Nickels not pennies: Explanations and implications of granularity in analysts' EPS forecasts

This paper describes a striking empirical regularity---the latest EPS forecasts made before annual earnings announcements are about five times as likely to be multiples of nickels as otherwise---and investigates the implications of this phenomenon for forecast accuracy, forecast bias, and stock market response to forecast revisions and earnings announcements. The evidence supports the view that analysts are more likely to issue nickel forecasts when they are less sure about their forecasts. There is also support for the view that nickel forecasts are more likely when managerial targets are multiples of nickels. Turning to the implications, I find that nickel forecasts are less accurate. Consistent with the finding on lower forecast accuracy, I observe lower market responses to forecast revisions when the new forecasts are nickel forecasts as well as to earnings announcements when the individual forecasts underlying the consensus are predominantly nickel forecasts.