Does foreign direct investment affect growth in MENA countries? A semi-parametric fixed-effects approach

ABSTRACT

It is often asserted with confidence that foreign direct investment (FDI) is beneficial for economic growth, especially in the host developing economy. Nevertheless, there is no empirical consensus on a positive effect of FDI on host-country growth, nor on the direction of causation. One of the reasons behind the lack of consensus is likely the presence of nonlinearities in the FDI and growth relationship. Most of the previous studies either used the linear empirical growth model or tried to bypass the nonlinearity issue by using ad hoc procedures. However, it is also true that growth theory provides little guidance about the exact nature of nonlinearity. Consequently, it is almost impossible to determine the exact form of nonlinear specification that would be appropriate for all data sets and data ranges. The paper investigates this challenging question in empirical growth literature that is the impact of FDI in promoting economic growth in developing economies without adopting any ad hoc procedure to capture the nonlinearity in the FDI–growth relationship. Based on a dualistic growth framework and a partial linear regression approach, it is possible to separate measures for sector externality and factor productivity effects between the two sectors (exports and non-exports sector). Sectoral externality is defined as a function of real FDI stocks per capita. Thereby, the adopted theoretical framework allows capturing both direct and indirect effects of FDI on economic growth across eight MENA countries during the period 1990–2016.

KEYWORDS:

• FDI
• growth of developing countries
• semiparametric and non-parametric methods
• panel data

JEL CLASSIFICATION:

• F21
• O11
• C14

Acknowledgements

I thank the journal editorial team and an anonymous referee for their insightful suggestions and helpful comments on an earlier version of the paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The considered MENA countries are Algeria, Egypt, Jordan, Morocco, and Tunisia.

2 A partial linear model is a model, part of which takes a parametric from, and the remaining part is non-parametric. This type of model assumes little about the shape of the regression function beyond some degree of smoothness. It constitutes an important advantage to deliver estimators and inferences that are less dependent on the assumptions about the functional from.

3 According to Feder’s original model, exporting activities encourage producers to improve their technology and adopt more efficient process management to face foreign competition. Moreover, foreign competition generates a sort of natural selection mechanism among firms, and throughout this mechanism, less efficient firms are constrained to adapt or to leave the market. At the end of this mechanism, the remaining firms are those whose marginal factor productivity is higher.

4 Hausman test favors the random effects over the fixed effects model.