Abstract
Many health behavior decisions involve immediate benefits or costs, while their consequences for future health may not appear until much later. Dynamic models of health behavior, however, typically assume that individuals follow a common, fully rational decision rule. I study how much this assumption matters for explaining persistent health behavior and predicting policy responses. I estimate a dynamic model of health behaviors in which individuals may follow fully rational, myopic, present-biased, or subjective-expectations decision rules. I first estimate the model separately under each rule and then allow the four rules to coexist as permanent latent types. Using longitudinal data from the Korea Labor and Income Panel Study, I find that the separately estimated models fit observed patterns of smoking and preventive health checkups similarly well, but imply different explanations for these behaviors. In the joint estimation, the myopic type accounts for 38.3% of the population, while the fully rational type accounts for 22.9%. Individuals with higher self-reported impulsivity have higher posterior probabilities on the myopic type. In a counterfactual that doubles cigarette prices, overall smoking falls under all four decision rules, but the size and age profile of the response differ across types. More generally, heterogeneity in economic behavior may reflect more than differences in preferences and constraints; individuals may also differ in the decision processes they use.
With Cecilia Rodriguez Ingold
SSM - Population Health, 2026
Abstract
This article evaluates the impact of social isolation on older adults' mental health, overcoming endogeneity concerns through an instrumental variables approach. Using data from the English Longitudinal Study of Ageing collected during the COVID-19 pandemic, we exploit self-reported National Health Service (NHS) contact advising vulnerable individuals to shield as an instrument for social isolation. We investigate whether the intensity of isolation, measured by its duration and strictness, exacerbates mental health decline. Our findings indicate that social isolation significantly worsens anxiety and depression, with greater isolation intensity amplifying these effects. The results remain robust across alternative model specifications and when relaxing the exclusion restriction.
Financial Planning Review (Korean Academic Society of Financial Planning), 12(2), 2019
Abstract
This study explores the relationships between Korean baby boomers’ retirement preparation and a) subjective financial literacy, b) psychological factors and c) demographic factors. In particular, this study focuses on how imperfect understanding of economic problems and psychological factors can affect retirement preparation and be associated with inaction. Psychological factors included future time perspective, aging anxiety, and subjective life expectancy. For demographic factors, education level, number of dependent children, subjective health status, age, gender, and income levels were considered. Using data collected from the Aging Institute of Seoul National University (n = 2933), this study analyzed how such factors are related to Korean baby boomers’ retirement savings rate, sufficiency of economic preparation, calculation of required retirement funds, and retirement planning and information gathering. Furthermore, Korean baby boomers’ economic preparation behavior was categorized into three types (preparing and sufficient, preparing but not sufficient, and not preparing) for further analysis of inactive behavior and shortage in economic preparation. OLS and logistic regression results showed that subjective financial literacy and psychological factors have significant relation to both economic preparation for retirement and retirement planning and information gathering. People with higher subjective financial literacy showed higher level of sufficiency of economic preparation, were more likely to have calculated required retirement funds, showed less inaction, and tended to be more active in information gathering. People with longer future time perspective and lower aging anxiety were more sufficiently prepared, whereas low subjective life expectancy was highly related to inaction. Demographic factors such as education level, income level, number of dependent children were important predictors of economic preparation. People with bad subjective health status showed inaction behavior, which can be associated with subjective life expectancy via ‘sense of control’.
Abstract
In health economics, there is ongoing debate over how bounded rationality influences risky health behaviors. We explore this issue by examining the impact of strategic sophistication, one’s capacity for backward induction and higher order reasoning, on health behaviors and health outcomes using experimental measures from the 2017 KLIPS (Korea Labor Income and Panel Studies) survey. To incorporate the dynamic nature of health behavior decision making, we jointly estimate a set of equations derived from a behavioral dynamic model of health behaviors with bounded rationality. The results show that people with sophisticated backward induction rationality engage more in health screening, smoke less heavily, and are healthier. Simulations show that backward induction sophistication can lead to differences in late life screening rates by up to 5 percent and shift smokers in their 30s to 40s toward lighter smoking by approximately up to 20 percent. Sequentially sophisticated individuals maintain higher health trajectories, while naive peers are up to approximately 12 percent more likely to be in poor health as they age.
Abstract
As populations age, healthy aging has become increasingly important, emphasizing good physical, mental, and cognitive health in later life. While previous research often focuses on medical treatments, community-based resources that support health maintenance, such as senior centers, have received less attention. This study examines how senior center utilization causally impacts older adults’ health, using nationally representative Korean data and instrumental variable approaches to address endogeneity. The findings, derived from partially identified bounds robust to misspecification, show meaningful health improvements, including enhanced cognitive function, reduced depression, and fewer loneliness-related suicidal ideations, particularly among older men. Although effects on physical health outcomes varied, senior center users generally experienced fewer chronic illnesses and greater muscle strength. Further analysis indicates these benefits potentially arise from reduced falls, improved access to meal services and health checkups, and increased participation in learning activities. The results emphasize senior centers as valuable community resources in promoting healthy aging and potentially reducing related economic burdens.
With Jane Fruehwirth, Adin Hammond, and Sydney Pugh
With Donna Gilleskie and Sean Zhang
With Yoobin Kim and Jieun Lee
KLI Panel Working Paper, 2018, 5