A windfall damages health, but only for the winners who did not need it. I exploit full-population Korean tax records linked to clinically measured biomarkers, medical claims, occupation, and death records for prime-age employees who won prizes of varying size and at different timing. Six-year mortality shows no effect in the full sample. That null is, however, an average over a group that is harmed and a group that is not. Above median baseline income, the prize becomes consumption: smoking rises in every post-treatment period, drinking rises on both margins, blood glucose and blood pressure deteriorate, and mortality rises by 175 deaths per 100,000, a 35 percent increase. Below the median, the same prize becomes an exit from physically demanding work: manual employment falls durably, blood pressure drifts down with it, and no biomarker deteriorates. Income damages health precisely where it relaxes no binding monetary constraint. This finding implies that the causal effect of income runs opposite to the cross-sectional gradient, so the gradient must be generated elsewhere.

We examine how physical health and mortality vary with income for prime-age adults. To do so, we leverage population-wide, individual-level cardiovascular health exams linked to income tax records and subsequent mortality for a random 8 percent of Koreans ages 47-57. For both men and women, there is a negative disease-income gradient for high blood pressure, high blood glucose, and high body mass index, but not for high cholesterol. Disease-income gradients are steeper for women than men, and are concentrated in the top half of the income distribution.  Subsequent mortality-income gradients are also concentrated in the top half of the income distribution, but are steeper for men than women. Although physical health measures are highly predictive of subsequent mortality, differences in these health measures by income explain very little of the income-mortality gradient, even for cardiovascular mortality. Our findings suggest that something other than differences in disease burden at prime age drives most of the income-mortality gradient at older ages.

We combine household survey micro data, tax data, and national accounts to construct annual pretax income inequality series for South Korea, which is coherent with macro aggregates. We provide the distribution of pretax national income over the time period from 1933 to 2022, with detailed breakdown by income composition in the years from 1996 to 2022. Our new series demonstrates that Korean top income shares decreased substantially from the 1930s up until the mid-1960s, following various wars, independence and land reform policies. Income inequalities then stabilized against the backdrop of rapid economic growth in the 1970s and 1980s, and decreased even further from the late 1980s onwards until the onset of the Asian Financial Crisis in 1997, which contradicts the ``Kuznets Curve''. In the aftermath of the crisis, income inequality worsened due to the rise of tax-exempted capital income concentration at the top. Compared to other East Asian countries, South Korea exhibits relatively lower levels of income inequality in terms of higher bottom 50% income shares, mostly due to a more equal distribution of national income growth at the stages of economic take-off in the 1980s, even though income concentration at the very top has strikingly worsened over the last two decades, with its top 1% income shares in 2022 returning back to the peak only observed during the colonial era.

This paper constructs the first long-run estimates of top wealth concentration in South Korea from 1970 to 2021, using newly assembled estate tax tabulations and applying a simplified mortality multiplier method. The series uncovers a pattern that diverges sharply from the gradual postwar increase in wealth inequality seen in many Western economies. Korea experienced two distinct regimes: a two-decade period of low and stable concentration from 1970 to 1990, when the top 0.1 percent held a roughly stable 3 to 5 percent of wealth, followed by a substantial rise beginning in the late 1990s to a new, higher plateau of around 10 percent. This abrupt “Great Unleveling”, plausibly linked to institutional and market changes surrounding the 1997 Asian Financial Crisis, coincided with a shift in elite portfolios from land-based to financial assets. Independent property-tax records show that land concentration stayed flat across this break, indicating that the rise originated in financial rather than landed wealth. In international perspective, Korea moves from a low-inequality profile typical of developing economies to a moderately high-inequality regime similar to contemporary France and Japan. The findings highlight how sudden institutional breaks, rather than gradual trends alone, shape the long-run distribution of wealth. 

This paper uses extended series on income and wealth inequality from the World Inequality Database (WID) covering all world regions over the 1800-2025 period, together with new series on hourly productivity and human capital expenditure, to revisit the relationship between equality and development, with a much broader comparative and historical perspective than previous studies. Over the long-run, we find a strong positive association between equality and productivity. Our proposed interpretation is that the rise of inclusive “social-democratic” institutions (including extended access to human capital, public services and democratic participation) led both to more equality and higher productivity, particularly in Western and Nordic Europe. We discuss the implications for future sustainable development strategies. 

Distance to hospitals is everywhere confounded with disadvantage: people who live far from care are also poorer, older, and worse-insured, so the effect of distance itself has never been cleanly estimated. I use the phased relocation of South Korea’s central government to Sejong (2012–2014), a planned city with no tertiary hospital until 2020, as administratively assigned variation in distance to care, linked to full-population medical care insurance records on utilization, provider location, biomarkers, health behaviors, and cause-specific mortality. I estimate how the response to distance differs between care that can be scheduled — which patients can defend by traveling — and care that cannot, and its effect on health outcomes.

Estimates of retirement's health effects point in opposite directions: mortality typically rises, while self-reported health improves. I use Korea's cohort-by-cohort increase in the normal and early retirement age, applied to full-population records linking cause-specific mortality to biomarkers, health behaviors, and medical claims, to observe the physiological margin that sits between these two literatures and to test which mechanism reconciles them.

We study a national resident physician strike in South Korea that disrupted physician supply for eighteen months, providing a rare natural experiment on how labor adjusts within and between hospitals, and how patients are prioritized and re-allocated in the face of sustained shortages.