Premium
An Approach to Adjusting Analysts' Consensus Forecasts for Selection Bias *
Author(s) -
HAYES RACHEL M.,
LEVINE CAROLYN B.
Publication year - 2000
Publication title -
contemporary accounting research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.769
H-Index - 99
eISSN - 1911-3846
pISSN - 0823-9150
DOI - 10.1111/j.1911-3846.2000.tb00911.x
Subject(s) - earnings , econometrics , economics , selection bias , sample (material) , undo , incentive , population , forecast error , consensus forecast , actuarial science , statistics , mathematics , finance , computer science , microeconomics , chemistry , demography , chromatography , sociology , operating system
Many recent empirical studies have concluded that analysts' earnings forecasts are optimistic on average. In this paper, we attempt to undo the effect of one potential source of optimistic bias in analysts' earnings forecasts. Assuming forecasts come from a truncated normal distribution, we estimate the “true” population mean using maximum likelihood. We find that our estimates of earnings are more accurate and less biased than standard measures of sample mean and median. However, we do not find a closer relationship between excess market returns and forecast errors from our maximum likelihood estimate than from the sample mean. This may suggest that the market does not fully incorporate analysts' incentives in generating expectations about future earnings.