Journal of World Business 39 (2004) 233243

Strategic factors affecting foreign direct investment decisions by multi-national enterprises in Latin America
Len J. Trevinoa,*, Franklin G. Mixon Jr.b
a

Department of Management and Decision Sciences, Washington State University, Pullman, WA 99164-4736, USA b Department of International Business and Economics, Box 5072, The University of Southern Mississippi, Hattiesburg, MS 39406-5072, USA

Abstract Cross-country differences in macroeconomic and institutional environments are used to explain MNE behavior, as proxied by foreign direct investments (FDIs) inows to seven Latin American countries, namely Argentina, Brazil, Chile, Columbia, Mexico, Peru and Venezuela for the period 19881999. Results indicate that the institutional approach is dominant, thus supporting recent FDI research that has included statistical measures of institutional reform in their models. Since MNEs must conform to the institutional environment prevailing in the host country, managers should undertake FDI where there is minimal institutional distance between the home and the host country environments. In addition, government ofcials should place increased emphasis on institutional reform if their objective is to increase inward FDI in their countries. Finally, any assistance provided by non-governmental organizations, such as the IMF and the World Bank, should also emphasize institutional reform. # 2004 Elsevier Inc. All rights reserved.

1. Introduction The opening of markets in developing countries in recent years has brought with it burgeoning foreign direct investment (FDI) ows. In the 1990s, FDI became the largest single source of external nance for developing countries. By 1997, FDI accounted for about half of all private capital and 40% of total capital ows into developing countries. In the past, governments in many developing countries often saw multinational enterprises (MNEs) as part of the development problem, due to assertions of exploitation of the environment and of the labor force. At present, MNEs are seen as part of the development solution for several reasons. First, governments in developing
Corresponding author. Tel.: 1-509-335-7850; fax: 1-509-335-7736. E-mail address: trevino@wsu.edu (L.J. Trevino).
*

countries acknowledge that they need outside capital to achieve their development objectives, partly because industrial nations have stabilized foreign aid and development loans. Second, export-oriented FDI brings relief from rampant foreign exchange shortages. Third, recognizing that reversal of portfolio investment is less costly, a fact that exacerbated recent nancial crises in a number of developing countries, governments now prefers FDI (UNCTAD, 1999). Fourth, host-country governments recognize that MNEs have access to resources other than capital, that can assist with their development (such as technology, management and access to foreign markets). Recognizing the long-term costs of failure to integrate their economies into the global environment, developing countries have opened up their markets in order to attract more FDI. There are numerous theories that have been advanced to explain this phenomenon. The macroeconomic approach (Aliber, 1970; Froot & Stein, 1991;

1090-9516/$ see front matter # 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jwb.2004.04.003

234

L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243

Grosse & Trevino, 1996) emphasizes why net investment among pairs or groups of nations tends to ow in certain patterns. This theory attempts to explain FDI behavior with macroeconomic variables, such as ination, national income and exchange rate behavior. With developing countries undertaking market reforms and becoming more receptive to FDI, researchers have begun to apply institutional theory from the strategic management literature (Kennedy & Sandler, 1997; Trevino, 1999; Trevino, Daniels, & Arbelaez, 2002) to understand this phenomenon. Institutional theory emphasizes the inuences of systems surrounding organizations that shape organizational behavior and decision making (Scott, 1995). As such, it attempts to explain the organizationenvironment interface. According to Hoskisson, Eden, Lau, and Wright (2000), the role of institutions in an economy is to reduce transaction and information costs by reducing uncertainty and by establishing a stable structure that facilitates interactions. Empirical research using an institutional theoretical approach has emphasized the study of political risk, bilateral investment treaties, foreign investment and trade regulations, and capital markets liberalization in an attempt to explain FDI. Although both of these approaches have been employed separately (Grosse & Trevino, 1996; Meyer, 2001) and together (Trevino et al., 2002), no effort has been made to determine the individual importance relative to each. This paper lls that void by presenting hypotheses concerning these two approaches. The empirical results support the superiority of the institutional construct, thus lending credence to recent studies that have employed institutional theory to explain FDI into developing and transitional economies (Trevino et al., 2002). Latin America is a useful region for our study because Latin American and Caribbean countries receive a signicant portion of FDI inows going to developing countries (UNCTAD, 1999). In addition, debt crises in this region resulted in reduced FDI inows during the rst half of the 1980s, after which they began a steady increase, partially resulting from macroeconomic and institutional liberalization policies. Although institutional reform has taken place in almost all countries of the region, Latin American countries liberalization policies, market reforms and inows of FDI have varied cross-sectionally and over time. The seven countries examined in the current

study were Argentina, Brazil, Chile, Colombia, Mexico, Peru, and Venezuela. Combined, they account for over 85% of FDI in Latin America.

2. Foreign direct investment in Latin America: a brief history Foreign direct investment in Latin America has a long history, often dating back to the 19th century. Early FDI was primarily export-oriented and/or driven by MNEs seeking natural resource supplies. Import substitution industrialization in the post-WWII era led to a shift in FDI toward manufacturing for domestic consumption. Lacking foreign exchange in the 1980s, many governments in Latin America began to open up their markets in a return to export-oriented FDI. Foreign direct investment in Argentina was prevalent from the late 19th century until the beginning of WWII. It was initially sought to improve the transportation infrastructure. However, the post-WWII era produced nationalist economies throughout most of Latin America and, in the case of Argentina, the closed nationalistic economy continued until the late 1980s, culminating in macroeconomic dislocation. This economic and institutional landscape led to capital ight and ultimately to market reform, characterized by trade openness, deregulation and privatization. These reforms led to increased FDI ows. The foundation of Argentinas economic transformation was the Convertibility Law, a currency board implemented in 1991. Other important elements included trade liberalization, privatization, tax reform, and deregulation (Petrocella & Lousteau, 2001). In the beginning of the 1990s, the vast majority of FDI ows came in the form of privatizations. After 1993, FDI ows increasingly took place in the private sector. The role that FDI plays in the modern Brazilian economy is different than that which it played in previous eras. Prior to WWII, FDI was concentrated in public utilities, including transportation, in the primary goods export economy, and in banking, with a small percentage in the manufacturing sector. Similar to Argentina, post-WWII FDI ows shifted to manufacturing as part of an import substitution industrialization strategy. In the 1990s, the role that FDI played changed considerably. Brazil adopted institutional and macroeconomic reforms, partially designed

 L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243

235

to stimulate FDI. Elements of these reforms included establishment of the real plan, privatization of stateowned enterprises, and implementation of the Mercosur free trade area. It has been estimated that of the worlds 500 largest corporations, 400 have investments in Brazil (Baer & Rangel, 2001), making Brazil a major source of investment for MNEs and emphasizing the importance of Brazil to the strategy of major MNEs. Chiles history of FDI is similar to that of other Latin American countries, having shifted from investment in infrastructure development and primary products to a post-WWII import substitution industrialization policy and, more recently, to manufacturing. Although Mexico is known for being one of the earliest countries in Latin America to undertake macroeconomic and institutional reforms, primarily because of the publicity surrounding NAFTA, Chile was the rst Latin American country to liberalize its foreign investment regulations. Decree Law 600 was promulgated in 1974 and it liberalized tax codes, repatriation restrictions, and the foreign exchange market. The countrys market reform policies appear to have been effective because Chile received over $7 billion in FDI inows between 1987 and 1994 (Ramirez, 2001). Continuing with institutional reform, the June 6, 2003 signing of a ChileU.S. free-trade agreement established clear and transparent rules for foreign investors, including an open system for dispute settlement. Although Colombias history of FDI parallels that of other Latin American countries, liberalization did not take place as a reaction to an economic or foreign exchange crisis, or at the beginning of a political administration. Nevertheless, radical reforms were implemented between 1990 and 1994. These included dramatic tariff reductions, liberalization of trade and foreign exchange transactions, and nancial and capital markets reform. Although Colombia had fewer state-owned enterprises than other Latin American countries, in the early 1990s, some of the banks the government had absorbed in the nancial crisis of the early 1980s were reprivatized, and the governments stake in the automobile manufacturing sector was sold (Birch & Halton, 2001). Mexico, along with Chile, was one of the rst countries to abandon import substitution industrialization in favor of an open, market-oriented model that

culminated with the passage of NAFTA. Although Mexicos FDI history is similar to that of other Latin American countries, its recent past differs substantially in that FDI has owed primarily to newly created rms in the maquiladora sector along the U.S. border. In other countries, FDI funds have been primarily channeled to traditional sectors, such as mining and energy. The impetus to change Mexicos FDI policy can be traced to the countrys desperate need for foreign exchange following the 1982 debt crisis. Elements of market reform included opening of markets previously closed to foreigners, privatization, deregulation, national treatment and, of course, NAFTA (Ramirez, 2001). Peru began to open its doors to external capital in the early 1990s, somewhat later than other Latin American countries. In an economy fraught with populist regimes, foreign debt crises, and political and economic instability, Peru had been effectively closed to FDI. However, since 1991, macroeconomic and institutional changes included the introduction of a free market policy, the deregulation of prices, the adoption of monetary and scal control measures (designed to reduce ination), and privatization (Rojas, 2001). All of these elements sent clear signals that Peru wanted to become a member of the global marketplace. Like other countries in the region, privatization has led to increased FDI ows in Peru. Oil has dened Venezuela and its FDI regime for over a century. In 1991 and 1992, after a number of technological and nancial failures, the state oil company, Petroleos de Venezuela, signed 11 letters of intent with transnational oil companies interested in pumping crude. Recently, the combination of severe economic crisis, coupled with foreign currency restrictions, has pushed Venezuela farther from reforms. The countrys desperate need for foreign capital continues to conict with its historical desire to exercise sovereignty over its natural resources.

3. Literature review of macroeconomic variables For developing countries to compete for FDI inows, they must implement macroeconomic policies designed to reduce ination, stabilize the exchange rate and increase the GDP of the host

236

L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243

country. With market-oriented economies in Latin America effectively in operation for little more than a decade, instability of prices, employment and output would be expected. A high rate of ination is a sign of internal economic instability and of a host governments inability to maintain expedient monetary policy. From the MNEs viewpoint, high ination creates uncertainty regarding the net present value of a costly, long-term investment. For these reasons, companies may avoid making investments in countries with high ination. Studies published before Latin American countries made signicant reforms (Schneider & Frey, 1985) as well as those published after reforms were enacted (Trevino et al., 2002) conrmed that companies invested less in developing countries with high ination rates. The value of a countrys currency may be undermined by monetary policy or by economic upheaval. Currency devaluation may result from such policy changes, and foreign investors must incur costs to prevent transaction and translation losses when host country currencies depreciate. Thus, ceteris paribus, a constant real exchange rate is preferred by MNEs in order to reduce the exchange rate risk inherent with investment in a foreign country. Another perspective suggests that currency under (over)valuation is an example of market disequilibrium. A currency may be dened as undervalued when, at the prevailing rate of exchange, production costs for tradable goods are, on average, lower than in other countries. In this case, MNEs would be inclined to locate production of internationally traded goods in countries with undervalued currencies and to purchase foreign production capacity with overvalued foreign exchange (Froot & Stein, 1991; Grosse & Trevino, 1996; Klein & Rosengren, 1994). From an additional perspective, the apparent under(over) valued exchange rate based on a rms production costs (i.e., wages for a laborabundant country) denotes that the country has a comparative advantage in producing the product and MNEs would be inclined to locate a plant within that host country. The demand-side of FDI theory argues that investment will go primarily to markets large enough to support the scale economies needed for production. This reasoning helps to explain why most FDI goes to developed countries rather than to developing countries (Grosse & Trevino, 1996), given that most

investment historically has been market seeking. Investment in developing countries has been in response to import substitution policies. Although Tuman and Emmert (1999) used GDP as a surrogate for market size and found it to be insignicant in explaining FDI among Latin American countries, more recently Trevino et al. (2002) found that GDP was a signicant and positive indicator of FDI ows in Latin America. Further, UNCTAD (1994) concluded that market size was the primary determinant of FDI.

4. Literature review of institutional variables Institutional theory in the strategic management literature suggests that institutions provide the rules of the game that structure interactions in societies and posits that organizational action is bound by these rules (North, 1990). Within this realm lies political risk, which may be dened as the risk that a host country government will unexpectedly change the institutional environment within which businesses operate (Butler & Joaquin, 1998). From a nancial perspective, political risk may alter operating cash ows via discriminatory policies and regulations. MNEs may deal with political risk by avoiding the risk altogether, by buying insurance, or by negotiating with the governing body prior to investment. Although previous studies reached mixed conclusions about the effect of political risk on FDI (Grosse & Trevino, 1996; Kobrin, 1979; Tallman, 1988), we expect a country with high political risk to be less appealing to foreign investors. Capital markets are responsible for mobilizing and allocating capital and for apportioning risk. It is benecial when a host country ensures that capital ows to its most optimal use and is allocated for economic, rather than for political reasons. In order for developing countries to attract FDI, they must attempt to enforce a capital allocation system with strict and transparent rules and regulations. At the same time, they should not exert excessive control over capital account transactions, such as via exchange-rate controls and/or repatriation or foreign ownership restrictions. In an effort to spur internal development, many Latin American countries have enacted capital market reform. If governments maintain strict control over capital transactions, such as via

 L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243

237

foreign exchange controls and restrictions on FDI, then MNEs may be reluctant to invest due to fears about restrictions on new capital formation, divestment and repatriation. Latin American countries, like their counterparts in other regions of the world, have been privatizing government-owned companies. The primary reason is that many of these companies operated inefciently under government ownership. These governments believe they can reduce their scal expenditures because they will no longer need to subsidize money-losing operations. In addition to signaling a more favorable investment climate, governments can receive tax revenue from them if they become protable. From the MNEs perspective, the potential for cost savings by transferring technology and management capabilities to the privatized rm is often present because most of these companies operated inefciently under government ownership. Recent studies of privatization in Latin America (Devlin & Cominetti, 1994; Hartenek, 1995; Trevino et al., 2002) concluded that privatization has helped to attract FDI to the region.

Based on the discussion in the previous two sections, we expect that:1 b1 ? b2 < 0 b3 > 0 b4 > 0 b5 > 0 b6 > 0 Table 1 presents the ndings of an OLS estimation of Eq. (1).2 The six regressors are jointly signicant in explaining FDI across countries/time. They produce a sizable R-square for panel data. In four of ve cases, our expectations are borne out regarding the signs of the variables. In the case of CPIPC (i.e., the unexpected sign), the parameter estimate is not signicant at any conventional level. Parameter estimates for RERT, GDPC, and PRSK all are signicant at the 0.05 level. These results support recent ndings by Trevino et al. (2002). Lastly, Eq. (1) passes a specication error F test (RESET)failing to reject the null hypothesis of no specication errorat the 0.10 level.3

5. Modeling FDI within the macroeconomic and institutional frameworks Eq. (1) presents a test of the macroeconomic and institutional theories: FDI b1 RERT b2 CPIPC b3 GDPC b4 CALI b5 PRIV b6 PRSK e: (1) In Eq. (1) FDI represents inward FDI for the seven Latin American countries under study for the period 19881999. RERT is the real exchange rate of Latin American currencies at year-end, per U.S. dollar. CPIPC represents the annual percentage change in consumer prices in the host countrys currency. GDPC is the host countrys per-capita gross domestic product, in U.S. dollars. CALI is the degree of the host countrys control over capital account transactions on an annual basis. PRIV is the value of domestic privatizations (less FDI) in each country. Lastly, PRSK is each host countrys political risk rating. See the Data Sources and Descriptions at the end of this study for data sources and a more detailed description of the variables.

1 This study uses the political risk measure published by Institutional Investor (various issues), where a higher number indicates more political stability. Although we expect a negative relationship between political risk and FDI, this would show up as a positive coefficient in the equation. Data sources for the other variables in Eq. (1) are provided below in a separate section. 2 As Table 1 notes, we used 56 observations in our regressions. For some variables (years), we had missing data. Specically, observations on CALI and PRSK cover the period 19881995, those on GDPC and CPIPC cover the periods 19881997 and 19881998, respectively, while those on RERT and PRIV cover the period 19881999. We selected all the observations at our disposal where there were no missing data for any of our variable series. This selection process facilitated the tests of non-nested hypotheses we detail below. 3 We use the standard RESET procedure from Ramsey (1969). First, predicted values for FDI are obtained from the OLS model. These predicted values are squared and cubed, and enter equation one as additional regressors (Gujurati, 1988). A RESET F statistic of 1.651 (2, 48 df) is produced regarding the joint signicance of these two additional regressors. The insignicance of the F statistic fails to reject the hypothesis of no specication error in the base regression model. Ramseys RESET is a useful test for detecting many types of specication error, such as omitted variables, irrelevant variables, nonlinearity, and errors in measurement (Gujurati, 1988; Kmenta, 1986).

238

L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243

Table 1 Summary of OLS and LSDV regression results dependent variable: FDI OLS Constant RERT CPIPC GDPC CALI PRIV PRSK Country dummies Brazil Chile Columbia Mexico Peru Venezuela No. of observations F statistic Adjusted R2 56 5.01 0.304 2.365.02 (1.67) 28.63 (2.01) 0.07 (0.25) 0.50 (2.10) 4.114.34 (2.36) 0.30 (1.78) 83.29 (2.95)

6. Macroeconomic and institutional theories of FDI: is either dominant? In order to discern the relative importance of the macroeconomic and institutional theories of FDI, additional empirical testing is necessary. Following Maddala (1992: 514518), we re-specify the two empirical approaches to FDI as Eqs. (2) and (3): H0 : FDI bXM u0 ; H1 : FDI dZIR u1 : (2) (3)

LSDV 8.737.99 (1.39) 5.62 (0.39) 0.15 (0.77) 2.32 (1.67) 785.18 (0.40) 0.23 (1.59) 208.80 (3.43) 5.997.70 10.990 13.723 4.846 7.669.54 9.265.72 56 9.53 0.651

Notes. The numbers in parentheses above are t values. For the country dummies, t values are not presented to conserve space, though signicance levels are included. Denotes signicance at the 0.10 level. Denotes signicance at the 0.05 level. Denotes signicance at the 0.01 level.

Because heterogeneity or country effects are very common in panel data, we also employ xed effects, least squares dummy variable (LSDV) approach to modeling Eq. (1). This approach tests for a common constant intercept term by including a dummy variable series for our seven Latin American countries. In our specication, Argentina serves as the omitted country. Results from the LSDV model also are presented in Table 1. The LSDV results mirror those of the OLS version in terms of the signs for CALI and PRSK. CPIPC retains its expected sign in the LSDV model, though not in the OLS specication. However, GDPC and PRIV retain anomalous signs in the LSDV specication, though neither is better than marginally signicant in the LSDV column. PRSK is statistically different from zero, as are most of the country dummy coefcients. In fact, a joint F test on the signicance of the set of country dummies produces a test statistic of 9.305 ([6, 44] df), suggesting that the LSDV model be used instead of the OLS to determine the relative importance of each FDI theory.

Again, FDI represents inward FDI for the seven Latin American countries under study for the period 1988 1999, XM is the vector of the three macroeconomic determinants (along with the country dummies) of FDI, and ZIR is the vector of the three institutional variables (along with the country dummies) under consideration. The separate hypotheses in Eqs. (2) and (3) above are said to be non-nested given that neither Eq. (2) nor Eq. (3) can be obtained from the other by imposing a restriction (Kennedy, 1998). That is, each equation has variables not included in the other (Davidson & MacKinnon, 1993; Kmenta, 1986; Ramanathan, 1998). It is often the case that there are competing theories that attempt to explain the same dependent variable, and the explanatory variables in the different theories are non-overlapping. Maddala (1992) suggests the use of various joint test (J tests) procedures for testing the competing theories, but notes that one limitation of [these] test[s] is that [they] sometimes . . . reject both H0 and H1 or accept [both] H0 and H1 (Maddala, 1992: 516). For the purposes of this study, Maddalas concerns are not relevant. We follow the novel use of J tests in Mixon and Gibson (2001), where complementarity of theories was tested, rather than competitiveness, and their J test nding that both H0 and H1 were accepted supported that contention. We also note that the models in Eqs. (2) and (3) are not necessarily competing theories, however they are non-nested and the tests suggested by Maddala (1992) may discover the dominant source of FDI. The procedure for testing H0 (the maintained hypothesis) against H1 is as follows: Eq. (3) is estimated by LSDV and the predicted values of FDI are obtained (predFDIIR). Eq. (4) below is then estimated: FDI bXM apredFDIIR v: (4)

 L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243

239

The hypothesis that a 0 is tested. If this hypothesis is not rejected, then H0 is not rejected by H1. If the hypothesis is rejected, then H0 is rejected by H1. A test of H1 (as the maintained hypothesis) against H0 is based on analogous steps, and examines the statistical signicance of l (where l is the counterpart to a). The results are presented in Table 1 as the Davidson MacKinnon J tests (Davidson & MacKinnon, 1981; MacKinnon, 1983). The results indicate that the institutional model signicantly adds to the macroeconomic model in explaining inward FDI in Latin American countries. This result supports the contention that recent empirical considerations of various measures of institution building have been benecial and that empirical use of these broad concepts has added to our understanding of the determinants of FDI. On the other hand, the J test for H1 vs. H0 fails to indicate that the macroeconomic model signicantly adds to the institutional model. In fact, the coefcient for l is negative (i.e., 0.402). Maddala (1992) explains that J tests are, in small sample cases, sometimes less rigorous than traditional F tests. He concludes, what . . . all this suggests is that in testing non-nested hypotheses, one should use the J test with higher signicance levels and supplement it with the F test on the comprehensive model . . . (Maddala, 1992: 518). F tests are performed by comparing an unrestricted model with all six regressors and the country dummies to the two models in Eqs. (2) and (3), each of which is restricted to three regressors and the country dummies. Unlike the one-degree-of-freedom J tests, the F tests reported in Table 2 are based on the error sum of squares (ESS) from LSDVestimation and represent [3, 44] degrees of freedom. Our F tests, as did the J tests
Table 2 Tests of non-nested hypotheses (H0 macroeconomic model; H1 institutional model) DavidsonMacKinnon J tests Test of H0 over H1 Test of H1 over H0 1.249 (4.50) 0.402 (0.89)

above, employ the same data panel as the OLS and LSDV regressions reported in Table 1. The F statistics in Table 2 support the conclusion of the J tests. That is, while institutional theory adds to macroeconomic theory in explaining FDI, macroeconomic theory fails to statistically extend institutional theory in this regard. This nding lends support to newer studies that have examined the relationship between institutional reform and FDI (see Hoskisson et al., 2000; Trevino et al., 2002).

7. FDI decisions by MNEs in Latin America in response to institutional reform The relationship between capital markets liberalization in Latin America and the expansion strategies of international banks into the region is one example of FDI decisions made by MNEs in response to institutional reform. Although policies introduced in Latin America over the last two decades have differed from country to country, internal reforms can be segmented into two phases. In the rst phase, interest rates were allowed to be determined by market forces, instead of by at, and nancial resources were allocated on the basis of supply and demand. During the rst phase of reforms, banks were invited to operate in markets in which they did not have prior access. These included areas such as leasing and factoring, brokerage underwriting and pension fund management. With economies of scale and scope becoming an increasingly important factor to foreign banks, the ability to operate in these areas was seen as increasingly important to their plans to globalize operations. In the second phase, banks were required to maintain more conservative capital to asset ratios, to impose more stringent requirements on loans, and to provide more uniform and transparent information to governing bodies. The nancial reforms enacted during the 1990s openly encouraged the entry of foreign banks into Latin America. The second phase of reforms created a regulatory environment similar to that of international banks home country environments, in the process creating a more certain investment climate and opening the door for foreign banks to operate in the local market. Capital markets liberalization, thus, helped to create an environment with appropriate institutions and removed the entry barriers for foreign

F tests 7.222 1.098

Notes. The numbers in parentheses above are t values for the J tests regression parameters. The F statistics above reect tests with [3, 44] df. Denotes signicance at the 0.01 level.

240

L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243

banks, enabling them to expand into Latin America and to gain an increasingly large market share in the region (ECLAC, 2003). One example of a multinational bank that has made signicant inroads into Latin America in the postreform era is Citigroup, one of the worlds largest nancial institutions. In fact, it is the largest bank in Latin America, based on share of regional lending (ECLAC, 2003). Although Citigroup has had a presence in Latin American for well over a century, its presence was largely limited to corporate and private banking services. This strategy was due in large part to its awareness of institutional weaknesses in Latin America that led to the debt crises of the 1980s and 1990s, and ultimately to institutional reform. The banks strategy shifted abruptly in 2001, when it moved full force into consumer banking with the well-publicized $12.5 billion acquisition of Mexicos BanamexAccival nancial group. This was the single largest FDI transaction in Latin America to date. Citibanks bold shift in strategy allowed it to gain control of over 25% of consumer and corporate banking in Mexico. Citibanks FDI decision making in Latin America is directly related to the liberalization and deregulation of international nancial markets in the region. Although Citigroup has made signicant inroads into Latin America, two Spanish banks, Banco Santender Central Hispano (SCH) and Banco Bilbao Vizcaya Argentaria (BBVA), expansion strategies in the region have been even more aggressive. Their initial strategy was to become a force in consumer banking, which had shed itself of most institutional controls and that was seen as having the greatest growth potential. Both of these Spanish banks went on a buying spree in Latin America during the time when rst and second generation reforms were implemented. Between 1997 and 2002, SCH made 26 acquisitions valued at over $13 billion in all of the countries that we studied, in addition to Bolivia and Paraguay. During the same timeframe, BBVA made 12 acquisitions valued at over $6 billion in six of the seven countries that were the focus of this study in addition to an investment in Uruguay. Spanish banks acquisition strategies in Latin America can be traced to liberalization policies within the European Union because these internal liberalization policies pushed European banks into expansionist modes to gain economies of scale. Coupling the need to globalize

operations resulting from liberalization policies within the European Union with institutional reform in Latin America, SCHs and BBVAs international expansion in Latin America was a natural outcome. Institutional reform was initiated in the early 1990s in much of Latin America in response to shortcomings in many sectors, including a lack of public funds for investment and gaps in technology. During this timeframe, governments introduced reforms designed to attract foreign private capital. Although reforms took place in many sectors, nowhere is this policy more evident than in the telecommunications sector. In the early years of reform, many Latin American countries privatized their public telecommunications companies and allowed unprecedented foreign participation, attracting large investments and substantial improvements in the telecommunications infrastructure. It was not long before the impact of privatization programs in Latin America was realized, with Telefonica of Spain making its presence felt in the region early on. In 1991, it became part of a consortium led by GTE Corporation (now Verizon Communications) that became majority owner of Telefonos de Venezuela, with a nearly $2 billion investment. Shortly thereafter, Telefonica of Spain expanded its Latin American operations, investing another $2 billion to acquire a 40% stake in two Peruvian companies, Empresa Nacional de telecommunicaciones, a domestic and international long distance monopoly, and Compania Peruana de Telefonos, the local telephone company of Lima, the countrys capital. The two companies subsequently merged to form Telefonica de Peru, and in 2000 Telefonica of Spain gained complete control of this company with an investment of over $3 billion. Telefonica of Spain is a prime example of a MNE that has responded to privatization reform and it has consolidated its position as a leading provider of xed and mobile telephony in both Venezuela and Peru. In Peru, it has retained its dominant position, but in the larger market of Venezuela, it has faced considerable competition due to liberalization of the mobile market combined with the expansionist strategy of Bell South. Institutional reform and privatization in the region has attracted other international investors into Peru and Venezuela, such as Bell South of the United States. Bell South responded to Latin Americas desire for increased FDI in telecommunications and, as such,

 L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243

241

it has become the leading provider of mobile telephone services in Venezuela and Colombia and number two in Peru. In fact, Bell Souths Venezuelan subsidiary has captured nearly 60% of the market and plans to invest an additional $1.5 billion by 2005 (ECLAC, 2003). More recent acquisitions made by Bell South in response to concessions made by the Colombian government include a $300 million stake in Celumovil, a mobile telecommunications company operating in 6 of Colombias 10 largest cities. This investment, coupled with the acquisition of Compania Celular de Colombia, for which it paid over $400 million, gave Bell South the largest share of the Colombian mobile telecommunications market (ECLAC, 2003). In Peru, where privatization of the telecommunications sector has been more liberal than in Colombia, Bell South bought Tele 2000, a multiservice telecommunications company, initiated its own ber optics network in the country and introduced the Bell South brand. The Bell South example highlights MNEs FDI strategic decision making in Latin America in response to host governments privatization strategies designed to attract FDI. Institutional reforms introduced in Latin American countries in the early 1990s were swift and decisive and the MNE strategic response was predictable. As a result of institutional reform in the region, the Peruvian government estimates that it has received over $4 billion in private investment in the telecommunications industry between 1997 and 2001 and that additional agreements have been signed with Telefonica of Spain and with AT&T of the United States worth another $5 billion. Similarly, over $5 billion in private capital owed into the telecommunications industry in Venezuela between 1997 and 2001.

8. Managerial implications Most agree that the economies of developing countries are inextricably tied to MNEs and private investment. In addition, it has been shown that MNEs can enhance their own bottom line by investing even in the poorest economies of the world (Prahalad & Hammond, 2002). While earlier explanations of FDI were dominated by economic theories, more recently institutional theory and institutional distance (i.e., the extent of similarity between the regulatory, cognitive

and normative institutions of two countries) provide alternate and complementary explanations for MNE behavior (see Kostova & Zaheer, 1999; Trevino et al., 2002; Xu & Shenkar, 2002). In this paper, we use cross-country differences in macroeconomic and institutional environments to explain MNE behavior, proxied by FDI inows to seven Latin American countries. Because it is well established that in order to survive, MNEs must conform to the institutional environment prevailing in the host country (DiMaggio & Powell, 1991), rms should understand the level of macroeconomic and institutional reform that has taken place in proposed host countries, such as those in Latin America. To assist in this understanding, managers of MNEs could develop and/or apply separate statistical indices comprising macroeconomic and institutional information for proposed host countries. Empirical results presented here suggest that managers should take particular care to examine host country institutional environments (reforms), and a longitudinal data base (of indices for individual countries) could be useful to managers in formulating FDI strategies. These indices could go well beyond those institutional factors examined here. For instance, aspects of labor market lawan important consideration in Brazil could be indexed to supplement information on political risk and other institutional factors. Our results may also prove useful to public administrators in proposed host countries, as well as to administrators and decision-making boards of international development agencies around the world. Given the statistical dominance of institutional theory in our model, government ofcials and politicians should place greater (lesser) emphasis on institutional reform (macroeconomic conditions) in order to attract greater levels of inward FDI. This refocusing of effort would likely exhibit a higher marginal productivity given the greater scope for political/public inuence on a countrys institutional framework rather than its macroeconomic environment (Barro, 1997, 1999). Additionally, our results also indicate that administrators and advisory boards of public agencies that are involved with international development efforts, such as UNCTAD, the International Monetary Fund and The World Bank, should also consider a similar refocusing of effort. Many of these agencies often are preoccupied with assisting LDCs in producing macroeconomic policies that are helpful with economic development. Our

242

L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243 Gujurati, D. N. (1988). Basic econometrics. New York: McGrawHill. Hartenek, G. (1995). Privatization as a catalyst for FDI: The case of Argentina. In Foreign direct investment: OECD countries and dynamic economies of Asia and Latin America. Washington, DC: OECD Publications and Information Center. Hoskisson, R., Eden, L., Lau, C. M., & Wright, M. (2000). Strategy in emerging economies. Academy of Management Journal, 43(3): 249267. Kennedy, P. (1998). A guide to econometrics. Cambridge, MA: The MIT Press. Kennedy, R. E., & Sandler, A. (1997). Transition to a market economy: The components of reform [Case No. 9-797-080, Mimeo]. Harvard Business School. Klein, M., & Rosengren, E. (1994). The real exchange rate and foreign direct investment in the United States: Relative wealth versus relative wage effects. Journal of International Economics, 36(3/4): 373389. Kmenta, J. (1986). Elements of econometrics. New York: Macmillan Publishing Company. Kobrin, S. J. (1979). Political risk: A review and reconsideration. Journal of International Business Studies, 10(1): 6780. Kostova, T., & Zaheer, S. (1999). Organizational legitimacy under conditions of complexity: The case of the multinational enterprise. Academy of Management Review, 24(1): 6481. MacKinnon, J. G. (1983). Model specication tests against nonnested alternatives. Econometric Reviews, 2(1): 85158. Maddala, G. S. (1992). Introduction to econometrics. New York, NY: Macmillan Publishing Company. Meyer, K. (2001). Institutions, transaction costs and entry mode choice in Eastern Europe. Journal of International Business Studies, 32(2): 357367. Mixon, F.G., Jr., & Gibson, M. T. (2001). The retention of state level concealed handgun laws: Empirical evidence from interest group and legislative models. Public Choice, 107(1/2): 120. North, D. (1990). Institutions, institutional change and economic performance. Cambridge: Cambridge University Press. Petrocella, D., & Lousteau, M. (2001). FDI in Argentina during the 1990s. Latin American Business Review, 2(1/2): 3354. Prahalad, C. K., & Hammond, A. (2002). Serving the worlds poor, protably. Harvard Business Review, 80(9): 4857. Ramanathan, R. (1998). Introductory econometrics with applications. New York: Harcourt Brace & Company. Ramirez, M. D. (2001). Foreign direct investment in Mexico and Chile: A critical appraisal. Latin American Business Review, 2(1/2): 5582. Ramsey, J. B. (1969). Tests for specication error in classical linear least squares regression analysis. Journal of the Royal Statistical Society, 31(Series B): 250271. Rojas, (2001). The increased role of capital in the Peruvian nancial system: 19901998. Latin American Business Review, 2(1/2): 101118. Schneider, F., & Frey, B. S. (1985). Economic and political determinants of foreign direct investment. World Development, 13(2): 161175.

results suggest that any assistance provided by either the IMF or the WB should, perhaps, be geared toward institutional reform. Not only does institutional theory appear to dominate macroeconomic theories concerning MNEs decision to locate outward FDI, these public lending enterprises have a vested interest in assuring that the marginal productivity of their own assistance efforts is increased.

Acknowledgments The authors thank two anonymous referees of this journal, Frank Hoy, and Kamal Upadhyaya for helpful comments.

References
Aliber, R. Z. (1970). A theory of foreign direct investment. In C. P. Kindleberger (Ed.), The international corporation. Cambridge, MA: The MIT Press. Baer, W., & Rangel, G. B. (2001). Foreign investment in the age of globalization: The case of Brazil. Latin American Business Review, 2(1/2): 8399. Barro, R. J. (1997). Getting it right: Markets and choices in a free society. Cambridge, MA: The MIT Press. Barro, R. J. (1999). Determinants of economic growth: A crosscountry empirical study. Cambridge, MA: The MIT Press. Birch, M. H., & Halton, G. (2001). Foreign direct investment in Latin America in the 1990s: Old patterns, new trends, and emerging issues. Latin American Business Review, 2(1/2): 1331. Butler, K. C., & Joaquin, D. C. (1998). A note on political risk and the required return on foreign direct investment. Journal of International Business Studies, 29(3): 599608. Davidson, R., & MacKinnon, J. G. (1981). Several tests for model specication in the presence of alternative hypotheses. Econometrica, 49(5): 781793. Davidson, R., & MacKinnon, J. G. (1993). Estimation and inference in econometrics. Oxford: Oxford University Press. Devlin, R., & Cominetti, R. (1994). La Crisis de la Empresa Publica, Las Privatizaciones y la Equidad Social. In Reformas de politica publica: 135 (Vol. 26) (LC/L 832). Santiago, Chile: CEPAL. DiMaggio, P. J., & Powell, W. W. (1991). Introduction. In W. W. Powell & P. J. DiMaggio (Eds.), The new institutionalism in organizational analysis: 138. Chicago: University of Chicago Press. Froot, K. A., & Stein, J. C. (1991). Exchange rates and foreign direct investment: An imperfect capital markets approach. Quarterly Journal of Economics, 106(November): 11911217. Grosse, R., & Trevino, L. J. (1996). Foreign direct investment in the United States: An analysis by country of origin. Journal of International Business Studies, 27(1): 139155.

 L.J. Trevino, F.G. Mixon Jr. / Journal of World Business 39 (2004) 233243 Scott, W. R. (1995). Institutions and organizations. Thousand Oaks, CA: Sage. Tallman, S. B. (1988). Home country political risk and foreign direct investment in the United States. Journal of International Business Studies, 19(2): 219234. Trevino, L. J. (1999). Market reform and foreign direct investment in Central and Eastern Europe: A comparative assessment. In D. K. Despotis & C. Zopounidis (Eds.), Proceedings of Decision Sciences Institute. Athens: New Technologies Publications. Trevino, L. J., Daniels, J. D., & Arbelaez, H. (2002). Market reform and FDI in Latin America: An empirical investigation. Transnational Corporations, 11(1): 2948.

243

Tuman, J. P., & Emmert, C. F. (1999). Explaining Japanese foreign direct investment in Latin America, 19791992. Social Science Quarterly, 80(3): 539541. United Nations Conference on Trade and Development. (1994). World Investment Report 1994: Transnational corporations, employment and the workplace. New York and Geneva: UNCTAD. United Nations Conference on Trade and Development. (1999). World Investment Report 1999: Foreign direct investment and the challenge of development. New York, NY: UNCTAD. Xu, D., & Shenkar, O. (2002). Institutional distance and the multinational enterprise. Academy of Management Review, 27(4): 608618.