Divergence of Opinion, Arbitrage Costs and Stock Returns

Abstract

In this paper we examine how divergence of opinion affect cross-sectional asset returns for different stocks with different arbitrage costs by employing a new proxy for divergence of opinion. We generalize Tauchen and Pitts' (1983) well-known Mixture of Distribution Hypothesis (MDH), which links asset volume and volatility in a way that derives a proxy for divergence of opinion among all individual investors. This new measure is a more reliable proxy for divergence of opinion among all individual investors than the existing proxies such as dispersion in analysts' earnings forecasts and turnover. We then use this measure of divergence of opinion in an empirical asset pricing analysis. In particular, we incorporate the crucial role of divergence of opinion in the determination of cross-sectional asset returns, establishing that when divergence of opinion is high, stock prices tend to be biased upwardly, resulting in lower future returns. These effects are especially pronounced for stocks with higher arbitrage costs including idiosyncratic risks, short sale costs, and other transaction costs, which are more difficult and costly to short sell. Hence the evidence for these stocks support Miller's (1977) view that, given short-sale constraints, observed prices overweight optimistic valuations. The predictions of recent theoretical work, such as Hong and Stein (2003), are valid only for stocks with less arbitrage costs. Also, our results suggest that the idiosyncratic risk, relative to other arbitrage cost measure, incrementally explain the divergence of opinon's effect on stock returns.