Coloring Inside the Lines: The Jagged Legacy of the HOLC Neighborhood Risk Maps

TL;DR

This study investigates whether the historic HOLC mortgage risk maps still accurately reflect racial residential boundaries in 2010, using novel Monte Carlo methods and entropy analysis to compare original and simulated maps across eleven US cities.

Contribution

The paper introduces a new Monte Carlo approach to generate plausible alternative maps and assesses their similarity to HOLC maps, revealing persistent racial boundary patterns over 80 years.

Findings

HOLC maps significantly align with racial boundaries in 2010

Monte Carlo simulations show the maps' racial boundaries are not random

Methodology is modular and adaptable for future spatial analyses

Abstract

There has been a large body of work exploring the discriminatory nature of the home mortgage risk maps produced by the Home Owners' Loan Corporation in the late 1930s. However, little attention has been paid to the question of whether these maps are still descriptive of racial residential boundaries in their cities 80 years after their creation. To address this gap, Markov Chain Monte Carlo, previously unutilized in the relevant literature, is employed to randomly generate many plausible alternative mortgage security maps. Then, the racial evenness of the HOLC maps and the generated maps is compared using Shannon's entropy. These findings indicate that the HOLC maps are significantly descriptive of the precise racial residential boundaries prevalent across eleven US cities in 2010. The methodology used here is highly modular and reproducible, allowing for future work measuring different outcome statistics, locations, and time periods.