Can Emerging Economies Rely on Foreign Direct Investment for Growth? Dynamic Heterogeneous Panel Analysis of Latin America

Such questions as 'does FDI assist growth? how important is foreign relative to domestic investment? what is the relative importance of FDI and trade openness?' are clearly important, given developing and emerging countries' need to respond to the overtures of investors from abroad. One key characteristic in favor of FDI is its stability relative to the highly volatile capital movements that have characterised financial crises. In contrast to such volatility, FDI has been remarkably persistent across all sectors of the economy and is clearly a candidate for having boosted growth above that which would otherwise have occurred. On the other hand, a pessimistic view might accuse FDI of serving merely to transfer ownership from domestic to foreign agents, so that its effect on growth is, at best, neutral or, at worst, immiserising. Despite the scale and persistence of FDI, and the weighty literature that has emerged in the area of economic growth, only a small number of studies have investigated empirically the relationship between the two. Moreover, there has been relatively little investigation of the effects on individual economies, as opposed to more aggregate groupings. A further issue concerns the methodologies that have been used to investigate the growth effects of FDI. Most studies of economic growth have involved panel data, using period averages, typically of five or ten years. The choice of start and end dates has often been arbitrary, despite the fact that alternative dates can produce considerable differences in the estimated results. This paper will examine the effects of FDI on growth using the case of Latin American countries. The paper will also examine the effects of other variables including trade openness, human capital, terms of trade and international debt. The paper will employ an innovative methodology, pooled mean group (PMG) estimation, that will overcome a number of problems. First, PMG makes full use of all available information in the form of individual observations, rather than period averages for each country, and also allows for short run growth responsiveness to differ between the individual Latin American countries. This dynamic modelling methodology permits the speeds of adjustment to the steady-state to differ between the countries. Second, the paper will use a new approach, termed rolling panels, to provide quantitative estimates of the sensitivity of the results to changes in the period averages stemming from alternative choices of start and end dates. The results demonstrate that the choice of time period does, indeed, matter as the results are far from invariant with respect to many of the time periods examined. FDI, in isolation from human capital, did not have a positive effect on growth in the Latin American countries. However, it did have a positive though small effect when interacted with human capital. Trade openness also had a significant positive effect on growth for some specifications. The results for Latin America show that the economic behavior of regions can differ considerably from the average behavior of wider groups. Hence, the paper demonstrates that it is inappropriate to generalize from the more aggregate level to smaller groups, such as all developing countries compared with different regions, at least in relation to the magnitudes of the effects of different variables on economic growth.