Research
Publications
Race, immunity, and lifespan: Unraveling the effect of early-life exposure to malaria risk on lifespan
Economics and Human Biology, 54, 101382. 2024. (with Sok Chul Hong) [Link]
Show abstract
We investigate a historical experience to measure the long-term effect of malaria on lifespan among infected survivors and identify a factor that mitigates malaria's effect. Using a sample of Union Army veterans born during the mid-19th century and their lifetime records, we show that exposure to high risk of malaria at birth or in early life substantially shortened their lifespan. The legacy of exposure to malaria is robust while controlling for lifetime socioeconomic and health conditions, fixed effects, and considering selection bias. Additionally, we include the US Colored Troops sample of black veterans to analyze racial differences in the effect of malaria exposure on lifespan. Exposure to malaria did not lead to a shorter lifespan among black veterans. Evidence suggests that genetic immunity to malaria in black veterans might contribute this heterogeneity.Mortality, Nutrition and Height in Early Colonial Korea
Review of Economic History, 67(2), 197–225. 2018. (in Korean) [Link]
Show abstract
The determinants of stature can be changed in one country by the change of cultural and economic situation. The cases of western countries in 19th century are good example. Therefore, identifying the determinants is needed to understand the change of stature in colonial Korea. In this study, I used Prisoners' cards made in prisons that provide a lot of information about their characteristic. What I found is that disease environments were important factors to determine people's stature in early colonial Korea. The change of the mortality rate that reflected disease environment explained the statural change.Working Papers
Avoiding the Risk, Bearing the Cost: Evidence from General Health Screening in Korea during the COVID-19 Pandemic
submitted (with Sok Chul Hong) [PDF] · Last updated: 2026-08-06
Show abstract
This study examines the unintended health consequences of voluntary responses to COVID-19. We focus on general health screening in Korea, using administrative data that link medical claims and screening records. At the national level, screening rates declined markedly in 2020, the first year of the pandemic, relative to counterfactual trends. Complementing this aggregate pattern, individual-level analysis reveals notable heterogeneity: declines were larger among those with higher predicted risk of chronic disease. We then assess the consequences of forgone screening, employing propensity score matching and event study designs. Our estimates show that, had they been screened, individuals who missed screening would have been more likely to initiate care for chronic diseases. The costs of missed screening were especially large among those at higher predicted risk of chronic disease. Such delays in management led to more advanced conditions at the time of care initiation. Our findings show that, even without strict quarantine policies, voluntary responses to infection can undermine preventive care, disproportionately affecting high-benefit groups. This underscores the importance of balancing infection control with the continuity of preventive care during health crises.Expanding Free Influenza Vaccination to Primary School Children: Evidence from Korea
[PDF] · Last updated: 2026-07-25
Show abstract
School-aged children play a central role in seasonal influenza transmission, with high attack rates and frequent contacts that amplify community spread. Publicly funded vaccination programs for this age group, however, remain uncommon internationally. This paper evaluates Korea's 2018 expansion of the National Immunization Program, which extended free vaccination to all children ages 5-12. Using nationally representative survey data, I find that difference-in-differences estimates show a sharp and persistent increase in vaccination coverage among newly eligible children. To evaluate the expansion's effect on influenza-related healthcare utilization, I turn to administrative claims data. Because pre-expansion utilization trends differ across the two age groups, I adopt a triple-difference design that exploits variation in vaccine match quality across time, and find a meaningful reduction in influenza-related claims during high-match months. A back-of-the-envelope calculation indicates that direct healthcare-cost savings offset a substantial share of the program's budgetary cost.Work in Progress
Negative Vaccine Information and Adverse Event after COVID-19 Vaccination: Evidence from the Official Recognition of Causality in Korea
(with Sangbeom Kim, Dae-il Kim, Sok Chul Hong) [Slides] · Last updated: 2026-04-24
Show abstract
Information plays a crucial role in shaping vaccine-related behaviors. However, because the timing of exposure to vaccine information can trigger responses at different margins, focusing solely on vaccination uptake risks underestimating the overall effect of information. We exploit a major information shock—the Korean government's recognition of a causal link between the Pfizer-BioNTech vaccine and myocarditis/pericarditis—to investigate its impact on healthcare utilization after vaccination. Using the K-COV-N cohort, an administrative dataset linking vaccination records and health insurance claims, we implement a difference-in-differences framework comparing individuals vaccinated around the announcement with not-yet-vaccinated controls. We find that exposure to the announcement increased adverse event–related healthcare utilization by about one percentage point from a 4.5% baseline.Evaluating Population-Based Cognitive Screening: Evidence from Korea
(with Jooho Moon, Sok Chul Hong)
Show abstract
Early diagnosis is the entry point to dementia care, and detection typically relies on brief cognitive screening as a first-stage filter. However, because screening cutoffs are derived to optimize diagnostic accuracy in case-control samples, and because pursuing confirmatory diagnosis after a positive screen is a voluntary choice, development-setting accuracy does not transfer cleanly to routine operation—real-world evidence is needed to inform the divergent policy choices health systems have already made. We exploit the cutoff of the KDSQ-C, a cognitive screening questionnaire embedded in Korea's National Health Screening Program, to investigate the effect of a positive screen on dementia-related diagnosis and care. Using the National Health Information Database, an administrative dataset linking health screening records, medical claims, and long-term care records, we implement a regression discontinuity design in the screening score. We find that crossing the screening cutoff increases dementia-related healthcare utilization within one year by about two percentage points from a 7.2% baseline, driven almost entirely by mild cognitive impairment. Among those who initiate care, screening-induced utilizers are no less likely than always-takers to receive a confirmed diagnosis or to sustain care, suggesting no systematic over-detection at the margin.Harder to Reach Care: Transport Access and Health Care Utilization in Old Age
(with Sangbeom Kim)
Show abstract
Non-Marital Births and Marriage Postponement in South Korea
(with Dae-il Kim, Sok Chul Hong)