Research

Working Paper

The First Era of American Federal Public Housing (1940-1960) — Effects on Neighborhoods

Runner-up for the Best Student Paper Prize at the 2025 Urban Economics Association North American Meeting

This paper examines how early American public housing affected the income and demographic composition of targeted neighborhoods. At its outset in the mid-1930s and roughly until 1960, public housing combined two missions: provide “decent” housing to working-poor families, and help cities curb urban decay through slum clearance. It was the first nationwide attempt to rely on housing policy to limit the extent of urban poverty concentration. To investigate its effects on neighborhoods, I construct a novel data set with development locations and construction dates in 44 cities. Using a shift-share-style instrument framework, I instrument for local public housing construction by exploiting exogenous aggregate fluctuations in the number of federally funded public housing units across regions and over time, interacted with a neighborhood’s likelihood of receiving public housing conditional on its conditional on its start-of-period characteristics. I find that in neighborhoods most likely to be targeted by the policy, public housing construction increased the shares of families in the lower half of the income distribution, with deciles 3 to 5 increasing by 5 to 6 percentage points each. The bottom decile share declined by about 7 percentage points, although this estimate is not statistically significant. I also show that the average family income below the city median increased by about 20 percent in typical recipient neighborhoods. Together, these effects indicate that public housing reduced the concentration of families in the lowest decile by drawing in families with slightly higher incomes. Put differently, targeted neighborhoods became “mixed-income” areas, which differs considerably from the high levels of poverty concentration that were downstream of changes in public housing policy beginning in the 1960s. These results underscore the way housing policy shapes neighborhood residential composition and income segregation. [pdf]

Work in Progress

Homeownership Dynamics and the Lock-In Effect — Evidence from the Federal Reserve’s Monetary Policy Tightening

With Amine Ouazad, Romain Rancière and Qitong Wang

We examine how search frictions and increasing mortgage interest rates interact to shape housing market dynamics, particularly through the ‘lock-in’ effect of fixed-rate mortgages. We develop and numerically solve a dynamic search and matching model that accounts for three key features of housing markets: (1) the dual buyer-seller role of most market participants, (2) costly housing search, and (3) widespread use of 30-year fixed-rate mortgages in the US. Using transaction-level data from Los Angeles and San Francisco to discipline the model, we aim to shed light on how high interest rates affect housing prices and their dispersion.

Local Knowledge Spillovers from Publicly Funded Research: The NIH Budget Doubling of 1998–2003

With Gautier Lenfant

Publicly funded research in universities is widely believed to spur innovation, yet whether large expansions of established federal research programs yield substantial returns remains an open question. This paper examines the effects of the NIH doubling–the dramatic expansion of National Institutes of Health funding from the mid-1990s to 2003–on local corporate innovation. Using a shift-share difference-in-differences design that compares innovation across 64 commuting zones containing R1 universities based on their historical NIH funding exposure (1985–1995), we find that a 1 percent increase in pre-doubling funding generated approximately $0.9 million in additional yearly funding at the 10-year horizon, along with 1.1 additional research projects and 5.7 additional NIH-supported publications. This research expansion spurred substantial local corporate innovation, with effects concentrated in hard science patenting. At 10 years, a 1 percent increase in pre-doubling exposure generated 1.12 additional corporate patent filings, representing a 0.24 percent increase. We identify two channels through which knowledge flows from universities to the corporate sector: direct commercialization by former NIH principal investigators and knowledge transfer through their research collaborators. Together, these channels account for roughly one-third of the total innovation response, with the indirect channel through coauthors an order of magnitude larger than the direct channel through principal investigators. These findings demonstrate that major funding expansions of longstanding federal research programs can generate substantial spillovers to corporate innovation.