بازگشت و خطر تعامل در بازار بورس چینی

This paper investigates interactions between Chinese A shares and B shares traded on the Shanghai stock exchange (SHSE) and the Shenzhen stock exchange (SZSE), using an asymmetric multivariate time-varying volatility model. We find that there is a causal relation from B share markets to A share markets in the second moment but no such relation is present in the first moment, suggesting B shares contain more prior information than A shares about risk but not return, due to differences in investment objectives and investment scopes between the two groups of investors and barriers between the two markets. Moreover, there exist stronger links between shares of the same type, i.e., between the two A (B) share markets, than those between shares of different types. All four markets exhibit leverage effects.

Over the last decade, Chinese stock exchanges (CSEs) have experienced rapid growth and development. There were only 10 listed companies when CSEs were established in 1990 with share trading being restricted to domestic investors only. By December 2000, the number of listed companies increased to 1088, the number of registered investors reached 58 million, and annual turnover was in excess of six trillion RMB yuan. Trading is fully computerised and can be conducted on screens from the offices of securities dealers. On 21 May 1992, the price and volume control imposed by regulations were abolished, so the exchanges are now a totally market driven system. The CSEs consist of two markets: the Shanghai stock exchange (SHSE) and the Shenzhen stock exchange (SZSE). SHSE was established on 26 November 1990 and started trading on 19 December of the same year, and SZSE was established on 1 December 1990 and started trading on 3 July 1991. SHSE is located in the coastal city of Shanghai, the largest city and one of the financial centres in the country, with most companies listed on it being large and state-owned firms. With its privileged position the local government has an ambition to reinstate it as an influential financial centre in Asia. SZSE is situated in the southern city of Shenzhen which is the first and most important special economic zone in China neighbouring Hong Kong, and the companies listed there are typically small and export-oriented firms, many of them joint ventures. Because of these characteristics, SHSE may be sheltered from the fluctuation in the world economy, and on the other hand, SZSE could be more vulnerable to a global slowdown. At present, two types of shares are traded on both SHSE and SZSE. One is A shares which are ordinary shares issued by mainland Chinese companies, denominated and traded in RMB yuan and designed for domestic investors. The other is B shares which are issued by mainland Chinese companies, denominated in RMB yuan but traded in foreign currencies, i.e., B shares carry a face value denominated in RMB yuan but are bought and sold in foreign currencies.3 Therefore, there are four major markets representing stock market investment in China-A share markets in Shanghai and Shenzhen, and B share markets in Shanghai and Shenzhen, respectively. The Chinese stock market is of special interest, as it has four major markets running in parallel, featured by geographically separated trading locations and different shares for different investors. The distinction between A shares and B shares constitutes different channels and ways with which information is received and processed by different groups of investors. This distinction also represents different ownership structures in which investors show interest in the control and running of the company, pursue rent-seeking activities, engage in diversification, or merely trade for liquidity. Finally, the distinction contributes to differed risks and risk attitudes due to restrictions imposed, until recently, on investors, and due to the adoption of different currencies with which shares are traded throughout the whole period. All of these, together with geographical features of Shanghai and Shenzhen, indicate that Chinese stock markets bear many hallmarks of asymmetry, and equilibrium and equilibrium prices are attained through interactions and information transmission among these markets in a rather complicated way. The subject of Chinese stock market with the above-demonstrated remarkable features indeed opens up an opportunity to investigate different price patterns and investors’ behaviour under an investment environment coming into existence for only a decade on the one hand. On the other hand, the subject poses a challenge with a serious task to examine complicated interactions between these markets and investment channels in a systematic way capable of capturing and explaining these features in Chinese stock markets. Although extensive efforts have been made in recent years, the research is still insufficient compared with that on Western mature markets. While multivariate models are applied to the inquiries of Chinese stock market interactions with regard to return, previous studies on the subject are mainly based on bivariate models with regard to volatility. Consequently, the results offer partial or piece meal pictures of, and explanations to, complicated mechanisms of interactions in Chinese stock markets which are beyond the reach these bivariate models are able to handle satisfactorily. The importance of the subject and the imperative need of systematic inquiries into the subject have therefore motivated the present research to investigate the issues in a truly multivariate circumstance.4 We estimate four return series simultaneously in this study, and in the mean time, we also allow the entire variance-covariance structure of the model to respond in an asymmetric fashion. Restrictions in trading in A shares and B shares give rise to market segmentation, an analytical framework with which a considerable prior empirical studies in the area are conducted. In this sense, the notion of mild segmentation proposed by Errunza and Losq (1985) provides a constructive theoretical context to follow. They assume that there is unequal access to capital markets due to restrictions imposed on a subset of investors. A subset of investors can trade in all the available securities, while the others can only trade in a subset of the securities, termed eligible securities. Under such a “mildly” segmented market structure, it is understandable that there exist “super” risk premia for a subset of securities, the ineligible securities which can only be held by the unrestricted investors. While Errunza and Losq (1985) offer explanations to unequal risk premia in a segmented market, the situation in Chinese A share and B share markets is different, at least until March 2001 when the B share market was opened to domestic investors.5 The two investor groups are mutually exclusive, and each group may yield “super” risk premia due to the restrictions imposed on its counterpart, leading to even richer risk-return patterns in Chinese stock markets. To establish a theoretical ground for the inquiry of these patterns, the market structure and institutional characteristics of Chinese stock markets will be deliberated in detail in the next section, offering explanations to expected investor behaviour and investment objectives in A share markets and B share markets. Empirical studies of Chinese stock markets in the extant literature pay attention predominantly to the linkage and comparison between A share and B share markets, which reflects the primary institutional features of Chinese stock markets. Applying the latent variable asset pricing model to examine the pricing of A shares and B shares in China stock markets, Fung et al. (2000) test whether the markets for the A share and B share of the same company are segmented. They claim that A share and B share markets are loosely related, based on their finding that the latent risk premia for A shares and B shares are only weakly correlated. Chakravarty et al. (1998) examine both information asymmetry and market segmentation in Chinese stock markets, and test whether A share returns lead B share returns, using a sample of 23 firms issuing both A and B shares on SHSE and 16 firms issuing both A and B shares on SZSE, together with the four market indices for Shanghai A and B shares and Shenzhen A and B shares. They document some evidence of two-way information flows between A share and B share markets, which are asymmetric with A share returns being more likely to lead B share returns on average. They also derive a relative pricing equation for A shares and B shares to explain the phenomenon that B shares trade at a discount, which is on average 60%, relative to A shares,6 and attribute the large price discount of B shares to the claim that foreign investors have less information on Chinese stocks than domestic investors. Bergstrom and Tang (2001), Chen et al. (2001), Fernald and Rogers (2002), Sun and Tong (2000) and Wo (1997) also make inquiries into the issue of price discounts of B shares. Whereas, Gordon and Li (1999) argue that the observed B share discounts in Chinese stock markets are consistent with a government choosing regulations and restrictions, through which indirect but more effective taxation is levied, to maximise a standard type of social welfare function. Investigating interactions between Chinese stock markets in terms of causalities, Kim and Shin (2000) find causality relationships among Shanghai A and B, Shenzhen A and B, and Hong Kong H shares and indicate that B Shares have become more influential in recent years. Although A shares tended to lead B shares before 1996, such relationships either disappeared or were reversed after 1996. Volatility dynamics in Chinese stock markets is inspected by Friedmann and Sanddorf-Kohle (2002), comparing different GARCH models. Their empirical results are claimed to reflect the different dynamics due to market segmentation in A share and B share markets. They find that a significant impact of the number of non-trading days on volatility, as well as a significant reduction of volatility by introducing the price change limit, is evident for the daily returns on A shares, but the result is mixed for B shares. Su and Fleisher (1999) estimate a dynamic model assuming the contemporaneous dependence of stock returns and trading volume to investigate the differences in volatility in A share and B share markets. They find that news enters A share markets more intensively than B share markets; news is more highly correlated with trading for A shares than for B shares; and news is more persistent for A shares than for B shares. Using GARCH models, Yeh and Lee (2000) investigate asymmetric reaction of return volatility. They find that the impact of bad news (negative unexpected return) on future volatility is smaller than the impact of good news (positive unexpected return) in the Shanghai and Shenzhen markets, suggesting incredibly “good-news-chasing” behaviour of the investors. Largely in line with most empirical evidence observed in Western mature markets, Su and Fleisher (1998) find that stock-market volatility in Chinese stock markets is time-varying, mildly persistent, and is best described by a fat-tailed distribution. Mookerjee and Yu (1999) focus on seasonal patterns in returns on the Shanghai and Shenzhen stock markets. They claim that, unlike studies for other stock markets, the highest daily returns on both exchanges occur on thursdays rather than on fridays, and daily stock returns appear to be positively correlated with risk. Wong et al. (1999) also test for stock return seasonality in terms of day of the week effects. The above analysis of the literature shows that extensive efforts have been made in recent years to examine return and risk patterns in Chinese stock markets, with the reviewed studies falling into two major categories. The first category is the examination of the linkage and comparison between A share and B share markets. While many studies have made attempts to uncover the mechanisms of interactions between A share and B share markets, they are predominantly confined to the relations and interactions between the prices or returns of different markets, or they are in the first moment only. In the second category where risk and volatility patters, or data characteristics in the second moment, are investigated, the research is almost univariate exclusively and at best bivariate. The summarised current state of research in Chinese stock markets indicates that information flows with regard to risk or volatility have yet to be examined structurally together with information flows with regard to returns, to which the present study dedicates. The rest of the paper is organised as follows. Section 2 presents the institutional background and structure of Chinese stock markets and market developments in the last decade. Section 3 discusses the data and presents preliminary analysis statistics. Section 4 specifies the models used in this study. Section 5 reports the empirical results and provides discussions of the results and implications. Finally, Section 6 concludes.

n this paper, we have investigated interactions in the mean and variance–covariance of returns on four Chinese stock market indices using an asymmetric BEKK model. The topic is of interest in that it opens up an opportunity to investigate different price patterns and investors’ behaviour in a market that is rather different from Western mature markets. As long as B share markets are concerned, they behave rather similarly to developed stock markets that there is strong linkage between the Shanghai B share market and the Shenzhen B share market. This linkage exists in both mean and variance–covariance through a two-way transmission mechanism, i.e., the causal relation is bi-directional. For A share markets, the linkage exists only in the second moment with bi-direction transmission. The most interesting part is nevertheless the relationship between A share markets and B share markets where major contributions have been made in this study. Our findings suggest that while B share markets lead A share markets in the sense of Granger causality, the lead is found in the variance–covariance structure only. Inferred from this evidence is that B shares contain more prior information than A shares about risk but not return, arising mainly from differences in investment objectives and investment scopes between the two groups of investors, but exchange control and other factors may play a role as well. Therefore, while the availability of B shares to domestic investors since March 2001 may alter overall B share investors’ behaviour and consequently reduce the degree of asymmetry, asymmetry will exist as long as there are barriers between A share and B share markets. Finally, we confirm that there exists sign effect or leverage effect in Chinese stock markets and extend the findings to a multivariate setting. That is, volatility increases more following a negative shock than a positive shock in the market itself, and to a certain degree, in the cross-section. The size and sign of return innovations are important features in determining the degree of spillovers in volatility.