Redistributive politics is often framed as a dispute about fairness: who deserves what, and how much society should care about inequality in itself. I argue that a second, analytically distinct motive is just as central; concerns about inequality as an externality. When the income distribution affects outcomes people value, such as crime, trust, democratic stability, or growth, redistribution is not only about justice, but also about internalizing social costs. I first lay out a preference and welfare framework that separates fairness and externality terms. I then summarize evidence that citizens across the world hold strong beliefs about inequality's consequences, and that shifting these beliefs causally shifts support for redistribution. Finally, I show that countries differ in redistributive rhetoric: some debates rely on fairness language, while others more often invoke externalities. The broader claim is that cross-country differences in redistribution likely reflects different beliefs about both fairness and inequality's societal consequences.

This talk asks whether economic inequality reflects not only differences in skills and constraints, but also systematic heterogeneity in preferences. Drawing on three linked papers, we scale incentivized experiments to population-representative samples and link experimentally elicited time, risk, and social preferences to administrative data.

We show that patience predicts wealth accumulation: more patient individuals save more and accumulate substantially higher wealth, accounting for a meaningful share of wealth inequality. Preferences also shape human capital investments: patient individuals choose longer programs with steeper earnings growth, while risk-averse individuals select fields with lower earnings dispersion, even after controlling for ability and family background. Finally, risk tolerance strongly predicts criminal behavior, particularly property crime.

Taken together, the evidence suggests that preference heterogeneity generates persistent divergence in economic outcomes and may imply heterogeneous—and potentially regressive—effects of incentive-based public policies.

Abstract We study whether policy can shift gendered beliefs, norms, and labor market outcomes by exploiting a major expansion of earmarked paternity leave in Denmark. The reform generated large first-stage effects, substantially reallocating leave from mothers to fathers. Using a regression discontinuity design combined with new survey data linked to administrative records, we show that the reform makes parents more supportive of paternity leave, shifts gender-role beliefs in a progressive direction, and reduces perceived differences in childcare ability. The reform also narrows gender gaps in earnings and hours worked. The earnings gap falls by 33pp in the first year following childbirth (during leave) and by 2.8pp in the second year (after leave). These results demonstrate that policy can meaningfully influence beliefs, norms, and gender inequality. On the other hand, earmarking restricts families’ ability to allocate leave freely and lowers leave satisfaction, highlighting a central trade-off inherent in paternalistic policies.

Socioeconomic inequality in longevity is typically measured using a single socioeconomic indicator such as education or income. We combine multiple indicators - education, income, occupation, wealth, and IQ scores - and apply machine learning to measure inequality in longevity. Using Danish population-wide data spanning 40 years, we track mortality for the 1942–44 birth cohorts from age 40 onwards to estimate life expectancy by socioeconomic status. Individuals at the top of the socioeconomic distribution live nearly 25 years longer than those at the bottom. The socioeconomic gradient in life expectancy becomes 50–150% steeper when using multiple indicators.