ادغام بازار سهام در « آ سه آن» پس از بحران مالی آسیا

This paper considers the degree to which the five stock markets in the original Association of Southeast Asian Nations countries (ASEAN-5) are correlated as a way to assess the feasibility of policy initiatives to enhance ASEAN stock market integration and the implications for portfolio investors. In particular, this paper considers whether the ASEAN-5 markets are integrated or segmented using the time series technique of cointegration to extract long-run relations. The empirical results suggest that the ASEAN-5 stock markets are cointegrated and are thus not completely segmented by national borders. However, there is only one cointegrating vector, leaving four common trends among the five variables. We therefore conclude that ASEAN-5 stock markets are integrated in the economic sense, but that integration is far from complete. On a policy level, initiatives to further integrate the stock markets are feasible, and in fact desirable. From the perspective of the international portfolio investor, benefits of international portfolio diversification across the five markets are reduced but not eliminated.

This paper examines stock market integration in Indonesia, Malaysia, the Philippines, Singapore, and Thailand in the aftermath of the Asian financial crisis. These five countries are the original members of the Association of Southeast Asian Nations (ASEAN), which now also includes Brunei Darussalam, Cambodia, Laos, Myanmar, and Vietnam. Over the past few years, ASEAN member countries have made tremendous progress in forming a free trade area and investment zone—witness the ASEAN Free Trade Area (AFTA) and the ASEAN Investment Area (AIA). The region is now examining the possibility of capital market integration for national bond markets and stock markets alike. ASEAN countries, in fact, are beginning to discuss the feasibility of a currency union as well, which suggests that the countries are interested in multilateral approaches to many regional economic and financial issues. The five original ASEAN countries (the ASEAN-5) are the most likely candidates to undertake integrative measures first, and therefore provide the focus for this paper. Integration in the ASEAN capital markets may include initiatives to coordinate the five national capital markets that already exit, or at an extreme may involve the creation of supranational regional bond and stock exchanges.1 The issue is integration as opposed to capital market development more generally, although one motivation for integration is typically to foster development of the market. ASEAN policymakers are weighing the benefits of financial market integration in the same way as European Union policymakers prior to monetary unification. With respect to unification in Europe, Kleimeier and Sander (2000) point out that there are really two kinds of benefits: first, benefits of regional integration of the type studied in the Cecchini Report (Cecchini, 1988); and second, benefits of integrating the region with global financial markets. Although Kleimeier and Sander (2000) examine banking, the point is generalizable to stock and bond markets as well. ASEAN policy makers are naturally wondering what gains might be available to their region. Interest in stock market integration arises primarily because financial theory suggests that an integrated regional stock market is more efficient than segmented national capital markets. Capital market efficiency in Southeast Asia has become even more important after the Asian financial crisis of 1997–1998, so we focus on the post-crisis period to-date, using high-frequency data. Southeast Asian countries are specifically seeking to reduce the traditional dependence of firms on bank loans rather than bond and stock issuances, and at the same time are seeking new capital from outside the region. With an integrated regional stock market, investors from all member countries will be able to allocate capital to the locations in the region where it is the most productive. With more cross-border flows of funds, additional trading in individual securities will improve the liquidity of the stock markets, which will in turn lower the cost of capital for firms seeking capital and lower the transaction costs investors incur. These suggest a more efficient allocation of capital within the region. From the perspective of a portfolio investor outside the region, stock market integration suggests that separate markets move together and have high correlations, so there is less benefit from portfolio diversification across countries. However, an integrated regional stock exchange will be more appealing to investors from outside the region who would find investment in the region easier or more justifiable. As shares become more liquid and transaction costs fall, fund managers become increasingly willing to take positions in the stocks. In addition, outside investors may take notice of the regional stock exchange instead of dismissing a collection of small national exchanges: the whole (one regional stock exchange) might be greater than the sum of the parts (individual country exchanges). For example, Freeman (2000) makes the argument that total equity market capitalization is important to investment managers outside the region: “Institutional investors with global portfolios may simply dispense altogether with equity markets that have low asset allocation recommendations, as resources—such as research—are limited (p. 2).” He suggests that, except for Malaysia and Singapore, equity markets in Southeast Asia may be edging toward irrelevance, and that one way to overcome the problem is to band together. Thus, an integrated stock market within the ASEAN-5 will help link the region with the world stock markets and bring more capital into the countries from abroad. This will allow ASEAN companies to expand their shareholder base and lower their cost of capital even further. In addition to interest from policy makers and investment practitioners, stock market integration also carries interest from an academic perspective. Recent advances in time series analysis allow investigation of “long run” equilibrium among stock markets using the methods of cointegration. As Kasa (1992) points out, stock markets that are cointegrated have a long-run relationship, so long-run correlations of returns are higher than short-run correlations typically examined. If n variables have p cointegrating relationships, they have n – p common trends. When n – p = 1, as in the case of the five developed-country stock indices investigated in Kasa (1992), correlations of returns converge to unity and there is no diversification potential in the long-run. In this situation, the individual stock markets are completely and perfectly integrated. However, Richards (1995) points out that a major reason for the findings in Kasa (1992) is an inappropriately long lag length used in the estimation process. With shorter lags, Richards (1995) finds that the five developed-country stock indices are not cointegrated, leaving a full set of n (=5) common trends. In this situation, the individual stock markets are completely segmented. Following Kasa (1992)—and, to a lesser extent, Richards (1995)—the technique of multivariate cointegration has been used extensively to study financial market integration around the world. For example, Corhay, Tourani Rad, and Urbain (1993) examine integration in five European stock markets over the period 1975–1991 and find one cointegrating vector. Chung and Liu (1994) consider a system including the US and five East Asian markets (Japan, Taiwan, Hong Kong, Singapore, and South Korea) over the period 1985–1992 and find two cointegration relationships. This and other studies of stock market integration in Asia are surveyed in Section 2.2 below. Recently, in what is apparently the only study of Latin America, Chen, Firth, and Rui (2002) apply the technique to study integration among Argentina, Brazil, Chile, Colombia, Mexico, and Venezuela over the period 1995–2000 and find that there is generally just one cointegrating vector. This paper specifically considers whether the stock markets of Indonesia, Malaysia, the Philippines, Singapore, and Thailand are currently cointegrated. We examine the period after the Asian financial crisis, specifically July 1, 1998 through December 31, 2002, in order to consider the recent experiences of the ASEAN-5 markets rather than a long history.2 The database thus suffers from being a short four-and-a-half year span of time from which to extract a long-run relationship, but also has the advantage of being a well-defined period during which we can reasonably say that there have been few structural breaks or shifts in the data. Since this is financial market data rather than macroeconomic data, four-and-a-half years should be a long enough time span to uncover the long-run equilibrium; financial markets are typically thought to achieve equilibrium quite quickly in contrast to macroeconomic markets characterized by price/wage stickiness, long-term contracts, adjustment costs, and other rigidities. There are three key features in our modeling strategy. One is that we consider both daily data and weekly (Friday or end-of-week) data over this period in order to examine what happens as analysts move from higher-frequency to lower-frequency data. In particular, we consider whether the lower-frequency data contains less noise and relatively more information to estimate a long-run relationship. A second feature is that we consider data denominated in local currencies, in US dollars, and in Japanese yen. Analysis is often done in local currencies, but investors outside the ASEAN countries have to convert local currency returns into their home currencies, of which the dollar and the yen are the most widely used. In the period before the Asian financial crisis, the ASEAN countries were typically pegging their exchange rates to the US dollar. This meant that the choice of local currency versus the US dollar did not matter much, but the stock market values denominated in yen were of course sensitive to fluctuations in the yen/dollar exchange rate. Currency issues have become more important in the aftermath of the Asian financial crisis, as countries have allowed their currencies to float against the US dollar (with the notable exception of Malaysia, discussed below). A third feature is that we carefully examine the lag structures of the models, and estimate cointegrating relationships in models with differing lag lengths. This allows us to determine whether our results are sensitive to the number of lags chosen. With five stock market variables, the number of common trends (n − p) can range from one to five, and this range forms something of a continuum from perfect integration to complete segmentation. 3 If the stock markets are not cointegrated, resulting in five common trends, we infer that they are nationally segmented in the economic sense, and are not yet suitable for a supranational regional stock market. However, if these stock markets are cointegrated in the econometric sense, we infer that they are integrated in the economic sense. If the number of common trends is more than one, we conclude that there is a degree of interdependence somewhat short of complete convergence, so policy initiatives to further integrate the stock markets are appropriate. If the number of common trends is exactly one, we conclude that the stock markets are completely, perfectly integrated and are ready for the establishment of a supranational regional stock market. The empirical results in this paper demonstrate that the ASEAN-5 stock markets in the period after the Asian financial crisis are cointegrated whether analyzed using daily data or weekly data, and whether analyzed in local currencies, the US dollar, or the Japanese yen. In addition, the finding does not depend on the number of lags used in estimation once a relevant range of lags is determined. The stock markets are thus not completely segmented by national borders. However, there is only one cointegrating vector among the five stock markets, leaving four common trends among the five variables. The ASEAN-5 stock markets are thus integrated in the economic sense, but integration is far from complete. Once again, this finding is robust to the frequency of the data, the currency denomination considered, and the lag lengths chosen. Perhaps surprisingly, the coefficients in the cointegrating vectors are remarkably similar across all forms of the model, suggesting that data frequency, currency denomination, and lag length have relatively little impact on the long-run equilibrium estimated. We therefore qualitatively examine and discuss the cointegrating vector in some depth, and conclude that the coefficients are reasonable for the stock markets being considered. On a policy level, we suggest that initiatives to further integrate the stock markets are feasible, and in fact desirable. From the perspective of the international portfolio investor, benefits of international portfolio diversification across the five markets are reduced but not eliminated. This paper is organized into four sections. After this introduction, Section 2 provides background on stock markets and integration in ASEAN. The first subsection considers public policies pertaining to ASEAN stock markets, with an emphasis on integrative efforts, and the second subsection surveys the academic literature on Southeast Asian stock market integration. Section 3 then considers the empirical analysis of stock market integration in ASEAN after the Asian financial crisis. The subsections consider data sources, short-run correlations of returns, unit root tests, lag length tests, and, most importantly, cointegration results. The final section is the conclusion.

The empirical results in this paper demonstrate that the stock markets of Indonesia, Malaysia, Philippines, Singapore, and Thailand in the period after the Asian financial crisis (July 1, 1998 through December 31, 2002) are cointegrated whether analyzed using daily data or weekly data, and whether analyzed in local currencies, the US dollar, or the Japanese yen. In addition, the finding does not depend on the number of lags used in estimation over a reasonable range. The stock markets are thus not completely segmented by national borders. However, there is only one cointegrating vector among the five stock markets, leaving four common trends among the five variables. We therefore conclude that ASEAN-5 stock markets are integrated in the economic sense, but that integration is not complete. Exclusion tests of the variables suggest that each country index participates in the long-run cointegrating vector, so no market should be removed from the analysis. In addition, the coefficients in the cointegrating vector are remarkably similar across all versions of the models, and are reasonable in magnitude and interpretation. One implication of cointegration is that there is less long-run diversification benefit from investing in all five countries than the short-run correlation coefficients indicate. On a policy level, cointegration suggests that initiatives to further integrate the stock markets are quite feasible, and in fact desirable from the standpoint of efficiency. In particular, since there is less long-run diversification benefit from investing across all five countries, a regional stock exchange will nudge investors to spread their money into smaller markets where they otherwise may not. In fact, investors from outside the region may value the benefits of a regional stock exchange (such as higher liquidity and lower transaction costs) and allocate more capital to the region than they otherwise would. This will allow ASEAN companies to expand their shareholder base and lower their cost of capital. From the stock market perspective, regional integration suggests that even currency unification would be feasible. Although this issue needs to consider other financial and macroeconomic issues as well,15 the point here is that efficient flows of capital across borders within the region have the capacity to mitigate the effects of any asymmetric macroeconomic shocks. The inverse relationships in cointegrating vectors among some stock market pairs suggest that such cross-border flows are already occurring. Stock market integration is thus an important component of overall economic integration and might be a useful precondition for monetary unification.