The Inferential Value of Quarterly Earnings Announcements*

ABSTRACT We propose a new methodology to assess the usefulness of information - its inferential value. We define inferential value as the ability of the capital markets to draw the correct inference from the information signals regarding future firm performance. We then examine the inferential value to investors of accounting versus non-accounting sources of information. Both average quarterly accounting rate of return on equity and excess stock returns are used as measures of firm performance. The findings indicate that a market-normalized accounting rate of return derived from stock prices around earnings announcement dates are more highly correlated with changes in future firm performance than similar measures in the non-disclosure periods. The findings support the prevalent view that accounting information disciplines information from other sources. Our findings also suggests that, despite the greater information environment of large firms, earnings announcements are as important in anticipating future profitability as they are for smaller firms.