Data description

The dataset used in this study covers 117 developing countries from 1980 to 2011. This dataset is compiled using four different sources: Energy Information Administration (EIA), International Energy Agency (IEA), World Development Indicators (WDI) and the BP Statistical Review of World Energy (BP). Our main source of data for the generation of renewable energy comes from the IEA. Data from the IEA on nonhydro electricity generation can be considered comprehensive, however it may have underestimated the amount of energy generated as off-grid generation are not included in the data set (Pfeiffer and Mulder 2013).

To identify factors that may affect the diversification of nonhydro renewable energy, we include several variables highlighted in the literature that influence the adoption of renewable energy. The independent variables included in the analysis are discussed as follows.

(1) The effect of income is widely captured in the literature (Marques et al. 2010; Popp et al. 2011; Zhao et al. 2013; Aguirre and Ibikunle 2014). Higher income will enable countries to invest in varied sources of renewable energy. To capture the nonlinear effect of income on renewable energy, which is often neglected in the previous studies, we include the squared term of income in the analysis.

(2) Energy imports capture the degree of dependence on external sources. We expect that greater dependence on foreign sources of energy will fuel investment on renewable energy to improve energy security (Marques et al. 2010; Popp et al. 2011;

Dong 2012; Zhao et al. 2013; Aguirre and Ibikunle 2014).

(3) While most of the available studies did not include population growth, we hypothesize that an increase in population will drive up demand for energy and this will positively contribute to the development of renewable energy as the country tries to address the increasing energy needs of its population.

(4) Adopting renewable technologies can be facilitated by the degree of technological advances in developing countries. We use total patents as a proxy for technological innovation in the developing countries. Popp et al. (2011) used patent data to assess the role of technological innovation in advancing investment on renewable

78 technologies and found that technological advances do lead to greater investment in renewable technologies. Though patents are an imperfect measure of the innovative performance of a country, it is still considered as a relevant measure of technological innovation (Johnstone et al. 2010).

(5) Access to finance plays a crucial role in renewable energy development. Investment in renewable energy requires large upfront costs. We capture this by using a financial development variable that measures the share of domestic credit to the private sector.

It’s the same variable used by Brunnschweiler (2010) in examining the role of financial development on renewable energy development.

(6) To control for the impact of education on renewable energy development, we use secondary school enrollment as a proxy for human capital (Pfeiffer and Mulder 2013;

Zhao et al. 2013). Venturing into renewable energy is knowledge or technology intensive and this can be facilitated by a certain level of human capital development.

(7) Foreign direct investments (FDI) and official development assistance (ODA) are two external sources of funding, which may potentially influence the adoption of renewable energy in developing countries. This is measured as a share of GDP.

(8) One of the first key steps in attempting to control global carbon emissions was the adoption of Kyoto protocol in late 1997. We control the impact of Kyoto protocol by introducing a time dummy from 1998 onwards.

(9) Renewable energy policies facilitate the adoption of renewable energy in developing countries. According to Pfeiffer and Mulder (2013) developing countries are increasingly implementing policies promoting renewable energy even though there are no binding agreements for them to reduce emissions. The IEA complies several policy types related to renewable energy including economic instruments, information and education, policy support, regulatory instruments, research and development and voluntary approaches (IEA 2014). We use a dummy variable if developing countries implement any of these policies.

(10) We include hydropower energy as an additional control variable. We suspect that countries having large hydropower may not be as keen in investing in new renewable energies as opposed those who do not have it. Alternatively, it could also be possible that they are more enthusiastic in adopting nonhydro energy given they already have the experience in dealing with renewable energies.

(11) To take into account the traditional sources of energy, we include in the analysis the country’s production of coal and crude oil (Marques et al. 2010). We want to include local production of gas but data is only available starting 1990, hence we dropped it.

Countries that have a relatively higher production of these traditional sources may be reluctant to invest in renewable energy. If renewable energy complements traditional sources, then an increase in coal and oil production is positively associated with renewable energy; while if it substitutes, we expect the opposite.

(12) We include crude oil price to capture the impact of market prices on renewables.

Since we are using the world price for crude oil, these prices vary only with time and

79 not across countries. It would have been ideal if we could have used the local prices for conventional fossil fuels in each developing country but unfortunately, that is not addressed in this current paper because of limited data availability.

(13) Aside from regional dummies, we also include in the analysis coastal dummy taking value of 1 if a country has a coastal area and 0 otherwise. This serves as proxy for generation potential of renewable energy or ease of trading because of its accessibility.

(14) Lastly, we control for time variations by including year dummies in the regression.

We try to capture several relevant variables that may potentially influence the diversification of nonhydro renewable energy in developing countries. However, we still cannot discount the fact that there might be other variables not included in the analysis that may influence the diversification. Tables 4.1 and 4.2 provide the data descriptions and summary statistics of the variables used in this study.

Table 4.1. Data descriptions.

Variable Definition Source

Dependent variables

Number of nonhydro Number of nonhydro renewable sources (wind, solar,

geothermal, waste and biomass) EIA

Diversity index Takes value 1 if a country adopts only 1 nonhydro source and converges to total number of nonhydro sources if each source generates electricity equally. If a country has not adopted any nonhydro sources, its value is zero

Own

computation Nonhydro energy

generation

Total generation of nonhydro renewable electricity in

billion kilowatt-hours per 1 million people EIA Independent variables

GDP per capita GDP per capita in constant 2005 USD WDI

Energy import Net energy imports (% of energy use) WDI

Population growth Annual population growth in (%) WDI

Patents Total patent application WDI

Access to finance Domestic credit to private sector (% of GDP) WDI Secondary enrollment Secondary school enrollment (% gross) WDI FDI Foreign direct investment, net inflows (% of GDP) WDI ODA Net official development assistance received (% of GNI) WDI Crude oil price Crude oil prices (West Texas intermediate) BP Kyoto protocol Dummy variable takes value of 1 from 1998 onwards IEA

Renewable policy Dummy variable takes value of 1 from the year of

implementation of a renewable energy policy IEA Hydro energy Total hydroelectric power generated in billion

kilowatt-hours / 1 million people EIA

Oil production Total oil production in thousands barrel / 1 thousand people EIA Coal production Total coal production in thousand tons / 1 thousand people EIA

Coastal Dummy variable takes a value of 1 if a country has a coast Google map

80 Table 4.2. Descriptive statistics for developing countries from 1980 to 2011.

Variable Obs Mean Std. Dev. Min Max

Dependent variables

Number of nonhydro 3946 0.53 0.90 0 4

Diversity index 3946 0.36 0.56 0 2.89

Nonhydro energy 3946 0.01 0.05 0 0.67

Independent variables

GDP per capita 4087 2154.27 2159.19 50.04 14777.22

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 dev 3707 28.59 24.60 0.56 167.54

Secondary enrollment 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

Kyoto protocol 4752 0.45 0.50 0 1

Renewable policy 4752 0.10 0.30 0 1

Hydro energy 4070 0.37 0.89 0 10.08

Oil production 3927 4.23 13.93 0 133.73

Coal production 4142 0.29 1.01 0 12.55

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 4.1