The Information Content of Analyst Forecasts— An Econometric Analysis of Informational Leadership

We measure the information content of monthly analyst consensus forecasts for one-year-forward earnings per share (EPS) based on two well-established price discovery measures drawn from the area of market micro-structure research. Employing a 36-year sample of large American companies listed in the S&P 100 Index, we compute (i) Hasbrouck’s information shares and (ii) Gonzalo and Granger’s common factor components to measure the relative share that the capital market and the analysts have in the process of price discovery. We find that while analysts do not lead the capital market, they have a small but significant share in the process of price discovery, amounting to 4.5% (Hasbrouck) or 18.0% (Gonzalo and Granger) on average. This share varies significantly in the cross-section. We identify a company’s analyst coverage as an explanatory factor: The larger the coverage, the lower is the analysts’ information share. This finding can be explained by analysts’ herding behavior, which lowers the information content of their estimates. JEL Classification: G12, G14, G17, G2, C32, C53, D82