Do Shifts in Renewable Energy Operation Policy Affect Efficiency: Korea’s Shift from FIT to RPS and Its Results

Abstract

South Korea’s new and renewable energy (NRE) policy experienced a drastic shift from the Feed-in Tariff (FIT) to the Renewable Portfolio Standard (RPS) in 2012. This study looks at the changes in the efficiency of NRE policy in this transition through DEA (Data Envelopment Analysis) and MI (Malmquist Index) methods, using investment for NRE technology development and for NRE dissemination as input factors and the number of firms, the number of employees, and the volume of NRE power generation as output factors. The results show a temporary drop in efficiency in 2012 during the transition period for the NRE industry as a whole. However, apart from those energy types with ulterior factors, the implementation of RPS increased the technical change (TC) of most NRE types. Furthermore, the findings highlight that, among South Korea’s three focal NRE industries—photovoltaic, wind power, and fuel cell energies—only fuel cell energies showed an increase in efficiency over time. South Korea’s policy shifts from FIT to RPS and the resulting effects on NRE policy’s efficiency provide a useful reference and guideline for government decision-making on NRE policy changes.

1. Introduction

South Korea’s new and renewable energy (NRE) policy was implemented full-scale in 2001 with the introduction of the Feed-in Tariff (FIT) policy to promote the NRE industry and to reduce the use of fossil fuels. One of the problems behind FIT is that the policy usually results in imposing financial burdens on taxpayers, distributors, and/or the government [1]. The increase in financial burden adversely affects the public acceptance of the policy, and consequently leads to pressures on the government or congress for policy change [2]. The South Korean government was no exception to this pressure, and in 2012, the government shifted its NRE policy from FIT to the Renewable Portfolio Standard (RPS) for reasons of high financial burden [3].

South Korea’s shift from FIT to RPS presents a meaningful case study for NRE policies not only because people were paying more and more interest to energy efficiency [4], but also as such a complete shift in national policy is rarely found. According to South Korea’s New and Renewable Energy Center, other than South Korea, only three countries—Belgium, Italy, and Japan—have undergone a similar policy shift from FIT to RPS [5]. The exceptional case of South Korea becomes further highlighted when considering that Italy revised its policy back to incorporate FIT sometime after its shift to RPS and Japan, who made the shift from FIT to RPS, moved back to FIT only after a few years. The uniqueness of South Korea’s policy shift opens a rare opportunity to investigate the difference in outcomes between the FIT and RPS policies. So far, however, while there is much research done from the public policy perspective with regard to this policy shift, few studies have been conducted on how the shift in policy affected the efficiency of South Korea’s NRE policy on the whole and/or by NRE type.

This study looks into the effect of South Korea’s NRE policy shift from FIT to RPS on the policy’s efficiency using data envelopment analysis (DEA) and the Malmquist index (MI). In particular, this study looks not only at the changes in the efficiency of the overall NRE policy due to the policy shift, but also in the efficiency of each NRE type. Through this investigation, we aim to provide novel insight into NRE policy-making for future reference to countries that are contemplating a shift between FIT and RPS in their NRE policy. Section 2 gives the background and literature review pertaining to NRE, followed by the details of the methodology and data in Section 3. The results and discussion in Section 4 elaborate the interesting findings made in this study, including the temporary drop in the efficiency of South Korea’s overall NRE policy in 2012 when the policy shift was made and the recovery of efficiency in the years afterward, and the effect of RPS on increasing technical change (TC) in most of NRE types. However, only fuel cell energy, among the South Korean government’s three strategic NRE areas, showed significant and consistent growth rate after the policy shift. Lastly, Section 5 presents the conclusion and policy implications of the study. Overall flow of this study is shown in Figure 1.

2. Background and Literature Review

2.1. South Korea’s New and Renewable Energy Policy

The importance of NRE has become widely acknowledged over the years [6]. However, due to its relatively late introduction as an alternative energy source, NRE lacks competitiveness in production cost compared to existing fossil fuel-based energy, which has been a barrier for its greater use and diffusion. Thus, at the present stage, government policies promoting the use of NRE are essential for the development and growth of the nation’s NRE industry [7]. Various policies are adopted by countries around the world towards this end, and these policies can be broadly categorized into two types: FIT and RPS (RPS is sometimes called Renewable Obligation (RO) in countries such as the United Kingdom and Australia, however, the basic concept behind RPS and RO is very similar (Dong [8]). For the purposes of this study, RPS will be understood to include RO.) [9].

Table 1. Comparison of FIT and RPS.

	Feed-in Tariff (FIT)	Renewable Portfolio Standard (RPS)
Concept	Price adjustment by amount of government subsidy, based on which the market decides the amount of NRE production	Demand adjustment by obligatory amount of NRE production set by government, based on which the market decides the price
Countries of implementation	European countries including Germany, Spain, and France	North America (USA, Canada) and the UK
Advantages	1) Guarantees mid-to-long term price, meaning high investment security 2) Promotes industry entry by firms, leading to fast growth of overall NRE industry 3) Industry growth centered around small-scale firms in various regions	1) Less financial burden on government 2) Promotes competition among NRE producers, thus driving technological development and competitive pricing 3) Easy to manage and predict production
Disadvantages	1) High financial burden on government 2) Little competition within industry, thus less drive for technological development to lower production cost	1) Market largely consists of big-scale firms 2) May lead to disproportionate growth concentrated in a few low-cost energy sources

below provides a summarized comparison of FIT and RPS based on Dong [8] and Trappey et al. [10] in terms of their concepts, countries of implementation, and advantages and disadvantages.

In South Korea, the implementation of FIT in 2001 laid provisions for subsidizing NRE power producers with the government’s Electric Power Industry Base Fund to make up the difference between the production cost and the market price for power produced and supplied using NRE. At the time FIT was implemented in 2001, the goal was to expand the production and dissemination of NRE-based power in order to alleviate the dependency on fossil fuel and to develop alternative energy sources. However, as those receiving subsidy under FIT expanded, so did the financial burden on the South Korean government. Consequently, on 25 April 2008, an economic policy adjustment meeting was held to review the FIT policy and the decision was made to adjust the price standard to a lower level with the goal to ultimately discontinue FIT at the end of 2011 to minimize the budgetary concerns caused by the policy. Instead, in 2012, in place of FIT, RPS was introduced [11]. This shift relieved the government’s financial burden as the trend in South Korea’s NRE-related government budget over the past few years show: From 876 billion KRW in 2010 and 1.3 trillion KRW in 2011 when the burden reached its height due to FIT, to 998 billion KRW in 2012, 851 billion KRW in 2013, and 802 billion KRW in 2014 with the shift to RPS.

The following paragraphs illustrate the details of RPS implemented in South Korea based on the “2014 New & Renewable Energy White Paper” published by the Ministry of Knowledge Economy and the South Korea Energy Management Corporation, the “Act on the Promotion of the Development, Use, and Diffusion of New and Renewable Energy,” and the Ministry of Trade, Industry, and Energy’s “Guidelines on the Management and Operation of Obligatory Renewable Energy Supply Program.”

South Korea’s RPS policy delineates, in the “Act on the Promotion of the Development, Use, and Diffusion of New and Renewable Energy,” the responsibility for all power producers with non-NRE power production facilities over a 500 MW capacity (obligatory producers) to meet an RSP target ratio for NRE-based power generation. As of 2015, a total of 17 power producers, including South Korea Hydro & Nuclear Power and South Korea South East Power are under RPS, and their obligatory renewable service supply ratio is set at 3% of their total power generation. This obligatory ratio is to be increased incrementally to reach 10% by 2024. Also, for photovoltaic energy, whose production unit cost is still relatively steeper than other NRE types, but which has the potential for great diffusion and growth, a separate obligatory ratio is given, independent from that for other NRE sources.

The power producer receives a renewable energy certificate (REC) if it produces a certain amount of power using NRE power facilities, which it is obligated to submit to prove it has fulfilled its mandatory supply ratio. REC is given to the power producers based on calculations using the NRE power (MWh) produced and supplied, multiplied by the REC weight given to the NRE type which is set by the government. For instance, if 1 MWh of NRE power was produced and supplied using geothermal energy, which has a REC weight of 2.0, the power producer will receive 2 RECs (1 MWh times 2.0). The REC weight was implemented to prevent excess investment in NRE types with low production unit costs, which can disrupt and deter the comprehensive use and diffusion of NREs. In addition, an REC trading market, under the supervision of the supply certification institution, was created to facilitate the buying and selling of RECs among power producers.

Under this new RPS policy, the South Korean government further specified three focal NRE areas, namely, photovoltaic, wind power, and fuel cell energies, and announced intentions to heavily invest in the development of core technologies and human resources for these NREs to gain key know-how on securing raw material, converting energy, and utilizing NRE power supply.

2.2. Literature Review

Previous literature on the topic of NRE and its efficiency has been reviewed for reference in setting the appropriate input and output factors for conducting DEA and MI. Unfortunately, while there are many researchers studying the efficiency of the NRE industry, we found a lack of those who investigated the efficiency of NRE policy using DEA or MI. Therefore, we focused on finding common input and output factors used by previous studies conducting DEA and/or MI on NRE-related topics to construct our research model on this novel topic.

Table 2. Previous literature on renewable energy and efficiency.

Authors	Year	Subject	Input	Output
Chien and Hu	2007	Technical efficiency of renewable energy in 45 countries	Labor employment, Capital stock, Energy consumption	Real GDP 2007
Barros	2008	Hydropower plants in Portugal	Number of workers, Capital, Operating cost, Investment	Production in MWh, Capital utilization
Jha and Shrestha	2008	Hydropower plants in Nepal	Installed capacity of the plant, Total operating and maintenance expenditure, Number of employees	Energy generated by the plant, Winter peaking capacity, Summer peaking capacity
Madlener et al.	2009	Biogas energy in Austria	Labor, Organic dry substance	Electricity, Heat
San Crisobal	2011	Renewable energy technology	Investment ratio, Implement period, Operating and maintenance cost	Power generation, Operating hours, Useful life, Electricity, Tons of CO2 avoided
Halkos and Tzeremes	2012	Renewable energy firms in Greece	Current ratio, Debt/Equity ratio, Assets turnover ratio	ROE, ROA, Margin, Gross profit margin
Lins et al.	2012	Alternative energy in Brazil	Greenhouse gases emission, Potential job creation, Potential distributed generation	Operating and maintenance cost, Investment cost
Kim et al.	2015	Investment for renewable energy in South Korea	Investment for renewable energy technology development, Investment for renewable energy dissemination	Number of patents, Power generation, Unit cost of power generation
Woo et al.	2015	Environmental efficiency of renewable energy among OECD countries	Total labor, Total capital, Renewable energy supply	GDP Carbon emission, Renewable energy generation
Meleddu and Pulina	2018	Performance of public spending on renewable energy in Italian regions	Research & development expenses, Other expenses, Radiation protection expenses, Electric power consumption	Photovoltaic power, Renewable energy
Lyu and Shi	2018	Financing efficiency of renewable energy industry	R&D investment, Stock market, Project financing, Venture capital, Private equity	Renewable energy generation

below summarizes the literature reviewed for this study.

Halkos and Tzeremes [12] used DEA to find out the efficiency of NRE firms in Greece, using the current ratio, debt to equity ratio, and assets turnover ratio as input factors and return on equity, return on assets, margin, and gross profit margin as output factors. Meanwhile, Cristobal [13], studying renewable energy technology, used investment ratio, implement period, and operating and maintenance cost as input factors, while setting power generation, operating hours, useful life, electricity, and tons of CO₂ avoided as output factors. Woo et al. [14] analyzed the environmental efficiency of NRE among OECD countries, using as input factors, total labor, total capital, renewable energy supply and as output factors, gross domestic product, carbon emission, and renewable energy generation for DEA and MI analysis. Barros [15] looked at the efficiency of hydropower plants in Portugal, using the number of workers, capital, operating cost, and investment as input factors and the amount of production in MWh and capital utilization as output factors. Lyu and Shi [16] used DEA to analyze the financing efficiency in global renewable energy industry and in different ways of financing.

Among the literature reviewed, Kim et al. [17] presented a helpful guideline for the present study due to its relevance to the topic on hand. They applied the South Korean government’s investment for renewable energy technology development and investment for renewable energy dissemination as input factors and the number of patents, power generation, and the unit cost of power generation as output factors for DEA to find out the efficiency of the South Korean government’s investment on NRE. As this study’s purpose is to analyze the efficiency of South Korea’s NRE policy, we follow Kim et al. in setting the input factors for DEA, namely, investment for renewable energy technology development and investment for renewable energy dissemination. However, for output factors, we use the number of firms and the number of employees in the NRE industry and the amount of NRE production. The reasons behind the choice of these output factors for the present analysis will be elaborated further in the next section on methodology.

3. Methodology

3.1. Data Envelopment Analysis and Malmquist Index

When measuring efficiency, SFA (Stochastic Frontier Approach), DFA (Distribution Free Approach), TFA (Thick Frontier Approach), DEA (Data Envelopment Analysis), and MI (Malmquist Index) are frequently used and each has its own advantages and disadvantages. Considering characteristics of this study, DFA, presented by Schmidt and Sikles (1984) and modified by Berger (1993), is not suitable for short period analysis since it affects the data by offsetting random errors [18,19]. TFA, introduced by Berger and Humphrey [20], needs a considerable number of DMUs to comprise the frontier. SFA, developed by Aigner et al. [21], requires an assumption of a specific function’s pattern and distribution. Thus, having data for a relatively short period and a small number of DMUs, this study used DEA for static analysis and MI for dynamic analysis in order to precisely capture the changes in energy efficiency.

Linear Programming (LP) is a mathematical technique used for optimizing the distribution of limited resources to achieve a decision-making goal. It is most often utilized in maximizing profit or minimizing cost, and in doing so, the objective function and constraints conditions are performed linearly [22]. Data envelopment analysis (DEA) makes use of linear programming to analyze the relative efficiency within a specific group, and has been applied in various areas of study since its introduction by Charnes et al. [23]. Charnes et al. [23] delineated that the ratio of the weighted sum outputs to the weighted sum of inputs of each decision-making unit (DMU) for evaluation cannot exceed 1, and defined DEA as a linear programming technique that aims to maximize the ratio of the weighted sum of outputs, assuming the condition that the weights of each input and output factor are greater than 0 [24]. The degree of inefficiency in the input-to-output ratio is most commonly analyzed in terms of ranking using the Charnes—Cooper—Rhodes (CCR) model or the further-developed Banker—Charnes—Cooper (BCC) model. The difference between the CCR model and the BCC model is that the BCC model allows an analysis of variable returns to scale [25].

Malmquist index (MI) is a dynamic analysis tool which was developed in order to overcome the shortcomings of DEA, that is, that DEA can only analyze a static situation in a specific time period. MI analyzes the ratio of input to output without assuming a specific production index, which allows for a dynamic evaluation of the changes in efficiency. MI is calculated using Malmquist analysis by multiplying the efficiency change (EC), which expresses the change in the proximity of the object of analysis to the production frontier from one period of time to another, and technical change (TC), which expresses how the change in technology occurring between one period of time and another has affected efficiency. When the concepts of EC and TC are applied to government policy, EC shows how efficiently resources are used to achieve the goal of a specific policy and TC shows the changes that occur outside the specific policy such as the introduction of a new policy.

In this study, we use the CCR-DEA model assuming constant return to scale (CRS) for static analysis and MI for dynamic analysis to understand the unique case of South Korea’s NRE policy shift from FIT to RPS from multiple angles. Detailed equation used in this study is shown below:

$$\begin{array}{c}CCR-DEA:\mathrm{Minimize}{E}_{j0}={\displaystyle \sum }_{r=1}^s{u}_r{y}_{rj0}\\ \mathrm{Subject}\mathrm{to}{\displaystyle \sum }_{i=1}^m{v}_i{x}_{ij0}=1\\ {\displaystyle {\displaystyle \sum }_{r=1}^s}{u}_r{y}_{ri}-{\displaystyle {\displaystyle \sum }_{i=1}^m}{v}_i{x}_{ij}\le 0, j=1,\dots ,j_0,\dots ,n\\ u_r\ge \epsilon >0, r=1,\dots ,s\\ v_i\ge \epsilon >0, i=1,\dots ,s\end{array}$$

(1)

$$MI:M_0\left(y_{t+1},x_{t+1},y_t,x_t\right)= {\left[\frac{D_0^t\left(x_{t+1},y_{t+1}\right)}{D_0^t\left(x_t,y_t\right)}\times \frac{D_0^{t+1}\left(x_{t+1},y_{t+1}\right)}{D_0^{t+1}\left(x_t,y_t\right)}\right]}^{1/2}.$$

(2)

3.2. Data Collection

The research model showing the input and output factors used for analysis is illustrated in Figure 2 below. Two input factors and three output factors are used in this study to find out the efficiency of NRE policy by energy source: Investment for NRE technology development and investment for NRE dissemination as input factors and the number of firms, the number of workers, and the volume of power generation as output factors.

Of the two input factors, investment for NRE technology development refers to the amount of investment made on research and development by the South Korean government for the NRE type, and investment for NRE dissemination is the government subsidy provided towards the promotion of NRE use and diffusion. In this study, the latter includes the South Korean government’s Housing Subsidy Project, Building Subsidy Project, Regional Subsidy Project, Financial Subsidy Project, and the FIT Project. The data for the two input variables were obtained from the “2014 New & Renewable Energy White Paper,” published by the Ministry of Knowledge Economy and the South Korea Energy Management Corporation [5].

The details of the investment for NRE dissemination included in this study are described below. The Housing Subsidy Project reflects the South Korean government’s intention to provide NRE power to all homes, and the “One Million Green Home Project” is its representative project. The government subsidy was implemented since 2009 to supply the most appropriate NRE to households based on comprehensive considerations, including the environment of each region and home (e.g., the amount of solar radiation, wind speed, amount of water), industrial impact, and potential job creation. The Building Subsidy Project has been in place since 1993 with the goal to expand the NRE market and to promote the commercialization of new NRE technology in public buildings, universities, and social welfare facilities by subsidizing a part of the installation cost for NRE equipment.

The Regional Subsidy Project takes into consideration the characteristics of each local self-governing body and their potential NRE resources, and provides educational opportunities to government officers and employees in charge of NRE-related projects and directly subsidizes the installation of NRE equipment. This project has been in place since 2006. The Financial Subsidy Project started in 2000 and provides low-interest, long-term loans for NRE equipment installation to promote private sector investment in the NRE industry. Lastly, the FIT (Project), implemented in 2001, subsidizes the difference between the market price and the government-adjusted price, but, as mentioned beforehand, was discontinued in at the end of 2011 with the change in policy shift to RPS. However, for existing recipients of FIT, the subsidy will be maintained or decreased incrementally until 2030 and thus has been accounted for in this study.

The number of firms and the number of workers in the NRE industry and the volume of NRE power generation were chosen as output factors. South Korea’s NRE policy is founded on the “Act on the Promotion of the Development, Use, and Diffusion of New and Renewable Energy” and thus, aims to achieve the purpose of this law, that is, “the development, dissemination and promotion of usage of new and renewable energy and the promotion of new and renewable energy industries.” In reflection of this focus on the NRE industry, the 4th NRE Plan, announced in September 2014 under the leadership of the Ministry of Trade, Industry, and Energy, uses the number of firms and their employees in the NRE industry as the indexes for measuring the growth of South Korea’s NRE. The details of the NRE Plan make it evident that the goal of South Korea’s NRE policy is to increase the volume of NRE power generation and to promote the growth of the NRE industry, while at the same time, expanding the number of NRE-related firms for job creation. The output factors for this study follows the logic behind South Korea’s NRE policy, and the data for these output factors were compiled from the industry statistics database provided by NRE South Korea.

This study used MaxDEA Ultra 6.6 to analyze the data. Brief description of the data is shown in

Table 3. Descriptive statistics of the data (the values are averaged from 2009 to 2013).

	Input		Output
Variables	Investment for NRE Technology Development	Investment for NRE Dissemination	Number of Firms	Number of workers	Volume of Power Generation
Unit	One million KRW	One million KRW	One unit	One person	TOE
Photovoltaic	436,170	74,540	88	8335.6	213,415
Solar Heat	30,165.2	4980	16.8	210.4	28,286.4
Geothermal	23,531.6	6560	13	288.4	51,128.8
Bio	12,976.2	15,420	49.6	807.4	1,038,324.2
Wind Power	18,330.4	32,520	35.8	2272	188,708.6
Fuel Cell	47,394.8	28,800	10	388.2	65,962

below.

4. Results and Discussion

4.1. Static Analysis: DEA

In this study, six energy types (photovoltaic, solar heat, geothermal, bio, wind power, and fuel cell) for the years from 2009 to 2013 have been chosen as the DMUs, making a total of 30 DMUs. CCR-DEA was conducted using investment for NRE technology development and investment for NRE dissemination as input factors and the number of firms, the number of employees, and the volume of power generation as output factors. The DEA results for relative efficiency are illustrated in

Table 4. Results of the CCR-DEA.

	2009	2010	2011	2012	2013
Photovoltaic	0.6555	0.8143	0.9936	0.8117	0.8719
Solar Heat	0.9967	0.4239	0.6705	0.7931	0.9456
Geothermal	0.5715	0.2977	0.3563	0.2434	1.0000
Bio	1.0000	0.8271	0.8962	0.8297	1.0000
Wind Power	0.8010	0.9353	0.8790	0.8170	1.0000
Fuel Cell	0.0740	0.0998	0.1306	0.2084	0.3008

, where the DMU is considered relatively more efficient if its value is closer to 1. For an easier comparison of relative efficiencies, the values in Table 4 were plotted on a graph, which is shown in Figure 3 below.

Figure 3 clearly indicates that in 2012, NRE policies for four energy types (photovoltaic, geothermal, bio, and wind power) experienced a decrease in efficiency, while those of fuel cell and solar heat showed an increase. As 2012 was the year when the shift from FIT to RPS was made, this decrease in efficiency may be interpreted as a short-term drop in efficiency occurring while the NRE industry was adopting to the transition. The relative efficiencies of all NRE policies by energy type, including that of fuel cell and solar heat, increased in 2013, giving weight to this interpretation and showing that the effect of the policy shift has stabilized within South Korea’s NRE industry.

4.2. Dynamic Analysis: MI

As DEA has limitations in analyzing the dynamic changes in efficiency, this study uses the results of the DEA as a supplementary method and focuses on Malmquist analysis to see what effects South Korea’s policy shift had on the country’s NRE industry by energy type over time.

Table 5. EC, TC, MI of South Korea’s NRE Policy from 2010 to 2013.

Year	EC	TC	MI
2010	0.9433	0.8731	0.8236
2011	1.2692	0.9384	1.1909
2012	1.0275	0.8531	0.8765
2013	1.1540	1.4091	1.6261
Mean	1.0915	0.9962	1.0874

tabulates the results of the Malmquist analysis on South Korea’s NRE policy by year, and shows that the mean values for MI and EC increased slightly to 1.0874 and 1.0915, respectively, while that for TC decreased slightly to 0.9962. Figure 4 visualizes the pattern of change in the EC, TC, and MI over time. A very clear pattern can be observed where, in 2012, EC and TC decreased, which led to a large drop in MI, which then recovered in 2013. Again, as in the results of DEA illustrated above, this can be interpreted to reflect the occurrence of the policy shift in 2012. The introduction of a new policy naturally requires a transition period for the objects of change, in this case the NRE power producers and the NRE market as well as those who implemented the change, namely the South Korean government and the government employees in charge. Thus, during the policy transition period, it can be difficult to achieve the same amount of efficiency even with the same amount of input. However, after the transition period, unlike FIT, it can be seen that RPS’ mechanism of forcing NRE power producers to meet the mandatory supply ratio came into effect despite the drop in government subsidy. Thus, as expected, the overall NRE efficiency increased after 2013, as demonstrated in Figure 4.

Another important trend observed from Figure 4 is that, around 2012, the factor that drove the efficiency (MI) of NRE policy changed from EC to TC. As MI is the product of EC multiplied by TC, the similarity in the pattern of EC and MI before 2012, that is, in 2010 and 2011, implies that EC was the influential factor for the changes in MI during these two years. However, after 2012 and into 2013, we can see that TC, not EC, was influencing MI. Thus, the policy shift from FIT to RPS consequently shifted the driver behind NRE efficiency from EC to TC. The reason behind this change can be interpreted as follows: under FIT, the government’s focus was on maintaining a level of technical efficiency and subsidy under a consistent policy for each NRE type which gave little motivation to develop new management strategies or policies. When RPS was implemented in 2012, the effect was such as though a new technology was introduced. Two reasons can be given for this effect. First, as illustrated above in the comparison between FIT and RPS, the introduction of RPS enabled the South Korean government to set and manage NRE supply goals with more ease. Second, the competition among NRE power producers further enhanced the efficiency of the RPS policy.

Table 6. EC, TC and MI for each NRE type under FIT.

NRE Policy	EC	TC	MI
Photovoltaic	1.0000	1.2125	1.2125
Solar Heat	1.0017	0.8001	0.8015
Geothermal	1.0076	0.8048	0.8109
Bio	1.0000	0.8335	0.8335
Wind Power	1.0940	0.9894	1.0825
Fuel Cell	1.5541	0.8541	1.3273
Mean	1.0942	0.9052	0.9904

and

Table 7. EC, TC and MI for each NRE type under RPS.

NRE Policy	EC	TC	MI
Photovoltaic	0.9449	0.9932	0.9385
Solar Heat	1.0000	1.2328	1.2328
Geothermal	1.2738	1.2566	1.6007
Bio	1.0000	0.8638	0.8638
Wind Power	1.0000	1.2922	1.2922
Fuel Cell	1.3848	1.0113	1.4005
Mean	1.0889	1.0964	1.1938

illustrate the EC, TC, and MI for each NRE type, the former for years 2010 and 2011 under FIT and the latter for years 2012 and 2013 under RPS.

Based on Table 6 and Table 7, Figure 5 compares the MI of NRE policy by energy type for the periods under FIT (A) and RPS (B) using TC as x-axis and EC as y-axis. Here, a clear transition can be found where the change in policy from FIT to RPS moved geothermal energy, wind power, solar heat, and fuel cell from the upper left box (where EC > 1 and TC < 1) to the upper right box (where EC > 1 and TC > 1). In other words, except for bioenergy and photovoltaic energy, the change to RPS triggered technical change in the NRE types to increase TC. Referring back to Table 6 and Table 7, the increase in TC led to a higher MI. It should be noted that bioenergy alone maintained its position from (A) to (B) in Figure 5. This reflects the South Korean government’s reluctance in investing in bioenergy due to the high reliance on imports and the fact that it can be used in combination with fossil fuels. This reluctance was expressed openly by the South Korean government, especially from the beginning of the policy shift to 2014, through consistent signals that it will limit the production of bioenergy as investment in this area will not further its NRE policy goals to promote growth in the national NRE industry [26].

In the case of photovoltaic energy, both EC and TC decreased to result in a diagonal shift from the upper right to the lower left box. South Korea’s RPS manages the amount of obligatory power production for photovoltaic energy independently from that of non-photovoltaic power production. Thus, even if photovoltaic power was produced beyond the mandatory amount, the excess in production is not counted towards non-photovoltaic power production. Because of this segregated management, the introduction of RPS led to a decrease in the incentive for NRE power producers to produce more photovoltaic power beyond the minimum amount. This has been a recurring issue for the solar energy industry, which has been arguing for an increase in the mandatory supply ratio for photovoltaic power and the transfer of excess photovoltaic power production towards the mandatory non-photovoltaic power supply ratio [27]. In this context, the movement of photovoltaic energy shown in Figure 5 can be interpreted as follows: The efficiency of South Korea’s photovoltaic energy was maintained at a high level thanks to the large government subsidy under FIT, but decreased with the change to RPS when there was no more subsidy coupled with a separate mandatory supply ratio leading to a segregation from other NRE sources.

To observe the chronological change in the MI of NRE policy by energy type in more detail, the cumulative values of EC, TC, and MI for each energy type was calculated under the assumption that the EC, TC, MI = 1 in 2009 and tabulated in

Table 8. Cumulative MI, EC, TC of NRE policy by energy type.

	NRE Policy	2009	2010	2011	2012	2013
Cumulative EC	Photovoltaic	1.0000	1.0000	1.0000	1.0000	0.8928
	Solar Heat	1.0000	0.5465	1.0033	1.0033	1.0033
	Geothermal	1.0000	0.6803	1.0152	0.7549	1.6472
	Bio	1.0000	1.0000	1.0000	1.0000	1.0000
	Wind Power	1.0000	1.1969	1.1969	1.1969	1.1969
	Fuel Cell	1.0000	1.5833	2.4153	3.8213	4.6318
Cumulative TC	Photovoltaic	1.0000	1.1722	1.4701	1.1771	1.4503
	Solar Heat	1.0000	0.7782	0.6402	0.6818	0.9730
	Geothermal	1.0000	0.8362	0.6477	0.5730	1.0228
	Bio	1.0000	0.7417	0.6947	0.4404	0.5184
	Wind Power	1.0000	0.9588	0.9789	0.9107	1.6346
	Fuel Cell	1.0000	0.8166	0.7295	0.6319	0.7461
Cumulative MI	Photovoltaic	1.0000	1.1722	1.4701	1.1771	1.2948
	Solar Heat	1.0000	0.4253	0.6424	0.6841	0.9762
	Geothermal	1.0000	0.5689	0.6576	0.4326	1.6850
	Bio	1.0000	0.7417	0.6947	0.4404	0.5184
	Wind Power	1.0000	1.1476	1.1717	1.0900	1.9565
	Fuel Cell	1.0000	1.2929	1.7618	2.4148	3.4555

.

The values of cumulative MI in Table 8 were plotted on a graph to visualize the change over time in Figure 6. The most notable trend in Figure 6 is that, over the four-year period, the MI of fuel cell energy increased at a considerable and consistent rate compared to those of other energy types. Referring back to Table 8, we can see that this significant growth was due to the consistent increase in EC.

Fuel cell energy is one of the three strategic NRE types chosen by South Korea’s Ministry of Knowledge and Economy (2008) for concentrated investment [28]. The trend in fuel cell shown in Figure 6 shows that the South Korean government’s focused efforts have been effective in promoting growth in the NRE. However, the trends found in the two other NRE types chosen by the South Korean government—wind power and photovoltaic energies—should also be noted. All three strategic industries show cumulative MIs above 1 in Figure 6, however, while fuel cell has been maintaining a consistent increase rate, wind power showed a slight downturn until 2012 followed by a growth-burst in 2013 and photovoltaic energy showed no significant change. Thus, the only NRE type that seems to have actually benefited from the South Korean government’s focus was fuel cell energy. In particular, the fact that the cumulative MI for photovoltaic energy failed to show growth over time, even with the additional government support for the NRE type through the implementation of a separate mandatory supply ratio under RPS, present doubts on the effectiveness of the government’s policy for photovoltaic energy.

5. Conclusions and Policy Implications

The results found in this study hold meaningful implications for predicting how the efficiency of NRE policy can change when a policy shift is made from FIT to RPS. In this regard, the present study can provide a useful guideline for governments whose NRE policy is FIT-based but who are considering a change to RPS due to the high financial burden FIT poses. The summary of the effect of the policy shift from FIT to RPS on South Korea’s NRE industry found in this study is as follows.

First, the DEA and MI results on the change over time in the efficiencies and productivities of NRE policy by energy type showed a temporary drop in 2012 when the policy shift from FIT to RPS was made. However, after the transition period between policies, RPS was observed to be effective in bringing technological change. Second, when plotting the MI of NRE policy by energy type before and after the introduction of RPS onto four quadrants using EC and TC as the axes, it could be seen that the TC of almost all energy types increased. However, in the case of photovoltaic energy and bioenergy, the South Korean government’s regulations and negative signals have led to a different trend from other energy types. Lastly, the results for cumulative MI showed that, among the three focal areas strategically supported by the South Korean government (photovoltaic, wind power, and fuel cell), the concentrated efforts were only effective in terms of fuel cell, and had almost no effect on photovoltaic and wind power energies.

One of the limitations of this study is that it only looked at the case of South Korea to analyze the impact of NRE policy shift. Japan, Italy, and Belgium also had shifted their NRE policies similarly to South Korea, and the study of their policy shifts and their effects on the efficiency of NRE policies will add deeper insights to this study. Also, while this study looked at only the effect of the shift from FIT to RPS, further implications may have been gained from investigating the case where RPS was changed to FIT as in the case of Japan. It is suggested for future studies to pursue investigation on these other countries for a greater understanding of how a policy shift can affect the efficiency of NRE policy.

The other limitation faced by this study is the time period put under analysis. While it was originally the plan to include the time period from three years before the policy shift in 2012 (years 2009, 2010, 2011) and three years afterwards (years 2012, 2013, 2014), the data for the year 2014 was unavailable at the point in time this study was conducted. Thus, only the years 2012 and 2013 were included for the years after the policy shift. Future studies will be able to look into how RPS has influenced the NRE industry over a longer term by including the data for 2014 and 2015 to the present analysis.