Disadvantaged by Design: The Long-Term Effects of Redlining on Census Tract-Level Outcomes in American Cities

Abstract

This thesis seeks to examine the role that redlining policies initiated by the Home Owners’ Loan Corporation (HOLC) in the late 1930s played in fundamentally shaping American neighborhoods and the groups that inhabited them. After establishing the basic relationships between redlining and a variety of socio-economic outcomes in 1960, I introduce a novel research design to estimate the causal effects of redlining at the census tract-level. By constructing a set of “control” tracts in non-surveyed cities just below an exogenous population threshold that determined which cities were surveyed by the HOLC and by comparing their trajectories to similar tracts that were actually surveyed, I estimate the causal effect of redlining using a regression discontinuity design. This design may better estimate the effects of redlining by avoiding potentially problematic assumptions of random grade assignment or continuity across borders of tracts that received different HOLC grades. My results suggest that redlining significantly increased the proportion of black and other non-white individuals in redlined tracts, devalued housing, and decreased educational attainment in redlined tracts relative to similar tracts in cities that were not surveyed by the HOLC. My insignificant findings on a variety of other housing characteristics contrast my basic associational results, which show strong relationships between redlining and these variables. This suggests that while disadvantaged neighborhoods were declining relative to better neighborhoods within cities on these dimensions, this decline was the result of factors other than redlining.