The Threshold Role of FDI Flows in the Energy-Growth Nexus: An Endogenous Growth Perspective.

• Date: 01 September 2023
• Author: Richard, Olayeni Olaolu

1. INTRODUCTION

   As the nexus between energy consumption and economic growth continues to attract the attention of researchers and policy makers, the role of foreign direct investment (FDI) in driving this relationship has gradually been garnering more attention. In most studies, the need to overcome the omitted variable bias necessitates the inclusion of other important variables in the multivariate modelling framework of the energy-economic growth relation. However, since the literature that includes FDI as a regressor in this relationship is still at its infancy, accounting for the role of FDI inflows could shed more light on the nexus between economic growth and energy consumption, as well as on the sudden stop, the capital reversals and flights, and other emerging policy issues.

   The importance of this study derives largely from the precariousness of FDI inflows in the sub-Saharan African (SSA) countries which have experienced both an increase and a decrease in FDI flows. The increased FDI inflows to the SSA sub-region over the last few years, driven by hydrocarbon discoveries, grew 16 percent to $43 billion in 2013 (World Bank, 2014). came to a halt in 2015, and then declined virtually in all facets (World Bank, 2015). The countries face a high and rising rate of population growth which requires high and steady capital inflows. This trend has been a concern for policy makers. First, the SSA countries have recently been experiencing uneven large hydrocarbon discoveries funded by unsteady FDI inflows as explained by the World Bank (2015). The dramatic rises and falls of FDI flows necessitate the need for an assessment of the consequent implications of a possible reversal that may expose the flank of these economies to further shocks. From here comes the importance of a foreign investment threshold: a low FDI threshold may not help much in developing appropriate technologies, whereas a higher threshold may well accentuate the risk of FDI reversals. In either case, the relationship between energy consumption and economic growth is affected.

   Second, one such possible implication for the energy-growth nexus bordering on the threshold effects for these countries forms the core of the present study. Although the literature has paid only a scant attention to the role of the levels of FDI flows in this relationship, particularly for the SSA countries, the question regarding the threshold effects of FDI has not yet been systematically addressed. Yet, the regimes of high and low FDI inflows and their precariousness are not likely to imply the same energy consumption or growth trends or links. It thus stands to reason that the data-generating process governing the relationship particularly for these countries will differ in those countries, depending on whether the FDI inflows are building up quickly or sluggishly. Indeed, the architecture of the data-generating process will change if the FDI flows are on the descent or the ascent. In the related literature, it is often instructive to know how a relationship evolves, and whether it has breached a critical point. Third, these countries with similar characteristics are also under-researched and stand as a good case study for the research in this paper.

   While we focus on the SSA countries, the choice of only eight countries for this study is largely informed by their characteristics as explained before and also by data availability (that is, the selected countries are those having available observations on key variables over the study period).

   Few distinctive features of the present study are noteworthy. (i) Although the paper does not employ a multivariate modelling strategy directly, yet by viewing the data-generating processes at the upper and lower regimes of the FDI flows as being dissimilar, we are able to minimize the omitted variables bias. Besides, the multivariate modelling leads unequivocally to parameter proliferation and, given the sample size of the study, this may not be an ideal approach to follow empirically, especially in a threshold regression. (ii) FDI inflows are modelled as a threshold variable rather than being included as a control variable in a multivariate setting. (iii) We devise a strategy to incorporate the Common Correlated Effects (CCE) estimator of Pesaran (2006) to account for the cross-sectional dependence, (iv) We set up a bootstrap strategy to overcome the finite sample problem associated with the asymptotic critical values and the p-values. (v) We investigate a cross-section of countries in a heterogeneous VAR framework (see Emirmahmutoglu and Kose, 2011) executed along the line of the Toda-Yamamoto (1995) lag-augmented model.

   The remainder of the paper is organized as follows. Section 2 presents a brief literature review to overview the existing state of the literature on the subject matter and the classification of the often-tested hypotheses in the nexus. Section 3 introduces the endogenous growth perspective on FDI flows in the energy-growth nexus. Section 4 presents the methodology used, underlying the threshold lag-augmented approach and the strategy to carrying out a bootstrap simulation for LR-based Granger causality, while Section 5 presents the empirical results, undertaking various extensions to the basic model. Section 6 presents the results relating to the sensitivity analysis and Section 7 concludes.

2. A BRIEF REVIEW OF THE LITERATURE: ENERGY-GROWTH HYPOTHESES AND FDI FLOWS:

   Since the pioneering study of Kraft and Kraft (1978), many papers on the relationship between energy consumption and economic growth have been published, where the distinguishing features are mainly in terms of the methodology and of course the countries investigated. Regarding the nexus between these variables, these studies have shown that four hypotheses usually play themselves out well (see, Tsai, 1994; Halicioglu, 2009; Soytas and Sari, 2009; Ghosh, 2010; Pao and Tsai, 2010; Apergis, 2011) which include: (i) the growth hypothesis, (ii) the conservation hypothesis, (iii) the feedback hypothesis, and (iv) the neutrality hypothesis. The growth hypothesis holds that energy consumption promotes economic growth. Indeed, this hypothesis is often supported theoretically since energy as an input complements other factors of production, and thus increases their productivities and their returns. Many studies including Tsai (1994), Halicioglu (2009), Soytas and Sari (2009), Ghosh (2010), Pao and Tsai (2010) also find empirical evidence that supports the growth hypothesis. On the other hand, rejection of the conservation hypothesis (i.e.. energy consumption does not Granger-causes economic growth) indicates that energy consumption may be reduced without harming economic growth. It therefore suggests that energy can be conserved without reducing total production - in fact, the reduction in energy consumption, the hypothesis argues, should induce congenial environment and possibly lead to economic growth.

   These two hypotheses are thus indicative of the possibility of a threshold point at which growth-enhancing effects of energy consumption will begin to take hold. The other two hypotheses, namely the feedback and the neutrality hypotheses, refer respectively to the instances of bidirectional and no causal effects, respectively.

   The desire to alleviate the problem of omitted variables pointed out by Lutkepohl (1982) has led to multivariate studies proliferating with many intervening variables. This is an important reason for including the FDI flows in driving the relation between energy consumption and economic growth. We here also emphasize the role of thresholds in this relation. Starting with the pioneering study of Tang (2009), who examined the effect of FDI flows on the nexus between energy consumption and economic growth for Malaysia, the literature now includes other studies such as Bento (2011) for Portugal, Kuo, Chang, Chen and Chen (2012) for China, Behket and Othman (2011) also for Malaysia, Alam (2013) in a comparative study for India and Pakistan, and Dalia (2015) for Egypt, using renewable electricity consumption as a proxy for energy consumption.

   These studies have all utilized the multivariate VAR approach to model the relationship between energy consumption and economic growth, while entertaining the FDI flows as the intervening variable. In this paper, we have used FDI as a threshold variable and not a control variable. In addition to these variables, Bekhet and Othman (2011) included the total consumption expenditure and the consumer price index, while Alam (2013) retained the three variables as used in Tang (2009) but proxied energy consumption using the time series on electricity consumption. Kuo et al. (2012) employed a trivariate VEC model involving energy consumption, economic growth and FDI flows. Bento (2011) dealt with the causal relation among the three variables using the ARDL model of Pesaran et al. (2009). Largely, the structural variables used in these studies have remained the same.

   The current paper departs from the existing multivariate causality setting by incorporating FDI flows as the third variable (see Omri and Kahouli, 2014; Pao and Tsai, 2011; Mielnik and Goldemberg, 2002; Lee, 2013; Tang and Tan, 2014).

3. THE ROLE OF FDI FLOWS IN THE ENERGY-GROWTH NEXUS: THE ENDOGENOUS GROWTH PERSPECTIVE

   In this section, we give a theoretical explanation for the connection between economic growth and energy consumption and use the proposition to shed light on the role of FDI flows in the energy-growth relation. Our framework adapts a similar model used by Esfahani and Pesaran (2012) to study the nonlinearity induced by the oil price growth.

   Let the technological progress index A, that depends on the global stock of knowledge [[OMEGA].sub.t] and energy use E, be given by the following Cobb-Douglas function:

   [A.sub.l]=[OMEGA].sup.l-[phi].sub.t]-[E.sup.[phi].sub.t] (1)

   where [phi] is the share parameter. Physical capital...