| With the rapid development of finance and economics, the market anomalies which go against the traditional financial theory are gradually revealed and it makes us to rethink about the efficient market hypothesis. Momentum, as one of the serious challenges that the risk based traditional asset pricing theory face, causes extensive study and discussion among domestic and foreign scholars since its discovery. It has become a hot topic and drew great attention in both theoretic and practical field.The earliest research on the momentum effect and reversal effect is Jegadeesh, Titman (1993) and De Bondt, Thaler (1985). Using monthly data from 1965 to 1989, Jegadeesh and Titman (1993) found that by building a zero-cost arbitrage portfolio which buys stock that had a good performance in the past and sales stock that had a poor performance in the past will have a significant positive return when holding period varies from three to twelve months. Besides, Jegedeesh, Titman (2001) used out of sample data to test the conclusion the drew in 1993, the result shows the positive return from momentum effect still exist which means the result did not come from data mining. In order to study if overreaction which comes from unexpected information and significant event will influence the price of stock, De Bondt and Thaler (1985) found that the loser portfolio which had poor performance in the past will exceed the winner portfolio which had good performance in the past in three to five years. Take the 36 months holding period portfolio as an example, the results showed that the return of loser portfolio is 25% higher than that of the winner portfolio. People can use zero-cost arbitrage portfolio to get positive return just by analyzing historical performance of stocks, that means weak form efficient market hypothesis is broken.During the literal review process of both home and abroad, we find that there are at least the following characteristics by comparing existing domestic research to foreign research:first time range of empirical research abroad is generally long, due to the early development of the foreign stock market, so their study sample is often richer than domestic study, and the great span of the time range also brought greater convenience for their research. Conversely, domestic research mainly based on data within a decade and the study sample is limited. Also due to frequent changes in policy, when comes to data acquisition we choose with blindness, but should do the data processing with clear target. Second, foreign scholars have reach a consensus that the momentum is in existence and there is no disagreement and ambiguity, this phenomenon is not only exist in the US, Even in many countries the same conclusion had been found such as the research done by Rouwenhorst (1998) studied 12 countries in Europe and Griffin, Martin (2003) selected 17 global representative markets. And some scholars have devoted to the analysis of robust test after friction, such as Korajczyk, Sadka (2004) proved that the momentum still exist after considering the market friction. But in the Chinese market, there is still no consensus on the existence of momentum effect, such as Wang Yonghong, Zhao Xuejun (2001) found that there is only overreaction phenomenon in Chinese stock market, Zhou Linjie (2002), and Huang Jun, Chen Ping (2009)regard that there is short-term momentum effect exists in China's stock market, and Lin Song Li, Tang Xu (2005) and Liu Bo, Pi Tianlei (2007)'research come to an opposite conclusion, which found that both short-term momentum effect and reversal effect does not exist, but found a long term reversal effect. Tan Xiaofen, Lin Yufei (2012) believes that the momentum effect exists in long term, but there is reversal effect exist in short-term. Lu Zhen, Zou Hengfu (2007) thinks that the reversal effect in China's stock market is stronger than the momentum effect. In short, we can find that no matter in the short-term or long-term, the existence of the momentum effect has not reach consensus in China. At the same time, we find that researches from both home and abroad had done some study on the characteristics of the momentum effect, such as momentum in different market conditions, in large and small companies and so on.Regard to the causes of momentum effect, although studies about causes of the momentum effect in domestic is as widely as that about existence of momentum, but scholars have not reached a common outlook which is the same with foreign studies. Disagreements still exist for the causes of the problem. Some people stand for the traditional risk compensation view, such as chordia, Tarun, and shivakumar (2002) found that by adding a series of macroeconomic variables that are related to business cycle for risk adjustment, return momentum disappeared, domestic studies such as Xiao Jun, Xu Xinzhong (2004) found that the FF three factor model had a strong explanation ability of momentum. Others stand for the view of behavioral finance theory. For instance, the BSV model established by Barberis, Shleifer and Vishny (1998), and the DHS model proposed by Daniel, Hirshleifer and subramanyam (1998),and Hong and Stein (1999) established the HS model, and in recent Da, Gurun, Warachka (2014) test a frog-in-the-pan(FIP) hypothesis. Also domestic research such as Chen Huanhua, Chen Rong, Zheng Zhenlong (2014), mainly focus on the anchoring deviation and disposition effect of behavioral finance to explaining the reason of momentum.As a typical emerging market, the stock market in China has some particularity compared with other developed countries, in addition to the differences of the government policies; the investor's behavior also has its own unique characteristics. By researching and studying the momentum in China, it is not only beneficial to explore the movement of stock price to verify whether the stock price can be predicted, but also to provide new thoughts for the analysis of investor's behavior patterns and psychological characteristics. Through the study of manifestations and characteristics of momentum, it has certain practical meaning to protect the interests of investors, and provide ideas to improve the development of our securities market. Thus in order to obtain a representative result, we choose stock data after January in 2006 considering the impact of stock ownership reform. We use the overlapping method which Jegadeesh and Titman (1993) adopted to build winner and loser portfolio. In research of the existence of momentum effect, this paper uses a rather long time span to study the existence of momentum effect in the short, medium and long period. In addition, despite of the single research of existence which can be seen in other studies of momentum, this paper will mainly focus on momentum effect features under different conditions, in other words, the study about what factors have an impact on momentum. Previous domestic research are concentrated in the main board market but ignored the phenomenon of other sectors, but considering the importance of SME and GEM board, we will analyze momentum in the differences among plates. Besides this, we will also examine other factors like momentum effect in different market, different size and different interval. Finally, this article will test whether risk compensation will explain the momentum effect by using the traditional CAPM and Fama-French three factor model to do risk adjustment. Besides, we will try to make some explanation from the perspective of behavioral finance.The results showed that in main board market in China, the momentum portfolios in short-term (1-4 months), midterm (5-11 months), or long-term (12-48 months is significant negative, shows reversal effect, and with the increase of holding period and formation the reversal effect has become increasingly significant. When compare the three different boards, we find that the reversal effect has the strongest performance in the main board, but momentum effect performed strong in the GEM board. and it is most significant in midterm for momentum profit is the biggest. When compare three different market conditions, we find in the bull market reversal effect is relatively weak, but in the bear market with the growth of holding period and the formation period, the reversal effect becomes significant, momentum effect perform strong in the adjustment stage. When compare the different company size, reversal can be found more strong in small companies than that in large companies, reflected both in the number of significant gains on the portfolio and in the absolute value of the negative returns. Finally, neither the CAPM model nor the FF three factor model can effectively explain the negative returns of the momentum portfolio, and the negative returns not only become bigger, but also very significant after risk adjusted excess. According to HS model, due to the frequent policy changes of our country, combined with investors are quiet sensitive to the policy information, investors motivation of using momentum strategies are reduced, thus the reversal effect is more obvious and the momentum profit is negative. Also because of the momentum traders pay more attention to small companies, the information of small companies is no longer valuable with the higher sensitivity, so small companies are more likely to show reversal than large companies. According to the DHS model, investors will determine which mechanism decides the profit, and in the bull market, a positive (negative) return will be accompanied by another positive (negative) return, after update the judge of which state they are with Bayes rule, they tends to think themselves in the trends mechanism and leads to the excessive reaction thus reversal effect. But in shock period, a positive (negative)return will be accompanied by another negative (positive) return, so investors will tend to judge themselves in the mechanism of regression, therefore cannot fully adjust their behavior thus lead to momentum. |
PDF Full Text Request