University of Wales Swansea, United Kingdom
Stochastic Trends in Stock Prices: Evidence from Latin American Markets *
This paper investigates the long-run relationship between stock indices from six Latin American markets and the United States. The empirical investigation is conducted using weekly data from January 1989 to December 1993, unit root tests, cointegration tests,and error-correction models. Results from the unit root tests provide evidence of a stochastic trend in all indices. Results from the cointegration tests indicate the presence of a long-run relationship between the six Latin American indices (with and without the United States index). Error-correction results indicate signiﬁcant causality among the stated indices.
1. Introduction In recentyears the quantity of research on interdependence of stock markets has been high and extensive (Corhay, Tourani, and Urbain 1993; Koch and Koch 1991). In an integrated world equity market, individual stock prices are expected to have long-run relationships, i.e. share common stochastic trend(s).1 There are several reasons why different countries’ stock prices may have a signiﬁcant long-runrelationship(s). The presence of strong economic ties and policy coordination between the relevant countries can indirectly link their stock prices over time. Stock prices are affected by real interest rate movements, thus real interest rates linkage between countries due to international capital ﬂows can contribute to long-term relationships between different stock prices. The importance of internationalinvestors
*I thank two anonymous referees for several helpful comments and suggestions. The remaining errors and omissions are my responsibility alone. 1 Baillie and Bollerslev (1989) and Hakkio and Rush (1989) claim that in an efﬁcient market, assets prices cannot have a long-run relationship(s). According to these studies, in an efﬁcient market, changes in asset prices cannot be predicted butdeviations of prices from a long-run relationship indicates predictable future changes. Dwyer and Wallace (1992) show that there is no general equivalence between market efﬁciency and lack of a long-run relationship between assets. Dwyer and Wallace base their analysis on zero transaction cost assumption and market efﬁciency deﬁned as the lack of arbitrage opportunities.
Journal of Macroeconomics,Spring 1997, Vol. 19, No. 2, pp. 285–304 Copyright 1997 by Louisiana State University Press 0164-0704/97/$1.50
Tauﬁq Choudhry can further induce co-movements in national stock prices. Jeon and Chiang (1991) state that international linkage between different equity markets could also be due to the recent deregulation and liberalization of different markets, improvement and developmentof communications technology, innovations in ﬁnancial products and services, increase in the international activities of multinational corporations, etc. This paper provides an investigation of long-run relationship(s) between six Latin American stock indices and the United States stock index. Stock price indices from Argentina, Brazil, Chile, Colombia, Mexico, and Venezuela are used in theempirical investigation. The stated markets are part of what is known as emerging markets.2 During the last decade or so the world ﬁnancial market experienced a rapid growth in emerging stock markets (IFC 1993). From 1982 to 1992, the emerging markets increased their market capitalization from $67 billion to $770 billion and nearly tripled their share of world equity capitalization from 2.5% to 7.0% (IFC1993). From 1988 to 1992, markets in Argentina, Chile, Colombia, Mexico, and Venezuela rose more than 100% in dollar terms as compared to 51% rise in the United States market. International portfolio investment in the developing countries has grown to $50 billion in 1992 from a few hundred million dollars in the early 1980s (IFC 1993). The volume of international equity issues in these markets...