Earnings and Stock Selection

Abstract

The pursuit of excess returns adds depth and momentum to the efficiency of stock markets, and yet accounting and finance literature lack a widely accepted model for stock selection. Equity selection risk, that is, the risk of making an adverse selection leading to a return lower than the average for comparable equities, is the investment risk element that is most relevant to earning excess returns under competitive conditions. This paper claims that, when making buy recommendations, financial analysts attempt to minimise equity selection risk by classifying companies into risk classes on the basis of economic similarities, and then select the companies that are expected to outperform on the criterion of expected earnings. Consistent with these arguments, it is shown that for predicting the risk perceptions of analysts, (1) expectational data are more useful than historical data and that (2) the earnings expectations proxy, expected earnings growth, is more useful than historical data-based accounting risk measures. The paper next addresses the issue of whether the analysts' approach to stock selection can be generalised to active investors. It is concluded that, for the purpose of generating excess returns, there are no theoretical impediments why investors also should not: (i) classify companies by perceived economic similarities and (ii) select the companies that are expected to outperform on the criterion of expected earnings. This role of earnings expectations, as the determinant of stock selection under competitive conditions, can help explain the importance attributed to earnings by investors and the accounting profession.