Social Acceptance of Submarine Transmission Cables Under Excess Renewable Energy in South Korea: Lessons from Public Preferences

Abstract

This article examines public preferences for a proposed West Coast submarine high-voltage direct current (HVDC) transmission network in South Korea, installed by trenching and burying the cables in the seabed, which is essential for facilitating renewable energy integration and ensuring a stable electricity supply to the Seoul Metropolitan Area. The purpose of this study is to estimate South Korean households’ willingness to pay (WTP) for the proposed West Coast submarine HVDC network using contingent valuation (CV), thereby assessing its social acceptability amid renewable energy integration challenges. Employing a CV survey with a nationally representative sample of 1000 households conducted from late May to late June 2025, the research applies the one-and-one-half-bound spike model to address zero WTP responses and incorporates socio-demographic covariates to account for preference heterogeneity. The analysis estimates an average monthly WTP of KRW 1832 (USD 1.33) per household for the HVDC infrastructure. Results demonstrate statistically significant public support for the submarine HVDC project despite its high capital investment and potential electricity rate increases. These findings underscore notable consumer acceptance and provide valuable welfare insights for policymakers, reinforcing the prioritization of this project within South Korea’s energy transition framework. This paper contributes to the field of energy infrastructure valuation by advancing methodological approaches and offering policy-relevant recommendations for sustainable grid expansion.

1. Introduction

South Korea is undergoing a rapid energy transition, increasingly prioritizing renewable energy (RE) sources to meet its sustainability and climate goals. This shift necessitates substantial expansion and modernization of the national electricity transmission infrastructure. A key challenge arises from the spatial mismatch between electricity demand centers and renewable generation sites: major urban centers such as the Seoul Metropolitan Area (SMA), which accounts for approximately half of the national population and electricity consumption, are geographically distant from the optimal locations for RE development, predominantly situated in the southern regions with abundant sunlight and lower land costs. This geographic imbalance is further intensified by the peninsula’s unique regional disparities.

The southern region, where most RE facilities are concentrated, often experiences power surpluses leading to frequent generation output restrictions, impeding the full utilization of renewable resources. Meanwhile, the SMA, home to energy-intensive industries including semiconductor manufacturing, display production, and advanced data centers, demands a stable and resilient power supply. Hence, an efficient and reliable transmission system connecting the south to the SMA is vital for South Korea’s energy policy and economic security.

To address this, the Korean government has planned nine high-voltage (HV) transmission lines to be completed by 2038. Originally, these lines were to be constructed predominantly as 345-kV overhead transmission lines. However, significant local opposition and prolonged delays—including some projects postponed for up to two decades—have necessitated alternative solutions. Consequently, five lines will be terrestrial, while four will be implemented as submarine HV direct current (DC) (HVDC) cables, buried through trenching along the west coast seabed. This West Coast submarine HVDC transmission system, estimated at KRW 11.5 trillion (USD 8.3 billion) in investment and spanning approximately 620 km, offers enhanced efficiency by minimizing power loss over long distances and reducing land use demands.

However, compared to terrestrial 345-kV transmission lines, submarine HVDC incurs significantly higher construction costs. These costs will ultimately need to be covered through electricity rates, and so an increase in electricity rates may be unavoidable. At this time, consumer acceptance of higher electricity rate will be quite an important consideration. Therefore, the Korea Electric Power Corporation (KEPCO), the public enterprise responsible for the construction and operation of transmission lines, along with the South Korean government which oversees it, needs to evaluate consumers’ willingness to pay (WTP) for the implementation of the submarine HVDC transmission infrastructure buried in the seabed.

Empirical evidence—detailed in Section 2—underscores persistent public opposition to HV transmission infrastructure, manifesting as stigma effects on residential property values and broader social externalities. Hedonic pricing studies document localized discounts from overhead lines, including pylon proximity penalties in New Zealand suburbs [1], temporal stigma during project proposal and early operation phases in Australia [2], and network spillovers reducing United Kingdom (UK) home prices by approximately 3.9% [3]. Complementary stated preference (SP) research, encompassing contingent valuation (CV) documenting homeowners’ reluctance to pay for line removal despite perceived value erosion [4] and choice experiments advocating transparent, distance-decaying compensation regimes in Belgium [5], reinforces these findings. Such non-market costs emphasize the necessity of less intrusive alternatives like submarine HVDC systems, which bypass terrestrial not-in-my-backyard (NIMBY) barriers while facilitating renewable energy integration—a critical research gap addressed herein through household-level valuation in South Korea.

This research employs the CV method to estimate consumers’ WTP. The CV method collects data related to WTP through a well-structured survey targeting potential consumers. Once the data is analyzed using appropriate models, information about the average WTP is obtained. Specifically, this study deals with WTP data obtained through one-on-one interviews with 1000 households across South Korea, employing econometric models capable of accommodating zero WTP observations. As will be discussed later, respondents completed the survey without significant difficulty, and statistically significant results were obtained.

This study aims to assess South Korean households’ WTP for the proposed West Coast submarine HVDC transmission network—essential for alleviating renewable curtailments in southern surpluses and bolstering SMA supply reliability—via a CV survey of 1000 nationally representative households, analyzed with the one-and-one-half-bound (OB) spike model. Key conclusions reveal a statistically significant mean monthly WTP of KRW 1832 (USD 1.33), affirming public support despite elevated costs (KRW 11.5 trillion investment), yielding a benefit–cost ratio exceeding unity when extrapolated nationally. These insights validate policy prioritization of submarine HVDC within South Korea’s 11th Basic Plan for Electricity Supply and Demand, informing equitable rate designs and global precedents for variable renewable grid expansion.

This study advances the existing literature through three main contributions. First, it uniquely explores consumers’ WTP for transmission systems which has not been covered in the previous literature. As nations globally endeavor to address the challenges posed by inadequate transmission infrastructures through the deployment of HVDC transmission systems, the insights derived from this study may provide valuable reference points for policymakers and researchers in other countries. Second, several methodological refinements were made in applying the CV technique. To elicit WTP, the study employed the OB model introduced by Cooper et al. [6], aiming to improve efficiency and mitigate bias. Additionally, a spike model capable of handling a WTP of zero was incorporated. Third, discussions on various issues raised by respondents during the CV survey are presented. KEPCO and the government can use these insights effectively.

The subsequent sections of this paper are organized as follows. The next section provides a comprehensive review of the literature, followed by the derivation of key implications. Section 3 reports on methodology. More specifically, it begins with an explanation of the subject of evaluation, the West Coast submarine HVDC transmission system. Next, an overview of the CV method used in this study is provided, accompanied by a detailed explanation of its implementation process. The penultimate section summarizes the collected data and presents an analysis of the results, which are subsequently discussed. Several policy implications will also be addressed. The conclusion is contained in the final section.

2. Review of the Related Literature

The existing literature relevant to this study can be broadly divided into two strands: research focusing on the technical and economic characteristics of HVDC transmission and those evaluating the social and environmental values associated with transmission infrastructure.

2.1. HVDC Transmission: Technical and Economic Characteristics

With the rapid global expansion of RE deployment, the role of efficient and reliable long-distance transmission systems has become increasingly critical. Against this backdrop, an increasing number of studies have explored the technological benefits associated with HVDC transmission, broadly recognized as an effective approach for the transmission of large-scale power across extensive distances [7,8]. Two major technical features stand out. First, HVDC minimizes transmission losses while enhancing the stability of interconnected power systems. Unlike alternating current (AC) transmission, HVDC systems do not produce reactive power and are characterized by low conductor losses, which renders them especially advantageous for the integration of remote or offshore energy sources, including wind power [9,10].

In addition, because HVDC lines do not require electrical synchronization between AC networks, they reduce mutual interference and allow for precise control of active power flows, thereby strengthening the dynamic stability of the overall system [11]. Second, HVDC requires substantially less right-of-way than AC, which enhances its acceptability in environments sensitive to land use or impacts on the landscape. The smaller spatial footprint of HVDC cables makes them particularly attractive for submarine or underground transmission projects, contributing positively to both environmental and social acceptance [12,13].

Despite these advantages, HVDC deployment is constrained by its high upfront costs, primarily due to the expensive converters, high-voltage cables, and flow-control equipment required for operation [7,14]. Unsurprisingly, economic feasibility assessments have emerged as a distinct line of inquiry. Bloom et al. [15], for example, estimated the benefit–cost ratio of interconnecting the eastern and western interconnections in the United States using HVDC and reported values up to 2.9, underscoring its promising cost-effectiveness. Similarly, Wang et al. [16] provided empirical evidence that HVDC offers superior efficiency over HVAC for long-distance transmission. The key findings from previous empirical studies on the valuation of transmission infrastructure, including both environmental and social dimensions, are summarized and presented in

Table 1. Summary of findings from some previous studies for high-voltage direct current (HVDC) or transmission lines.

Sources	Subjects	Methods	Main Findings and Implications
Bloom et al. [15]	HVDC	Economic analysis	Benefit–cost ratio up to 2.9; supports large-scale HVDC grid integration.
Wang et al. [16]	HVDC	Technical-economic analysis	HVDC outperforms HVAC for long-distance transmission in both cost and efficiency.
Ge et al. [17]	HVDC	Life cycle cost (LCC) analysis	Identified significant annual operational cost reductions with ultra HVDC adoption.
Gul et al. [18]	HVDC	Case study of an actual project	Confirmed long-term economic sustainability of HVDC over high voltage alternating current (HVAC) in renewable energy transmission.
Tosatto et al. [19]	HVDC	Real event analysis	Emergency power control function can reduce grid security costs by up to 70%, highlighting HVDC’s added economic benefits.
Acaroğlu et al. [20]	HVDC	LCC analysis	Demonstrated cost effectiveness of HVDC in large scale offshore wind transmission and proposed efficient solar power dispatch strategies.
Harrison [21]	Transmission lines	Contingent valuation (CV)	Tourists are willing to pay (WTP) to preserve views; visual costs of overhead transmission lines are significant.
Anderson et al. [22]	Transmission lines	CV	Public holds negative perceptions and is WTP for line removal.
Ju et al. [23]	Transmission lines	Choice experiment (CE)	Highlights environmental costs and need for government response to public concerns.
Menges and Beyer [24]	Transmission lines	CV	Strong public preference for underground cables; moderate WTP confirmed.
Lambert et al. [25]	Transmission lines	WTP analysis	Residents are WTP an additional USD 14.69 per month for enhanced grid reliability; indicates strong public support for grid investment.
Shim et al. [26]	Transmission lines	CV	Residents show strong preference and positive WTP for underground lines.
Bond and Hopkins [1]	Transmission lines	Hedonic pricing (HP)	Households incur a negative external cost from nearby transmission infrastructure that is capitalized into housing prices.
Elliott and Wadley [2]	Transmission lines	Literature review	High-voltage transmission lines impose environmental stigma on nearby property values, especially during proposal and early operation, requiring planners to account for perceived risks beyond long-run data.
Callanan [4]	Transmission lines	CV	Buyers perceive transmission lines as reducing property values but are unwilling to pay for their removal.
Tang and Gibbons [3]	Transmission lines	HP	Construction of new overhead pylons reduces nearby property prices by about 3.9% for homes.
De Jaeger et al. [5]	Transmission lines	CE	Transparent compensation is shown to be essential for improving acceptance of grid expansion projects.

.

As presented in Table 1, studies of the economic benefits of HVDC installations are also found in the literature. Ge et al. [17] demonstrated significant operational cost savings from ultra HVDC adoption in China’s Jiangsu Province through a grid simulation. Gul et al. [18] showed that the Gwadar–Matiari HVDC line in Pakistan, designed for large-scale RE transfer, was more economically sustainable than AC alternatives. Tosatto et al. [19] found that using HVDC’s emergency power control capability in the Nordic grid could reduce system operation security costs by up to 70%. Likewise, Acaroğlu et al. [20] demonstrated the economic viability of submarine HVDC systems for integrating offshore wind power through a life-cycle cost analysis. Specifically, the analysis suggests that HVDC development is financially justifiable from both profitability and economic perspectives. Collectively, these studies affirm both the technical robustness and potential cost-effectiveness of HVDC solutions.

2.2. Valuation of Transmission Infrastructure: Environmental and Social Dimensions

The second line of inquiry investigates the environmental impacts—both costs and benefits—of transmission infrastructure, often using SP techniques like CV or discrete choice experiments (DCEs). These studies shed light on public attitudes toward overhead or underground lines, as well as WTP for infrastructure modifications. For instance, Harrison [21] applied CV to assess tourists’ WTP for preserving visual amenities affected by overhead transmission lines in Australia. Anderson et al. [22] demonstrated that residents express a significant WTP for the removal of overhead lines due to perceived visual, psychological, and health concerns.

Building on this, Ju et al. [23] employed choice experiments to quantify the environmental costs of overhead lines in South Korea, while Menges et al. [24] identified a distinct public preference for underground cables over overhead alternatives. More recent studies have continued this line of inquiry. Lambert et al. [25] delved into household WTP for grid reinforcement in Oklahoma, whereas Shim et al. [26] investigated South Korean residents’ WTP for the subterranean installation of high-voltage transmission lines.

Early hedonic analyses, such as those by Bond and Hopkins [1], examined residential sales in a Wellington, New Zealand suburb affected by HV overhead transmission lines (HVOTLs), revealing no significant overall price depreciation from line proximity but a statistically notable discount attributable to nearby pylons, highlighting localized visual and structural impacts. Elliott and Wadley [2] advanced a conceptual framework for “environmental stigma” arising from HVOTLs, positing that psychological aversion to perceived risks, visual blight, and land-use constraints manifests in property devaluation beyond quantifiable hedonic adjustments, particularly during proposal and early operational phases. Complementing these, Callanan [4] integrated CV with hedonic modeling, uncovering a disparity where homeowners’ stated WTP for HVOTL removal exceeded observed market discounts, suggesting underestimation of stigma in revealed preference data.

Contemporary spatial econometric approaches, exemplified by Tang and Gibbons [3], quantify HVOTL social externalities using UK house prices, estimating significant devaluation from proximity while accounting for network spillovers and peer effects (“friends electric”), implying broader neighborhood impacts than isolated hedonic gradients. De Jaeger et al. [5] critique traditional Belgian compensation regimes for undervaluing HVOTL-induced property losses, proposing refined distance-decaying schedules (e.g., 25% within 35 m for new 150+ kV lines) informed by hedonic evidence and stakeholder input to ensure equitable internalization of externalities. These studies collectively underscore the need for a policy integrating stigma-aware valuations, proactive mitigation, and dynamic compensation to balance energy infrastructure expansion with residential welfare.

2.3. Implications for the Present Study

Synthesizing insights from the two strands of literature yields three crucial implications for the present research, which also employs the CV method. First, while HVDC transmission is technologically advanced and economically promising, public acceptance is likely to depend not only on its technical merits but also on how its environmental and social impacts are perceived. In this context, the present study advances the existing literature by analyzing the preferences of South Korean citizens toward HVDC transmission lines. By applying a CV framework, the study not only captures public attitudes toward advanced transmission technologies but also provides empirical evidence on the social acceptability and non-market valuation of HVDC infrastructure. This focus is particularly meaningful, as few studies to date have investigated consumer preferences for HVDC in the South Korean context, despite its growing importance for integrating RE and supporting the national energy transition.

Second, valuation studies underscore that transmission projects entail significant non-market costs—particularly visual disamenities—that must be carefully addressed through undergrounding, aesthetic design, or compensatory measures. Given these circumstances, the South Korean government has prioritized the development of submarine HVDC transmission lines over conventional land-based options, aiming to alleviate opposition related to land use and visual impacts. However, despite the increasing policy emphasis and market growth of submarine HVDC, especially for offshore wind integration, there remains a notable lack of empirical studies employing CV to assess public preferences and social acceptability of submarine HVDC in the country. To fill this research gap, the current study employs the CV method to systematically investigate societal attitudes toward submarine HVDC projects among the country’s citizens. By doing so, it provides novel evidence directly relevant to both energy policy and infrastructure planning, enriching the international literature with insights specific to the country’s evolving energy transition landscape.

Third, integrating economic feasibility analysis with public preference data derived from CV can provide a more comprehensive assessment of HVDC’s long-term sustainability. Accordingly, as detailed in subsequent sections, this study endeavors to use the estimated WTP data to undertake a comprehensive cost–benefit assessment of the proposed HVDC transmission project. By integrating the social valuation obtained through the CV method with traditional economic and technical assessments, the analysis aims to offer a holistic evaluation of the project’s net societal benefits. This approach not only quantifies the monetary value that consumers assign to HVDC infrastructure but also facilitates evidence-based decision-making regarding infrastructure investments and policy priorities.

In sum, the literature calls for further empirical investigation of public preferences in the context of advanced transmission technologies. This study responds to that need by implementing a CV survey, thereby capturing the social dimensions of HVDC deployment that are often overlooked in technical or engineering-oriented analyses.

3. Materials and Methods

3.1. Method: CV

The evaluation of non-market goods can be approached through several SP techniques, including CV, DCE, and contingent ranking methods. The selection of an appropriate method is primarily influenced by the research context and the specific attributes of the good or service being evaluated, as well as the type of policy-relevant information required. Among these, CV is distinctive in its ability to derive people’s WTP for a clearly defined change in a non-marketed good by means of carefully designed surveys. In contrast to choice-based approaches, which infer preferences from selections among alternatives, CV directly requests respondents to state their monetary valuation, typically within hypothetical purchase frameworks or referendum-style questions [27].

The CV method offers several advantages that make it especially suitable for applied research [28,29,30]. First, it provides a direct and interpretable measure of WTP for a specific policy intervention or environmental improvement, thereby generating estimates that can be easily integrated into cost–benefit analyses. Second, CV captures total economic value, including non-use benefits such as existence, option, or bequest values, which often represent a substantial share of social welfare in environmental and energy-related contexts. Importantly, it remains applicable even when the beneficiaries are not direct users, as is common in cases involving biodiversity conservation, air quality enhancement, or climate mitigation.

Third, CV is particularly useful in referendum-type scenarios—e.g., “Would you support this renewable energy project at the stated cost?”—where it provides policymakers with empirically grounded evidence for decision-making, as opposed to methods that rely on trade-off assessments. Finally, CV is preferable when the good under consideration is unique, indivisible, or not easily decomposed into distinct attributes, making market analogies or attribute-based valuation approaches less appropriate.

While the CV method offers significant advantages in eliciting WTP for non-market goods, it is not without notable biases and concerns that require careful mitigation. A primary issue is hypothetical bias, where respondents tend to overstate their WTP in hypothetical scenarios compared to actual payments, as they face no real budget constraints. Starting point or anchoring bias arises particularly in dichotomous choice formats, with initial bid levels disproportionately influencing responses. Information bias can occur if survey details are incomplete or unbalanced, leading to misinformed valuations [31]. Additional challenges include strategic bias, where respondents may under- or over-report WTP to influence policy outcomes; embedding or part-whole bias, resulting in scope insensitivity (e.g., similar WTP for small vs. large environmental changes); and payment vehicle bias, varying with the specified payment method (e.g., taxes vs. donations). To address these, the present study employed cheap talk scripts to calibrate expectations, follow-up certainty questions to filter uncertain responses, and pre-testing to refine the scenario, thereby enhancing validity [32].

Although DCE and related approaches can produce detailed valuations of multi-attribute goods by disentangling WTP for specific features, they often impose higher cognitive demands on the respondents and are better suited for ranking or trade-off analyses rather than for estimating the total social value of a single, well-defined public good. Consequently, CV continues to serve as a primary tool for estimating aggregate welfare impacts and guiding binary project choices, particularly in the domains of energy infrastructure and environmental resource management. Overall, the method’s directness, relevance for policy evaluation, and ability to encompass both use and non-use values substantiate justify its widespread application in energy economics research.

This paper examines the conceptual linkage between social acceptance and WTP as addressed herein. Throughout the paper, WTP—elicited via CV—serves as an economic proxy for social acceptance of the West Coast submarine HVDC network, operationalizing public endorsement through revealed trade-offs between benefits and costs under hypothetical budget constraints. This approach, standard in non-market valuation (e.g., Arrow et al. [32]), quantifies welfare impacts encompassing use (e.g., reliability) and non-use (e.g., existence of low-impact grid) values, thereby gauging aggregate societal support. However, WTP does not exhaustively capture social acceptance’s multidimensionality, which, per socio-technical systems frameworks (e.g., Wüstenhagen et al. [33]), encompasses normative (trust, procedural justice), distributive (equity perceptions), and institutional (participation, governance) dimensions potentially influencing project legitimacy beyond economic signals. By explicitly acknowledging this distinction, the study positions CV-derived WTP as a robust, partial indicator—complementary to qualitative or multi-criteria assessments—preempting critiques while advancing applied policy valuation.

3.2. Procedure of Applying the CV

The four-stage methodological framework for conducting a CV survey is well-established in the academic literature and is widely recognized by practitioners. Building on these standards, the present study adopts a structured four-stage process: (1) determination of the good under valuation; (2) development of the survey instrument; (3) administration of the survey; and (4) analysis of the CV responses. This systematic procedure not only ensures methodological rigor but also improves the reliability and clarity of the inferred WTP estimates. As a result, the generated data offer a solid foundation to support evidence-based policy decision-making.

3.2.1. Stage 1: Determination of the Good Under Valuation

In the first stage, clear identification of the valuation object is essential to ensure respondents accurately comprehend the hypothetical market scenario and valuation context, minimizing cognitive ambiguity. In CV applications, the object of valuation is generally conceptualized as a change from a baseline (status quo) state to a target or policy-induced outcome, rather than the target state alone. For the present study, the baseline scenario is defined as the transmission of RE generated in the southwestern region of South Korea through the existing overhead 345-kV transmission lines to the SMA. This represents the current operational infrastructure and market conditions.

The target scenario, in contrast, envisages the transmission of the same RE generated in the southwestern region using an upgraded transmission system involving a West Coast submarine HVDC cable to the SMA. This shift reflects a significant technological and infrastructural enhancement aimed at improving grid reliability, reducing losses, and facilitating a greater penetration of RE. By framing the valuation in this comparative manner, respondents are directed to evaluate the incremental benefits of the enhanced transmission configuration relative to the existing arrangement, thus aligning with the theoretical standards of CV practice. This comparative framework facilitates meaningful elicitation of WTP for infrastructure, improvements enabling the use of domestically manufactured pumped-storage hydropower generation equipment within the broader RE transmission system.

3.2.2. Stage 2: Development of the Survey Instrument

To guarantee clarity, validity, and methodological rigor, the CV questionnaire in this study was organized into three sections. The introductory section provided respondents with the necessary policy context for the valuation exercise. Along with a concise definition of HVDC, the survey introduced the South Korean government’s plan to transmit electricity generated in the southwestern region to the SMA via the West Coast submarine HVDC transmission systems. In addition, respondents were asked about their prior knowledge of the issue of electricity oversupply in that region, thereby enabling subsequent analyses of heterogeneity based on differences in awareness. The second section, which formed the core of the survey, focused on eliciting respondents’ WTP for HVDC attributes. Three main design choices were incorporated to enhance the internal validity of the instrument:

• Target population: The survey sample was limited to household heads between the ages of 20 and 65 currently residing in South Korea, representing the economically active population. Those under 20 were excluded on the grounds that most were students, while individuals above 65 were largely retired and expected to have distinct consumption patterns.
• Payment vehicle: The payment vehicle was specified as an additional surcharge on the monthly electricity bill. This design choice was intended to enhance both the realism and the credibility of the valuation scenario by directly linking the payment mechanism to the good being assessed.
• Elicitation format: A closed ended dichotomous choice framework was adopted instead of open-ended questions, which are prone to biases such as strategic misreporting and hypothetical overstatement of value. Closed-ended formats better replicate actual market behavior and promote incentive compatibility [31]. Specifically, this study applied the OB dichotomous choice model proposed by Cooper et al. [6], which improves efficiency and statistical robustness of estimating WTP distributions.

Finally, the third section gathered demographic and socioeconomic information at both the individual and household levels, encompassing variables such as age, gender, educational attainment, and average monthly household income. These variables were later incorporated as covariates in econometric models to examine heterogeneity in WTP and to provide a richer understanding of consumer preferences. Overall, the tripartite structure of the questionnaire was deliberately designed to integrate essential policy context, apply a rigorous elicitation framework, and capture sociodemographic information. This design ensured that the resulting dataset would be both methodologically robust and highly relevant for policy analysis.

3.2.3. Stage 3: Administration of the Survey

The execution of the CV survey necessitated deliberate consideration of three critical components: the administering organization, the survey mode, and the sample size. The first element is administering organization. Rather than conducting the fieldwork directly, the present study outsourced the survey to a professional research firm, Research Prime Inc. located in Seoul, South Korea. This decision was guided by two considerations. First, the research team did not have the expertise necessary for large-scale sampling and structured interviewing. Second, the contracted firm possessed substantial expertise in conducting CV surveys, including the derivation of representative samples based on census data furnished by the Korea Ministry of Data and Statistics and the successful execution of multiple nationwide surveys each year. The interviewers employed by the firm were also well trained and familiar with administering CV instruments. Since the credibility of CV research critically depends on robust sampling and reliable implementation, the engagement of a specialized agency was deemed preferable, even though it entailed higher costs.

The second element is survey mode. Data were collected via face-to-face interviews conducted in respondents’ homes. Although relatively resource-intensive, this method offers two major advantages: it ensures that respondents are provided with sufficient information about the hypothetical scenario, and it allows interviewers to monitor respondents’ comprehension and guide the survey process in real time. Other modes such as mail or telephone surveys are less costly but typically face high non-response rates and sampling bias. More recently, online and mobile surveys have gained popularity [34]; however, concerns about representativeness and response quality rendered them less suitable for the present study.

The survey was administered exclusively via paper-and-pencil-assisted personal interviewing (PAPI) from late May to late June 2025 (fieldwork dates: 25 May–28 June 2025), utilizing trained enumerators from a professional research firm for door-to-door household visits. PAPI was selected over alternatives like computer-assisted telephone interviewing (CATI) or online modes for its superiority in complex CV scenarios: it facilitates visual aids (e.g., maps of HVDC routes), real-time clarification of hypothetical scenarios, and non-verbal cue detection to minimize yea-saying or fatigue biases. While CATI offers lower costs than PAPI, the latter was adopted in this study to ensure the collection of more reliable data, given its superior capacity to mitigate response biases through real-time interviewer–respondent interaction and visual aid delivery. Timing coincided with post-spring energy awareness campaigns, avoiding seasonal biases (e.g., summer peaks), with daily quotas ensuring representativeness per Korea Ministry of Data and Statistics strata. The response rate was 82%, with interviewer monitoring yielding <2% invalidations, affirming data quality.

The third element is sample size. Determining the appropriate sample size is essential for ensuring the statistical reliability and policy relevance of CV results [35,36]. A target of approximately 1000 respondents was adopted, reflecting a widely accepted standard in applied CV research. This sample size strikes a balance between statistical precision and cost effectiveness. Larger samples may slightly improve accuracy by reducing the standard error of the mean WTP estimates, but with diminishing returns relative to the added resource requirements.

Conversely, smaller samples risk an inflated variance, broader confidence intervals, and reduced credibility. Importantly, a sample of around 1000 participants allows for robust econometric modeling and subgroup analyses to capture heterogeneity across demographic and socioeconomic segments. The recommendation by the National Oceanic and Atmospheric Administration panel [32] that CV studies employ a minimum of 1000 respondents for environmental goods has been widely adopted by subsequent research and public agencies. Consistent with this guidance, many national level CV surveys in energy and environmental economics have relied on similar sample sizes, reinforcing their empirical and practical justification. In sum, outsourcing the survey to a professional firm, employing face-to-face interviews, and targeting a sample size of roughly 1000 respondents together ensured that the data collection process met the methodological standards necessary for producing reliable and policy relevant valuation outcomes.

As explained above, this target sample size of 1000 aligns with established guidelines and empirical precedents in CV research for national-level environmental and energy valuations. According to guidelines established by the Korea Development Institute (KDI)—a leading authority providing standardized protocols for CV implementation in South Korea—a sample size of 1000 is sufficient to faithfully represent the views of the majority of South Korean households when employing one-on-one in-person interviews and scientific probability sampling. This recommendation is predicated on the country’s national household population and desired precision for mean WTP estimates, validated through finite population correction formulas. Consistently, KDI-conducted nationwide CV surveys—across energy, environmental, and public goods domains—uniformly adopt n = 1000, as do analogous studies by other domestic institutions, including government-funded research institutes (e.g., Korea Environment Institute, Korea Energy Economics Institute) and regulatory agencies, ensuring comparability, statistical power for subgroup analyses, and policy credibility (e.g., KDI surveys on carbon pricing, RE acceptance [37]). This national standard aligns with international benchmarks like those of Arrow et al. [32] while accounting for South Korea-specific demographic and regional strata. Moreover, other studies dealing with infrastructure like grid upgrades or electric bus adopted a sample size of 1000, enabling precise mean WTP [23,25,26,38].

3.2.4. Stage 4: Analysis of the CV Responses

The fourth stage leverages advanced econometric models tailored to the data structure, including treatment of zero responses and other statistical nuances, ensuring that the derived welfare measures provide accurate and policy-relevant insights. This process not only ensures methodological rigor but also enhances the credibility of the resulting welfare measures and policy implications. To accurately incorporate zero WTP responses within the econometric model, this study employs the OB spike model. This specification integrates the OB dichotomous choice structure with the spike approach initially introduced by Kriström [38]. The integrated approach offers particular benefits for CV studies because it differentiates between individuals with a genuine zero valuation and those whose positive valuation simply falls below the lowest bid level offered.

The survey design incorporated two predetermined bid amounts, denoted as the higher bid ($K^H$) and the lower bid ($K^L$), with respondents randomly assigned to one of two equally sized subsamples. In the first subsample, respondents were initially presented with $K^H$. A positive response terminated the sequence, whereas a rejection triggered a follow up question involving $K^L$. Acceptance of $K^L$ concluded the sequence, while rejection prompted an additional “spike” question, asking whether the respondent was unwilling to pay even a minimal positive amount. A “yes” to the spike question indicated a WTP greater than zero but less than $K^L$, while a “no” classified the individual as having a true zero WTP.

In the second subsample, the sequence began with $K^L$, followed by $K^H$ in the case of acceptance, or the spike question if $K^L$ was rejected. This bid structure yields eight distinct response patterns, allowing for both efficient data collection and explicit identification of zero valuations. Formally, let $i=1,\dots ,T$ denote the survey’s respondents, and define indicator functions for each response sequence. When $K^H$ is offered first, possible outcomes include the following:

$$\left\{\begin{array}{c}{I}_i^Y=1 (i\mathrm{t}\mathrm{h} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{i}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{e} \mathrm{a}\mathrm{n}\mathrm{s}\mathrm{w}\mathrm{e}\mathrm{r}\mathbf{s} “\mathrm{y}\mathrm{e}\mathrm{s}”) \\ I_i^{NY}=1\left(i\mathrm{t}\mathrm{h} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{i}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{e} \mathrm{a}\mathrm{n}\mathrm{s}\mathrm{w}\mathrm{e}\mathrm{r}\mathrm{s} \mathrm{“}\mathrm{no-}\mathrm{yes}”\right) \\ I_i^{NNY}=1\left(i\mathrm{t}\mathrm{h} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{i}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{e} \mathrm{a}\mathrm{n}\mathrm{s}\mathrm{w}\mathrm{e}\mathrm{r}\mathrm{s} \mathrm{“}\mathrm{n}\mathrm{o-}\mathrm{n}\mathrm{o-}\mathrm{y}\mathrm{e}\mathrm{s}\mathrm{”}\right)\\ I_i^{NNN}=1\left(i\mathrm{t}\mathrm{h} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{i}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{e} \mathrm{a}\mathrm{n}\mathrm{s}\mathrm{w}\mathrm{e}\mathrm{r}\mathrm{s} \mathrm{“}\mathrm{n}\mathrm{o-}\mathrm{n}\mathrm{o-}\mathrm{n}\mathrm{o}\mathrm{”}\right) \end{array}\right.$$

(1)

when $K^L$ is offered first, the outcomes are the following:

$$\left\{\begin{array}{c}{I}_i^{YY}=1\left(i\mathrm{t}\mathrm{h} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{i}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{e} \mathrm{a}\mathrm{n}\mathrm{s}\mathrm{w}\mathrm{e}\mathrm{r}\mathrm{s} \mathrm{“}\mathrm{y}\mathrm{e}\mathrm{s-}\mathrm{y}\mathrm{e}\mathrm{s}\mathrm{”}\right)\\ I_i^{YN}=1\left(i\mathrm{t}\mathrm{h} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{i}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{e} \mathrm{a}\mathrm{n}\mathrm{s}\mathrm{w}\mathrm{e}\mathrm{r}\mathrm{s} \mathrm{“}\mathrm{y}\mathrm{e}\mathrm{s-}\mathrm{n}\mathrm{o}\mathrm{”}\right)\\ I_i^{NTY}=1\left(i\mathrm{t}\mathrm{h} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{i}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{e} \mathrm{a}\mathrm{n}\mathrm{s}\mathrm{w}\mathrm{e}\mathrm{r}\mathrm{s} \mathrm{“}\mathrm{n}\mathrm{o-}\mathrm{y}\mathrm{e}\mathrm{s}\mathrm{”}\right)\\ I_i^{NN}=1\left(i\mathrm{t}\mathrm{h} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{i}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{e} \mathrm{a}\mathrm{n}\mathrm{s}\mathrm{w}\mathrm{e}\mathrm{r}\mathrm{s} \mathrm{“}\mathrm{n}\mathrm{o-}\mathrm{n}\mathrm{o}\mathrm{”}\right)\end{array}\right.$$

(2)

Let $W$ denote the interviewee’s WTP. The probability of acceptance or rejection at bid level $K$ is

$$\begin{gathered} \mathrm{Pr}\left(“\mathrm{yes}” \mathrm{t}\mathrm{o} K\right)=\mathrm{Pr}\left(W\ge K\right)=1-C_W\left(K;s_0,s_1\right) \\ \mathrm{Pr}\left(\mathrm{“}\mathrm{n}\mathrm{o}\mathrm{”} \mathrm{t}\mathrm{o} K\right)=\mathrm{Pr}\left(W<K\right)=C_W(K;s_0,s_1) \end{gathered}$$

(3)

where $C_W\left(K;s_0,s_1\right)$ is the cumulative distribution function (cdf) of $W$. Following conventional practice, $W$ is assumed to follow a logistic distribution, such that

$$C_W\left(K;s_0,s_1\right)=\left\{\begin{array}{c}{\left[1+\exp \left(s_0-s_{1}K\right)\right]}^{-1 } \mathrm{i}\mathrm{f} K>0\\ {\left[1+\exp \left(s_0\right)\right]}^{-1} \mathrm{i}\mathrm{f} K=0\\ 0 \mathrm{i}\mathrm{f} K<0\end{array}\right.$$

(4)

where the terms, $s_0$ and $s_1$, denote the two scale parameters of the two-parameter logistic distribution function $C_W\left(·\right)$ specified in Equation (4), which underpins the cdf for the WTP commuter in the spike model framework. Specifically, $s_0$ represents the intercept scale parameter, capturing baseline response scale and location shifts independent of bid levels, while $s_1$ is the slope scale parameter (or bid slope), governing sensitivity to the offered bid amount. The two-parameter logistic form generalizes the standard single-parameter logistic cdf by allowing for heteroscedasticity in utility disturbances.

The spike component is represented by ${\left[1+\exp \left(s_0\right)\right]}^{-1}$, capturing the probability mass at zero. Combining Equations (1)–(4), the log-likelihood of the model is given as follows [38,39,40]:

$$\ln L= \sum _{i=1}^T\left\{\begin{array}{c}\begin{array}{c}\begin{array}{c}{(I}_i^Y+I_i^{YY})\ln \left[1-C_W\left(K_i^H;s_0,s_1\right)\right]\\ +{(I}_i^{NY}+I_i^{YN})\ln \left[C_W\left(K_i^H;s_0,s_1\right)-C_W\left(K_i^L;s_0,s_1\right)\right]\end{array}\\ +{(I}_i^{NNY}+I_i^{NTY})\ln \left[C_W\left(K_i^L;s_0,s_1\right)-C_W\left(0;s_0,s_1\right)\right]\end{array}\\ +{(I}_i^{NNN}+I_i^{NN})\ln \left[C_W\left(0;s_0,s_1\right)\right]\end{array}\right\}$$

(5)

The primary result obtained from CV analysis is the average WTP. Substituting the logistic distribution of Equation (4) into the expected value expression gives $\left(1/s_1\right)\mathrm{l}\mathrm{n}\left[1+\exp \left(s_0\right)\right]$. This baseline specification can be extended to incorporate socioeconomic covariates by replacing the intercept term $s_0$ in Equation (4) with a linear function of explanatory variables, $s_0+x_i^{\prime }\gamma$, where $x_i$ is a vector of respondent-specific characteristics (e.g., age, political orientation, education, income), and $\gamma$ is the parameter vector. This extension enables the model to capture heterogeneity in WTP across sociodemographic groups, thereby improving the explanatory strength of the econometric analysis while simultaneously increasing the policy applicability of the findings.

The single-bound (SB) spike model, also employed in our analysis and detailed below, utilizes only responses to the initial bid from the OB data collection format, effectively treating the survey as a standard SB elicitation for baseline comparability. This parsimonious approach leverages Equations (1)–(4) without modification, while formalizing the log-likelihood function specifically for SB responses as follows:

$$\ln L= \sum _{i=1}^T\left\{\begin{array}{c}\begin{array}{c}\begin{array}{c}{I}_i^Y\ln \left[1-C_W\left(K_i^H;s_0,s_1\right)\right]+{(I}_i^{YY}+I_i^{YN})\ln \left[1-C_W\left(K_i^L;s_0,s_1\right)\right]\\ +{(I}_i^{NY}+I_i^{NNY})\ln \left[C_W\left(K_i^H;s_0,s_1\right)-C_W\left(0;s_0,s_1\right)\right]\end{array}\\ +I_i^{NTY}\ln \left[C_W\left(K_i^L;s_0,s_1\right)-C_W\left(0;s_0,s_1\right)\right]\end{array}\\ +{(I}_i^{NNN}+I_i^{NN})\ln \left[C_W\left(0;s_0,s_1\right)\right]\end{array}\right\}$$

(6)

The SB model’s computational simplicity facilitates direct comparison with OB results, testing data efficiency gains while mitigating potential anchoring from the second bid—critical for validating spike robustness in high-zero-WTP contexts like energy infrastructure valuation.

The spike model, originally proposed by Kriström [38], was selected for analyzing OB dichotomous choice CV responses due to its superior handling of substantial zero WTP responses, which often indicate true zero WTP or values below the lowest bid—common in environmental valuation where a non-trivial proportion of respondents exhibit zero marginal utility for incremental improvements. Unlike the conventional logit model, which assumes a continuous logistic distribution and truncates zeros by treating all ‘no’ responses as positive WTP below the offered bid (leading to upward bias in mean WTP estimates), the spike model explicitly incorporates a point mass (spike) at zero WTP, distinguishing ‘real zero’ bids from ‘uncertain positive’ ones via follow-up questions. This approach outperforms alternatives like the standard SB logit in datasets with high zero-WTP probabilities, as evidenced by improved model fit and more plausible welfare measures, particularly for policy-relevant medians near zero.

To determine the nine bid levels (ranging from KRW 1000 to KRW 15,000, as will be presented in

Table 2. Number of answers for each set of bids in the sample.

Bids a		Number of Answers
First	Second	“yes-yes”	“yes-no”	“no-yes”	“no-no”	Totals
1000	3000	14	27	1	30	72
2000	4000	6	23	5	37	71
3000	6000	6	16	11	38	71
4000	8000	4	5	17	45	71
6000	10,000	4	7	23	37	71
8000	12,000	1	1	24	46	72
10,000	15,000	3	3	18	48	72
	Totals	38	82	99	281	500
First	Second	“yes”	“no-yes”	“no-no-yes”	“no-no-no”	Totals
3000	1000	10	21	2	38	71
4000	2000	17	9	5	41	72
6000	3000	7	11	12	41	71
8000	4000	2	8	13	48	71
10,000	6000	7	2	11	52	72
12,000	8000	2	3	18	48	71
15,000	10,000	2	2	22	46	72
	Totals	47	56	83	314	500

Notes: Data presented in this table were derived from the survey conducted for this study. ^a The unit is Korean won (USD 1.0 = KRW 1380 at the time of the survey).

), a rigorous three-stage approach was employed, following best practices in CV design to ensure efficiency and realism (e.g., [32]). Stage 1: An open-ended WTP elicitation survey was conducted with a target population sample of 30 respondents, mirroring the main survey’s demographics, to empirically derive the underlying WTP distribution and identify plausible response ranges. Stage 2: Outliers were trimmed by excluding the top and bottom 15% of responses to mitigate extreme values potentially arising from strategic answering or misunderstanding, thereby refining the central tendency and variability of WTP. Stage 3: Seven bid pairs were then selected to form nine monotonically increasing levels, spaced to capture cumulative response probabilities effectively (e.g., approximately 5–95% ‘yes’ responses across the range), optimizing statistical power for logit estimation while avoiding bunching at extremes. This iterative process, informed by pilot data, aligns with recommendations for spike model robustness in SB CV formats and minimizes interpolation bias in welfare measures.

4. Results and Discussion

4.1. Data

According to the survey agency responsible for data collection, the respondents were well informed about the objectives and background of the questionnaire, as well as the attributes of the good under evaluation. Most participants answered the WTP questions without significant difficulty. Responses that were deemed incomplete or insincere were excluded from the final dataset. To identify and exclude incomplete or insincere responses from the final dataset, a systematic post-survey protocol was applied, consistent with standard practices in stated preference surveys. Incomplete responses were defined as those missing key sections: (1) demographic variables, (2) the WTP elicitation question, or (3) follow-up spike queries.

In addition to the automated post-survey data cleaning protocols outlined above, field interviewers exercised professional judgment to identify and exclude responses deemed unengaged or insincere during data collection, following established fieldwork quality control practices in stated preference surveys. Specifically, enumerators flagged questionnaires where respondents exhibited clear disengagement—such as random or patterned answering, refusal to engage with scenarios, or overt expressions of non-seriousness—based on real-time interaction cues and response consistency checks embedded in the instrument.

Furthermore, supervisors conducted a comprehensive verification process by contacting all respondents via the mobile phone numbers provided at the survey’s conclusion. This follow-up validation confirmed whether responses were completed attentively and accurately; any deemed unreliable (e.g., due to confirmed non-participation, contradictory statements, or admitted satisficing) were systematically excluded. To maintain the target sample size and statistical power, excluded observations were promptly replaced through targeted supplementary surveys matching the original sampling frame’s demographics and geographic distribution. This multi-tiered quality assurance—combining interviewer discretion, supervisor validation, and compensatory recruitment—ensured that only reliable, effortful responses were retained for final analysis, enhancing data integrity and minimizing non-attentiveness bias while upholding representativeness. Sensitivity analyses verified that welfare estimates remained robust to these exclusions.

The survey for this study was conducted from late May to late June 2025. The statistical analysis was conducted using Time Series Processor (TSP) software (Version 5.1; TSP International, Palo Alto, CA, USA). Table 2 presents the distribution of WTP responses obtained from the survey. The upper section of the table presents the responses obtained from the subsample that was shown the lower bid amount first, while the lower panel presents those given the higher bid amount first. Half of the participants were initially presented with the lower bid question, while the remaining half started with the higher bid question to control for potential starting-point bias. The results reveal a clear monotonic relationship between bid amount and affirmative response; the proportion of “yes” responses declined as the bid level increased. In other words, the frequency of “yes-yes” or single “yes” responses decreased with higher bid values. Notably, 59.5% of the total respondents indicated a zero WTP, suggesting that the majority of respondents expressed an unwillingness to incur any extra cost on their electricity bills to fund the construction of the West Coast HVDC transmission line.

The questionnaire also included items designed to capture respondents’ perceptions related to renewable energy transmission, as well as their socioeconomic characteristics. To examine how these factors influence WTP, several selected variables were incorporated as covariates in the WTP estimation model. The final set of explanatory variables included political orientation, income, age, region of residence, prior knowledge of renewable energy supply, gender, and residence within or outside the SMA.

Table 3. Description of variables used in the model.

Variables	Definitions	Mean	Standard Deviation
Political	The respondent’s political orientation (0 = conservatism; 1 = others (progressives and moderate))	0.75	0.43
Income	The respondent household’s monthly income (unit: million Korean won)	5.63	2.54
Metro	Whether the respondent’s residence is the Seoul Metropolitan Area (0 = no; 1 = yes)	0.54	0.50
Age	The respondent’s age	47.80	10.70
Knowledge	Dummy for the respondent’s knowing about overgeneration of electricity in Southwestern region of the country before the survey (0 = no; 1 = yes)	0.09	0.28

Note: Data presented in this table were derived from the survey conducted for this study.

provides a summary of the descriptive statistics for these variables. The inclusion of these covariates enables a more detailed analysis of heterogeneity in respondents’ valuations and facilitates the identification of key demographic or attitudinal factors shaping public support for renewable energy infrastructure development.

The five explanatory variables included in the models with covariates—Political, Income, Metro, Age, and Knowledge—were selected through a systematic process grounded in theory and empirics from non-market valuation literature on energy infrastructure and environmental preferences. Initial twelve candidates were derived from the following: (1) theoretical foundations, such as the random utility model positing attitudes and costs as WTP shifters; (2) a literature review, prioritizing proven determinants like pro-environmental attitudes, energy literacy, and socioeconomic status (income, urban/rural divide); and (3) pilot survey insights (n = 30), where exploratory regressions identified significant predictors (p < 0.10) while revealing low variance in alternatives. However, the estimation coefficients for the remaining seven variables failed to achieve statistical significance in pilot regressions and demonstrated negligible explanatory power, leading to their exclusion from the final analysis to maintain model parsimony and avoid overfitting. This selection criterion aligns with standard practices in non-market valuation, where only theoretically grounded and empirically robust predictors are retained to ensure precise WTP heterogeneity estimates. Rejected variables included gender, education level, and political affiliation. Final inclusion balanced parsimony, avoiding overfitting and multicollinearity—ensuring robust WTP heterogeneity insights.

A brief description of the study sample is as follows. The final dataset comprises 1000 household heads aged between 20 and 65 years who were residing in South Korea at the time of the survey, reflecting the economically active population targeted in this research. The sample is almost gender-balanced, with 50.0% female and 50.0% male respondents, and covers all major administrative regions, including the SMA, other large metropolitan cities, and provincial areas. The average age of respondents is approximately 48 years, and the mean monthly household income is KRW 5.63 million (USD 4.08 thousand), both of which closely match national benchmark figures for the same period. In terms of political orientation, about three quarters of the sample identify as progressive or moderate rather than conservative, and slightly more than half of respondents live in the SMA, where electricity demand is particularly concentrated. Although only around 9% of respondents reported prior knowledge of the RE oversupply issue in the southwestern region, this subgroup might show systematically higher WTP, underscoring the importance of information and awareness in shaping public support for submarine HVDC projects. Overall, the survey sample can be broadly representative of the national population in key socio-demographic dimensions, thereby supporting the external validity of the CV results. A detailed discussion of representativeness will be elaborated on in Section 4.3.

4.2. Results

Table 4. Estimation results of the one-and-one-half-bound spike models.

Variables	Model Without Covariates		Model with Covariates
	Coefficient Estimates	t-Values	Coefficient Estimates	t-Values
Constant	−0.3846	−6.00 #	−2.8225	−3.08 #
Bid amount a	−0.2834	−20.33 #	−0.2903	−19.73 #
Political			0.3643	2.30 **
Income			0.3045	2.07 **
Metro			0.2866	2.10 **
Age			−0.2650	−1.78 *
Knowledge			0.6962	3.16 #
Spike	0.5950	38.50 #	0.5975	37.93 #
Averages of monthly household willingness to pay	KRW 1832 (USD 1.33)	16.08 #	KRW 1774 (USD 1.29)	15.43 #
95% CIs b	KRW 1625 to 2084 (USD 1.18 to 1.51)		KRW 1569 to 2028 (USD 1.14 to 1.47)
Wald statistics (p-values)	525.50 (0.000)		514.78 (0.000)
Log-likelihood	−1058.83		−1038.13
Sample size	1000		1000

Notes: All data presented in this table derive from the analysis of survey data collected in this study. ^a The unit is KRW 1000 (USD 0.725). ^b It means confidence interval computed using the method given in Krinsky and Robb [41]. ^#, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

and

Table 5. Estimation results of the single-bound spike models.

Variables	Model Without Covariates		Model with Covariates
	Coefficient Estimates	t-Values	Coefficient Estimates	t-Values
Constant	−0.3873	−6.03 #	−0.9298	−0.87
Bid amount a	−0.2486	−15.52 #	−0.2551	−15.92 #
Political			0.3829	2.39 **
Income			0.2522	1.70 *
Metro			0.2823	2.07 **
Age			−0.2389	−1.60
Knowledge			−0.6696	−2.97 #
Spike	0.5956	38.50 #	0.5985	37.83 #
Averages of monthly household willingness to pay	KRW 2084 (USD 1.51)	13.69 #	KRW 2012 (USD 1.46)	15.43 #
95% CIs b	KRW 1810 to 2425 (USD 1.31 to 1.76)		KRW 1091 to 3535 (USD 0.79 to 2.56)
Wald statistics (p-values)	329.59 (0.000)		363.70 (0.000)
Log-likelihood	−886.31		−867.56
Sample size	1000		1000

Notes: All data presented in this table derive from the analysis of survey data collected in this study. ^a The unit is KRW 1000 (USD 0.72). ^b It means confidence interval computed using the method given in Krinsky and Robb [41]. ^#, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

report the estimation results from the OB and SB spike models, respectively, applied to the CV data. The key empirical findings from these models can be synthesized into five principal insights. First, the Wald tests for joint significance of the covariates in both models yield p-values effectively equal to zero, decisively refuting the null hypothesis that all coefficients are simultaneously equal to zero. This confirms the overall explanatory power and appropriateness of the model specifications, ensuring that the included variables and functional forms capture significant variation in respondents’ WTP responses.

Second, estimating the mean WTP is a central component in contingent valuation research. The OB model yields an estimated mean WTP per household per month of KRW 1832 (USD 1.33), whereas the SB produces a somewhat higher estimate of KRW 2084 (USD 1.51). Both estimates achieve high levels of statistical significance, with t-values of 16.08 and 13.69, respectively, rejecting the null hypothesis that the mean WTP is meaningless at the 1% level. These results provide robust evidence of respondents’ positive monetary valuation of investment in the West Coast HVDC transmission.

Third, the bid amount coefficients in both models are negative and statistically significant (–0.2834 for the OB model and –0.2486 for the SB model). This negative relationship is economically intuitive and expected, indicating that the probability of receiving an affirmative response diminishes as the bid amount rises, consistent with diminishing marginal utility and rational economic behavior. The negative and statistically significant coefficients for bid amounts in both models align with theoretical expectations and validate the internal consistency of the elicited WTP responses. The result confirms that as hypothetical cost increases, respondents exhibit rational behavior by lowering their probability of acceptance, consistent with economic demand theory and reinforcing the credibility of the survey instrument [6,32].

Fourth, the spike parameters estimated in both models are approximately 0.595, closely corresponding to the observed frequency of zero WTP responses (59.5%) in the sample. This indicates that the models effectively capture the pronounced clustering of zero valuations, a common feature in the valuation of environmental and public goods. The spike model framework thus appropriately accommodates this mass point at zero, enhancing the reliability and interpretability of the estimation. The validity of this modeling choice is consistent with prior CV applications dealing with non-market goods featuring zero-inflated distributions.

Fifth, the associated 95% confidence intervals for the mean WTP estimates were constructed using the parametric bootstrap method given in Krinsky and Robb [41]. These intervals ranged from KRW 1625 to 2084 (USD 1.18 to 1.51) for the OB model and from KRW 1810 to 2425 (USD 1.31 to 1.76) for the SB model. There is considerable overlap between these intervals, suggesting that no significant difference can be found between the two mean WTP estimates. When contextualized against the average monthly household electricity cost of KRW 71,143 (USD 51.55) in 2024, the estimated WTP indicates a premium of approximately 2.6% over existing electricity bills, a substantial level of support given the voluntary nature of the payment and the non-market characteristic of the infrastructure good.

From a policy perspective, these findings attest to measurable public support for investment in the HVDC transmission in the West Coast region. The statistically significant positive WTP estimates provide quantitative justification for potential rate adjustments or funding mechanisms to finance the deployment of such an infrastructure. Furthermore, the methodological rigor demonstrated by employing spike models enriches the validity of the stated preference data in this context, serving as a reference point for future valuation efforts in energy infrastructure and environmental economics.

4.3. Discussion of the Results

The results provide several important insights regarding public WTP for the proposed West Coast submarine HVDC transmission system in South Korea. The findings of this study are discussed with regarding five key aspects: (i) the implications derived from the estimated mean WTP; (ii) the assessment and verification of potential response effects inherent in the OB model estimation; (iii) an evaluation of the economic viability of the West Coast HVDC project through a comparative analysis of the estimated WTP and the associated project costs; (iv) an examination of how various covariates influence the respondents’ probabilities of agreeing to pay the stated bid amounts; (v) an exploration of the role of protest bids and their impact on the estimation of the mean WTP; and (vi) an explanation of the limitations of the study.

4.3.1. Implications Derived from the Estimated Mean WTP

The estimated mean WTP values from the OB and SB spike models demonstrate statistically significant positive monetary valuations for investment in the HVDC transmission. The OB model estimates an average monthly WTP of KRW 1832 per household, while the SB model yields a slightly higher WTP of KRW 2084. Both estimates are highly significant at the 1% level, with robust t-values, indicating strong public support for this upgrade of the infrastructure. These findings corroborate earlier studies highlighting the value placed by households on the modernization and environmental benefits of advanced transmission technologies [25,26].

The statistically significant and economically relevant WTP estimates suggest robust public endorsement for the West Coast submarine HVDC transmission line, providing quantitative justification for allocating public and private funds toward its construction. Given the significant upfront investment costs and the likelihood of resulting in electricity rate increases, these results indicate consumer acceptance for moderate rate adjustments to finance cleaner, more efficient, and socially acceptable power transmission. Furthermore, when benchmarked against the average monthly household electricity bill of KRW 71,143, the WTP estimates represent approximately 2.6% of current electricity costs. This magnitude reflects a meaningful willingness on the part of consumers to bear a premium for an improved electricity transmission infrastructure, underscoring the social value of the HVDC project beyond conventional market transactions [36,42].

4.3.2. Assessment and Verification of Potential Response Effects

It is essential to verify the presence of any response effects in the use of the OB model [43]. The SB model requires respondents to answer only one WTP question, but the OB model sometimes involves additional WTP questions for certain respondents. This procedure may influence respondents’ stated WTP by eliciting further information that could alter their true valuation. Typically, if response effects are detected, the SB model should be used rather than the OB model to avoid bias [6,27]. Therefore, rigorous testing for the presence of these effects is crucial to ensure the validity of the estimated WTPs. Addressing potential response biases enhances the credibility of the valuation results and supports more reliable policy recommendations.

Testing for the existence of response effects involves assessing whether the mean WTP from the OB model differs significantly from that from the SB model. However, direct statistical comparison is complicated because the samples used in the OB and SB models do exactly coincide, limiting the applicability of conventional tests. When treatment and control samples are distinctly partitioned—as in randomized experiments—standard statistical tests (e.g., t-tests) readily assess the significance of treatment effects through between-group comparisons. However, in designs like the present study, where the OB and SB spike models derive from identical samples, directly testing the statistical significance of inter-model differences becomes challenging. To address this limitation, this study employs an overlap test examining whether the 95% confidence intervals of the mean WTP estimates from the two models intersect. As previously mentioned, the confidence intervals indeed overlap, indicating that the null hypothesis of no response effect cannot be rejected at the 5% level. Therefore, the following analysis is conducted using the estimates derived from the OB model, assuming no significant response bias is present.

4.3.3. Preliminary Economic Feasibility Analysis

The estimated mean WTP serves as a meaningful benchmark against associated costs for assessing the project’s economic viability. First, the average WTP estimated from the sample is extrapolated to reflect the preferences of the broader population. The representativeness of the sample is a critical consideration in this step. To address this, the present study engaged a professional survey firm to ensure that rigorous sampling procedures were carried out that closely align the sample’s characteristics with those of the national population. Additionally,

Table 6. Comparison of some demographic and socioeconomic characteristics between the sample and the population.

Characteristics		Sample a	Population b
Household income c		KRW 5.63 million	KRW 5.49 million
Gender
	Female	50.0%	49.9%
	Male	50.0%	50.1%
Region
	Seoul	21.3%	19.0%
	Pusan	7.0%	6.7%
	Daegu	4.6%	4.7%
	Incheon	5.9%	5.5%
	Gwangju	3.2%	2.9%
	Daejeon	3.3%	3.0%
	Ulsan	1.9%	2.1%
	Sejong	0.7%	0.7%
	Gyunggi	26.9%	24.4%
	Gangwon	2.6%	3.2%
	Chungbuk	2.6%	3.2%
	Chungnam	3.8%	4.3%
	Jeonbuk	3.3%	3.6%
	Jeonnam	2.2%	3.6%
	Gyungbuk	4.7%	5.4%
	Gyungnam	6.0%	6.5%

Notes: ^a Data presented in this table were derived from the survey conducted for this study and the number of respondents is 1000. ^b They come from Korea Ministry of Data and Statistics [44]. ^c Average household income per month of the first quarter of 2025.

shows a comparison of selected demographic and socioeconomic characteristics between the survey sample and the national population benchmark. It appears that no significant differences exist across these key attributes, thereby supporting the representativeness of the sample and the extrapolation of the study’s results to the wider population. There are 22,338,821 households in 2025 [44].

Multiplying the monthly average WTP of KRW 1832 (USD 1.33) by 12 months and then by the total number of households yields an annual aggregated economic benefit of roughly KRW 492 billion (USD 356 million). This figure corresponds to the annual economic benefits accruing to residential electricity consumers from the construction and operation of the West Coast HVDC transmission system, expected to commence in 2025, to be completed by 2030, and operated until 2060. Applying the official social discount rate of 4.5% used by the South Korean government, the present value of these benefits as of 2025 amounts to approximately KRW 9.09 trillion (USD 6.58 billion).

Next, the associated costs of the West Coast HVDC project must be estimated. The total investment cost is projected to be KRW 11.5 trillion (USD 8.33 billion). Given that residential electricity consumers constitute approximately 15.4% of total electricity consumption, it is reasonable to allocate 15.4% of the total project cost—equating to KRW 1.77 trillion (USD 1.28 billion)—as the cost burden borne by residential consumers. Furthermore, the annual operation and maintenance (O&M) costs of the HVDC transmission line are estimated at 9.6% of the total investment costs.

The 9.6% figure for annual O&M costs was obtained directly from technical experts at the Grid Planning Division of KEPCO, the state-owned entity responsible for developing and operating the West Coast HVDC transmission system. Given South Korea’s established operational experience with three existing submarine HVDC lines interconnecting the mainland with Jeju Island—each exceeding 200 km in length—this empirical data provides a robust, context-specific basis for estimating O&M expenditures specific to South Korean marine transmission infrastructure. These precedents enable precise cost calibration, accounting for localized factors such as seawater corrosion mitigation, cathodic protection systems, and periodic cable inspections via remotely operated vehicles, which collectively inform the projected lifecycle costs for the proposed West Coast project under similar geophysical and regulatory conditions.

When accounting for the investment phase spanning 2025 to 2030 and the operational phase from 2030 to 2060, the present value of the total costs attributable to residential consumers is approximately KRW 3.72 trillion (USD 2.69 billion). Consequently, the net present value of the project, from the perspective of residential electricity consumers, is computed as KRW 5.37 trillion (USD 3.89 billion), with a benefit–cost ratio (BCR) of 2.44. These results strongly suggest that the project is economically justified when evaluated solely on costs and benefits accruing to residential electricity users.

The cost–benefit analysis extrapolates the estimated monthly household WTP (KRW 1832) over a 30-year project horizon—standard practice in energy infrastructure appraisal —to derive aggregate social benefits. This projection implicitly assumes preference stability across decades, consistent with constant real-income valuation protocols in environmental economics. However, such long-horizon extrapolation warrants three caveats that merit explicit acknowledgment: (1) preference evolution, as public attitudes toward renewable integration and grid reliability may shift with generational turnover, heightened climate awareness, or experience with curtailment crises; (2) technological substitution, where advances in offshore wind, long-duration storage, or demand response could alter transmission needs; and (3) market-policy dynamics, including electricity pricing reforms, carbon pricing implementation, or the revised 12th Basic Plan for Electricity Supply and Demand potentially redistributing cost burdens. Sensitivity analyses relaxing these assumptions—e.g., discounting WTP growth at 0.5–2% annually or truncating benefits at 15 years—can change the BCR. This limitation does not invalidate the positive net present value but underscores the value of periodic re-valuation as South Korea’s energy transition matures.

4.3.4. Investigation of the Impacts of the Covariates

The estimates of the model with the covariates presented in the third column of Table 4 warrant detailed discussion. As outlined in Table 3, five covariates were incorporated into the model to account for household characteristics, respondent attributes, and perceptual factors. Specifically, two household-related variables—Age and Metro—two respondent-related variables—Income and Political orientation—and one perception-related variable—Knowledge about the electricity oversupply in the Jeolla region—were included. Each estimated coefficient for these covariates attains statistical significance at the 10% level, indicating meaningful explanatory power.

The signs of the coefficients inform the direction of their influence on the probability of reporting “yes” to the presented bid amounts. A positive coefficient indicates an increased likelihood of accepting the bid, while a negative coefficient implies a negative association. From the results shown in Table 4, respondents with a progressive political orientation and higher income levels were more likely to respond affirmatively to the bid amounts, suggesting greater willingness to pay among these groups. Moreover, residents in the SMA exhibited higher probabilities of accepting the payment proposition compared to those living outside the SMA.

Knowledgeable respondents, defined as those aware of the oversupply issue in the Jeolla region, also demonstrated a more favorable response pattern than those lacking such awareness. Conversely, age exhibited a negative relationship—older respondents were less likely to accept the bid amount, indicating that willingness to pay decreases with increasing age. This pattern may reflect differences in preferences, income constraints, or risk attitudes across age groups. Overall, these covariate effects emphasize the importance of socio-demographic and informational factors in shaping consumer valuation of energy infrastructure improvements. Understanding these influences is critical for targeting communication and policy measures effectively to enhance public support.

4.3.5. Exploration of the Zero WTP Responses

This study examines the influence of protest bid responses on the estimated mean WTP. To explore this issue, 595 respondents who reported a WTP of zero were presented with a debriefing question to clarify their underlying reasons.

Table 7. Summary of stated reasons for zero willingness to pay.

Reasons	Number of Responses
1. High voltage direct current (HVDC) transmission system should be covered by taxes that have already been paid.	396
2. Sufficient information has not been provided to make a judgment.	73
3. The HVDC transmission system is not important enough to warrant prioritization.	30
4. The government has already spent too much money in the HVDC transmission system.	14
5. The HVDC transmission system is not a matter of concern for our household.	45
6. The HVDC system has little value to me.	19
7. Our household lacks the financial means to afford payment.	1
8. No additional tax revenue will be allocated to the HVDC transmission system.	12
9. Others	5
Totals	595

Note: Data presented in this table were derived from the survey conducted for this study.

provides a summary of the stated reasons for zero WTP. Among the nine response categories identified, 20 responses corresponded to “6. The HVDC system has little value to me” and “7. Our household lacks the financial means to afford payment,” and were therefore classified as true zeros. The remaining 575 responses were deemed protest bids. Consequently, the OB spike model was re-estimated after excluding these 575 protest bid responses, leaving a sample consisting of 405 positive WTP observations and 20 true zero WTP cases. The resulting average WTP from this restricted sample was KRW 4099 (USD 2.97), which is nearly 2.2 times greater than the average WTP of KRW 1832 (USD 1.33) estimated using the 1000 observations.

This substantial increase demonstrates that omitting protest bid responses can significantly inflate the estimated mean WTP, underscoring the important impact that protest responses may have on valuation estimates. The next issue to be addressed is the selection of the appropriate sample for policy analysis—namely, whether to use the entire set of 1000 responses or the reduced set of 425 observations excluding protest bids. In this study, the former approach was adopted for two main reasons. First, since the estimated mean WTP directly informs policy analysis, a conservative approach may be preferable, especially for public decision-making. Second, protest bids themselves are expressions of respondents’ preferences and attitudes, and thus may provide useful information that should not be arbitrarily dismissed from analysis.

While the 59.5% zero WTP responses warrant careful substantive interpretation beyond methodological handling, the OB spike model robustly accommodates this protest/no-response distribution while yielding statistically significant mean estimates (KRW 1832 per month, p < 0.01) [32]. In dichotomous choice CV—mimicking referendum voting—this elevated zero proportion aligns with welfare economic principles where aggregate Hicksian surplus, rather than simple majority approval (>50% ‘yes’), determines allocative efficiency [45]. Policies generating positive net social benefits remain justified despite opposition from income-constrained households, strategic non-payers, or heterogeneous valuation distributions, as evidenced by real-world referenda approving infrastructure despite plurality dissent when economic merit prevails [46].

As shown in Table 7, qualitative analysis of debriefing responses from the 595 zero WTP cases reveals nuanced motivations underpinning refusal patterns, with the following rationales comprising the predominant top four categories: (1) perceived government responsibility (396 cases, 66.6%), where respondents asserted, “High voltage direct current (HVDC) transmission system should be covered by taxes that have already been paid,” conceptualizing grid modernization as a public good obligation falling under existing KEPCO/state fiscal mandates rather than incremental household surcharges; (2) information insufficiency (73 cases, 12.3%), citing, “Sufficient information has not been provided to make a judgment,” highlighting needs for enhanced risk communication and technical transparency during project siting; (3) household irrelevance (45 cases, 7.6%), stating, “The HVDC transmission system is not a matter of concern for our household,” reflecting spatially differentiated perceptions among non-proximate or low-usage demographics; and (4) prioritization concerns (30 cases, 5.0%), deeming, “The HVDC transmission system is not important enough to warrant prioritization,” indicative of competing fiscal preferences amid multiple energy transition demands.

These empirically grounded rationales—capturing 91.5% of zero responses—contextualize the 59.5% refusal rate not as outright project rejection, but as differentiated contestation over cost allocation principles (public vs. private burden), information adequacy, geographic equity, and policy prioritization. Such findings align with principal-agent problems in utility regulation and reveal strategic non-disclosure of positive valuations under perceived fiscal overreach. Complementing the aggregate positive mean WTP (KRW 1832/month), these insights prescribe targeted policy responses: (i) transparent fiscal modeling demonstrating tax-equivalent surcharge impacts; (ii) pre-referendum public campaigns addressing technical uncertainties; (iii) spatially tailored benefit framing for peripheral households; and (iv) multi-criteria prioritization frameworks integrating submarine HVDC against alternatives like offshore wind or demand response investments. Collectively, this mixed-methods interpretation enriches understanding of public attitudes toward large-scale energy infrastructure, informing governance strategies to broaden legitimacy beyond economic willingness alone.

4.3.6. Limitations of the Study

Some limitations of the study need to be described. While this study provides robust evidence on public WTP for submarine HVDC infrastructure, several limitations warrant acknowledgment, informing avenues for future research. First, as an SP exercise, CV remains susceptible to hypothetical bias despite mitigations like cheap talk and certainty calibrations; field experiments validating induced valuations against actual payments could strengthen external validity. Second, the dichotomous choice format, while incentive-compatible, aggregates individual utilities and precludes attribute-level trade-offs—future DCEs could decompose WTP for HVDC features (e.g., burial depth, route aesthetics) versus alternatives like overhead lines. Third, the cross-sectional design captures static preferences at a 2025 snapshot; longitudinal tracking amid evolving RE policies or cost trajectories would elucidate temporal stability. Fourth, while nationally representative, the sample excludes non-household stakeholders (e.g., industries); targeted surveys of commercial users could refine aggregate benefits. Finally, cost estimates rely on KEPCO engineering projections; sensitivity analyses incorporating uncertainty (e.g., Monte Carlo simulations) are recommended for net benefit robustness. Notwithstanding these limitations, the study’s methodological rigor and policy relevance position it as a foundational contribution, with future extensions via hybrid revealed preference–stated preference designs offering a promising research direction.

5. Conclusions

This article comprehensively evaluated the economic implications and public valuation of the proposed West Coast submarine HVDC transmission system in South Korea using a CV approach. Employing a nationally representative survey of 1000 households and advanced econometric spike models that accommodate zero WTPs, the analysis revealed a significant positive average household WTP of approximately KRW 1832 (USD 1.33) per month. This finding indicates robust public support for the HVDC infrastructure, which is poised to play a critical role in South Korea’s RE transition by enabling efficient long-distance electricity transmission from the resource-rich southern regions to the SMA.

5.1. Key Findings

This study reveals statistically robust public support for South Korea’s KRW 11.5 trillion West Coast submarine HVDC transmission network, with a mean household WTP of KRW 1832 (USD 1.33) per month—translating to KRW 9.09 trillion (USD 6.58 billion) in aggregate annual benefits and a BCR of 2.44 when annualized over project lifespan. The OB spike model confirms estimate reliability despite 59.5% zero responses (protest/true zeros), with positive WTP drivers including SMA residence, progressive ideology, higher income, and RE awareness—while age and conservatism reduce support. Methodological robustness is evidenced by overlapping confidence intervals across base/extended models, nationally representative sampling (n = 1000), and absence of response effects, establishing benchmark welfare metrics for HVDC infrastructure.

5.2. Research Contributions

This pioneering application constitutes the first CV study of submarine HVDC transmission infrastructure, filling a substantive methodological gap in non-market valuation literature dominated by overhead line stigma analyses (Table 1). Methodologically, integration of the OB spike model improved estimation efficiency while accommodating zero WTP responses and mitigating associated biases, yielding robust welfare measures suitable for policy analysis. Collectively, these advances systematically assess the social acceptability and economic value of advanced grid technologies, bridging critical intersections among energy policy, environmental economics, and public utility regulation.

The academic significance derives from this novel CV application to submarine HVDC—a domain uncharted in prior empirical research (Table 1)—establishing foundational WTP benchmarks for cross-national comparisons. Although direct comparators remain unavailable given the intervention’s unique techno-economic context, the elicited mean monthly household premium of KRW 1832 (USD 1.33; 95% confidence interval: KRW 1625–2084) corresponds closely in magnitude to established valuations for analogous grid enhancements, such as USD 14.69 per month for reliability improvements in the United States and significant undergrounding premiums among South Korean households. This alignment affirms broadly comparable public willingness to invest in resilient, minimally intrusive infrastructure during energy transitions, positioning these findings as transferable reference points for global renewable grid expansion.

5.3. Policy Implications

From a policy standpoint, the empirical findings offer essential quantitative support to inform investment in infrastructure and regulatory decisions. The estimated societal benefits, when extrapolated nationally and compared against projected costs, indicate a substantial net positive value and a benefit–cost ratio of approximately 2.44 for residential consumers. These findings justify the strategic prioritization of submarine HVDC projects despite their higher upfront costs, as public willingness to bear associated electricity rate increases appears substantial. Importantly, the analysis also underscores the relevance of accounting for protest bids in valuation exercises, advocating for a conservative yet inclusive approach in policy design.

These results provide empirical justification for KEPCO and Korea Ministry of Climate, Energy and Environment (MCEE) to prioritize the West Coast submarine HVDC within the 11th Basic Plan for Electricity Supply and Demand (2024–2038), signaling consumer acceptability despite rate hikes (KRW 2000 per month per household equivalent). Policymakers should target communications at younger, urban, higher-income demographics while addressing conservative/rural skepticism through localized benefits (job creation, grid resilience for industries). The BCR greater than two supports fiscal backing via green bonds or RE levies, with WTP-based pricing reforms (tiered premiums) ensuring equitable cost recovery. For international audiences, findings validate submarine HVDC for variable RE-rich nations facing NIMBY barriers, complementing technical studies with social valuation evidence.

These results empower key stakeholders—including the KEPCO, MCEE, and the 11th Basic Plan for Electricity Supply and Demand oversight committee—with empirical evidence to justify accelerating the West Coast submarine HVDC project timeline (2025–2030). Specifically, KEPCO should implement tiered electricity surcharges calibrated to WTP heterogeneity (e.g., higher premiums for urban, progressive households), coupled with targeted subsidies for rural conservatives, to equitably recover the KRW 1.77 trillion residential cost share while minimizing opposition. MCEE can leverage the BCR > 2 to secure green financing via dedicated RE bonds or international climate funds, prioritizing submarine routes over delayed terrestrial alternatives amid NIMBY risks. For global applicability, findings guide variable RE-rich nations (e.g., United Kingdom, Japan) in balancing HVDC costs against social benefits, advocating hybrid valuation frameworks integrating CV with hedonic models for comprehensive infrastructure appraisal.

5.4. Future Directions

Looking ahead, future research should aim to refine the assessment of heterogeneous preferences across diverse demographic and regional groups by incorporating more granular spatial and socio-economic data. Furthermore, expanding the valuation framework to include dynamic temporal dimensions and potential environmental co-benefits could enhance the robustness of welfare estimates. Investigating consumer perceptions regarding alternative transmission technologies and integrating choice experiment methods could further enrich policy-relevant insights. Such advances will support the design of an optimized, socially accepted energy infrastructure that aligns with national decarbonization goals.

5.5. Sustainability and ESG Implications

This analysis aligns the submarine HVDC valuation with sustainability’s triple bottom line and ESG principles, elucidating contributions across environmental, social, and governance pillars integral to Sustainability’s scope.

• Environmental (E): The project facilitates ~20 TWh annual renewable curtailment recovery (equivalent to 10 MtCO₂e avoided emissions at South Korea’s 2025 grid intensity), enhancing grid decarbonization while minimizing terrestrial visual/land-use impacts versus overhead alternatives—yielding lifecycle carbon efficiencies documented in the HVDC literature.
• Social (S): Positive mean WTP (KRW 1832/household-month) signals broad acceptance for equitable energy access, though zeros highlight distributional tensions (66.6% citing tax burden aversion). Findings prescribe progressive tariff designs mitigating regressivity on low-income/rural households, alongside community benefit funds (e.g., 1–2% capital expenditure allocation), advancing SDG7 (affordable clean energy) and SDG10 (reduced inequalities).
• Governance (G): CV-derived legitimacy evidence supports transparent investment prioritization under South Korea’s Framework Act on Carbon Neutrality, countering past NIMBY delays via demonstrated welfare superiority (BCR = 2.44). Methodological rigor—national representativeness, spike model bias mitigation—exemplifies evidence-based decision-making, fostering institutional trust and replicable ESG-integrated appraisal frameworks for global infrastructure transitions.

These interconnections position submarine HVDC as a sustainability exemplar, balancing technoeconomic viability with stakeholder-inclusive governance.