The Betrayal They Kept

I. The First Withdrawal: The Freedmen’s Bureau, 1865–1872

The Bureau of Refugees, Freedmen, and Abandoned Lands was established by Congress in March 1865 inside the War Department. It was the federal government’s first sustained effort to administer relief, labor contracts, land redistribution, and education for the four million people emancipated by the end of the Civil War. Within five years it supported more than 1,500 schools serving over 100,000 pupils, with attendance rates around 80 percent at a time when white public school attendance in much of the South was substantially lower. By the end of 1865 alone, 90,000 formerly enslaved people were enrolled. In Georgia, 8,000 Black students were in Bureau-supported schools within a year of emancipation; eight years later the figure was near 20,000. The Bureau also sponsored roughly twenty-five institutions of higher learning for freedpeople — the founding moment for what became the HBCU network.

The Bureau’s total budget for all activities between 1865 and 1870 was approximately $7 million. The schools themselves ran on a blended model: Bureau money paid for some teacher salaries and building repair, but the bulk of operational funding came from Black communities themselves — through direct donations, community taxation, and the Black churches. From the start, the federal commitment was thinner than the mythology remembers, and the gap between federal commitment and actual instruction was filled by the people the federal government claimed to be helping.

The withdrawal came in three steps, each documented and named.

February 19, 1866. President Andrew Johnson vetoed a bill extending and expanding the Bureau. His veto message called it an encroachment on states’ rights and warned that it would render freedpeople “dependent.” Congress overrode the veto on a second version of the bill later in 1866. The override was bipartisan in the sense that Republicans crossed party lines from both directions; the political coalition for sustained federal investment in Black education was already showing fractures.

1869. Southern Democrats returning to Congress, together with Northern Democrats who characterized Bureau work as making African Americans “lazy,” stripped the Bureau of most of its funding. Staff were cut. The agency’s educational division — the most successful and most politically valuable of its functions — was the slowest to be defunded, but the funding cuts foreshadowed what was coming.

Summer 1872. Congress, under pressure from white Southerners and exhausted Reconstruction-era Republicans, refused to reauthorize the Bureau. It was dismantled. Most of its educational functions ceased. The 1,500 schools that had been built in seven years did not all close immediately — many continued under church or community sponsorship — but the federal infrastructure that had created them, paid the teachers, supplied the textbooks, and protected the students from Klan violence was gone.

What replaced the Bureau was not nothing. What replaced it is what the historian James D. Anderson, in The Education of Blacks in the South, 1860–1935, calls the “double taxation” regime. Black Southerners paid the same property taxes as white Southerners but then taxed themselves again — through churches, mutual aid societies, and direct community collections — to build and sustain schools the states refused to fund. W.E.B. Du Bois, in Black Reconstruction in America (1935), documented the fierce commitment of Black communities to schooling and argued that the Freedmen’s Bureau’s short life was not a failure of Black capacity but a failure of federal will. The Bureau worked. The Black communities worked. What stopped working was the Republican coalition’s appetite for the project.

The episodic point here is the one Du Bois made and Anderson documented and Eric Foner’s Reconstruction placed in its full context: the federal government briefly committed itself to universal public education in the South, watched Black communities take that commitment seriously, and then withdrew. The withdrawal was a political choice made by identifiable people. Andrew Johnson signed his veto. The Democratic congressional minority kept growing. The Republican majority stopped caring. None of this happened by accident. None of it requires explanation by structural historical forces. It was a vote. It is in the Congressional Record. The Bureau died in the summer of 1872, and what came next was forty years of double-taxed, parallel-funded, structurally inferior Black schooling that no federal institution would touch until 1954.

II. The Permission Slip: Plessy v. Ferguson, 1896

Plessy v. Ferguson, 163 U.S. 537 (May 18, 1896), was not a school case. It was a railroad case. Homer Plessy, a man of mixed ancestry living in New Orleans, had deliberately boarded a “whites only” car to test Louisiana’s 1890 Separate Car Act. The case was a planned constitutional challenge by the Comité des Citoyens, a New Orleans civil rights organization, with Plessy as the plaintiff who had agreed in advance to be arrested. The Court, 7–1, held that the Fourteenth Amendment did not bar the separation of races in public accommodations as long as the separate facilities were “equal.” Justice Henry Billings Brown wrote for the majority.

Justice John Marshall Harlan — a former Kentucky slaveholder who had come to civil rights through a winding path that included his own brother’s anti-slavery activism — wrote the lone dissent. His sentences have outlived the majority’s, and they should be quoted at length because the gap between what he wrote in 1896 and what he would write in 1899 is the central forensic fact of this section of the article.

The white race deems itself to be the dominant race in this country… But in view of the Constitution, in the eye of the law, there is in this country no superior, dominant, ruling class of citizens. There is no caste here. Our Constitution is color-blind, and neither knows nor tolerates classes among citizens. In respect of civil rights, all citizens are equal before the law. The humblest is the peer of the most powerful. The law regards man as man, and takes no account of his surroundings or of his color when his civil rights as guaranteed by the supreme law of the land are involved.

And:

The thin disguise of “equal” accommodations for passengers in railroad coaches will not mislead anyone, nor atone for the wrong this day done.

The railroad frame mattered less than the doctrine. Once the Court had blessed “separate but equal” as a general constitutional principle, the rest of the Jim Crow architecture — schools, hospitals, cemeteries, drinking fountains, swimming pools, courtrooms, hospitals — slid neatly under the same umbrella. “Equal” in practice was never measured. Funding ratios in the Jim Crow South by the 1910s and 1920s ran as high as 10:1 in favor of white schools in some counties. Anderson documented per-pupil expenditure ratios of roughly 5:1 statewide in Mississippi by 1930, with rural Delta counties far worse. The Court had said “equal” and meant nothing in particular by the word; the state legislatures took the permission and discarded the qualifier.

Niklas Luhmann’s framework of binary codes is unusually useful here. Luhmann argues that the legal system operates on the binary code legal/illegal, and that the code is operationally closed: it cannot evaluate inputs by criteria from outside its own code. Plessy is the case where the legal code accepted “separate but equal” as legal because the code had no operational capacity to ask whether separate had ever been equal in any documented case. The empirical question — were the railroad cars actually equal? — was outside the code’s ability to process. The Court could only ask: was the statute facially racially neutral? The answer, by the doctrine the Court had constructed for itself, was yes. The substantive question was unaskable inside the code. Luhmann would call this structural blindness. Harlan, in 1896, called it “the thin disguise.” The two descriptions agree.

The episodic point: the Court did not write school segregation into the Constitution in 1896. It wrote a permission slip, and every state legislature that wanted one could find it on the desk. Article 1 of this series traced what the architects of public education built. Article 2 traces what the courts permitted other people to build on top of it. Plessy is the moment the permission was issued.

III. The Two-Harlan Moment: Cumming v. Richmond County, 1899

Three years and seven months after Plessy, the Court returned to a related question and made it sharper. Richmond County, Georgia, had operated Ware High School — the only Black public high school in the county — since 1880. In 1897 the Richmond County Board of Education voted to close it and redirect the funds to Black primary schools. The county continued to operate its white high school. Black taxpayers, represented by J. W. Cumming, sued for an injunction not to desegregate the white school but merely to force the county to keep Ware open until it could provide any high school for Black children at all. They were not asking for integration. They were asking the county to spend their property tax money on a school they were already paying for.

The Supreme Court ruled unanimously against the Black plaintiffs in Cumming v. Richmond County Board of Education, 175 U.S. 528 (December 18, 1899). The opinion was written by Justice John Marshall Harlan — the same Harlan who had written the “color-blind Constitution” dissent in Plessy three years and seven months earlier. Harlan held that the plaintiffs had to prove the board’s decision was motivated “exclusively” by racial hostility. They could not carry that burden. The injunction was denied. Ware High School stayed closed. The white high school stayed open. The Constitution that Harlan had described as “color-blind” in 1896 turned out, in 1899, to be color-blind in exactly the way that left the closed Black high school closed.

This is the two-Harlan moment, and it deserves to be sat with. The author of the most-quoted civil-rights dissent in Supreme Court history wrote the unanimous opinion that made Plessy’s “equal” promise unenforceable in practice. The two opinions are not by different men. They are by the same man, three years and seven months apart, and the gap is not a change of mind that anyone has documented. There is no recorded explanation. There is no biographical pivot. There is only the record of what Harlan wrote in May 1896 and what Harlan wrote in December 1899, and the structural fact that the second opinion gutted the principle the first opinion announced.

The historian’s most generous reading is that Harlan distinguished the two cases on doctrine: Plessy was about whether segregation was constitutional, and Harlan said no; Cumming was about whether a federal court could compel a county to fund a particular school, and Harlan said no on federalism grounds. The reading is technically defensible. It is also what the legal system generates when it is operating its code correctly and reaching a monstrous result. The doctrinal distinction was available; the doctrinal distinction was used; the doctrinal distinction made the moral content of the 1896 dissent inoperative against the documented facts of a closed Black high school in 1897. That is what binary codes do. They generate distinctions that allow the system to reach the result it needed to reach. Luhmann predicted exactly this. Harlan demonstrated it.

The practical effect of Cumming: for the next several decades, a Black high school education in the South was something you obtained only if a local school board decided to let you. Many didn’t. As late as 1932, three-quarters of Southern counties with significant Black populations had no public Black high school of any kind. The federal government, through Justice Harlan, had said this was constitutional. The state governments, through their boards of education, had taken the permission. The Black communities, through their churches and mutual aid societies, paid for whatever schooling existed. Du Bois’s Black Reconstruction would later describe this period as the second betrayal — the first was the dismantling of the Bureau in 1872, the second was the constitutional doctrine that authorized the dismantling to be permanent.

The episodic point: Cumming is the case that made Plessy’s “equal” promise unenforceable in practice. Even when Black taxpayers did not demand integration — merely the continuation of a school they were helping to pay for — the Court said no. And the hero of the color-blind Constitution wrote the opinion. This is the first place in the sequence where an institution said aloud that racial sorting in schooling was not only permissible but judicially unreviewable.

IV. The Maps: HOLC Redlining, 1933–1940

The Home Owners’ Loan Corporation was created under the Federal Home Loan Bank Act of 1933 to refinance mortgages in default during the Depression. Between late 1935 and 1940, as part of its City Survey Program, HOLC produced residential “security maps” for 239 American cities. Neighborhoods were graded on a four-color scale: A — “Best” (green); B — “Still Desirable” (blue); C — “Declining” (yellow); D — “Hazardous” (red). The red color is the origin of the verb to redline.

The grades were assigned by HOLC staff using data supplied by local real estate professionals — appraisers, lenders, developers. The criteria explicitly encoded race. The University of Richmond’s Mapping Inequality project, which digitized the HOLC archive, has made the area descriptions publicly searchable. The “detrimental influences” column on the appraisers’ standardized form cites “infiltration of Negroes,” “subversive racial elements,” “Jewish,” “foreign-born,” and similar phrases across hundreds of majority-Black, majority-Jewish, and majority-immigrant neighborhoods in the major American cities of the late 1930s. The phrases are not paraphrase. They are the original handwriting of the appraisers, in the original boxes on the original forms, scanned from the National Archives, posted online, searchable by city, indexed by census tract.

HOLC’s own loan activity used these maps to some extent, and scholarly debate continues about the precise transmission belt between HOLC’s direct lending and subsequent private-sector practice. The decisive channel, however, was the Federal Housing Administration, founded in 1934, and later the Veterans Administration. Richard Rothstein, in The Color of Law (2017), documents how the FHA Underwriting Manual — distributed to every appraiser evaluating mortgage insurance applications — explicitly warned against “inharmonious racial groups” and refused to insure mortgages in or near Black neighborhoods. The Manual also required that federally insured subdivisions include racially restrictive covenants in their deeds, until the Supreme Court struck down covenant enforcement in Shelley v. Kraemer in 1948. The covenants did not stop appearing in new deeds in 1948. They simply stopped being enforceable in court. The neighborhoods that had been built with them were already built.

The connection to schools is direct and mechanical. Public school district boundaries in twentieth-century America were drawn to match residential patterns. Residential patterns were drawn to match the HOLC and FHA maps. Property-tax school finance then converted the residential pattern into a per-pupil funding pattern. By the time Brown was decided in 1954, the pipeline from federal mortgage underwriting to unequal school funding had been running for twenty years and had already placed Black families in the neighborhoods that would, inside two decades, become the “urban core” the Court would tell us was constitutionally untouchable.

Lawrence Lessig’s framework of architecture as the binding constraint on power is the right diagnostic tool here, and not metaphorically. Architecture, in Lessig’s usage, means the structural features of a system that determine what is and is not possible inside it. The HOLC map is architecture in the most literal sense: a paper artifact that determined which buildings would be built where, which mortgages would be insured for whom, which streets would have sewers, which neighborhoods would have parks, and — downstream of all of these — which children would attend which school. The architecture was federal. The architecture was paper. The architecture is still operative in 2026. You can pull up the 1937 map of any American city of medium size or larger and overlay it on the 2024 per-pupil spending data, and the correspondence will be uncanny enough to feel like cheating. It is not cheating. It is cartography. The maps that the Department of the Interior commissioned in 1935 are the maps the school finance system is still running on.

The episodic point: the federal government did not merely fail to stop residential segregation. It wrote the instruction manual, printed the color-coded maps, and handed them to the lenders. The redlined neighborhoods of 1940 and the low-funded school districts of 2024 are the same places. That is not metaphor. It is documented cartography, available to any citizen with an internet connection, refreshed and georeferenced for the modern web. The receipts are public.

V. The Plumbing: The GI Bill and the Mississippi Two

The Servicemen’s Readjustment Act of 1944 — the GI Bill — is the single most important policy in the creation of the twentieth-century American middle class. It subsidized college tuition for 2.2 million veterans by 1956, guaranteed home loans for 4.3 million more, and provided unemployment benefits to millions. Suzanne Mettler’s Soldiers to Citizens (2005) argues, correctly, that the Bill’s formally universal benefits created a new political consciousness among working-class Americans and enhanced postwar democratic participation. Article 3 of this series will trace the macroeconomic conditions under which the Bill produced the brief window of partial meritocratic plausibility. The job of this section is to document, with the sources, what the Bill did when it ran through the racially structured plumbing the federal government had spent the previous decade installing.

The mechanism, as Ira Katznelson documents in When Affirmative Action Was White (2005), was not a racial clause in the statute. The Bill contained no such clause. The mechanism was the administrative architecture: Southern Democrats, as the price of supporting Roosevelt and then Truman, had insisted that New Deal and Fair Deal programs — Social Security, the Fair Labor Standards Act, and now the GI Bill — be administered locally. Veterans Administration loan officers, FHA appraisers, and local employment offices applied the statute through existing discriminatory practice. The federal program was race-neutral on its face. The plumbing through which the federal program flowed was racially structured by twenty years of HOLC, FHA, and local custom. The output was the predictable function of the input and the plumbing.

The statistical residue is unambiguous, and unambiguous is the word historians of this period use because the numbers are not approximations. They are counted records.

Two of 3,229. The number is small enough that it deserves to be read out loud. Two Black veterans, in a state where Black veterans had served in the same war as white veterans, with the same benefits on paper, in a year when the federal government issued 3,229 mortgages, business loans, and farm loans through the same statute. The number is not an estimate. It is a count. Katznelson found it in the Veterans Administration’s own files. He published the page number. Anyone can look it up.

The episodic point is the one Katznelson made and Rothstein corroborated and Mettler conceded: the GI Bill is the clearest case in the entire sequence of a formally race-neutral statute producing a racially catastrophic distribution of benefits — not because of individual prejudice in isolated cases but because the statute was designed to pass through discriminatory plumbing, and everyone who wrote it knew the plumbing existed. The post-war American middle class was not an organic product of abundance. It was a subsidized creation, and the subsidy was race-gated by the administrative architecture of a 1944 statute that ran through the 1934 FHA Manual that ran through the 1937 HOLC map that ran through the 1899 Cumming opinion that ran through the 1872 defunding of the Freedmen’s Bureau. Each step is documented. Each step has a name on it. Each step was a choice.

Mark Blyth’s argument about ideas as institutional weapons is the diagnostic frame for this section. The phrase that did the political work in 1944 was local administration. The phrase sounded technical and procedural. It described a real and defensible value: the idea that federal programs should be administered close to the citizens they serve. It was also, in the specific context of Southern Democratic insistence on it as a non-negotiable price for any New Deal legislation, a code for racial sorting. Everyone who wrote the GI Bill knew this. The Southern senators said so. The phrase was the weapon. Blyth’s contribution to political economy is the observation that ideas are not neutral descriptions of pre-existing realities; they are political tools deployed at specific moments to do specific work and then retroactively naturalized as American traditions. Local administration was deployed in 1944 to lock in racial sorting and then retroactively naturalized as a federalism principle. The naturalization is what allows us, in 2026, to talk about “local control” of schools as if the phrase had no documented genealogy. It does. It runs through Mississippi. The number is two.

VI. The Decade That Didn’t Happen: Brown and Massive Resistance, 1954–1964

Brown v. Board of Education, 347 U.S. 483 (May 17, 1954), was unanimous. Chief Justice Earl Warren held that “separate educational facilities are inherently unequal” and that segregated public schooling violated the Equal Protection Clause of the Fourteenth Amendment. The decision did not explicitly overrule Plessy — it distinguished it — but it functionally ended the doctrine of separate-but-equal in public education. The mythology that grew up around Brown in subsequent decades treats it as the moment the schools were desegregated. They were not. Brown declared the principle. The remedy was held over for reargument.

On May 31, 1955, the Court issued Brown II, 349 U.S. 294, which remanded enforcement to the federal district courts with the famously elastic instruction to proceed “with all deliberate speed.” Chief Justice Warren had personally drafted the phrase. According to his biographers, he chose it precisely for its lack of definition. He believed the South needed latitude to avoid violent resistance. The South took the latitude. It did not use it to integrate. It used it to organize.

Jürgen Habermas’s diagnosis of the colonization of communicative action is the right framework for what “all deliberate speed” actually did. Habermas argues that a healthy public sphere processes contested questions through deliberation — through argument among citizens about norms, values, and practical decisions. The colonization happens when system imperatives (administrative procedure, market logic, bureaucratic timelines) overrun the deliberative space and substitute procedural questions for normative ones. “All deliberate speed” is the textbook case. The normative question — how fast should we end the constitutional violation we just found? — was reframed as a procedural question about the timing tolerances of district courts, and the procedural question was systematically deferred until the political will to integrate had eroded. The deferral was not a deliberation. It was the absence of a deliberation, dressed in legal language. The South understood this immediately. So did the NAACP. Only the white moderates who quoted “all deliberate speed” as if it were a neutral phrase failed to notice.

The Southern Manifesto, March 12, 1956

Eighty-two members of the United States House of Representatives and nineteen United States Senators — nearly the entire Southern congressional delegation — signed a document titled “Declaration of Constitutional Principles.” History has called it the Southern Manifesto. It called Brown “a clear abuse of judicial power” and pledged the signatories to use “all lawful means” to resist desegregation. Every signatory was a Democrat except two Virginia Republicans. The Manifesto was not a fringe document. It was the official position of the political party that had dominated the South since Reconstruction, signed by the elected representatives of the states in question, distributed in the Congressional Record, reprinted in the major newspapers. It said the quiet part in all capitals.

Prince Edward County, 1959–1964

When federal courts ordered Prince Edward County, Virginia, to desegregate its public schools, the county Board of Supervisors refused to appropriate any funds for the 1959–60 school year. The county closed all of its public schools for five years. A private “Prince Edward Academy” opened immediately for white children, funded by tuition grants the state and county provided. Black children — approximately 1,700 of them — either left the county to live with relatives, enrolled in makeshift “training centers” run by churches, or received no schooling at all. A generation lost its education as a matter of official policy. The Supreme Court finally ended the closure in Griffin v. County School Board of Prince Edward County, 377 U.S. 218 (1964). The five years are not recoverable. The cohort of Black children who would have been in those five years of school is in their seventies and eighties now. The lost education is in their resumes, their wages, their retirement accounts, and their grandchildren’s expectations.

The segregation academies

By 1969, more than 200 private “segregation academies” had opened across the eleven Southern states, enrolling roughly 300,000 white students who had fled desegregation orders. Virginia pioneered the model in the late 1950s. Mississippi took it furthest: between the 1969–70 and 1970–71 school years alone, approximately 41,000 Mississippi white students left the public schools for segregation academies. The arithmetic of this is worth dwelling on. Forty-one thousand white children, in a single state, in a single school year, were withdrawn from public education by their parents because the public schools had begun to integrate. Many of these academies were initially subsidized through state tuition grants in the 1960s. When federal courts blocked direct state aid, the academies survived on private tuition and, later, on state-level tax-credit scholarship programs that continue today. ProPublica’s 2024 reporting documents that many of the same academies still operate and, in several Southern states, still receive indirect public funds through voucher and tax-credit programs designed to be facially race-neutral.

The episodic point: Brown was not self-executing, and the decade between 1954 and 1964 is not a period of slow social adjustment. It is a period in which every Southern state government actively chose resistance, the federal government largely declined to override them, and a private school infrastructure was built specifically to receive white flight and outlast any desegregation remedy the courts might eventually impose. The infrastructure outlasted the remedy. The remedy is now mostly gone. The infrastructure is still here.

VII. The Brief Moment It Worked: Swann, 1971

Swann v. Charlotte-Mecklenburg Board of Education, 402 U.S. 1 (April 20, 1971), was the last unanimous school-desegregation decision the Supreme Court ever issued. Chief Justice Warren Burger — a Nixon appointee, no friend of busing, the man who had been picked by a president running on a Southern Strategy — wrote the opinion. He circulated it through six drafts. Each draft was stronger than the last, pushed by Justices William O. Douglas, John Marshall Harlan II (the grandson of the Plessy dissenter), William Brennan, and Thurgood Marshall. In the end, the Court held that federal district courts could order busing, redistricting, and racial quotas as remedial tools to dismantle the dual school systems left by de jure segregation. The vote was 9–0.

Charlotte-Mecklenburg, which had already consolidated its city and county school districts in 1960 — a decision that turned out to be decisive for everything that came next — implemented the resulting busing plan over the 1971–72 school year. For roughly the next two decades, Charlotte-Mecklenburg operated one of the most successfully integrated large school systems in the country. Test score gaps narrowed. White flight was limited because there was nowhere within the consolidated district to flee to. The district became a case study in what Brown could look like when actually enforced, and the case study was widely cited at the time as evidence that the Burger Court was, despite its conservative reputation, willing to follow the logic of Brown to its remedial conclusion.

This is the interval that matters for the rest of the article. Swann shows the system was not incapable of delivering on Brown. The tools existed. The unanimous Court existed. The local political will, in a handful of places, existed. Burger himself existed, writing for nine, on April 20, 1971. The next case — written by the same Burger, three years and three months later — would take it all apart. What changes between Swann and Milliken is not the law. It is not the facts. It is not the Constitution. What changes is whether the Court is willing to extend the remedy across the district line that the previous fifty years of federal residential policy had drawn. Swann was a city-and-county consolidated district. Milliken would be Detroit and its fifty-three white-flight suburbs. The difference between the two cases is the line.

VIII. The Two-Burger Moment: Milliken v. Bradley, 1974

Milliken v. Bradley, 418 U.S. 717 (July 25, 1974), is the case that made the zip-code-property-tax system permanent. It is the load-bearing decision for this article and for the entire arc of post-Brown American school finance. Everything before it pointed somewhere. Everything after it has had to live with where it pointed.

The facts. In 1970, the NAACP, on behalf of Ronald Bradley and other Black Detroit students, sued the Detroit Board of Education. The federal district court — Judge Stephen Roth presiding — found, after extensive evidence, that Detroit’s schools had been deliberately segregated through decades of school-board decisions, attendance-zone gerrymandering, and direct interaction with HOLC and FHA residential policy. Judge Roth’s finding of de jure segregation by the Detroit Board was not contested on appeal. That was the easy part. The hard question was the remedy. By 1970, the city of Detroit was approximately 64 percent Black. The suburbs surrounding Detroit — fifty-three separate school districts — were overwhelmingly white. A desegregation plan limited to Detroit alone would achieve paper integration at best and would accelerate the remaining white flight to the suburbs, producing in five years a school system that was 90 percent Black and surrounded by school systems that were 95 percent white. Judge Roth ordered an interdistrict remedy: a metropolitan busing plan covering Detroit and the fifty-three surrounding districts.

The decision. The Supreme Court, 5–4, reversed. Chief Justice Warren Burger — the same Burger who had written Swann three years and three months earlier — wrote the opinion. The holding: absent proof that the suburban districts had themselves engaged in de jure segregation, or that the state of Michigan had manipulated district boundaries for a racial purpose, an interdistrict remedy was “wholly impermissible.” Burger invoked “local control” as “the single most important tradition of public education.” District lines, once drawn, were presumptively sovereign. The fact that the district lines had been drawn during and after the period when HOLC and FHA were producing the residential pattern those lines tracked — the fact that Judge Roth had documented in his district court findings — was not material to the constitutional question Burger was deciding. The question was whether the suburban districts had themselves committed a cognizable violation. They had not. The remedy could not reach them.

Marshall’s dissent

Justice Thurgood Marshall — who as chief NAACP counsel had argued Brown itself two decades earlier — wrote a dissent whose most-quoted passage has become the epitaph for the entire post-Brown enforcement project. It deserves to be quoted in full because the structural argument is in the syntax:

School district lines, however innocently drawn, will surely be perceived as fences to separate the races when, under a Detroit-only decree, white parents withdraw their children from the Detroit city schools and move to the suburbs in order to continue them in all-white schools. The very evil that Brown I was aimed at will not be cured, but will be perpetuated.

And, in the closing movement of the dissent:

In the short run, it may seem to be the easier course to allow our great metropolitan areas to be divided up each into two cities — one white, the other black — but it is a course, I predict, our people will ultimately regret. I dissent.

Marshall’s dissent made the structural point directly: the residential segregation that produced Detroit’s school demographics was itself the product of federal policy — HOLC, FHA, GI Bill — and of state-level complicity. Treating the district lines as innocent was a legal fiction that required the Court to un-know the previous half century of documentary evidence. Marshall named it as such. He had the documents. Judge Roth had them in the district court record. The majority did not contest the documents. The majority simply held that the documents were not a constitutional violation by the entities the remedy needed to bind. The legal code reached its result. The structural reality reached its conclusion. The two had nothing to do with each other.

This is Luhmann at his most brutal. The legal system’s binary code — de jure / de facto, intentional / incidental, cognizable violation / mere effect — cannot represent the upstream causality the federal government itself had documented for decades. Burger’s opinion treats the district lines as “innocently drawn.” Marshall’s dissent insists they are not. The legal code requires the fiction of innocence in order to function as a binary code at all. The fiction is not a lie. It is a structural feature of how the legal system processes information that doesn’t fit its categories. The information is real. The system cannot see it. The system reaches the result the system can reach. The result is monstrous. The system is doing what it was built to do. Luhmann predicted exactly this. Marshall, in dissent, demonstrated it.

What Milliken did structurally

The decision locked in the arithmetic of American school finance. Because (1) school district boundaries were treated as presumptively innocent, and (2) property tax revenue stays inside district boundaries, and (3) residential patterns — produced by HOLC, FHA, GI Bill, and subsequent private discrimination — dictate the property-tax base, the result was that unequal school funding, mapped precisely onto the residential patterns that federal policy had built, became constitutionally unreachable through federal desegregation remedies. The pipeline from 1935 HOLC maps to 2024 per-pupil spending disparities passed through Milliken without touching the sides. This is the architecture Article 1 introduced as Lessig’s binding constraint, in its most consequential historical form. The architecture won. The remedy lost. The Constitution was not amended. A vote was taken. It was 5–4. The same Chief Justice who had written the previous unanimous decision wrote the new one in the opposite direction.

Mark Blyth’s diagnosis of “local control” as an institutional weapon is the right frame for what Burger’s majority opinion did with the phrase. “Local control” is a real value. It has a real history in American governance, traceable to colonial-era town meetings and the genuine democratic intuition that people who live in a community should have voice over the schools their children attend. It is also the phrase Burger reached for in 1974 to lock in the residential sorting that fifty years of federal policy had produced. The phrase did political work for the institutional interest of suburban whiteness while sounding like a neutral principle of federalism. Blyth’s point is that ideas function as institutional weapons at specific moments and are then retroactively naturalized as traditions. “Local control” was a weapon in 1974. It is now a tradition. The naturalization is what makes the doctrine impossible to dislodge: any reform of school finance has to either argue against local control or work around it, and the rhetorical cost of arguing against an American tradition is the cost the doctrine was designed to impose.

The episodic point: this is the case in which the Court chose to stop. Every prior case in the sequence — Plessy, Cumming, Brown, Swann — left the door partially open. Milliken closed it. And it closed it with the same Chief Justice who had written Swann, which is a fact this article should sit with one more time before moving on. Same man. Three years and three months apart. Unanimous in 1971; 5–4 in 1974. The 1971 opinion authorized the remedy. The 1974 opinion confined the remedy to the smallest possible jurisdiction. Nothing in the law had changed. The composition of the Court had shifted slightly with Nixon’s appointments. The political environment had hardened around busing in white suburban communities. Burger followed the political environment. The Court reached the result the political environment required. Marshall named it. The dissent is the receipt. The majority is the architecture.

IX. The Continuous Arc: Rodriguez, the State Cases, and the Present

A year before Milliken, in San Antonio Independent School District v. Rodriguez, 411 U.S. 1 (March 21, 1973), the Supreme Court had already foreclosed the other available remedy. Plaintiffs in Edgewood Independent School District — 98 percent Mexican-American, in San Antonio — had challenged Texas’s property-tax school-finance system under the Equal Protection Clause. Per-pupil spending in Edgewood was roughly two-thirds of per-pupil spending in the wealthier Alamo Heights district next door, despite Edgewood residents paying property taxes at higher rates. The district court found the system unconstitutional.

The Supreme Court, 5–4, reversed. Justice Lewis Powell, writing for the majority, held that education is not a fundamental right under the U.S. Constitution — “neither explicitly nor implicitly” guaranteed — that wealth is not a “suspect class” triggering strict scrutiny, and that the property-tax system therefore needed only to survive rational basis review, which it did, because local control of schools was a legitimate state interest. Marshall, again in dissent, called the decision “a retreat from our historic commitment to equality of educational opportunity.” The 5–4 split was the same as Milliken’s would be the following year: the Nixon bloc — Burger, Blackmun, Powell, Rehnquist, plus Stewart in Rodriguez — against the remnants of the Warren Court.

After Rodriguez, the only remaining path for challenging property-tax school finance ran through state constitutions. California’s Serrano v. Priest (1971, 1976, 1977), decided under the California Constitution, was an early victory, but the passage of Proposition 13 in 1978 capped property taxes and shifted school funding to the state in a way that produced new inequities through “basic aid” districts. Texas litigants returned in Edgewood ISD v. Kirby (1989), prevailing under the Texas Constitution. New Jersey’s Abbott v. Burke litigation, beginning in 1981, produced more than two dozen rulings and remains the most extensive state-level school-finance jurisprudence in the country. The state-by-state patchwork has narrowed some gaps and created others. None of the state-level victories has touched the federal architecture Milliken built.

The current arithmetic, from EdBuild’s 2019 analysis still cited in 2024 and 2025 reporting:

The chronology continues. The Reagan administration, beginning in 1981, actively retreated from federal desegregation enforcement — the Justice Department under Assistant Attorney General William Bradford Reynolds filed briefs asking courts to reduce existing busing plans. The Supreme Court’s Parents Involved in Community Schools v. Seattle School District No. 1, 551 U.S. 701 (2007), held that school districts could not use race as an explicit factor in student assignment even to maintain voluntary integration, with Chief Justice John Roberts writing the now-famous line: “The way to stop discrimination on the basis of race is to stop discriminating on the basis of race.” The trajectory from Milliken to Parents Involved is a single arc: the federal judiciary steadily narrowing the range of constitutional tools available to address a segregated system the federal government itself helped build. The arc has not bent back. It has continued in the same direction for fifty-two years.

X. The Strongest Defense

The strongest case against the framing of this article deserves to be made in its strongest form. There are four versions of the counterargument worth engaging.

First. Local control is a real value, not a pretext. The Michigan Law Review’s revisit essay on Milliken acknowledges the critique while noting that local control of schools is not an invented doctrine. It predates Brown, it predates Plessy, it is genuinely traceable to colonial-era town governance, and it reflects a real democratic value: the belief that parents and communities should have meaningful voice in their children’s education. A court that ordered the redrawing of fifty-three school district boundaries by federal decree would have been, in Burger’s phrase, a “de facto legislative authority” and “school superintendent” for metropolitan Detroit. Whether you think that would have been good policy, it would have been a dramatic expansion of federal judicial power, and reasonable people can believe that the line had to be drawn somewhere. The fact that Marshall’s dissent reads as more compelling in 2026 does not mean Burger’s majority was acting in bad faith.

Second. The GI Bill genuinely expanded opportunity, and its universalism was not cynical. Suzanne Mettler has spent her career arguing that the GI Bill should not be dismissed as a white-only entitlement. Her interview research with Black veterans found that many experienced the Bill as an exceptionally positive encounter with the federal government — the first such encounter for many of them — and that its benefits did reach significant numbers of Black veterans, particularly in the North. The Bill is partly a story of expanded citizenship for groups previously excluded — Jewish, Catholic, Italian, Irish — who had been kept out of earlier middle-class formation. On this reading, the GI Bill is partly a story of expanded citizenship and partly a story of how Southern Democrats insisted that any New Deal legislation route through their segregationist administrative apparatus. The second story is the indictment, but the first story is also true.

Third. Property-tax financing enables genuine community investment. The strongest affirmative case for property-tax school finance is that it produces accountability and responsiveness. Communities that value schooling can tax themselves to fund it. Communities that want better schools can reliably get better schools by passing bond referendums. The alternative — full state or federal funding — uncouples school quality from community commitment and, in the experience of several states that have moved toward equalized funding, has produced mediocre convergence rather than leveling-up. California after Serrano and Proposition 13 is the cautionary example: equalized funding plus a statewide tax cap produced a funding system that is more uniform but, by most measures, worse than what preceded it. A historian defending the property-tax system could point out that the worst Jim Crow funding disparities were in states that did not fund schools through local property taxes — they funded them through state-controlled systems that state legislatures used to actively discriminate. Local control, on this reading, was a check on worse possibilities.

Fourth. Post-Brown enforcement was hard because social change is hard. The argument is that the decade between Brown and meaningful desegregation reflects the ordinary difficulty of implementing major social transformation rather than a deliberate institutional failure. Warren’s “all deliberate speed” was a mistake in retrospect, but at the time it was a plausible attempt to manage the real risk of violent white backlash — which did materialize, in places like Little Rock and across the Deep South, and did include the killings of civil rights workers. A Court that had ordered immediate desegregation in 1955 would have faced the question of what to do when Southern states simply refused, and the available enforcement tools were genuinely limited until the Civil Rights Act of 1964 and the Elementary and Secondary Education Act of 1965 created new federal leverage. On this reading, the Warren Court was not complicit in massive resistance; it was constrained by the institutional realities of mid-century federalism.

Each of these counterarguments contains a truth, and none of them rescues the overall indictment. Local control is a real value, but it cannot do the moral work Milliken assigned it once you know how the district lines were drawn. The GI Bill did expand opportunity, but the exclusions were structural, foreseeable, and documented in the Veterans Administration’s own records. Property-tax financing does enable community investment, but it also enables the capture of community investment by the communities that can afford to invest. Post-Brown enforcement was hard, but the Warren Court chose “deliberate speed” before it knew how hard it would be, and then the Burger Court stopped trying. The honest position is the conjunction. Each step in the sequence had a defensible-sounding justification at the time. Each step also produced the documented outcome the indictment names. The defensible justifications and the documented outcomes are not in tension. Both are true. Both can be held in the same sentence. The architecture is the result of the conjunction, not the result of either side of it taken in isolation.

XI. The Place This Article Was Written From

One paragraph of honest reckoning is owed before the sources. This series is published from a state — Texas — whose 1954 response to Brown included its own segregation academies and its own use of the zip-code-property-tax system to lock racial sorting into the school finance architecture for the remainder of the twentieth century and into the twenty-first. The author of this series writes from inside that history, not above it. Texas in 2026 is the state where Edgewood ISD v. Kirby won at the state supreme court level in 1989 and where the structural inequity it found still has not been fully corrected. The Edgewood district that brought the original Rodriguez case in 1973 still spends substantially less per pupil than Alamo Heights. The article is not written from a perch of moral clarity. It is written from inside the architecture it is describing. The series is the work of a Texan documenting a national betrayal that his own state participated in fully.

Michael Sandel’s framework of meritocratic moral corrosion is the right scholar for the closing of this article, and Sandel will be the central voice of Article 5. The relevant move here is the one Sandel calls the rhetoric of rising. The mythology that “I worked hard for my house in the good school district” is, in many cases, true. The dignity of the work is real. What it leaves out is the prior, unworked-for advantage of inherited residential sorting. The good school district was not built by the work of its current residents. It was built by HOLC and FHA and the GI Bill, which produced the residential pattern, and by Milliken, which made the residential pattern constitutionally unreachable. The current residents inherited the architecture and then added their work on top of it. Treating the inheritance as if it were absorbed into the work is the move Sandel diagnoses as meritocratic hubris. It is morally corrosive at scale because it tells the families on the other side of the district line that they are on the other side because they did not work as hard, and it tells the families inside the line that they earned what was given. The dignity of the work is real and the inheritance underneath the work is real. Both have to be held. The series will return to this in Article 5.