Analyst Behavior, Crowdsourced Research and the Capital Markets

Abstract

This thesis consists of three papers revolving around stock analysts’ behavior, crowdsourced research and the capital markets. Using a news-based index of economic policy uncertainty (EPU), the first paper documents that EPU is positively related to analyst forecast dispersion and the number of analysts following the firm, and negatively related to analyst forecast accuracy. The uncovered relations are robust to various endogeneity corrections and hence are likely causal. Cross-sectional analyses indicate that the impacts of EPU on analyst behavior are more pronounced for firms with higher exposure to EPU, higher information asymmetry among investors, and higher transient institutional ownership. These findings suggest that EPU increases analyst following possibly because it increases investor demand for analyst research. The second paper studies the dispersion effect that aggregate analyst forecast dispersion negatively predicts future market returns. I find that the market return predictive power of aggregate dispersion only exists prior to 2005. The dispersion-return relation is partly driven by its correlation with the conditional equity premium. I find that the impact of investor sentiment on dispersion effect is also significant only prior to 2005. I do not find the short-sale constrained stocks experience higher dispersion effect. The third paper examines whether increased monitoring stemming from an innovation in financial technology affects corporate managers’ earnings management behavior. I find that firm managers significantly manipulate earnings upward after their firms being covered by Estimize, an open platform which crowdsources short-term earnings forecasts. This suggests that coverage by Estimize exerts pressure on firm managers to meet short-term earnings targets. I further show that this effect works through two channels: Estimize sets higher earnings targets than sell-side analysts do, and also reduces analyst short-term downward bias.