تکامل یکپارچه سازی بازار سهام در دوره پس از آزاد سازی - نگاهی به امریکا لاتین

I use ADRs to examine if the equity markets of Argentina, Chile, and Mexico have become internationally integrated in the post-liberalization period and, if not, whether direct and/or indirect barriers are the cause of segmentation. In addition, I assess the evolution of the level of integration over time to determine if these markets are converging to or diverging from integration. I find that these markets have not become integrated. More revealing is that there is no secular trend towards greater integration. In fact, the Brazilian and Mexican currency crises temporarily increased the level of segmentation of Argentina and Chile, and appear to have had a more persistent effect on the level of integration of Mexico, as this market has become increasingly segmented in the post-crisis period. It appears that both direct and indirect barriers are responsible for the segmentation.

In the late 1980s, several developing economies liberalized their stock markets in order to precipitate their integration into the world capital market.1 The process of reform was radically different from anything that had previously been attempted in these markets and held out the promise of the emerging markets reaping the substantial benefits of integration. The scale of these changes is reflected in a 1993 LatinFinance article that stated, Latin America “… has changed so much in the last three years …” because “… since 1990, one country after another has chipped away at regulatory regimes that blocked or, at the very least, discouraged the easy inflow and outflow of capital.” 2 Yet, a decade later there are doubts that liberalization has led to integration. For instance, The Economist reports that the American depositary receipt (ADR) of Taiwan semiconductor (TSMC) traded at a 70% premium relative to its underlying stocks, while an Indian high-tech ADR traded at a 150% premium, and Korean ADRs frequently traded at substantial premiums. The premium on ADRs, the article asserts, arises because these markets are not integrated into the world capital market. 3 If markets are perfectly integrated, then the ADR and its underlying stocks are perfect substitutes, and arbitrage opportunities between these securities cannot persist. However, if there exist investment barriers between the U.S. and emerging markets, then this arbitrage condition might not be immediately eliminated (see, e.g., Adler and Dumas, 1983), giving rise to persistent price differences. Academic researchers have also questioned the success of liberalization. For example, Bekaert and Harvey (1995, p. 405) find that, in the first two to three years after liberalization, “… some countries (e.g., Chile and Mexico) are becoming less integrated into the world market …,” and Bekaert and Harvey (2000, p. 601) find that liberalization “… reduces the cost of capital but perhaps by less than we expected.” In this paper, I use ADRs to examine the level of integration of a sample of emerging markets in the post-liberalization period. Specifically, I answer the questions: Have these markets become integrated in the decade after liberalization? If not, are direct and/or indirect barriers the cause of segmentation?4 In addition, I assess the evolution of the level of integration of these emerging markets over the post-liberalization period to determine if they are converging to or diverging from integration. I utilize the idea that in fully integrated markets investors are indifferent between the national markets to which they allocate funds because there is a common reward – price of risk – per unit of systematic risk exposure (see, e.g., Stulz, 1981, Harvey, 1991 and Bekaert and Harvey, 1995). I test the null hypothesis that in the post-liberalization period the prices of risks for portfolios of Latin American ADRs and the U.S. market portfolio are equal. To achieve the above, I use an asset-pricing model in which conditionally expected returns on portfolios of ADRs and the U.S. market are jointly modeled as a product of time-varying prices and quantities of equity and currency risks (see, e.g., De Santis and Gerard, 1998). Because the null hypothesis of integration is based on the equality of risk prices, the magnitude of the deviation from zero of the differences in risk prices indicates the changing level of integration over time, consistent with the time-varying integration of Bekaert and Harvey (1995), and others. If markets have not become integrated and indirect barriers were the only cause of segmentation, then underlying stocks would not price risks equal to the U.S. market, but the ADRs would (Jorion and Schwartz, 1986; and others). This is because the stringent reporting requirements for Levels II and III ADRs (used in this study) eliminate the most important indirect barriers.5 Thus, if only the ADRs price risks differently from the broad U.S. market, then direct barriers are the cause of segmentation. Hence, I include the underlying stocks in the system of estimated equations to determine which type of barrier is the cause of segmentation, if in fact the markets are segmented as suggested above. Addressing the above issues is important for several reasons. First, the potential benefits of becoming integrated are substantial for the integrating market (see, e.g., Adler and Dumas, 1983). Hence, policymakers are interested in knowing if these markets are positioned to reap these benefits. For example, if the emerging markets were to become integrated they could significantly lower their cost of capital, increase social welfare through higher economic growth rates and more efficient risk sharing, facilitate the pricing of their assets by international investors, and improve cross-border regulatory coordination (e.g., with the U.S. market).6 For U.S. investors, integration increases the benefits of international diversification (Lee and Sachdeva, 1977 and Stulz, 1981) and these benefits are more efficiently obtained by direct investment than by investing in U.S. multinational firms (Ragazzi, 1973). Second, the uncertainty about the effect of liberalization on integration offers little guidance to currently liberalized markets facing the problem of whether to extend or curtail the process. This is a dilemma because further liberalization may not lead to the promised benefits of integration if thus far it has not led to integration, but could expose the emerging markets to more of the negative consequences of liberalization. These include the destabilizing effect of “hot money” flowing across borders (Bhagwati, 1998 and Eichengreen and Mussa, 1998) and increased susceptibility to currency crises (Wyplosz, 1998). On the other hand, curtailment could reduce the opportunity for further benefits. Furthermore, if it is not clear that the last decade of liberalization led to integration, then other emerging and transitional economies that are contemplating the initiation of liberalization may be overly cautious of doing so. This is especially the case given the evidence that liberalization caused some markets to become less integrated (Bekaert and Harvey, 1995) and that further liberalization via the expansion of an ADR program retards, rather than facilitates, the development of the local market (Karolyi, 2003). Third, if the markets remain segmented due to the existence of direct barriers, then the authorities can repeal or amend the existing laws, if they consider the possible gains from further liberalization enough to outweigh the potential negative consequences. Conversely, if indirect barriers are the cause of segmentation and some are related to foreign investors' fear and irrationality (Gultekin et al., 1989 and Bosner-Neal et al., 1990), then further liberalization may not be able to effectively address these concerns. This distinction cannot be made by merely observing that restrictive laws exist, because if foreign investors can circumvent them they will not cause segmentation. For instance, while Muzere (2001) shows theoretically that imposing a minimum investment period (like Chile over the sample period) induces segmentation, Kaminsky and Schmukler (2001) note innovative ways that this restriction has been circumvented and suggest that capital restrictions do not lead to segmentation. Though it is not possible to determine the effect of each and every barrier on integration, a contribution of the paper is that I examine if currency crises impact the level of integration. This is important because the increasing frequency and pervasiveness of currency crises in the liberalized economies could induce segmentation by, inter alia, reducing the magnitude and increasing the volatility of net capital inflows and increasing the uncertainty about future government policy as they respond to new or potential crises. Foreign investors' fear of currency crises may cause them to expect large risk prices for exposure to this risk (Bailey and Chung, 1995). If investors price risks in emerging markets that they do not price in industrialized markets or if they require larger risk prices from emerging market securities, then this is by definition segmentation. Indicative of the importance of this line of enquiry are the recent calls from prominent policymakers and academics, such as Joseph Stiglitz (former chief economist at the World Bank) and Paul Krugman, for the emerging markets to re-impose capital flow restrictions that were lifted during the liberalization process in order to limit the devastation of currency crises.7 As if heeding these calls, after the Asian currency crisis Malaysia and Taiwan (temporarily) reversed some of the steps taken under liberalization (see, e.g., Kaminsky and Schmukler, 2001 and Kim and Singal, 2000). Earlier tests (e.g., Bekaert and Harvey, 1995) were unable to address these issues because they coincide with the period in which the major reforms were still underway (1989–1992). Consequently, they capture only the immediate impact of liberalization. Moreover, even where earlier studies cover the same period and markets and use broadly similar methodologies, they are fraught with conflicting evidence as to how integrated these markets are.8 Recently, Bekaert et al. (2002) dates the integration of the emerging markets by identifying an endogenous structural break in multiple financial and economic series that are likely to be related to the integration process. They find that these break dates are within three years after some official liberalization date. Though insightful, as the authors note, a structural break “… is a necessary but not sufficient condition for concluding that capital market liberalizations actually served to integrate the capital market (p. 208).” For instance, if foreign portfolio investments in the emerging markets increased significantly at a particular date, then this date is identified as the break date for that variable. However, if after the structural break foreign investors demand a different compensation for a unit exposure to risk from emerging market assets relative to that from developed market assets then these markets cannot be regarded as being integrated. Nonetheless, in this paper I take a conservative position and test for integration using data with starting dates corresponding to a minimum of two years after the initiation of the respective market liberalization. If the markets in fact became integrated after the structural break dates of Bekaert et al. (2002), then the current paper's results should indicate this.9 In this study, I focus on Argentina, Chile, and Mexico (for reasons discussed below). I find that these emerging markets have not become integrated into the international capital markets in the post-liberalization period. More revealing is that there is no distinct trend towards greater integration. In fact, for the sample of countries, currency crises significantly reduce the level of integration, regardless of the type of exchange rate regime. The region's currency crises temporarily increase the level of segmentation of Argentina and Chile, while Mexico has become increasingly segmented in the post-crisis period, suggesting that currency crises might have a more persistent effect on the level of integration. It appears that both direct and indirect barriers cause the remaining segmentation. The remainder of the paper is as follows. I discuss the use of ADRs in tests of integration and justify the selection of the sample of markets in Section 2. In Section 3, I explain the methodology and in Section 4 discuss the data. The main results are reported in Section 5 and the paper's summary is in Section 6.

In this paper, I use ADRs to examine if the equity markets of Argentina, Chile, and Mexico have become integrated into the world equity market in the decade after liberalization and, if not, whether direct and/or indirect barriers are the cause of segmentation. In addition, I assess the evolution of the level of integration of these markets over the post-liberalization period to determine if they are converging to or diverging from integration. The test is based on the null hypothesis that if the markets are integrated, then the prices of systematic risks of portfolios of the region's ADRs are the same as the risk prices of the U.S. market portfolio. I find that liberalization has not successfully led to a high and sustained level of integration of these Latin American markets into the international capital market. This is unlike the experience of industrialized markets such as Canada (Mittoo, 1992), Germany and the U.K. (Cumby, 1990), and Japan. For instance, Japan became fully integrated with the U.S. within five years after opening up their market through the implementation of the Foreign Exchange Law in 1980 (Gultekin et al., 1989 and Campbell and Hamao, 1992). However, despite the failure of liberalization to lead to full integration, the emerging markets, as a whole, have experienced a small increase in economic growth rates (Bekaert et al., 2001), a small reduction in the cost of capital (Bekaert and Harvey, 2000), a temporary rise in stock prices (Henry, 2000a), and a short-term increase in private physical investments (Henry, 2000b) as a result of liberalization. Unfortunately, these benefits are not as significant as expected, prompting Bekaert and Harvey (2000, p. 601) to proclaim that the reduction in the cost of capital is “… less than we expected.” The results also indicate that there is no tendency towards an increase in the level of integration over the sample period and that the level of integration in all three markets is negatively affected by currency crises. In fact, Mexico has become less integrated following the peso crash of December 1994. Further, it appears that though many government-imposed impediments and other barriers have been lifted, there are remaining binding direct and indirect barriers. The results have implications for foreign investors seeking international portfolio diversification by investing in the emerging markets. Foreign investors must be cognizant of the fact that these markets have evolved to the point where the pricing of their securities is substantially influenced by international factors. However, that they are segmented suggests that they continue to be influenced also by local factors. Additionally, emerging and transitional economies contemplating opening their markets to foreign investors for the first time are to be aware that liberalization does not necessarily lead to full integration and its potential benefits. Moreover, when currency crises occur they negatively affect the level of integration. It appears that the type of exchange rate regime is immaterial, as the level of integration of markets with fixed and floating rate regimes is affected by these crises. This raises the issue of whether or not the increased inflow of capital following liberalization is worth the potential devastation of currency crises. No doubt this will be the subject of future studies.