- Political Governance and Macroeconomic Variables in Determining Foreign Direct Investment FlowsA Reply to John P. Tuman
I thank Professor John Tuman for the opportunity to pursue questions I raised in my article but could not engage further in the original study (Montero 2008). I will attempt here to address his commentary with an eye toward clarifying how additional testing of the data not only raises questions about Tuman and Emmert's (2004) findings concerning human rights and regime type but also refines my own findings about macro-economic performance and particularly the role of the current account. I hasten to underscore that it is still not my primary purpose to disprove the results of Tuman and Emmert (2004) or to engage the broader literature on regime type and investment flows. However, as I conclude in my article (Montero 2008, 76), the data thus far do not sustain consistently the finding that human rights violations and regime type affect foreign direct investment (FDI) flows in Latin America. Moreover, my study is not meant to challenge Tuman's point about the importance of the domestic political and economic institutions of the home countries of multinational corporations (MNCs). I argue here that exploring these factors requires a different kind of study than the one I initially designed. Yet I also raise doubts about whether such a study can be sustained with the data available and implemented using time-series cross-sectional techniques.
Regarding my study's findings that regime type and human rights violations proved inconsistent predictors of FDI, Tuman responds that a flaw in research design—namely, the pooling of FDI inflows per Latin American country-year—explains the performance of these political variables. He argues that the disaggregation of flows by sending countries would introduce the effects of different domestic institutions and pressures to direct FDI in certain ways in the Latin American region. The examples he gives in his commentary suggest that differences between liberal America and social democratic Europe explain the conditions under which [End Page 195] regime type and human rights matter in the investment decisions of MNCs.1 Yet even when the proposed remedy of disaggregating flows is pursued, the findings for the political variables remain uncertain. In the simplest time-series cross-sectional models in which U.S. FDI and European FDI per country-year are used as the dependent variables, the coefficients for the lagged Polity score and the Gibney terror scale continue to produce inconsistent results. The Polity score is insignificant in every test.2 The Gibney terror scale is significant and negative for U.S. FDI (b = –.2422, p < .05) and for European FDI (b = –.2876, p < .05). This finding suggests that foreign companies are dissuaded from investing in polities that engage in state terror. In the original article, I ran the Polity and Freedom House (FH) scales in separate specifications as they are notoriously collinear. Keeping the Gibney terror scale in the model, the FH tests including the political rights and civil liberties scales offer similarly inconsistent findings for regime type. For neither U.S. FDI nor European FDI are the FH scales significant. Notably, the Gibney terror scale is insignificant for U.S. FDI and significant and, once again, negative for European FDI (b = –.3531, p < .05). European firms wish to avoid human rights violators, but we can come to no conclusions about the role of this factor for American firms.
When control variables are added to the preceding specifications, the story becomes more complex. Because two of the three governance models in the original article showed significant and correct-sign coefficients for Polity and the Gibney terror scale for pooled FDI, it is most appropriate to retest these specifications using disaggregated FDI—specifically, U.S. and European data. The first governance model in the original article was retested in two specifications using U.S. FDI and European FDI as separate dependent variables.3 All of the governance variables were insignificant in these specifications. When tax burden and financial regulations were added, as in the Governance + Cost1 model in the article, and market size was dropped but factor controls such as urban density, manufacturing intensity, mining...