Market segmentation and information diffusion in China’s stock markets

Journal of Multinational Financial Management

10 (2000) 421–438

Market segmentation and information diffusionin China’s stock markets

Boo Sjoo a,b, Jianhua Zhang a,*a Department of Economics, School of Economics and Commercial Law, Goteborg Uni6ersity,

P.O. Box 640, Vasagatan 1, 405 30 Goteborg, Swedenb Department of Economic and Political Sciences, Uni6ersity College of Sko6de,

541 28 Sko6de, Sweden

Received 15 July 1999; accepted 20 March 2000

Abstract

This study analyses the information diffusion between Chinese A shares (restricted todomestic investors) and B shares (restricted to foreign investors). The results show that thereis an important long-run information diffusion between A and B shares. In the Shanghaistock market, information flows from foreign to domestic investors. However, in the smallerand less liquid Shenzhen stock market, the information diffusion goes in the opposite way.The direction of the information diffusion is determined by the choice of stock exchangerather than firm size. © 2000 Elsevier Science B.V. All rights reserved.

JEL classification: G12; G14

Keywords: Information flow; Information diffusion; A and B shares; Premium

www.elsevier.com/locate/econbase

1. Introduction

Firms often issue different types of equity to discriminate between differentinvestors. In China, firms are required to discriminate between domestic andforeign investors to ensure that ownership remains under Chinese control. Domes-tic investors can only buy A shares and foreign investors can only buy B shares.The shares are identical in terms of voting power and dividend claims. Due to the

* Corresponding author. Tel.: +46-31-7732689; fax: +46-31-7731043.E-mail address: [email protected] (J. Zhang).

1042-444X/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.

PII: S1042 -444X(00 )00035 -9

B. Sjoo , J. Zhang / J. of Multi. Fin. Manag. 10 (2000) 421–438422

existing regulations, the amount of outstanding B shares is always smaller, soforeign investors are forced to be minority shareholders. The outcome is that theequity of the same firm is traded at the same time, at the same exchange, but by twodifferent investor groups and at quite different prices. Typically, A shares trade ata premium over B shares. Moreover, the premium is not constant. It changes overtime in a way that resembles an integrated stochastic process.

This study tests a number of aspects concerning the observed informationdiffusion between A and B shares in China’s emerging stock market (ESM). Theobjectives are to learn more about the role of foreign investors in ESMs and toinvestigate where price information is produced.

Several factors can cause information diffusion between domestic and foreigninvestors in emerging markets. First, the foreign investors in China are mainly bigfinancial institutions. Compared with the domestic investors, foreign institutionalinvestors can in general be assumed to be more experienced, have better means ofobtaining information, and have access to more advanced technology to analyzedata. Thus, the presence of foreign investors can be a buy signal for the relativelyuninformed domestic investors. In this situation, the prices of B shares would leadthose of A shares reflecting that domestic investors get information from foreigninvestors.

Second, the domestic investors might have the information advantage. They canbe better in acquiring relevant news from local sources. In this case, the prices ofA shares would lead the prices of B shares, because of foreign investors learningfrom domestic investors. Third, it follows from the discussion that the priceinformation can flow in both directions. Different investor groups can havedifferent comparative advantages in acquiring information. Finally, as an extremecase, the markets for A and B shares might be completely segmented, showing nocorrelation and lead-lag relations what so ever. Foreign investors can face severepolitical risk in emerging financial markets, and they might form quite differentconditional expectations about the future prospects of the Chinese economy ingeneral and of the cash flow of the individual firms in particular.

Earlier studies on Chinese stock markets have focused on either the price premia,or on the lead-lag structure between the returns of A and B shares. Chui and Kwok(1998) investigate the cross-autocorrelation structure of A and B share returns inChina. Their conclusion is that the returns on B shares lead the returns on Ashares. This result, however, is based on an implicit assumption of a completelong-run segmentation between A and B shares. There is no ground for makingsuch an assumption about the relationship between the prices of A and B shares. Infact, it is natural to assume that the difference between the price levels of A and Bshares contains information of coming returns. We test this hypothesis and investi-gate its consequences on the flow of information by modeling the prices as amultivariate vector error correction process1.

1 Harris et al. (1995) discuss cointegration in stock transactions data. They focus on specification andestimation of an error correction mechanism for IBM price on different exchanges in order to investigatethe price-discovery theory.

B. Sjoo , J. Zhang / J. of Multi. Fin. Manag. 10 (2000) 421–438 423

Our results support the view in Chui and Kwok (1998), that information flowsfrom foreign investors to domestic investors, but only for the Shanghai market. Weobtain this result for both the short and the long run. In the smaller Shenzhenmarket, the causality is more ambiguous. Here, foreign investors affect returns onlyin the short run. In the long run, information flows from domestic to foreigninvestors. Our results suggest that the most important factor for whether informa-tion flows from domestic or foreign investors is the choice of stock exchange ratherthan firm size.

The study is organized as follows. Section 2 discusses theories and hypothesesrelated to this study. Section 3 explains the use of cointegration and vector errorcorrection models for analyzing information diffusion. Section 4 presents theempirical results. Finally, Section 5 concludes the study.

2. Theoretical framework and hypotheses

The type of information diffusion that we observe between A and B shares on theChinese stock markets is related to the price-discovery theory, the small-firm effectand studies on the general behavior of institutional investors. The theory of pricediscovery attempts to determine ‘the process whereby markets attempt to findequilibrium’ (Schreiber and Schwartz, 1986)2. We address a similar type of problemin this study by asking whether new information is produced in A-share markets orin B-share markets.

The information flow between firms with large market capitalization and firmswith small capitalization is investigated by Lo and MacKinlay (1990), Chan (1993),among others. The existing evidence shows that the returns of large firms’ stockslead those of small firms’ stock. Large firms’ stocks are more liquid than smallfirms’ stocks. Thus, new information reflected in the shares of large firms by the endof the trading day will be reflected in small firms’ stocks in the following day.

Other explanations build on imperfect information and institutional investors.Chan (1993) argues that information from large firms is of better quality than thatfrom small firms. Thus, investors usually focus on large firms. Market makersadjust the prices of small stocks after observing previous price changes of largestocks. Badrinath et al. (1995) argue that the returns on institutionally favoredstocks are leading the returns on stocks not favored by institutional investors. Theirempirical study on the US stock markets supports this hypothesis. According to theinformation hypothesis (Bailey and Jagtiani, 1994), foreign investors prefer toinvest in larger domestic firms where the financial disclosure and informationavailability are better.

Based on the discussion in this section, our hypotheses are the following: ChineseB shares are likely to lead A shares, because the prices of B shares are the outcome

2 See Grunbichler et al. (1994) (Germany), Harris et al. (1995) (US), Kleidon and Werner (1993) (UK),Pallmann (1992) (Germany) and Roell (1992) (France and UK). Also, see Bailey (1995), Garbade et al.(1979) and Hausbrouck (1991).


of more active investment decisions. Furthermore, foreign investors have alternativeinvestment opportunities and are more experienced in evaluating the expectedfuture prospects of firms. Large firms’ B shares with high liquidity could beexpected to lead A shares. Finally, these effects might be stronger in the initialperiod when the stock markets are more ‘emerging’ than later.

3. Cointegration, error correction and information diffusion

This section discusses the formulation of a stochastic representation of Aand B share prices. If domestic and foreign investors are identical, and haveaccess to the same information, the prices of A and B shares would be the same,adjusted for some minor transaction costs. Obviously, this is not the case judgingfrom Figs. 1–3 showing the share prices and the premium of A shares over Bshares3.

Assume that the prices of A and B shares form a stochastic vector with thefollowing vector error correction representation,

G(L)(1−L)xt=Pxt−1+m+ot, (1)

Fig. 1. Average weekly A- and B-share prices in Shanghai. This figure plots the weekly average A-shareprice and the weekly average B-share price in natural log form in Shanghai (22 firms) from July 1993 toJune 1997.

3 The Dickey–Fuller tests do not reject the hypothesis of a unit root for the price series, or for thepremiums. The results of these tests are available on request from the authors.


Fig. 2. Average weekly A- and B-share prices in Shenzhen. This figure plots the weekly average A-shareprice and the weekly average B-share price in natural log form in Shenzhen (19 firms) from July 1993to June 1997.

Fig. 3. Average A-share price premium. This figure plots the weekly average A-share premium inShanghai (22 firms) and the weekly average A-share premium in Shenzhen (19 firms) from July 1993 toJune 1997.


where, L is the lag operator, m is a vector of deterministic components (if theyexist), and ot is a residual white noise vector with a normal distribution. The G(L)and P matrices describe the flow of information in the system.

Transitory shocks going from the return of one market to the other are indicatedby significant off-diagonal elements in G(L). Permanent shocks common to theprices of A and B shares result in cointegration, and that P=ab %, where a is amatrix of adjustment coefficients, b % is the cointegrating vector.

The off-diagonal parameters in G(L), together with the adjustment parameters(a), reflect how information is spread — or not spread — between the markets.One significant off-diagonal element in G(L) indicates a one-way short-run causalityin the return series. The vector of adjustment parameters (a) indicates which marketis driving the price levels in the long run. If both markets have access to the sameinformation at the same time and process news in the same way, both coefficientsin a will be significant and of the same magnitude. If one market is driving theother, only one of the coefficients is significant. In the case of complete segmenta-tion in the long run, both coefficients in a are 0. If domestic and foreign investorshave access to different information, the price of the share, which is based onsuperior information, is expected to lead that of the other shares.

The main methodological difference between our study and Chui and Kwok(1998) is the error correction term ab %xt in Eq. 1. If A and B share prices form along-run steady state relation, reflecting the fact that the prices are formed fromexpectations on the same firm’s cash flow, inference based only on historical returnseries [G(L)(1−L)xt ] could be misleading because it leaves out important informa-tion in ab %xt.

The efficient market hypothesis (EMH) says that the prices of two assets cannotbe cointegrated because cointegration implies predictability in at least one direction.If share prices can be predicted from historical prices, either market efficiency isviolated or the model captures a stationary risk premium. In this study, with A andB shares of the same firm, can the two share prices be cointegrated under theEMH? Given that stock prices are based on the expected future cash flow of thesame firm, cointegration is expected. For the types of A and B shares traded on theChinese markets, finding a cointegrating price vector means that domestic andforeign investors have the same information in the long run. If only one type ofshare price is predicted by the cointegrating vector, one investor group can assumedto have superior information.

In this context, we argue that the most likely cause for no cointegration would bethe presence of a non-stationary political risk premium. Under the EMH, mosttypes of information differences should be temporary. As investors in one groupstart trading on their superior information, they will transmit their information notonly to their own market but to the other market as well. In an ESM, foreigninvestors might have to be extremely sensitive to changes in the economic andpolitical environment. If information regarding the political risk appears repeatedlyand in a stochastic way, the outcome could be a permanent non-stationarystochastic price premium, rather than stationary process. Accordingly, the prices ofA and B shares would follow separate stochastic trends in the end. This permanent


diffusion might exist only during the ESM period and disappear, if or when theinstitutions behind the markets become more credible. We find it unlikely that thepolitical risk premium is an integrated stochastic process. It is more likely that thepremium shifts at discrete intervals, if it shifts at all4.

4. Data and empirical results

4.1. Data

Chinese enterprises began to raise capital by issuing bonds and stocksin the 1980s. Since then, China’s financial markets have evolved quickly. Thereare two stock exchanges in China, the Shanghai stock exchange and the Shenzhenstock exchange. Both were inaugurated in the early 1990s. The Shenzhenexchange is a relatively smaller and less liquid market. The market for B sharesopened in 1992, which was more than 1 year after the A shares were first listed inthe Shanghai stock exchange. Table 1 presents basic statistics of the twoexchanges5.

The sample in this study includes weekly time series of 41 firms issuingboth A and B shares from July 1993 to June 1997 in either the Shanghaistock exchange or the Shenzhen stock exchange. Among them, 22 are fromthe Shanghai stock exchange and 19 from the Shenzhen stock exchange. Usingfirm-specific data, we construct average price series (in natural log form) fordifferent exchanges, by evenly weighting the share prices of the firms in eachexchange.

Table 1Descriptive statisticsa

Shenzhen stockShanghai stockexchangeexchange

Number of A shares listed 3003284445Number of B shares listed

A- and B-share market capitalization (billion RMB) 101.2 76.34.26Average daily trading volume of B to A shares (%) 2.99

a This table contains basic statistics of China’s stock markets. The sample period is July 1993–June1997.

4 A possible consequence of changes in the political risk premium is segmented trends in the priceseries. This could bias our tests towards finding I(1) processes (see Perron, 1989). Without any detailedinformation about when these possible shifts might occur, we have not investigated this idea further.

5 A detailed institutional background of the Chinese stock markets is given in Zhang (1999).


4.2. Results for the Shanghai and Shenzhen stock markets

Empirical tests of cointegration are sensitive to the properties of the estimatedmodel, and tabulated critical values are valid only for normally distributed whitenoise residuals. To construct a VAR representation of our samples, such that theassumption of normally and independently distributed white noise residuals cannotbe rejected, takes around 10–20 dummy variables. A closer investigation of ourmodels reveals that cointegration depends critically on two observations, 51 and 526.

Figs. 1 and 2 reveal a downward trend in the prices of B shares during the firstyear of the sample, up to observation 52. Imposing a dummy for what is the endof a downward trend, after which the prices quickly adjust back again, seems ad hoc.Removing these critical outliers will make A and B shares look more alike than theyreally are7.

The approach here is to avoid a huge number of dummy variables. We focus ona sufficient number of lags to ensure that the null of no autocorrelation in themodels cannot be rejected. This is achieved by using two and three lags in themodels.

For the aggregated Shanghai and Shenzhen price series, Johansen’s trace test andthe max test statistics reject cointegration at the 5% risk level, in Table 2, based onJohansen (1995) and Hendry and Doornik (1996). The result changes if we includesome dummy variables, but we cannot reject the hypothesis of no cointegration witha margin. Thus, we find the cointegration test inconclusive about whether the seriesare cointegrating or not.

The fact that we cannot easily establish cointegration is an important resultbecause it tells us that there are substantial differences between A and B share prices.A long-run stationary relation between A and B share prices cannot be taken forgranted. As discussed above, we are skeptical to the no-cointegration hypothesis. Inthis situation, with just two variables, an alternative test is to impose a reduced rankof the P matrix in Eq. (1), and test the significance of the adjustment parameters8.

After imposing one cointegrated vector in the system, we find one significant errorcorrection mechanism. The a1 parameter is significant for the Shanghai market,showing that foreign B-share prices drive the domestic A-share prices. In theShenzhen market, the relationship is reversed. In this market, the A-share pricesdrive the B-share prices.

Tables 3 and 4 show the estimated parameters of the VECMs. In both markets,B shares affect A shares, but there are important differences between the short andthe long run. In the short run, we find an uni-directional link from historical returns

6 These observations are c51, July 22, 1994 and c52, July 27, 1994.7 Some empirical studies deal with this problem by deleting observations according to some statistical

properties, for example, Xu and Wang (1997).8 With only one I(1) or I(0) variable, the t-statistics of the adjustment vector will have an asymptotic

N(0, 1) distribution under the null, see Banerjee et al. (1993).

B.

Sjoo

,J.

Zhang

/J.

ofM

ulti.F

in.M

anag.10

(2000)421

–438

429

Table 2Cointegration test statistics of the Shanghai and Shenzhen average price seriesa

Adjustment parameters (a) forEigenvalues (m) Cointegration test statisticsNormalized Eigenvectors(b. ) r=1

H0 m maximum m maximum Trace Trace (95%) B sharesSH–A A shares leadingSH–BB shares (a2)(95%) leading A

shares (a1)

Panel A: Shanghai stock exchange14.10 14.85 15.400.066 0.120**0.011 −0.0101.000 −0.348 r=0 12.80

3.80 2.05 3.80 (3.27) (0.46)2.051.000−0.073 r51

Panel B: Shenzhen stock exchanger=0 11.16 14.10 13.31 15.40 −0.032 −0.071**1.0000.059 −1.1860.012r51 2.15 3.80 2.15 3.80 (1.29) (3.27)1.017 1.000

a This table reports the Johansen cointegration test statistics. The variables are SH–A, Shanghai A-share average price; SH–B, Shanghai B-share averageprice; SZ–A, Shenzhen A-share average price; SZ–B, Shenzhen B-share average price. The t-values are in parentheses.

** Significant at the 0.01 level or better.


Table 3Vector error correction model results for the Shanghai stock exchangea

Dependent variable

DSH–B (ii)DSH–A (i)

−0.002DSH–At−1 0.015(−0.031) (0.339)

0.014 −0.038DSH–At−2

(−0.900)(0.191)0.0820.470**DSH–Bt−1

(3.507) (1.059)−0.405**DSH–Bt−2 −0.064

(0.810)(−2.942)0.216** −0.016Constant

(2.955) (−0.389)−0.110** 0.008b %xt−1

(−2.971) (0.389)Vector residual tests Vector AR 1-2 F(8, 344)=1.756 [0.085]

Vector normality x i2(4)=91.72 [0.000]

Vector Xi2 F(42, 475)=2.384 [0.000]

Vector Xi×Xj F(105, 417)=2.505 [0.000]

a This table reports the VECM results for Shanghai. The variables are, DSH–A, Shanghai A-shareaverage price, first difference; DSH–B, Shanghai B-share average price, first difference. The t-values arein parentheses. The P-values are in brackets.


on B shares to A shares. In the long run, B-share prices drive the A-share prices inShanghai. In Shenzhen, the long-run effect is just the opposite; here A-share pricesdrive the B-share prices.

These results support the assumption that foreign investors in the Shanghai stockexchange have better information, and that domestic investors adjust towards theprices of B shares. However, in the smaller and less liquid Shenzhen market, thedomestic investors have better information about the future long-run prospects ofthe firms. This could be a type of neglected firm effect, if foreign institutional investorsdo not find it worthwhile or too costly to examine the firms listed in this exchange.

4.3. Sensiti6ity tests

In the following, we test the sensitivity of these results with respect to variousassumptions regarding the information process. First, the sample is split into differentregimes9. Second, the series from Shanghai and Shenzhen are pooled into one model.Pooling the data will permit us to ask more detailed questions about the informationflow.

9 When viewing the results based on different sub-periods, it is important to remember that the testsof cointegration are based on the asymptotic properties of assumed infinite processes. Therefore, whenwe split the sample into sub-periods, we are not formally testing for structural breaks; we are onlydemonstrating the consequences, assuming that there are different regimes in the sample period.


Figs. 1 and 2 reveal that the prices of A shares fall during the first 52 weeks. Itcould be that foreign investors play a larger role in the early stages of an ESM.Suppose that this first part of the sample represents a different regime, withdifferent behavior of domestic investors. To analyze this possible regime change,the models are re-estimated with the first 60 observations truncated from thesample. The results are reported in Table 5. The cointegration test statistics are nowsignificant for both exchanges. More interesting, the significance of the a parame-ters is changing. As for the whole sample, foreign investors determine domesticA-share prices in Shanghai, and domestic investors determine long-run B-shareprices in Shenzhen. The change is that foreign investors also drive A shares inShenzhen. If we assume a regime shift, the role of foreign investors seem moreimportant over time, in the sense that their influence spreads to the Shenzhen stockexchange as the markets develop10.

Table 4Vector error correction model results for the Shenzhen stock exchangea

Dependent variable

DSZ–B (ii)DSZ–A (i)

DSZ–At−1 0.038−0.055(−0.680) (0.536)−0.029 0.057DSZ–At−2

(−0.366) (0.823)DSZ–At−3 −0.071 −0.025

(−0.365)(−0.921)0.178**DSZ–Bt−1 −0.141

(−1.876)(2.073)DSZ–Bt−2 −0.131−0.153

(−1.750) (−1.705)−0.145 −0.102DSZ–Bt−3

(−1.672) (−1.346)−0.025**−0.010Constant

(1.213) (−2.721)0.072**0.030b %xt−1

(3.320)(−0.923)Vector residual tests Vector AR 1-2 F(8, 344)=0.734 [0.662]

Vector normality x i2 (4)=90.63 [0.000]

Vector Xi2 F(48, 470)=1.463 [0.027]

Vector Xi×Xj F(132, 390)=2.185 [0.000]

a This table reports the VECM results for Shenzhen. The variables are, DSZ–A, Shenzhen A-shareaverage price, first difference; DSZ–B, Shenzhen B-share average price, first difference. The t-values arein parentheses. The P-values are in brackets.


10 Another observation from Figs. 1 and 2 is that the markets can be characterized as bear marketsuntil the end of 1995, and bull markets thereafter. We also estimated these periods separately, but theresults did not lead us to change our conclusions from above.


Table 5Vector error correction model results for both Shenzhen and Shenzhen stock exchanges — sub-samplea

Shanghai stock exchange Shenzhen stock exchange

3 3LagYes YesCointegration

ECMs−0.067** (2.28)0.181** (3.01)B shares leading A shares (a1)

A shares leading B shares (a2) −0.011 (0.31) −0.113** (3.67)AR (1-2)=0.902 [0.516]AR (1-2)=0.839 [0.570]Vector residual tests

Normality=100.7 [0.000] Normality=62.64 [0.000]Vector Xi

2=1.827 [0.036]Vector Xi2=0.779 [0.816]

Vector Xi×Xj=2.112 [0.000]Vector Xi×Xj=0.805 [0.876]

a This table summarizes the cointegration tests and the VECM results for both Shanghai andShenzhen in the period of October 1994–June 1997, with 130 observations. The t-values are inparentheses. The P-values are in brackets.


Our second sensitivity test is to pool the A and B shares of the two exchanges inone model. The cointegration test statistics from this ‘pooled’ model suggest two, orpossible one cointegrating vector, depending on the choice of risk level, see Table6. In the following, we explore the different hypotheses that follow by assuming oneor two cointegrating vectors.

Suppose there is only one cointegrating vector in the system. The system wouldconsist of three common stochastic trends and one stationary relation. The lattercould be a common risk premium for A shares over B shares in both exchanges.Alternatively, there is a stationary risk premium between the Shanghai and theShenzhen markets.

To test for these hypotheses, we start by testing for exclusion of exchanges ortypes of shares from the vector. All these hypotheses are rejected; means all fourvariables are needed to form the stationary relation. Next, we test if the premiumof A shares over B shares in Shanghai together with the premium in Shenzhen forma stationary relation. This joint hypothesis, the vector is made up of two premia, isrejected by the data. The x2(2) statistics is 10.702, with probability value of0.004711.

The alternative is that the two non-stationary A-share series cointegrate with thetwo non-stationary B-share series. This hypothesis assumes a joint risk premium ofA shares over B shares in the two markets. To test this hypothesis, we impose therestrictions of a ratio of A-share series and a ratio of B-share series on the

11 The exclusion tests are not presented here since they are all insignificant. In the tests, x= [SH–A, SH–B, SZ–A, SZ–B], the b vector is restricted as [b1= −b2] and [b3= −b4=1]. The test [b1= −b2] is not rejected with probability of 0.1928, [b3= −b4] is rejected with probability of 0.0011.

B.

Sjoo

,J.

Zhang

/J.

ofM

ulti.F

in.M

anag.10

(2000)421

–438

433

Table 6Cointegration test statistics of the Shanghai and Shenzhen pooled price seriesa

Cointegration test statisticsRestricted cointegration vectors Adjustment parameters (a) for r=2

H0 m maximum m maximum SH–B (ATrace Trace (95%) SH–A (BSH–A SH–B SZ–A (BSZ–A SZ–B (ASZ–Bshares leading(95%) shares leadingshares leading shares leading

B shares)A shares) A shares) B shares)

27.10 54.04**0.000 47.20 a11, −0.146**−1.2351.000 a21, 0.0370.000 r50 a31, −0.06126.33 a41, −0.154**(1.01) (2.93)(1.09)(2.38)

21.10 29.70* 29.70 a12, 0.108−1.446 a22, −0.0191.000 a32, 0.0470.0000.000 a42, 0.178**r51 15.61(1.35) (3.76)(0.60)(1.20)

14.10 14.09 15.40r52 8.83r53 5.26 3.80 5.26 3.80

Vector residual testsVector VectorVector AR 1-2 Vectornormality x i

2 Xi2=1.756F(32, 602) Xi×Xj=2.00

[0.000] [0.000]=1.001 (8)=127.8[0.000][0.468]

a This table reports the Johansen cointegration test statistics. The variables are: SH–A, Shanghai A-share average price; SH–B, Shanghai B-share average price; SZ–A, Shenzhen A-share average price;SZ–B, Shenzhen B-share average price. The optimal lag length in this model is 3. The t-values are in parentheses. The P-values are in brackets.

* Significant at the 0.05 level.** Significant at the 0.01 level or better.


cointegrating vector. The data does not reject the hypothesis that these two ratioscointegrate, x2(2)=1.570, with probability value of 0.4566. Our conclusion fromthese tests is that one cointegrating vector is not sufficient to correctly describe thesystem. There is a stationary risk premium between A and B shares, but thispremium is not necessarily of the same magnitude in both markets.

Assume that there are two cointegrating vectors, one vector represents thestationary premium in the Shanghai stock exchange and the other premium in theShenzhen stock exchange12. The next question is how the two markets interact witheach other in the long run.

The non-significant adjustment parameters of the pooled system suggest thatthere are two long-run exogenous prices in the system, the B-share prices inShanghai and the A-share prices in Shenzhen. The significant parameters confirmthe findings above that the foreign investors drive the Shanghai market, and thatthe domestic investors are more important in Shenzhen. The new result from thepooled system is that the price information in the Shanghai exchange spills over tothe B share market in Shenzhen, as suggested by the significant a41-parameter. Sincethe first vector occurs in two exchanges, but for different types of stocks, theforeign investors seem to use the same information to price B shares in Shenzhenas domestic investors used in Shanghai.

4.4. The flow of information between the markets

Why should the prices of B shares lead the prices of A shares in Shanghai? Chuiand Kwok (1998) suggest that foreign investors are better informed and receivenews faster than domestic investors because of the information barriers in China.An additional factor is that B-share investors are mostly big financial institutions,while domestic A-share investors are relatively smaller. Thus, the returns of theinstitutional favored shares could lead those of institutional unfavored shares, assuggested by Badrinath et al. (1995).

If information barriers are crucial, domestic investors have a problem in obtain-ing information, mainly because of the low creditability of domestic media. Thecost of obtaining information about the stock market in general and the prospectsfor individual firms is high for domestic investors. Therefore, a cost-effective way ofgetting information is to observe the price movements of the foreign B shares.Then, the question is why A-share prices follow B-share prices in Shanghai, but notin Shenzhen. The answer could be that the Shenzhen exchange is relatively smallerin terms of total market capitalization and number of listed firms, or because theShenzhen stock exchange is dominated by small firms.

In Table 1, we see that the total market capitalization of the Shanghai stockexchange is 101.2 billion RMB, and that of the Shenzhen stock exchange is 76.3billion. By June 1997, in Shanghai, the number of A-share listing firms is 328, while

12 This hypothesis is not rejected by the data. The test statistic is x2(2)=0.4735, with probability valueof 0.7892. Here, the two cointegrating vectors are restricted as [b11=1, b13=b14=0] and [b23=1,b21=b22=0].


Table 7Summary of the cointegration test results of individual firms

Stock exchange Cointegrated (%) Not cointegrated (%)

(N=22)Shanghai stock exchange 45.5 55.5Shenzhen stock exchange (N=19) 73.7 26.3

(N=41)Total 58.5 41.5

in Shenzhen, this number is 300. If we calculate the ratio of the average dailytrading volume of B shares to A shares in 1997, we find that this ratio is 4.26% forShanghai, and 2.99% for Shenzhen13. The Shanghai market is bigger and the Bshares are much more liquid than those in Shenzhen. The result that foreigninvestors are leading domestic investors in Shanghai could be in line with varioussmall firms and liquidity effects found in other markets. The next section analysesthe firm size effect in detail.

4.5. The firm size effect

The lead-lag effect and the information hypotheses suggest that firm size could bean important factor for foreign institutional investors. Therefore, we test if theprices of B shares have a tendency to lead those of A shares for firms with largermarket capitalization. Table 7 summarizes the cointegration test results, whichshow that more than half of the A and B shares are cointegrated. The share of firmswith cointegration among the assets is 58.5%.

Table 8Summary of the vector error correction model results — classified by firm size and stock exchangea

a2 Significant: A a1 and a2a1 and a2a1 Significant: Bshares leading B significant insignificantshares leading A

shares shares

Panel A: firm size62.5% – 12.5%Large 25.0%36.0% 36.0%Medium 20.0% 22.0%

Small 87.5% 12.5% – –Panel B: stock exchange

72.7% 9.1% 4.5%13.6%Shanghai stockexchange

52.7%26.3% 10.5%Shenzhen stock 10.5%exchange

a This table summarizes the estimated results from individual firms classified by firm size and stockexchange, respectively. Large firms are in the top 20%; small firms are in the bottom 20%; and in themiddle 60%, they are the medium ones. Significant stands for significant at the 0.05 level or better.

13 Source, Shanghai and Shenzhen Stock Market Data, 1997, respectively.


Table 8 summarizes the VECM results, classified by firm size and exchange,respectively14. Firm size is measured by adding the market capitalization of A andB shares (all in local currency, RMB) at the end of June 1997. The sample is thensplit into three groups, big firms are the top 20%; small firms are the bottom 20%;medium firms 60%. Panel A of Table 8 reveals that B shares lead A shares for bigfirms (62.5%) as we expected. However, for small firms, B shares lead A shares aswell (87.5%), which is inconsistent with our expectation. We check firm size indifferent exchanges in our sample and find that most of the firms in both the topand the bottom 20% are from the Shanghai stock exchange. We proceed to test ifthe choice of exchange determines the investment decisions of the foreign investors.Panel B of Table 8 shows that in the Shanghai stock exchange most of the B shareslead A shares (72.7%), while in the Shenzhen stock exchange, on the contrary, mostof the firms’ A shares lead B shares (52.7%). The results demonstrate that it is theShanghai stock exchange that determines that B shares are leading A shares. InShanghai, a larger number of A shares is driven by B shares compared with theShenzhen stock exchange. Thus, the choice of stock exchange is the most importantfactor behind the conclusion of B shares driving A shares.

5. Summary and conclusions

This study tests various aspects of the information diffusion resulting in differentprices on domestic investors’ A shares and foreign investors’ B shares in theemerging Chinese stock markets.

If both investor groups have the same information, information will flow in bothdirections between domestic and foreign investors. We expect A and B shares of thesame firm to be correlated, both in levels (prices) and in first differences (returns).If one investor group is leading the other, due to superior information, informationwill go in one direction only. If the markets are totally segmented, no informationwill be passed between the markets of A and B shares.

Our main conclusion is that the information diffusion between A and B sharesgoes from foreign investors to domestic investors in the larger and more liquidShanghai stock exchange. However, in the smaller Shenzhen stock exchange, thecausality is more ambiguous. In Shenzhen, foreign investors affect returns only inthe short run. In the long run, the information goes from domestic to foreigninvestors. These conclusions are quite stable under various assumptions of regimechanges, and after taking account of firm size. In the end, the most importantfactor for determining the information flows is the choice of stock exchange ratherthan firm size.

We argue that foreign investors drive the prices of A shares in China’s stockmarkets because domestic investors have problems in acquiring relevant andtrustworthy firm information from domestic and foreign media. Domestic investors

14 The cointegration tests and the VECM results for the individual firms are available on request fromthe authors.


therefore condition their investment decisions on observed B-share prices to find agood long-run valuation of the stock. In a smaller exchange, this foreign informa-tion advantage might not exist, other than in the short run. In this type of market,foreign investors rely on the locals to determine the correct long-run futureprospects of the firms. Future research will reveal if the information diffusionbetween A and B shares in Shanghai and Shenzhen is stable, or changes as themarkets evolve over time.

Acknowledgements

The authors thank Richard J. Sweeney and Clas Wihlborg for discussions andsuggestions, Roger Huang and Lars Meuller for valuable comments. The remainingerrors are the fault of the authors. Part of this study was written when the firstauthor was visiting the Department of Management, McGill University.

References

Badrinath, S.G., Kale, J.R., Noe, T.H., 1995. Of shepherds, sheep, and cross-autocorrelations in equityreturns. Rev. Financial Studies 8, 401–430.

Bailey, W., 1995. One security, many markets: determining the contributions to price discovery. J.Finance 50, 1175–1199.

Bailey, W., Jagtiani, J., 1994. Foreign ownership restrictions and stock prices in the Thai capital market.J. Financial Econ. 36, 57–87.

Banerjee, A., Dolado, J., Galbraith, J.W., Hendry, D.F., 1993. Co-Integration, Error-Correction, andthe Econometric Analysis of Non-Stationary Data. Oxford University Press, London.

Chan, K., 1993. Imperfect information and cross-autocorrelation among stock prices. J. Finance 48,1211–1230.

Chui, A.C.W., Kwok, C.C.Y., 1998. Cross-autocorrelation between A shares and B shares in theChinese stock market. J. Financial Res. 21, 333–335.

Garbade, K.D., Pomrenze, J.L., Silber, W.L., 1979. Dominant and satellite markets: a study of duallytraded securities. Rev. Econ. Stat. 61, 455–460.

Grunbichler, A., Longstaff, F.A., Schwartz, E.S., 1994. Electronic screen trading and the transmissionof information: an empirical examination. J. Financial Intermediation 3, 166–167.

Harris, F.H., Mclnish, T.H., Shoesmith, G.L., Wood, R.A., 1995. Cointegration, error correction, andprice discovery on informationally linked security market. J. Financial Quant. Anal. 4, 563–578.

Hausbrouck, J., 1991. Measuring the information content of stock trades. J. Finance 46, 179–208.Hendry, D.F., Doornik, J.A., 1996. Empirical Econometric Modeling Using PcGive 9.0 for Windows.

International Thomson Business, London.Johansen, S., 1995. Likelihood-Based Inference in Cointegrated Vector Auto-Regressive Models. Oxford

University Press, London.Kleidon, A.W., Werner, I.M., 1993. Round the clock trading: evidence from UK cross-listed securities.

NBER Working Paper, 4410.Lo, A., MacKinlay, C., 1990. When are contrarian profits due to stock market overreaction? Rev.

Financial Studies 3, 175–206.Pallmann, N., 1992. Market structure and the speed of incorporating new information into security

prices: evidence from Germany. Working Paper, New York University.Perron, P., 1989. The great crash, the oil price shock, and the unit root hypothesis. Econometrica 57,

1361–1401.


Roell, A., 1992. Comparing the performance of stock exchange trading systems. In: Fingleton, J.,Shoenmaker, D. (Eds.), The Internationalization of Capital Markets and the Regulatory Response.Graham and Trotman, London.

Schreiber, P.S., Schwartz, R.A., 1986. Price discovery in securities markets. Journal of PortfolioManagement 12, 43–48.

Xu, X., Wang, Y., 1997. Ownership structure, corporate governance, and corporate performance: thecase of Chinese stock companies. The World Bank, Policy Research Working Paper, No. 1794.

Zhang, J., 1999. Essays on Emerging Market Finance. Economic Study No. 94, Department ofEconomics, Gothenburg University.

.

Documents

Market segmentation and information diffusion in China’s stock markets