Portal de Periódicos da CAPES

An empirical study of the asymmetric cointegration relationships among the Chinese stock markets

2007; Taylor & Francis; Volume: 39; Issue: 11 Linguagem: Inglês

10.1080/00036840600606302

1466-4283

Chung‐Hua Shen, Chien-Fu Chen, Li-Hsueh Chen,

Financial Markets and Investment Strategies

Abstract The Enders and Siklos asymmetric cointegration test is employed to examine the long-run asymmetric equilibrium relationships between the Chinese Shanghai and Shenzhen stock markets. Three samples are adopted, which are the whole sample (October 1992 to September 2002); the first subsample before B shares were opened up to the Chinese public (October 1992 to February 2001); and the second subsample after B shares were opened up (February 2001 to September 2002). The estimated results are as follows. First, when the conventional Engle–Granger symmetric cointegration test is used, only the A shares in Shanghai and Shenghen stock exchange market are cointegrated when using the whole sample and the first subsample. However, with the Enders–Siklos M-TAR asymmetric cointegration test, Shenzhen A and B shares stock prices have an asymmetric cointegration relationship after B shares were open, suggesting that openness increases the market efficiency. Furthermore, the two A shares in Shanghai and Shenzhen stock exchanges also have an asymmetric cointegration relationship in the whole sample and the first subsample, implying that although the asymmetric relationship is crucial, it has long been neglected. Finally, it is found that the adjustment speed of Shanghai A shares is faster when deviation from the long-run equilibrium is positive than when it is negative. Notes 1 Chinese citizens have also been allowed to purchase B shares since 19 February 2001. Because of obvious price differences between A and B shares, this policy change may have induced arbitrage opportunities between the two markets. 2 H-share denotes the firms listed in Hong Kong. 3 For detailed description of the Chinese stock markets, refer to the World Bank (Citation1995). 4 Results from other exchange markets are the same; that is, the adjustment speed is faster for bad news than good news. For example, Li and Lam (Citation1995) used a threshold ARCH model to study possible asymmetric behaviour of the Hong Kong Hang Seng index during bull and bear markets. Their findings indicated that the adjustment speed of the Hang Seng index is asymmetric when returns are positive but not when they are negative. 5 Mt in Equations Equation5 and Equation6 can be replaced by It in the TAR model. 6 η12 is insignificantly different from zero, with t value −0.562.