The momentum effects and market timing of
momentum strategies in Chinese stock market
Thesis supervisor: Prof.Dr.Wolfgang Bessler
1
The momentum effects and market timing of momentum strategies in
Chinese stock market
Abstract: This paper provides a basic analysis of the momentum effects on the Chinese stock
market and simulates the returns of momentum strategies within a one-year investment period, including the formation period and the holding period. There are some different features among the three submarket and reasons of this are complicated.
Keywords: Fama-French three factors model, momentum factors, return rate, momentum
2
1. Introduction
Momentum is one of the most well-known anomalies on most of stock markets in the world. It was first demonstrated by Jegadeesh and Titman (1993). They stated that many stocks have a sustained performance over the following one-year period and this finding was not the result of data mining. The momentum anomaly seems to be inconsistent with the efficient market hypothesis. However, a growing number of studies provide evidence that the “momentum effects” are widely observed in financial markets. This means that stock returns are affected not only by market returns but also by cumulative stock returns over the past period. In addition, The Fama-French three factors model is unable to interpret the considerable profits generated from momentum strategies. So Carhart (1997) extended the three-factor model by introducing a new “momentum” factor to capture the abnormal returns. The “momentum factor” is the acceleration rate of change of stock prices. It measures how past returns of stocks influence their future performance and becomes a valuable indicator which help investors to develop profitable investment strategies. On the one hand, if momentum factors is directly proportional to future returns, then investors are able to earn higher and abnormal profits from momentum investment strategies. In contrast, if momentum factors has a negative impact on future returns, investors should follow a reversal strategy, selling the winners of the previous period and buying the losers. The very interesting issue is that investors can formulate various strategies to build their portfolios based on different features of momentum effects.
Most literature demonstrates that a persistent momentum affects the short-term returns and that a reversal momentum affects the long-term returns. A research by Antoniou, Lam and Paudyal (2007) shows that the momentum effects can explain future stock returns in a stock portfolio analytical framework and help predict the persistence of mutual fund performance. Because of the momentum effects, many investors find that the momentum investment strategy of holding the past high return rate stocks and short selling the past low return rate stocks to be a profitable one. The belief in momentum strategies is that the past stock returns can influence the stocks’ future return therefore the past winners will continue to maintain and provide a higher return rate than the past losers in the short term. The profitability of the momentum strategies have been verified by Rouwenhorst (1998). From 1980 to 1995, he tested momentum strategies in the European financial markets by establishing various portfolios based on the belief of momentum effects. What is more, Hameed and Kusnadi (2002) also discovered the same phenomenon in Asian financial markets. When a momentum effect exists in the stock market, investors can obtain abnormal returns by analyzing the past performance of each stock and taking a long (short) position on winners (losers) if there are no short selling constraints.
3
effect is triggered by the overreactions of irrational investors and accelerated by herding effects. Both overreaction and herding effects are important factors in analyzing stock market anomalies and reflecting how investors react when facing new market information and information asymmetries and what actions they will take. Addoum et al. (2019) find that momentum profits may result from out-of-market information, such as the new political regime. Because some industries may get government support under specific governments, the stock price of those industries will rise once the relevant political information is reflected in the stock market. Other scholars have tried to explain momentum in different ways. For instance, Chordia et al. (2002) decompose momentum factors and find that momentum effects are decided by some lagging economic variables.
The majority of research focuses on the US stock market and the rest of regions is lacking some thoroughly investigations. Compared with western countries, China, the second-largest economy in the world, has attracted more and more attention because of its special political and economical characteristics. It has maintained rapid economic growth over the past 30 years and is expected to become a major contributor to the world economy in the future. The Chinese market has pronounced influence on the world and world markets. Mathews (2016) points out that the Australian financial market is strongly affected by the Chinese economy and Ge et al. (2019) find that there is a strong link between Chinese and New Zealand stock market. Nowadays, the Chinese economy is undergoing a huge transformation and the Sino-US trade war has also brought more challenges to the economic development of China. Nevertheless, it is worth investigating the momentum effects on the Chinese stock market and offer a deep insight of its characteristics. The most important purpose of this article is to improving the investors’ timing ability of momentum investment strategies.
The Chinese stock market is relatively immature compared to the western stock markets and it is still under the process of development. As of June, 8, 2019, there are in total 2,704 listed companies’ shares traded on the Shanghai Stock Exchange and the Shenzhen Stock Exchange. Carpenter et al. (2015, 2017) report that the Chinese stock market is less correlated with other stock markets in the world and provides high abnormal returns or alpha to its domestic investors. It also plays an important role in international financial markets and represents the strongest power among emerging markets. An understanding of the Chinese stock market is critical not only for domestic investors but also for the international investors who are looking for hedging and diversifying their risks.
4
important issue is that it may help both domestic and international investors find feasible momentum investment strategies on the Chinese stock market.
The rest of this paper is organized as follows. The section 2 discusses the momentum literature and its theory. The section 3 demonstrates the methodology and measurement of variables. The section 4 shows the empirical cross-section findings. The section 5 presents the simulation results of profits generated from momentum strategies. Conclusions and discussion of the results are presented in the last section.
2. Literature review
In 1993, Fama and French established a three-factors model using company size and book-to-market ratios to capture excess returns of firm stocks, and they showed that the three factors model has better explanatory power than the general capital asset pricing model (CAPM) for the US stock market. Gaunt, (2004) conducts a survey of the Australian stock market and comes up with consistent findings that adding book-to-market factors can significantly improve the accuracy of pricing models. However, previous studies have also provides substantial evidence against the Fama-French three-factors model in emerging markets. For instance, Drew et al. (2003) state that the cross-section of returns on Chinese stock market is not caused by the market factor. Zhang (2014) hold a different opinion that the Fama-French three-factors model is still applicable on the Chinese stock market because this model can explain 93% variation in Chinese stock portfolio returns. But they also present that there are some special features affect three factors and reduce the influence of those factors considerably. Wang and Xu (2004) also stated that unlike the book-to market effect, which is pervasive in the United States, it is insignificant in China. Narayan and X. Zheng (2010, 2011,2012) try to explain try to explain returns of Chinese stock market by liquidity and asymmetric information.
Although the explanatory power of the Fama-French three-factors model is recognized, it cannot explain the abnormal returns generated by holding the past well-performing stocks and short-selling the past poorly-well-performing stocks. The empirical evidence suggests that past transactions contain useful information that can help investors to achieve considerable investment returns. For instance, Chan and Lakonishok (1995) state in their study that past returns of a stock can predict its return over six to twelve months in the future. This global phenomenon is called momentum anomaly. Momentum runs counter to the EMH developed by Fama (1991), but it is robust economically and statistically.
5
the US stock market. Lee and Swaminathan (2000) find that the effects of price momentum, will be reversed in the next five years. Thaler (1985); De Bondt and THALER (1987, 1990) also provide evidence which is persistent with the findings that returns reversal in the long term. Besides, the continuation of medium-term returns also appears on the international equity markets.
In contrast, Jegadeesh and Titman (1993) document a different observation: the momentum effects are persisted in the medium-term and long-term horizons on the US stock market. They find that winners of the past three months up to one year continue to outperform losers in the same period. Daniel et al., (1997) finds that the mutual funds have a persistent performance over a short-term period of one to three years.
There are many reasons why researchers may find such different results. Lo and Mackinlay (1990) and Hameed and Kusnadi (2002) conclude that the definitions of short-term, mediate-term, and long-term periods are different in each paper. As a result, researchers may draw totally opposite conclusions from the same regression result. Secondly, the sample sizes are not the same. Moreover, investors among different markets have different personalities and each market has its own characteristics. They will behave differently even faced with the same situation. Because of those reasons, a large number of samples are needed to ensure the power of the research.
Although the discoveries of momentum are not exactly the same, it is reasonable to conclude that in most western countries, there is a positive momentum in a relative short-term period (within 6 months) and a negative momentum in relative longer period (about 3 years). There is also some evidence that came from the Chinese stock market. For instance, Kang, Liu and Ni (2002) find that the short-term momentum effects are negative, while the medium-term momentum effects are positive, based on data collected from the Chinese stock market.
Based on these findings, momentum strategies have been created and became popular not only for individual investors but also for portfolio managers. In previous research, both individual investors and institutional investors from western financial markets and the Chinese stock market have recorded the abnormal profits of momentum strategies. For example, Rouwenhorst (1998) reports that momentum investment strategies can generate profits in European equity markets. Kang, Liu and Ni (2002) find statistically significant abnormal profits on the Chinese stock market when they applied intermediate-horizon momentum strategies. Miffre and Rallis (2007) build their own momentum strategies in the commodity futures market and their results suggest that 13 momentum strategies provide 9.38% average return per year.
6
they find that momentum has a positive and significant correlation with high turnover, which is an important measure of overreaction behavior. Besides, Huang (2008); Brown and Cliff (2005) find that investors’ sentiment is pessimistic and unresponsive to market movements during a bear market. This means that the economic environment may influence the profitability of momentum strategies directly and indirectly.
Some scholars find that there is a positive relationship between momentum effect and stock returns during a bull markets but this relationship disappears during a bear markets. In order to analyze the impact of the macro-economic environment and look for better momentum strategies, a further study is required. Kiyotaki and Moore (1997) predicts that a change in credit market conditions have an impact on stock market performance and Moskowitz and Grinblatt (1999) conclude that industry factors can explain the profitability of momentum strategies. Chordia et al. (2002) presents that momentum can generate positive returns only when the business cycle is in an expansion phase. Addoum et al. (2019) suggest that investors can generate a higher return than normal momentum strategies by combining political environment and momentum strategies.
There are two hypotheses we investigate in this paper. The first hypothesis is to test whether momentum effects exist on the Chinese stock market and to provide evidence of the explanatory power of the momentum factors within the Fama-French three-factor model on the Chinese stock market.
Hypothesis 1: There is momentum effect existing in the Chinese stock market.
The second hypothesis focuses on the role of the macroeconomic environment in momentum investment strategies for the Chinese stock market. Momentum is sensitive to the macroeconomic environment. This paper divides the sample into different macroeconomic environments by using different macro-economic indicators as proxies.
Hypothesis 2: The momentum strategies based on macroeconomic environment generate more profits.
3. Research method and data
1. Data collection
7
Most of domestic investors are trade at the A share market and all transactions are paid in RMB, while foreign investors invest in the B share market in US dollars or Hong Kong dollars. Stocks in the A-shares and the B-shares are relatively more mature than stocks in the growth market and Cheung and Liu (2014) present that there is a large information asymmetry among investors in the growth market.
2. Summary statistics
Our sample contains 3,718 firms: 2,869 firms in the A-shares market; 108 firms in the B-shares market; 741 firms in the Growth market. It is an unbalanced data panel because there are constantly new companies entering the stock markets and old companies exiting the stock market based on their operating conditions and willingness to stay listed. Overall, there are 321,952 observations in our sample.
Several important indicators are measured as following:
𝑅𝑓,𝑡 is the risk-free rate at time t and the proxy of risk-free rate is the monthly one-year periodic fixed-deposit rate.
𝑅𝑖,𝑡 is the return adjusted for cash dividends at time t and calculated by the following
equation:
𝑅𝑖,𝑡 = (𝑃𝑖,𝑡+ 𝐷𝑖,𝑡) 𝑃⁄ 𝑖,𝑡−1− 1 (1),
where 𝑃𝑖,𝑡 is the closing price of i at time t and 𝐷𝑖,𝑡 is the cash dividends paid during the period from time t-1 to t.
ri,t = 𝑅𝑖,𝑡− 𝑅𝑓,𝑡 is the adjusted return on stock i at time t.
𝑅𝑀𝑅𝐹𝑡 measures the excess return rate of market portfolio, weighted by total market value. The Fama-French factors, 𝑆𝑀𝐵𝑡 and 𝐻𝑀𝐿𝑡, are replicated as small cap minus big cap and
high book-to-market minus low book-to-market. This paper first sort stocks into two groups based on their market value. Stocks which have market values higher than the median are regarded as big cap (B), and stocks which has a lower than average market value are in the small cap group (S). Then we divided each group by ranking book-to-market ratio and treating the 30% quantile and 70% quantile as breakpoint into three subgroups, L, M, H. After that, six portfolios are constructed (BL; BM; BH; SL; SM; SH). The variable SMB is the average return of S portfolios subtracting the average return of B portfolios. The variable HML is the average return of H portfolios minus L portfolios.
8
developing in a positive direction. If the confidence index is lower than 100, it indicates that the economy is developing in an unfavorable direction.
The second indicator is the OECD based Recession Indicators for China (RD). It takes the value of 1 during recessionary periods and 0 during expansionary period applying NBER standard. NBER defines that ‘A recession is a significant decline in economic activity spread across the economy, lasting more than a few months’. Recessions obviously affects the GDP, national income, the employment rate as well as investments.
Using two proxies should be meaningful, since the RD is an ex-post indicator while the BIM is an ex-ante indicator.
Table 1 presents the data description and correlation within three independent variables for sample from January 2008 to December 2018 in different submarkets.
Table 1. Descriptive Statistics.
The A-shares market
Variable Mean Std.Dev. Min Max Correlation
R .007 .173 -.851 22.051 RMRF SMB HML RMRF -.002 .077 -.254 .174 1.00
SMB .005 .039 -.101 .265 0.189 1.00
HML .001 .033 -.099 .096 -0.211 --0.016 1.00 The B-shares market
Variable Mean Std.Dev. Min Max Correlation
R .005 .127 -.617 1.842 RMRF SMB HML RMRF .002 .087 -.319 .265 1.00
SMB .011 .055 -.235 .242 0.338 1.00
HML -.001 .035 -.145 .159 -0.330 -0.715 1.00 The Growth market
Variable Mean Std.Dev. Min Max Correlation
R .015 .22 -.721 6.398 RMRF SMB HML RMRF .005 .105 -.298 .401 1.00
SMB .007 .03 -.058 .109 0.313 1.00
HML 0 .029 -.075 .096 0.161 0.049 1.00 Macro-economy indicators
Variable Mean Std.Dev. Min Max BIM 97.597 3.666 92.29 104
9
As Table 1 shows, stocks in the Growth market have the highest average return (0.15), although the maximum value of stocks return is generated at the A-share stock market (22.05). Generally, the Growth market have the highest market return (0.005) and average stock’ returns. The volatility of both individuals and market return of the Growth market is also higher than that of the A-shares and the B-shares stock markets, while the difference between the A-shares market and B-shares market is not significant. Another difference between the Growth market and other two markets is the positive correlation between the SMB and HML factors. Overall, the correlations between the three independent variables are moving in a relatively low level, except for the relation between SMB and HML in the B-shares market. The correlation coefficient of SMB and HML is -0.715, which shows a strong negative correlation. In order to keep the consistency and comparability of our regression models, we still keep the HML factors in our regression model when analyzing the B-shares market.
The macroeconomic indicator for BIM measurement has a mean value of 97,597, which indicates that there is a more unfavored macroeconomic environment. However, the mean (0.446) of RD measurement presents a different finding: There are more expansionary periods than recessionary periods. This means that there is a gap between expectation and reality. People have a conservative and relatively pessimistic estimation of the future.
3. Methodology 3.1 regression models
There is a substantial literature documenting systematic risk effects, size effects as well as book-to-market effects in US stock market. The focus of this research is to test the explanatory power of these effects in the Chinese stock market and capturing the momentum effects. This paper follows earlier studies by using both the Fama-French three-factor model and the Carhart four-factor model. The basic regression equation is:
𝑅𝑖,𝑡− 𝑅𝑓,𝑡 = 𝛼𝑖,𝑡+ 𝛽𝑖𝑅𝑀𝑅𝐹𝑡+ 𝜀𝑖,𝑡 (2)
𝑅𝑖,𝑡− 𝑅𝑓,𝑡 = 𝛼𝑖,𝑡+ 𝛽𝑖𝑅𝑀𝑅𝐹𝑡+ 𝑠𝑖𝑆𝑀𝐵𝑡+ ℎ𝑖𝐻𝑀𝐿𝑡+ 𝜀𝑖,𝑡 (3)
𝑅𝑖,𝑡− 𝑅𝑓,𝑡 = 𝛼𝑖,𝑡+ 𝛽𝑖𝑅𝑀𝑅𝐹𝑡+ 𝑠𝑖𝑆𝑀𝐵𝑡+ ℎ𝑖𝐻𝑀𝐿𝑡+ 𝛾𝑖𝑀𝑂𝑀𝑡−𝑘+ 𝜀𝑖,𝑡 (4) As explained earlier, the 𝑆𝑀𝐵𝑡 and 𝐻𝑀𝐿𝑡 is the proxy of the size effect and the
book-to-market effect, respectively. The variable 𝑀𝑂𝑀𝑡−𝑘 is constructed to measure the momentum, and the procedure of forming 𝑀𝑂𝑀𝑡−𝑘 is based on average returns over a K months formation
10
the gap between formation period and holding period. We sort and rank our sample by average returns during the k-formation period. Then the stocks in the top 30% is equally weighted to construct winner portfolio with an average return rate of Wt−k. The stock in the bottom 30% also equally weighted to construct loser portfolio with an average return rate Lt−k. The 𝑀𝑂𝑀𝑡−𝑘 is calculating as following:
{ 𝑊𝑡−𝑘 = 1 𝑘∑ 𝑟𝑖,𝑡−𝑘 𝑘 𝑘=1 𝑖𝑓 𝑟𝑖,𝑡−𝑘 𝑖𝑛 𝑡ℎ𝑒 𝑡𝑜𝑝 30% 𝐿𝑡−𝑘 = 1 𝑘∑ 𝑟𝑖,𝑡−𝑘 𝑘 𝑘=1 𝑖𝑓 𝑟𝑖,𝑡−𝑘 𝑖𝑛 𝑡ℎ𝑒 𝑏𝑜𝑡𝑡𝑜𝑚 30% (5) 𝑀𝑂𝑀𝑡−𝑘 = 𝑊𝑡−𝑘 − 𝐿𝑡−𝑘 (6)
where k denotes the length of formation period and takes values of 3,6,9,12.
Hypothesis 2 is based on the CAPM model, but we divide our sample into a “good” economic period and a “bad” economic period and simulate the return rate of momentum strategies in each period.
The first step is to construct a measure of the macroeconomic environment. This paper uses the BIM and RD as two proxies to determine the macroeconomic environment. Based on these two measurements, we are able to identify “good” macroeconomic periods and “bad” macroeconomic periods under each standard.
Then a dummy variable 𝑚𝑎𝑐𝑟𝑜𝑡𝐺 is added to the CAPM model and the equation is as
following:
𝑟𝑖,𝑡− 𝑟𝑓,𝑡 = 𝛼𝑖+ 𝛽𝑖(𝑟𝑚𝑘𝑡,𝑡− 𝑟𝑓,𝑡)+ 𝜃𝑖𝑚𝑎𝑐𝑟𝑜𝑡𝐺+ 𝜀𝑖,𝑡 (7)
where
𝑚𝑎𝑐𝑟𝑜𝑡𝐺 is equal to 1 when the macro-economy is good and zero when the macro-economy is bad. The 𝜃𝑖 estimation can capture the sensitivity of individual stocks to the macroeconomic environment. In this regression model, we use time series regression to measure the 𝜃𝑖 estimation. A positive 𝜃𝑖 estimate indicates that stocks earn higher returns
during good economic periods. Based on the value of 𝜃𝑖, we sort individual stocks in descending order to form macroeconomic sensitive portfolios and un-sensitive portfolios.
3.2 momentum strategy construction
11
based momentum strategies consists of holding winner stocks and macroeconomic sensitive and selling stocks that are losers and macroeconomic un-sensitive. As stated early, there is a one-month gap between the formation period and the holding periods.
By comparing the returns generated from two momentum strategies, we are able to test the second hypothesis. To verify the profitability of macroeconomic environment based momentum strategies, a reverse test can be done by holding a long position on winner stocks that belong to un-sensitive portfolios and holding a short-position on losers stocks that belong to sensitive portfolios.
4. Empirical cross-section results
1. CAPM and Fama-French model
Table 2 presents the regression result of the full sample. The finding is that the market factor is one key driver of cross-sectional returns in China, although it is not enough to explain all the variation but it is consistent with the finding of Drew et al. (2003). The size effects and book-to-market effects are significant, although the effect of book-to-market ratio is found to be negative and relative smaller than the effects of the market and the size ratios. This is in contrast to the finding of Wang and Xu (2004) who conclude that the effects of book-to-market is insignificant.
Table 2 : Regression results of full sample
Full sample CAPM Fama-French three model RMRF 1.123c 1.064c (0.003) (0.003) SMB 0.579c (0.007) HML -0.254c (0.008) Alpha 0.009c 0.006c (0.000) (0.000) Obs. 321952 321952 R-squared 0.265 0.283
12
Table 3: Regression results of submarket
The A-shares Market The B-shares Market The Growth Market CAPM Fama CAPM Fama CAPM Fama RMRF 1.130c 1.066c 1.046c 0.994c 1.117c 1.068c (0.004) (0.004) (0.009) (0.009) (0.009) (0.009) SMB 0.636c 0.283c 0.412c (0.007) (0.020) (0.032) HML -0.358c 0.063b 0.287c (0.009) (0.031) (0.032) Alpha 0.009c 0.006c 0.003c 0.000 0.009c 0.007c (0.000) (0.000) (0.001) (0.001) (0.001) (0.001) Obs. 266715 266715 13127 13127 42110 42110 R-squared 0.253 0.278 0.516 0.527 0.286 0.290 Note: the suffix a, b, c denotes the significance level of 10%, 5%, 1% respectively.
13 2. Momentum effects
Table 4: Regression results of full sample
K=3 K=6 K=9 K=12 RMRF 1.089c 1.084c 1.080c 1.073c (0.003) (0.003) (0.004) (0.004) SMB 0.576c 0.570c 0.583c 0.586c (0.007) (0.007) (0.007) (0.007) HML -0.213c -0.218c -0.207c -0.191c (0.008) (0.008) (0.008) (0.009) MOM(K) 0.123c 0.221c 0.260c 0.145c (0.009) (0.017) (0.023) (0.031) Alpha -0.010c -0.014c -0.013c -0.001 (0.001) (0.001) (0.002) (0.001) Obs. 315301 310068 304590 298966 R-squared 0.288 0.281 0.265 0.262 Note: the suffix a, b, c denotes the significance level of 10%, 5%, 1% respectively.
The momentum effects are calculated for different holding periods, K=3, 6, 9, 12. The variable MOM(K) calculated as the difference between the average return rate of winner portfolios which constructed by the stocks that are in the top of the 30% highest return during the K month period from t-k to t-1 and the average return rate of loser portfolio which contains stocks that have the lowest 30% of returns during K month period.
Table 4 shows that MOM (K) is significant with explanatory power in the Chinese stock market, although the explanatory power seems declining with the increase of the length of formation period. By adding the momentum factors, the absolute values of beta for the market risk and for the HML factors is decreasing, while the decrease of HML factor’s beta is more significant than the decrease of the market risk’s beta. The value of the size factor betas is consistent around 0.58. The momentum factor with 9-months formation periods has the highest beta value among the momentum factors with a K months formation period, K=3, 6, 9, 12. In contrast, the momentum factor with a 3 months formation period has the lowest beta value. Most of the coefficients are significant at the 1% level.
14
Table 5: Regression results of submarket
Panel A
K=3 K=6
A-shares B-shares Growth A-shares B-shares Growth
RMRF 1.099c 0.992c 1.068c 1.091c 0.985c 1.076c (0.004) (0.010) (0.009) (0.004) (0.010) (0.009) SMB 0.637c 0.288c 0.394c 0.630c 0.285c 0.351c (0.007) (0.022) (0.032) (0.008) (0.023) (0.035) HML -0.315c 0.064a 0.311c -0.317c 0.050 0.293c (0.009) (0.033) (0.032) (0.009) (0.035) (0.032) MOM(K) 0.090c 0.054a 0.169c 0.195c -0.038 0.266c (0.011) (0.028) (0.020) (0.020) (0.053) (0.043) Alpha -0.006c -0.004 -0.017c -0.011c 0.003 -0.018c (0.001) (0.003) (0.003) (0.002) (0.004) (0.004) Obs. 260906 12706 41689 256414 12393 41261 R-squared 0.284 0.523 0.291 0.276 0.521 0.290 Panel B K=9 K=12
A-shares B-shares Growth A shares B-shares Growth
RMRF 1.086c 1.000c 1.065c 1.076c 1.012c 1.066c (0.004) (0.012) (0.009) (0.004) (0.011) (0.010) SMB 0.642c 0.298c 0.406c 0.640c 0.304c 0.393c (0.008) (0.024) (0.033) (0.008) (0.024) (0.034) HML -0.307c 0.073b 0.281c -0.289c 0.088b 0.336c (0.009) (0.035) (0.033) (0.009) (0.036) (0.034) MOM(K) 0.187c -0.057 0.314c 0.074b 0.005 0.502c (0.026) (0.063) (0.061) (0.035) (0.066) (0.088) Alpha -0.007c 0.003 -0.017c 0.002 0.000 -0.015c (0.002) (0.004) (0.005) (0.001) (0.002) (0.004) Obs. 251804 12072 40714 247162 11755 40049 R-squared 0.256 0.477 0.289 0.252 0.475 0.289 Note: the suffix a, b, c denotes the significance level of 10%, 5%, 1% respectively.
15
(6) and MOM (9) have insignificant negative values in the Bshare market of 0.038 and -0.057, respectively. Secondly, the B-share market seems not sensitive to the momentum effects. Only the coefficient of MOM (3) in the B-shares market is significant at 90% confidence level. The effects of the book-to-market factors in the B-share market is also not as remarkable as that of other markets. Third, the size effects are more pervasive than momentum and book-to-market effects in the A-share market and the B-share market, while the momentum effects have larger influence on the Growth market than on the A-share market and the B-share market. The betas of MOM (K) also show that the longer of the formation period, the larger the effects.
3. Momentum effects and quantile regression
16
Table 6: Quantile regression results of MOM (3)
The A-shares Market
Low-performing stocks Media-performing stocks High-performing stocks
RMRF 0.999c 1.031c 1.053c 1.071c 1.088c 1.107c 1.129c 1.159c 1.211c (0.005) (0.004) (0.003) (0.003) (0.003) (0.004) (0.004) (0.005) (0.008) SMB 0.381c 0.464c 0.520c 0.566c 0.610c 0.657c 0.714c 0.791c 0.923c (0.012) (0.009) (0.008) (0.008) (0.008) (0.009) (0.011) (0.013) (0.019) HML -0.451c -0.407c -0.377c -0.352c -0.329c -0.304c -0.273c -0.233c -0.162c (0.011) (0.009) (0.008) (0.007) (0.008) (0.008) (0.010) (0.013) (0.018) MOM -0.518c -0.321c -0.189c -0.079c 0.026c 0.139c 0.273c 0.456c 0.771c (0.015) (0.012) (0.010) (0.010) (0.010) (0.011) (0.013) (0.017) (0.024)
The B-shares Market
Low-performing stocks Median-performing stock High-performing stock
RMRF 0.874c 0.915c 0.941c 0.963c 0.983c 1.004c 1.029c 1.061c 1.122c (0.016) (0.012) (0.010) (0.010) (0.010) (0.011) (0.013) (0.016) (0.022) SMB 0.174c 0.214c 0.239c 0.260c 0.280c 0.300c 0.324c 0.356c 0.414c (0.034) (0.026) (0.022) (0.021) (0.022) (0.024) (0.027) (0.034) (0.049) HML -0.015 0.012 0.030 0.045 0.058a 0.072b 0.089b 0.111b 0.152b (0.051) (0.039) (0.034) (0.032) (0.032) (0.035) (0.041) (0.051) (0.073) MOM -0.067 -0.025 0.001 0.024 0.045a 0.066b 0.091c 0.124c 0.186c (0.042) (0.032) (0.028) (0.026) (0.027) (0.029) (0.034) (0.042) (0.061)
The Growth Market
Low-performing stocks Median-performing stocks High-performing
17
The coefficients of RMRF show the same trend in the B-share market, the Growth market as well as in the A-shares market. They are positive, significant and upward from low-performing stocks to high-low-performing stocks. However, the characters of coefficients of the HML factor and the MOM factor in the B-shares market and the Growth market are different from A-shares market. For the B-shares market, the HML effects and momentum effects are only significant at stocks which have higher than average returns. The coefficients of these two factors keep a relatively small positive value, below 0.2, compared to the coefficient in the A-share market, although they are increasing with the stocks’ performance. The effects of the size factors in the B-shares market is also smaller than that in the A-shares market. For the Growth market, the size effects are higher than for the B-share market’s size effects but lower than the A-share market’s. The HML effects in the Growth market effects reveal a huge difference. They are more marked and positive. Its’ coefficient takes the value of 0.726 at the 90% quantile ranked by stock returns, comparing to -0.162 and 0.152 at the A-share market and the Growth market in the same quantile, respectively. It is even higher than the size factor in the high performance stocks, which takes values of 0.663 at the Growth market. The coefficients of MOM in the Growth market shows a similar distribution in the A-shares market. There is a significant negative momentum effect among low-performing stocks and a positive momentum effect among high-performing stocks, but the negative momentum effects among low-performing stocks is smaller in the Growth market comparing to the A-shares market while the positive momentum effects among high-performing stocks in the Growth market is higher in the A-share market.
18
Table 7: Coefficient list of Momentum factors
The A-shares Market
Low-performing Median-performing High-performing
MOM6 -0.832c -0.497c -0.275c -0.092c 0.087c 0.278c 0.504c 0.812c 1.346c (0.026) (0.020) (0.017) (0.017) (0.017) (0.019) (0.023) (0.029) (0.042) MOM9 -0.771c -0.458c -0.252c -0.082c 0.085c 0.262c 0.474c 0.762c 1.265c (0.036) (0.028) (0.024) (0.023) (0.023) (0.025) (0.029) (0.037) (0.053) MOM12 -0.946c -0.610c -0.392c -0.211c -0.034 0.154c 0.378c 0.683c 1.212c (0.042) (0.032) (0.028) (0.026) (0.027) (0.030) (0.036) (0.046) (0.066)
The B-shares Market
Low-performing Median-performing High-performing
MOM6 -0.355c -0.246c -0.177c -0.118b -0.062 -0.006 0.062 0.150a 0.317c (0.083) (0.063) (0.055) (0.052) (0.053) (0.058) (0.068) (0.084) (0.121) MOM9 -0.131 -0.106 -0.090 -0.076 -0.062 -0.049 -0.033 -0.012 0.026 (0.092) (0.070) (0.061) (0.058) (0.059) (0.064) (0.075) (0.093) (0.133) MOM12 0.032 0.022 0.017 0.012 0.007 0.002 -0.004 -0.011 -0.025 (0.098) (0.074) (0.065) (0.061) (0.063) (0.069) (0.081) (0.102) (0.145)
The Growth Market
Low-performing Median-performing High-performing
MOM6 -0.837c -0.480c -0.262c -0.076a 0.114b 0.327c 0.585c 0.950c 1.573c (0.067) (0.051) (0.045) (0.044) (0.046) (0.053) (0.064) (0.085) (0.124) MOM9 -0.755c -0.415c -0.197c -0.018 0.165c 0.371c 0.618c 0.970c 1.571c (0.090) (0.068) (0.060) (0.058) (0.062) (0.070) (0.086) (0.113) (0.165) MOM12 -1.730 -1.028 -0.574 -0.194 0.189 0.621 1.146a 1.874c 3.185c (2.209) (1.827) (1.581) (1.375) (1.168) (0.937) (0.659) (0.305) (0.525) Note: the suffix a, b, c denotes the significance level of 10%, 5%, 1% respectively.
4. Momentum effects and rolling window regression
19
Graph 1. The MOM(K) in the A-shares market
Graph 2. The MOM(K) in the B-shares market
Graph 3. The MOM(K) in the Growth market
Note: The trade month begin with 1, which means the January 2008 and end up with 132, the November 2018.
As the three graphs represent, the average value of MOM (K) is fluctuating with time. In most of the periods, the MOM (3) have the highest average value in the four momentum factors and the average value is decreasing with the increase of K. The peak value occurs between 2014-2016 among the three markets. The result of MOM (K) in rolling window regressions shows that the majority of betas take values between [-2, 2] and the betas of MOM (12) are the most volatile and take some extreme values.
5. The difference between the A-shares market and the B-shares market
A-20 share market and the B-share market in China.
The difference between the A-share market and the B-share market has attracted some scholars’ attention. Chui and Kwok (1998) investigate the cross-autocorrelation between the A-share market and the B-share market by assuming that there is information asymmetry in these two markets. They suggest that foreign investors may receive information faster than the domestic investors. Therefore, the stock price of the B-share market will change faster than that of the A-shares market. Chiang, Tan and Li (2007) report that the herding behavior occurs in both the A-share and the B-share markets and the impact of herding is more obvious in the A-share market because investors in the A-share market have very limited knowledge about the stock markets while the typical investors of the B-shares market are knowledgeable and experienced foreign and institutional investors. Yan et al. (2012); Lien and Zhang (2015) investigate the relationship between the herding effect and momentum effects at the institutional level and they find that herding effects negatively influences the profits generated by the momentum strategies.
In order to investigate what may cause those different momentum effects, we perform additional tests using the overlapping stocks of the A-share market and the B-share market. In 2008-2018, there are 89 firms’ stocks that can be traded both in the A shares market and the B-shares market. In detail, there are 42 firms listed in the Shenzhen Stock Exchange and the rest of dual-list stocks are traded at Shanghai Stock Exchange.
21
Table 8: regression results of overlapped stocks in the Chinese stock market
Non-overlapped stocks in the A-shares market Overlapped stocks in the A-shares market Non-overlapped stocks in the B-shares market Overlapped stocks in the B-shares market RMRF 1.100c 1.081c 1.081c 0.972c (0.004) (0.015) (0.027) (0.011) SMB 0.648c 0.370c 0.506c 0.240c (0.008) (0.030) (0.059) (0.023) HML -0.316c -0.277c -0.047 0.088b (0.009) (0.034) (0.090) (0.035) MOM3 0.085c 0.205c -0.177b 0.100c (0.011) (0.043) (0.079) (0.030) Alpha -0.005c -0.020c 0.015a -0.008c (0.001) (0.005) (0.008) (0.003) R-squared 0.282 0.366 0.540 0.525 Note: the suffix a, b, c denotes the significance level of 10%, 5%, 1% respectively.
There are some interesting findings. First of all, only the non-overlapped stocks at the B-shares market take a negative value of the coefficient of MOM (3). Then, even if the stock of the same firm shows a certain degree of difference in the two different markets. The table above means that the different momentum effects between the A-share market and the B-share market is not only caused by the difference between the individual firms trading in two different markets, but also caused by other reasons.
The table below shows the difference in price, volume, turnover as well as the monthly individual stocks’ return rate between the two markets. The price ratio is calculated by the price of a dual-listed stock in the A-shares stock market dividing by the price of the same stock in the B-shares market. The Shares ratio is calculated in the same way using the data of transaction volume. The dif-return and dif-turnover refer to the difference between the returns and turnovers obtained by subtracting the B-share market data from the A-share market data.
22
market. In most cases, the trading frequency of the A-shares market is higher than the trade frequency of the B-share market and the Shanghai B-share market is more active than Shenzhen B-share market. There is not a huge difference on the monthly return rate between these two markets.
Table 9: comparison between the A-shares market and the B-shares market
Indicators
Mean Standard deviation. Min Max
SH SZ SH SZ SH SZ SH SZ
Price ratio 13.667 1.693 4.318 .498 4.559 .578 38.237 6.593
Shares ratio 14 17.554 27.132 26.39 .006 .161 479.655 1007.507
Dif-return .002 0 .103 .093 -.816 -.553 1.283 2.214
Dif-turnover 17.717 20.119 25.941 25.487 -51.759 -55.345 406.734 288.515
Table 10: momentum effects of overlapped stocks
Panel A
Shenzhen Shanghai
A-shares B-shares A-shares B-shares RMRF 1.077c 0.958c 1.084c 0.985c (0.020) (0.016) (0.022) (0.014) SMB 0.384c 0.128c 0.356c 0.342c (0.041) (0.035) (0.044) (0.031) HML -0.211c -0.024 -0.338c 0.190c (0.047) (0.054) (0.050) (0.047) MOM3 0.155c 0.080a 0.252c 0.119c (0.058) (0.046) (0.062) (0.039) Alpha -0.013a -0.004 -0.026c -0.012c (0.006) (0.005) (0.005) (0.004) R-squared 0.389 0.493 0.348 0.556
Panel B coefficient list of Mom(3) of overlapped stocks in two markets
Shenzhen
Low price ratio High price ratio Low shares ratio High shares ratio Low dif-turnover High dif-turnover
A B A B A B A B A B A B
23
Low price ratio High price ratio Low shares ratio High shares ratio Low dif-turnover High dif-turnover
A B A B A B A B A B A B
0.190b 0.106a 0.315b 0.168b 0.285a 0.136b 0.163 0.102 0.246b 0.108 0.372c 0.135b Note: the suffix a, b, c denotes the significance level of 10%, 5%, 1% respectively. The A means the result of the A-shares market and the B means the result of the B-shares market.
Compared with stocks traded on the Shenzhen Stock Exchange, stocks traded on the Shanghai Stock Exchange are more sensitive to the momentum factor. Sorting stocks by the value of the price ratio, the shares ratios and differences of turnover rates, we find that stocks which have high price ratios are more sensitive to momentum factor and when the difference between price increases, the difference between coefficient also increases. There is not a clear pattern in the Dif-turnover indicator groups and share ratio groups. The discounted price of stocks in the B-share market might be partial reason for the differences in momentum effects between the B-shares market and the A-shares market. The investors’ behavior behind the discounted price need more investigation in the future.
5. Momentum strategies construction
1. Simulation of momentum strategy
To analyse the relation between the macro-economic environment and the momentum, we first construct 3 portfolios: a momentum portfolio, a macro-economic sensitive portfolio, and a sensitive consistency momentum portfolio. Then we compare the profits and volatility of two momentum strategies for the three submarkets.
. At the beginning of every month, stocks are sorted to winner groups or loser groups based on their past three-month accumulated returns and momentum portfolios contain the winners and the losers. The normal momentum strategy is constructed by equally holding the winner group’s stock and selling the loser group’s stock for the next 6 months. Then we calculate the return of all momentum portfolios and record the main characteristics of the returns.
The way to construct sensitive measurement is to rank the sensitive measurement at the beginning of every month. The sensitive measurement is based on the value of θi. To ensure the measurement is directly related to “good” economy periods, the θi= -θi, when the economy is in a “bad” state. The state of the “good” or “bad” economy is defined by the two proxies, RD and BIM.
24
Graph 4. The simulation result of momentum strategy at the A-shares market
Graph 5. The simulation result of momentum strategy at the B-shares market
Graph 6. The simulation result of momentum strategy at the Growth market
Note: The trade month begin with 1, which means the January 2008 and end up with 132, the November 2018.
25
Table 11. the return of two momentum strategies.
Strategy Market type Mean Std. Dev. Min Max
MSR A-shares .241% 10.485 -39.376% 20.794% B-shares 5.527% 15.511 -32.799% 57.216% Growth 2.454% 9.984 -21.47% 27.424% CMSR Under Macro1 A-shares .365% 15.795 -60.714% 181.241% B-shares 2.125% 17.882 -112.219% 92.557% Growth -.212% 8.584 -32.425% 40.432% CMSR Under Macro2 A-shares 10.537% 21.444 -37.323% 114.454% B-shares 12.139% 21.528 -36.140% 71.312% Growth 9.128% 27.290 -58.577% 136.090%
The table presents the results of two strategies in the Chinese stock market. The macroeconomic environment consistent strategies do help investors to gain higher profits since the average returns of all three markets is increasing. However, playing such a consistent strategy would be riskier. Although the maximum profits increase significantly, the standard deviation and the maximum loss also increase under the macroeconomic environment consistent strategy. Interestingly, the portfolio returns at the A-shares market generated the highest value when the sensitive measurement is an ex-ante indicator while the portfolio returns at the Growth market generated the highest value when the sensitive measurement is an ex-post indicator. Overall, the macroeconomic environment consistent strategies create higher profits than the normal momentum strategy and show more extreme values. The returns of the B-shares market does not increase as much as that of the A-shares and Growth market since it is less sensitive to the macro-economy. We also test other momentum strategies by changing the length of formation period and holding period. The result is listed in the Appendix.
2. Momentum strategy and short sell constraint
26
Table 12. momentum strategies with short selling constraint
Return
rate Market type Mean Std. Dev. Min Max
A-shares 4.1% 21.1 -40.1% 76.5%
B-shares 10.8% 24.2 -36.7% 119.2%
Growth 4.3% 25.8 -38.5% 107.2%
The table above shows the returns for the normal momentum strategy with short selling constraints included. Compared to Table 11, we can find that investors can generate higher profits by only holding winners without selling losers. The maximum returns increase significantly and the minimum returns also slightly increases. This means that most of past winner still keep a good performance in the next 6 months while most of past losers get rid of the downward trend at the following period and begin to create positive returns.
6. Conclusion
27
Reference
Addoum, J. M., Delikouras Stefanos, Ke Da, Kumar Alok, 2019. Underreaction to political information and price momentum. Financial Management 48(3), 773–804.
Antoniou, A., Lam, H. Y. T., Paudyal K., 2007. Profitability of momentum strategies in international markets: The role of business cycle variables and behavioural biases. Journal of Banking and Finance 31(3), 955–972.
Bondt, W. De, Thaler, R., 1990. Do security analysts overreact? The American Economic Association 80(2), 52-57.
DeBondt, Werner F.M., Richard H.Thaler, 1985. Does the stock market overreact? Journal of Finance 40, 793-805.
De Bondt, W. F. M., Thaler R. H., 1987. Further evidence on investor overreaction and stock market seasonality. The Journal of Finance 42(3), 557–581.
Brown, G. W., Cliff, M. T., 2005. Investor sentiment and asset valuation. Journal of Business, 78(2), 405–440.
Carhart, M. M., 1997. On persistence in mutual fund performance. The Journal of Finance, 52(1), 57–82.
Carpenter, J. N., Lu Fangzhou, Whitelaw R.F., 2015. The real value of China’s stock market. NBER Working paper.
Carpenter, J. N., Whitelaw R. F., 2017. The development of China’s stock market and stakes for the global economy. Annual Review of Financial Economics, 9, 233-257.
Chan, L. K. C., Jegadeesh, N., Lakonishok, J., 1999. The profitability of momentum strategies. Financial Analysts Journal, 55(6), 80–90.
Chan, L. K. C., Lakonishok J., 1995. The behavior of stock prices around Institutional trades. The Journal of Finance, 50(4), 1147–1174.
Cheung, W., Liu, K., 2014. A comparison of China’s main board and growth enterprise market board - Market microstructure approach. Review of Pacific Basin Financial Markets and Policies. 17(2).
Chiang, T. C., Tan, L., Li, H., 2007. Empirical analysis of dynamic correlations of stock returns: Evidence from Chinese A-share and B-share markets. Quantitative Finance, 7(6), 651–667.
Chordia, T., Shivakumar Lakshmanan, 2002. Momentum, business cycle, and time-varying expected returns. Journal of Finance 57(2), 985-1019.
Chui, A. C. W., Kwok, C. C. Y., 1998. Cross-autocorrelation between A shares and B shares in the chinese stock market. Journal of Financial Research, 21(3), 333–353.
28
Daniel, K., Grinblatt. M., Titman. S., Wermers. S., 1997. Measuring mutual fund performance with benchmarks’, the Journal of Finance, 52(3), 1035–1058.
Daniel, K., Hirshleifer, D., Subrahmanyam, A., 1998. Investor psychology and security market under- and overreactions. Journal of Finance, 53(6), 1839-1885.
Drew, M., Naughton, T., Veeraraghavan, M., 2003. Firm size, book-to-market equity and security returns: Evidence from the Shanghai Stock Exchange. Australian Journal of Management, 28(2).
Fama, E., 1991. Efficient Capital Markets: II. The Journal of Finance, 46(5), 1575–1617. Huang Dayong, 2006. Market states and international momentum strategies. The Quarterly Review of Economics and Finance, 46(3), 437-446.
Gaunt, C., 2004. Size and book to market effects and the Fama French three factor asset pricing model: Evidence from the Australian stockmarket. Accounting and Finance, 44(1), 27–44.
Ge Yongbo, Zou, L., Wu, J., Zhang, J., 2019. The role of Chinese economic variables on Australian and New Zealand equity returns. Accounting and Finance, 2–47.
Hameed, A., Kusnadi, Y., 2002. Momentum strategies: Evidence from pacific basin stock markets. Journal of Financial Research, 25(3), 383–397.
Huang, C., Goo Yeong-Jia, 2008. Are happy investors likely to be overconfident? Emerging Markets Finance and Trade, 44(4), 33-39.
Jegadeesh, N., Titman, S., 1993. Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency, The Journal of Finance, 48, 65-91.
Kang, J., Liu, M. H., Ni, S. X. 2002. Contrarian and momentum strategies in the China stock market: 1993-2000. Pacific Basin Finance Journal, 10(3), 243–265.
Kiyotaki, N. and Moore, J., 1997. Credit cycles. Journal of Political Economy, 105(2), 211– 248.
Lee, C. M. C., Swaminathan, B., 2000. Price momentum and trading volume. Journal of Finance, 55(5), 2017–2069.
Lien, D., Zhang, H. 2015. Industry Herding and Momentum Strategies. Pacific-Basin Finance Journal, 32, 95-110.
Lo, A., Mackinlay Craig, 1990. An econometric analysis of nonsychronous trading. Journal of Econometrics, 45(1-2), 181-211.
Mathews, T., 2016. The Effect of Chinese Macroeconomic News on Australian Financial Markets. RBA Bulletin, 53–62.
Miffre, J., Rallis, G., 2007. Momentum strategies in commodity futures markets. Journal of banking and finance 31(6), 1863-1886.
29 54(4), 1249–1290.
Narayan, P., Zheng Xinwei, 2011. The relationship between liquidity and returns on the Chinese stock market. Journal of Asian Economics, 22(3), 259-266.
Narayan, P., Zheng Xinwei, 2010. Market liquidity risk factor and financial market anomalies: Evidence from the Chinese stock market. Pacific-Basin Finance Journal, 18(5), 509-520. Narayan, P., Zheng Xinwei, 2012. Asymmetric information and market decline: Evidence from the Chinese market. Review of Pacific Basin Financial Markets and Policies, 15(3). Peng, C. L., Lee Jen-Sin, Chan, K. C., Shyu, S., 2010. Momentum effect of high-and low-performing stocks in Taiwan: A quantile regression analysis. International Research Journal of Finance and Economics, 52, 104-112.
Rouwenhorst, K. G., 1998. International Momentum Strategies. The Journal of Finance, 53(1), 267–284.
Wang, F., Xu, Y., 2004. What determines Chinese stock returns? Financial Analysts Journal, 65–77.
Wermers Russ, 2000. Mutual fund performance: An empirical decomposition into stock-picking talent, style, transactions costs, and expenses. The Journal of Finance, 55(4), 1655-1695.
Yan Zhipeng, Zhao Yan, Sun Libo, 2012. Industry herding and momentum. The Journal of Investing Spring, 21(1), 89-96.
30
Appendix
Table 1. Data description
Variable Mean Std.Dev. Min Max
R .008 .181 -.851 22.051 RMRF 0 .075 -.262 .175 SMB .008 .05 -.221 .229 HML .002 .036 -.144 .16 RMW -.001 .027 -.08 .1 CMA .003 .02 -.061 .054
The table above presents the basic information about Fama-French five factors of the A-shares market and the Growth market. The B-A-shares market are not included because of data available.
Table 2. the regression result of Fama-French five factors model on the A-shares market and the Growth market
K=3 K=6 K=9 K=12 RFRM 1.019c 1.027c 1.023c 1.023c 1.021c (0.004) (0.004) (0.004) (0.005) (0.005) SMB 0.900c 0.864c 0.855c 0.861c 0.855c (0.012) (0.013) (0.013) (0.014) (0.013) HML -0.126c -0.151c -0.168c -0.156c -0.162c (0.014) (0.014) (0.015) (0.015) (0.015) RMW -0.005 -0.019 -0.021 -0.038a -0.034 (0.021) (0.020) (0.021) (0.021) (0.021) CMA -0.005 -0.015 -0.017 -0.068c -0.059c (0.018) (0.018) (0.018) (0.018) (0.019) MOM(K) 0.070c 0.057c 0.048b 0.111c (0.009) (0.018) (0.024) (0.031) Alpha 0.001c -0.008c -0.004b -0.002 -0.003b (0.000) (0.001) (0.002) (0.002) (0.001) R-squared 0.288 0.289 0.282 0.267 0.262
Note: the suffix a, b, c denotes the significance level of 10%, 5%, 1% respectively.
31
significant in Fama-French five factor model. All of momentum factors have positive betas. In more detail, the one year momentum has the highest beta value (0.111). As for short-term momentum factor (less than one year period), the three month momentum factor has a higher beta (0.075) than other two momentum factors (0.057 and 0.048, respectively).
Table 3. the simulation results of momentum strategies
The A-share market
Formation-holding 1-1 1-3 1-6 3-1 3-3 3-6 6-1 6-3 Mean -.003 -.005 -.006 -.002 -.006 .004 -.003 .001 Min -.103 -.214 -.289 -.121 -.234 -.41 -.109 -.207 Max .138 .236 .292 .235 .177 .285 .123 .195 Std.Dev .037 .061 .088 .045 .068 .105 .042 .067
The B-shares market
Formation-holding 1-1 1-3 1-6 3-1 3-3 3-6 6-1 6-3 Mean -.003 -.002 .013 -.002 .005 .036 .007 .031 Min -.114 -.197 -.277 -.149 -.309 -.393 -.116 -.16 Max .125 .191 .313 .307 .283 .413 .181 .372 Std.Dev .042 .076 .117 .054 .093 .135 .056 .104
The Growth market
Formation-holding 1-1 1-3 1-6 3-1 3-3 3-6 6-1 6-3 Mean -.002 .001 .011 .005 .019 .034 .006 .019 Min -.129 -.221 -.215 -.146 -.28 -.264 -.135 -.218 Max .144 .219 .294 .145 .189 .263 .138 .263 Std.Dev .045 .072 .097 .049 .074 .089 .053 .077
32
market provided relatively higher returns compared to the A-shares market and the Growth market. It is worth mentioning that the higher return rates always coming with the higher standard deviation, which means the momentum strategies is risky. The highest simulation return rate of the Growth market is 29.4%, which is higher than 29.2% of the A-shares market but lower than 41.3% of the B-shares market.
