Towards A Better Measurement of Redlining: Digitizing the Federal Housing Administration’s Real Property Survey Maps

• In Event: Beyond the Red Lines: New Directions in the Study of Housing Policy and Racial Inequality

Abstract

This project provides a new resource to study the implications of mid-20th mortgage discrimination by digitizing a heretofore unstudied set of pre-WWII 140 Federal Housing Administration (FHA) maps. While the Homeowners’ Loan Corporation’s (HOLC) maps are more commonly studied and used to delineate areas of “redlining”, recent studies have complicated their role in shaping inequal housing and neighborhood outcomes. Using computer vision tools augmented by human refinement, we create a dataset of FHA rent grade designations from the Real Property Survey maps. The FHA used this data, most importantly the average rental rate and percentage of non-White residents on the block, to define economic areas in cities by gauging economic and racial homogeneity. The results of this effort are visible as red lines dividing cities on many of these maps. Our descriptive analysis compares these FHA maps with HOLC neighborhood grade designations. We find city-level disparities in the areas that were surveyed and graded by the FHA and HOLC, as well as regional disparities in which there are relatively more maps in the Western and Midwestern regions of the U.S. While there are a wide range of FHA rent grades for each HOLC grade designation, though no clear biases, indicating that the use of HOLC maps to proxy redlining may result in unbiased but imprecise estimates.

Author

• Wenfei Xu, University of California, Santa Barbara
  Presenting Author