Enhancing Earnings Predictability Using Individual Analyst Forecasts

Publisher Website
Google Scholar
Abstract
There is considerable evidence suggesting that stock election based on firms' anticipated earnings can generate excess returns. The earnings predictor model (EPM) introduced in this article uses individual analyst forecasts to generate an earnings forecast that is more accurate than the consensus in over 1,200 (non-independent) backtests using three alternative metrics. The model determines those firm-specific components that can best generate superior earnings forecasts for each company at each point in time. The EPM is shown to have been very effective for stock selection purposes, generating a total annualized Q1 minus annualized Q5 return differential of 15.57% over the period of the study.

This publication has 7 references indexed in Scilit: