THE THEORY OF VALUE RELEVANCE

I. INTRODUCTION

Accounting value relevance is a concept that has admitted a number of definitions and
measures. Lev (1989) asserted that the relevance of accounting value was characterised by the quality of accounting information. For Lev, earnings quality was measured by the coefficient of determination in a regression of market returns on earnings. The strength of association between market returns and earnings is the basis of most measures of value relevance. For example, Collins et al (1997) and Lev and Zarowin (1999) both used the coefficient of this association (the earnings association coefficient) to estimate value relevance. Chang (1998) suggested the variance of the log of the value-price ratio as a measure of value relevance, with value determined from an earnings-based valuation model. In some cases, for example Chang (1998), the association between earnings and market returns is assessed with a lag rather than contemporaneously, reflecting different rates at which information is impounded. And in some cases, the variability in market returns is controlled for by forming portfolios sorted exogenously, with the association between earnings and returns then assessed across portfolios (see for example Francis and Schipper (1999) and Nwaeze (1998)).