Investor Sentiment, Limits to Arbitrage, and The Cross-Section of Expected Returns: Empirical Evidence From Developed and Emerging Equity Markets

Abstract

In this paper, we investigate the interaction of two forces—investor sentiment and the limitation of arbitrage—on the cross-section of expected equity returns in developed and emerging markets. The main idea is that sentiment does not impact on every stock. It is likely to have the greatest impact on hard-to-value and expensive-to-arbitrage stocks, such as small, young, volatile, distressed, non-dividend paying, high growth and intangible-intensive stocks. The paper builds upon the noise-trader risk framework and the behavioural asset pricing literature and draws a distinction between investor sentiment in the short- and long-run and adds it to the distinction between idiosyncratic volatility, illiquidity, bid-ask spread and funding spread proxies of arbitrage frictions. A secondary empirical synthesis is constructed around the cross sectional portfolio tests, Fama-French and Carhart risk adjusted alphas, interaction regressions and around an AMOS style structural equation model. Three major conclusions can be drawn from the results. First, the sentiment of short run changes is positively related to the sentiment-prone portfolio returns at the same time. Second, long run sentiment is a negative predictor of any future returns, particularly when it leads to an exuberant price run-up from fundamentals. Third, when sentiment is high and limits to arbitrage are high, returns from anomalies are more likely to be high, suggesting that the mispricing associated with sentiment remains when arbitrage can be costly or risky or institutionally limited. Evidence is also more prevalent in less liquid markets, such as emerging markets, where both short selling markets and retail participation are less developed and have a more significant effect on price; in the developed markets, this effect is visible in the short leg of anomaly strategies. The paper makes a contribution by developing a behavioural asset pricing model that combines the concepts of sentiment decomposition, interaction between limits-to-arbitrage and comparison across markets.