theUnitedStates,however, firms routinely convey
private information to analysts following them.- Modelsforforecastingearningsthatdependentirely
upon past earningsdata mayignore otherpublicly
available information that is useful in forecasting
futureearnings.Ithasbeenshown,forinstance,that
otherfinancialvariables suchasearningsretention,
profit margins, and asset turnover are useful in
predicting future growth. Analysts can incorporate
information from these variables into their forecasts.
Quality of Earnings Forecasts
Iffirmsarefollowedbyalargenumberofanalystsandthese
analysts are indeed better informed than the rest of the
market, the forecastsof growth that emerge from analysts
should be better than estimates based on either historical
growth orother publicly available information.But is this
presumption justified? Are analyst forecasts of growth
superior to other growth forecasts?
The general consensus from studies that have looked at
short-termforecasts ofearnings(one quarterahead to four
quarters ahead) isthat analysts provide better forecasts of
earningsthanmodelsthatdependpurelyonhistoricaldata.
The mean relative absolute error, which measures the
absolute difference between the actual earnings and the
forecastforthenextquarter,inpercentageterms,issmaller
foranalystforecaststhanitisforforecastsbasedonhistorical
data.Twostudiesshedfurtherlightonthevalueofanalysts’
forecasts. Crichfield, Dyckman, and Lakon-ishok (1978)
4 examinetherelativeaccuracyofforecastsintheEarnings
Forecaster, a publication from Standard & Poors that