Working Papers
This paper estimates the causal effects of federal "redlining" – the mapping and grading of US neighborhoods by the Home Owners’ Loan Corporation (HOLC) – with a novel empirical strategy. In the late 1930s, a federal agency developed color-coded maps to summarize the financial risk of granting mortgages in different neighborhoods, together with forms describing the presence of racial and ethnic minorities as "detrimental". Our analysis exploits an exogenous population cutoff: only cities above 40,000 residents were mapped. We employ a difference-in-differences design, comparing areas that received a particular grade with neighborhoods that would have received the same grade if their city had been mapped. The control neighborhoods are defined using a machine learning algorithm trained to draw HOLC-like maps using newly geocoded full-count census records. HOLC maps had a negative impact on neighborhoods colored in red, reinforcing patterns of urban disadvantage, with a particular burden on African-American communities. In the short term, we estimate a reduction in property prices and an increase in the percentage of African American residents. Significantly higher percentages of Black Americans can be detected in D communities in the long term, up until the early 2000s. For property prices, we find negative effects in C and D neighborhoods until the early 1980s. Our empirical results show that a government-supplied, data-driven information tool can coordinate exclusionary practices and amplify their consequences.Racial residential segregation in U.S. cities rose sharply during the first half of the twentieth century. We study whether early federal public housing contributed to this rise by examining the first projects built by the Public Works Administration in the mid-1930s, most of which were racially designated. Using newly assembled data on project locations linked to full-count Census records from 1910 to 1950 and comparing built to planned-but-not-built projects, we find that public housing reinforced the racial composition of project sites but had little effect on surrounding neighborhoods. The results suggest that early public housing played a limited role in shaping neighborhood-level segregation.Work in Progress
The Long Term Effects of Exposure to Non-Traditional Family Structures.
Environmental inequalities, such as unequal exposure to pollution and climate risks, persist across racial and socioeconomic groups in the United States. This paper re-examines the role of the Residential Security Maps created by the Home Owners’ Loan Corporation (HOLC) in the 1930s, which graded neighborhoods according to perceived mortgage risk and have been widely linked to long-run racial segregation and environmental disadvantage. A common view holds that these maps not only reinforced residential segregation but also directly shaped the spatial distribution of environmental hazards, including air pollution, flood risk, and extreme heat. We evaluate this claim using a causal framework that combines machine-learning predictions of counterfactual HOLC grades in unmapped cities with a spatial difference-in-differences design. Our results confirm that the maps modestly increased racial sorting and segregation, consistent with prior work. However, we find no evidence that HOLC mapping independently affected the siting of environmental or climatic hazards. Differences in air pollution, flood risk, heat exposure, and mortality across historical grades are quantitatively similar in mapped and unmapped cities. Single-mother households have become common in the US over the past fifty years. Economists, sociologists, and psychologists have documented that children from single-headed families have lower intergenerational mobility because of a lack of resources and the type of parenting they receive. However, little is known about the effects of children from single-mother families on their school peers. Taking advantage of the Add Health panel data structure, I estimate the effect of this feature of the adolescents’ social environment on educational achievement and long-run labor market outcomes. My identification strategy is based on cohort-to-cohort variation in the percentage of children without a father figure within a school. The preliminary estimates indicate that exposure to peers with a higher rate of father absence does not have much of an effect on education, employment, or wages.