ABSTRACT
The economic impact of FDI is a widely researched topic. In line with the growing interest in the role of externalities in processes of economic growth, the last two decades have seen a rapidly growing body of quantitative research that tries to identify and quantify FDI externalities in a wide variety of host economies. Despite the large amount of evidence on these FDI effects, this research field is facing several important challenges. One challenge concerns the continuing debate on the question how important FDI externalities are, and whether these effects are of a positive or negative nature. Cross-country growth regressions provide little evidence in support of a general positive FDI growth effect, whereas the body of evidence from micro-economics research on FDI externalities in individual host economies is very heterogeneous in nature. Although recent empirical studies have broadened their scope, both by increasing the number of host economies for which FDI spillovers are estimated, as well as by estimating for FDI effects within and between industries, the large degree of heterogeneity of the evidence has not diminished. Another important challenge concerns the question whether there are factors that affect the materialisation and perhaps also the nature of FDI externalities. The most commonly accepted factor is the level of absorptive capacity of domestic firms, which points at the requirement that these firms must possess sufficient knowledge and skills to be able to absorb new technologies from FDI firms. The empirical translation of this concept of absorptive capacity into the technology gap between FDI and domestic firms, which is interpreted as a direct inverse indicator of the level of absorptive capacity of domestic firms, is problematic as it rests on an incomplete interpretation of the underlying catch up thesis. The findings from several studies that show that a large technology gap may be conducive rather than detrimental to positive FDI externalities further indicate the need to address this problem.
