استراتژی با شتاب 52 هفته ای در بازارهای بین المللی سهام

We study the 52-week high momentum strategy in international stock markets proposed by George and Hwang [George, T., Hwang, C.Y., 2004. The 52-week high and momentum investing. Journal of Finance 59, 2145–2176.]. This strategy produces profits in 18 of the 20 markets studied, and the profits are significant in 10 markets. The 52-week high momentum profits exist independently from the Jegadeesh and Titman [Jegadeesh, N., Titman, S., 1993. Returns to buying winners and selling losers: implications for market efficiency. Journal of Finance 48, 65–91.] individual stock and Moskowitz and Grinblatt [Moskowitz, T.J., Grinblatt, M., 1999. Do industries explain momentum? Journal of Finance 54, 1249–1290] industry momentum strategies. These profits do not show reversals in the long run. We find that the 52-week high is a better predictor of future returns than macroeconomic risk factors or the acquisition price. The individualism index, a proxy to the level of overconfidence, has no explanatory power to the variations of the 52-week high momentum profits across different markets. However, the profits are no longer significant in most markets once transaction costs are taken into account.

Many studies document that stock returns are predictable based on past price information. Among them, Jegadeesh and Titman (1993) (JT henceforth) show that when stocks are ranked into deciles based on past medium-term returns, the top decile portfolio (winners) continues to outperform the bottom decile portfolio (losers) in a medium-term period; DeBondt and Thaler (1985) document long-term reversals of portfolio returns.3Moskowitz and Grinblatt (1999) (MG henceforth) claim that the medium-term momentum in individual stock returns is dominated by momentum in industry returns. Jegadeesh and Titman (2001) provide more recent evidence supporting the explanations of behavioral models, which incorporate both medium-term momentum and long-term reversals (e.g., Barberis et al., 1998, Daniel et al., 1998 and Hong and Stein, 1999) over risk-based explanations. In a recent paper, George and Hwang (2004) (GH henceforth) propose an investing strategy different from that of JT. At the end of each month, GH rank stocks based on the ratio of the current price to the past 52-week high. They then construct a zero-investment portfolio from purchasing the top 30 percentile (winners) and selling the bottom 30 percentile (losers) stocks, and hold these positions for a medium term of 6 or 12 months. GH find that this trading strategy generates profits comparable to those of the JT strategy. They also show that this 52-week high strategy explains a large portion of the JT momentum profits and that there are no return reversals in the long run, inconsistent with the predictions of behavioral models. Their results, therefore, call for different theoretic models for the medium-term momentum and long-term reversals. It is crucial that any anomaly in financial markets pass the out-of-sample test to show that the results are not an artifact of data-mining. The JT and MG momentum strategies have been extensively studied with international data in previous studies.4 To our knowledge, however, only a few papers have studied the 52-week high momentum strategy in international stock markets.5Marshall and Cahan (2005) find significant profits of the 52-week high momentum strategy in the Australian stock market. They do not examine whether the 52-week high momentum profits reverse in the long run. Du (2008) finds that the 52-week high momentum profits reverse in the long run using 18 stock market indexes. Alsubaie and Najand (2008) find short-term price reversals, rather than price continuation, after stocks reach their 52-week highs in the Saudi stock market. Burghof and Prothmann (2009) show that the 52-week high strategy is profitable in the UK market and the profit is correlated with information uncertainty. In contrast, we conduct a more comprehensive study on the 52-week high investing strategy with individual stocks in 20 major stock markets. We examine the profitability and long-term performance of the GH strategy in international stock markets, and investigate the sources of these profits. Our main findings are as follows. First, the 52-week high momentum effect is robust in international markets. In our sample of twenty major stock markets, eighteen show profits from this trading strategy. Among them, ten exhibit statistically significant profits, including nine out of thirteen European markets and the Hong Kong market. Their average monthly returns range from 0.60% to 0.94%, compared to 0.45% in the U.S. market in George and Hwang (2004). Eight and seven of these ten markets also have statistically significant JT and MG momentum profits, respectively. In seven markets (Japan, Norway, Russia, Singapore, South Korea, Spain, and Taiwan), none of the three momentum strategies has significant profits. Our results show that GH, JT, and MG momentum effects tend to coexist in a market. Our second finding shows that, although the returns of these momentum strategies are highly correlated, the GH momentum profits still exist conditional on the JT or MG momentum. Unlike the U.S. market in George and Hwang (2004), the 52-week high momentum strategy does not dominate the JT momentum strategy. Instead, the profits of the GH and JT momentum strategies exist independently. Therefore, the 52-week high phenomenon documented in GH is a genuine one in financial markets. Third, based on the event-study approach in Jegadeesh and Titman (2001), our results on the long-run performance of the GH momentum strategy support the view that the 52-week high momentum profits do not reverse, consistent with GH’s finding in the U.S. market. Fourth, we investigate the sources of the 52-week high momentum profits. The models that we adopt include the macroeconomic risk model in Chordia and Shivakumar (2002), and the anchoring model in Grinblatt and Han (2005). Our results show that these models cannot fully capture the GH momentum profits, and the 52-week high is a better predictor of future stock returns than macroeconomic risk factors or stock acquisition price. In Chui et al. (2010), the Hofstede individualism index, a proxy to the level of overconfidence across countries is shown to be closely related to JT momentum profits. However, we find that it has no explanatory power to the variations of the GH momentum profits across markets, suggesting that cultural difference in terms of overconfidence is not the cause of the 52-week high momentum profits. Our evidence supports the view that the 52-week high momentum profits are caused by investors’ anchor-and-adjust bias when they use the 52-week high price as an anchor. To understand whether this behavioral bias is exploitable in financial markets, we examine the profitability of this momentum strategy net of transaction costs at last. Following the approach in Lesmond et al. (1999), we estimate the transaction costs for the momentum portfolios. We find that once the transaction costs are taken into account, the GH momentum profits are no longer significant in most markets. Therefore, the profitability of the 52-week high momentum strategy documented earlier does not mean that markets are not semi-strong efficient. The remainder of the paper is organized as follows. In Section 2, we discuss the data set used in our empirical analysis. In Section 3, we present empirical evidence of the 52-week high momentum profits in international markets. We examine the relation of the three momentum strategies by portfolio analysis and cross-sectional regressions in Section 4. We also investigate the risk-adjusted returns and long-term returns to compare the GH and JT momentum strategies. In Section 5, we investigate the sources of the 52-week high momentum profits, and the momentum profits after trading costs. Section 6 concludes.

George and Hwang (2004) find that an investing strategy based on the nearness to the past 52-week high generates comparable profits as the momentum strategy proposed by Jegadeesh and Titman (1993) and the Moskowitz and Grinblatt (1999) industrial momentum strategy in the U.S. market, and these profits do not revert in the long run. In this paper, we conduct a comprehensive study on this 52-week high momentum strategy in 20 major stock markets. Our main findings are as follows. First, we find that the 52-week high momentum profits are robust in international stock markets. Eighteen out of twenty markets in our sample present evidence of a profitable 52-week high momentum strategy, with ten having significantly positive profits. Second, our portfolio and cross-sectional regression analyses show that the MG industry momentum effect is weaker after controlling for the 52-week high momentum effect. But the GH and JT momentum effects tend to coexist independently in a market. Third, our results from the event-study approach indicate that there are no return reversals on the 52-week high momentum strategy, especially in the most recent period. Fourth, the 52-week high momentum profits cannot be explained by the macroeconomic risk model or the anchoring model based on the acquisition price. Though the individualism index, as a proxy to the level of overconfidence, is correlated to the returns from the GH momentum strategy, it does not have explanatory power to the variations of the momentum profits across different markets after controlling for other country characteristics. The robustness of the 52-week high momentum strategy in international stock markets seems to suggest that markets are not semi-strong efficient since the nearness of a stock’s price to its past 52- week high is public information. However, once the transaction costs are taken into account, the 52- week high momentum profits are no longer significant in most markets. Consistent with George and Hwang (2004), we find that the 52-week high as a price level is a better predictor of future stock returns than macroeconomic risk factors or the acquisition price. It has predictive power no matter whether stocks have had extreme past returns or in the extreme industries. Our evidence suggests that investors may use the 52-week high as a reference point against which they evaluate the potential impact of news and therefore incur an anchor-and-adjust bias when a stock price is near or far from its past 52-week high. This leads to a short-term underreaction. When the information eventually prevails, the market resolves the bias without the overreaction that results in longterm reversals. How this cognitive bias affects the decision making of agents in financial markets will draw interest from more researchers in the future.