Renewable energy and economic growth in the MENA region: empirical evidence and policy implications

ABSTRACT

We estimate the impact of an increase of installed capacity for electricity generation from renewable sources and from increasing renewable electricity generation on economic growth in the MENA region, using a neoclassical growth function that includes capital, labor and energy use as additional input factors. Our working hypothesis is that there could be a negative impact from renewable electricity on growth, given the high initial investments associated with alternative energy technologies. We could not prove this hypothesis and even found some evidence for a positive relation between renewable electricity and growth, mainly for renewable electricity generation; however, causality remains unclear. The results hold for several robustness checks. We conclude that investing in renewables does not hinder growth in MENA countries.

KEYWORDS:

• Energy
• electricity
• growth
• renewables

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 We included the following countries to our definition of the MENA region: Algeria, Bahrain, Egypt, Iran, Iraq, Israel, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Palestinian Territories (West Bank and Gaza), Qatar, Saudi Arabia, Syria, Tunisia, Turkey, United Arab Emirates, Yemen.

2 The term ‘renewable sources’ refers to biomass, hydro, geothermal, solar, wind, ocean thermal, wave action, and tidal action. This follows the definition of the US Energy Information Administration.

3 A second reason for the low usage of renewables we do not discuss here might be abundant (and cheap) oil and natural gas reserves in many MENA countries.

4 Data on the capital stock is not available in WDI. It is reported in Penn World tables, but only in current PPPs and constant 2011 US$; the necessary calculations to combine the capital stock from World tables with the WDI data on GDP in constant 2010 US$ would lead to inaccuracies. Due to this, we decided to follow the common way in the literature and to proxy the capital stock with gross fixed capital formation (see for example Apergis & Payne, 2012; Apergis et al., 2010; Omri, Daly, & Nguyen, 2015).

5 WDI gives the latter in per capita values, we recalculated it to total energy consumption by multiplying it with total population, also from WDI.

6 As an alternative, we considered Worldbank’s World Governance Indicators, but as they are available only for 16 years in our time period (1998, 2000, 2002 and 2002 and later), this lead to a large drop of observations, and interpolating for the missing years to inaccuracies.

7 Note that including a constant, , in Equation (2) means that there is a time trend in the output if is significant. This is a result of differencing, as a non-time dependent constant would drop out (as is obviously zero), while a time-dependent coefficient remains with factor 1 (as ). This time trend might for example be explained with an increase of total productivity.

8 Both aspects are a consequence that might be transformed to .

9 For example, one could think that the share of FDI inflows and not the change of the share of FDI inflows influences the growth rate.

10 Our consideration holds for the use of generation or capacity in absolute values, and primarily for electricity. For the share of renewables, as used by Ocal and Aslan (2013), a negative causality running from growth to renewables might be supposed: A higher output might result in a higher demand in energy (or vice versa), and if renewables stay constant in absolute values, their share decreases.

11 We decided not to report the results of the Phillips-Perron test as it does not provide us with additional information.

12 To our knowledge, there is no commonly used unit root test for unbalanced panels with Hadri’s H₀.

13 As we have countries for which data is available only from 1995 onwards, we took 1996 for trimming to be consistent in the first difference.

14 For this, we rearranged Equation (2), using the first lag of the explaining variables instead of the present value (i.e. and a second equation where is explaining variable for : as well as a simpler model with just . The consideration behind this is that if a variable has a causal effect on another variable, this effect might persist to the following period.

15 As the labor force is not in the focus of our paper, we do not discuss this result deeper, although it is contrary to economic theory.

16 Several studies exploring decarbonization strategies for the region point out that the use of renewables means that ‘a near complete decarbonization of the power sector can be achieved at moderate costs’ (Haller, Ludig, & Bauer, 2012; see as well for example German Aerospace Center, 2005; Karakosta & Psarras, 2013; Supersberger & Führer, 2011).