among alternatives, or when rankings by information content are desired (e.g., when comparing alternative disclosures). Questions of both incremental and relative information content arise frequently in accounting. However, few previous studies have examined questions of relative information content. Possible explanations include unfamiliarity with the relative versus incremental distinction, and the additional statistical complexity involved in testing for relative information content.

First, we examine analytically the relation between incremental and relative information content, demonstrating that they address different research questions and require different tests for statistical significance. Second, we identify accounting research contexts in which questions of relative and incremental information content arise. Third, we propose a new regression-based test for relative information content. This test applies to both returns and valuation studies, generalizes to any number of predictor variables, and can be used in conjunction with White's (1980) adjustment for heteroskedasticity. Fourth, we illustrate tests for relative and incremental information content in a familiar research setting that compares the information contents of net income, cash flows, and net sales in 40 industries.