“ The January effect: Evidence from Chinese markets ”

1

Faculty of Economics and Business

MSc Finance

Master Thesis

“

The January effect: Evidence from Chinese markets

”

By

Yehan Wang (S2370174)

Supervised by:

Dr Henk von Eije June, 2014

Esdoornlaan 838 9741 ML Groningen

2

The January effect: Evidence from Chinese market

Abstract: I examine the January effects in China with indexes and individual firms data of the

Shanghai and the Shenzhen stock markets respectively. I find that the Chinese B share indices show a significant January effect. Moreover, all four segments of the Chinese stock markets show significant January effects with firms data. Size effects also exists in most of the Chinese stock market segments.

Key words: January effect, Chinese Stock Market, A shares, B shares JEL classification: C12, G14

1. Introduction

Based on the Efficient Market Hypothesis (EMH), stock prices should follow a random walk according to traditional financial theory. However, more and more domestic and international empirical studies found monthly or seasonality effects in the stock markets; one of these is the so called “January effect”. This effect was found in major industrial countries in the world stock markets.

3

on the subscriber, the remaining stocks are segmented into “A shares” and “B shares”. A shares are the shares traded in mainland of China and trades are with Renminbi. B shares are the shares available to abroad investors. Third, in the mature security markets, the IPO price is decided by intrinsic value and the relation of supply and demand but this is not the case in Chinese market because of the government influence. In summary, the Chinese stock market is not a result of reproduction of the “normal” stock market . it is strongly influenced by the government. Whether the “January effect” exists in Chinese markets still lacks consensus.

The goal of this paper is, therefore, to test whether the “January effects” exists in the Chinese stock markets despite the different market particularities and cultural differences. The Semi-strong Efficiency Market hypothesis (Fama, 1970) can be rejected for China if the results show that the January effects exist. If so, the qualified foreign institution investors and other domestic investors might have arbitrage opportunities making use of the January effects. In this paper, firstly, OLS and GARCH (1, 1) models are used following previous study methods that studying the existence of a “January effect” for indices. Many previous articles used OLS as the model to research monthly effects. Here, firstly, I also use OLS as a comparison. But OLS must meet the requirement of the constant and finite variance and that is very unlikely to occur with high frequency data. Therefore, many later researches use the Autoregressive Conditional Heteroskedasticity (ARCH) model. In this paper, I choose to use the GARCH (1,1) model.

4

individual stocks has no such a problem as each shares will have equal weight. From a practical view, when “January effects” are significant with index regression the investors can gain abnormal return by a portfolio which can track the market and buy and hold a portfolio in December and sell it later in January. And when January effects are significant with individual firms, this means that investors can gain abnormal return with portfolios which equally weight the stocks. As many articles (e.g., Banz,1981 and Smirlock 1985) show that small size firms have more significant January effects, I use an interaction variable with market value and January dummies to find out whether market size has an influence on January effects in each of the four markets. The four markets I study are Shanghai A share market, Shanghai B share market, Shenzhen A share market and Shenzhen B share market.

In section 2, there is a review of the articles of monthly effects in mature industrial countries’ stock markets and a summary of the empirical results as well as an overview for the Chinese stock markets. In section 3, I present the data and the methodology. Section 4 gives the results of these tests. I use OLS and GARCH analysis for the indices. Moreover, I re-examine January effects of the four markets using individual firms’ data. In Section 5, I provide the conclusion of this study.

2. Literature review

5

prices are independent and random. There are three forms of EMH: weak-form efficiency, semi-strong form efficiency and strong-form efficiency. In the weak-form, future price cannot be predicted by an analysis of prices from the past. In the semi-strong form, stock prices reflect publicly available information quickly and unbiasedly, so no abnormal return can be earned by trading on that information. In the strong-form efficiency, stock prices reflect all information, both public and private, such that no one can earn abnormal return. The study analysis whether once the information related is published, the stock price is able to incorporate the information quickly and accurately.

Since the 1960s, scholars found many market anomalies, like Calendar Anomalies (Lakonishok & Smidt,1988), Size effect (Keim,1983), IPO’ Anomalies (Chin, Lee and Chen, 2006) and Equity Premium Puzzle (Weil, 1989). These anomalies suggest that the capital market is not working fully in consistent with EMH. The regular anomalies are an important topic for scholars and important for investment strategy designs.

Calendar Anomalies include seasoned effects, monthly effects, weekly effects and holiday effects. Monthly effect means the financial markets have significant positive return or abnormal volatility during some months of the year. According to empirical study, most countries do have Monthly effect. In America, the January effect is notable. It was first found by Wachtel, 1942. Until 1976, Rozeff and Kinney studied it systematically. Some scholars found January effect mainly shown in small cap companies (e.g.Banz (1981)). And explanations like tax loss selling hypothesis (Reinganum, 1983), is come up to explain the anomalies.

6

the equal weighted New York Stock Market Index found that New York Stock Market Index average return in January is 3.5% while the other eleven months’ average return is only 0.5%. They found more than 1/3 of the annual return occurred in January of each year. They regard that this is caused by “Tax Loss Selling”.

7

For western countries’ January effect, besides tax loss selling hypothesis, scholars also come up the Window Dressing Hypothesis. The Window Dressing Hypothesis suggest that institution investors like to sell loss stocks in December and buy gain stocks to dress the annuitized report. These trades will increase the gaining stocks prices and decrease the losing stock prices. When these trades stop, the losing stock prices will go up to restore in line with value. Since 1980, studies of the January effects are no longer limited to stock markets. There are also studies of bonds market and funds markets. Tinic &West (1984) examined the January effects with the CAPM model. They found that the stocks with higher beta get the risk premium in January but not in other months. Smirlock (1985) used long-term Treasury bonds, long-term high-grade and low-grade corporate bonds from 1953 to 1981 and found that small firms with low-grade corporate bond have significant January effects. But high grade firms’ bonds and treasury bonds have no January effects. He drew a tentative conclusion from the study that both the equity and debt market are affected by January small firm effects. Keim (1986) found the companies pay high dividend have markeable January effects compared with companies had pay low dividend. Lakonishok & Smidt (1986) study the DOJ INDEX and found no January effects. They explain that this is due to DOJ index composition is big cap companies. Besides, Kato and Schallheim (1985) study the Japanese Stock market from 1964 to 1980 and their study found Japanese Stock market have January effect had size effects because small firms outperformed larger firms.

8

the January effects is decreasing. Until now, the most popular explanation of January effects is Tax Loss Selling hypothesis and windows dressing.

In summary, we can draw following conclusion from the literature:  Abnormal return in January is higher than in other months.  Small firms have a higher January abnormal return than big firms.  The abnormal return usually happened on the first week of January.  Tax Loss Selling is only a partial explanation to January effects.

9

select the data from January 1996 to December 2003 and found firms with a low market to book ratio have more significant January effects. Li (2010) used the CAPM model focusing on specific industries: pharmacy, transportation and clothing and draw the conclusion that there is no January effect in those industries.

The study of January effect in China can be concluded as following:

 The studies about Monthly effects of China are not that much and in recent years the number of such studies increases.

 There is still not a final verdict of whether January effects exist in China. Most scholars do not find a January effect in China stock market.

 There is different opinion about whether small cap firms have larger monthly effects.

Based on previous literature, my study first re-examines Shanghai and Shenzhen Stock market indices using OLS and GARCH (1, 1) as most previous studies. I choose four market segments, namely the Shanghai A share market, the Shanghai B share market, the Shenzhen A share market and the Shenzhen B share market. Then I examine the four markets with individual firms’ data to find whether the January effects will be different with panel data. Finally, I study whether there are size effects in any January effects in the Chinese markets. I do this by including market size and an interaction variable.

3. Data and Methodologies

10

Nowadays, there are A shares, B shares, H shares, N shares, L shares and S shares of listed firms. They are divided by the IPO place and investor types. A shares are the shares IPO at mainland of China and trade with Renminbi. The investor of A shares is Chinese mainland institutions and trade available for Chinese mainland investors. B shares are the shares is issued and listed in mainland China but the investors are abroad. Since 2001, Chinese residents can also invest in B shares. We define the A share market as a semi-closed market which is little affected by international markets. As there are many international mutual fund and hedge fund invest in Chinese B Share market. We consider that the B share market is affected by international stock markets.

11

3.2 Methodologies

Most previous empirical studies about the January effects used OLS with monthly dummy variables. In order to compare with previous articles results I first replicate the analysis and do the OLS regression with index. Some articles do the analysis using the CAPM model (e.g., Dbouk, Jamali and Kryzanowski, 2013). But I do not use it. The reason is that the model performs poor in the Chinese market (e.g., Li, 2010). First, I test the following equation for January effects:

(1)

In the above equation, the dependent variable Rt is weekly return of market indices i, Jant is the Dummy variable, Jant=1 when date t falls in January, otherwise 0. and ε is the random error term.

Hypothesis 1 is

H0: α1=0 H1: α1≠0

However, although OLS is the most commonly used estimation technique, but it has the precondition that the variance of random error should be constant and not change with time. However, this precondition cannot be satisfied. In fact the random residual may change with time heteroscedasticity in high frequency like weekly data. French, Schwert and Stambaugh (1987) examine the SP 500 index and found that there exists conditional heteroscedasticity. Pindyck & Rubinfeld (1998) further pointed out that if the annualized variance is constant, we may accept the null hypothesis mistakenly. This is the reason why I also use GARCH test. In general, GARCH (1, 1) is enough. As supplement of equation (1) variance equation to original equation in added to the following way.

(2)

12

H0: α1=0 H1: α1≠0

is the expected variance based on previous period information, ω is the constant term of the variance equation, εt-12 is the error term of equation (1) and (2). is variance of the period.

All above methods are for market indexes. But the four market indexes are weighting the constituents by market value. That means that the big firms takes dominate. From literature reviews, however, we know that the small cap firms have more significant January effects than big cap firms. Therefore, I also make analysis of individual firms and then each firms returns is weighted equally. I use panel least squares with cross-section fixed effects. To solve for heteroscedasticity, White diagonal robust standard errors are used. The following equation is used:

(3)

Rit is the firm’s weekly return, α0 is the average return, Dit is time dummy variable that take value 1 for each day in the evaluated event window in January for each firms and 0 otherwise. To further study the January effect, the event window period used is the first week, first two weeks, first three weeks and all January weeks. In this way, we can see how January effect changes with different event windows during January. α1 is the average abnormal return in January. εit is the residual.

Hypothesis 3 is:

H0: α1=0 H1: α1≠0

As the literature shows that firm sizes have effect on January return. I also control the firm size to see whether any findings of January effects change. The formula is:

13

Where, Rit, α0, α1 and ε have the same definition as in previous equation. MVit is the market value of firm i at time t. It represents the firm size. MVit*Dit is the interaction variable used to measure the effect of market value on the January effects.

Hypothesis 4 is:

H0: α1=0 H1: α1≠0

As January effects occur at the beginning of each year, the market value of previous year may be more important. Therefore, I add the previous year’s market value as a new control variable.

(5)

Where, other variables are the same as previous equation. T is the year. MVi,T-1 is the end of the previous year’s market value of firm i.

Hypothesis 5 is again:

H0: α1=0 H1: α1≠0

4. Results

4.1 OLS regression of market index results

Table 1 presents the results of equation (1) with OLS estimations. We can see from the table that no stock market indices shows significant January effects with 0.05 significant levels. We cannot reject the null hypothesis of hypothesis 1. There is no January effect according to the above results.

4.2 GARCH regression of market indexes

14

have a positive significant January dummy coefficient, while neither Shanghai nor Shenzhen A market indices show significant January dummy coefficients. As the GARCH model overcomes the shortcoming of the assumption of a constant variance, I believe that the GARCH model results are more convincing. That means that the Chinese B share market indexes have significant January effects. Moreover, the January effects in the Shanghai B stock market is larger than the January effect in the Shenzhen B stock market.

4.3 Panel data regression of individual firms

All the previous results are the results of market indices regressions. As mentioned before, theses index constituents are weighted by market value. Therefore, the results of individual firms’ regressions may be different from the indexes regression. Table 3 estimates the weekly returns for individual firms for the month of January obtained from equation 3. The coefficients associated with the January are the average abnormal returns for each event windows in January. From this table an interesting phenomenon can be found: almost all the coefficient of January Days dummy are positive and significant except these of 14 days and 21 days windows of the Shenzhen A stock market. Interestingly, the peaks of average weekly abnormal returns are primarily those that are formed within the first seven days of January. moreover, in cases the January effects are larger in the Shanghai stock markets than in Shenzhen Stock markets, while also B share January effects are larger than A share January effects.

15

January effects for the four segment markets in 7, 14, 21 and January windows after including market size and the interaction variable. It can be seen that market values have a negative contribution to firms’ returns. This indicates that larger firms may have more difficulties in generating high return than small firms. More importantly, all the coefficients of the interaction variable are also negative. They are significant in Shanghai A, Shanghai B and Shenzhen A stock markets except in the 7 days window of the Shanghai A stock market. And the coefficients of the interaction variable are not significant in the Shenzhen B stock market. We can draw the conclusion that the interaction variables have a significant negative value in most of firms. That means small firms have larger January effects than big firms. And the small size effects are clearest in the windows of 21 days and for the whole month of January.

Further analysis is done by replacing the MV (market value) with MV (-1) which means the market value at the end of last year. The reason do that is January effects may not be influenced by the market size of that time but by the time at the end of the previous year. The results of equation 5 are shown in table 5. Similar results are shown that there are January effects in all four segment stock markets and in all four windows. And the coefficients of the interaction variables are negative in most the windows of Shanghai A, Shanghai B and Shenzhen A stock markets. The market size has negative contribution to January effects in above three stock market segments. And this effect is larger in windows 21 days and for a full month of January.

16

5. Conclusion

This study first re-examine the January effects in Chinese stock market from 2001 to 2013 with four segment stock markets indices. Based on the OLS regression results, there is no January effect in Chinese stock markets. This is inconsistent with most previous studies (e.g., (Li & Sun, 2009)). But GARCH regression results indicate that there is a January effect in B share stock markets, but not in A share stock markets. Because GARCH regressions take the heteroscedasticity into consideration, I conclude that Chinese B share indices have January effects.

Further, I analysis the January effect and its relation with firms size for individual firms. I then find that there are positive significant January effects in all Chinese stock market segments, moreover, there is a small size effect in the Shanghai A, Shanghai B and Shenzhen A stock markets. In general, small firms have larger January effects than large firms. And such effects are clearer in the 21 the days and January windows. The results cannot be compared with previous Chinese studies because there is no other article using individual firms data with cross section fixed effects. However others have found such size effects in other markets (e.g., Banz, 1981 and Kato & Kiyoshi, 1985). Therefore, investors have opportunities to gain abnormal return by buying portfolios in December and sell them later in January.

17

18

References

Banz, R. W. (1981). The relationship between return and market value of common stocks. Journal of Financial Economics, 9(1), 3-18.

Brown, P., Keim, D. B., Kleidon, A. W., & Marsh, T. A. (1983). STOCK RETURNS SEASONALITIES AND THE TAX-LOSS SELLING HYPOTHESIS: Analysis of the arguments and australian evidence. Journal of Financial Economics, 12(1), 105-127.

Chang, E. C., & Pinegar, J. M. (1989). Seasonal fluctuations in industrial production and stock market seasonals. Journal of Financial & Quantitative Analysis, 24(1), 59-74.

Chang, E. C., & Pinegar, J. M. (1990). Stock market seasonals and prespecified multifactor pricing relations. Journal of Financial & Quantitative Analysis, 25(4), 517-533.

Cross, F. (1973). The behavior of stock prices on fridays and mondays. Financial Analysts Journal, 29(6), 67.

Eakins, S., & Sewell, S. (1993). TAX-LOSS SELLING, INSTITUTIONAL INVESTORS, AND THE JANUARY EFFECT: A NOTE. (cover story). Journal of Financial Research, 16(4), 377.

Engle, R. F., & Mustafa, C. (1992). Implied ARCH models from options prices. Journal of Econometrics, 52(1-2), 289-311. doi:9

Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. Journal of Finance, 25(2), 383-417.

19

Gultekin, M. N., & Gultekin, N. B. (1983). Stock market seasonality: International evidence. Journal of Financial Economics, 12(4), 469-481.

doi:http://dx.doi.org/10.1016/0304-405X(83)90044-2

Kato, K., & Schallheim, J. S. (1985). Seasonal and size anomalies in the japanese stock market. Journal of Financial & Quantitative Analysis, 20(2), 243-260.

Keim, D. B. (1986). The CAPM and equity return regularities. Financial Analysts Journal, , 19-34.

Keim, D. B. (1983). SIZE-RELATED ANOMALIES AND STOCK RETURN SEASONALITY: Further empirical evidence. Journal of Financial Economics, 12(1), 13-32. Keim, D. B. (1983). Size-related anomalies and stock return seasonality: Further

empirical evidence. Journal of Financial Economics, 12(1), 13-32. doi:http://dx.doi.org/10.1016/0304-405X(83)90025-9

Kramer, C. (1994). Macroeconomic seasonality and the january effect. Journal of Finance, 49(5), 1883-1891.

Lakonishok, J., & Smidt, S. (1988). Are seasonal anomalies real? A ninety-year perspective. Review of Financial Studies, 1(4), 403.

Mehdian, S., & Perry, M. J. (2002). Anomalies in US equity markets: A re-examination of the january effect. Applied Financial Economics, 12(2), 141-145.

doi:10.1080/09603100110088067

Reinganum, M. R. (1983). The anomalous stock market behavior of small firms in january: Empirical tests for tax-loss selling effects. Journal of Financial

Economics, 12(1), 89-104. doi:http://dx.doi.org/10.1016/0304-405X(83)90029-6

20

International Library of Critical Writings in Financial Economics, vol. 3;

Cheltenham, U.K. and Lyme, N.H.:; Elgar; distributed by American International Distribution Corporation, Williston, Vt.

Smirlock, M. (1985). Seasonality and bond market returns*. The Journal of Portfolio Management, 11(3), 42-44.

Stoll, H. R., & Whaley, R. E. (1983). Transaction costs and the small firm effect. Journal of Financial Economics, 12(1), 57-79.

Tinic, S. M., & West, R. R. (1984). Risk and return: Janaury vs. the rest of the year. Journal of Financial Economics, 13(4), 561-574.

doi:http://dx.doi.org/10.1016/0304-405X(84)90016-3

Wachtel, S. B. (1942). Certain observations on seasonal movements in stock prices. Journal of Business, 15, 184-193.

Wachtel, S. B. (1942). Certain observations on seasonal movements in stock prices. Journal of Business, 15, 184-193.

Chin, C. L., Lee, P., Kleinman, G., & Chen, P. Y. (2006). IPO anomalies and innovation capital. Review of Quantitative Finance and Accounting, 27(1), 67-91.

Weil, P. (1989). The equity premium puzzle and the risk-free rate puzzle. Journal of

Monetary Economics, 24(3), 401-421.

Ross, S. A. (1976). The arbitrage theory of capital asset pricing. Journal of Economic Theory, 13(3), 341-360.

21

李红兵,孙丽敏. 中国股票市场一月效应的实证研究[J]. 财贸研究,2009,02:96-100. ( translation: Li Hongwei, & Sun Limin. (2009). Empirical study of January Effect on Chinese Stock Market. Finance and Trade Research, (02), 96-100.)

李呈娇. CAPM 模型对中国上证 A 股的实证分析[J]. 武汉冶金管理干部学院学 报,2010,04:14-16. ( translation: Li Chengjiao. (2010). Empirical analysis of Shanghai A-share index by CAPM. Journal of Wu Han Metallurgical Management School, (04), 14-16.)

邓长荣,马永开. 三因素模型在中国证券市场的实证研究[J]. 管理学

报,2005,05:591-596. ( translation: Deng Changrong, & Ma Yongkai. (2005). Empirical study of Three Factors Model on Chinese Stock Market. Journal of Management, (05), 591-596.)

徐炜,张兵. 中国股市月份效应研究[J]. 经济管理,2005,24:63-68. ( translation: Xu Wei, & Zhang Bing. (2005). Research of monthly effect on Chinese Stock Market. Economic Management, (24), 63-68.)

刘凤元,陈俊芳. 换月效应的窗饰解释:基于上海市场的实证[J]. 数量经济技术经 济研究,2004,03:149-154. ( translation: Liu Fengyuan, & Chen Junfang. (2004). Windows dressing hypothesis of Monthly Effect- based on demonstration of Shanghai Stock market. Research of Quantity Economics and Technical Economics, (03), 149-154.)

22

董文卓. 中国股票市场的“月份效应”研究——“三月与十二月份”效应[J]. 中山大 学研究生学刊(社会科学版),2005,03:79-90. ( translation: Dong Wenzhuo. (2005). Study of monthly effect on Chinese Stock Market- March Effect and December Effect. Postgraduate Journal of Zhong Shan University (social science version), (03),79-90.)

张颖,刘桂荣. 上证综合指数一月效应的实证分析[J]. 金融经济,2013,24:124-125. ( translation: Zhang Ying, & Liu Guirong. (2013). Empirical analysis of January Effect on Shanghai composite index. Financial Economy, (24), 124-125.)

何晓光,许友传. 中国 A 股市场的月份效应研究[J]. 生产力研究,2006,01:80-82. ( translation: He Xiaoguang, & Xu Youchuan. (2006). Study of monthly effect on Chinese A share market. Study of Productivity, (01), 80-82.)

范辛亭,董文卓. 中国股票市场的月份效应——基于消费习惯的解释[J]. 中山大学 学报(社会科学版),2007,03:97-104+128. ( translation: Fan Xinting, & Dong Wenzhuo. (2007). Monthly effect on Chinese Stock Market- based on the explanation of consumption habit. Journal of Zhong Shan University (social science version), (03), 97-104+128.)

严雄. 中国股票市场换月效应及窗饰假说的实证研究[D].暨南大学,2010. ( translation:

Yan Xiong. (2010). Empirical study of monthly effect and window dressing hypothesis on Chinese Stock Market. Ji Nan University.)

邓长荣. 三因素模型在中国证券市场的实证研究[D].电子科技大学,2006.

23

刘洪. 股票市场月份效应[D].浙江大学,2011. ( translation: Liu Hong. (2011). Monthly effect on stock market. Zhe Jiang University.)

Table 1 OLS regression results of market index for January effects (2001 – 2013)

24

Probability 0.2662

January 0.0058

Probability 0.3889

Notes: This table presents the results from OLS regression of equation (1). The January is dummy variable that take the value 1 if the date is in the January, and 0 otherwise. * indicates statistical significance at the 0.05 level. Numbers in brackets represent negative number.

Table 2 GARCH regression results of market index for January effects (2001 – 2013)

Shanghai A Constant (0.0007) Probability 0.5509 January 0.0045 Probability 0.2921 Shanghai B Constant (0.0015) Probability 0.4139 January 0.0249 Probability 0.0000* Shenzhen A Constant (0.0006) Probability 0.6523 January 0.0037 Probability 0.4415 Shenzhen B Constant 0.0014 Probability 0.4108 January 0.0173 Probability 0.0001*

25

Table 3 Panel data regression results of individual firms for January effects (2001 – 2013)

January days 7 14 21 January

Shanghai A Constant (0.0002) (0.0002) (0.0003) (0.0005) Probability 0.0222* 0.0768 0.0008* 0.0000* January Days 0.0164 0.0070 0.0072 0.0076 Probability 0.0000* 0.0000* 0.0000* 0.0000* Shanghai B Constant 0.0010 0.0007 0.0005 0.0004 Probability 0.0026* 0.0429* 0.1208 0.2467 January Days 0.0187 0.0180 0.0148 0.0115 Probability 0.0000* 0.0000* 0.0000* 0.0000* Shenzhen A Constant (0.0003) (0.0001) (0.0001) (0.0003) Probability 0.0014* 0.2633 0.4181 0.0006* January Days 0.0099 (0.0002) 0.4181 0.0026 Probability 0.0000* 0.7615 0.1435 0.0000* Shenzhen B Constant 0.0015 0.0015 0.0014 0.0011 Probability 0.0000* 0.0000* 0.0000* 0.0011* January Days 0.0077 0.0047 0.0050 0.0068 Probability 0.0031* 0.0153* 0.0012* 0.0000*

Notes: This table presents the results from panel regression of equation (3). Event windows of 7, 14, 21 and one month are employed. The January Days is dummy variable that take the value 1 if the date is in the event window, and 0 otherwise. * indicates statistical significance at the 0.05 level. Numbers in brackets represent negative numbers.

Table 4 Panel data regression results of individual firms for January effects with market value and interaction variable (2001 – 2013)

January days 7 14 21 January

Shanghai A Constant 0.0003 0.0003 0.0002 (0.0001)

Probability 0.0028* 0.0012* 0.1043* 0.5451*

26 Probability 0.0000* 0.0000* 0.0000* 0.0000* MV (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0000* 0.0000* 0.0000* 0.0000* MV*January Days (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.6358 0.0355* 0.0004* 0.0013* Shanghai B Constant 0.0023 0.0019 0.0017 0.0015 Probability 0.0000* 0.0000* 0.0000* 0.0002* January Days 0.0205 0.0206 0.0180 0.0135 Probability 0.0000* 0.0000* 0.0000* 0.0000* MV (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0000* 0.0000* 0.0000* 0.0000* MV*January Days (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0171* 0.0002* 0.0000* 0.0001* Shenzhen A Constant 0.0014 0.0016 0.0016 0.0013 Probability 0.0000* 0.0000* 0.0000* 0.0000* January Days 0.0112 0.0015 0.0016 0.0042 Probability 0.0000* 0.0088* 0.0007* 0.0000* MV (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0000* 0.0000* 0.0000* 0.0000* MV*January Days (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0001* 0.0000* 0.0000* 0.0000* Shenzhen B Constant 0.0034 0.0033 0.0032 0.0029 Probability 0.0000* 0.0000* 0.0000* 0.0000* January Days 0.0088 0.0055 0.0054 0.0075 Probability 0.0026* 0.0122* 0.0022* 0.0000* MV (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0000* 0.0000* 0.0000* 0.0000* MV*January Days (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.1458 0.2611 0.4997 0.2016

Notes: This table presents the results from panel regression of equation (4). Event windows of 7, 14, 21 and one month are employed. The January Days is dummy variable that take the value 1 if the date is in the event window, and 0 otherwise. * indicates statistical significance at the 0.05 level. Numbers in brackets represent negative numbers. MV is the market value. MV*January Days is interaction variables which are multiple market values by January Days dummy variables

Table 5 Panel data regression results of individual firms for January effects with market value of end of previous year and interaction variable (2001 – 2013)

January days 7 14 21 January

Shanghai A Constant 0.0003 0.0003 0.0001 (0.0001)

Probability 0.0113* 0.0057* 0.2323 0.3542

January Days 0.0165 0.0070 0.0074 0.0077

Probability 0.0000* 0.0000* 0.0000* 0.0000*

27 Probability 0.0000* 0.0000* 0.0000* 0.0000* MV(-1)*January Days (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.5584 0.0268* 0.0002* 0.0025* Shanghai B Constant 0.0024 0.0020 0.0017 0.0016 Probability 0.0000* 0.0000* 0.0000* 0.0001* January Days 0.0257 0.0233 0.0197 0.0145 Probability 0.0000* 0.0000* 0.0000* 0.0000* MV(-1) (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0000* 0.0000* 0.0000* 0.0000* MV(-1)*January Days (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0050* 0.0000* 0.0000* 0.0000* Shenzhen A Constant 0.0014 0.0016 0.0015 0.0013 Probability 0.0000* 0.0000* 0.0000* 0.0000* January Days 0.0114 0.0014 0.0016 0.0043 Probability 0.0000* 0.0126* 0.0006* 0.0000* MV(-1) (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0000* 0.0000* 0.0000* 0.0000* MV(-1)*January Days (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0001* 0.0000* 0.0000* 0.0000* Shenzhen B Constant 0.0034 0.0034 0.0033 0.0029 Probability 0.0000* 0.0000* 0.0000* 0.0000* January Days 0.0113 0.0067 0.0063 0.0081 Probability 0.0003* 0.0032* 0.0004* 0.0000* MV(-1) (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0000* 0.0000* 0.0000* 0.0000* MV(-1)*January Days (0.0000) (0.0000) (0.0000) (0.0000) Probability 0.0762 0.1555 0.2809 0.1148