Essays on Analysts' Long-term Earnings Growth Forecasts

Prior research documents a large dispersion in long-term earnings growth forecasts among financial analysts covering the same stock. This suggests that while analysts believe long-term earnings growth is predictable, they face significant uncertainty when estimating it. This thesis examines the effect long-term earnings growth forecast dispersion has on the firm that receives long-term earnings growth forecasts, and investors who follow analysts’ stock recommendations based on valuations that incorporate long-term earnings growth. The first essay examines determinants of voluntary earnings growth guidance by management, and investigates analysts’ reactions to the guidance. The second essay examines trading strategies based on superior analysts’ long-term earnings growth forecasts. The results of the first essay indicate that there is pervasive upward bias in managers’ earnings growth guidance. As a result of this bias, managers are more likely to issue earnings growth forecasts in years of negative short-term earnings surprises and when analysts have low initial long-term earnings growth forecasts. The finding that managements’ earnings growth forecasts primarily convey good news is in contrast to the generally negative nature of management short-term earnings guidance, and suggests that different incentives drive firms’ disclosure of different financial information. Moreover, analysts respond to managements’ earnings growth guidance by increasing their long-term earnings growth forecasts. This finding sheds new light on the over-optimism hypothesis in analysts’ long-term earnings growth forecasts, suggesting the bias could be management induced. However, with regard to analysts’ response, the findings suggest that analysts sacrifice precision in long-term earnings growth forecasts for deliberate upward bias. This bias is not subject to ex post scrutiny and is designed to maintain or increase analysts’ target prices, even though short-term earnings may warrant a lower target price. Results in the second essay show a contemporaneous association between long-term earnings growth forecast accuracy, one-year ahead earnings forecast accuracy, and stock recommendation profitability. Despite these results, a model that predicts the profitability of future stock recommendations based on the analyst’s historical one-yea ahead earnings forecast accuracy and long-term earnings growth forecast accuracy, is not found.