Redlining and the FHA (United States, 1933 to 1968 and after)

Federal agencies in the 1930s graded neighborhoods for lending risk and the FHA's manual told appraisers to avoid "inharmonious racial groups"; economists now disagree over whether the HOLC maps or the FHA's own rules did the most to shape lending, and how much of today's wealth gap traces to them.

The record

The agencies and the maps

• The Home Owners' Loan Corporation (HOLC) refinanced mortgages in the 1930s and, in the later 1930s, had local appraisers produce "residential security" maps that graded neighborhoods, in four grades. The scans and area descriptions in the Mapping Inequality project come mostly from the National Archives City Survey Files, 1935 to 1940. The scans are public domain. We could open only the project's data page and a metadata stub, not its explanatory text, so the number of cities and the grade colors are not confirmed here from that site. (Mapping Inequality data page, University of Richmond; opened; the main page blocked automated reading)
• The Federal Housing Administration (FHA) insured private mortgages. Its Underwriting Manual set the rules for rating a property and its neighborhood. The HUD archive page says the manual "references policies of racial discrimination." (HUD USER, Underwriting Manual page; opened; the manual PDF itself could not be read)
• Quotations attributed to the manual, as shown on a Trinity College teaching page: paragraph 229, "Protection From Adverse Influences," lists "prevention of the infiltration of business and industrial uses, lower-class occupancy, and inharmonious racial groups"; paragraph 284 praises "The prohibition of the occupancy of properties except by the race for which they were intended." The same page paraphrases paragraphs 228, 233 and 289 (racial deed restrictions viewed as favorable; valuators told to check for incompatible racial or social groups; schools not attended by "inharmonious racial groups"). We did not open the manual, so the wording and paragraph numbers rest on that page. (Trinity College, FHA Underwriting Manual; opened)
• Date of the manual: the Trinity page dates it 1936. The HUD page calls it a 1938 document. Both could be correct for different editions. Not reconciled.

What happened in court and Congress

• Shelley v. Kraemer, 334 U.S. 1 (decided May 3, 1948): private racially restrictive covenants do not themselves violate the Fourteenth Amendment, but state courts may not enforce them. Chief Justice Vinson wrote for the Court; three justices took no part. (Justia; opened)
• The Fair Housing Act, Title VIII of Public Law 90-284, enacted April 11, 1968, bars refusing to sell or rent, refusing to negotiate, or discriminating in terms, conditions or privileges because of race, color, religion, sex, familial status or national origin. The page notes amendments in 1974 and 1988; we did not open them, so which grounds were in the 1968 text is unverified. (42 U.S.C. 3604, Cornell LII; opened)

Timeline

The debate: did the HOLC maps drive FHA underwriting?

The long-standing account, associated with historians such as Kenneth Jackson (1985), is that HOLC's maps were the model the FHA and private lenders then followed. Two bodies of research challenge or qualify it.

| Study | Method | Finding on the maps | Source | |---|---|---| | Hillier, "Redlining and the Home Owners' Loan Corporation," Journal of Urban History (2003) | Address-level Philadelphia mortgage data, GIS and spatial statistics | HOLC grades do not explain most differences in lending, except interest rates, which were higher in red-colored areas. Archives suggest lenders avoided red areas before the maps, the maps were not widely distributed, and lenders had other sources of risk information | Abstract (Penn repository), opened | | Fishback, Rose, Snowden and Storrs, "New Evidence on Redlining by Federal Housing Programs in the 1930s," NBER Working Paper 29244 (2021); Journal of Urban Economics (2022) | Over 16,000 loans in Baltimore, Peoria and Greensboro | The maps had "little effect" on either agency's geographic lending. The FHA "crafted and implemented its own redlining methodology prior to the HOLC." FHA's exclusion of low-income urban neighborhoods began in its first years | NBER paper, opened | | Aaronson, Hartley and Mazumder, "The Effects of the 1930s HOLC Redlining Maps," American Economic Journal: Economic Policy 13(4), 355 to 392 (2021) | Boundary design comparing neighborhoods on either side of grade lines; propensity score methods; a city population cutoff | Lower-rated neighborhoods had lower homeownership, lower house values and rents, and more segregation in later decades. The authors attribute this to reduced credit access and disinvestment | RePEc abstract, opened |

Details from the Fishback paper (NBER, opened):

| City | Lender | Share of loans in A+B areas (percent) | Share in D areas (percent) | |---|---|---| | Baltimore | FHA | 67.3 | 1.6 | | Baltimore | HOLC | 32.2 | 23.2 | | Peoria | FHA | 53.1 | 2.2 | | Peoria | HOLC | 14.6 | 19.4 | | Greensboro | FHA | 73.9 | 1.2 | | Greensboro | HOLC | 34.8 | 17.1 |

• Sample: 10,145 HOLC and 3,540 FHA loans in Baltimore, 982 HOLC and 556 FHA in Peoria, 624 HOLC and 238 FHA in Greensboro.
• About 10 percent of HOLC borrowers in Baltimore were Black, against about 4 percent of homeowners there in 1930. About 2 percent of Black owner-occupants in Baltimore received FHA loans against about 8 percent of White owner-occupants. Greensboro had one FHA loan to a Black borrower and Peoria none.
• The authors say studies by Aaronson and by Xu found about 60 to 80 percent of city area received the same rating from both agencies, and treat that as weak evidence about influence. They acknowledge the FHA may have consulted the HOLC maps after they were shared around 1937 to 1940, and that they did not study HOLC's treatment of applicants or its foreclosure sales.
• The Fishback paper's reading of Hillier's work: it "reinforced" his Philadelphia findings. The two sides therefore agree that lower-graded areas still received HOLC loans, and disagree on whether the maps caused long-run harm.

Measured effects: homeownership and wealth

• Aaronson, Hartley and Mazumder report lasting effects on homeownership, values and segregation (above). They describe these as "meaningful and lasting." The Fishback authors do not dispute that the FHA excluded low-income urban neighborhoods; they dispute the source of the exclusion. Neither side we opened claims to measure the whole racial wealth gap.
• Federal Reserve Survey of Consumer Finances, 2022: median family wealth was $285,000 for White families, $44,900 for Black families and $61,600 for Hispanic families. Black homeownership was about 45 percent. (Federal Reserve FEDS Note, Oct. 2023; opened) The note does not attribute any share of the gap to the 1930s programs.
• Census Bureau, second quarter 2026: homeownership was 74.5 percent for non-Hispanic White alone householders, 45.4 percent for Black alone and 48.1 percent for Hispanic (any race). (Census Housing Vacancy Survey press release; opened; the Hispanic figure was read from a table)

What this case shows

1. The criteria were written down, and they named race. The FHA manual, as quoted on the teaching page, treated "inharmonious racial groups" as an adverse influence and praised racial deed restrictions. A rule that decided who could get a federally insured loan was a government document, which people can now read. Whether it was public in 1936 or 1938 is not shown in what we opened.
2. The people graded had no way to see or contest the grade. Hillier's abstract says the HOLC maps "were not widely distributed." If so, neighborhoods were rated without notice to the people living in them. This is a point about access, not about motive.
3. The most careful studies disagree on the cause, not on the exclusion. Fishback and colleagues find the FHA excluded low-income urban neighborhoods early and on its own method. Aaronson and colleagues find lasting neighborhood effects from the map grades. Both findings describe a federal credit system that left some neighborhoods out. (NBER; AEJ abstract)
4. Remedy came late and came by court and statute. Shelley (1948) barred court enforcement of covenants but left private covenants lawful; the Fair Housing Act came in 1968, decades after the manual. (Justia; Cornell LII)
5. Gaps are measurable today. Median wealth and homeownership by race differ by large amounts in Federal Reserve and Census data. What those sources do not show is how much traces to the 1930s programs.

The strongest objections, and our answers

1. Objection: the HOLC maps are overblown; HOLC lent widely, including in lower-graded areas and to Black borrowers. The Fishback data support this: in Baltimore, 23.2 percent of HOLC loans were in D areas, and HOLC borrowers there were 10 percent Black against 4 percent of homeowners. Mitigation: we show the table. We do not say HOLC maps alone caused the pattern.
2. Objection: the FHA only reflected existing private market practice. Hillier's archival evidence suggests lenders avoided red areas before the maps. Mitigation: that supports the view that private practice came first, but the FHA manual still set federal underwriting rules in writing. Who followed whom is not settled by what we opened.
3. Objection: later effects like lower homeownership and values may come from other causes: income, local schools, jobs, discrimination by private actors, or later policy. Mitigation: Aaronson and colleagues use a boundary design to address this, and the Fishback paper questions their link to the maps. We report both, and do not assign a share of today's wealth gap to any one cause.
4. Objection: the FHA expanded homeownership for millions. Mitigation: Not examined in our fetches. The Fishback tables show FHA loans concentrated in A and B areas (67.3 percent in Baltimore), consistent with benefits that went mostly to favored neighborhoods, but we did not measure national totals. Unverified.
5. Objection: the Trinity page is a teaching page and Wikipedia-level sources should not carry the manual's words. Mitigation: agreed. We flag the quotations as secondary and ask readers to check the PDF on the HUD site. We could not open that PDF.

Open questions: what would change our minds

• The FHA manual PDF was not opened. Paragraph numbers and exact wording of paragraphs 229, 233, 284 and 289 are unverified, as is the edition year (1936 or 1938).
• The Mapping Inequality explanatory text was not read, so the number of cities mapped and the grade scheme are unconfirmed here.
• We did not open the Hillier paper itself (only its abstract) or the Aaronson paper (only an abstract), so effect sizes are not reported.
• Whether private lenders had access to the HOLC maps, and when, is the core dispute. Archival proof either way would change Section 3, objection 2.
• No measured share of today's racial wealth gap attributable to 1930s federal programs was found.
• The 1968 Act's original enforcement powers and the 1988 amendments were not opened.
• Evidence on benefits of FHA insurance nationally (total loans, share to Black families) was not fetched.

Sources

Evidence status and correction history

Key claims

Correction history

Found an error or a better source? We keep every correction on the record with its date and reason. Nothing here is quietly rewritten.