Wenzhou-Kean University Herding Behaviors of China’s Stock Market Lou Qiyue

Wenzhou-Kean University
Herding Behaviors of China’s Stock Market
Lou Qiyue (0952598)

TOC o “1-3” h z u Abstract PAGEREF _Toc512455217 h 1 HYPERLINK l “_Toc512455218” I. Background PAGEREF _Toc512455218 h 1
HYPERLINK l “_Toc512455219” II. Problem Statement and Methodology PAGEREF _Toc512455219 h 3
HYPERLINK l “_Toc512455220” III. Literature Review PAGEREF _Toc512455220 h 4
HYPERLINK l “_Toc512455221” IV. Descriptive Testing of Herding PAGEREF _Toc512455221 h 12
4.1 introduction of methodology and data sources PAGEREF _Toc512455222 h 124.2 data ; charts PAGEREF _Toc512455223 h 12 HYPERLINK l “_Toc512455224” V. Summary and Conclusion PAGEREF _Toc512455224 h 18
Appendix Data of Trading Volumes and Listed Securities from SSE and SZSE PAGEREF _Toc512455225 h 19Reference PAGEREF _Toc512455226 h 21

AbstractThis paper aims to launch a descriptive analysis of herding behaviors in China’s security markets by collecting and observing trading volumes data and comparing with the corresponding market conditions. The two methods employed by this paper is literature review and descriptive statistical method. Five famous existing researches of herding behaviors are reviewed in this paper and leaves two opposite conclusions about if there is herding in Chinese stock market.

The descriptive analysis of trading volumes data retrieved from Shanghai Stock Exchange (SSE) and Shenzhen Stock Exchange (SZSE) gives a clearer indication of herding in China’s stock market. The observations find that trading volume swings as the market condition changes, which indicates that herding do exist in China’s market. <back>I. BackgroundIn today’s financial world, behavioral finance and phenomena under this category was more and more emphasized by scholars because people realized that human beings is the main part of the market, and humans pursue biological activities instead of mechanical motions.

In the behavioral finance, herding is usually used to describe the correlation in trades resulting from interactions between investors CITATION Chi10 l 2052 (Chiang ; Zheng, 2010). In another word, herding behavior represents a phenomenon that most people in security markets will follow several big investors’ steps to determine their investment decisions. The result of herding, as Dr. Nofsinger and Dr. Sias noting, is “a group of investors trading in the same direction over a period of time” CITATION Nof99 l 2052 (Nofsinger & Sias, 1999). This paper focuses on herding behaviors in China’s stock market since China has more private investors than western countries. It seems rational that herding behavior appears more frequently in less develop financial markets. However, according to Dr. Chiang and Dr. Zheng’s research CITATION Chi10 l 2052 (Chiang & Zheng, 2010), the statistical significance of herding behavior is relatively low in some less developed areas such as Latin America. Thus, I am curious whether the herding in China’s stock market is significant or not. If the herding behavior is significant, then it is important to figure out the reason (or unique reasons of China’s stock market) and to determine if it is a good nature to the market.

This research paper differs from previous researches in the following aspects. Previous researches emphasize more on data analysis and economics modeling. This research introduces another perspective of testing herding behaviors from the scope of investors, which makes materials more intuitive and understandable. Also, this research is aim to specifically Chinese stock market and will have a discussion on why or why not herding exists in Chinese market and whether herding is good to Chinese investors or not.

The rest of this paper is organized as follows. Part II states the problems of current researches and the methodology employed by this research; part III is the literature review of variety existing researches and their conclusions; part IV is the research works done by myself; and the last part, part V, is the conclusion from this research. <back>II. Problem Statement and MethodologyThe research objective is to use a more understandable and straightforward way to interpret whether the herding behaviors in Chinese market it significant or not. So far, there are already some researches that focus on or mention China’s stock market. However, the methodology they used were all econometric modeling and statistics test. Such kind of paper requires readers to have some knowledge about statistics and econometrics, which is not easy to understand the process of how to derive the results. Also, the uniqueness of China’s stock market, for instance, strong interference from the government and prevalent of private speculators, makes the problem of the herding in China’s market more complex and some models did not consider those factors. So far, it is not clear in the past researches that what kinds of features of stock market can affect the significance of herding behavior. In Dr. Chiang and Dr. Zheng’s research, they established a model that only contains Rm, t (as well as its absolute term and its square term), where Rm, t is “the cross-sectional average stock of N returns in the portfolio at time t” CITATION Chi10 l 2052 (Chiang & Zheng, 2010). Dr. Cipriani and Dr. Guarino’s model considered variety kinds of factors, such as type of assets, the market, the market maker, and traders CITATION Cip12 l 2052 (Cipriani ; Guarino, 2012).

The main methodology that this research employs is descriptive analysis of trading volumes in Chinese stock market. This is the primary part of the paper that different from existing researches. Data analysis and econometrics modeling will appear in the literature review part. What’s more, literature review is also an important methodology appeared in this research. Previous researchers had already done lots of examinations of herding in other markets, which are very good templets for this research. Data used in this research will be all secondary data from annual statistics books of Shanghai and Shenzhen Stock Exchanges. Spread sheets and statistical graphs will be made and put into the main body of part IV as well as the appendix. If the trading volume experienced a relatively large swing during extreme market conditions, either boom or recession, then it indicates that there is herding behaviors in Chinese stock markets. Otherwise, we cannot say there is a significant evidence of the herding from descriptive results and deeper quantitative test are required. <back>III. Literature ReviewThere are four famous existing researches that are related to testing herding behaviors in Chinese market. They are Christie and Huang (1995), Chang et al. (2000), Demirer and Kutan (2006), and Tan et al. (2007). In addition, Chiang et al. (2010) (hereafter referred as CHI) also mentioned Chinese stock market as one of the emerging markets within that researches’ range of study.

Christie and Huang (1995) (hereafter referred as C;H) suggests that market condition is the primary factor which affects the decision-making process of investors. In normal periods, different people have different analysis to the market and they may take diverse investment strategies under the logics of the rational asset pricing model. However, in extreme conditions, such as stock market bubble and market depression, investors may lose their own ideas and mimic what others did, as stated by C;H. Based on this, they established a model that can be represent by the following equation:
St=?+?LDtL+?UDtU+?twhere St is return deviation at time t, DtL is a dummy variable at time t taking on the value of unity when the market return at time t lies in the extreme lower tail of the distribution and taking the value of 0 otherwise. DtU is also a dummy variable with a value of unity when the market return at time t lies in the extreme upper tail of the distribution and taking the value 0 otherwise CITATION Chr95 l 2052 (Christie ; Huang, 1995). C;H also carried out the cross-sectional standard deviation (CSSD) to measure the return deviation St, which can be expressed by the following formula:
CSSDt=i=1nRit-Rmt2n-1where n is the sample space, that is, the numbers of companies in portfolio, Rit is the stock return from the firm i at time t, and Rmt is the average of returns of those n stocks at time t. C;H asserted that investors will make similar decisions when herding occurs, which could lead to lower return deviation. Therefore, statistically significant negative values for ? L and ? U would tell us that the herding happens.

Demirer and Kutan (2006) (hereafter referred as D;K) employed the model established by C;H to test if there is herding behaviors in China’s stock market. Their first data set was retrieved daily return data of 375 Chinese stocks on Shanghai and Shenzhen Stock Exchanges from January 1999 to December 2002 from Sinofin (http://www.sinofin.net). Their second data set contains daily sector indexes of Shanghai and Shenzhen Stock Exchanges obtained from Taiwan Economic Journal Financial Database within the period of May 3, 1993 to November 16, 2001 (totally 1860 days). These data were consisted of four sectors: industry, commerce, realty, and utility CITATION Kut06 l 2052 (Kutan & Demirer, 2006). After D&K run the regression model employed from C&H’s research, they concluded that there is no evidence of herding behaviors in Chinese stock market.

It is surprising that D;K made such a conclusion. As for the reason behind why they came to this conclusion, Tan et al. (2007) (hereafter referred as TAN) provided us some clues. According to TAN, one of the challenges of C;H’s model is how to define the extreme returns. TAN discovered that C&H had noted that this definition was arbitrary, and they used 1% and 5% as the cutoff points to identify the upper and lower tails of the return distribution (which is the ?-value if using the language of statistics). TAN also noted that, in real-world, investors might have different opinions about what is extreme returns, which might change characteristics of return distribution. What’s more, TAN also asserted that herding may “occur to some extent over the whole return distribution but become more obvious during market stress periods”. Hence, TAN indicated that the reason why C;H’s model could leading to the conclusion that no herding evidence in Chinese stock market is that C&H’s method only captures herding during periods of extreme returns. TAN also proposed another challenge of C;H’s model, which was the relatively short history of Chinese stock market comparing to developed markets, which makes it is hard for investors to identify when extreme returns occur.

Chang et al. (2000) (hereafter referred as CHA) proposed an alternative model to test the herding. CHA’s research was not limited to Chinese market only. They examined variety markets around the world. Some are developed markets, like US and Hong Kong; some others are emerging markets, such as Taiwan and South Korea. Instead of using C;H’s CSSD to measure the return dispersion, CHA’s paper carried out another measurement, that was, the cross-sectional absolute deviation of returns (CSAD). CHA demonstrated that “rational asset pricing models predict not only that equity return dispersions are an increasing function of the market return but also that the relation is linear.” They proved this statement by using the CAPM model as well as the absolute value of the deviation (AVD) of security’s expected return. Let Ri denote the return of the asset I in the portfolio, Rm denote the return of market portfolio, and Et (*) be the expected value of variable * at time t, the CAPM model CHA’s paper used can be expressed as the following equation:
Et (Ri) =?0+?i E (Rm – ?0)
where ?i is the beta-value (measures systematic risks) of asset i, ?0 is the return of risk-free portfolio. In addition, CHA also defined ?m as the beta-value of an equally-weighted market portfolio, which has the following relationship between ?i:
?m=1ni=1n?iwhere n is the total numbers of assets in the portfolio. Then, CHA defined AVD of asset i’s expected return at time t:
AVDit =| ?i-?m | E (Rm – ?0)
Finally, CHA carried out the expected value of CSAD (ECSAD) at time t:
ECSADt=1ni=1nAVDit=1ni=1n?i-?mERm-?0Through this formula, one can easily show that ECSAD, which measures return dispersion, is an increasing linear function of the expected market return E (Rm) by taking the first-order and the second-order partial derivatives with respect to E (Rm):
?ECSADt?ERm=1ni=1n?i-?m>0?2ECSADt?ERm2=0Based on the above patterns, CHA established an alternative model to test herding that can be expressed by the following equation:
CSADt=?+?1Rmt+?2Rmt2+?rwhere | Rmt | is “the absolute value of an equally-weighted realized return of all available securities on day t” CITATION Cha00 l 2052 (Chang, Cheng, & Khorana, 2000). CHA noted that during the periods of relatively large price volatility, investors do indeed herding behaviors around market indicators, which may result a decreasing non-linear relation between CSADt and the average market return. Thus, herding will be captured by a negative and statistically significant of ?2 according to CHA’s model. After running the regression, CHA concluded that there was no evidence of herding in “part of market participants in the US and Hong Kong”, “partial evidence of herding in Japan” CITATION Cha00 l 2052 (Chang, Cheng, ; Khorana, 2000), and they did find evidence of herding in South Korea and Taiwan, two emerging markets in samples.

Tan et al. (2007) (TAN as referring in previous paragraphs) introduced CHA’s model into their research. However, they replaced CSADt in CHA’s model with CSSDt (still denoted as CSADt in TAN) proposed by Christie and Huang (1995) (C;H as referring in previous paragraphs). According to TAN, his CSAD differs from CHA’s in two aspects: 1) TAN’s CSAD does not need to estimate beta-value (the estimation may cause error), and 2) TAN’s model removes the assumption of constant risk over time. TAN’s research collected data of four portfolios of Chinese stocks: Shanghai Stock Exchange A shares (SHA), Shanghai Stock Exchange B shares (SHB), Shenzhen Stock Exchange A shares (SZA), and Shenzhen Stock Exchange B shares (SZB). The period was from July 12, 1994 to December 31, 2003. The stock returns in rime t was computed by Rt=100×lgPtPt-1, where Pt stands for the stock price at time t, and lg is the notation of common logarithm (logarithm with base 10). To consider the Asian financial crisis in 1997 into the model to make it more robust, TAN modified their model a little bit based on CHA’s model by adding a dummy variable DMt. TAN’s model can be formulated by the following equation:
CSADt=?+?1Rmt+?2Rmt2+?3Rmt2DMt+?rwhere DMt takes the value of 1 during the Asian Crisis from July 2, 1997 to November 17, 1997, and the value of 0 otherwise. TAN tested the herding for daily, weekly, and monthly data and reached the conclusion that both A and B shares in Shanghai and Shenzhen Stock Exchanges have significant evidence of herding based on daily data, and the evidences are less significant based on weekly and monthly data. This result indicates that “herding is a phenomenon confined to short time horizons.” CITATION Tan08 l 2052 (Tan, Thomas, Mason, ; Nelling, 2008) TAN also tested potential asymmetries in herding with respect to market returns, trading volume, and volatility. Their conclusions were: 1) herding presents in all four markets when market was rising, and trading volume and volatility are high, SHA was much stronger significant, and 2) there was no such asymmetric characteristic in B-share markets. TAN thought that the causation of these two conclusions was the different groups of investors of Chinese A- and B-share markets. In Chinese A-share market, most investors are private speculators, which are more likely to pursue the herding when extreme return occurs; whereas B-share market was dominated by foreign institutional investors that are much more rational when facing with extreme conditions.

Chiang et al. (2010) (hereafter referred as CHI) is another paper that tested herding behaviors of Chinese stock market. CHI examined variety markets around the world, which were much broader than TAN. Those markets were listed by the following table:
Regions Full Market Names Abbreviation
Advanced Markets Australia AU
France FR
Germany GR
Hong Kong HK
Japan JP
United Kingdom UK
United States US
Latin American Markets Argentina AR
Brazil BR
Chile CL
Mexico MX
Asian Markets China CN
South Korean KR
Taiwan TW
Indonesia ID
Malaysia MY
Singapore SG
Thailand TH
The model that CHI used to test Latin American and Asian markets is similar to TAN (Thomas C. Chiang is the second author of TAN), which can be expressed by the following equation:
CSADt=?+?1Rmt+?2Rmt2+?3Rmt2+?4CSADUS, t+?5RUS, mt2+?rCSADt=?+?1Rmt+?2Rmt2+?3Rmt2+?4CSADUS, t-1+?5RUS, mt-12+?rThe reason why the term with ?4 and ?5 appears was that there is a time lag between American markets and Asian and European markets. Therefore, for US and Latin American markets, CHI used the upper model with CSADUS, t and Rus, mt and otherwise using the alternative model with subscript “US, t-1″. The testing results of CHI was kind of interesting. It was contrast to Chang et al. (2000) (CHA as referring in the previous paragraphs) ” no herding in advanced markets” and Demirer and Kutan (2006) (D;K as referring in the previous paragraphs) ” no herding in Chinese markets”. CHI found that most investors not only herd with their domestic markets, but also with US markets, which indicates that previous researches excluding foreign markets in their models could produce biased estimations. There was a special case in CHI’s sample, that was, Latin American investors usually herd with US markets only instead of their domestic markets. That is why there was no significant evidence of herding found in Latin American markets. Another important conclusion of CHI was that financial crisis was often play as the trigger of herding behaviors in the original country and then spread to nonboring countries and caused herding in more and more countries stock markets.

As we can see from those existing researches, they all employed the methodology of establishing modeling and statistical test. The advantage of this methodology is that it is the method that is universally accepted by the academic fields. However, I also noted that different papers using different models (although they are very similar, but still a little differences) and may come up with different conclusions (D&K and CHI). In addition, these papers require readers to have knowledge from calculus, statistics and econometrics, which are relatively difficult for ordinary readers to read. Hence, a more straightforward and understandable examination of the herding behavior is needed. <back>IV. Descriptive Testing of Herding4.1 introduction of methodology and data sourcesAs stated in part II, descriptive analysis is one of the methodologies employed by this paper (another is literature review). Data were collected from the official website of Shanghai Stock Exchange (http://www.sse.com.cn) (hereafter referred as SSE) and Shenzhen Stock Exchange (http://www.szse.cn) (hereafter referred as SZSE).

The first part of descriptive analysis is the collection of annually trading volume. Two spread sheets of annually trading volumes for each types of securities in SSE and SZSE would be made and corresponding line charts will show the changes of every year’s trading volume for each security. The period of data is from 2007 to 2016. Due to the limitation of data accessibility, there are no data of 2017. During this 10-year period, there were two main crises occurred, the 2008 Crisis and the 2015 Crash. We can compare the trading volumes and see if we can find the evidence of the herding during extreme market conditions.

The second part of descriptive analysis is a microscopic analysis. In this part, monthly trading volumes data would be carried out. I will pick 2008’s, 2009’s, 2015’s, and 2016’s monthly trading volumes data to compare because Chinese stock market had experienced from booming to crash in 2008 and 2015 as well as from crisis to recover in 2009 and 2016. The source of data are the monthly statistics books from SSE and SZSE.

4.2 data ; chartsThe spreadsheets of raw data are put in the appendix. The following are several charts created from the annual data of trading volumes using Excel:

From the first charts (SSE Total Trading Volumes), we can observe that for total trading volume, there are two local extrema occurring in 2007 and 2009, which was the year before 2008 Crisis and the year after 2008 Crisis. Checking past news, it was not hard to find that Chinese stock market experienced a booming period in 2007 (because of the market bubble) and another booming period in 2009 (because Chinese government declared to inject 4 trillion RMB into market in 2 years). During the 2008 Crisis, the total trading volume drops about 100 thousand comparing with both 2007 and 2009.

Chart 2 (SSE Share, Bond, ; Fund’s Trading Volumes) gives us a more specific interpretation. We can see that there are three peaks of the stock trading volumes in 2007, 2009, and 2015 respectively. They are exactly corresponded with three periods of market booming (China experienced another market booming in the first half year of 2015 and then suffered crash). Chart 3 (SZSE Total Trading Volumes) presents the same patterns of Chart 2’s stock trading volumes. From these three charts, we can see that investors traded their security a lot during the booming period and traded less during the recession. It is a typical phenomenon of herding behavior since herding behavior is defined as investors do what market indicators did in extreme market conditions. Thus, it is fair to conclude that there is relatively significant evidence of herding in China’s security markets, especially in the stock market.

SZSE’s annual statistics books also provided the top 10 records of trading volumes from 1991 to the year that particular book was published. The following three tables show the top 10 records from 1991 to 2008, 2009, and 2016, respectively:
Top 10 Trading Volume Records (1991~2008)
Rank Value (M) Month/Year ? Rank Value M/D/Y
1 2,812,454.41 06/2007 ? 1 178,190.24 06/13/2007
2 2,269,151.79 08/2007 ? 2 172,155.30 06/12/2007
3 2,170,142.03 05/2007 ? 3 163,340.96 05/30/2007
4 1,864,460.23 04/2007 ? 4 157,182.30 05/31/2007
5 1,841,907.87 09/2007 ? 5 154,836.66 06/14/2007
6 1,736,423.16 01/2007 ? 6 148,353.81 06/11/2007
7 1,586,413.04 07/2007 ? 7 148,045.91 06/01/2007
8 1,310,958.01 10/2007 ? 8 141,734.31 06/20/2007
9 1,206,582.86 03/2007 ? 9 140,273.81 05/24/2007
10 1,118,531.11 12/2007 ? 10 140,036.16 05/28/2007
Top 10 Trading Volume Records (1991~2009)
Rank Value (M) Month/Year ? Rank Value M/D/Y
1 2,812,454.41 06/2007 ? 1 181,998.90 11/24/2009
2 2,546,604.14 11/2009 ? 2 178,190.24 06/13/2007
3 2,457,043.76 07/2009 ? 3 172,155.30 06/12/2007
4 2,269,151.79 08/2007 ? 4 166,071.76 11/26/2009
5 2,170,142.03 05/2007 ? 5 163,340.96 05/30/2007
6 2,091,862.31 12/2009 ? 6 157,182.30 05/31/2007
7 1,864,460.23 04/2007 ? 7 154,836.66 06/14/2007
8 1,841,907.87 09/2007 ? 8 154,543.68 12/04/2009
9 1,831,001.77 08/2009 ? 9 148,353.81 06/11/2007
10 1,736,423.16 01/2008 ? 10 148,045.91 06/01/2007
Top 10 Trading Volume Records (1991~2016)
Rank Value (M) Month/Year ? Rank Value M/D/Y
1 18,260,260.24 06/2015 ? 1 1,204,895.11 05/28/2015
2 16,125,789.17 05/2015 ? 2 1,111,248.57 06/05/2015
3 14,180,724.63 11/2015 ? 3 1,072,422.67 05/27/2015
4 13,882,748.60 04/2015 ? 4 1,053,263.33 05/26/2015
5 13,617,935.58 07/2015 ? 5 1,052,824.11 06/08/2015
6 12,285,933.73 12/2015 ? 6 1,047,552.09 06/03/2015
7 10,840,092.53 08/2015 ? 7 1,024,763.13 05/22/2015
8 10,596,895.22 03/2015 ? 8 1,017,349.39 06/04/2015
9 9,472,690.14 10/2015 ? 9 1,014,840.92 06/15/2015
10 9,427,553.40 11/2016 ? 10 1,014,597.52 06/02/2015
These three tables give us clearer evidences of herding behaviors in China’s security markets. We can observe that from 1991 to 2008, all the top 10 trading volumes occurred in 2007. When moving into 2009, although some records of 2009 came into the top-10 records, 2007’s records was still the most among all the top-10 trading volumes. As for the records from 1991 to 2016, the most records occurred in 2015, which was still the year that China’s security market experienced extreme conditions. <back>
V. Summary and ConclusionThis paper pursued a descriptive analysis of herding behaviors in China’s security markets by collecting and observing trading volumes data and comparing with the corresponding market conditions. The two methods employed by this paper is literature review and descriptive statistical method. The literature review parts reviewed five famous existing researches that testing the herding in Chinese market by using econometrics modeling. They came up with two opposite conclusions: Demirer and Kutan (2006) concluded that there is no evidence that Chinese investors have herding behaviors, but results of Tan et al. (2007) and Chiang et al. (2010) showed that there do have evidence of herding in Chinese market.

The descriptive analysis and testing proposed by this paper gives more straightforward clues to the herding behaviors in Chinese market. From data provided by Shanghai Stock Exchange (SSE) and Shenzhen Stock Exchange (SZSE), we can see that the trading volumes reached extreme values once the market conditions reach the extreme (either booming or recession). Consider the definition of herding, it is fair to conclude that there are relatively significant evidences of herding in Chinese market. And the evidence is more significant especially in stock market. ;back;Appendix Data of Trading Volumes and Listed Securities from SSE and SZSESSE Trading Volumes and Listed Securities
? 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
No. of Trading days 242 246 244 242 244 243 238 245 244 244
No. of Listed Securities 1125 1184 1351 1500 1691 2098 2786 3758 5914 9647
Share 904 908 914 938 975 998 997 1039 1125 1226
A-share 850 854 860 884 921 944 944 986 1073 1175
B-share 54 54 54 54 54 54 53 53 52 51
Bond 198 246 411 536 680 1059 1731 2646 4538 8130
T-bond 62 77 160 199 213 191 218 267 893 1620
C-bond 82 112 192 284 417 830 1468 2336 3596 6457
Repo 54 57 59 53 50 38 45 43 49 53
Fund 17 16 18 25 36 41 58 68 135 161
Closed Fund 14 13 13 13 13 12 9 3 4 3
ETF 3 3 5 12 23 29 47 61 73 76
LOF N/A N/A N/A N/A N/A N/A N/A N/A 47 59
Money Market Fund N/A N/A N/A N/A N/A N/A 2 4 11 23
Warrant 6 14 8 1 0 0 N/A N/A N/A N/A
Preferred Share N/A N/A N/A N/A N/A N/A N/A 5 16 24
Option N/A N/A N/A N/A N/A N/A N/A N/A 100 106
? ? ? ? ? ? ? ? ? ? ?
Trading Volume (100M) 380,025.57 271,842.03 441,874.65 398,395.72 454,646.41 547,450.83 865,094.23 1,281,492.46 2,663,610.59 2,838,711.93
Share 305,434.28 180,429.95 346,511.91 304,312.01 237,555.31 164,460.86 230,266.03 377,162.12 1,330,992.10 501,700.42
A-share 301,960.29 179,762.44 345,443.26 303,215.93 236,809.12 164,047.38 228,918.82 375,149.95 1,323,231.16 496,880.34
B-share 3,473.99 667.51 1,068.65 1,096.08 746.19 413.48 689.94 484.45 2,357.12 984.49
Repo N/A N/A N/A N/A N/A N/A 657.27 1,527.72 5,403.82 3,835.59
Bond 20,399.38 28,090.64 39,806.33 74,914.43 210,714.87 379,818.85 625,839.41 866,848.58 1,228,533.71 2,247,175.21
T-bond 1,262.20 2,075.90 2,055.50 1,590.04 1,243.11 905.56 771.60 1,247.47 4,330.72 7,773.00
C-bond 528.25 1,707.97 1,821.58 3,306.79 4,850.47 7,537.43 14,540.88 24,198.95 26,350.38 36,050.28
Repo 18,608.93 24,306.77 35,929.25 70,017.60 204,621.29 371,375.86 610,526.93 841,402.16 1,197,852.61 2,203,351.93
Fund 4,298.25 3,700.23 6,549.05 4,771.70 2,901.41 3,171.12 8,988.79 37,477.49 103,799.88 89,359.23
Closed Fund 3,100.51 961.16 790.29 569.89 202.28 144.53 231.86 193.00 684.04 212.80
ETF 1,197.74 2,739.07 5,758.76 4,201.81 2,699.13 3,026.59 6,706.52 10,142.68 28,890.51 6,510.84
LOF N/A N/A N/A N/A N/A N/A N/A N/A 1,367.36 610.62
Money Market Fund N/A N/A N/A N/A N/A N/A 2,050.41 27,141.81 72,857.97 82,024.97
Warrant 49,893.66 59,621.21 49,007.36 14,397.58 3,474.82 0.00 N/A N/A N/A N/A
Preferred Share N/A N/A N/A N/A N/A N/A N/A 4.27 48.24 45.18
Option N/A N/A N/A N/A N/A N/A N/A N/A 236.66 431.89
(Source: Shanghai Stock Exchange)
SZSE Trading Volumes and Listed Securities
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
No. of Trading days 242 246 244 242 244 243 238 245 244 244
Share 712 782 872 1211 1453 1581 1577 1657 1784 1908
Bond 101 131 237 286 334 381 460 527 1179 2057
Fund 48 48 55 93 151 228 291 339 476 511
Warrant 7 3 1 0 0 0 0 0 0 N/A
Preferred Share N/A N/A N/A N/A N/A N/A N/A N/A N/A 5
No. of Listed Securities 868 964 1165 1590 1938 2190 2328 2523 3439 4481
Total Trading Volume (M) 18,764,557.09 9,938,843.82 19,873,386.59 24,742,661.96 19,318,832.67 17,865,969.05 29,667,146.08 44,470,819.04 136,105,131.51 93,442,586.13
(Source: Shenzhen Stock Exchange)
Reference BIBLIOGRAPHY l 2052 Chang, E. C., Cheng, J. W., ; Khorana, A. (2000). An Examination of Herd Behavior in Equity Markets: An International Perspective. Journal of Banking ; Finance, pp. 1651-1679.

Chiang, T. C., ; Zheng, D. (2010, Jan. 4). An Empirical Analysis of Herd Behavior in Global Stock Markets. Journal of Banking ; Finance, pp. 1911-1921.

Christie, W. G., ; Huang, R. D. (1995, July). Following the Pied Piper: Do Individual Returns Herd around the Market? Financial Analysts Journal, pp. 31-37.

Cipriani, M., ; Guarino, A. (2012, June 9). Estimating a Structural Model of Herd Behavior in Financial Markets. Social Science Electronic Publishing, pp. 224-251.

Kutan, A. M., ; Demirer, R. (2006, Feb.). Does Herding Behavior Exist in Chinese Stock Markets? Journal of International Financial Markets Institutions and Money, pp. 123-142.

Nofsinger, J. R., ; Sias, R. W. (1999, Dec.). Herding and Feedback Trading by Institutional and Individual Investors. Journal of Finance, pp. 2263–2295.

Shanghai Stock Exchange. (2018). Statistical Annual. Retrieved from Shanghai Stock Exchange: http://www.sse.com.cn/aboutus/publication/yearly/
Shenzhen Stock Exchange. (2018). Annual Statistical Books. Retrieved from Shenzhen Stock Exchange: http://www.szse.cn/main/marketdata/wbw/marketstat/
Tan, L., Thomas, C. C., Mason, J. R., ; Nelling, E. (2008). Herding Behavior in Chinese Stock Markets: An Examination of A and B Shares. Pacific-Basin Finance Journal, pp. 61-77.