The Political and Economic Determinants of Foreign Direct Investment in Latin America: A Brief Comment on "Macroeconomic Deeds, Not Reform Words: The Determinants of Foreign Direct Investment in Latin America," by Alfred P. Montero, LARR, Volume 43, Number 1

In a recent article, Montero (2008) sought to clarify the determinants of foreign direct investment (FDI) in Latin America. Testing a number of competing hypotheses, he found that macroeconomic stability, as measured by the current account, had the most consistent effect on FDI flows across countries in the region. Although Montero was interested in the role of macroeconomic stability, he also explored the impact of governance factors, including human rights and regime type. His results suggest that the effects of rights and regime type are inconsistent. Briefly, in his models that focused on governance and cost-related factors, the coefficient for the terror scale (rights abuse) had a positive and significant effect in two of three trials; in the same trials, the coefficient for Polity IV (regime) was negative and statistically significant, which suggests that politically competitive regimes received less FDI.¹ Nevertheless, when Montero modeled the effects of economic reform, rights abuse and regime type were no longer statistically significant (Montero 2008, Table 1). Given the inconclusive nature of his findings with regard to rights and regime type, and the ongoing controversy in the literature, a brief comment on his article is potentially instructive.

As a result of problems with the research design used in his study, Montero's analysis of human rights and risk factors remains open to question. In particular, Montero employed FDI data (World Bank, n.d.) that pool observations of outward FDI flows from firms based in many different home markets. As I have argued previously (Tuman 2006), pooled FDI data can obscure the effects of political factors in FDI decisions. Within the home country of a multinational firm, a variety of different actors—ranging from unions (particularly those unions embedded in co-determination systems) to consumer groups or the state—may have the ability to influence the FDI strategies that corporate decision makers adopt. The influence of such actors within the home market might cause German and Swedish multinational firms, for example, to weigh labor and political rights in potential recipient (host) countries of FDI in ways that are different from the evaluations of U.S. firms. For this reason, a preferred methodological approach is to use FDI data that disaggregate flows by the home market of the multinational firm.²

The effects of rights abuse and political risk are clearly discernable when controls for the country of origin of FDI flows are introduced to the analysis. In a previous article, Tuman and Emmert (2004) presented a model of U.S. FDI to Latin America that tested many of the same (or similar) macroeconomic and governance-related variables that Montero employed in his models. To examine Montero's claims, I conducted a sensitivity analysis by reestimating the full model of U.S. FDI developed in Tuman and Emmert (2004)³ with the addition of two covariates: current account/gross domestic product (GDP)_t–1 and economic reform.⁴

The results of the sensitivity analysis indicated that the coefficient for the current account was positive and significant (b = .05, z = 2.69, p < .01), but economic reform had no effect (b = –.02, z = –.12, p < .90). Nevertheless, in contrast to Montero's findings, the sensitivity analysis also demonstrated that even when the current account, economic reform, and other macroeconomic influences were included in the U.S. FDI model, the coefficient for rights/liberties abuse_t–1 was positive and statistically significant (b = .10, z = 2.03, p < .05),⁵ while the coefficients for other risk factors remained statistically significant (revolution deaths [per 1,000 population]_t–1, b = –.47, z = –2.37, p < .02; riots_t–1, b = –.45, z = –2.20, p < .03; military coups_t–1, b = .37, z = 1.87, p < .07). In addition, the results called into question Montero's claim that the current account is the only consistent economic predictor. Trade, growth in real GDP per capita, and secondary school enrollment...