DISPERSION IN ANALYSTS’ EARNINGS FORECASTS AND FUTURE STOCK RETURNS: EVIDENCE FROM JAPAN

Abstract

Some recent empirical studies in the US document a negative relation between the dispersion in analysts’ earnings forecasts and future stock returns. Both the empirical validity and the explanations for this negative relation have been a topic of continuing debate. This study aims to contribute to this debate by providing out-of-sample evidence from Japan. In contrast to recent US-based evidence, we find a positive relation between analysts’ forecast dispersion and future stock returns. We demonstrate that a zero-investment strategy that is long in the high dispersion stocks and short in the low dispersion stocks yields statistically significant monthly risk-adjusted returns of 0.368% (the CAPM model). The results are robust to sensitivity tests involving the use of an alternative dispersion measure, an alternative sample of firms with increased analyst coverage, a lag in portfolio formation and industry controls. Further analysis indicates that the dispersion premium we document is concentrated in medium and large size stocks. Our findings are in contrast to recent US evidence of a negative dispersion-returns relation. We identify and discuss some of the differences between the US and Japan markets, that might account for the differences in the results. Analysts’ forecasts for high dispersion stocks seems less optimistic in Japan than is the case for the US, indicating that, unlike in the US, high dispersion stocks may not be overpriced in Japan. Further regression analysis suggests that dispersion is strongly associated with known risk factors and could be a proxy for risk, which possibly explains our documented dispersion premium effect. Overall, our findings serve to show that it might be premature to assume the negative dispersion-returns relationship to be a universal phenomenon.