Research library

Redlining, then and now

In 1937 a federal agency colored Philadelphia's neighborhoods green, blue, yellow and red to tell lenders where mortgages were safe. The red areas were the ones where Black families lived. The lines still predict who is shot, who owns a home, and how long people live.

What the Home Owners' Loan Corporation did

Between 1935 and 1940 the Home Owners' Loan Corporation, a New Deal agency, produced "residential security" maps for more than 200 American cities. Field agents, working with local bankers and realtors, graded every neighborhood on a four-step scale and wrote an area description to justify it. Philadelphia's map was completed in 1937.

A"Best." New, homogeneous, in demand.
B"Still desirable." Stable, but older.
C"Definitely declining." Aging housing, "infiltration" noted.
D"Hazardous." Marked red. Routinely the neighborhoods where Black residents lived.

The area descriptions are explicit about race. The presence of Black families, and often of recent immigrants, was recorded as a detrimental influence and pushed a neighborhood down the scale. Banks and the Federal Housing Administration then used the grades, and the same logic, to decide where they would and would not lend for decades. That is the practice the word redlining names.

Redlining was one tool among several in Philadelphia. The Federal Reserve Bank of Philadelphia has documented thousands of racially restrictive covenants written into city deeds between 1919 and 1932. Urban renewal in the 1950s through 1970s then cleared neighborhoods such as the Black Bottom in West Philadelphia, displacing between 4,496 and more than 10,000 residents for the University City Science Center. The map, the deed and the bulldozer worked together.

What the 1937 lines still predict

75%
of Philadelphia's 2019 homicides occurred in areas graded C or D on the 1937 map, according to the City Controller's 2020 report on structural racism.
1919 to 1932
The years covered by the Philadelphia Fed's spatial dataset of racially restrictive covenants found in city deeds.

The Controller's report overlays Census disadvantage indicators and residential sale prices on the HOLC grades and finds the same pattern across measures: the red and yellow areas of 1937 are, with few exceptions, the areas of concentrated poverty, low home values and violence today. The Department of Public Health's reports show the same geography for life expectancy and maternal and infant outcomes.

None of this is unique to Philadelphia. It is unusually well documented here, which is what makes the city a good place to learn how to read it.

Check your own block

The address lookup draws the 1937 grades on a live map of the city. Type an address, or tap the map, and it reports the grade for that spot alongside the property's current assessment, its recorded deed history, and counts of demolitions, unsafe-building orders, shootings and reported crime within a quarter mile. Each grade links to the original area description at Mapping Inequality, so you can read what the field agent wrote about your street.

What to do with what you find

If your home sits in an A or B area, you are likely living on the appreciation that policy manufactured. If it sits in C or D, you are living with what it withheld. Either way, the ways to contribute page has concrete next steps, and the task force tracker shows where the city's own repair process stands.

Reproduce the analysis

The joining of 1937 grades to present-day outcomes is a short, repeatable computation. The method below is implemented in a Python script in this project's companion research folder and produces a summary table, a per-tract file and a local map. It is deliberately simple so that anyone can check it.

  1. Get the crosswalk. The University of Richmond publishes HOLC polygons already split by 2020 Census tract, with each fragment's share of the tract. Filter to Philadelphia, Pennsylvania and grades A through D.
  2. Assign each tract a grade. Sum the fragment shares by grade. Keep tracts at least half inside the 1937 map. The dominant grade is the one with the largest share; the HOLC score is the area-weighted average where A is 1 and D is 4.
  3. Pull the outcomes. From the 2022 American Community Survey five-year estimates at tract level: median household income (B19013), tenure (B25003), race (B02001), poverty (B17001) and median home value (B25077). Negative values are suppressed estimates and are dropped.
  4. Join and summarize. Merge on tract GEOID. Report means and medians of income, home value, ownership rate, Black share and poverty rate by dominant grade, plus the correlation of each with the HOLC score.
  5. Run it. The Census API now requires a free key, requested at api.census.gov. Set it in your environment and pass it on the request.
    pip install geopandas pandas requests pyarrow
export CENSUS_API_KEY=your_key
python3 philly_redlining_analysis.py

The analysis was last attempted on November 9, 2026. The Census API refused the request for lack of a key, so no results are published here yet. This page will carry the table once a keyed run completes; until then, the Controller's figures above are the published reference.

Sources