Introduction Directory UMM :Data Elmu:jurnal:M:Multinational Financial Management:

1. Introduction

The economic and financial turmoil that struck Asia in mid 1997 was representa- tive of both crisis and panic. Asia’s financial systems had continued to rely upon what Professor Merton Miller has called ‘19th century technology’. Capital markets were poorly developed and, on the upside of the business cycle, banks extended credit excessively, misallocated capital and employed weak credit controls. When the Thailand Baht was devalued on July 2, 1997, panic spread among both foreign and domestic creditors and investors. In the following months, this was transmitted to other Asian Countries. Eventually, a group of countries Thailand, Korea, Indonesia, Malaysia, Japan and the Philippines suffered severe collapse of curren- cies and stock markets. Other countries China, India, Hong Kong, Taiwan and Singapore have, so far, suffered only the indirect consequences of the crisis. Earlier evidence was provided by Malliaris and Urrutia 1992 that informational linkage among national capital markets is a factor responsible for financial crisis. They detected the causal relationships among national stock markets during the October 1987 stock market crash, finding a dramatic increase in bi-directional and uni-directional causality during the month of the crash. In contrast, no lead-lag relationships were detected for the periods before and after the crash. Eun and Shim 1989 empirical results also indicate that there exists a substantial degree of interdependence among national stock indices. Unexpected movements in interna- tional stock markets seem to have become influential news event that affect domestic stock markets. King and Wadhwani 1990 investigated why, in October 1987, almost all the stock markets fell together despite widely differing economic circumstances. They examined a rational expectations price equilibrium and mod- eled ‘contagion effects’ between markets. They found that an increase in volatility in one market, leads in turn to an increase in the size of the contagion ellects on other markets. They claim that their theory can provide a partial of explanation why the world stock markets uniformly fell during the October 1987 crash. Hamao et al. 1990 examined the prices and price volatility spillover effect among New York, London and Tokyo markets using an ARCH-type model. Evidence of price volatility spillovers from New York to Tokyo, London to Tokyo and New York to London was observed, bot no price volatility spillover effects in other directions were apparent for the pre-October 1987 period. Recently, financial market liberalization and rapid development of telecommuni- cations networks has increased significantly the ability to transmit and disseminate information between markets. Bekaert and Harvey 1997 found that capital market liberalization did increase the correlation between local market returns and the world market, but did not drive up local market volatility. Also, Longin and Solnik 1995 found an upward trend in international correlations over the past 30 years, particularly in the rise of the correlations in periods of high volatility. The interdependence of regional stock markets has also become the subject of extensive research. Divecha et al. 1992 investigated ten emerging Asian stock markets and found that they were homogeneous with a dominating strong market force and less correlated with each other and with the developed markets. Cheung and Ho 1991 and Cheung 1993 examined the correlation structure among eleven emerging Asian stock markets and developed markets. They concluded that the correlation between the emerging Asian stock markets group and the developed market group was smaller than among the developed markets. However, based on various analytical techniques, the correlation matrix is not stable over time. Using cointegration analysis, Corhay et al. 1993 also investigated the price indices of five European stock markets and found that they displayed a common long-run trending behavior over the period 1975 – 1991. Choudhry 1997 investigated the long-run relationship between the stock indices of six Latin American countries and the United States, and found evidence of cointegration and significant causality among the six Latin American indices with and without the United States index. Christofi and Pericli 1999 explored the short-run dynamics between five major Latin American stock markets. They modeled the joint distribution of stock returns using a vector autoregression model with errors following a multivariate exponen- tial GARCH process and found evidence of first and second moment interactions among these markets. The purpose of this paper is to analyze the linkages among national stock markets before and during the period of the Asian financial crisis. It is well known that the linkages among stock markets vary over time, and we examine whether there exist different degrees of linkages before and during the period of the Asian financial crisis. In contrast to the procedure used by Engle and Granger 1987 in Malliaris and Urrutia 1992, we employed Johansen 1988 multivariate cointegra- tion and errorcorrection model to investigate the nature and extent to which national stock markets contribute to the crisis process 1 . Moreover, we decompose the forecast error variance to show the proportion of the movements in one market due to its own shocks versus shocks from other markets. This can help us to understand how markets are ‘fully exogenous’ to the financial crisis, and whether the ‘degree of exogeneity’ differs before and during the period of the financial crisis. The multivariate cointegration and error-correction tests provide some evidence to support the existence of cointegrational relationships among the national stock indices during, but not before, the period of the financial crisis. The relationship in the South-East Asian countries seems to be stronger than that in the North-East Asian countries. The forecast error variance decomposition demonstrates that the ‘degree of exogeneity’ for all indices has been reduced, implying that no countries are ‘exogenous’ to the financial crisis. In addition, Granger’s causality test suggests that the US market still ‘causes’ all the Asian countries during the period of crisis, reflecting the US market’s persisting dominant role. The paper is organized as follows: Section 2 describes the data. Section 3 presents the test of unit root, cointegration and error correction, causality and their results. 1 Although Engle and Granger 1987provide the cornerstone research linking cointegrated series that move together through time to the concept of error correction. However, their two-step procedure has since been shown to be most apprpriate for systems of only two variables with one possible cointegrating vector. The multivariate test by both Johansen 1988 and Stock and Watson 1988 for cointegration is preferred for several variables, since it identifies the space spanned by the cointegrating vectors. Section 4 introduces the forecast error variance decomposition model and its result. Section 5 summarizes and concludes the paper.

2. Data description