Social Security pays spousal and survivor benefits, which are based on a spouse’s earnings record. Black women have less access to these benefits and receive smaller amounts conditional on access because of differences in marital histories and spouse earnings. I combine micro-level evidence with a structural life-cycle model to evaluate how these benefits affect labor supply and welfare among Black and White households. I find that spousal benefits discourage married women’s participation more than survivor benefits do. Moreover, these benefits reduce labor-force participation more among White married women than among Black married women. Replacing spousal benefits with higher own-worker benefits raises average welfare for both racial groups, with larger gains for Black households. In contrast, replacing survivor benefits with higher own-worker benefits increases welfare only for White households. Black households’ welfare declines because they value these benefits more due to higher widowhood risk and shorter life expectancy. The results indicate that shifting resources from spousal to own-worker benefits while preserving survivor benefits would reduce racial welfare inequality.

This paper studies how AI affects productivity and labor markets in the Middle East and North Africa and the Caucasus and Central Asia. We develop a general-equilibrium model combining occupational AI exposure, country preparedness, and wage rigidity. AI raises output in the countries studied, but its effects on wages and employment differ across countries and workers by skill, age, and gender. Productivity gains increase labor demand, while the replacement of workers’ tasks can reduce employment for some groups. Country preparedness and labor-market institutions shape whether productivity gains translate into higher wages, more jobs, and lower unemployment.

This project examines how AI adoption at work and at home shapes labor supply, gender disparities, and welfare. We develop a general equilibrium model incorporating unpaid household work and endogenous labor-force participation to assess how AI-related time savings translate into paid work or leisure, depending on employment opportunities and care responsibilities. The analysis explores how AI preparedness and access to childcare and eldercare can broaden the gains from technological change.

This paper examines the geography of U.S. venture capital during its formative decades of the 1960s and early 1970s. Drawing on newly assembled data on Small Business Investment Companies (SBICs), VC firms, PhD production, and industrial R&D laboratories, we show that specialized risk capital emerged where innovation-supporting institutions and educated populations were already concentrated. State-level analyses find that population, urbanization, and federal R&D spending predict SBIC presence. At the metropolitan level, wartime OSRD-funded patents and PhD production predict both whether and how extensively venture capital took root. These historical findings underscore that policies expanding venture capital supply without the prerequisite human capital infrastructure face fundamental limitations.