The GI Bill of 1944 (Servicemen's Readjustment Act)

The record

• July 28 1943 to January 1944. President Roosevelt proposes veterans' benefits in a July 28, 1943 speech and messages to Congress on Oct. 27 and Nov. 23, 1943; the bill is introduced in the House on Jan. 10, 1944 and in the Senate the next day. (VA, "GI Bill" history)
• 1943 to 1944. The American Legion "designed most of the main features" and pushed the bill through Congress; Legion publicist Jack Cejnar coined "the GI Bill of Rights." (National Archives)
• Spring 1944, drafting. Rep. John Rankin of Mississippi chaired the House veterans committee. History.com reports he pushed for state rather than federal administration and got it, and that he stalled the committee over the $20-a-week unemployment provision until the American Legion flew in an absent member to break the deadlock. (History.com, Blakemore) Historian Stephen Ortiz told PolitiFact that Rankin feared jobless pay would discourage Black veterans from working. (PolitiFact, 2012)
• June 12-13, 1944. Senate (June 12) and House (June 13) approve the final bill. (VA)
• June 22, 1944. Roosevelt signs Public Law 78-346. (National Archives; VA)
• Terms as enacted (per the July 1944 Social Security Bulletin summary of the new law, SSA):
• Eligibility: generally 90 days of service (or earlier discharge for disability in line of duty) and a discharge "under conditions other than dishonorable."
• Education: tuition and fees up to $500 per ordinary school year paid to the school, plus a subsistence allowance of $50 a month ($75 with dependents).
• Loans: federal guarantee of 50% of an approved home, farm or business loan, up to $2,000, at no more than 4% interest.
• Unemployment: $20 a week for up to 52 weeks (the "52-20 club"), claimed through public employment offices and paid by cooperating state agencies, with federal reimbursement.
• The law contains no nondiscrimination clause in the summaries and text excerpts reviewed. A copy of the act text hosted by George Mason University includes a section stating that no federal department "shall exercise any supervision or control, whatsoever, over any State educational agency" or educational institution. (GMU text; note this copy shows later-amended amounts, $65 subsistence and a $4,000 real-estate guarantee cap, so it is not the original 1944 text.)
• April 1945. A VA directive says a blue (undesirable) discharge for homosexual acts is "under dishonorable conditions and a bar to entitlement," although the statute only barred dishonorable discharges. (Jennifer Dominique Jones, Ambivalent Affinities, UNC Press excerpt)
• Dec. 1, 1941 to June 30, 1945. The War Department issued 48,603 blue discharges, 10,806 (about 22%) to Black soldiers, per a January 1946 memo by Maj. Gen. Edward Witsell cited by Jones. (UNC Press excerpt) GMU's veterans clinic puts Black soldiers at 6.5% of the Army and 22.2% of blue discharges. (GMU Law veterans clinic)
• October 1945. The Pittsburgh Courier, a Black newspaper, reports concern that the Army and VA were using blue discharges to deny Black soldiers GI benefits. (Learning for Justice reprint)
• January 1946. A House Military Affairs Committee report calls the blue discharge's treatment "anomalous and illogical," says it had "much of the practical effect of a dishonorable discharge" without a court-martial, and recommends automatic review. (GMU Law veterans clinic)
• Feb. 12, 1946. Hours after his honorable discharge, Black Army veteran Isaac Woodard is beaten and blinded in Batesburg, South Carolina; he testified that police chief Lynwood Shull struck his eyes with a nightstick. The state declined to prosecute; an all-white federal jury acquitted Shull. The case influenced President Truman, who created the President's Committee on Civil Rights (report "To Secure These Rights," October 1947) and on July 26, 1948 ordered the armed forces desegregated. (Military Times)
• March 1946. Only 6 of 246 on-the-job training programs for veterans in Atlanta had Black participants (Onkst, cited by Turner and Bound). Onkst also found about a dozen Black VA counselors for all of Georgia and Alabama and none in Mississippi. (Turner and Bound, NBER w9044, pp. 6-7)
• July 2, 1946. On his 21st birthday, Army veteran Medgar Evers, his brother Charles and other Black World War II veterans try to vote in Decatur, Mississippi, and are turned away by an armed white mob. (Mississippi Today; National WWII Museum)
• July 17-20, 1946. Army veteran Maceo Snipes, reportedly the only Black person to vote in the Taylor County, Georgia primary, is shot at his home the next day and dies July 20. A coroner's jury accepted the shooter's self-defense claim. The FBI later ruled the killing unrelated to voting and the Justice Department closed its review in 2010; EJI describes the attackers as a white mob and notes family recollections of Klan threats before the election. (PBS Frontline; EJI)
• October 1946. Of 6,500 former soldiers placed in nonfarm jobs by Mississippi's employment service, 86% of skilled and semiskilled jobs went to whites and 92% of unskilled jobs to Blacks (Katznelson's figures, as reported in a 2005 New York Times review of his book). (NYT review, copy hosted at alansinger.net; same figures in History News Network)
• 1947. Veterans make up 49% of college admissions at the peak. (VA) Southern Black colleges lack space: a survey of 21 found 55% of veteran applicants turned away for lack of room (vs. about 28% for all colleges), and Olson estimated 20,000 Black veterans turned away. (Turner and Bound, p. 9) Katznelson estimates Black colleges lacked places for about 70,000 Black veterans in 1947. (NYT review copy)
• New York and northern New Jersey. Katznelson reports fewer than 100 of 67,000 GI-Bill-insured mortgages supported home purchases by nonwhites. (NYT review copy)
• By 1952 to 1956, scale. About 8 million veterans received education benefits within seven years: about 2.3 million in college, 3.5 million in school training, 3.4 million in on-the-job training. By 1955, 4.3 million home loans worth $33 billion; veterans bought 20% of new homes built after the war; $14.5 billion spent on education and training by 1956. (National Archives) The VA counts 7.8 million of 16 million World War II veterans in education or training by July 25, 1956, and nearly 2.4 million VA-backed home loans to WWII veterans from 1944 to 1952. Less than 20% of funds set aside for unemployment pay were used. (VA)
• 1954. A Senate Finance Committee report notes most service in the Women's Army Auxiliary Corps (before the WAC gave women full military status in 1943) does not count for VA benefits because "the 1943 act did not purport to grant retroactive military status or benefits," and advances a narrow fix for disabled WAAC members only. (Senate Report 83-1910; NPS)
• 1955. In Longernecker v. Higley, 229 F.2d 27 (D.C. Cir.), the court says the VA "should have felt himself bound" by the Army's classification of a blue discharge as other than dishonorable, but holds courts could not review the VA's benefit denials. (opinion text, FlexLaw)

What this case shows

1. Local control of a federal benefit tracked local political power. The statute was race-neutral, but it routed jobless pay through state agencies, home loans through private lenders, and education through colleges that states controlled, with a clause barring federal supervision of state educational agencies and institutions. In the South, those state and local institutions were run by officials chosen in elections from which most Black citizens were excluded. (SSA 1944 summary; GMU text; History.com on Rankin)
2. The best causal study finds no college gain for Southern-born Black veterans. Turner and Bound find the war and GI Bill substantially raised education for white men and for Black men born outside the South, but had "little effect on collegiate outcomes" for Black veterans likely limited to Southern colleges, widening the Southern Black-white gap. (NBER w9044; NBER Digest)
3. Capacity and gatekeeping, not demand, were the binding limits. 55% of veteran applicants to 21 Southern Black colleges were turned away for lack of space; Mississippi had no Black VA counselors; 6 of 246 Atlanta job-training programs had Black participants. (Turner and Bound, pp. 6-9)
4. Home loans depended on banks' willingness, and that was where gaps were starkest. The VA guaranteed loans but did not make them. Katznelson's Mississippi (two of 3,229) and New York/New Jersey (under 100 of 67,000) figures, and the National Archives' own statement that Black veterans "often could not get bank loans" and were largely kept out of white suburbs, point the same way. (National Archives; Katznelson) A 2025 American Journal of Sociology study using restricted VA data finds the loan guaranty raised homeownership for both groups but "exacerbated extant racial inequalities." (Agbai, AJS 2025)
5. Discretionary exclusions fell hardest on Black soldiers. Blue discharges, which the VA treated as a bar despite the statute, went to Black soldiers at roughly three times their share of the Army. (UNC Press / Jones)
6. Returning veterans who sought the vote met force. Evers and fellow veterans were driven from the Decatur courthouse in July 1946; Snipes was shot the day after voting; Woodard was blinded in uniform. Those who controlled ballots in the South also controlled the offices that administered veterans' benefits there. (Mississippi Today; PBS; Military Times)

The strongest objections, and our answers

1. Objection: The GI Bill helped Black veterans a great deal, and outside the South it raised Black college attainment as much as white. Turner and Bound estimate about 0.4 extra years of college for non-Southern Black men, compared with about 0.3 for white men. (NBER Digest) Interviews collected by Suzanne Mettler include "Every black we knew used the GI Bill." (Jacobin excerpt) Mitigation: This is correct and should be said plainly. It actually supports the local-control argument: the same federal law worked where Black citizens could vote and attend integrated or better-resourced schools, and fell short where they could not. The claim is about uneven delivery, not that Black veterans got nothing.
2. Objection: Measured in dollars, Black veterans received as much or more. Eden (2023) estimates government spending per veteran of $877 for Black veterans vs. $784 for white veterans (1944 dollars), with Black veterans getting more in education and unemployment benefits and slightly less in housing ($128 vs. $135). (Eden, Explorations in Economic History 2023) Mitigation: Eden herself finds the cash-equivalent value was 18% higher for white veterans ($554 vs. $468), that Black veterans were channeled more into non-college training (21% vs. 12% in "other schools"; 11% vs. 17% in college), and she stresses "significant uncertainty." Spending is not the same as value received. Her work is national and does not test North vs. South.
3. Objection: The North-South difference in Turner and Bound is not statistically strong. Their test of equality between Southern and non-Southern Black effects has p-values of 0.10 (years of college) and 0.21 (completion). (NBER w9044, Table 5) Mitigation: Fair. The Southern estimate is indistinguishable from zero and the non-Southern one is positive, and the qualitative record (capacity limits, counselors, job programs) points the same way. But write "the evidence suggests," not "proves."
4. Objection: The famous loan figures are thin. The Mississippi "two of 3,229" figure comes from a 1947 Ebony survey of 13 cities that we could not locate, and is often misstated as statewide. National survey data compiled by Logan and Parman (1980 National Survey of Veterans) show 42.8% of Black veterans vs. 38.0% of white veterans reporting a VA home loan, across all eras. (Logan and Parman slides) Mitigation: Use the Mississippi figure only with its source ("an Ebony survey of 13 Mississippi cities, cited by Katznelson") and do not generalize it nationally. The 1980 survey mixes later cohorts, including Vietnam-era veterans when lending rules had changed. The 2025 AJS study, using VA administrative data, still finds the WWII program widened racial housing gaps.
5. Objection: Discrimination came from Jim Crow generally, not from the GI Bill's design. Segregated colleges, redlining banks and hostile employers existed independently. Mitigation: Partly true. The design point is narrower: Congress chose to deliver the benefit through those very institutions and barred federal supervision of educational ones, rather than building federal delivery or a nondiscrimination rule. Note too that later federal civil rights laws, not the GI Bill, changed the underlying conditions.
6. Objection: Linking Snipes's death to voting is disputed. The FBI ruled it unrelated to voting. (PBS) Mitigation: Report both the official finding and EJI's account of Klan threats and the timing. Lead with the Evers episode, which is not disputed.

Open questions: what would change our minds

• Original Ebony survey. Locate the 1947 Ebony article (issue, page, method). If it covered something narrower than VA-guaranteed loans, or if the count is wrong, drop or revise the figure.
• VA administrative records by race and state. Agbai (2025) used restricted VA data; her state-level breakdowns, if published, would directly test whether gaps were largest where Black disenfranchisement was greatest. A finding of similar gaps North and South would weaken the local-control argument.
• Turner and Bound robustness. Newer work replicating the North/South split with full-count census data could strengthen or erase the regional difference.
• Readjustment allowances by race and state. Eden finds Black veterans received more unemployment pay nationally. Southern state-level data would show whether state administration restricted it (History.com reports, citing a June 1, 1947 NYT story, accusations that Southern postmasters withheld forms; not independently verified here).
• Women veterans' take-up. Lennon (2021) finds large education and earnings gains for the roughly 330,000 women who served (JEH via RePEc), but we did not find reliable counts of how many used benefits, or how WAAC exclusion affected Black women specifically.
• Rankin's role. We rely on History.com and PolitiFact (quoting historian Stephen Ortiz) for Rankin's push for state administration; the committee record itself was not checked.

Sources

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