I am a Research Associate at the Korea Institute of Public Finance
My research interests lie in Labor and Health Economics, with a focus on Applied Econometrics for Causal Inference and the application of Machine Learning and Large Language Models in empirical economic research.
CV ; Contact me at: byeorhee@gmail.com
Publications
“COVID-19 Vaccine Misinformation in Middle Income Countries,” (w. Jongin Kim, Aditya Agrawal, Jiaxi Wu, Veronika J. Wirtz, Traci Hong, Derry Wijaya), In Proceedings of The 2023 Conference on Empirical Methods in Natural Language Processing. Link
"A Cohort Analysis on the Patterns of Birth in South Korea" (w, Wonsik Ko), 여성경제연구, 2019, 16(1), 1-25. Link
Working Papers
“Parental Absences, Children, and Their Children” Sep 2024. Link
A substantial proportion of children in the U.S. experience parental absence, a significant event whose impact may extend to future generations. However, methodological challenges and data limitations have hindered analysis on the multigenerational implications of parental absence. This paper addresses this gap by integrating the National Longitudinal Survey of Youth 1979 (NLSY79) with its Child and Young Adult (CYA) data, where the mothers of CYA respondents are the female respondents in NLSY79. I propose novel specification tests and apply them to augmented inverse probability weighting to estimate the average treatment effects, thereby enhancing the reliability of the results. I find suggestive evidence that the impacts of parental absence on children before they turned 18 are transmitted to their grandchildren. Additionally, I find that parental absence has a substantially larger negative impact on the educational attainment and early wages of the younger generation from the CYA compared to its effects on the older generation from the NLSY79.
“Son Preference: Childbearing and Asian American Women's Labor Market Outcomes,” May 2024. Link
This study employs an IV based on son preference to investigate the impact of having more than one child on the labor supply of Asian American females in the IPUMS USA dataset. Under son preference, mothers with a first-born girl are more likely to have another child in the hope of having a son. Therefore, I use the sex of the first child to instrument for having more than one child. The IV estimates reveal that having more than one child substantially reduces women's employment, number of weeks worked, weekly working hours, probability of working in the last year, and annual earnings. I provide suggestive evidence supporting the exclusion restriction and monotonicity assumptions. The results are robust to allowing for defiers and heterogeneous treatment effects.
“Cultural Norm, Shotgun Marriage, and Birth Outcomes,” (w. Sokchul Hong, and Wonsik Ko), Oct 2021. Draft available upon request.
This paper examines the impact of shotgun marriages on birth outcomes, with a focus on the mediating role of Confucianism in South Korea using administrative birth records. Although previous studies have investigated the effects of marital status on birth outcomes, the cultural mechanisms underlying these effects remain largely unexplored. Confucianism, which has shaped South Korean society for over a thousand years, stigmatizes sexuality and places a strong emphasis on chastity. This cultural influence helps explain why out-of-wedlock births and births from shotgun marriages have historically been rare, as they defy societal norms. South Korea's recent generational shift away from Confucian values, along with substantial regional variation, provides a unique setting to investigate how culture moderates the effects of shotgun marriages on birth outcomes. We find that shotgun marriages are associated with a 9-gram decrease in birth weight and a 0.09 percentage point increase in the probability of low birth weight. Using the third-child sex ratio (TSR) as a proxy for Confucian cultural strength, we find that in regions and years with stronger Confucian influence, the negative impact of shotgun marriages is more pronounced. A one standard deviation increase in TSR further reduces birth weight by 6 grams and raises the probability of low birth weight by 0.14 percentage points. Alternative explanations, such as differences in prenatal care, selective abortion, and selection into shotgun marriages do not fully account for these findings.
“Comparative Analysis of Social Media Moderation Strategies in Combating COVID-19 Misinformation,” (w. Traci Hong, Nina Mazar, Pujan Paudel, Gianluca Stringhini), May 2024.
The rapid spread of misinformation during the COVID-19 pandemic posed significant challenges for social media platforms, prompting them to implement various moderation strategies. This paper examines Instagram’s and Twitter’s differing approaches to mitigating COVID-19 misinformation and the resultant impact on user sentiment extracted using a large language model, FLAN-T5. Specifically, we focus on Instagram's unilateral moderation strategy, introduced on February 9, 2020, and Twitter's evolving moderation measures, culminating in a curated moderation mechanism in May 2020. We found that Instagram’s moderation was more effective than Twitter’s in both promoting positive sentiments and reducing negative sentiments toward trustworthy organizations and people.
Research in Progress
“How Does Education Affect Financial Planning Horizon?” August 2023.
“Measuring Temporal Discounting Independently from Measuring Utility,” March 2021.