JANUARY EFFECT AND CHINESE NEW YEAR EFFECT IN CHINESE STOCK MARKET

(1)

JANUARY EFFECT AND CHINESE NEW YEAR

EFFECT IN CHINESE STOCK MARKET

University of Groningen

Faculty of Economics and Business

ERLAN ZHOU

Supervisors:

Dr. A.J. (Aljar) Meesters

(2)

Abstract

This study tests the existence of January effect and Chinese New Year effect and the evolution pattern of these anomalies in Chinese stock market. The dataset is based on SSE Composite Index and SSE Component Index with the period of 1997 to 2010. The results indicate that Chinese stock market does not have January effect, while Chinese New Year effect evolves to be more evident. Moreover, there is no evidence to prove the trading volume is related with the anomalies in January and Chinese New Year vacation. Lastly, the test of predicting role of these anomalies to the returns of next 11 months shows that Other January effect, which means the return of January predicts the returns of next 11 months, does exist and Chinese New Year effect also predicts the returns of next 11 months.

(3)

Acknowledgement

(4)

Introduction

This thesis will focus on the January effect and Chinese New Year effect in the Chinese stock market. The January effect1 is a calendar-related anomaly in the financial market, which indicates that financial security prices increase in the month of January. Therefore, investors can seize this opportunity to buy stock for lower prices before January and sell them after their value has increased. Obviously, this type of pattern in price behavior on the financial market indicates that financial markets are not fully efficient. Another famous “January effect” 2 is involved with the Chinese New Year. It is a fact that this vacation, which is determined by a lunar calendar, is the most important holiday in China. In this holiday, most economic and commercial activities come to a halt. Employees will usually receive year-end bonuses and overdue loans are normally settled before lunar New Year Eve. As Lee, Yen and Chang (1993) concluded, Chinese New Year has significant impact on the stock market in Taiwan, which is named as the Chinese New Year effect. Yang et al. (2006) has documented that there is significant evidence supporting the Chinese New Year effect.

It is clear that the January effect is identified in American and most of European stock markets and some studies indicate the Chinese New Year effect exists in Taiwan and Hong Kong stock markets. Since the fast growth of China has attracted the international investors' interests from the past decade, the study on the existences and evolutions of these anomalies are important for the investors. Currently, the economy of China is the third largest in the world, after the United States and Japan with a

(6)

nominal GDP of US$4.91 trillion, and China is the fastest-growing major economy in the world, and has had the fastest growing major economy for the past 30 years with an average annual GDP growth rate of over 10%. Besides, China is the second largest trading nation in the world and the largest exporter and second largest importer of goods, which contributes to drive China to lead the role as the engine of the world economical development. As the increasing importance of Chinese area and the increasing internationalized stock markets, the motivation to investigate and to analyze the January effect and the Chinese New Year effect of China is becoming more and more prominent for investors in Chinese stock market.

As former studies indicate, most of them have identified the January effect and the Chinese New Year effect. Some of the researches have documented that the January effect is declining. Moreover, the foreign studies have also investigated the relation between trading volume and the anomaly and the predicting role of the January effect (The Other January effect3_{). However, the researches of China merely focus on testing}

the existences of the anomalies, while the study on the issues of evolutions of these anomalies, trading volume and the Other January effect are still a gap in the researches of China. This study tests the existences of the January effect and the Chinese New Year effect, the evolution patterns of these anomalies, relations between anomalies and trading volumes and the predicting role of anomalies.

(7)

(8)

II Literature Review

A. The research on foreign financial markets

Investors should not be able to use public information to receive abnormal profits under the efficient market hypothesis (EMH)5, as security prices should reflect all available information. Nevertheless, there are a number of persistent challenges to this famous hypothesis that have been recorded in the literature. One of the most widely recognized of these anomalies is the January effect, which means the observation that the month of January shows to have systematically higher returns than other months of the year.

Rozeff and Kinney (1976) provided the first empirical evidence of the January effect in the United States. Since then, the anomaly has attracted much attention from academics and practitioners alike. These findings are supported in subsequent research like Keim (1983), Roll (1983) and Reinganum (1983), which identified that the effect is mainly concentrated in smaller firms. Branch and Chang (1990) have concluded that the January effect is particularly significant in small firms with low share prices and De Bondt and Thaler (1987) have documented that small firms, which have underperformed in the past, have January effect. Other studies like Roll (1981) and Roll (1983) have

(9)

approved that market microstructure effects6, such as the bid-ask spread and thin-trading, which are more prominent for lower priced and small stocks, are important factors in explaining the abnormally high returns in January. Including the United States, many studies like Gultekin and Gultekin (1983) have also concluded a significant January effect in other countries and other securities. More recent studies, such as Booth and Keim (2000), Hensel and Ziemba (2000) and Rendon and Ziemba (2005), have focused on the ability of investors to take advantage of the effect using mutual funds or stock index futures.

For the abnormally high returns in January, the literature comes up with several explanations. The tax-loss selling hypothesis and the window dressing hypothesis are two of the most pronounced among these theories. Dyl (1977) has suggested the tax-loss selling hypothesis, which means investors sell the losing stocks in their portfolio in order to gain a tax benefit at the end of the year. According to Haugen and Lakonishok (1988), the window dressing hypothesis suggests that investors sell some stocks to construct a more acceptable portfolio of stocks to fund holders in their year-end reports when it comes to the end of the year. Both theories suggest that as a result of these investors repurchasing the stocks in the new year, it will create the abnormal returns observed in January. According to Chen and Singal (2004), evidence has approved that the tax-loss selling hypothesis is the most likely explanation for the January effect. Nevertheless, other studies like Brown et al. (1983) and Fountas and Segredakis (2002) have indicated that the theory may not be sufficient to support the abnormally high returns observed in many foreign countries where the tax year for individuals is not the calendar year.

(10)

As the research of this anomaly goes on, the most interesting issue is that recent research has argued that the magnitude of January effect has declined. As Riepe (1998) concludes, investors not only increase general market knowledge about the January effect, but also take the advantage of the rising of futures contracts on many of the major US market indices. As a result, this has created a low-cost investment strategy for investors to profit from the effect. In more recent times, investors have the knowledge of January effect and they are inclined to take advantage of the anomaly. Mehdian and Perry (2002) and Gu (2003) have documented that the January effect of major US market indices has disappeared at a statistical level or has been in decline after 1987, which means that investors may have begun taking advantage of the January effect. As Moller and Zilca (2008) notes, they have concluded that January effect has declined, and judging from trading volume of January, some investors have become more knowledgeable and sophisticated. As a result, they are less willing to buy stocks when the January effect peaks. Easterday et al. (2008) have documented that trading volume for small firms in December and January is not different from other months, indicating that traders do not actively arbitrage the anomaly.

(11)

suggests a ‘temporary anomaly’ interpretation.

B. The research on Chinese financial markets

When it comes to China, the research on January effect is relatively lagged behind. The study of January effect began in the 1990s. Cheuang (1997) has concluded that small-cap stocks in January yielded greatly different returns from other months in the Hong Kong stock market and the January effect of large capitalization stocks is more evident. Feng (2003) has identified that January effect, which is found in the majority of developed countries and other emerging stock markets, does not exist. But Shanghai and Shenzhen stock markets have a significant Turn-of-the-Month effect7, indicating a lack of efficiency in the stock market. Li (2003) examined the Shanghai Composite Index and Shenzhen Component Index from the period of Jan. 2 1993 to Dec. 3 2002, confirming that January effect does not exist in the stock market of China, but a December effect exists.

Obviously, the data used in these articles are mainly in the early stage of stock market, which lacks reasonable regulations. As some regulations have been adopted recently and stock markets are getting more and more standardized, previous results for guiding the investors in the stock market are increasingly weak. In recent years, the range of data used to test January effect is expanded and methods are constantly updated. Xu (2004) has documented that the Shanghai stock market has significant Season effect, while Season effect of the Shenzhen stock market is not obvious, and the study also find

(12)

that December effect exists in the Shanghai and Shenzhen markets. They also believe that the existence of the Season effect, from a point of view, reflects the low efficiency of stock market in China. Zhang (2004) indicates there is a significant Week effect, but the January effect is weak. Seeing from the researches above, the researches do not confirm the January effect in Chinese stock market.

(13)

normally jump before and during Chinese New Year holiday season. Considering the supply side, the central bank pumps new money into the banking system to accommodate the year end seasonal demand (lunar calendar system). Interestingly, the rising interest rate environment does not stop the rise of stock prices in Chinese New Year. Zhang and Gu (2005) did not find any significant evidence of the Chinese New Year effect, which is not expected. Conversely, Yang et al. (2006) and Zhang et al. (2007) did find the Chinese New effect. The former is based on the whole stock market level, while the later is the research on stocks from the tourism companies.

In nutshell, most of the literature shows that the January effect and the Chinese New Year effect in the stock market of China do not exist or are not evident, due to some limitations like the less developed stock market and inefficient stock market supervision. As the stock market proves more standardized and formative, the supervision of old researches is decreasing. Besides, it is clear that researches on the decline patterns of the stock market anomalies, the relations between trading volumes and the stock market anomalies and the predicting roles of stock market anomalies like the Other January effect are still blank. As a result, this study will focus on these parts to fill in the blank of the research of the Chinese stock market.

(14)

(15)

III Methodology

Before starting to test the hypotheses, some parameters should be defined. The return of stock markets is constructed by the following equation8:

-1

100 * log(

/

)

t t t

R

=

P

(1) Where: P is the daily price of index. _t

In order to test the existence of the January effect and the Chinese New Year effect, this paper will employ two methods. The first one involves plotting the relationship between January return or February return and the monthly average return for the other months. By analyzing the graphs we will see whether the “January effects” exist. The second method uses formal statistical analysis to test for the existence of these anomalies. If the returns of the indexes have similar trends for other months and only the January return or February return shows very different trend from all the other months in the first method, the second method involves testing the null hypothesis of equal January return or February return and returns for the rest of the year. For the test of January effect, the equation9 will be given:

0 1 t t

R

=

β

+

β

Jan

+

ε

(2)

(

2

)

~

0,

t

NID

ε

σ

(3) Where: R is the daily return in day t and Jan is a dummy variable, which is set equal to _t

one if the day is in January and to zero otherwise. The intercept β₀ indicates average

(16)

daily return of the other 11 months while the coefficient β₁ represents the difference between the average daily return in January and the other 11 months. If the mean return is the same for each month, then the estimate β₁ would be close to zero and the F- statistic would be insignificant. Apparently, the test of the Chinese New Year effect employs the same method. In order to research the decline patterns of these stock market anomalies, the data are split into three categories. The first range is from Jan. 1st 1997 to Dec.1st 2003, and the second one is from Jan.1st 2004 to Apr. 30th 2010. The third one is the whole range from Jan.1st 1997 to Apr. 30th 2010. By employing the research methods above, the decline patterns are presented from the results of these three categories.

The test of trading volumes also involves two methods above. For testing the hypothesis of that the magnitudes of trading volumes in December and January or in January and February are higher than in other months of the year, it is obvious that using plotting analysis will present a clear image of the relationships between the magnitudes of trading volume in two months and the other months. For testing the correlation between returns and trading volumes, the investigative approach utilizes simple intuitive regression models10. Setting the January effect part as an example, as my premise is that investors attempting to arbitrage the January effect also influence December trading behavior, I exclude January (December) observations from the regression equation testing the December (January) dummy variable in order to avoid confounding the tests:

1 1

|

2 1 3

( *|

1

|)

t

TV

= +

α β

ret

+

β

D

+

β

D

ret

+

ε

(4)

2 4

|

5 2 6

( *|

2

|)

t

(17)

In Equation (4) and (5), TV is the trading volume in month i, |ret is the absolute value | of the return in month i, D equals 1 if the month is January, 0 otherwise (December ₁

observations excluded from sample), D equals 1 if the month is December, 0 otherwise ₂

(January observations excluded from sample). The definition of the trading volume variable TV is the daily trading volumes of SSE Composite Index and SSE Component Index. Interpretation of the regression coefficients is as follows:

•α₁, α₂ = average trading volume for the 11-month period;

•β₁, β₄= average correlation between trading volume and return for the 11-month period; •β₂, β₅ = incremental trading volume in January (December) relative to the intercept; • β₃ , β₆ = incremental correlation between trading volume and return in January

(December) relative to β₁ (β₄).

Evidence of investors trading on the January effect would manifest in β₂ and β₅ being significantly positive (elevated trading volume relative to other months) and/or β₃ and

6

(18)

Testing hypotheses of the January effect or the Chinese New Year effect predicting returns of other months requires regression model. Similarly, I take the hypothesis of January effect as an example. The required equations11 are as follows:

1 1 t t t

r

=

α β

+

J

+

ε

(6) 2 2 , t J t t

r

=

α β

+

r

+

ε

₍₇₎

In Equation (6), r is the 11-month excess return from February to December in year t for _t

the respective stock market. J = 1 if the January excess return in year t is positive and _t t

J = 0 otherwise. I test the null hypothesis that β1 is significantly different from zero. If

1

β is significant and positive, this is viewed as evidence in support of the Other January effect. Noting that β₁ can also is interpreted as the spread in returns between positive and negative Januarys. Equation (6) has the drawback that potentially useful information might be lost when only considering a dummy variable. For completeness, I also run the regression (7) where the excess January return is used in lieu of the January dummy as the explanatory variable. In Equation (7),

r

_{J t}_, is the January excess return for year t and

the other terms are as defined for Equation (6). Here, a reliably positive estimate for the

2

(19)

IV Data and descriptive statistics

A. Data Description

There are two stock exchanges in China. One is Shanghai Stock Exchange and the other one is Shenzhen Stock Exchange. In order to research the behaviors of Chinese Stock Market, these two stock exchanges are considered and SSE Composite Index and SSE Component Index, which represent two exchanges respectively, are studied. The data of this study is drawn from Google finance12_.

Before 1997, the stock market of China was still in the primary stage. The lack of regulations and suitable restrictions caused the stock prices easily to jump or fall significantly. This phenomenon brought much noise into the research on stock prices. On Dec. 16th_{1996, CSRC (China Security Regulation Commission) brought out a new}

regulation of 10% limit of daily stock price movements to control the sudden rise or sharp fall of stock prices. In order to generate a more accurate research result, this paper will select the data from the period of Jan. 1st 1997 to Apr. 30th 2010. Based on the researches on January effect of stock markets in financial markets, most of the studies are limited by the fact that they use only monthly returns to draw their conclusions. Investigating the daily pattern of the January effect, rather than measuring the effect at the monthly level, may give us a better picture of January effect and the decline pattern of it. As a result, the selected data are daily Shanghai Stock Exchange Composite Index and

(20)

Shenzhen Stock Exchange Component Index13, and daily trading volumes of these two stock exchanges are also selected.

B. Main data descriptive statistics

After calculating the returns of these two indexes, the Figure 4.1 and Figure 4.2, which present the daily return of security markets of the period from 1997 Jan.1st to 2010 Apr. 30th, show the general pattern of returns.

Figure 4.1 The daily return of SSE Composite Index from Jan. 1st_{1997 to Apr. 30}th₂₀₁₀

SSE Composite Index

-5 -4 -3 -2 -1 0 1 2 3 4 5 1997/1/2 1998/11/9 2000/9/21 2002/8/14 2004/6/30 2006/5/12 2008/3/14 2010/1/13 Re tu rn (% )

Figure 4.2 The daily return of SSE Component Index from Jan. 1st_{1997 to Apr. 30}th₂₀₁₀

SSE Component Index

(21)

It is clear that SSE Composite Index and SSE Component Index have the similarities in the pattern of returns. During the period of 1997 to 2000, they fluctuated more significantly compared with the range from 2000 to 2007. Nevertheless, compared to the first three-year period, they even presented more evident fluctuation range from 2007 to 2009. Obviously, it has something to do with the unreasonable fluctuations of stock exchange indexes. Only seeing from the general pattern can not get an obvious result of stock market anomalies. In order to see the individual performance of every month, I list the main results in Table 4.1, which are the descriptive statistics of the mean and standard deviation of daily stock markets returns, the largest and the smallest value, T-statistic, Jarque-Bera test and number of observations.

Table 4.1 Monthly average returns of Composite Index and SSE Component Index Summary Statistics of SSE Composite Index and SSE Component Index Return (%)

SSE Composite Index SSE Component Index

Month Mean Stdev. Max. Min. T-Stat JB Mean Stdev. Max. Min. T-Stat JB Num.

All 0.015 0.757 4.083 -4.054 1.154 2348.7 0.017 0.828 4.139 -4.315 1.149 1643.2 3222 Jan. 0.019 0.839 2.86 -3.253 0.365 77.4 0.075 0.944 3.066 -3.441 1.271 57.7 256 Feb. 0.08 0.924 3.763 -4.054 1.241 220.1 0.078 0.979 3.886 -4.315 1.142 170 203 Mar. 0.056 0.64 2.578 -2.418 1.533** 47.9 0.068 0.712 2.903 -2.43 1.675** 31.8 310 Apr. 0.065 0.688 3.86 -2.511 1.627** 217.8 0.055 0.766 3.979 -3.333 1.248 239.8 298 May 0.034 0.839 1.971 -4.016 0.631 170.2 0.036 0.978 2.625 -4.009 0.565 79.2 236 Jun. 0.006 0.887 3.843 -3.743 0.119 175.6 0.021 0.969 3.879 -3.739 0.353 111.8 277 Jul. 0.000 0.749 2.718 -3.44 0.001 99.1 -0.012 0.877 3.486 -3.634 -0.24 97.1 288 Aug. -0.052 0.764 3.192 -3.79 -1.145 237.1 -0.069 0.809 2.99 -3.433 -1.441 141.3 286 Sep. -0.016 0.757 3.924 -3.059 -0.359 341.9 -0.041 0.771 3.744 -3.009 -0.887 193.5 275 Oct. -0.051 0.787 4.083 -2.835 -0.991 128.3 -0.043 0.805 4.139 -3.099 -0.81 117.2 230 Nov. 0.018 0.675 3.049 -2.829 0.437 133.1 0.031 0.74 2.735 -3.255 0.688 64.7 277 Dec. 0.028 0.548 1.787 -2.024 0.864 27.5 0.011 0.583 2.322 -2.105 0.332 33.5 286 Significance level *5% **10%

(22)

not significantly high. Conversely, January and February have the highest returns, 0.075% and 0.078% respectively, in SSE Component Index. It is obvious that the returns of March, April and May share with the similarities and August, September and October have the negative returns in both indexes. In contrast to my prediction, the return of January and February in both indexes fail to pass the T-test, even at the significance level of 10%.

This part presents the description and main statistics of trading volumes of both indexes. In order to see individual trading volume of each month, the main descriptive statistics of trading volume are given in Table 4.2.

Table 4.2 Monthly average trading volumes of Composite Index and SSE Component Index Summary Statistics of SSE Composite Index and SSE Component Index Trading Volume (mln)

Month Mean Stdev Max. Min. JB Num. Mean Stdev Max. Min. JB Num.

All 39.625 49.761 303.175 0.000 2220.9 3222 22.834 25.043 156.974 0.000 2169.2 3222 Jan. 42.478 48.491 185.213 1.938 51.7 256 24.506 25.245 104.829 1.638 64.4 256 Feb. 48.077 55.058 230.583 3.148 112.2 203 27.711 28.134 117.198 3.074 90.6 203 Mar. 40.586 46.792 203.033 0.000 131.2 310 24.221 23.213 92.483 0.000 89.6 310 Apr. 49.105 55.692 213.985 4.236 90.7 298 29.550 29.337 110.376 5.074 81.9 298 May 44.869 56.777 218.563 3.662 96.6 236 26.011 27.845 110.223 3.488 81.5 236 Jun. 40.461 49.039 189.883 2.751 132.3 277 23.961 24.104 97.864 3.276 107.6 277 Jul. 38.558 54.459 303.175 3.033 491.8 288 21.689 24.809 112.827 2.907 273.1 288 Aug. 34.698 46.897 230.009 2.507 310.2 286 19.066 21.357 106.903 2.474 249.7 286 Sep. 32.986 42.852 186.678 2.018 188.0 275 18.167 20.020 93.629 2.092 172.1 275 Oct. 30.592 36.821 161.398 1.538 130.0 230 17.711 19.747 94.223 1.614 302.3 230 Nov. 36.924 51.386 272.142 2.347 382.0 277 21.307 28.305 156.974 2.573 626.6 277 Dec. 37.280 47.097 243.138 1.875 157.2 286 20.669 24.084 127.435 1.789 197.3 286

(23)

(24)

V Results Analysis

A. Examinations of January effect and Chinese New Year effect

In this part, the main results of test for the existences of January effect and Chinese New Year effect and the evolution patterns of these anomalies are presented.

Figure 5.1 Individual month return of SSE Composite Index SSE Composite Index Return

-0.100 -0.050 0.000 0.050 0.100 0.150 Janu ary Febu rary Marc h April May June July Augu st Sept ember October Novem ber Decm ber Re tu rn (% )

Jan.1997-Apr.2010 Jan.1997-Dec.2003 Jan.2004-Apr.2010

Figure 5.2 Individual month return of SSE Component Index

SSE Component Index Return

-0.150 -0.100 -0.050 0.000 0.050 0.100 0.150 Janu ary Febu rary Mar ch April May June July Aug ust Sept embe r Octob er Novemb er Decm ber Ret ur n( % )

Jan.1997-Apr.2010 Jan.1997-Apr.2003 Jan.2004-Apr.2010

(25)

in January while SSE Component has the second highest return in January. Seeing from the other two categories, it is obvious that the return of January from presents a weaker trend while the returns of February exists a stronger performance for both indexes. Assuming the anomalies of January and February are significant, the January effect evolves weaker while the Chinese New Year effect has a stronger influence over time.

Table 5.1 Results of the examination of January effect and Chinese New Year effect Tests of January effect and Chinese New Year effect

β1 T-stat β1 T-stat

January effect

Jan.1997-Apr.2010 0.004 0.083 0.063 1.171

Jan.1997-Dec.2003 0.051 0.812 0.097 1.413

Jan.2004-Apr.2010 -0.044 -0.573 0.027 0.325

Chinese New Year effect

Jan.1997-Apr.2010 0.069 1.266 0.066 1.096

Jan.1997-Dec.2003 0.044 0.590 0.009 0.115

Jan.2004-Apr.2010 0.088 1.098** 0.104 1.188**

Significance level *5% **10%

Notes: β1 represents the difference between the average daily return in January or February and the other 11 months.

(26)

the SSE Composite Index and SSE Component Index, the January effect does not exist and the Chinese New Year effect proves to be stronger. Therefore, investors can not profit from January effect, but the Chinese New Year effect. Since the previous studies have proved that the Chinese financial market is not efficient, the inexistence of January effect does not mean the investors employ the relevant knowledge to arbitrage the anomaly away.

B. Empirical test of the relationships between trading volumes and

anomalies

The results of the relationships between trading volumes and anomalies are shown in this part. Figure 5.3 and 5.4 present the trading volume of individual month of SSE Composite Index and SSE Component Index.

Figure 5.3 Trading volume of individual month of SSE Composite Index SSE Composite Index Trading Volume(mln)

0 10 20 30 40 50 60 Janu ary Febu rary March Ap ril May June July Aug ust Septem ber Octob er Nove mbe r Decm ber T ra din g V olu me ( m ln ) Jan.1997-Apr.2010

(27)

SSE Component Index Trading Volume(mln) 0 5 10 15 20 25 30 35 Janu ary Febu rary Mar ch April May June July Augu st Septe mber Octob er Nov ember Decm ber T ra din g V olu m e( m ln ) Jan.1997-Apr.2010

As Figure 5.3 and Figure 5.4 show, it is obvious that SSE Composite Index and SSE Component Index have a second highest trading volume in February and a forth highest trading volume in January. Moreover, the trading volume in December is merely higher than that of August, September and October. The results indicate that the magnitudes of trading volumes in December and January or in January and February are not higher than in other months of the year.

Table 5.2 Results of the relationship between trading volumes and anomalies Tests of Trading Volumes

January effect T-stat T-stat

β2 5.628 1.229 β2 3.360 1.461

β3 -6.822 -1.239 β3 -3.821 -1.552

β5 -9.606 -2.201* β5 -4.910 -2.218*

β6 28.717 3.686* β6 11.786 3.172*

Chinese New Year effect T-stat T-stat

β2 6.744 1.405 β2 1.657 0.681

β3 0.387 0.073 β3 4.114 1.627

β5 6.667 1.479 β5 3.760 1.657**

β6 -7.948 -1.460 β6 -3.850 -1.578

(28)

trading volume and return in February (January) relative to β1 (β4).

In order to test the correlations between the anomalies and trading volumes, the regression results are presented in Table 5.2. For the January effect, two insignificant β2 indicate that both indexes have insignificantly positive incremental trading volume in January, and two significant β5 show that they both have significantly negative incremental trading volume in December relative to the average trading volume for the 11-month period. Relative to average correlation between trading volume and return for the 11-month period, two insignificant β3 indicate the incremental correlation between trading volume and return in January is insignificantly negative for both indexes. Conversely, two significant β6 mean that the incremental correlation between trading volume and return in December is significantly positive for both indexes. Based on these results, the trading volume in January is not related to the anomaly for both indexes. Moreover, the trading volume in December is not higher than average trading volume in other 11 months, even though the correlation between the return and trading volume in December is higher than other months for both indexes, which indicates that the return in January does not influence the trading volume in December. All these findings are consistent with the previous research that the trading volume is not related with the anomaly.

(29)

February, but significant β5 mean that the incremental trading volume in January is significant at the significance level of 10%. The results of both indexes conclude the trading volume is unrelated with the anomaly in Chinese New Year vacation.

In order to confirm the formal statistic results, I present the trading volume correlations between two serial months in Table 5.3. For example, if long positions built in January are reversed in February to lock the profit, then trading volumes in January and February should be more highly correlated than in February and March, March and April, etc.

Table 5.3 Trading volume correlations of both indexes Summary of Trading Volume Correlation

Month Correlation Correlation

Dec.&Jan. 0.751 0.775 Jan&Feb. 0.570 0.588 Feb.&Mar. 0.264 -0.073 Mar.&Apr. 0.760 0.767 Apr.&May 0.582 0.549 May&Jun. 0.612 0.556 Jun.&Jul. 0.718 0.651 Jul.&Aug. 0.894 0.862 Aug.&Sep. 0.705 0.663 Sep.&Oct. 0.863 0.878 Oct.&Nov. 0.713 0.614 Nov.&Dec. 0.855 0.825

(30)

Chinese New Year does exist, it indicates the investors in Chinese stock market lack the knowledge to arbitrage the abnormally high return in Chinese New Year vacation.

C. Documentations of predicting roles of the anomalies

This part presents the results for the test of predicting role of the anomalies in January and February of Chinese stock market. The regression results are summarized in Table 5.4.

Table 5.4 Results of the examination of predicting roles of two anomalies

Tests of Predicting Roles

January effect T-stat T-stat

β1 0.091 3.240* β1 0.121 4.020*

β2 0.015 4.096* β2 0.017 4.289*

Chinese New Year effect T-stat T-stat

β1 0.068 2.292* β1 0.096 2.984*

β2 0.008 1.234 β2 0.009 1.490

Notes: For the predicting role of return in January, β1 represents the coefficient of January Dummy variable. β2 represents the coefficient of the return of January. For the predicting role of return in Chinese New Year vacation, β1 represents the coefficient of February Dummy variable. β2 represents the coefficient of the return of February.

(31)

(32)

VI Conclusion

This empirical study tests the January effect and Chinese New Year effect in Chinese stock market. Moreover, the evolution patterns of these anomalies are also examined in this paper. According to the results, the January effect does not exist and the Chinese New Year effect is not very significant in Chinese stock market. As regards the evolution pattern, the return in January is in decline while the return of February is getting higher. Besides, the Chinese New Year effect evolves to be stronger. In order to verify the relationship between the anomalies and trading volume, this paper tested several hypotheses. First of all, the trading volume magnitudes of January and February are not higher than other months. Secondly, the correlations between trading volume and the return in January and December or February and January are not significant. Lastly, the trading volume correlations between January and December or February and January are not the highest. Based on all these results, I deem the trading volume is not related with the anomalies in January and February. The third main part of the empirical test is about the predicting roles of the anomalies to the returns of the rest 11 months. As results indicate, the returns in the other 11 months vary positively with the returns of January and February in Chinese stock market.

(33)

market in China is not efficient, here comes up with one potential explanation. It is well known that January effect is significant in small-cap firms. It is possible that the January effect is not significant on SSE Composite Index and SSE Component Index.

With respect to the issue of trading volume, the results show that the trading volume is unrelated with the anomalies in January and February, which are consistent with the studies of foreign financial markets. As stated in the literature, the unrelated trading volume with abnormal return indicates that investors do not actively arbitrage the anomaly. For the January effect does not exist in Chinese stock market, it is reasonable for investors choose not to trade much on the return of January. Unexpectedly, traders also do not trade much on the highly abnormal return of February. This finding also manifests that investors are less sophisticated to arbitrage the abnormal return in Chinese New Year vacation.

Just like the previous studies in foreign financial markets instate, the return in January and February show a strong predicting role of the return in the next 11 months of the year. If investors employ this knowledge, it is very useful and practical for investors to profit from the stock market in China.

(34)

Appendix

A.1 The daily SSE Composite Index from Jan. 1st 1997 to Apr. 30th 2010

SSE Composite Index

0 1000 2000 3000 4000 5000 6000 7000 1997/1/2 2000/1/2 2003/1/2 2006/1/2 2009/1/2 In de x P ri ce

A. 2 The daily SSE Component Index from Jan. 1st 1997 to Apr. 30th 2010

SSE Component Index

0 5000 10000 15000 20000 25000 1997/1/2 2000/1/2 2003/1/2 2006/1/2 2009/1/2 Inde x P ri ce

(35)

significantly since 2006 and achieved the highest point in 2007. However, after 2008, they fell dramatically until 2009.

A.3 The daily trading volume of SSE Composite Index from 1997 to 2010

SSE Composite Index

0 50 100 150 200 250 300 350 1997/1/2 2000/1/2 2003/1/2 2006/1/2 2009/1/2 Tr ad in g v olu m e( m ln )

A.4 The daily trading volume of SSE Component Index from 1997 to 2010 SSE Component Index

0 20 40 60 80 100 120 140 160 180 1997/1/2 2000/1/2 2003/1/2 2006/1/2 2009/1/2 T ra di ng vol um e( m ln)

(36)

(37)

Refernce

Periodicals:

Fama, Eugene, 1970, Efficient Capital Markets: A Review of Theory and Empirical Work, Journal of Finance 25 (2), 383–417.

Lee, C.F., G. Yen, and C. Chang, 1993, Informational efficiency of capital market revisited: anomalous evidence from a refined test, Advances in Quantitative Analysis of

Finance and Accounting 2, 39-65.

Rozeff, M.S., Kinney Jr., W.R., 1976, Capital market seasonality: The case of stock returns, Journal of Financial Economics 3, 379–402.

Keim, D.B., 1983, Size-related anomalies and stock return seasonality: Further empirical evidence, Journal of Financial Economics 12, 13–32.

Keim, D.B., 1989, Trading patterns, bid-ask spreads, and estimated security returns: The case of common stock returns at calendar turning points, Journal of Financial Economics 25, 75–98.

Roll, R., 1981, A possible explanation of the small firm effect, Journal of Finance 36, 879–888.

Roll, R., 1983, Vas ist das? The turn-of-the-year effect and the risk premia of small firms,

Journal of Portfolio Management 9, 18–28.

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, 89–104. Branch, B., Chang, K., 1990, Low price stocks and the January effect, Quarterly Journal

of Business and Economics 29, 90–118.

De Bondt, W.F.M., Thaler, R., 1985, Does the stock market overreact? Journal of

Finance 40, 793–805.

De Bondt, W.F.M., Thaler, R., 1987, Further evidence on investor overreaction and stock market seasonality, Journal of Finance 42, 557–581.

Gultekin, M.N., Gultekin, N.B., 1983, Stock market seasonality: International evidence,

Journal of Financial Economics 12, 469–481.

(38)

Journal of Financial Economics 33, 3–56, 169–178.

Dyl, E., 1977, Capital gains taxation and year end stock market behavior, Journal of

Finance 32, 165–175.

Chen, H., Singal, V., 2004, All things considered, taxes drive the January effect, Journal

of Financial Research 27, 351–372.

Brown, P., Keim, D.B., Kleidon, A.W. and Marsh, T.A., 1983, Stock return seasonalities and the tax-loss selling hypothesis: Analysis of the arguments and Australian evidence,

Journal of Financial Economics 12, 105–127.

Fountas, S., Segredakis, K.N., 2002, Emerging stock markets return seasonalities: The January effect and the tax-loss selling hypothesis, Applied Financial Economics 12, 291– 299.

Riepe, M.W., 1998, Is publicity killing the January effect? The Journal of Financial

Planning 11, 64–70.

Mehdian, S., Perry, M.J., 2002, Anomalies in US equity markets: A reexamination of the January effect, Applied Financial Economics 12, 141–145.

Gu, A.Y., 2003, The declining January effect: Evidence from the US equity markets,

Quarterly Review of Economics and Finance 43, 395–404.

Nicholas Moller, Shlomo Zilca, 2008, The evolution of the January effect, Journal of

Banking & Finance 32, 447–457

Kathryn E. Easterday, Pradyot K. Sen and Jens A. Stephan, 2009, The persistence of the small firm/January effect: Is it consistent with investors’ learning and arbitrage efforts?

The Quarterly Review of Economics and Finance 49, 1172–1193.

Cooper, M., McConnell, J., Ovtchinnikov, A., 2006, The other January effect, Journal of

Financial Economics 82, 315–341.

Chris Stivers, Licheng Sun and Yong Sun, 2009, The other January effect: International, style, and subperiod evidence, Journal of Financial Markets 12, 521–546.

Renliang Zhang, Bin Hu, 1997, The Analysis of Small-cup Stock Effect and Season Effect in Hong Kong Stock Market, Journal of Science 2, 30-34.

Baoxian Zhu, Zhiguo He, 2002, Empirical tests of the relationship between β value and book value, market value and stock returns, Journal of Science 1, 5-63.

(39)

Rui Li, 2003, Month effect of Chinese stock market, Journal of Quantitative & Technical

Economics 7, 37-44.

Shaofu Zhou, 2004, Test of Weekday effect on volatility of Shanghai stock market,

Journal of Science 5, 45-48.

Guodong Xu, Xiangxin Tian and Binhong Lin, 2004, Empirical test of Season effect in Chinese stock market, Journal of Guangxi Financial Colleage 1, 63-66

Xuan Zhang, 2004, A based on Quantile Regression model research of risk measurement of Chinese stock market, Journal of WUT 7, 25-31.

Jones, S. l., W. Lee, and R. Apenbrink, 1991, New Evidence on the January Effect before Personal Income Taxes, Journal of Finance 46, 1909-1924.

Lee, C.F., G. Yen, and C. Chang, 1993, Informational Efficiency of Capital Market Revisited: Anomalous evidence from a refined test, Advances in Quantitative Analysis of

Finance and Accounting 2, 39-65.

Jianxin Gu, Jun Zhang, 2005, Empirical Test Of Chinese New Year Effect In Chinese Stock Market, Journal of WUT (INFORMATION&MANAGEMENT ENGINEERING) 6, 141-144.

Xi Zhang, Canqiu Wei, 2007, Empirical test of Chinese New Year Effect In Chinese Tourism Stocks, China Academic Journal Electronic Publishing House 1,133-134. Mills, T.C., Coutts, J.A., 1995, Calender Effects in the London Stock Exchange FT-SE Indices, European Journal of Finance 1, 79-93.

Monograph (Books):

Haugen, R.A., Lakonishok, J., 1988, The incredible January effect (Dow Jones Irwin, Homewood, IL).

O'Hara, Maureen, 1995, Market Microstructure Theory (Blackwell, Oxford).

Contributions to Collective Work:

Booth, D.G., Keim, D.B., 2000, Is there still a January effect? in Keim, D.B., Ziemba, W.T. (Eds.), Security Market Imperfections in Worldwide Equity Markets (Cambridge University Press, New York).

Hensel, C.R., Ziemba, W.T., 2000, Anticipation of the January small firm effect in the US futures markets, in: Keim, D.B., Ziemba, W.T. (Eds.), Security Market Imperfections in Worldwide Equity Markets (Cambridge University Press, New York)

(40)

Rendon, J., Ziemba, W.T., 2005, Is the January effect still alive in the futures markets? Unpublished working paper, University of British Columbia.

Yue Hu, 2005, Some empirical tests for the January effect, Unpublished working paper, Simom Fraser Univeristy.

Institutes and Foundations:

Yuhsuan Yang, Xinyi Hu, Yirong Chen and Shijie Jiang, 2006, An empirical study of Chinese New Year effect in stock markets, Grey System Theory and Application Seminar, October.

Research/Data Sources:

JANUARY EFFECT AND CHINESE NEW YEAR EFFECT IN CHINESE STOCK MARKET