The Implications Of Accounting Conservatism For The Relation Between Earnings And Stock Returns

Abstract

<p class="MsoNormal" style="text-justify: inter-ideograph; text-align: justify; margin: 0in 0.5in 0pt;"><span style="font-size: 10pt;"><span style="font-family: Times New Roman;">Characterizing accounting conservatism as the accountants’ tendency to require a higher degree of verification for recognizing good news than bad news, Basu (1997) predicts that the slope coefficient and R<sup>2</sup> in a regression of earnings on concurrent stock returns will be higher for bad news (negative stock returns) than for good news (positive stock returns). However, standard econometric analysis indicates that the R<sup>2</sup> is a function of the sensitivity of earnings to returns and the noise ratio, which is defined as the ratio of the variance of noise in earnings to the variance of noise in returns. I show that the R<sup>2</sup> from the regression of earnings on stock returns is not necessarily higher for bad news than for good news. So the test of R<sup>2</sup> is not a robust test of accounting conservatism. Consistent with the prediction, I find that the slope coefficient is higher for bad news firms reporting losses than for good news firms reporting profits, but R<sup>2</sup> is lower for bad news firms reporting losses than for good news firms reporting profits. </span></span></p>