Analysts who have more consistent forecast errors over time influence the movement of stock prices significantly more than analysts who forecast accurately but less consistently, especially when sophisticated investors are involved. Consistent analysts receive more recognition, such as being nominated as “all-stars.”
The authors examine whether consistent estimates over time are more informative to investors than accurate estimates. They find that consistency is significantly more informative than accuracy, especially if institutional investors are involved. Their evaluation of the effect of consistency adds to the current literature because previous studies viewed accuracy, not consistency, as the critical measure. Accuracy is defined as the absolute distance from the forecast to actual earnings at a point in time. Consistency is defined as the volatility of the unexpected error in the forecast and assumes that investors are Bayesian.
The authors’ main findings are as follows: Consistent estimates provide more useful information to sophisticated investors than accurate ones with unpredictable forecast errors; more consistent forecast errors are associated with greater price moves; analysts showing more consistent forecast biases are more likely to be nominated as “all-star” analysts; and analysts tend to use downwardly biased forecasts in order to deliver more consistent but less accurate estimates, especially when covering stocks with more institutional ownership.
How Is This Research Useful to Practitioners?
Consistent forecasts are more informative than accurate forecasts. Specifically, the authors find that consistency is two to four times more economically and statistically significant than accuracy with respect to the abnormal reaction in stock prices when earnings revisions are announced. A valuable takeaway for practitioners is the finding that if consistency increases by one standard deviation, the market reaction increases by about 50% of the median beta.
The authors test three implications of their conclusion and find that all of their deductions are valid. The first deduction is that analysts’ welfare improves as the consistency of their forecasts improves. Consistent analysts are more likely to be designated as all-stars by institutional investors and less likely to be demoted.
The second deduction is that analysts tend to low-ball their true expectations to increase consistency, and the authors confirm that low-balling does increase consistency. Prior literature indicated that analysts can enable managers to beat consensus forecasts. By issuing biased reports, analysts may continue to gain favor with company management and receive better information than others, thus allowing them to improve their consistency.
The third deduction is that consistency is more useful to analysts who provide estimates to sophisticated investors. Institutions are able to untangle the consistent bias and identify analysts who have predictable forecasting capability. But accuracy is more important to analysts who cater to the retail investor. The retail investor seems less equipped to identify bias and thus desires more transparency.
Consistency, accuracy, and the tendency to low-ball have all declined since the enactment of Regulation Fair Disclosure in 2000, which requires companies to publicize material information to all investors simultaneously.
Practitioners and regulators could benefit from understanding the motives of the analysts they monitor.
How Did the Authors Conduct This Research?
The authors regress abnormal stock returns on quarterly forecast revisions to test consistency and accuracy after controlling for many factors, such as analyst boldness, horizon, experience, and breadth of firms covered. They use actual earnings and analyst forecast data from the Thompson Reuters I/B/E/S database for quarterly forecasts beginning in 1994 and ending in 2006. The accounting data are from Compustat, and stock return information is from CRSP daily files.
The authors’ findings are robust, and they are able to confirm their conclusions when using alternative specifications. They repeatedly find that consistency is more informative than accuracy.
The authors provide support for the belief that the relationship between an analyst and company management can be mutually beneficial. The analyst is rewarded as an all-star if he or she provides consistent forecasts and institutional investors are present, and management may be rewarded by beating low-balled consensus forecasts. Institutional investors also benefit from this process because they can identify and measure the consistent bias. Small individual retail investors may not benefit.