Applied Energy 137 (2015) 301–309

Contents lists available at ScienceDirect

Applied Energy

journal homepage: www.elsevier .com/locate /apenergy

Korean public’s perceptions on supply security of fossil fuels:A contingent valuation analysis

http://dx.doi.org/10.1016/j.apenergy.2014.10.0160306-2619/� 2014 Elsevier Ltd. All rights reserved.

Abbreviations: CV, contingent valuation; DBDC, double-bounded dichotomouschoice; IEA, International Energy Agency; KRW, Korean won; LNG, liquefied naturalgas; NOAA, National Oceanic and Atmospheric Administration; USD, United Statesdollars; WTP, willingness to pay.⇑ Corresponding author. Tel.: +82 31 420 2103; fax: +82 31 420 2161.

E-mail addresses: jinsookim@hanyang.ac.kr (J. Kim), jihyokim@keei.re.kr(J. Kim).

1 Tel.: +82 2 2220 2241; fax: +82 2 2281 7769.

Jinsoo Kim a,1, Jihyo Kim b,⇑a Department of Natural Resources and Environmental Engineering, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul 133-791, Republic of Koreab Energy Demand Management Division, Climate Change Policy Research Group, Korea Energy Economics Institute, 132 Naesonsunhwan-ro, Uiwang-si, Gyeonggi-do437-713, Republic of Korea

h i g h l i g h t s

� This study investigates the Korean public’s perceptions of the fossil fuels.� The public’s WTP will increase if their understanding of the policy is enhanced.� Electricity charge is a promising source of funding for energy security policies.

a r t i c l e i n f o

Article history:Received 3 June 2014Received in revised form 9 September 2014Accepted 4 October 2014

Keywords:Energy securityFossil fuelPublic’s perception, willingness to payContingent valuation

a b s t r a c t

The supply security of fossil fuels is one of the most important policy issues in South Korea. This isbecause of South Korea’s economic condition, geopolitical status, and lack of energy resources. This studyinvestigates the Korean public’s perceptions of the supply security of fossil fuels using the contingent val-uation method. It derives the Korean public’s willingness to pay (WTP) for policy measures that enhancethe supply security of fossil fuels, such as overseas energy development and strategic reserves. It alsocompares the WTP in terms of an increase in taxes on gasoline or diesel with the WTP in terms of anincrease in the tax on electricity. The results demonstrate that the WTP will increase if the Korean public’sunderstanding of the importance of policy measures that boost the supply security of fossil fuels as wellas the role of government in implementing these measures are enhanced. The comparison of the WTPshows that the Korean public will be less against an increase in the tax on electricity than an increasein the tax on gasoline or diesel. This is so even though the WTP amount for a tax increase on electricityis smaller than that for a tax increase on gasoline or diesel.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction (IEA) [3] clarified that energy security consists of a physical

The importance of energy security has resurfaced recentlybecause of the increased volatility of energy prices, the scarcityof fossil fuels, climate change, and geopolitical supply tensions.Energy security is generally defined as the uninterrupted availabil-ity of energy sources at an affordable price [1–3]. However, it is dif-ficult to precisely define energy security because it is very contextand perspective dependent [4,5]. The International Energy Agency

unavailability component and a price component and discussedthat the relative importance of both components differs accordingto the market structure in a country. Kruyt et al. [5] also pointedout that the definition of energy security can differ across countriesand time periods. Thus, it is important to discuss policies forenhancing energy security in the context of a specific country.

In South Korea (hereafter Korea), energy security is defined asthe ‘‘availability of a stable supply of energy at an affordable price’’[6], following the IEA’s definition [7]. Like many other govern-ments, the Korean government has also implemented several pol-icies to enhance energy security. These include renewable energydevelopment, energy efficiency improvement, energy systemreform, and supply securement of fossil fuels. However, supplysecurement of fossil fuels has been considered one of the mostimportant policies on energy security in Korea for the followingthree reasons. First, fossil fuels such as coal, oil, and liquefied

2 This master plan has been renewed every three years. Now, in 2014, the fifth planis about to be announced.

302 J. Kim, J. Kim / Applied Energy 137 (2015) 301–309

natural gas (LNG) accounted for 85.2% of the total primary energyconsumption in 2012 [8], and fossil fuels largely had to beimported since Korea lacks energy resources. Second, it is not easyto reduce the consumption of these fossil fuels, because most areconsumed for industrial materials or transportation [8]. Energy-intensive industries (such as petrochemicals and basic metals) thatuse fossil fuels as primary industrial materials are crucial to theKorean national economy. Conditions in Korea for cultivating cropsfor biofuel production are also unfavorable. Thus, supply disrup-tions and price change risks have a negative impact on the Koreannational economy. Third, a relevant geopolitical factor is that Koreacan be regarded as an island although it is located in the southernpart of the Korean Peninsula. This implies that Korea is vulnerableto supply interruptions in a crisis.

To internalize the externality occurring when the stable supplyof fossil fuels at affordable prices is threatened [2], the Korean gov-ernment has implemented several measures based on taxes col-lected from the purchases of petroleum products, such asgasoline and diesel [9]. These measures include overseas energydevelopment and strategic reserves; however, an insufficient bud-get has hindered the promotion of these measures. Motivated bythis policy problem, this study attempts to examine the Koreanpublic’s perceptions of the supply security of fossil fuels by apply-ing the contingent valuation (CV) method. This method enables usto analyze the Korean public’s willingness to pay (WTP) for supplysecurement of fossil fuels through overseas energy developmentand strategic reserves.

In particular, the following two research questions are investi-gated. First, what are the characteristics of the Korean public’s per-ceptions of the supply security of fossil fuels? To answer thisquestion, this paper discusses their awareness and opinions aboutsupply security of fossil fuels. It also identifies the attributes thathave statistically significant effects on the Korean public’s WTPfor such supply security of fossil fuels. Second, apart from taxingpetroleum products, are there other ways of funding policy mea-sures on the supply securement of fossil fuels? To answer thisquestion, this study compares the WTP for an increase in the taxon electricity charges with that for an increase in the tax on gaso-line or diesel. Although the Korean public has paid social expensesfor the supply security of fossil fuels, its perceptions on this issuehave not yet been investigated. This information will help to gainan insight into the Korean public’s perspectives on policy measuresdealing with the supply security of fossil fuels.

Many studies on national energy security have focused on pol-icy issues from the perspective of the government, the principalagent of policy measures on energy security. Gracceva and Zeniew-ski [10], Yao and Chang [11], Sharifuddin [12], and Portugal-Pereiraand Esteban [13] assessed the level of energy security of thenational energy system. Augutis et al. [14], Martchamadol andKumar [15], and Wu et al. [16] developed methodologies for mea-suring the level of energy security. Jun et al. [17] measured the costof energy security in terms of supply disruption and price volatil-ity, considering the degree of concentration in energy supply anddemand. Månsson et al. [18] and Gouveia et al. [19] assessed theeffects of renewable resources on the supply security of energy.In those studies, how energy users (i.e., the beneficiaries of thepositive externality of energy security) recognize the effect of animprovement in energy security has not received sufficient atten-tion. Thus, this study attempts to measure the effect of an improve-ment in energy security from the perspective of energy users.

Some studies have investigated national energy security fromthe perspective of energy users based on stated preference meth-ods, such as CV. Damigos et al. [20] derived Greek households’WTP for securing a continuous flow of gas to the Greek electricitysystem using a CV approach. Chou et al. [21] examined Europeanconsumers’ preferences for supply security of gas based on the

choice experiment method. Demski et al. [22] compared publicperceptions of energy security risks with those on climate changerisks in the United Kingdom, focusing on their level of concern withsuch risks. Li et al. [23] measured US households’ WTP for researchand development (R&D) investment in energy sources not relianton fossil fuels using the CV approach. Jensen et al. [24] examinedconsumers’ preferences for ethanol-blended fuel consideringenergy security in the United States using the contingent choicemodel. Jang et al. [25] investigated Koreans’ WTP for a reliableLNG supply through a policy measure for improving the naturalgas storage rate. Thus, these studies justify using the CV approachin this study. On the other hand, this study differs from those thatuse stated preference methods, because it attempts to measure thevalue of the supply security of fossil fuels through overseas energydevelopment and strategic reserves, which have been the primarypolicy measures for improving the supply security of fossil fuels inKorea.

This study contributes to the assessment of policy measures forsecuring the supply of fossil fuels using the public’s perspective.The supply security of fossil fuels has received relatively littleattention because it is a less important measure for enhancingenergy policy in many other countries. As shown for previous stud-ies, renewable energy development, energy efficiency improve-ment, and energy system reform are all alternatives to thesupply securement of fossil fuels in the sense that they all enhanceenergy security. However, the supply securement of fossil fuels iscritical for some Northeast Asian countries (e.g., Korea, Japan,and Taiwan) because of their economic conditions, geopolitical sta-tus, and lack of energy resources. Thus, this paper contributes tothe body of policy literature concerning energy security in thesecountries. In addition, this study contributes to existing work onsurvey-based valuation of energy security as it compares WTPswith different payment vehicles. The results will be useful forcountries where policy measures on energy security are fundedby taxes.

The remainder of this paper is organized as follows. Section 2discusses Korean policy measures on the supply security of fossilfuels. Section 3 demonstrates the survey design and the value-elic-iting method employed in this study. Section 4 describes the sur-vey results and responses to value-eliciting questions. Section 5presents the WTP estimation results and discusses the tworesearch questions. Section 6 concludes.

2. Policy measures on the supply security of fossil fuels in Korea

After the two oil crises of the 1970s, the Korean government hasimplemented two policy measures, i.e., promoting overseas energydevelopment and building up strategic reserves, to protect thecountry from price risks and supply disruptions. The importanceof these measures was reinforced during the period of high oilprices in the mid-2000s.

To promote overseas energy development, the Korean govern-ment legislated the ‘‘Act on Overseas Energy and Resources Devel-opment Business’’ [26] in 1983. It has also implemented the‘‘Master Plan for Overseas Energy and Resources Development’’[27] as a guideline as of 2001.2 Overseas energy development pro-jects have been conducted by national energy corporations (e.g.,Korea National Oil Corporation and Korea Gas Corporation) and pri-vate enterprises. The Korean government has supported overseasenergy development projects by giving loans or by issuing a sover-eign financial guarantee for debt. From 1981 to 2012, Koreannational energy corporations and private enterprises participated

J. Kim, J. Kim / Applied Energy 137 (2015) 301–309 303

in 359 overseas energy development projects, and almost 50% ofthese projects received governmental support. In addition, the gov-ernment suggested an index called the ‘‘self-sufficiency rate of oiland gas.3’’ Such government support contributed to increasing theself-sufficiency rate of oil and gas from 4.2% in 2007 to 13.8% in2012. The government set a goal of achieving a 40% self-sufficiencyrate for oil and gas by 2040 [28].

Most of the strategic reserves of fossil fuels have been stockedby the three national energy corporations: Korea National Oil Cor-poration, Korea Gas Corporation, and Korea Coal Corporation. Thepetroleum stockholding amount in 2013 was 135 million barrels.This amounts to 246 days of the prior years’ net oil imports forKorea and is two and a half times higher than the IEA stockholdingrequirements [29]. Although there are no specific targets for gasand coal stockholding, the government has tried to sustain a cer-tain level of gas and coal stocks for several decades. The liquefiednatural gas stockholding amount in 2012 was 1840 kton, about5% of the yearly natural gas demand [8]. The amount of coal stock-holding in the same year was 1628 kton, almost 70% of the yearlycoal demand [8].

These policy measures have been mainly funded by taxes onpetroleum products such as gasoline and diesel. On retail busi-nesses, the Korean government has imposed a KRW4 36 (USD0.03)5 sales tax per liter of petroleum products [30]. In addition tothis sales tax, the government has imposed a KRW 475 (USD 0.43)consumption tax per liter of gasoline and a KRW 340 (USD 0.31) con-sumption tax per liter of diesel on end consumers [31]. A part ofthese taxes has been used to support overseas energy developmentprojects and to build up strategic reserves of fossil fuels. Given thatthe average retail price was KRW 1924 (USD 1.76) per liter of gaso-line and KRW 1730 (USD 1.58) per liter of diesel in 2013 [32], taxesassociated with these two policy measures accounted for more than20% of the retail prices for these products.6 On the other hand, taxeson LNG products and electricity are much smaller than the taxes onpetroleum products and they are not even used to secure the supplyof fossil fuels.

To accomplish the target of a 40% self-sufficiency rate of oil andgas by 2040 and to maintain the amount of strategic reserves in theface of increasing fuel prices, it is necessary to find additionalfinancial resources. There are two alternative ways of doing this.The first is to increase taxes on petroleum products, such as gaso-line and diesel. This way is consistent with the current taxationsystem on energy products. However, it may induce public opposi-tion, considering the current tax rates. Actually, social concernsthat the tax rates on petroleum products are too high have beenraised. The second alternative is to impose a new tax on otherenergy products. Energy products responsible for fossil fuel con-sumption are possible candidates. Currently, taxes account for19% of the retail price of LNG products and 12% of electricitycharges but have not been used to support policy measures onthe supply security of fossil fuels [33]. Coal briquette is not taxedas it is mainly used by a low-income group. No tax has beenimposed on bituminous coal for electricity generation, either. Thus,the second alternative has been seriously discussed to balance taxequity among different energy products and to find an additionalfunding source for policy measures regarding the supply securityof fossil fuels [33].

3 The self-sufficiency rate of oil and gas is defined as the amount of oil and gasdeveloped by Korean agents divided by the total amount of oil and gas imports.

4 KRW stands for Korean won, USD for United States dollars.5 The average standard trading rate (KRW 1,095=USD 1.00) announced by the Bank

of Korea in 2013 was applied in this study.6 Besides the taxes mentioned here, there are other taxes on petroleum products. In

total, taxes account for approximately 50% of the retail price of gasoline and 40% ofthe retail price of diesel.

3. Survey design and value elicitation method

3.1. Survey design

This section analyzes the Korean public’s WTP to secure a stablesupply of fossil fuels resulting from both overseas energy develop-ment and strategic reserves. A CV approach is applied. Thisapproach elicits people’s preferences for non-market goods (publicgoods) by measuring their WTP for specified improvements ofthese goods [34]. In a CV survey (which is the most commonlyapplied among stated preference methods), each respondent isasked about his/her WTP for certain scenario [35]. It is easy touse a CV approach, because there are many references using CVand standardized manuals on such an approach. CV also accordswith microeconomic theory, because it directly measures Hicksianwelfare. However, this approach has been criticized, because itrelies on respondents’ statements that may not be reliable [36].Despite this limitation, the results from a CV approach have beenknown to be generally reliable and precise [37,38]. The applicationrange of the CV approach has been continuously extended, and thisapproach has recently been applied to measure the value of energysecurity.

Two payment vehicles are assumed in this study: the tax on theprices for petroleum products such as gasoline or diesel and the taxon electricity charges [20,39]. The reason behind the selection ofthe former payment vehicle is that most taxes on petroleum prod-ucts have been used to finance policy measures regarding the sup-ply security of fossil fuels. If the majority of the Korean publicagrees to pay more taxes for their gasoline or diesel purchases tosecure the supply of fossil fuels, additional financial resourcescan be raised within the current taxation system. That is, the gov-ernment can simply increase the tax rates on petroleum productsinstead of creating a new tax. The second payment vehicle wasselected because the current tax rate on electricity is lower thanthat on petroleum products. Although electricity generationaccounts for over 60% of fossil fuel consumption [8], taxes on elec-tricity have not been used to finance policy measures on the supplysecurity of fossil fuels. Taxing electricity also has the indirect effectof imposing a tax on bituminous coal and on LNG for electricitygeneration.

This study specifically investigates how respondents’ WTPs dif-fer by payment vehicles based on two different surveys, that is,survey A and survey B. In survey A, the payment vehicle is anincrease in the tax imposed on gasoline or diesel. The surveyselected 500 households from the national sample of owners ofpersonal vehicles aged 19–65. They were asked how much addi-tional tax per liter of gasoline or diesel they were willing to payto gain a supply security of fossil fuels through overseas energydevelopment and strategic reserves. In survey B, the payment vehi-cle is an increase in the tax on electricity. The target population inthis survey is significantly different from that in survey A becausewhether a respondent has a personal vehicle is not considered insurvey B. Survey B extracts 500 households from the national sam-ple of heads of households or housewives aged 19–65. They wereasked how much additional tax they would be willing to pay ontheir monthly electricity bills to gain the supply security of fossilfuels through overseas energy development and strategic reserves.Both samples were extracted by cluster sampling following themethods recommended by Bateman et al. [35]. The sample sizesare acceptable [34].

Except for the difference in payment vehicle, the questionnairesfor surveys A and B contain the same questions. The questionnaireconsists of three parts [35]. The first part contains attitudinal andgeneral knowledge questions on current policy issues with a focuson the supply security of fossil fuels. Research efforts to survey

304 J. Kim, J. Kim / Applied Energy 137 (2015) 301–309

perceptions of the supply security of fossil fuels in South Koreahave been minimal thus far. Therefore, these questions aredesigned to derive some information on public preferences aboutpolicies for securing the supply of fossil fuels. This preparesrespondents for the hypothetical scenario. The second part pre-sents the hypothetical scenario, the value-eliciting questions, andsome follow-up questions that are presented to identify therespondents with zero WTP and to find out why they have zeroWTP. The third part investigates the socioeconomic and demo-graphic characteristics of the respondents.

In presenting the hypothetical scenario, the respondents werefirst informed about the predicted share of fossil fuels in the totalprimary energy consumption of Korea in 2030. The current and tar-geted rates of self-sufficiency of oil and gas were demonstrated.Using this information, the respondents were told that overseasenergy development and strategic reserves would be required toachieve the target self-sufficiency rate. Based on this hypotheticalscenario, they were asked value-eliciting questions that about theirWTP for a gain in the supply security of fossil fuels. To develop arealistic WTP, questions about the average monthly expenses ongasoline or diesel for survey A and on electricity for survey B wereasked right before presenting the value-eliciting questions. Therespondents of survey A were asked how much they would be will-ing to pay in the form of fuel costs, whereas the respondents of sur-vey B were asked how much they would be willing to pay in theform of their electricity bill.

The value-eliciting questions are presented in accordance with adouble-bounded dichotomous choice (DBDC) model [40] format. Inthe DBDC model, the respondents the second dichotomous choicequestion that respondents are asked is dependent on theirresponses to the first question. A respondent is asked whether heor she would pay the first bid amount (Bi). He or she is then askedabout willingness to pay a second bid amount. If he or she responds‘‘yes’’ to the first bid, the amount of this second bid (denoted as BU

i )is some amount greater than the first bid ðBi < BU

i Þ. If he or sheresponds ‘‘no’’ to the first bid, the amount of the second bid (ðBL

i Þis some amount smaller than the first bid ðBL

i < BiÞ.Table 1 shows the bid amount by survey type. To determine the

bid amount, we gathered twenty responses in a preliminary sur-vey. The preliminary survey was also randomly sampled from theSeoul region. In the preliminary survey, three bid amounts weresuggested for the first bid, Bi. For survey A, KRW 500, 1000, and2000 were suggested; for survey B, the suggested amounts wereKRW 40, 80, and 160. Based on the results of the preliminary sur-vey, KRW 1000 was set as an initial bid amount for survey A, andKRW 80 for survey B.

3.2. CV data analysis

Both the DBDC model [40] and the spike model [41] are used toanalyze the CV data in this study. The DBDC model is based on therandom utility model [42] and can be written as

uðj;m; sÞ ¼ vðj;m; sÞ þ ej; j ¼ 0;1; ð1Þ

where j = 1 if the respondent values the supply security of fossilfuels through overseas energy development and strategic reserves

Table 1Bid amount by survey type.

BLi

Bi

Survey A KRW 40 per liter (USD 0.04 per liter) KRW 80 per literSurvey B KRW 500 per month (USD 0.46 per month) KRW 1000 per m

The payment vehicle is an increase in the tax on gasoline or diesel for survey A and anThe average standard trading rate (KRW 1095 = USD 1.00) announced by the Bank of Ko

and j = 0 if he or she does not. u(j, m; s) and v(j, m; s) representthe true utility function of the respondent and the utility functionobserved by the investigator, respectively. The true utility functionincludes the error term ej, which is independent and identically dis-tributed because it is not known to the investigator. m representsthe respondent’s income, and s denotes his or her socioeconomicstatus, such as income and age, and demographic characteristics,such as knowledge of and opinions on energy security

When a respondent is offered an amount of money B to increasethe supply security of fossil fuels, he or she accepts this offer if

vð1;m� B; sÞ þ e1 � vð0;m; 2Þ þ e0; ð2Þ

For the investigator, a respondent’s response is a random vari-able. Then, the probability that he or she accepts the offer is givenby

Pr½vð1;m� B; sÞ � vð0;m; sÞ � e0 � e1� ¼ Pr½C � B�¼ 1� FC ½B; h�; ð3Þ

where C denotes the respondent’s WTP and FC[B; h] is the cumula-tive probability distribution function of C with parameter vector h.This study assumes that FC[B; h] follows a logistic distribution.

To treat the responses of zero WTP separately, the spike modelis employed. Given a logistic distribution, the spike model assumesthat FC[B; h] has the following form:

FCðBÞ ¼ 0 if B < 0

¼ ½1þ expðaÞ��1 if B ¼ 0

¼ ½1þ expða� bBÞ��1 if B > 0;

ð4Þ

where a and b are parameters of FC[B; h]. When the DBDC and thespike models are combined, the sample can be divided into fivegroups. For each respondent i, the discrete indicator which tellsus which of the five groups the individual belongs to, is defined as

IYYi ¼ 1 if WTP � BU

i ð0 otherwiseÞIYN

i ¼ 1 if Bi �WTP < BUi ð0 otherwiseÞ

INYi ¼ 1 if BL

i �WTP < Bið0 otherwiseÞINN

i ¼ 1 if 0 < WTP < BLi ð0 otherwiseÞ

I0i ¼ 1 if WTP ¼ 0ð0 otherwiseÞ:

ð5Þ

Given a sample of N respondents, the log-likelihood function ofthis model is given by

ln L ¼XN

i¼1

IYYi ln½1� FCðBU

i Þ� þ IYNi ln½FCðBU

i Þ � FCðBiÞ� þ INYi

� ln½FCðBiÞ � FCðBLi Þ� þ INN

i ln½FCðBLi Þ � FCð0Þ� þ I0

i

� ln½FCð0Þ� ð6Þ

The maximization of Eq. (6) yields estimates for a and b. Specif-ically, a can be substituted with a ¼

Pkj¼1cjXj, where Xj(j = 1, . . ., k)

represents a vector of covariates. The calculation of the mean WTPand the spike follows Kim et al. [43].

BUi

(USD 0.07 per liter) KRW 160 per liter (USD 0.15 per liter)onth (USD 0.91 per month) KRW 2000 per month (USD 1.83 per month)

increase in the tax on electricity for survey B.rea in 2013 was applied.

J. Kim, J. Kim / Applied Energy 137 (2015) 301–309 305

4. Data description

The survey was conducted through personal interviews, follow-ing the recommendation of the National Oceanic and AtmosphericAdministration (NOAA) panel [44], and covered the entire region ofKorea. Sampling and fieldwork were conducted using a Koreanprofessional polling firm. The response rate was 100%. None ofthe responses was found to be inadequate and the sample statisticsare presented in Table A1 in the Appendix A.

4.1. Perceptions of energy security

At the beginning of the survey, several questions regardingattitudes to and general knowledge about the security of energysupply were asked. The first set of questions asked whether therespondents had heard of the terms ‘‘energy security’’ and ‘‘self-sufficiency rate of oil and gas.’’ Interestingly, 41.4% of the respon-dents for survey A and 37.0% of the respondents for survey Banswered that they had heard of the term, ‘‘energy security.’’Meanwhile, 22.0% of the respondents for survey A and 19.4% ofthe respondents for survey B answered that they had heard ofthe term, ‘‘self-sufficiency rate of oil and gas.’’ The proportion ofrespondents who were aware about energy security and self-suffi-ciency of fossil fuels was higher in survey A than in survey B. Thatis, direct consumers of gasoline or diesel have higher awarenessregarding supply security of fossil fuels relative to indirect con-sumers of fossil fuels. This result seems consistent with the factthat the Korean government has used taxes on petroleum productsto finance the supply security of fossil fuels.

The second set of questions tested the respondents’ knowledgeof the energy consumption structure in Korea. The first question isabout the most heavily consumed energy source, and the correctanswer is oil [8]. 67.6% and 64.6% in surveys A and B, respectively,answered the question correctly. The second question is about themost heavily consumed energy source in electricity generation,and the correct answer is coal [8]. The proportion of respondentswho provided the correct answer was 3.6% and 2.0% for surveysA and B, respectively. This result implies that the Korean publicdoes not know the energy-consumption structure in electricitygeneration. Furthermore, the respondents for both surveys A andB had similar knowledge about the energy consumption structurein Korea.

The third set of questions tried to elicit the respondents’ atti-tudes towards common issues about energy security; the resultsare shown in Table 2. The first three questions of this set aimedto determine which policy measure among renewable energydevelopment, technology research and development (R&D) in the

Table 2Opinions on the security of the energy supply.

[Unit:%] Strongly disagree Slightly disagree

Opinion 1: renewable energy will be able to meet the entire energy demand in KoreaSurvey A 21.0 41.4Survey B 22.6 42.0

Opinion 2: technology R&D in the energy field is necessary to enhance supply securitySurvey A 0.6 –Survey B 0.4 0.6

Opinion 3: overseas energy development is necessary to enhance energy security1.0 1.4 11.00.4 3.4 12.4

Opinion 4: energy security is necessary for economic growthSurvey A 0.6 0.8Survey B 0.4 1.0

The payment vehicle is an increase in the tax on gasoline or diesel for survey A and an0.2% Of the respondents for survey B did not respond to opinion 1.

energy field, and overseas energy development was consideredmost important by the Korean public. The final question of thisset aimed to find their opinion on the relationship between energysupply and economic growth. In total, 13.2% and 14.8% of therespondents for surveys A and B respectively agreed with opinion1 that renewable energy would be able to meet the entire energydemand in Korea. 95.6% and 97.4% of the respondents for surveysA and B agreed with opinion 2 that technology research and devel-opment (R&D) in the energy field is necessary for enhancing energysecurity. 86.6% and 83.8% of the respondents for surveys A and B,respectively, agreed with opinion 3 that overseas energy develop-ment is necessary for enhancing energy security. Finally, 95.4% and94.6% of the respondents for surveys A and B respectively agreedwith opinion 4 that energy security is necessary for economicgrowth. In sum, most of the respondents generally agreed on theneed to secure energy supply for continuing economic growth,but they were skeptical about the prospect of renewable energy.It is likely that the high level of agreement on the need for overseasenergy development resulted from this skeptical attitude towardsrenewable energy.

4.2. Responses to value-eliciting questions

According to Eq. (5), the possible answers to the value-elicitingquestions can be categorized into five types: yes/yes, yes/no, no/yes, no/no, and zero. The response ‘‘yes/yes’’ means that theanswers to both the first and second bids are ‘‘yes’’; all the otherresponses are similarly defined. Table 3 summarizes the responsesfor the five groups according to the type of survey. It is noticeablethat there are huge differences in the proportions of the respon-dents stating zero WTP between the two samples. This differenceis attributed to differences in perceptions of the payment vehicles.For example, total taxes account for approximately 50% of the retailprice of gasoline whereas they accounts for only about 12% of theretail price of electricity [33]. Thus, it is likely that the Korean pub-lic is more against raising the tax on gasoline or diesel relative tothe tax on electricity.

To find why respondents stating zero WTP refuse to pay for again in the supply security of fossil fuels through overseas energydevelopment and strategic reserves, a follow-up question wasasked. These respondents were asked to choose among the follow-ing five possible reasons behind their refusal to pay:

1. I cannot afford to pay an additional energy price.2. I do not feel the necessity for overseas energy development and

strategic reserves.3. The government has to find other financial resources.

Neutral Slightly agree Strongly agree

24.4 10.8 2.420.4 11.2 3.6

3.8 28.6 67.01.6 30.0 67.4

44.6 42.0 67.041.8 42.0 67.4

3.2 30.2 65.24.0 27.6 67.0

increase in the tax on electricity for survey B.

Table 3Number of responses to the value-eliciting questions.

Response WTP level Survey A Survey B

Yes/yes WTP P BUi

35 (7.0%) 104 (20.8%)

Yes/no Bi 6WTP < BUi

53 (10.6%) 113 (22.6%)

No/yes BLi 6WTP < Bi 48 (9.6%) 66 (13.2%)

No/no 0 < WTP < BLi

38 (7.6%) 17 (3.4%)

Zero WTP = 0 326 (65.2%) 200 (40.0%)

The payment vehicle is an increase in the tax on gasoline or diesel for survey A andan increase in the tax on electricity for survey B.

306 J. Kim, J. Kim / Applied Energy 137 (2015) 301–309

4. I cannot trust the government’s policy.5. Firms are responsible for overseas energy development and

strategic reserves.

Responses regarding the reasons behind their refusal to pay aresummarized in Fig. 1. Each bar denotes the ratio of respondentschoosing a specific option regarding refusal to pay. Because therespondents were allowed to choose multiple reasons amongexamples that were not mutually exclusive, the sum of these ratiosexceeded 100%. For survey A, the reason most frequently indicatedwas ‘‘the government has to find other financial resources,’’ forwhich 57% of the respondent with zero WTP refused to pay. Thesecond most-frequent reason was ‘‘I cannot trust government pol-icy,’’ for which 48% of the respondents refused to pay. For survey B,both these reasons were also the most frequently stated, for which46% of the respondents refused to pay. Furthermore, very fewrespondents refused to pay in both surveys because they did notfeel that overseas energy development and strategic reserves wererequired.

The respondents refusing to pay for the first and second reasonsare classified as ‘‘legitimate zero’’ bids, whereas those refusing topay for the other three reasons are classified as ‘‘protest bids’’[35]. Protest bids occur whenever individuals who oppose or donot approve of the survey fail to respond, give invalid but positivebids (outliers), or place a zero value on a good that they actuallyvalue [45]. The latter category seems to describe the behavior ofrespondents classified as protest bids in our study. In particular,the reason they protested the bid seems to be their resistance tothe payment vehicle. Therefore, 228 respondents in survey A and131 respondents in survey B may actually be protesting againstthe payment vehicle being proposed. The result that the number

Fig. 1. Reasons for refusal to pay for t

of respondents protesting the bid is much smaller for survey B thanfor survey A supports our supposition that the Korean public is lessresistant to increasing the tax on electricity relative to raising thetax on gasoline or diesel.

Empirical studies suggest three principal means of treating pro-test bids [45]. The first is to drop them from the data set. This tech-nique reduces the downward bias caused by the inclusion ofprotest zero bids. A major side effect of this approach is the con-comitant loss of potentially useful information and the possibleintroduction of a self-selection bias. The second method is to treatprotest bids as legitimate zero bids and include them in the dataset[46]. This technique assumes that the strict valuation of the goodcannot be separated from public policy issues associated with itsprovision, such as the payment method for the good being offered[45]. The third is to assign protest bidders mean WTP values basedon their socio-demographic characteristics relative to the rest ofthe sample group. If the protest zero bidders’ socio-demographiccharacteristics are similar to those of the general population, theremay be a basis for extrapolating the mean sample WTP to the pop-ulation as a whole [47]. One of the purposes of this study is to com-pare the two different proposed payment vehicles in investigatingthe Korean public’s WTP for the supply security of fossil fuels.Thus, we assume that the payment vehicle itself cannot beexcluded from the commodity being assessed and conclude thatthe second technique of addressing protest bids is appropriate.Thus, regarding the protest bids as legitimate zeroes entailedincluding the respondents who protested bids.

5. Empirical results and discussions

5.1. WTP estimation results

The set of respondents’ perceptions of and opinions about thesupply security of energy, their socioeconomic status, and theirresponses to the value-eliciting questions was examined by apply-ing the maximum likelihood estimation approach. Table 4 showsthe definitions and sample statistics for explanatory variables inthe WTP models. Table 5 summarizes the estimates of the meanWTP for the supply security of fossil fuels through overseas energydevelopment and strategic reserves. It also shows the spike thatindicates the proportion of respondents with zero WTP, and theeffects of the explanatory variables on the WTP. The delta method,which enables the derivation of an approximate probability distri-

he supply security of fossil fuels.

Table 4Names, definitions, and sample statistics of explanatory variables.

Variable Definition Sample mean

Survey A Survey B

Awareness 1 Awareness of the term ‘‘energy security’’ (0 = no; 1 = yes) 0.414 (0.493) 0.370 (0.483)Awareness 2 Awareness of the term ‘‘self-sufficiency rate of oil and gas’’ (0 = no; 1 = yes) 0.220 (0.415) 0.202 (0.412)Knowledge 1 Knowledge of the overall energy mix in South Korea (1 = yes; 0 = no) 0.676 (0.468) 0.646 (0.479)Knowledge 2 Knowledge of the energy mix in the electricity generation of South Korea (1 = yes; 0 = no) 0.472 (0.500) 0.482 (0.500)Opinion 1 Attitude towards the opinion that renewable energy will meet the entire energy demand in South Korea

(1–5: 1 = mostly disagree; 5 = mostly agree)2.322 (1.000) 2.306 (1.058)

Opinion 2 Attitude towards the opinion that technology R&D is necessary for enhancing energy security (1–5:1 = mostly disagree; 5 = mostly agree)

4.614 (0.621) 4.634 (0.591)

Opinion 3 Attitude towards the opinion that overseas energy development is necessary for enhancing energysecurity (1–5: 1 = mostly disagree; 5 = mostly agree)

4.252 (0.783) 4.216 (0.819)

Opinion 4 Attitude towards the opinion that energy security is necessary for economic growth (1–5: 1 = mostlydisagree; 5 = mostly agree)

4.252 (0.783) 4.216 (0.819)

Expense Type A: monthly gasoline expense; Type B: monthly electricity charge 4.540 (1.604) 4.600 (1.651)Age Respondent’s age (1–4: 1 = 20s; 4 = 50s and 60s) 3.086 (0.879) 2.910 (0.974)Sex 1 = man; 2 = woman 1.496 (0.500) 1.514 (0.500)Education Education (1 = below middle school graduate; 2 = high school graduate; 3 = college student;

4 = bachelor’s degree holder; 5 = masters’ or doctoral degree holder)2.832 (1.100) 2.834 (1.184)

Householder 0 = no; 1 = yes 0.760 (0.428) 0.668 (0.471)Early adopter Tendency to buy a new product (1–5: 1 = the lowest; 5 = the highest) 2.648 (1.063) 2.582 (1.137)Income Monthly household income (1–10: 1 = below KRW 1 million; 10 = over KRW 10 million) 5.870 (1.817) 5.518 (1.870)

The payment vehicle is an increase in the tax on gasoline or diesel for survey A and an increase in the tax on electricity for survey B.Standard deviations are described in parentheses.

7 This is shown by a negative estimate of the coefficient on the ‘‘Early adopter’’variable, and positive estimates of the coefficients on the ‘‘Income’’ and ‘‘Education’’variables.

J. Kim, J. Kim / Applied Energy 137 (2015) 301–309 307

bution for a function of an asymptotically normal statistical esti-mator from knowledge of the limiting variance of that estimator[48], was applied to compute the means and standard errors forthe WTP and for the spike. The 95% confidence intervals for themean WTP and for the spike are calculated by using the Monte Car-lo simulation technique developed by Krinsky and Robb [49]. Theestimates of the mean WTP and the spike are significant at the1% level.

The mean WTP in survey A, where the payment vehicle is anincrease in the tax on gasoline or diesel, is KRW 37 (USD 0.03)per liter. This WTP is equivalent to KRW 4519 (USD 4.13) on amonthly basis. The mean WTP in survey B, where the paymentvehicle is an increase in the tax on monthly electricity charges, isKRW 1130 (USD 1.03). The spike, that is, the probability of statingzero WTP, is 0.648 for survey A and 0.380 for survey B. The esti-mates for the spike are quite similar to the proportions of respon-dents stating zero WTP in both samples.

The WTP in survey A is higher than in survey B, although theproportion of respondents stating zero WTP is also higher in surveyA than in survey B. At first glance, this casts doubt on the anchoringeffect [50], which hypothesizes that the WTP difference in the twosurveys results from the fact that the initial bid in survey A ishigher than the initial bid in survey B. The initial bid in survey Ais equivalent to KRW 9818 (USD 8.97) on a monthly basis, whichis about ten times the initial bid in survey B. However, the degreeof the bias from the anchoring effect is not serious in this study.First, this difference itself is likely to reflect the higher WTP interms of an increase in the tax on gasoline or diesel as these bidamounts were determined by the preliminary survey. Second, thegap between the WTP in the two surveys is smaller than the gapbetween the initial bids in the two surveys. The WTP in survey Ais about four times the WTP in survey B, and the number of respon-dents with the answer ‘‘no/no,’’ which stated that they were will-ing to pay less than the lower bid but more than zero in surveyA, is more than twice the corresponding number in survey B. Thisresult implies that the respondents stated their WTP according totheir true preferences. Therefore, although the anchoring effectmight bias our results on WTP, the degree of the bias is not severe.

Besides the anchoring effect, there is another cause of thedifference in the WTP between the two surveys. The fact that therespondents’ monthly expenses for gasoline or diesel are

approximately six times the monthly expenses for electricitymay explain this difference. The survey results reveal that therespondents’ average monthly expense for gasoline or diesel isKRW 235,750 (USD 215.30), whereas the average monthly expensefor electricity is KRW 41,000 (USD 37.44). An individual’s responseto a request for an additional payment may depend on the initiallevel of payment. The result that the WTP in terms of an increasein the tax on gasoline or diesel is 1.9% of the average monthly gas-oline expense, whereas the WTP in terms of an increase in the taxon electricity is 2.8% of the average monthly electricity charge, sup-ports the above supposition.

A positive sign on a coefficient associated with an explanatoryvariable generally indicates that respondents who are more favor-ably disposed towards a gain in the supply security of fossil fuelshave a higher WTP. Three explanatory variables concerning therespondents’ attitudes toward security in energy supply and socio-economic status are significant in survey A, whereas six explana-tory variables are significant in survey B. The number ofsignificant explanatory variables is smaller in survey A than in sur-vey B because of greater proportion of respondents refusing to payin survey A.

When the respondents were asked about their WTP for anincrease in the tax on gasoline or diesel, the respondents whoclaimed that there was less need for technology R&D for energysupply security (Opinion 2) and more need for overseas energydevelopment (Opinion 3) stated a higher WTP. When they wereasked about their WTP in terms of an increase in the tax on elec-tricity, the respondents who had awareness of the term ‘‘energysecurity’’ (Awareness 1) and who claimed that there was moreneed for energy supply security for economic growth (Opinion 4)stated a higher WTP. In addition, the respondents who were spend-ing less on electricity, those who were conservative in buying newhigh-tech products, and those with higher household incomes sta-ted a higher WTP.7 In both surveys, the more the level of education,the higher the WTP stated.

Table 5Mean WTP and covariate estimates.

Variables Estimates

Survey A Survey B

Mean WTP ***KRW 37 per liter ***KRW 1130 per monthUSD 0.03 per liter USD 1.03 per month(10.062) (16.044)

95% Confidence interval [KRW 31, KRW 45] [KRW 1006, KRW 1278]Spike ***0.648 (30.426) ***0.380 (17.51)

95% Confidence interval [0.606, 0.689] [0.339, 0.424]

Effect of covariates on WTPAwareness 1 0.079 (0.373) **0.458 (2.347)Awareness 2 0.057 (0.224) 0.274 (1.168)Knowledge 1 �0.276 (�1.256) 0.245 (1.262)Knowledge 2 �0.307 (�1.496) �0.012 (�0.068)Opinion 1 �0.019 (�0.198) �0.133 (�1.618)Opinion 2 *�0.377 (�1.801) 0.225 (1.232)Opinion 3 **0.365 (2.248) �0.022 (�0.186)Opinion 4 0.088 (0.444) **0.368 (2.293)Expense �0.023 (�0.365) *�0.102 (�1.842)Age 0.185 (1.448) 0.025 (0.215)Gender 0.174 (0.855) 0.049 (0.278)Education **0.252 (2.459) ***0.339 (3.760)Householder �0.224 (�0.966) �0.088 (�0.422)Early adopter �0.085 (�0.895) **�0.179 (�2.109)Income 0.091 (1.561) ***0.170 (3.122)Constant *�2.097 (�1.807) ***�3.200 (�2.988)Bid ***0.012 (12. 85) ***0.001 (15.86)Log likelihood �551.7 �683.4

The payment vehicle is an increase in the tax on gasoline or diesel for survey A andan increase in the tax on electricity for survey B.t-Statistics are shown in arentheses.

* Denote the significance at the 10% levels.** Denote the significance at the 5% levels.

*** Denote the significance at the 1% levels.

Table A1Sample characteristics of Survey A and Survey B.

Survey A (N = 500) Survey B (N = 500)

GenderMale 50.4% 48.6%Female 49.6% 51.4%

Age19–29 3.8% 7.8%30–39 23.2% 28.8%40–49 33.6% 28.0%50–65 39.4% 35.4%

Level of educationMiddle school education 6.6% 12.0%High school education 47.8% 40.8%College graduate or equivalent 43.0% 44.4%Post graduate 2.6% 2.8%

Level of monthly incomeBelow KRW 200 million 10.0% 15.2%KRW 200–300 million 29.8% 26.4%KRW 300–400 million 25.0% 28.2%KRW 400–700 million 29.8% 27.6%Above KRW 700 million 5.4% 2.6%

The payment vehicle is an increase in the tax on gasoline or diesel for survey A andan increase in the tax on electricity for survey B. Thus, the population of survey Aconsists of owners of personal vehicles aged 19–65 in Korea. The population ofsurvey B consists of heads of households or housewives aged 19–65 in Korea.

308 J. Kim, J. Kim / Applied Energy 137 (2015) 301–309

5.2. Discussion

This section discusses the two research questions raised in Sec-tion 1 in the light of the empirical results presented in the previoussections.

The first research question addresses the characteristics of theKorean public’s preferences for the supply security of fossil fuels.The most important findings regarding this question are summa-rized below. First, the Korean public was unfamiliar with the con-cept of supply security of fossil fuels, similar to the UK public [22].Less than half the respondents had heard of the term ‘‘energy secu-rity’’ and only about 20% of them had heard of the term ‘‘self-suf-ficiency rate of oil and gas’’ in spite of the latter being the key indexreflecting the supply security of fossil fuels. Second, the Koreanpublic tended to lack an understanding of the policy measuresdealing with the supply security of fossil fuels. It also tended to dis-trust the government. As seen in Fig. 1, almost half the respondentswith zero WTP refused to pay for reasons related to misunder-standing and distrust. Overall, the Korean public is likely to lackan understanding of the importance of policy measures, such asoverseas energy development and building strategic reserves. Itmay also be ill informed about the role of the government in imple-menting these measures. This supposition implies that if the Kor-ean public’s understanding can be enhanced, their WTP willincrease through a reduction in the ratio of respondents whorefuse to pay. The coefficients on Awareness 1 and Awareness 2shown in Table 5 also imply that an increase in the understandingof the Korean public will contribute to an increase in the WTP.

The second research question deals with possible fundingsources, apart from the tax on petroleum products, that could helpsecure the supply of fossil fuels. There are two contradictory piecesof evidence regarding this question. As seen in Table 5, the WTP foran increase in the tax on gasoline or diesel is greater than the WTPfor an increase in the tax on electricity. This result seems to

support the idea that it is better to increase the tax on petroleumproducts than to find other sources for taxation sources. However,we also found that the Korean public tended to resist increases inthe tax on gasoline or diesel. The ratio of respondents who werewilling to pay for the greater supply security of fossil fuels was34.8% in survey A and 60% in survey B. That is, the Korean publicis significantly more likely to be against an increase in the tax ongasoline than one in the tax on electricity. Also, electricity genera-tion is highly dependent on fossil fuels in Korea [8]. Thus, the taxon electricity is likely to be a promising funding source for policymeasures to increase the supply security of fossil fuels, when weconsider the respondents’ WTP and the role of electricity genera-tion in fossil fuel consumption.

6. Concluding remarks

This study investigates the Korean public’s perceptions of thesupply security of fossil fuels. The application of CV methodenables us to derive their awareness, knowledge levels, and opin-ions about the supply security of fossil fuels and to estimate theirWTP to implement two policy measures relating to the supplysecurity of fossil fuels, i.e., overseas energy development and build-ing strategic reserves. In particular, the WTP for an increase in thetax on gasoline or diesel is compared with that for an increase inthe tax on electricity. The empirical results demonstrate that theWTP will increase if the Korean public’s understanding of theimportance of policy measures dealing with the supply securityof fossil fuels as well as their knowledge about the role of the gov-ernment in implementing these measures is enhanced. The com-parison of the WTP across different payment vehicles shows thatthe Korean public will be less against an increase in the tax on elec-tricity than an increase in the tax on gasoline or diesel although theWTP in terms of electricity is smaller than the WTP in terms of gas-oline or diesel. This result implies that a tax on electricity is apromising source for funding policy measures boosting the supplysecurity of fossil fuels.

Acknowledgement

This work was supported by a research fund from HanyangUniversity (HY-2013).

J. Kim, J. Kim / Applied Energy 137 (2015) 301–309 309

Appendix A

See Table A1.

References

[1] APERC. A quest for energy security in the 21st century: resources andconstraints. Asia Pacific Energy Research Center; 2007.

[2] Bohi D, Toman M. Energy security: externalities and policies. Energy Policy1993;21:1093–109.

[3] IEA. Energy security and climate policy: assessinginteractions. Paris: International Energy Agency; 2007.

[4] Alhajji AF. What is energy security? Definitions and concepts. Oil, Gas & EnergyLaw; 2008.

[5] Kruyt B, van Vuuren D, de Vries H, Groenenberg H. Indicators for energysecurity. Energy Policy 2009;37:2166–81.

[6] Do H-J. Review of energy security and its strengthening plan in 21st century.Korea Energy Economics Institute; 2003.

[7] IEA. Towards a sustainable energy future. Paris: International Energy Agency;2001.

[8] KEEI. Yearbook of energy statistics. Gyunggi-Do: Korea Energy EconomicsInstitute; 2013.

[9] Kim H, Shin E-s, Chung W-j. Energy demand and supply, energy policies, andenergy security in the Republic of Korea. Energy Policy 2011;39:6882–97.

[10] Gracceva F, Zeniewski P. A systemic approach to assessing energy security in alow-carbon EU energy system. Appl Energy 2014;123:335–48.

[11] Yao L, Chang Y. Energy security in China: a quantitative analysis and policyimplications. Energy Policy 2014;67:595–604.

[12] Sharifuddin S. Methodology for quantitatively assessing the energy security ofMalaysia and other southeast Asian countries. Energy Policy 2014;65:574–82.

[13] Portugal-Pereira J, Esteban M. Implications of paradigm shift in Japan’selectricity security of supply: a multi-dimensional indicator assessment.Appl Energy 2014;123:424–34.

[14] Augutis J, Krikstolaitis R, Martisauskas L, Peciulyte S. Energy security levelassessment technology. Appl Energy 2012;97:143–9.

[15] Martchamadol J, Kumar S. An aggregated energy security performanceindicator. Appl Energy 2013;103:653–70.

[16] Wu G, Liu L-C, Han Z-Y, Wei Y-M. Climate protection and China’s energysecurity: win–win or tradeoff. Appl Energy 2012;97:157–63.

[17] Jun E, Kim W, Chang SH. The analysis of security cost for different energysources. Appl Energy 2009;86:1894–901.

[18] Månsson A, Sanches-Pereira A, Hermann S. Biofuels for road transport:analysing evolving supply chains in Sweden from an energy securityperspective. Appl Energy 2014;123:349–57.

[19] Gouveia JP, Dias L, Martins I, Seixas J. Effects of renewables penetration on thesecurity of Portuguese electricity supply. Appl Energy 2014;123:438–47.

[20] Damigos D, Tourkolias C, Diakoulaki D. Households’ willingness to pay forsafeguarding security of natural gas supply in electricity generation. EnergyPolicy 2009;37:2008–17.

[21] Chou W-J, Bigano A, Hunt A, la Branche S, Markandya A, Pierfederici R.Households’ WTP for the reliability of gas supply. Milan: Fondazione EniEnrico Mattei; 2011.

[22] Demski C, Poortinga W, Pidgeon N. Exploring public perceptions of energysecurity risks in the UK. Energy Policy 2014;66:369–78.

[23] Li H, Jenkins-Smith H, Silva C, Berrens R, Herron K. Public support for reducingUS reliance on fossil fuels: investigating household willingness-to-pay forenergy research and development. Ecol Econ 2009;68:731–42.

[24] Jensen K, Clark C, English B. Effects of demographic and attitudes on WTP forfuel import reductions through ethanol purchases. Southern agriculturaleconomics association annual meeting. Birmingham; 2012.

[25] Jang J, Lee J, Yoo S-H. The public’s willingness to pay for securing a reliablenatural gas supply in Korea. Energy Policy 2014;69:3–13.

[26] MOTIE. Act on overseas energy and resources development business; 2014.[27] MOTIE. 5th Basic plan for overseas energy and resources development; 2014.[28] KEEI, MOTIE. 2nd Energy master plan; 2014.[29] IEA. Closing oil stoak levels in days of net imports; 2014.[30] MOTIE. Act on petroleum products and susbstitute fuels business; 2014.[31] MOSF. Act on taxes on transportation, energy, and environment; 2013.[32] Opinet. Oil information; 2014.[33] NRCS. Comprehensive approach to reasonable energy price system:

examination and improvement of the energy price system Korean NationalResearch Council for Economics, Humanities, and Social Sciences; 2013.

[34] Mitchell R, Carson R. Using surveys to value public good: the contingentvaluation method. Washington, D.C.; 1989.

[35] Bateman I, Carson R, Day B, Hanemann M, Hanley N, Hett T. Economicvaluation with stated preference techniques: amanual. Massachusetts: Edward Elga; 2002.

[36] Bergstrom JC, Randall A. Resource economics: an economic approach tonatural resource and environmental policy. 3rd ed. Northampton: EdwardElgar; 2010.

[37] Brookshire DS, Thayer MA, Schulze WD, d’Arge RC. Valuing public goods: acomparison of survey and hedonic approaches. Am Econ Rev 1982;72:165–77.

[38] Loomis JB. Comparative reliability of the dichotomous choice and open-endedcontingent valuation techniques. J Environ Econ Manage 1990;18:78–85.

[39] Li H, Jenkins-Smith HC, Silva CL, Berrens RP, Herron KG. Public support forreducing US reliance on fossil fuels: investigating household willingness-to-pay for energy research and development. Ecol Econ 2009;68:731–42.

[40] Hanemann M, Loomis J, Kanninen B. Statistical efficiency of double-boundeddichotomous choice contingent valuation. Am J Agric Econ 1991;73:1255–63.

[41] Kriström B. Spike models in contingent valuation. Am J Agric Econ1997;79:1013–23.

[42] Hanemann W. Welfare evaluations in contingent valuation experiments withdiscrete responses. Am J Agric Econ 1984;66:332–41.

[43] Kim J, Park J, Kim H, Heo E. Assessment of Korean customers’ willingness topay with RPS. Renew Sustain Energy Rev 2012;16:695–703.

[44] Arrow K, Solow R, Portney PR, Leamer EE, Radner R, Schuman H. Report of theNOAA panel on contingent valuation. Resources for the Future. Washington,DC; 1993.

[45] Halstead J, Luloff A, Stevens T. Protest bidders in contingent valuation.Northeast J Agricult Resour Econ 1992;21:160–9.

[46] McGuirk A, Stephensen K, Taylor D. The use of Tobit analysis in the valuationof nonmarket resources. Blacksburg, VA: Virginia Polytechnic Institutes andState University; 1989.

[47] Walsh R, Loomis J, Gillman R. Valuing option, existence, and bequest demandsfor wilderness. Land Econ 1984;60:14–29.

[48] Greene W. Econometric analysis. 5th ed. New York: Prentice Hall; 2002.[49] Krinskty I, Robb A. On approximating the statistical properties of elasticities.

Rev Econ Stat 1986;68:715–9.[50] Green D, Jacowitz K, Kahneman D, McFadden D. Referendum contingent

valuation, anchoring, and willingness to pay for public goods. Resour EnergyEcon 1999;20:85–116.