Energy security through diversification of non-hydro renewable energy sources in developing countries

Abstract

The increasing diversity of renewable energy sources in developing countries is receiving attention in discussions about the future of energy security and climate change. Given the strong relationship between energy demand and economic growth, this paper explores the factors that influence the diversification of non-hydro renewable energy sources across 117 developing countries covering more than 30 years. We contribute to the literature by using a new measure capturing diversification of non-hydro renewable energy sources and explore several estimation techniques in investigating determinants of diversification. Controlling for regional variations, results show that higher per capita income, implementation of policies promoting renewable energy, technological innovations and human capital improvement encourage diversification. In addition, the squared term of income was included to capture nonlinear effects. The results depict a U-shaped kind of relationship suggesting non-monotonic changes in renewable energy diversification in relation to increasing affluence. This implies that greater environmental concern in terms of energy use can be expected as countries developed. Other determinants suggest that high dependence on imported fuels and increasing world market price for crude oil will motivate developing countries to diversify non-hydro renewable energy sources. In contrast, the local abundance of hydropower and the availability of natural resources like oil impede diversification. Finally, we conclude that the progressive integration of renewable energy in developing countries energy mix can be hastened with environmental awareness, relevant policy, and favorable economic conditions.

Keywords

Diversification measure non-hydro renewable energy environmental Kuznets curve energy security climate change

Introduction

For the past century, economic activities heavily relied on fossil fuels, gravely affecting the stability of our climate system. The Intergovernmental Panel for Climate Change (IPCC)¹ reported that it is extremely likely that human influence has been the dominant cause of the recent climate change. Increased energy access on fossil fuels has contributed to the improvement of the lifestyle of billions of people, particularly from developing countries, but it has adversely affected the climate system.²,³

Today, policymakers are confronted with the challenge of stabilizing the climate system without hindering economic growth. Delinking economic growth and carbon emissions is crucial in limiting the adverse effect of climate change.⁴,⁵ In 2016, approximately 1 billion people still did not have access to household energy, particularly electricity.⁶ One of the feasible approaches that can be explored to provide energy access for all is to increase the share of renewable energy in the developing countries’ energy mix. Renewable energy can augment the need for fuel and electricity, boost energy security, support economic growth and reduce carbon emissions.^6–9 In addition, the adoption of renewable energy will offer an opportunity for developing countries to leapfrog to a developed status as a result of harnessing energy from environment-friendly sources before a lock-in fossil fuels occur.^10–12

Though conventional fossil fuels are still the primary source of energy worldwide, renewables are steadily becoming part of the global energy mix and have been gaining attention in the past decade¹³ According to the World Bank,¹⁴ the share of renewable energy consumption to total final energy consumption steadily increased from 16.91% in 2007 to 18.05% in 2015. This renewable energy consumption comes from the traditional biomass and modern renewables such as solar, wind, geothermal, hydropower, wave, and biofuels for transport. The combined share of renewable energy (modern and traditional biomass) to total energy consumption in 2016 is around 18.2%.⁶ Renewables have grown rapidly in recent years, accompanied by sharp reductions in cost for solar photovoltaics and wind power.¹⁵

Renewable energy is crucial in transitioning to a less carbon-intensive economy and a more sustainable energy system.¹⁵ However, despite positive developments, enormous environmental and economic benefits associated with renewable energy,¹⁶,¹⁷ the global renewable energy transition is still progressing far too slowly and unevenly.⁶ Further integration of renewable energy in developing countries requires the transfer of climate-friendly technologies.¹²,¹⁸ Although there are already commercially available renewable technologies, they account for only a limited share of the total energy generation. This could be because of several market, economic, institutional, technical and socio-cultural barriers that hinder developing countries in moving away from conventional energy sources.¹⁹,²⁰ Correspondingly, several government policies like feed-in tariffs, tax credits, tradable certificates, investment incentives and production quotas play an essential role in promoting the diffusion of renewable energy.²¹,²²

The diversification and optimal use of renewable energy will help reduce dependence on the importation of oil and reduce the adverse effects of climate change.⁴ To aid policy decisions in accelerating renewable energy diffusion, we investigate the issue of diversification. According to Brunnschweiler,²³ achieving a diversified and sustainable energy supply for the future is one of the major challenges for today's policymakers. Diversification of renewable energy sources has direct implications for energy security. When a county's energy system relies heavily on one source, it becomes vulnerable to energy shocks. The diversification of energy sources is an essential strategy for ensuring energy security.

Our study improves on the recent literature and takes a different approach by investigating what motivates developing countries to diversify non-hydro renewable energy sources. The main contribution of our paper to the literature is threefold. First, we use a new variable in capturing diversification. The usual practice in the literature is to aggregate the energy produced from different sources (for example, Brunnschweiler,²³ Pfeiffer and Mulder,¹² Zhao et al.²²) or by counting the number of renewable energy sources²⁴ to capture diffusion of renewable energy in developing countries. We capture diversification by creating a novel variable or measure that will put weight on each source by using the share of energy produced from a particular source to the total renewable energy generated. The index takes a value of zero if a country has not invested in any non-hydro renewable technology and takes a positive value if the energy generated is coming from one or more sources of renewable energy. To the best of our knowledge, this is the first paper that uses this approach in measuring diversification. Second, we focus only on modern non-hydro renewable energy and exclude hydropower because hydropower is a mature technology.¹²,¹⁸ In the last five years, renewable energy growth mostly comes from non-hydro sources such as solar, wind, geothermal, biomass, and waste.¹⁵ Third, we use a substantially larger data set covering more than one hundred developing countries spanning from 1980 to 2011. The empirical approach uses innovative estimation methods such as the two-part model and Poisson pseudo-maximum likelihood to account for a large number of samples with zero observations.

Diversification of renewable energy for energy security

In business, it is crucial to have a diversified portfolio to help manage risks and volatility of investments. The saying “do not put all your eggs in one basket” is a common expression encouraging diversification. When a country's energy supply relies heavily on one particular energy source, it becomes extremely vulnerable to exogenous supply shock.²⁵ The energy crisis in the 1970s had tremendous economic, political and social consequences not just in developed countries but also in developing countries. Since then, policymakers have paid increasing attention to energy security.

Li²⁶ argued that diversification and localization of energy sources are essential for future energy system because it promotes sustainable development as well as energy security. Li²⁶ stressed that the idea of diversified energy is useful not just for the people but also for the environment. He cited several analogies in other fields pointing out the advantage of diversification. For example, biodiversity is an excellent strategy to prevent the spread of pests and diseases, a diversified portfolio will guarantee a better investment return and in governance, the success of democracy has a diversification of ideas at its core. Similarly, for renewable energy, diversifying its sources is viewed as an attractive option as it can help stabilize the energy supply.

The main disadvantage of renewable energy, aside from its enormous capital cost, is the reliability of energy supply. Most renewable energy relies on the weather as its primary source. For example, hydropower needs rain to fill dams and keep the supply of water flowing; wind turbines need wind to turn blades; solar panels need clear skies and sunshine to generate electricity. These natural sources are somehow variable, unpredictable, and inconsistent.²⁷ These non-hydro sources cannot be simultaneously dispatched as they involve climatic factors like wind velocity and solar irradiation.²⁸,²⁹ When these sources are unavailable, the supply of energy will be affected. However, diversifying these renewable energy sources will allow for a steady and reliable energy supply. For example, dry and sunny weather may not be suitable for generating hydropower but will be great for generating electricity from solar panels; stormy weather may reduce the generation of solar energy but will be good for hydropower, wind energy, and tidal energy. Hence, diversifying renewable energy sources is essential to achieve energy security through a consistent and localized supply of energy.²⁶

Methodology

Diversity measure

One of the innovations of this study is to come up with a new indicator of diversification. To do this, we use the share of renewable energy generated from each source to the total energy generated from all non-hydro sources. The diversity measure takes a value of zero if a country has not invested in any non-hydro renewable technology and takes a positive value if a country has invested in any of the non-hydro renewable technologies. Technically, the diversity measure takes a value of 1 if the energy generated is coming only from one particular source and if the country is more diversified, then the measure is greater than 1. If each source carries the same share of energy generated, the measure converges to the total number of non-hydro sources adopted by developing countries. We propose to estimate the diversity measure as follows

diverse = \{\begin{array}{l} \frac{1}{\sum_{i = 1}^{n} {(\frac{{nhre}_{j t}}{NHRE})}^{2}} i f NHRE > 0 \\ 0 i f NHRE = 0 \end{array}

(1)

where diverse captures the diversity of a developing country’s non-hydro renewable energy mix, NHRE is the net generation of non-hydro electricity measured in billion kilowatt-hours and the nhre_jt is the individual net energy generation from either of the non-hydro renewable energy sources (wind, solar, geothermal, biomass and waste) in country j in year t.

Pfeiffer and Mulder¹² measure the variety of renewable energy technologies by calculating the energy mix and uses this variable as a measure of diversification. Our proposed calculation differs from Pfeiffer and Mulder¹² because their index of diversification takes a value of zero if a country produces energy from only one source and converges to one when a country is more diversified. We could have opted for this measure, but this approach cannot accommodate zero observations or those developing countries at a particular time that did not invest in non-hydro renewable energy. Figure 1 below shows that our data set has a substantial number of zero observations. Using Pfeiffer and Mulder's¹² index will lead to a lot of missing observations because it will not compute for zero observation. The proposed measure of diversification handles this problem. For example, there are four countries named A, B, C, and D. Countries A and B produce energy from wind and solar. Country C produces energy from wind only while Country C has not invested in any non-hydro renewable energy. Assuming countries A, B, and C produce 10 gigawatts (GW) total renewable energy. Country A produces 5 GW from solar and 5 GW from wind energy, country B produces 7 GW from solar and only 3 GW from wind and country C produces 10 GW from wind. Applying our proposed formula for diversification, country A's diversity measure is 2, country B is 1.72, country C is 1 and country D is zero. This implies that if there is an equitable generation of energy from various renewable sources, the diversity value converges to the total count of renewable energy sources.

Figure 1.

The number of non-hydro sources of renewable energy adopted by developing countries.

Factors affecting diversification

The main dependent variable captures diversification in non-hydro renewable energy. Notably, our dependent variable contains substantial number of zero observations or those countries that did not invest in any non-hydro renewable energy (Figure 1). Following Seriño,³⁰ we specify the regression analysis as follows

{diverse}_{i t} = α + β_{1} {GDP}_{it} + β_{2} {GDPsq}_{it} + β_{3} {energ_imp}_{it} + β_{4} {popgrow}_{it} + β_{5} {patents}_{it} + β_{6} {fin_dev}_{it} + β_{7} {secondary}_{it} + β_{8} {fdi}_{it} + β_{9} {oda}_{it} + β_{10} {oilprice}_{it} + β_{11} {renpolicy}_{it} + β_{12} {hydenerg}_{it} + β_{13} {oil_prod}_{it} + β_{14} {coal_prod}_{it} + β_{15} {coastal}_{i} + u_{i} + v_{t} + e_{it}

(2)

where diverse_it is the measure of diversification for country i at year t, GDP_it is the gross domestic product capturing income, GDPsq_it is the squared of GDP to incorporate the nonlinear effect of income. The rest of the covariates is summarized in Table 1. In addition, u_i is the country fixed effects used to capture time-invariant country heterogeneity, v_t is time fixed effects and ε_it is the remaining error.

Table 1.

Data descriptions and sources.

Variable names	Definition	Source
Dependent variable
diverse_it	Diversity measure. It takes a value of 1 if a country adopts only 1 renewable source and a higher value if more diversified. The measure is weighted on the share of each renewable energy source. If a country did not invest in renewable energy, the value is zero	Own computation
Explanatory variables
GDP_it	GDP per capita in constant 2005 USD (in logarithmic form)	World Development Indicator (WDI)
GDPsq_it	GDP per capita squared capturing the nonlinear effect of income
energ_imp_it	net energy imports in % of energy use	WDI
popgrow_it	annual population growth in %	WDI
patents_it	total patent application taken in log form	WDI
fin_dev_it	financial development measured as domestic credit to private sector in % of GDP	WDI
secondary_it	secondary school enrollment % gross	WDI
fdi_it	foreign direct investment, net inflows in % of GDP	WDI
oda_it	net official development assistance received in % of GNI	WDI
oilprice_it	crude oil prices (West Texas intermediate)	BP
renpolicy_it	dummy variable taking value 1 from the year of implementation of a renewable energy policy	IEA
hydenerg_it	total hydroelectric power generated (in thousand kilowatthours) per capita	IEA
oil_prod_it	total oil production in thousands barrel/1 million people	Energy Information Administration (EIA)
coal_prod_it	total coal production in thousand tons/1 million people	EIA
coastal_i	dummy variable taking value 1 if a country has a coastal area	Google Maps⁵²

The dataset is compiled using four different sources: International Energy Agency,²¹ World Bank’s World Development Indicators (WDI),³¹ Energy Information Administration),³² and BP Statistical Review of World Energy³³ (Table 1). Our main source of data for the generation of renewable energy comes from IEA. Data from IEA on non-hydro power generation can be considered comprehensive; however, it may have underestimated the electricity generated as an off-grid generation may not be included in the data set.¹² Table 2 presents the descriptive statistics of the variables used in the analysis.

Table 2.

Descriptive statistics.

Variable	n	Mean	SD	Min	Max
Dependent variables
Diversity index	3946	0.37	0.31	0	2.89
Explanatory variables
GDP per capita (in log)	4090	7.17	1.07	3.91	9.60
Energy import	2688	−41.92	195.54	−1982.88	99.96
Population growth	4727	1.81	1.35	−10.96	11.18
Patents	1473	6.25	1.93	0.69	13.17
Financial development	3708	3.01	0.87	−0.58	5.12
Secondary enrolment	2944	54.79	30.03	2.48	122.20
FDI	3849	3.69	11.79	−82.93	366.36
ODA	3654	9.29	12.95	−2.70	242.29
Renewable policy	4752	0.10	0.30	0	1
Hydro energy	4070	0.37	0.89	0	10.08
Oil production	3929	4281.3	14,337.73	0	216,634.1
Coal production	4144	289,337.7	1,012,096	0	1.25e + 07
Crude oil price	4752	37.66	25.00	14.39	100.06
Coastal	4752	0.78	0.42	0	1

Source: Author’s calculations based on the data described in Table 1.

Empirical approach

A two-stage estimation technique is used to model the choice of whether to adopt or not adopt non-hydro sources of renewable energy. And then, conditional on adopting, we examine the factors influencing the diversification of non-hydro renewable energy sources. Take note, there is a large number of zero values in the dependent variable (Figure 1). However, we assume that these zero observations are true zero or actual outcomes. These zero observations reflect countries that did not invest in non-hydro renewable energy sources.^a The two-part model is a methodological strategy pioneered by Duan et al.³⁵ designed to deal with a large percentage of zero value observations. The two-part model decomposes an observed random variable into two observed variables. In the first stage, the methodology consists of estimating the observation with zero values and positive outcomes and in the second stage, it will estimate the subset with positive outcomes. We explicitly assumed that the decision to adopt a non-hydro renewable source and whether to diversify it or invest in different types of non-hydro resources are independent of each other.

The two-part model (2PM) consists of two equations; the first equation estimates the entire sample and the second equation focuses on the subset with positive outcomes. The positive outcomes reflect the diversity index. We specify the first stage equation as follows

\Pr ({diverse}_{i t}^{*} = 1 | Z_{i t}) = \Pr [{dre}_{i t} > 0 | Z_{i t}] = Ω (β Z_{i t} + ε_{i t})

(3)

where diverse*_it takes a value of 1 if the measure of diversification is greater than zero and zero otherwise. Z_it is the vector of control variables, β is the associated coefficient, Ω the standard normal distribution and ε_it is the remaining error term. In the second equation, we use the diversity measure defined in equation (1) as the dependent variable. Diversifying sources of non-hydro renewable resources on the condition that a country invests on non-hydro renewable energy is defined as

{diverse}_{i t} = β Z_{i t} + ω_{i t}

(4)

where diverse_it captures how diversified is a particular country i at a given time t, Z_it captures explanatory variables and ω_it is the remaining error. We estimate equations (3) and (4) using the method of logit and ordinary least squares, respectively.

For robustness check, the methods of Poisson pseudo-maximum likelihood (PPML) estimation technique is used. PPML is suitable for modeling a large proportion of zero observations. Santos Silva and Tenreyro³⁶ did pioneering work on Poisson pseudo-maximum likelihood (PPML) estimation technique. PPML gives consistent estimates in the presence of large zero observations. This was based on the simulation conducted by Santos Silva and Tenreyro.³⁶ Besides, PPML also effectively handles heteroskedasticity by using a robust covariance matrix.

Results and discussions

Table 3 presents the baseline results from different specifications using the new measure of diversification while controlling for time and regional fixed effects. Results show the marginal effects of the two-part model estimation. While the effect of income on renewable energy has been well documented in the literature, none of the studies explored the nonlinear effect of income.⁹,¹²,²²,^37–39 Results show that while income plays a significant positive role in explaining the diversification, the nonlinearity of its effect is strongly evident. Higher income makes developing countries more capable of diversifying non-hydro sources but the effect is non-monotonous. Our results showed evidence of the U-shaped kind of relationship between income and diversification. This implies that as developing countries grow, diversification of renewable energy sources tends to decline, but diversification increases as their economies become more affluent. Results suggest that as countries develop, demand for renewable energy is substituted by conventional energy sources to fuel a growing economy but as economies mature, it will address environmental concerns through the use of cleaner sources of energy.

Table 3.

Average marginal effects from two-part model estimation (ATPM) with diversity measure as the dependent variable.

Variables	(1)	(2)	(3)	(4)

Log GDP per capita	0.059**	0.103***	−1.745***	−1.856***
	(0.0290)	(0.0327)	(0.3751)	(0.6467)
Log GDP per capita_sq			0.123***	0.126***
			(0.0249)	(0.0425)
Energy import	0.003***	0.0001	0.0001	0.001***
	(0.0002)	(0.0003)	(0.0003)	(0.0003)
Population growth	0.114***	0.124***	0.116***	0.175**
	(0.0257)	(0.0246)	(0.0243)	(0.0376)
Oil price	0.009***	0.012***	0.012***	0.009***
	(0.0022)	(0.0027)	(0.0028)	(0.0032)
Policy on renewables	0.185***	0.080*	0.049	0.072
	(0.0440)	(0.0474)	(0.0480)	(0.0453)
No. of patents	0.084***	0.105***	0.114***	0.132
	(0.0098)	(0.0110)	(0.0112)	(0.0142)
Financial development	0.004	0.017	0.001	−0.095***
	(0.0251)	(0.0272)	(0.0262)	(0.0329)
Hydro energy		−0.104***	−0.113***	−0.295***
		(0.02184)	(0.0234)	(0.0409)
Oil production		−0.021***	−0.025***	−0.031***
		(0.0035)	(0.0036)	(0.0039)
Coal production		−0.021	−0.006	0.078***
		(0.0147)	(0.0148)	(0.0158)
Secondary enrolment				0.007***
				(0.002)
FDI				−0.015***
				(0.0056)
ODA				−0.012
				(0.0072)
Coastal dummy	0.245***	0.150***	0.117**	0.194**
	(0.0494)	(0.0534)	(0.0532)	(0.0862)
Regional dummies	Yes	Yes	Yes	Yes
Time dummies	Yes	Yes	Yes	Yes
Observations	1216	1091	1091	732

Robust standard errors in parentheses.

*p < 0.1, **p < 0.05, ***p < 0.01.

The nonlinear effect of income on the diversification of renewable energy presents additional evidence on the relevance of the environmental Kuznets curve (EKC). The EKC hypothesis that ecological problems will worsen for a given society until reaching its peak and then further increase in income will translate to a reduction in environmental issues as society takes initiatives in cleaning the environment.^40–42 Table 3 shows that in models 3 and 4 the log of GDP per capita is negative but the squared term is positive. The sign of the coefficients reflects the presence of environmental Kuznets curve, suggesting that at the initial stage of development, countries choose to fuel economic activities with unclean energy but as the countries accumulate more wealth, then they will take care of the environment by using cleaner sources of energy.

The other determinants such as population growth, energy imports and oil price are positively correlated with diversification (Table 3). A growing population means rising energy demand and this can motivate renewable energy investment. Abban and Hasan⁴³ reported that population size has a positive effect on renewable energy investment. However, Hao and Shao⁹ argued that less populous countries may have more potential to promote renewable energy. In terms of the world price for crude oil, results show that increasing price is positively correlated with the diversification of non-hydro renewable energy sources. Countries that are more dependent on foreign energy sources are more likely to diversify non-hydro renewable energy. Developing countries are sensitive to price increases in fuel because this can influence higher inflation and hamper economic growth. Renewable energy presents an attractive option in dealing with energy security; hence, diversifying sources of renewable energy is a feasible action for dealing with fluctuations in oil prices. While Marques et al.³⁸ and Aguirre and Ibikunle³⁹ suggest that energy prices are not relevant factors in explaining the diffusion of renewable energy, we argued the opposite. Our results are consistent with Chang et al.⁴⁴ and Lin and Omoju⁷ that an increase in fossil fuel price is positively correlated with renewable energy use.

The associated relationships of policy variable, number of patents for technological advancements and improved human capital through secondary enrollment are expected (Table 3). Implementing policies that facilitate diffusion of renewable energy is positively correlated with diversification. This is consistent with Zhao and Luo⁴⁵ indicating that some government policies have a positive impact on promoting renewable energy. However, Zhao et al.²² and Romano et al.⁴⁶ warn about policy crowdedness, wherein the effectiveness of policies diminishes as more renewable energy policies are implemented. Technological improvements can facilitate diversification. This is quite plausible because modern non-hydro renewable energy is technology dependent. This implies that adopting a variety of renewable energy requires a certain grasp of the related technology. This result is complemented by the evidence showing that the accumulation of human capital, as measured by secondary enrollment, positively contributes to the diversification. Our results aligned well with the previous findings on the influence of policy,²²,⁴⁷ technology,¹⁸ and human capital¹² in renewable energy. However, the effect of financial development on diversification is not robust. Given that renewable energy requires a high level of financing, we would have expected conclusive results. Painuly and Wohlgemuth⁴⁸ noted that the absence of well-developed financial intermediaries and the consequent financing difficulties impede the renewable energy development in developing countries. The paper of Pfeiffer and Mulder¹² claimed that financial development has no influence on renewable energy diffusion in developing countries, but Brunnschweiler²³ and Freitas et al.⁴⁹ argued the opposite. Further studies can be done to clarify this issue.

Developing countries that are already generating renewable energy from hydropower do not show evidence of diversification. This suggests that developing countries are reluctant to invest in non-hydro sources if they have hydropower. This reflects the claim that countries with relatively low carbon intensity in their energy system are likely to diminish incentives in investing in other renewables.¹²,¹⁸ Similarly, the abundance of natural resources like oil decreases the attractiveness of investing in other energy sources. According to Popp et al.,¹⁸ local oil production reduces energy security concerns, increases the relative price of other renewable energy, and undermines support for reducing emissions, making the investment in renewable energy unattractive. However, results show the effect of coal production is inconclusive.

For external sources of funding such as foreign direct investment (FDI) and official development assistance (ODA), the associated correlation with diversification is negative. While these two external sources of funding are essential for technology transfers, neither of them facilitates the diversification of non-hydro renewable energy sources. The same findings were reported by Pfeiffer and Mulder.¹² This result adds to the vast literature casting doubt on the effectiveness of aid in developing countries. Seriño²⁴ presented some possible explanations citing that the ODA variable is aggregated value and does not capture the energy-specific projects. Similarly, FDI does not promote diversification of non-hydro renewable energy sources in developing countries because most FDI projects are related to conventional fuel sources.

Aside from regional fixed effects, we also control for geographic locations of countries, taking a value of 1 if a country has a coastline and 0 if land-locked. This controls some countries' geographic advantage in harnessing renewable energy and captures the ease of access to trade. Results show that countries with coastal areas are more likely to diversify renewable energy sources as compared to landlocked countries. To control for variations in renewable energy associated with time, we also included year dummies.

Robustness of analysis

We check the robustness of our results by pooling all the observations and run a regression using the same specifications. Since many developing countries have yet to invest in renewable energy, our dependent variable, the diversity measure, has a large proportion of zero observations. To address this problem, we follow the suggestion of Zhao et al.²² to use the Poisson pseudo-maximum likelihood (PPML) estimation technique. In addition, PPML effectively handles heteroskedasticity by using a robust covariance matrix.^b The PPML approach gives consistent estimates regardless of how the data is distributed.²²,³⁶

Table 4 presents the estimation results using Poisson pseudo-maximum likelihood estimation. The results of PPML generally complement our previous findings. Higher income is associated with diversification and the nonlinear effect of income still holds, suggesting the robustness of the relevance of environmental Kuznets curve in developing countries. Higher dependence on external sources of energy as proxied by energy imports contributes positively to the diversification of non-hydro renewable energy sources. Higher crude oil prices and a growing population is associated with increasing diversification. Diversification of non-hydro sources of renewable energy is further supported by the adoption of renewable energy policies, technological innovation as proxied by the number of patents, and improvement in human capital. In contrast, oil production and abundance of hydropower are negatively associated with diversification. As with our previous results, FDI and ODA do not support diversification.

Table 4.

Estimation results using the Poisson pseudo-maximum likelihood estimation with total energy generated from non-hydro sources as the dependent variable.

Variables	(5)	(6)

Log GDP per capita	0.102**	−2.077***
	(0.0451)	(0.4686)
Log GDP per capita_sq		0.133***
		(0.0305)
Energy import	0.0001	0.001*
	(0.0003)	(0.0007)
Population growth	0.076*	0.175***
	(0.0413)	(0.0439)
Oil price	0.015***	0.013***
	(0.0033)	(0.0036)
Policy on renewables	0.097	0.113
	(0.0652)	(0.0741)
No. of patents	0.066***	0.041***
	(0.0094)	(0.0128)
Financial development	0.291***	0.254***
	(0.0430)	(0.0551)
Hydro energy	−0.330***	−0.724***
	(0.0684)	(0.0635)
Oil production	−0.004	−0.022**
	(0.0038)	(0.0106)
Coal production	−0.017	0.091**
	(0.0235)	(0.0427)
Secondary enrolment		0.011***
		(0.0024)
FDI		−0.047**
		(0.0207)
ODA		−0.020*
		(0.0105)
Coastal dummy	0.305***	0.897***
	(0.0911)	(0.1560)
Constant	−4.749***	3.397*
	(0.4441)	(1.8648)
Regional dummies	Yes	Yes
Time dummies	Yes	Yes
Observation	1908	1242
R ²	0.393	0.521

Standard errors in parentheses.

*p < 0.1, **p < 0.05, ***p < 0.01.

Conclusion

In this paper, we investigate what motivates developing countries to diversify modern renewable energy sources despite its huge upfront investment cost. In contrast to most studies in the literature that use the amount of energy generated from renewable energy, we develop a new measure capturing diversification using the quantity of energy generated from each source as weights. To the best of our knowledge, this is the first paper that uses this kind of measure for diversification. The analysis is focused only on the modern sources of renewable energy such as solar, wind, geothermal, biomass, and waste, excluding hydropower.

The proposed measure of diversification is comparable with the conventional method of pooling all sources of renewable energy. The results of the robustness check displayed similar findings using the new indicator. This suggests that the proposed new measure using the share of individual sources to the total non-hydro power generation can be a potential measure of renewable energy diversification. In addition, compared with the other indicators of diversification, our proposed measure of diversification can handle zero observations. The zero observations are treated as actual observations and not missing observations. These zero values capture developing countries that are yet to invest in non-hydro sources of renewable energy.

The regression results consistently show that income has a non-monotonous effect on diversification. While most available studies in the literature do not include the nonlinear effect of income, our results show robust evidence that income has a significant nonlinear effect portraying a U-shaped kind of relationship with non-hydro renewable energy. This implies that as society develops, demand for renewable energy is declining because it is replaced by conventional energy to fuel a growing economy, but as the economy accumulates more wealth, it will address environmental concerns through various sources of clean and renewable energy. Results show the relevance of the environmental Kuznets curve, suggesting that greater environmental awareness comes with increasing affluence.

Results show that countries dependent on foreign energy sources are more likely to diversify non-hydro renewable energy sources. Energy security is a major policy agenda in developing countries. Diversification is one of the feasible approaches to localize energy supply and improve energy security. This result is complemented by the effect of the higher world price for crude oil. Developing countries are sensitive to price increases in fuel because it can trigger higher inflation and hamper economic growth. Renewable energy presents an attractive option in dealing with energy security and diversifying its sources can help minimize the risks related to fluctuations in world market prices for oil.

Results show that the coefficient of policy variable, patents and secondary enrolment is as expected. This means that the implementation of renewable energy policies, technological innovation and accumulation of human capital positively contributes to the diversification of non-hydro sources of renewable energy in developing countries. Several policies provide incentives for promoting renewable energy; however, Zhao et al.²² cautioned about the potential crowding-out effect of these several policies. The significant positive coefficient of patents suggests that the advances in technology facilitate diversification of non-hydro renewable energy. Similarly, improvement in human capital does promote diversification.

Our empirical investigations identified drivers and barriers to the diversification of non-hydro renewable energy sources. Results present various dimensions where policymakers can promote renewable energy. Although there is a rising trend of renewable energy adoption in developing countries, it still has a long way to go in integrating renewables into the energy system. According to REN21,⁶ the share of modern renewable energy (excluding traditional biomass) to total final energy consumption is only around 10.4% in 2016 a little higher from its 2015 levels. This shows that there is progress in integrating renewable energy into the countries' energy systems, albeit slow. Given the strong link between economic development and energy consumption, developing countries' future growth will heavily rely on massive energy use.⁵⁰,⁵¹ In the last decade, we see continued economic growth from developing countries and this could be their opportunity to transform their economy and integrate more diversified sources of cleaner energy. Our empirical analysis shows that one of the primary drivers behind rising diversification is a greater environmental concern. Using cleaner sources of energy, developing countries can transition to a low-carbon economy while ensuring energy security and sustainability.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Moises Neil V Seriño

Notes

Author Biographies

Moises Neil V. Seriño is an associate professor of the Department of Economics, Visayas State University, Visca, Baybay City, Leyte, Philippines. He earned his doctorate degree in economics at the University of Göttingen, Germany in 2014. He was a visiting research fellow at the Center for Southeast Asian Studies, Kyoto University in 2019. He is currently the director of the Visayas Socio-Economic Research and Data Analytics Center and the dean of the College of Management and Economics, Visayas State University. His research interests include development economics, climate change, agricultural economics, environmental conservation and sustainable development.

References

Intergovernmental Panel for Climate Change (IPCC). Climate Change 2014: synthesis report. Contribution of working groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (RK Pachauri, LA Meyer and Core Writing Team (eds). IPCC, Geneva, Switzerland, 151 pp.

Abdelkareem

Elsaid

Wilberforce

, et al. Environmental aspects of fuel cells: a review. Sci Total Environ 2021; 752: 141803.

Elsaid

Kamil

Sayed

, et al. Environmental impact of desalination technologies: a review. Sci Total Environ 2020; 748: 141528.

Reddy

Bijwe

PR.

Day-ahead and real time optimal power flow considering renewable energy resources. Int J Electr Power Energy Syst 2016; 82: 400–408.

Seriño

MNV.

Is decoupling possible? Association between affluence and household carbon emissions in the Philippines. Asian Econ J 2017; 31: 165–185.

REN 21:2018. Renewables 2018 global status report, http://www.ren21.net/status-of-renewables/global-status-report/ (accessed 21 January 2019).

Lin

Omoju

OE.

Focusing on the right targets: Economic factors driving non-hydro renewable energy transition. Renewable Energy 2017; 113: 52–63.

REN21:2020. Renewables 2020, global status report, https://www.ren21.net/wp-content/uploads/2019/05/gsr_2020_full_report_en.pdf (accessed 18 March 2021).

Hao

Shao

What really drives the deployment of renewable energy? A global assessment of 118 countries. Energy Res Soc Sci 2021; 72: 101880.

10.

Popp

International technology transfer, climate change, and the clean development mechanism. Rev Environ Econ Policy 2011; 5: 131–152.

11.

Watson

Sauter

Sustainable innovation through leapfrogging: a review of the evidence. IJTG 2011; 5: 170.

12.

Pfeiffer

Mulder

Explaining the diffusion of renewable energy technology in developing countries. Energy Econ 2013; 40: 285–296.

13.

International Renewable Energy Agency. IRENA's renewable energy trends graph, https://public.tableau.com/shared/TZWM5HDDM. ?:toolbar=n&:display_count=n&:origin=viz_share_link&:embed=y (2020, accessed 19 March 2021).

14.

World Bank. Renewable energy consumption, https://data.worldbank.org/indicator/EG.FEC.RNEW.ZS (2018, accessed 8 December 2018).

15.

International Energy Agency (IEA). Renewables 2018, 2018, https://www.iea.org/topics/renewables/ (accessed 21 February 2019).

16.

Abdelkareem

El Haj Assad

Sayed

, et al. Recent progress in the use of renewable energy sources to power water desalination plants. Desalination 2018; 435: 97–113.

17.

Sayed

Wilberforce

Elsaid

, et al. A critical review on environmental impacts of renewable energy systems and mitigation strategies: wind, hydro, biomass and geothermal. Sci Total Environ 2021; 766: 144505.

18.

Popp

Hascic

Medhi

Technology and the diffusion of renewable energy. Energy Econ 2011; 33: 648–662.

19.

Verbruggen

Fischedick

Moomaw

, et al. Renewable energy costs, potentials, barriers: conceptual issues. Energy Policy 2010; 38: 850–861.

20.

Dulal

Shah

Sapkota

, et al. Renewable energy diffusion in Asia: can it happen without government support? Energy Policy 2013; 59: 301–311.

21.

International Energy Agency (IEA). IEA – renewable energy. http://www.iea.org/policiesandmeasures/renewableenergy/ (2014, accessed 29 April 2014).

22.

Zhao

Tang

Wang

Do renewable electricity policies promote renewable electricity generation? Evidence from panel data. Energy Policy 2013; 62: 887–897.

23.

Brunnschweiler

CN.

Finance for renewable energy: an empirical analysis of developing and transition economies. Envir Dev Econ 2010; 15: 241–274.

24.

Seriño

MNV.

Diversification of non-hydro renewable energy sources in developing countries. Energ Ecol Environ 2018; 3: 317–329.

25.

Van Hove

Diversification of primary energy consumption in six west European countries: quantification and analysis by means of measures of concentration. Energy Econ 1993; 15: 239–244.

26.

Diversification and localization of energy systems for sustainable development and energy security. Energy Policy 2005; 33: 2237–2243.

27.

Reddy

Bijwe

Abhyankar

AR.

Real-time economic dispatch considering renewable power generation variability and uncertainty over scheduling period. IEEE Syst J 2015; 9: 1440–1451.

28.

Reddy

Momoh

JA.

Realistic and transparent optimum scheduling strategy for hybrid power system. IEEE Trans Smart Grid 2015; 6: 3114–3125.

29.

Reddy

SS.

Optimal scheduling of thermal-wind-solar power system with storage renew. Energy 2017; 101: 1357–1368.

30.

Seriño

MNV.

What motivates developing countries to diversify sources of renewable energy?

Japan-ASEAN Transdisciplinary Studies Working Paper Series (TDWPS) 2019; 6: 1–24.

31.

World Bank. World development indicator, https://data.worldbank.org/products/wdi (2014, accessed 18 November 2014).

32.

Energy Information Administration. Data tools, apps and maps, https://www.eia.gov/tools/ (2014, accessed 25 November 2014).

33.

BP Statistical Review of World Energy. Energy economics, https://www.bp.com/en/global/corporate/energy-economics.html (2014, accessed 8 October 2014).

34.

Dow

Norton

EC.

Choosing between and interpreting the heckit and two-part models for corner solutions. Health Serv Outcomes Res Methodol 2003; 4: 5–18.

35.

Duan

Manning

Morris

, et al. A comparison of alternative models for the demand for medical care. J Bus Econ Stat 1983; 1: 115–126.

36.

Santos Silva

JMC

Tenreyro

On the existence of the maximum likelihood estimates in Poisson regression. Econ Lett 2010; 107: 310–312.

37.

Menz

Vachon

The effectiveness of different policy regimes for promoting wind power: experiences from the states. Energy Policy 2006; 34: 1786–1796.

38.

Marques

Fuinhas

Pires Manso

JR.

Motivations driving renewable energy in European countries: a panel data approach. Energy Policy 2010; 38: 6877–6885.

39.

Aguirre

Ibikunle

Determinants of renewable energy growth: a global sample analysis. Energy Policy 2014; 69: 374–384.

40.

Stern

DI.

The rise and fall of the environmental Kuznets curve. World Dev 2004; 32: 1419–1439.

41.

Seriño

MNV

Klasen

Estimation and determinants of the Philippines' household carbon footprint. Dev Econ 2015; 53: 44–62.

42.

Irfany

Klasen

Affluence and emission tradeoffs: evidence from Indonesian households' carbon footprint. Environ Dev Econ 2017; 22: 546–570.

43.

Abban

Hasan

MZ.

Revisiting the determinants of renewable energy investment – new evidence from political and government ideology. Energy Policy 2021; 151: 112184.

44.

Chang

T-H

Huang

C-M

Lee

M-C.

Threshold effect of the economic growth rate on the renewable energy development from a change in energy price: Evidence from OECD countries. Energy Policy 2009; 37: 5796–5802.

45.

Zhao

Luo

Driving force of rising renewable energy in China: environment, regulation and employment. Renew Sustain Energy Rev 2017; 68: 48–56.

46.

Romano

Scandurra

Carfora

, et al. Renewable investments: the impact of green policies in developing and developed countries. Renewable Sustain Energy Rev 2017; 68: 738–747.

47.

Johnstone

Haščič

Popp

Renewable energy policies and technological innovation: evidence based on patent counts. Environ Resour Econ 2010; 45: 133–155.

48.

Painuly

Wohlgemuth

Renewable energy financing – what can we learn from experience in developing countries?

Energy Stud Rev 2008; 14: 154–170.

49.

Freitas

IMB

Dantas

Iizuka

The Kyoto mechanisms and the diffusion of renewable energy technologies in the BRICS. Energy Policy 2012; 42: 118–128.

50.

Jakob

Steckel

Klasen

, et al. Feasible mitigation actions in developing countries. Nature Clim Change 2014; 4: 961–968.

51.

Seriño

MNV.

Rising carbon footpring inequality in the Philippines. Environ Econ Policy Stud 2020; 22: 173–195.

52.

Google Maps, https://www.google.com/maps/ (2014, accessed 17 December 2014).