Causal dynamics between renewable energy consumption and economic growth in South Korea: Empirical analysis and policy implications

Abstract

This paper examines the causal relationship between renewable energy consumption and economic growth in South Korea using a framework of the conventional neo-classical production function of capital, labor, and renewable energy. We use cointegration technique of the autoregressive distributed lag bounds test and vector error correction mechanism causality tests to determine the econometric relationship, using data for the period 1990–2012; the results support the conservation hypothesis for South Korea. The results of the autoregressive distributed lag bounds test show that renewable energy consumption has a negative effect on economic growth, and the results of a vector error correction mechanism causality tests indicate a unidirectional relationship from economic growth to renewable energy consumption. The empirical results imply that economic growth is a direct driver expanding renewable energy use. In terms of policy implications, it is best for policy makers to focus on overall economic growth rather than expanding renewable energy to drive economic growth.

Keywords

Renewable energy consumption economic growth conservation hypothesis ARDL bounds test VECM causality test

Introduction

The renewable energy industry is recognized as an important sector in terms of future growth, given its promise to sustain global economic development, although considerable time and effort are needed to mainstream renewable energy use.¹ Renewable energy is tapped from diverse natural resources and includes energy derived from the wind, sun, biomass, ocean currents, heat from the earth (geothermal energy), and the motion of water (hydropower). As renewable energy is already used by various existing industries, we can expect diffusion and spillover effects from the renewable industry to other industries.^2–4 Although fossil fuels have contributed much to economic development, they have fatal weaknesses, given their linkages to global warming and natural resource depletion.

Both pollution (emissions) and environmental depletion resulting from fossil fuel energy have driven research on the relationship between energy consumption and economic growth, confirming the environmental Kuznets curve (EKC). Renewable energy is recognized as an alternative energy source that can overcome the shortcoming of fossil fuels.^5–7 The expansion of renewable energy in terms of the ratio of renewable energy to total energy consumption raises a new question about the causal relationship between renewable energy and economic growth. Knowledge on this relationship can provide policymakers with new and improved policy ideas.

The relationship between renewable energy and economic growth may be explained in four ways.¹ The causal relationship between these aspects has been investigated by a number of researchers, involving various countries, periods, variables, and econometric methodologies. Tables 1 and 2 summarize the recent studies analyzing the causal relationship between renewable energy and economic growth. Unlike the previous literature, we classify the existing studies by referring to the research subjects (countries) either as a group (in Table 1) or an individual country (in Table 2).

Table 1.

Summary of recent literature on renewable energy consumption and economic growth (countries considered as groups).

Study	Methodology	Period	Country	Confirmed hypothesis
Aissa et al.⁸	Panel cointegration techniques<Group>	1980–2008	11 African countries	Neutrality (in the short run)Growth (in the long run)
Apergis and Payne⁹	Panel cointegration and panel causality tests<Group>	1985–2005	20 OECD countries	Feedback (short and long)
Apergis and Payne¹⁰	Panel cointegration and panel causality tests<Group>	1992–2007	13 Eurasian countries	Feedback (short and long)
Apergis and Payne¹¹	Panel cointegration and panel causality tests<Group>	1980–2006	6 Central American countries	Feedback (short and long)
Apergis and Payne¹²	Panel cointegration and panel causality tests<Group>	1990–2007	80 countries	Feedback (short and long)
Apergis et al.¹³	Panel cointegration and panel causality tests<Group>	1984–2007	19 developed and developing countries	Feedback (short and long)
Bhattacharya et al.¹⁴	Panel estimation techniques<Group>	1991–2012	38 top renewable energy consuming countries	Neutrality (short run)
Kahia et al.¹⁵	Panel Error Correction Model<Group>	1980–2012	11 MENA(Middle East and North Africa) Net Oil Importing Countries	Feedback (short and long)
Kazar and Kazar¹⁶	Granger causality<Group and Human development>	1980–20102005–2010	154 countries	Long: Conservation (all countries, high human development)Neutrality (very high and low human development)Feedback (middle human development)Short: Conservation (high human development)Neutrality (very high and low human development)Feedback (all countries)Growth (middle human development)
Menegaki¹⁷	One-way random effect model, panel causality tests<Group>	1997–2007	27 European countries	Neutrality
Ohler and Fetters¹⁸	Panel error correction model<Group>	1990–2008	20 OECD countries	Feedback
Sadorsky¹⁹	Panel Cointegration<Group>	1980–2005	G7 countries	Conservation
Sadorsky²⁰	Panel error correction model<Group>	1994–2003	18 emerging countries	Neutrality

Note: Renewable Energy Source (RES) in Menegaki¹⁷ is all RES including both electricity and thermal, and other articles are considering electricity RES.

Table 2.

Summary of recent literature on renewable energy consumption and economic growth (individual countries).

Study	Methodology	Period	Country	Confirmed hypothesis
Al-mulali et al.²¹	Fully modified OLS<Country specific>	1980–2009	108 countries(low, lower and upper middle, high)	Feedback (79%, long)Neutrality (19%, long)Conservation and growth (2%, long)Higher income – Feedback
Alper and Oguz²²	ARDL<Country specific>	1990–2009	New EU member countries	Growth (Bulgaria)Neutrality (Cyprus, Estonia, Hungary, Poland, and Slovenia)Conservation (Czech Republic)
Bobinaite et al.²³	Cointegration and Granger causality	1990–2009	Lithuania	Neutrality (Long)Growth (Short)
Chang et al.²⁴	Emirmahmutoglu and Kose’s method25<Country specific>	1990–2011	G7 countries	Neutrality (Germany, Italy, UK, and US)Growth (Canada, France, and Japan)excluding hydroelectricity
Dogan²⁶	ARDL<Country specific>	1990–2012	Turkey	Neutrality (in the short run)Growth (in the long run)
Kocak and Sarkgunesi²⁷	Panel cointegration<Group and country specific>	1990–2012	9 Black Sea and Balkan countries	Feedback (Group)Growth (Bulgaria, Greece, Macedonia, Russia, and Ukraine)Feedback (Albania, Georgia, and Romania)Neutrality (Turkey)
Lin and Moubarak²⁸	ARDL<Country specific>	1977–2011	China	Neutrality (in the short run)Feedback (in the long run)
Mbarek et al.²⁹	Granger causality<Country specific>	2001/Q1–2012/Q3	France	Growth (in the short run)
Menyah and Wolde-Rufael³⁰	A modified version of Granger causality test<Country specific>	1960–2007	US	Conservation
Ocal and Aslan³¹	ARDL<Country specific>	1990–2010	Turkey	Conservation
Pao and Fu³²	Error correction model<Country specific>	1980–2010	Brazil	Feedback (total renewable)Growth (excluding hydro)
Payne³³	Toda–Yamamoto procedure<Country specific>	1949–2006	US	Neutrality
Rafindadi and Ozturk³⁴	VECM<Country specific>	1971/Q1–2013/Q4	Germany	Feedback
Salim and Rafiq³⁵	Granger causality<Country specific>	1980–2006	Brazil, China, India, Indonesia, Philippines, and Turkey	Feedback (Brazil, China, Philippines, and Turkey)Growth (India and Indonesia)
Sebri and Ben-Salha³⁶	VECM<Country specific>	1971–2010	BRICS countries	Feedback (Brazil and South Africa)Growth (India)
Shahbaz et al.³⁷	VECM Granger causality<Country specific>	1972–2011	Pakistan	Feedback
Tiwari³⁸	Structural VAR<Country specific>	1960–2009	India	Growth (only hydroelectricity as renewable)
Tugcu et al.³⁹	Hatemi-J causality test⁴⁰<Country specific>	1980–2009	G7 countries	Classical Production: FeedbackAugmented Production:Feedback (England and Japan)Neutrality (France, Italy, Canada, and USA)Conservation (Germany)
Yildirim et al.⁴¹	Toda–Yamamoto and Hatemi-J causality tests<Country specific>	1949–2010	US	Neutrality (all kinds of renewable energy)Growth (biomass waste)

Note: Renewable Energy Source (RES) in Al-mulali et al.,²¹ Chang et al.,²⁴ Dogan,²⁶ Pao and Fu,³² and Rafindadi and Ozturk³⁴ is electricity RES, Tiwari³⁸ is hydro-electricity RES, and other articles are considering all RES including both electricity and thermal.

Table 1 shows a relatively consistent result; the causal relationship between renewable energy and economic growth mostly follows the feedback hypothesis.^9–13,18 Even if the results of a few studies differ from the above-mentioned norm, it is notable that there is generally no evidence of causality between renewable energy and economic growth for short time periods of about 10 years. It is highly probable that these results would change if a longer period was considered. Thus, we may safely draw a consensus from Table 1, in that the relationship between renewable energy and economic growth follows the feedback hypothesis, provided the survey period is long enough and some countries are considered together as a group.

However, the results of the survey for individual countries, seen in Table 2, tell a different story. Past studies have not reached a consensus on this issue. Al-mulali et al.²¹ analyzed the data of 108 countries and concluded that 79% of the countries have a positive bidirectional long-run relationship between renewable energy and economic growth. Some of the previous studies considering individual countries show that the relationship between renewable energy and economic growth varies greatly from country to country, depending on the length of time and methodology considered.^{24,26,31,32,35,36,39}

Thus, according to previous studies, the causal relationship between renewable energy and economic growth differs depending on whether countries are considered in groups or individually. The empirical results in the former case mainly support the feedback hypothesis. However, studies considering individual countries show very diverse opinions.

Policymakers interested in establishing optimal renewable energy policies in an individual country are mainly interested in the results of previous studies referring to individual countries rather than those considering groups, because the causality analysis for a particular country is strictly relevant to that country’s policy implications.² The main contribution of this paper is to provide guidance to policymakers toward establishing optimal renewable energy policies for South Korea.

The stage of development is also an important factor in determining the relationship between renewable energy and economic growth.^16,21 Therefore, it is crucial for policymakers to understand the relationship between renewable energy and economic growth given the changing environment during the various development stages of a country. South Korea has changed from a developing country to a developed one, but to date, no study³ has analyzed the relationship between renewable energy and economic growth in the context of South Korea. Consequently, this paper aims to fill this gap using individual country selection.

The purpose of this study is to examine the causal relationship between renewable energy consumption and economic growth using the autoregressive distributed lag (ARDL) cointegration approach of Pesaran et al.,⁴² and the causality tests of Toda and Yamamoto⁴³ for the period of 1990–2012 in South Korea. Production is specified as a function of labor, capital, and renewable energy. The remainder of the paper is organized as follows. The next section introduces the overall trend of renewable energy policy and the status of renewable energy in South Korea followed by descriptions on the data source and methods used in this study. Next, we present the results of the empirical strategy, and we conclude the paper and provide policy implications.

Trends and policies pertaining to renewable energy in South Korea

Two major acts govern renewable energy policies in South Korea. The first is the Act on the Promotion of the Development, Use and Diffusion of New and Renewable Energy, and the second is the Basic Plan on New and Renewable Energy Technology Development and Deployment, which was implemented under the existing Act on renewable energy. A simple and compact description of the two renewable energy policies appears below.

Known as the “miracle,” South Korea has made remarkable progress in the 1970s and 1980s, with its economic growth rate rising over 10% despite two negative oil shocks. However, these two oil shocks have raised concerns about energy security and sustainable economic growth for South Korea, which imports almost 100% of its energy. This situation has led the country to implement policies on renewable energy.

South Korea established its first renewable energy policy in 1987, calling it the “Act on the Promotion of the Development of Alternative⁴ Energy.” The main purpose of this Act is to focus on R&D in alternative energy technology. After the announcement of the United Nations Framework Convention on Climate Change in 1992, the existing law was revised to “Act on the Promotion of the Development, Use and Diffusion of Alternative Energy” in 1997, which emphasized both utilization and deployment. Another main purpose of this revision was to add environmental protection as part of the Law. In 2004, the Act was amended to “Act on the Promotion of the Development, Use and Diffusion of New and Renewable Energy,” marking the change in terminology from “Alternative” to “New and Renewable Energy.” This Law intends to help South Korea meet its potential growth rate, which has seen a continuous drop, while aiming for both energy security and environmental protection through the renewable energy industry.

The “Basic Plan on Alternative Energy Technology Development and Deployment”⁵ was first established in 2001. It was based on the Renewable Energy Act. This Plan had a pre-designed cycle of 5 years and a planning period of over 10 years. However, the second “Basic Plan on New and Renewable Energy Technology Development and Deployment,” was deployed after two years (in 2003), because the official name of the Basic Plan was changed to include the words “new” and “renewable” instead of “alternative.” The second Basic Plan spanned the planning period from 2003 to 2012, with the mandate to achieve a total primary energy supply of 3% from new and renewable sources in 2006 and raising the same to 5% in 2011. After five years (in 2008), the Korean government established the third “Basic Plan on New and Renewable Energy Technology Development and Deployment.” The planning period for this Plan spans 2008 to 2030, a bit longer than that of the second Basic Plan. The third Basic Plan contains two main goals. First, it aims to raise the ratio of new and renewable energy supply within the total primary energy supply to 11% by 2030, and second, it aims to foster the new and renewable energy industry as a future industry. The most recent (fourth) “Basic Plan on New and Renewable Energy Technology Development and Deployment” was released in 2014, with a planning period from 2014 to 2035. This Plan has two goals. The first is to achieve a new and renewable energy supply of 11% within the total primary energy supply by 2035, and the second is to increase the supply of new and renewable sources in total electricity production to 13.4% by 2035. Notably, the main notion of the Fourth Basic Plan is to shift the new and renewable energy sector from government control to a market orientation. To ensure this, the Basic Plan suggests increasing proactive private investment⁶ in this industry and its expansion into foreign markets. Specifically, the Plan includes various deployment and infrastructure programs⁷ to accomplish these goals, which mirror the Act on renewable energy.

This process of policy changes for new and renewable energy point to certain important facets. The most obvious aspect is that increasing new and renewable energy use has become an increasingly important goal. Second, the main goals of both the Act and the Basic Plan were revised from focusing on energy security alone, to include environmental issues and sustainable economic growth. Lastly, even if the exact nature of the relationship between renewable energy and economic growth is unknown, it is a commonly accepted view that these variables are undoubtedly interlinked, further increasing the importance of our research. There are two major effects that we can expect in the electricity market as the share of renewable on electricity market increases although it is impossible to confirm in Korea, which has a low share of renewable. One is the merit-order effect, which is the most widely known effect of renewables on electricity markets in the short term. The marginal costs of most renewables are low or they are supported without market integration, and thus the higher marginal costs make conventional plants out of the market when renewable electricity becomes available and leads to a decrease in price. The other is the increased price volatility caused by the residual load expressed as the difference between demand and renewable generation even if this is true for high share of renewable on electricity market.⁴⁴

Table 3 explains the performance of South Korea’s new and renewable energy policies, showing the trends in electricity production from renewable sources (excluding hydroelectric power). Table 3 and Figure 1 indicate that electricity production from renewable sources is very low in terms of both actual production and the proportion to total electricity production. However, we note both rates have continued to increase over time. The slope of the curve was very steep in the mid-2000s, and the actual production and proportion to total electricity production increased by over 100 times and over 50 times between 1995 and 2012, respectively.⁸

Table 3.

Trend of electricity production from renewable sources.

	Electricity production – excluding hydroelectric(unit: GWh)	Ratio to total electricity production(unit: %)
1995	25.2	0.01
2000	101.0	0.04
2005	379.0	0.10
2007	926.0	0.22
2009	1874.0	0.41
2010	2504.0	0.50
2011	2881.0	0.55
2012	3018.0	0.57

Figure 1.

(a) Trend of electricity production from renewable sources. (b) Proportion of electricity production by renewable energy sources. No data on renewable energy sources is available prior to 1995.

Data and methodology

Data

We model the short- and long-run relationships between renewable energy consumption and economic growth in South Korea. To study the relationship between renewable energy consumption and economic growth, we use a framework of the conventional neo-classical one-sector production technology, where production is a function of capital, labor, and renewable energy, as proposed by Apergis and Payne.¹¹ The product function at time t is as follows:

Y = f (L, K, R)

(1)

where Y is the economic output, L is the labor force, K is the capital stock, and R is renewable energy.

By including capital and labor in the production function, we avoid the possible omitted variable bias in the estimation of the causal relationship between renewable energy consumption and economic growth. The production function describes the long-run relationship between real output and factor inputs such as capital, labor, and renewable energy. Assuming that labor and capital are exogenous variables, the short-run dynamics between renewable energy consumption and economic growth can be examined using the causality test in the vector error correction mechanism (VECM).

We explain the role of South Korea’s renewable energy consumption policy in the country’s economic growth using a classical production model, where labor, capital, and renewable energy are inputs. Thus, we test the hypothesis that the gross domestic product (GDP) and renewable energy consumption policies have a long-run steady state relationship as specified below:

l n y_{t} = β_{0} + β_{1} l n l_{t} + β_{1} l n k_{t} + β_{1} l n r_{t} + ε_{t}

(2)

where prefix ln represents the natural logarithm, y is the real GDP in billions of constant 2005 U.S. dollars, l is the total labor force in millions, k is the real gross fixed capital formation in billions of constant 2005 U.S. dollars, and r denotes total renewable electricity consumption in millions of kilowatt hours (excluding hydroelectricity). The data for this study are sourced from the World Bank’s World Development Indicators (WDI) and cover the period 1990–2012.

Test methods

This paper first examines the time series stability of the dynamic causal relationships between the variables. To do so, we consider the augmented Dickey–Fuller (ADF) and Phillips–Perron (PP) tests. We also apply the Elliott–Rothenberg–Stock (ERS) test to improve the power of the unit root tests.⁴⁵

Then, we apply Pesaran et al.’s⁴² cointegration technique of the ARDL bounds test. The Johansen’s cointegration test may result in biases with a small sample size.⁴⁶ Moreover, the ARDL estimators of the long-run coefﬁcients are super-consistent with a small sample sizes.⁴⁷ Many previous studies have employed the ARDL bounds test.^48–51 The equation for the ARDL bound tests is

\begin{matrix} Δ l n y_{t} = ρ_{0} + β_{0} l n y_{t - 1} + β_{l} l n l_{t - 1} + β_{k} l n k_{t - 1} + β_{r} l n r_{t - 1} \\ + \sum_{i = 1}^{p} ϕ_{0 i} Δ l n y_{t - i} + \sum_{i = 0}^{p} ϕ_{l i} Δ l n l_{t - i} + \sum_{i = 0}^{p} ϕ_{k i} Δ l n k_{t - i} + \sum_{i = 0}^{p} ϕ_{r i} Δ l n r_{t - i} + {ε_{i}}_{i} \end{matrix}

(3)

Next, after confirming the existence of the cointegration relationship between the variables, we estimate the VECM to infer the causality among the variables. The dynamic error correction model with lagged residual terms from the long-run cointegration estimation can be expressed as

\begin{matrix} l n y_{t} = δ_{y} + θ_{y} e c t_{y t - 1} + \sum_{j = 1}^{p - 1} α_{y j} Δ l n y_{t - j} + \sum_{j - 1}^{p - 1} α_{ylj} Δ {lnl}_{t - j} + \sum_{j - 1}^{p - 1} α_{ykj} Δ {lnk}_{t - j} + \sum_{j - 1}^{p - 1} α_{yrj} Δ {lnr}_{t - j} + ε_{yt} \\ l n l_{t} = δ_{l} + θ_{l} e c t_{l t - 1} + \sum_{j - 1}^{p - 1} α_{l j} Δ l n_{t - j} + \sum_{j = 1}^{p - 1} α_{lyj} Δ {lny}_{t - j} + \sum_{j = 1}^{p - 1} α_{lkj} Δ {lnk}_{t - j} + \sum_{j = 1}^{p - 1} α_{lrj} Δ {lnr}_{t - j} + ε_{lt} \\ l n k_{t} = δ_{k} + θ_{k} e c t_{k t - 1} + \sum_{j = 1}^{p - 1} α_{k j} Δ l n k_{t - j} + \sum_{j = 1}^{p - 1} α_{kyj} Δ {lny}_{t - j} + \sum_{j = 1}^{p - 1} α_{klj} Δ {lnl}_{t - j} + \sum_{j = 1}^{p - 1} α_{krj} Δ {lnr}_{t - j} + ε_{kt} \\ l n r_{t} = δ_{r} + θ_{r} e c t_{r t - 1} + \sum_{j = 1}^{p - 1} α_{r j} Δ l n r_{t - j} + \sum_{j = 1}^{p - 1} α_{ryj} Δ {lny}_{t - j} + \sum_{j = 1}^{p - 1} α_{rlj} Δ {lnl}_{t - j} + \sum_{j = 1}^{p - 1} α_{rkj} Δ {lnk}_{t - j} + ε_{rt} \end{matrix}

(4)

where

{e c t}_{i t - 1}

is error correction term. As per the short-run causality test, the null hypothesis is

H_{0} : α_{i i j} = 0

for all j, and i could be replaced by y, l, k, and r.

Empirical results

To begin with the stationarity analysis, we employ the ADF, PP, and ERS unit root tests for the level and then for the first difference. The Akaike information criterion (AIC) is used to select the optimal lag length for the ADF test. For the PP test, the Bartlett kernel is used as the spectral estimation method, and the bandwidth is selected using the Newey–West method. The autoregressive (AR) spectral-ordinary least squares (OLS) is used as the spectral estimation method for the ERS test. Table 4 shows the results of the unit root tests. Most of the tests for the variables are non-stationary in the level, but all the tests are stationary in the first differences. Based on these results, all the series are integrated series of order 1.

Table 4.

Unit root tests.

		lny	lnl	lnk	Lnr
ADF	Level	−2.179	−2.291	−2.016	−1.411
ADF	First difference	−4.818***	−4.037***	−3.689**	−5.180***
PP	Level	−7.416***	−2.667*	−2.161	−1.411
PP	First difference	−4.886***	−4.037***	−4.582***	−5.167***
ERS	Level	−0.051	0.318	−0.996	−0.579
	First difference	−4.618***	−3.851***	−3.834***	−4.933***

Note: AIC is utilized to select the lag length for the ADF test. For the PP test, we use the Bartlett kernel as the spectral estimation method, and the bandwidth is selected using the Newey–West method. For the ERS test, the AR spectral OLS is used as the spectral estimation method.

*10%, **5%, and ***1% levels of significance.

Since the results of the unit root tests show that the variables of real GDP, labor force, real capital, and renewable energy consumption are I(1), we then test the existence of the cointegration relationship among the variables. We employ the ARDL bounds test with maximum lags of three, as introduced by Pesaran et al.⁴² Table 5 shows the results of the ARDL bounds test. The null hypothesis is rejected at the 1% significance level based on the F-statistic provided by Pesaran et al.⁴² We can conclude from the result of the ARDL bounds test with this small sample size that there exists a cointegration relationship among the variables.⁹

Table 5.

Results of the ARDL bounds test.

Calculated F-statistic	22.52**
Critical bounds of the F-statistic: Unrestricted intercept and no trend	1% level		5% level
	I(0)	I(1)	I(0)	I(1)
Pesaran et al.⁴²	4.29	5.61	3.23	4.35

Note: The AIC selection criteria are utilized to select the lag length. The optimal lag length was found to be ARDL (3,1,3,3). The sample size of the present study (N) is 23.

**The level of significance is 5%.

Next, we infer the short-run causality using a VECM since we have confirmed the existence of a cointegration relationship between the renewable energy consumption variable and the economic growth variable after including the variables of labor and capital from the ARDL bounds test.

The cointegrating equation extracted from the VECM for the four-variable system, where ${l n y}_{t - 1}$ is the normalized variable, is as follows:

\begin{matrix} l n y_{t - 1} = - 67.04 + 2.91 l n l_{t - 1} + 1.78 l n k_{t - 1} - 0.09 l n r_{t - 1} \\ (7 .0 4) (16.84) (- 16.39) \end{matrix}

(5)

where the values in parentheses denote the z-statistics. The renewable energy consumption index has a negative long-run effect on economic growth. To test whether there is autocorrelation at the lag order, we would ideally perform the Lagrange multiplier test, employing three lags, as inferred from the ARDL bounds test. However, our small sample size in this study does not allow us to perform the test.¹⁰

Table 6 shows the results of the causality test for the VECM using a four-variable system with three lags. There is a weak (at the 10% significance level) bidirectional short-run causality between the variables but a strongly unidirectional long-run causality from real GDP to renewable energy consumption. The statistically significant and negative coefficients of the ECT (Error correction term) variables in the cointegrating equations suggest that the dynamic interactions of the variables restore the long-run equilibrium from a deviation. The estimated coefficients of the ECT variables for both the real GDP and the renewable energy consumption equations have negative coefficients but only the renewable energy consumption equation has a statistically significant coefficient at the 1% significance level. The results suggest that there is a long-run causality from real GDP to the renewable energy consumption policy. Thus, the unidirectional long-run causality between renewable energy consumption and economic growth supports the “conservation hypothesis.” The long-run causality from economic growth to renewable energy consumption implies that a renewable energy production or consumption policy can be implemented in an advanced economy having a large real GDP, such as South Korea.¹¹ ^, ¹²

Table 6.

Results of the causality tests.

	Source of causation
	Short-run				Long-run
Dependent variable	Δly	Δll	Δlk	Δlr	ECT
Δly		9.58**	8.30**	7.57*	−0.134
Δll	3.25		4.00	2.31	0.051
Δlk	25.15***	35.24***		36.99***	−0.336
Δlr	356.61***	476.86***	515.95***		−10.602***

*10%, **5%, and ***1% levels of significance.

Table 7 shows the results of the alternative ARDL bounds test. The null hypothesis is rejected at the 5% significance level, and thus, we can conclude from the ARDL bounds test that a cointegration relationship exists among the variables. The alternative cointegrating equation with lags of two is as follows.

\begin{matrix} l n y_{t - 1} = - 60.39 - 6.59 l n l_{t - 1} + 2.92 l n k_{t - 1} + 0.10 l n r_{t - 1} \\ (- 6.95) (13.11) (4.81) \end{matrix}

(6)

where the values in parentheses denote z-statistics. The renewable energy consumption index has a statistically significant positive long-run effect on economic growth. To test whether there is autocorrelation at the lag order, we perform the Lagrange multiplier test. According to the null hypothesis, there is no autocorrelation at the lag order. The statistic for the test shows that χserial2(1) = 11.31 with a p value of 0.79, which indicates no first serial correlation in the model. The Jarque–Bera statistic for normality is 11.59 with a p value of 0.17.

Table 7.

Results of the ARDL bounds test with maximum lags of two.

Calculated F-statistic	3.86**
Critical bounds of the F-statistic: Unrestricted intercept and no trend	1% level		5% level
	I(0)	I(1)	I(0)	I1)
Pesaran et al.⁴²	4.29	5.61	3.23	4.35

Note: The AIC selection criteria are utilized to select lag length. The optimal lag length is ARDL (2,1,1,2). The sample size of the present study (N) is 23.

**The level of significance is 5%.

Table 8 shows the causality test for the alternative VECM having a four-variable system with two lags. There is no short-run causality between the variables. The estimated coefficients of the ECT variables for both the real GDP and the renewable energy consumption equations are positive, suggesting that the dynamic interactions of the variables do not restore the long-run equilibrium from a deviation. Thus, the result indicating no long-run causality between renewable energy consumption and economic growth supports the “neutrality hypothesis,” which implies that Korea’s renewable energy policy is not related to its real GDP when we consider two lags and the small sample size in this study.

Table 8.

Results of the causality tests with two lags.

	Source of causation
	Short-run				Long-run
Dependent variable	Δly	Δll	Δlk	Δlr	ECT
Δly		1.24	2.83	2.54	0.264**
Δll	1.30		4.77*	2.17	0.056
Δlk	5.49*	2.34		5.20	0.771***
Δlr	1.07	1.54	0.92		0.149

*10%, **5%, and ***1% levels of significance.

Conclusions and policy implications

We examined the short- and long-run dynamic causal relationships between renewable energy consumption and economic growth in South Korea, including labor and capital to the estimation. We used data for the time period 1990–2012. First, we conducted the unit root test for time series stability. The cointegration test was then performed with the ARDL bounds test, and causality tests were employed using VECM.

Given the small sample size, the ARDL bound test showed the existence of a long-run relationship between the renewable energy consumption and economic growth variables. Notably, labor and capital were added to the estimation to avoid possible omitted variable bias. Further, the findings of causality test that employed a VECM with three lags indicated that there is statistically significant short- and long-run causality, supporting the “conservation hypothesis” between economic growth and renewable energy consumption with the inclusion of the labor and capital variables in the estimation. To be clear, we identified that there is unidirectional long-run causality form economic growth to renewable energy consumption.

Despite the limited sample size, we can infer several policy implications from the results. First, there are two main purposes of the expansion of renewable energy consumption. One is focusing on environmental issues, such as pollution emissions reduction, and the other is to drive economic growth. Although we do not examine yet, the renewable energy policy should be focusing on environmental issues in addition to economic growth, because there is no any emission from renewable energy. Second, our result suggests that renewable energy policy should be implemented when the real GDP is enough large to overcome the negative impact from renewable energy, because the causality from economic growth to renewable energy consumption in the long run as one of our result is caused by both low productivity of renewable energy production and expansion of government-led renewable energy. Third, to drive economic growth from the expansion of renewable energy we need various economic policies to increase the productivity of renewable energy production as well as energy policy that can be implemented by market-driven rather than government-led expansion of renewable energy. Fourth, a wide variety of renewable energy policy should be implemented depending on the share of renewable energy in total energy. Finally, future analyses with larger sample size could help us confirm the hypothesis so as to establish an optimal renewable energy policy for South Korea.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2015S1A5A2A03049131), and this article was also partially financially supported by College of Public Policy at Korea University in 2018.

Notes

Seong-Hoon Lee received his PhD in University at Buffalo and is currently working in the Department of Economics at Korea University (Sejong). His research has been on various aspects of productivity change, knowledge spillover, market structure, financial markets, energy efficiency, and exporting activity of firms. He has several papers published in the journals such as Small Business Economics and Energy Economics.

Yonghun Jung received his PhD in University at Buffalo and is currently working in the Department of Economics at Korea University (Sejong). His research interests include the dynamic of exporting firms, productivity and efficiency, and applied econometrics. His research has been published in the journal such as Energy policy and Energy Economics.

References

Pleune

The role of renewable energy sources in a sustainable world. Energy Environ 1992; 3: 430–443.

Klobasa

Winkler

Sensfub

et al . Market integration of renewable electricity generation – the German market premium model. Energy Environ 2013; 24: 127–146.

Kardooni

Yusoff

Kari

Barriers to renewable energy development: five fuel policy in Malaysia. Energy Environ 2015; 26: 1353–1361.

Vrba

Spacek

Jez

et al . Integration of electricity from renewable energy sources – the Czech story. Energy Environ 2015; 26: 157–166.

Zhuang

Xie

Energy consumption, economic development and environmental improvement in China. Energy Environ 2014; 25: 1345–1357.

Niu

Chang

Yang

et al . The long-run relationships between energy efficiency and environmental performance: Global evidence. Energy Environ 2017; 28: 706–724.

Duscha

Rio

An economic analysis of the interactions between renewable support and other climate and energy policies. Energy Environ 2017; 28: 11–33.

Aissa

MSB

Jebli

Youssef

SB.

Output, renewable energy consumption and trade in Africa. Energy Policy 2014; 66: 11–18.

Apergis

Payne

JE.

Renewable energy consumption and economic growth: evidence from a panel of OECD countries. Energy Policy 2010; 38: 656–660.

10.

Apergis

Payne

JE.

Renewable energy consumption and growth in Eurasia. Energ Econ 2010; 32: 1392–1397.

11.

Apergis

Payne

JE.

The renewable energy consumption-growth nexus in Central America. Appl Energy 2011; 88: 343–347.

12.

Apergis

Payne

JE.

Renewable and non-renewable energy consumption-growth nexus: Evidence from a panel error correction model. Energ Econ 2012; 34: 733–738.

13.

Apergis

Payne

Menyah

et al . On the causal dynamics between emissions, nuclear energy, renewable energy, and economic growth. Ecol Econ 2010; 69: 2255–2260.

14.

Bhattacharya

Paramati

Ozturk

et al . The effect of renewable energy consumption on economic growth: evidence from top 38 countries. Appl Energy 2016; 162: 733–741.

15.

Kahia

Aissa

Lanouar

Renewable and non-renewable energy use – economic growth nexus: the case of MENA Net Oil Importing Countries. Renew Sustain Energ Rev 2017; 71: 127–140.

16.

Kazar

The renewable energy production-economic development nexus. Int J Energ Econ Pol 2014; 4: 312–319.

17.

Menegaki

Growth and renewable energy in Europe: a random effect model with evidence for neutrality hypothesis. Energ Econ 2011; 33: 257–263.

18.

Ohler

Fetters

The causal relationship between renewable electricity generation and GDP growth: a study of energy sources. Energ Econ 2014; 43: 125–139.

19.

Sadorsky

Renewable energy consumption, CO2 emissions and oil prices in the G7 countries. Energ Econ 2009; 31: 456–462.

20.

Sadorsky

Renewable energy consumption and income in emerging economies. Energy Policy 2009; 37: 4021–4028.

21.

Al-Mulali

Fereidouni

Lee

et al . Examining the bi-directional long run relationship between renewable energy consumption and GDP growth. Renew Sustain Energ Rev 2013; 22: 209–222.

22.

Alper

Oguz

The role of renewable energy consumption in economic growth: evidence from asymmetric causality. Renew Sustain Energ Rev 2016; 60: 953–959.

23.

Bobinaite

Juozapaviciene

Konstantinaviciute

Assessment of causality relationship between renewable energy consumption and economic growth in Lithuania. Eng Econ 2011; 22: 510–518.

24.

Chang

Gupta

Inglesi-Lotz

et al . Renewable energy and growth: evidence from heterogeneous panel of G7 countries using Granger causality. Renew Sustain Energ Rev 2015; 52: 1405–1412.

25.

Emirmahmutoglu

Kose

Testing for Granger causality in heterogeneous mixed panels. Econ Model 2011; 28: 870–876.

26.

Dogan

The relationship between economic growth and electricity consumption from renewable and non-renewable sources: a study of Turkey. Renew Sustain Energ Rev 2015; 52: 534–546.

27.

Kocak

Sarkgunesi

The renewable energy and economic growth nexus in black sea and Balkan Countries. Energy Policy 2017; 100: 51–57.

28.

Lin

Moubarak

Renewable energy consumption – economic growth nexus for China. Renew Sustain Energ Rev 2014; 40: 111–117.

29.

Mbarek

Nasreen

Feki

The contribution of nuclear energy to economic growth in France: short and long run. Qual Quant 2015; 1–20.

30.

Menyah

Wolde-Rufael

CO2 emissions, nuclear energy, renewable energy and economic growth in the US. Energy Policy 2010; 38: 2911–2915.

31.

Ocal

Aslan

Renewable energy consumption-energy growth nexus in Turkey. Renew Sustain Energ Rev 2013; 28: 494–499.

32.

Pao

Renewable energy, non-renewable energy and economic growth in Brazil. Renew Sustain Energ Rev 2013; 25: 381–392.

33.

Payne

JE.

On the dynamics of energy consumption and output in the US.

Appl Energy 2009; 86: 575–577.

34.

Rafindadi

Ozturk

Impacts of renewable energy consumption on the German economic growth: evidence from combined cointegration test. Renew Sustain Energ Rev 2017; 75: 1130–1141.

35.

Salim

Rafiq

Why do some emerging economies proactively accelerate the adoption of renewable energy?

Energ Econ 2012; 34: 1051–1057.

36.

Sebri

Ben-Salha

On the causal dynamics between economic growth, renewable energy consumption, CO2 emissions and trade openness: fresh evidence from BRICS countries. Renew Sustain Energ Rev 2014; 39: 14–23.

37.

Shahbaz

Zeshan

Afza

Is energy consumption effective to spur economic growth in Pakistan? New evidence from bounds test to level relationships and Granger causality tests. Econ Model 2012; 29: 2310–2319.

38.

Tiwari

KA.

A structural VAR analysis of renewable energy consumption, real GDP and CO2 emissions: evidence from India. Econ Bull 2011; 31: 1973–1806.

39.

Tugcu

Ozturk

Aslan

Renewable and non-renewable energy consumption and economic growth relationship revisited: evidence from G7 countries. Energ Econ 2012; 34: 1942–1950.

40.

Hatemi-J

Asymmetric causality tests with an application. Empir Econ 2012; 43: 447–456.

41.

Yildirim

Saraç

Aslan

Energy consumption and economic growth in the USA: evidence from renewable energy. Renew Sustain Energ Rev 2012; 16: 6770–6774.

42.

Pesaran

Shin

Smith

RJ.

Bounds testing approaches to the analysis of level relationships. J Appl Econ 2001; 16: 289–326.

43.

Toda

Yamamoto

Statistical inference in vector autoregressions with possibly integrated processes. J Econom 1995; 66: 225–250.

44.

Winkler

Gaio

Pfuger

et al . Impact of renewables on electricity markets – do support schemes matter? Energy Policy 2016; 93: 157–167.

45.

Elliott

Rothenberg

Stock

JH.

Efficient test for an autoregressive unit root. Econometrica 1996; 64: 813–836.

46.

Cheung

Lai

KS.

Finite-sample sizes of Johansen's likelihood ration tests for conintegration. Oxford Bulletin of Economics and Statistics 1993; 55: 313–328.

47.

Pesaran

Shin

Generalized impulse response analysis in linear multivariate models. Econ Lett 1998; 58: 17–29.

48.

Fatai

Oxley

Scrimgeour

Modeling and forecasting the demand for electricity in New Zealand: a comparison of alternative approaches. Energy J 2003; 24: 75–102.

49.

Narayan

Smyth

The residential demand for electricity in Australia: an application of the bounds testing approach to cointegration. Energy Policy 2005; 33: 467–474.

50.

Narayan

Singh

The electricity consumption and GDP nexus for the Fiji Islands. Energ Econ 2007; 29: 1141–1150.

51.

Wolde-Rufael

Electricity consumption and economic growth: a time series experience for 17 African countries. Energy Policy 2006; 34: 1106–1114.