Research

My research follows families through major life events. When someone is diagnosed with dementia, what happens to their household's work and earnings? How do people fare financially in the years around a spell of homelessness? And when an adult involved in crime joins a household, how do the children do in school?

Job market paper

The Labor-Market Costs of Dementia for Patients and Families: Evidence from Colombia

with Andrea Franco Hernandez and Christian Posso

Serious illness can reduce a person’s ability to work while raising the household’s need for care. Other household members may respond by working less to provide care or by working more to replace lost income. We study these competing responses after a first recorded diagnosis of Alzheimer’s disease or a related dementia (ADRD) in Colombia. Linking administrative health records to monthly formal-employment records and household rosters, we use a stacked difference-in-differences event study to compare newly diagnosed patients and their households with similar people and households that never receive a diagnosis. Two years after diagnosis, patients’ formal employment is about 10 percentage points lower for women and 14.8 points lower for men. Because symptoms may affect work before diagnosis, we interpret this as the decline surrounding diagnosis rather than the effect of disease onset. We find no evidence that other household members increase formal work to offset these losses. Instead, the share of other members with formal jobs is about 5 percentage points lower, and their formal earnings fall as well.

Paper (PDF)

Working papers

Homelessness and the Persistence of Deprivation: Income, Employment, and Safety Net Participation

with Bruce D. Meyer, Angela Wyse, Gillian Meyer, and Derek Wu
NBER Working Paper 32323, 2024

Homelessness is arguably the most extreme hardship associated with poverty in the United States, yet people experiencing homelessness are excluded from official poverty statistics. This paper provides the first national estimates of their income, employment, and safety net participation based on administrative data. Starting from the first large and nationally representative sample of adults recorded as sheltered and unsheltered homeless, taken from the 2010 Census, we link restricted-use longitudinal tax records and administrative data on safety net programs, housing assistance, and mortality. Nearly half of these adults were employed in the year they were observed as homeless, and nearly all either worked or were reached by at least one safety net program. Nevertheless, their incomes remained low for the decade surrounding an observed period of homelessness, suggesting that homelessness tends to arise in the context of long-term, severe deprivation rather than large and sudden losses of income.

NBER ↗

State Income Taxes and Retiree Migration: Evidence from Household-Specific Tax Incentives

States compete for retirees through tax preferences for Social Security, pensions, and other retirement income, but it is unclear whether these policies induce retirees to move or only shape where movers settle. Using CPS ASEC data from 1982–2019 and TAXSIM, I calculate the after-tax income each retiree household would receive in every state given its own mix of income, and estimate conditional logit models of location choice. Among retirees overall, after-tax income has little effect on where they live, and distance from the origin state is a much stronger predictor. Among interstate movers, however, retirees are significantly more likely to choose states where they would receive higher after-tax income. State tax policy appears to do little to induce retirees to relocate, but it can shape competition for those who have already decided to move.

Household Criminality and the Intergenerational Transmission of Disadvantage: Evidence from Medellín, Colombia

with Christian Posso

This paper examines how children’s outcomes change when an adult who becomes involved in crime enters their household. Using linked household survey, education, and crime records from Medellín, we compare children who experience the arrival of such an adult with children who experience the arrival of an adult who does not commit a crime. Children exposed to a criminally involved adult are 9–10 percentage points more likely to drop out of school and 1–1.3 percentage points more likely to have a recorded criminal offense, and those who take the high-school mathematics exam score 0.05–0.06 standard deviations lower. The estimates are similar after matching on child, household, and mover characteristics.