Financial Analysts' Underreaction and Reputation-Building Incentives

Abstract

This thesis examines the role of reputation in financial analysts' underreaction in earnings forecasts. Prior research suggests that the reputation effect mitigates short-term economic incentives that lead to overly optimistic forecasts, and hence, increases forecast accuracy (i.e., an aspect of high quality forecasts). In contrast, I hypothesise that certain factors affecting analyst reputation lead to analysts' underreaction. Specifically, when faced with uncertainty, analysts employ underreaction as a mechanism to improve consistency between their forecast revisions and subsequent news (i.e., another aspect of high quality forecasts), so as to protect themselves from incurring a higher reputation cost of inaccuracy for inconsistent versus consistent consecutive forecast revisions and forecast errors (i.e., asymmetric reputation cost). In my first research question, I examine the asymmetric reputation cost theory that predicts underreaction increasing with uncertainty and asymmetric reputation cost. I contextualise my study in business cycles where both factors change. I predict and find that uncertainty is greater during recessions than expansions whereas asymmetric reputation cost is greater during expansions than recessions (i.e., reputation concerns are greater during expansions). Further, I find that analysts' underreaction is greater during expansions than recessions. The implication is that the asymmetric reputation cost, rather than the uncertainty, drives analysts' underreaction. In my second research question, I investigate the differential underreaction to good news versus bad news in relation to short-term economic incentives and the reputation-building incentives simultaneously. If analysts put more emphasis on short-term gains, they will underreact more to bad news than good news, particularly during recessions where the short-term economic incentives are heightened. On the contrary, if analysts are more concerned with their reputations, they will underreact less (more) to bad news than good news during recessions (expansions), because bad (good) news is more likely to follow in bad (good) times and, accordingly, they can incorporate the current bad (good) news with greater confidence. My findings are consistent with the reputation-building incentive theory, but inconsistent with the short-term incentive theory. Robustness tests and further research considering industry/firm specific information provide consistent results. Overall, the thesis suggests that analysts underreact to information due to their reputation concerns.