Dispersion in Analysts' Earnings Forecasts and Market Efficiency

Abstract

Recent studies show that firms with higher analysts' earnings forecasts dispersion subsequently have lower returns than firms with lower forecasts dispersion. This paper evaluates alternative explanations for the dispersion-return relation using a stochastic dominance approach. We aim to discriminate between the hypothesis that some asset pricing models can explain the puzzling negative relation between dispersion and stock returns, and the alternative hypothesis that the dispersion effect is mainly driven by investor irrationality and thus is an evidence of a failure of efficient markets. We find that low dispersion stocks dominate high dispersion stocks by second- and third-order stochastic dominance over the period from 1976 to 2012. Our results imply that any investor who is risk-averse and prefers positive skewness would unambiguously prefer low dispersion stocks to high dispersion stocks. We conclude that the dispersion effect is more likely evidence of market inefficiency, rather than a result of omitted risk factors.