A new measure for differences of opinions: Institutional trade dispersion

Using 13F quarterly holdings data of institutional investors, we introduce a novel measure for differences of opinions among investors by analyzing divergence in institutional investors’ trades. We find that increases in institutional trade dispersion predict a significant decline in future abnormal returns. Moreover, we find that this relationship persists after controlling for other proxies of divergence of investor opinions, such as the dispersion in analysts’ earnings forecasts, the change in the number of institutions holding the stock, and the dispersion in institutional investors’ holdings relative to their benchmarks. Consistent with Miller (1977), underperformance of high-dispersion stocks is found to be the strongest among stocks that experienced recent significant price increases, the stocks with the highest mispricing scores, and the stocks that face increases in short-interest positions after the onset of dispersion.