ABSTRACT
According to the predominant conclusion from the recent FDI-growth literature, the conflicting evidence concerning the effect of foreign direct investment on economic growth in host countries might be due to the fact that the gains from FDI are conditional on certain domestic factors. Among these factors is the quality of institutions. The aim of this study is therefore to investigate the role of institutional development in FDI-growth nexus in African countries. Using a panel smooth transition regression model with a sample made up of 27 African countries over the period 1990–2017, the results show that FDI promotes economic growth in countries where the level of institutional development is beyond a certain threshold. In countries that fall below the threshold, FDI has either a negative or null effect on economic growth. Specifically, we find that countries should be above the 65% threshold regarding government stability score, 55% for investment profile, 50% for democratic accountability, 45% for law and order, 35% for corruption, and above the 25% threshold for bureaucracy quality score to benefit from FDI in terms of economic growth.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 Host countries’ domestic factors that are believed to condition the effect of FDI on economic growth are the level of financial and institutional development, human capital, trade policy, and macroeconomic stability (see, Alfaro, Citation2014; Kose, Prasad, & Taylor, Citation2011).
2 Economic Community of West African States.
3 Due to data constraints, in the final period, the data is averaged over four-year periods (2014–2017).
4 Author’s calculation using UNCTAD online database.
5 Adams (Citation2009), Agbloyor et al. (Citation2016), and Mensah et al. (Citation2018), among others, find a direct effect of institutions in promoting economic growth in African countries. Thus the institutional quality indicators used in this study are sequentially included in the PSTR model as a threshold variable and also as an explanatory variable for all the tests and estimations to avoid erroneous switching (see, for example, Fouquau et al., Citation2008, p. 291).
6 See Jude and Levieuge (Citation2016) for a similar approach.
7 The coefficients of the control variables which are also regime-dependent are available upon request.