آیا معامله گران سود سهام می توانند بازده آتی سهام در ژاپن را پیش بینی کنند ؟

A growing body of literature suggests that investor sentiment affects stock prices both at the firm level and at the market level. This study examines the relationship between investor behavior and stock returns focusing on Japanese margin transactions using weekly data from 1994 to 2003. Margin trading is dominated by individual investors in Japan. In analysis at the firm level, we find a significant cross-sectional relationship between margin buying and stock returns. Both market-level and firm-level analyses show that margin buying traders follow herding behavior. They seem to follow positive feedback trading behavior for small-firm stocks and negative feedback trading behavior for large firm stocks. Our results show that information about margin buying helps predict future stock returns, especially for small-firm stocks at short horizons. The predictive power does not diminish even after controlling for firm size and liquidity.

A growing section of the finance literature suggests that investor sentiment affects stock prices both at the firm level and at the market level. This paper extends this literature by linking investor sentiment to margin trades and examines how these quantities predict future stock returns in Japan. De Long et al. (1990) demonstrate that if risk-averse arbitrageurs know that prices may diverge further away from their fundamental values before they converge, they will take smaller positions when betting against mispricing. Therefore, if the sentiments of noise traders are systematically correlated and there are constraints on arbitrage, their investment behavior may predict future market prices. Yet, the direction of causality is not entirely clear because the behavior of noise traders may be influenced by the market. Fisher and Statman (2000) examine the usefulness of a variety of sentiment variables in predicting short-horizon market returns. Baker and Wurgler (2006) examine how investor sentiment affects the cross-section of stock returns. When sentiment is pessimistic, subsequent returns are relatively high for smaller stocks, high-volatility stocks, unprofitable stocks, non-dividend-paying stocks, extreme-growth stocks, and distressed stocks. When sentiment is optimistic, these patterns attenuate or, in several cases, fully reverse. Brown and Cliff (2004) document that returns cause sentiment rather than vice versa. In his analysis of volatility, Brown (1999) finds deviations from the mean level of investor sentiment are positively related to volatility during the same period. Wang et al. (2006) examine the relationship between sentiment, returns, and volatility and find that investor sentiment is caused by returns and volatility rather than vice versa. In addition, lagged returns cause volatility. Odean's (1998) model shows that investor overconfidence will increase trading volume. Gervais and Odean (2001) argue that high past market returns may cause overconfidence in individual investors if they happened to invest in stocks in the same period. Using monthly market data, Statman et al. (2004) show that investor overconfidence is positively related to trading volume. Baker and Stein (2004) propose a model that explains why increases in liquidity are associated with lower subsequent returns at both the firm level and the aggregate level. When short sales are constrained, unusually high liquidity is a symptom of market domination by irrational investors who underreact to the information contained in order flow. Individual investors have long been considered to be noise traders. They are less informed or trade for non-informational reasons. Nevertheless, if their trades are correlated and arbitrage is limited in some way, their investments will change asset prices. Lee et al. (1991) argue that the discount on closed-end funds can be explained by the irrational behavior of individual investors. Because of leverage, margin transactions are sometimes considered speculative and major players in these transactions tend to be individual investors. Therefore, we argue that margin transactions tend to reflect individual investor sentiment. This study examines the relationship between investor behavior and stock returns by focusing on margin transactions in Japan. Margin trades are widely thought to be dominated by individual investors in Japan. First, we confirm that margin transactions are indeed dominated by individual investors. Second, we examine how margin transactions are related to stock returns. We look for specific patterns that are consistent with apparently irrational behavior. Our market-level analysis shows that margin buying is dominated by individual investors, but that margin selling is not. In analysis at the firm level, we find a significant cross-sectional relationship between margin buying and stock returns. We do not find significant patterns for margin selling. Both the market-level and firm-level analyses show that margin buying traders follow herding behavior. They seem to follow positive feedback trading behavior for small-firm stocks and negative feedback trading behavior for large firm stocks. As predicted, margin traders heavily impact the stock prices of small firms over a certain period of time. The deviation from previous value exists longer and is more pronounced for small-firm stocks that are mainly owned by individual investors. Our results show that information about margin buying shares outstanding helps predict future stock returns, especially for small-firm stocks. The predictive power does not diminish even after controlling for liquidity. This is consistent with De Long et al. (1990), who show that stock prices deviate from their fundamental values for a certain period of time due to excess demand by noise traders. This is the first comprehensive study of Japanese margin trading using weekly data over a long period of time. These weekly data cover most stocks eligible for margin trades. Standardized margin trades have been practiced in Japan for more than fifty years. In contrast to the United States, the Japanese margin trading system is advanced and highly centralized. The Japanese system allows stockbrokers to borrow securities and funds from specialized securities finance companies when there is a shortage of securities and funds. Because of this highly evolved system, margin traders almost always use the standardized margin trading system when they can satisfy its requirements. Japanese margin data are complete and market-wide compared to U.S. data, which include margin transactions for only the largest brokerage firms. Furthermore, individual firm data are not available in the U.S. The structure of the paper is as follows. The next section describes and compares margin transactions in Japan and the U.S. The third section discusses Japanese margin data. The fourth section discusses results for the aggregate market. The fifth section examines margin transactions and stock returns at the firm level. A brief conclusion follows.

This study examines the relationship between investor behavior and stock returns focusing on Japanese margin transactions. We use weekly margin transactions data from 1994 to 2003 for our analysis. Our market-level analysis shows that Japanese margin buying is dominated by individual investors. Individual investors appear to follow positive feedback trading behavior because the change in margin buying shares outstanding is positively autocorrelated, and is positively related to stock market performance in the recent past. Aggregate margin selling transactions, however, are practiced by all investors and may be strongly influenced by institutions. This is consistent with the conventional wisdom that institutional investors do not need to borrow money to purchase stocks. We do not find evidence of positive feedback behavior for margin selling. Our individual firm-level analysis shows that margin buying investors do not follow positive feedback trading behavior. Instead, they seem to follow negative feedback trading. Margin buying investors increase their positions in particular stocks when the recent performance of the market was high but the recent performance of these stocks was poor. Interestingly, excess returns of these stocks in the following week are significantly positive. On the other hand, the subsequent excess returns of stocks in which margin traders reduce their positions are significantly negative. Margin traders' herding behavior seems to impact stock prices in the following week. One possible explanation is that the trading decisions of risk-averse market-makers and constrained arbitragers contribute to the relation between margin buying and subsequent returns. Analysis of firm size suggests that margin traders follow positive feedback trading behavior for small-firm stocks and negative feedback trading for large firm stocks. Yet, predictability persists regardless of firm size. In addition, the predictive power of margin trades does not diminish after adjusting for liquidity. It is extremely puzzling that individual Japanese margin traders follow positive feedback trading behavior for small-firm stocks while also following negative feedback trading behavior for large firm stocks. How individual Japanese margin traders can so effectively time the market and the related microstructure issues for market-makers is an intriguing topic for future research.