34 Pages Posted: 19 Jul 2000
Date Written: April 17, 2000
This paper develops a theory of the frequency of financial analysts' forecast revisions and then tests the empirical predictions of the model. Financial analysts act as information intermediaries for firms and investors and therefore their forecast revision frequency helps explain the equilibrium of the supply of and demand for earnings predictions and assessments of firm value. The theory is based on the analyst's costs of information gathering and the profits obtained from selling the information to investors. Our analysis is conducted in two stages. In the first stage, a single-period, Kyle (1985) model is used to determine the profits generated by privately informed investors who trade on the analyst's forecast revision. The analyst is assumed to be compensated as a function of these profits. In the second stage, the analyst's optimal revision frequency to collect and sell private information is determined. We find that the analyst's optimal revision frequency is increasing in the variance of liquidity trading volume, the volatility of the underlying earnings process, and the earnings-response coefficient and decreasing in the total number of informed traders who invest in the firm and the cost of revision. These theoretical results are developed into empirical hypotheses that the frequency of analysts' forecast revisions between earnings announcements is positively associated with variability of the earnings process, average prior trading volume, and earnings response coefficients, and negatively associated with skewness of prior trading volume, after controlling for firm size and prior average daily stock price changes. These hypotheses are tested cross-sectionally and we find significant support each of the hypothesized relations.
JEL Classification: D40, G29, M41
Suggested Citation: Suggested Citation
Stuerke, Pamela S. and Holden, Craig W., The Frequency of Financial Analysts' Forecast Revisions: Theory and Evidence (April 17, 2000). Available at SSRN: https://ssrn.com/abstract=232023 or http://dx.doi.org/10.2139/ssrn.232023