Individual Investor Sentiment and Stock Returns: Evidence from the Korean Stock Market

ABSTRACT

We investigate the dynamic relationship between individual investor sentiment and stock returns in the Korean stock market. The evidence indicates that individual investor sentiment has no significant explanatory power for cross-sectional stock returns. However, individual investors’ trades can move stock prices in certain stocks by their contrarian behavior, which leads them to implicitly provide liquidity to other market participants. In addition, individual investors earn a small market-adjusted excess return in the short-horizon future as compensation for liquidity provision. Our findings show that short-horizon return predictability of individual investors does not come from their private information.

KEY WORDS:

• individual investors
• investor sentiment
• return predictability
• stock returns

Acknowledgments

The authors are grateful for helpful comments from Ali M. Kutan (the editor) and three anonymous referees.

Notes

1. By contrast, a few studies suggest that some trades by individual investors are profitable (Choe, Kho, and Stulz 2005; Coval, Hirshleifer, and Shumway 2005; Ivković and Weisbenner 2005).

2. The Korea Exchange (KRX) reports that individual investors dominate trading volumes on the KRX, accounting for 88.19 percent of total market activity (institutions 5.37 percent, foreign investors 4.99 percent) over the period 2001–2009. Even on the basis of trading values, individuals’ share of total trading remains dominant at 61.32 percent of total market activity (institutions 17.27 percent, foreign investors 18.23 percent).

3. Our analysis is based on a clientele-based model documented by Kumar and Lee (2006), in which different investor groups restrict themselves to trading within different natural “habitats,” or groups of stocks. Kumar and Lee (2006) conjecture that the returns of individual stocks reflect not only fundamental risk, but also changes in sentiment of important investor groups.

4. Literature documents that individual investors tend to be contrarian traders while institutional investors rather tend to be momentum traders (see Choe, Kho, and Stulz 1999; Chordia, Goyal, and Jegadeesh 2011; Goetzmann and Massa 2002; Grinblatt and Keloharju 2000; Kaniel, Saar, and Titman 2008; Nofsinger and Sias 1999).

5. A mispricing is the result of an uninformed demand shock in the presence of a binding arbitrage constraint. In practice, these two distinct channels lead to quite similar predictions because stocks that are sensitive to speculative demand and subjective valuation also tend to be the riskiest and costliest to arbitrage (Baker and Wurgler 2006).

6. Existing literature can be classified by how they measure investor sentiment: first, measuring the sentiment by investor surveys (Brown and Cliff 2004), and second, examining order and trade imbalances that have become popular in recent studies (Barber, Odean, and Zhu 2009; Dorn, Huberman, and Sengmueller 2008; Hvidkjaer 2008; Jackson 2003; Kaniel, Saar, and Titman 2008; Kumar and Lee 2006). Others have looked into closed-end fund discounts (Lee, Shleifer, and Thaler 1991) or formed a composite sentiment index built on sentiment proxies: the closed-end fund discount, NYSE share turnover, the number and average first-day returns on IPOs, the equity share in new issues, and the dividend premium (Baker and Wurgler 2006).

7. Wurgler and Zhuravskaya (2002) document that the high idiosyncratic risk for a stock makes arbitrage especially risky.

8. We calculate the equity holdings ratio of individual investors as shares held by individual investors with less than 1 percent over the floating shares. Aitken and Comerton-Forde (2003) contend that this is particularly important for the emerging stock markets, where company founders typically hold large portions of stocks that are not freely traded.

9. We are interested not only in which stocks individual investors own, but also in how important their trades are in these stocks relative to the trades of other market participants. The trading intensity of individual investors in our sample is defined as the number of shares traded by individuals divided by the number of shares traded in the market.

10. The summary statistics for the factors as well as for the test portfolios are available upon request.

11. Our findings contrast with those of Kumar and Lee (2006), who examine correlations among trade imbalances of stocks traded by clients of a single U.S. discount broker. They find that their measure of trade imbalance is moderately correlated across stocks and conclude that there is evidence of a systematic component in retail investor trading. They also find that investor sentiment may have significant effects on the cross section of stock returns. One distinct difference between our article and Kumar and Lee (2006) lies in the data, in that Kumar and Lee (2006) look into trading data of clients of a specific discount brokerage house, whereas we analyze extensive retail trading data available for all firms listed on the KRX. Trading activities of investors from a particular brokerage firm are correlated because investors from the same brokerage are more likely to share a similar set of information or follow the same investment recommendations supplied by the brokerage firm (Barber, Odean, and Zhu 2009).

12. In addition to the common directional component, limits to arbitrage exist for sentiment to affect equilibrium stock prices (Shleifer and Vishny 1997). Barber, Odean, and Zhu (2009) document that trading by individuals is highly correlated, which is consistent with systematic noise trading that does not wash out in the aggregate.

13. It can be seen that statistically significant excess returns persist and grow in magnitude up to sixty days following a week when retail investors made buy transactions in the portfolio with the lowest fraction of retail investors (P4-Q5); we could not present it here due to space constraints.