Research

I study how public policy and technological change shape work, health, and regional economies in Korea, using large administrative and survey data.

Labor Economics Public Economics Health Economics Economics of AI & Automation Economic Geography

Publications 14

International journals

Welfare Expenditure Growth and its Impact on Elderly Suicide Rates

with Changsu Ko · Health Economics, 2026 · DOI  Corresponding author

Abstract

We investigate the causal effect of increased elderly welfare expenditure in South Korea on the suicide rate among older adults, using predicted expenditure changes induced by the Basic Pension reform as an instrumental variable. Our findings show that higher welfare spending explains 74.5% of the decline in the suicide rate among individuals aged 65+ between 2013 and 2019, with especially large reductions among men aged 80+. Household-level evidence indicates that while total income remained stable, consumption increased following the reform, particularly in transportation and leisure. We further show that the reform reduced unmet medical needs and stress and increased participation in community and leisure activities, while leaving depressive symptoms unchanged. Overall, the results are consistent with the interpretation that stronger public income transfer programs reduce income uncertainty and expand opportunities for consumption, leisure, and social participation, thereby contributing to lower elderly suicide rates.

Measuring Regional Economic Activity through Electricity Sales

with Changhui Kang, Sangkon Park, and Suyeon Cha · Energy Economics, 163: 109635, 2026 · DOI  Corresponding author

Abstract

Gross Regional Domestic Product (GRDP) is used as a proxy for regional economic activity, but its delayed publication limits its usefulness for real-time policy assessment. To address this, we propose electricity sales as a timelier, more reliable alternative in developed economies, where night lights—often used in data-scarce contexts—face limitations due to urban saturation. Using county-level electricity sales from Korea, we find that electricity sales consistently explain variations in GRDP across urban and rural areas and accurately predict GRDP one year ahead. This suggests that electricity sales can serve as a complementary indicator to GRDP in real-time policy evaluation.

Public Certification in the Age of Online Reviews: Signalling Efficacy in Tourism Markets

with Sangkon Park and Kyoungbo Sim · Applied Economics, 2026 · DOI  Corresponding author

Abstract

Tourism quality certification aims to reduce information asymmetry in service markets. This study empirically examines whether Korea’s national tourism certification programme serves as an effective economic signal of service quality. Using hedonic pricing models and discrete choice experiments, we jointly evaluate supply-side pricing behaviour and demand-side preferences of international tourists. Certified accommodations were found to charge an average premium of KRW 21,000, indicating that providers incorporate the certification in pricing. However, tourists’ willingness to pay is statistically indistinguishable from that for a placebo label, suggesting that the signal is weakly recognized on the demand side. Online reviews, by contrast, exhibit consistently higher explanatory power for perceived quality. These findings, interpreted through the lens of signalling theory, highlight a structural imbalance between provider incentives and consumer interpretation. To improve effectiveness, certification should target credence attributes—such as hygiene and safety—that consumers cannot easily assess, alongside stronger mechanisms for reliability, selectivity, and awareness.

Predicting Suicide Attempts among the Elderly: A Machine Learning Evaluation of a Staged Ideation-to-Action Approach Based on the Three-Step Theory

with Kahyun Lee · Journal of Affective Disorders, 404: 121514, 2026 · DOI  First author

Abstract

Suicide among older adults remains a major global public health concern, particularly acute in South Korea. Effective prevention requires prediction methods grounded in robust theoretical frameworks. This study evaluates the utility of a Three-Step Theory (3ST) based approach—which posits a sequential progression from suicidal ideation to attempts—using advanced machine learning on nationally representative data from South Korea’s Elderly Survey. We compare two predictive strategies: (a) a stepwise model that explicitly estimates suicidal ideation before attempts, and (b) a direct model that predicts attempts without this intermediate stage. Results strongly favor the stepwise approach (sensitivity = 0.531, specificity = 0.953, AUC = 0.828), suggesting both practical and theoretical advantages of modeling suicidal risk within a staged ideation-to-action framework. Error analysis further indicates that the two models capture complementary risk dimensions—progression-consistent versus early, non-progression markers—while a linear ensemble combining them achieves higher sensitivity with minimal loss of specificity. To address data imbalance, we evaluate oversampling (SMOTE) and undersampling techniques. Finally, SHapley Additive exPlanations (SHAP) analysis enhances interpretability and reveals distinct predictors across stages that are consistent with the sequential ideation-to-action logic proposed by the 3ST: depressive symptoms and life satisfaction in the first, and social connectedness and functional capacity in the second. Our findings provide new evidence consistent with 3ST and may inform the development of more precisely targeted interventions within large-scale suicide prevention policy.

How Will Automation Reshape Worker Well-Being? Evidence from a Highly Automated Economy

with Changkeun Lee and Olivia Hye Kim · Humanities and Social Sciences Communications, 13: 111, 2026 · DOI  Co-first author

Abstract

The rapid growth of industrial robotization in South Korea—one of the world’s most automated economies—offers a unique setting to study the psychological as well as economic impacts of automation. Drawing on county-level robot-intensity measures linked to individual data from the Community Health Survey and the Korean Labor & Income Panel Study, we employ two-way fixed-effects instrumental-variables models over one-, two-, and three-year intervals to isolate causal effects. We find that higher robot adoption swiftly elevates workers’ stress and depression, undermines self-rated health, and prompts increased alcohol consumption. In contrast, overall job-satisfaction ratings remain essentially unchanged, though select dimensions—daily task enjoyment, perceived meaningfulness, confidence in benefits, and long-term job commitment—erode in a staggered pattern consistent with our technostress framework. By extending technostress theory from digital tools to factory automation, our study reveals how shifting performance norms and cognitive burdens generate mental spillovers beyond the shop floor. These findings underscore the need for accompanying robotic investments with worker-centered supports—targeted upskilling, job redesign, and mental-health resources—to ensure that productivity gains do not come at the cost of human well-being.

Identifying High-Risk Elderly for Suicide Using Machine Learning

with Kahyun Lee · Empirical Economics, 69(4): 2467–2499, 2025 · DOI  First author

Abstract

South Korea faces a significant public health issue with its high elderly suicide rate. Despite efforts by the government to mitigate this issue, the challenge remains in effectively targeting the highest risk groups, which hampers policy implementation. This study introduces a machine learning-based approach to accurately identify high-risk groups for elderly suicide, aiming to enhance intervention effectiveness. Given the practical necessity of utilizing readily accessible variables for prediction—specifically, those that can be acquired not through face-to-face interviews but from administrative records—we categorized the variables into administrative, passive, and active data according to accessibility. We developed the predictive model with various machine learning methods, sampling techniques, and data combinations. Using only administrative data, the best model achieved an area under the curve (AUC) of 0.742. Adding passive data increased the AUC to 0.749, and using all data boosted it to 0.818. To enhance the identification of high-risk groups, we prioritized increasing the model’s sensitivity over accuracy. This adjustment maintained an 80% accuracy rate, achieving sensitivities of 0.510 with just administrative data, 0.516 with the addition of passive data, and 0.662 with all data types. A SHAP analysis indicated that disease status and personal income were most influential when using administrative data alone. With the addition of passive data, financial status also emerged as significant, while the depression index and life satisfaction became crucial in models utilizing the full data.

Local Employment Multiplier: Evidence from Relocation of Public-Sector Entities in South Korea

with Changsu Ko and Wookun Kim · Labour Economics, 88: 102515, 2024 · DOI  First author

Abstract

We exploit a series of public-sector entity relocations in South Korea as an exogenous source of variation in public sector employment to estimate the local employment multiplier. We find that the introduction of one public sector employment position increases private sector employment by one unit, primarily driven by the service sector. Consistent with existing literature, we document that the effect of public employment on private employment is highly localized. In addition to changes in private employment, we also discover that the relocations led to a positive net influx of residents into the treated neighborhoods; this effect is also localized. Lastly, by estimating the local employment multiplier for each relocation site, we document the heterogeneity of the local employment multiplier and provide suggestive evidence that this heterogeneity is shaped by the local economic environment’s capacity to accommodate additional general equilibrium responses.

The Effect of Spousal Loss on the Cognitive Ability of the Elder

with Changsu Ko · The B.E. Journal of Economic Analysis & Policy, 22(3): 475–525, 2022 · DOI  First author

Abstract

Given that economic growth is associated with increased life expectancy, declines in cognitive ability among the elder is a critical problem across the developed world. In this paper, we analyze the causal effect of the death of a spouse on the surviving spouse’s cognitive ability using the fixed effect model. The reliability of the estimates is enhanced by robustness checks, such as an event study model, to attend to potential threats to identification. Results show that, on average, spousal loss significantly reduces the cognitive functioning of the surviving spouse. We also study heterogeneity in the effect of spousal loss, finding that co-residing with children greatly mitigates the negative effect of bereavement.

In the media: 연합뉴스 · 조선일보 · 세계일보 · 경향신문 · 메디컬투데이

Public Finance Responses to COVID-19 in Korea

with Jaehyun Jung and Kyung-ho Song · National Tax Journal, 73(3): 879–900, 2020 · DOI  Co-first author

Abstract

This article summarizes the current economic status of the Republic of Korea amid the coronavirus pandemic and the government’s policy responses. The authors propose the areas of the government’s fiscal policy that entail in-depth research to prepare a better response in the future.

Korean journals (in Korean)

Employment Multiplier Effects of Local Government Fiscal Expenditure: A Panel Data Analysis of Korean Municipalities

지방재정지출의 고용승수 효과: 시군구 패널자료를 이용한 분석

with Wonmun Shin · The Journal of Eurasian Studies (유라시아연구), 23(3): 166–187, 2026 · DOI  First author

Abstract

This study estimates the effect of local government fiscal expenditure on regional employment—that is, the employment multiplier. Because Korea lacks reliable data on the regional allocation of central government spending, identification strategies that exploit exogenous variation in central government expenditure are constrained. This study therefore constructs a panel of 221 municipalities by combining 2016–2019 Local Finance 365 expenditure data at the city/county/district (si-gun-gu) level with employment data from the Census on Establishments, and mitigates the endogeneity of fiscal expenditure by controlling for a Bartik variable that allocates national-industry-level employment changes to each region. The results show that a KRW 100 million increase in local government spending raises employment by approximately 1.88 persons, with the effect larger in metropolitan cities and relatively larger for spending in the social welfare and health categories. In addition, the increase is concentrated among workers in other employment categories, which tend to offer relatively lower-quality jobs, suggesting that the qualitative dimension of job creation should be considered alongside its quantitative scale when evaluating the employment effects of fiscal programs.

Estimating the Effects of Agricultural Infrastructure Improvement Projects on Rice Production and Economic Benefits

농업 생산기반 정비사업의 쌀 생산량 증대효과와 경제적 편익 추정

with Myung-Heon Lee, Jinwoo Park, Cheong-Ryong Lim, and Miran Ahn · Korean Journal of Agricultural Management and Policy (농업경영·정책연구), 53(2): 327–347, 2026  Corresponding author

Abstract

In light of increasing abnormal weather events and shifts in the agricultural landscape, it has become crucial to evaluate the impact of agricultural infrastructure improvement projects and estimate their associated economic benefits. This study examines the productivity gains from such projects and estimates the resulting economic benefits. Employing a fixed-effects model, the analysis found that maintaining irrigation facilities increased rice production by an average of 5.2%. To quantify the economic benefits of this productivity boost, the methodology of Chavas (2017) for assessing food consumption benefits was utilized. The findings suggest that agricultural infrastructure projects contribute to an annual increase of 2.9 trillion KRW in food consumption benefits. These results offer valuable insights for enhancing the evaluation framework of agricultural SOC projects and informing future policy decisions.

Agricultural Infrastructure and Cropping System Shifts: An Economic Evaluation

농업 생산기반 정비사업이 작부체계에 미친 효과와 경제적 편익 추정

with Jinwoo Park, Miran Ahn, Sangkon Park, and Suyeon Cha · Journal of Rural Development (농촌경제), 48(4): 25–41, 2025 · DOI  First author Best Paper Award, 2025

Abstract

This study empirically examines the impact of the Agricultural Infrastructure Improvement Project, emphasising rural water development and cropping systems and their economic implications. Parcel-level cultivation data were constructed by integrating the Agricultural Management Entity database with GIS information on beneficiary areas, and the Difference-in-Differences framework of Callaway and Sant’Anna (2021) was applied to account for heterogeneous project completion times. The results indicated that, after facility completion, the share of rice cultivation within the same village decreased by 0.374 percentage points, while upland crop cultivation increased by 0.144 percentage points, reflecting a structural shift in cropping patterns. A benefit–cost analysis incorporating this shift produced a B/C ratio of 0.02. Although modest in conventional terms, the findings highlight the policy relevance of agricultural infrastructure in promoting crop diversification, enhancing farm household income and stability, and providing empirical evidence for feasibility assessments and policy decisions.

Types of AI Technology and Labor Market Outcomes: Evidence from Population Level Employment Data

AI 기술 유형별 노동시장 영향 분석: 한국 고용보험 자료를 활용한 실증 연구

with Hyeongjun Bang · Journal of Korean Economic Analysis (한국경제의 분석), 31(2): 129–184, 2025 · DOI  First author

Abstract

This paper examines the impact of AI technology on the Korean labor market by comparing the effects of traditional AI (Webb Index) and generative AI (Felton Index). Using employment insurance data covering nearly the entire wage-earning population, we document that traditional AI primarily affects older workers in technology-intensive sectors such as manufacturing and construction, while generative AI is more associated with younger and female workers in knowledge- and service-oriented sectors like education and information technology. To estimate the relationship between AI exposure and labor market outcomes, we implement a long-difference model, controlling for the influence of traditional automation technologies. The results indicate that exposure to traditional AI is associated with declines in employment, whereas exposure to generative AI has no such effect and is instead linked to wage gains, suggesting a productivity-enhancing role. These findings highlight the heterogeneous labor market implications of different AI technologies and offer empirical foundations for designing targeted employment policies.

Economic Effect of Local Currency (Local Consumption Coupon): Focusing on the Impact on Sales of Small Businesses

지역화폐(지역사랑상품권)의 경제적 효과: 소상공인 매출액에 미친 영향을 중심으로

with Kyungho Song · Journal of Korean Economic Analysis (한국경제의 분석), 27(3): 1–53, 2021  Corresponding author

Abstract

Local currency was introduced to revitalize underdeveloped local economies and protect small businesses by limiting the regions and industries that accept it. With financial support from the Ministry of the Interior and Safety, almost all local governments had adopted a local currency by 2020. We assess whether the policy met its goals using a theoretical model and empirical analysis. A linear city model shows that when one local government introduces a local currency, others are compelled to follow, and that when all do, smaller local governments may be worse off than before. Empirically, a triple-difference design that uses municipalities without a local currency as an additional control group finds no evidence that issuance raised overall sales of eligible retail businesses. The effects vary widely across industries, and the benefits were concentrated in a few sectors.

In the media: 한국일보 · KBS · 주간경향

Working Papers 7

The Economic Effects of a Rapid Increase in the Minimum Wage: Evidence from South Korean Experiments

with Taeyoung Doh, Kyooil Kim, Sungil Kim, and Kyungho Song · Federal Reserve Bank of Kansas City Working Paper  R&R, Oxford Economic Papers

In the media: 연합뉴스 · 뉴스1 · 중앙일보 · 문화일보

Shrinking Cohorts, Fixed Earmarks: An Information Experiment on Public Preferences over Education Finance in Korea

with Kyoung Hoon Lee and Sungho Park  R&R, Journal of Economic Behavior & Organization

Cash Transfer and Fertility: Evidence from an Ultra-Low Birth Rate Society

with Kanghyock Koh  R&R, Review of Economics of the Household

Automation and the Composition of Jobs: Empirical Evidence from Population Data

with Dongik Kang  R&R, The B.E. Journal of Economic Analysis & Policy

In the media: 한겨레 · 연합뉴스

Rising Temperatures, Falling Trust: The Impact of Heat Exposure on Institutional Trust

with Dongyoung Kim  Submitted

Can Minimum Wage Increases Reduce Wage Inequality? Empirical Evidence from Korea

with Chung Choe, Yoonyoung Choi, and Minuk Song  Submitted

The Local Economic Effects of Fixed-Link Infrastructure: Township-Level Evidence from Korean Sea-Crossing Bridges

with Jinwoo Park, Changsu Ko, and Sangkon Park

Work in Progress 5

The Erosion of Entry-Level Jobs: AI Adoption and Firm Outcomes from Korea’s AI Voucher Program

with Hoon Choi

The Employment Effects of Extending the Statutory Retirement Age: Evidence from the 2013 Labor Reform in South Korea

with Changhui Kang

In the media: 파이낸셜뉴스 · 세계일보 · 연합뉴스

The Aggregate and Regional Effects of Minimum Wage Changes

with ShinHyuck Kang and Joonseok Oh

Regional and Aggregate Fiscal Multipliers in Korea

with ShinHyuck Kang and Joonseok Oh

Does Fiscal Decentralization Amplify or Mitigate the Impact of Population Aging? Evidence from Local Government Spending in Korea

with Changsu Ko