Are the ASEAN Equity Markets Interdependent?

Abstract

This article investigates the extent and structure of price linkages among five Association of Southeast Nations (ASEAN) markets (Malaysia, Singapore, Philippines, Indonesia and Thailand), both in the long run and in the short run using cointegration based on the Johansen (1988) procedure, Granger causality and variance decomposition and impulse response analyses. The study finds that in the long term these markets are not significantly linked. However, in the short term, with the exception of Indonesia, all the ASEAN markets have significant linkages with each other. Malaysia is the most influential market while Singapore and Thailand are the markets with most linkages with other markets. Indonesia is not linked at all with any other ASEAN market. Introduction The growth rates of the ASEAN economies are among the highest in the world (DFAT 1992). Starting in 1988, as a result of financial reforms arising out of deregulation (Cargill et. al. 1986; Greenwood 1986; Pringle 1987; Drake 1986; Cole 1988), the financial markets of these countries have also grown significantly (Fry 1995; Allen 1991). Thus, this group of markets has been attracting the attention of investors world-wide. Crucial information that investors need for purposes of portfolio diversification is the degree of linkage among these markets. Unfortunately, this information is unknown at present. There has been an increasing intensity of trade between these countries (Ariff 1996) suggesting that their financial markets are also becoming increasingly linked. The relatively few studies on financial integration in Asia have primarily focused on Japan and the so-called Asian tigers or dynamic economies of Hong Kong, Singapore, Taiwan and Korea (see, for instance, Hung and Cheung 1995; Kwan et al. 1995; Kwok 1995; Cheung and Mak 1992 among others). The current authors are unaware of any systematic study which has been undertaken on equity market linkages among ASEAN countries. This study therefore seeks to redress this deficiency. Specifically, the study seeks to find answers to the following questions: 1. To what extent are equity prices in these markets significantly linked, both in the short term and in the long term? 2. Which markets lead and which ones lag? 3. How fast and for how long do interactions occur between these markets? The issue of financial market integration is of crucial importance in the theories in finance (for example portfolio investment theory) and economics (for example Mundell-Fleming theories). Empirically, however, this issue remains unsettled. Previous studies have produced mixed results depending on the theory, methodology, time period, and data used' (Jeon and von Furstenberg 1990; Czerkawsky 1995; Kwan et al. 1995; Ben-zion et al. 1996; Chung and Liu 1994). Many of these studies are plagued by several problems (Jeon and von Furstenberg 1990). Some problems pertain to the use of differencing in order to make variables stationary, the inability to simultaneously analyse short-term and long-term relationships among variables, and the use of non-comparable stock market data (see, for instance, Kwan et al. 1996). The current study overcomes these problems through the use of recent advances in econometric time series analysis-cointegration analysis (Engle and Granger 1987) using the Johansen (1988) procedure, combined with Granger causality, and impulse response and forecast variance analyses. Differencing achieves stationarity of data but important long-term relationships contained among levels of the variables are lost. Cointegration allows the determination of long-term relationships among non-stationary variables. Its associated error-correction model, together with Granger causality and variance decomposition and impulse response analyses, allows a more rigorous scrutiny of the short-run linkages between the markets in terms of identifying the markets that lead or lag, and the speed and duration of interactions between markets. …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 58 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.