The emerging stock markets of Brazil, Russia, India and China (BRIC) grew rapidly throughout most of the 2000s. This rapid growth of the emerging BRIC stock markets raises the question as to whether or not these markets are becoming increasingly integrated with the leading developed stock market of the US. This is important because that international portfolio investment strategies depend on the degree of integration of the stock markets (Barari, 2004), and transmission of stock price among equity markets can affect ability to hedge risk via international diversification (Asgharian and Nossman, 2011). Since more and more evidence show that the participation of foreign institutional investors in emerging equity markets increased dramatically (Ilyina, 2007); thus, for the sake of potential benefit of international risk diversification, the integration and co-movement of stock markets between the US and the BRIC needs more investigation.
Within the research on stock market integration there is a vast body of literature devoted to potential risk diversification benefits from an international investment portfolio. Due to several factors, such as the rapid expansion of international trade in commodities, services and financial assets (Kearney and Lucey, 2004) and the liberalization of financial market (Awokuse et al., 2009), emerging stock markets are facing financial meltdown and binding each other (Aktan et al., 2009). Though empirical evidence from previous studies using conventional linear cointegration models has shown stock market integration in some regions, the existing empirical evidence remains inconclusive and there are conflicting results regarding the nature of dynamic interdependence between developed and/or emerging markets (Awokuse
et al., 2009). There are two weaknesses which are often overlooked. One is the case where the nonlinear type of cointegration may be ignored; and the other is where the important element of time variation of integration is missing (Kearney and Lucey, 2004), so that the instability problem of long-run relationships is insufficiently considered, leading to ambiguous results and conflicts (Awokuse et al., 2009; Aktan et al., 2009). In this situation, research in the field of nonlinear cointegration and time-varying cointegration requires further development.
The use of cointegration measures to assess the degree of international integration in equity markets has been verified by previous studies (see inter alia Kearney and Lucey, 2004). In this vein, we used the Engle–Granger (E–G; Engle and Granger, 1987) cointegration test, and the Enders–Siklos (E–S; Enders and Siklos, 2001) threshold cointegration test along with the threshold error correction model (TECM) to investigate the interdependences between the developed US stock market and those of developing BRIC countries. This combination of developing Brazil, Russia, India and China has very large population, which expects astonishing growth in consumer markets in the near future, making them the largest emerging markets in the world.
The E–G cointegration test is a special case of the E–S threshold cointegration test, which implies symmetric adjustment behaviour (Enders and Granger, 1998; Enders and Siklos, 2001). The alternative E–S cointegration model allows for the case that adjustment speeds differ in two regimes based on an estimated threshold, called asymmetric threshold cointegration. We use the both linear and nonlinear models in this study for comparative analyses.
Further, in order to have a dynamic analysis in comparison with the comparative analysis, we expand the E–S asymmetric threshold cointegration test by using two
alternative time-varying approaches, recursive estimation and rolling estimation, to gain insight into the dynamic evolving process of nonlinear cointegration between the stock markets studied. If the evolving pattern of cointegration shows no changes and/or stability, then there are no differences for making conclusions when comparing the comparative analyses. If it presents volatile and/or unstable situations, this indicates the time-varying nature of the cointegration relationship, as argued by Awokuse et al. (2009) and Lucey and Aggarwal (2010). The advantage of the dynamic analyses here helps to prevent confusing and partial results.
This paper extends the existing literature in the following aspects. First, the study explores the long-run cointegration relationship using a nonlinear framework with asymmetric adjustment behaviour, which has generally been overlooked in earlier studies. Second, to our knowledge, this paper is the first to expand the consistent momentum threshold autoregressive (consistent M-TAR) model (Enders and Siklos, 2001) by time-varying approaches to the stock market integration between the US and the BRIC. Third, the study is the first to explore the short-run instantaneous price transmission between the developed market of US and the developing markets of BRIC by the Granger-causality test in a dynamic manner.
The results demonstrate that time-varying long-run nonlinear cointegration relationships exist between the Dow Jones and each of the BRIC markets. In the short-run, the Dow Jones continues playing a leading role, Granger-causing each of the emerging BRIC indices with increasing trends. The findings confirm the time-varying nature of cointegration relationships, contending the propositions by Awokuse et al. (2009) and Lucey and Aggarwal (2010), though in a nonlinear manner, and we also found time-varying Granger-causality relationships. Moreover, it is noteworthy that the Brazil and China stock markets began exerting significant
influence on the Dow Jones index after 2006.
This article is organized as follows: Section II briefly considers previous studies on stock market integration. Section III presents the data and the methodology of this study. Empirical results are discussed in Section IV, and Section V concludes this article.