Chapter 2 — Statehood Without the Federal Floor

Vermont to Texas, 1791–1845

“This isn’t just government land; it’s trust land. And just like any other trust, the trustee — which is the state — has a fiduciary duty to act in the best interest of the beneficiary.” — Margaret Bird, August 2025 video transcript

What this chapter does. I want to take the reader through six states that came into the Union after the federal land system was already running — and that nevertheless got no federal section-sixteen grant when they joined. The cohort is a natural experiment the country ran without meaning to. By the time you finish this chapter you should be able to say, with the evidence in hand, that a federal land grant is neither necessary nor sufficient to build a school trust — and to name what does the work in its place.

A trust that should not, on the standard story, exist

I want to start with a number, because the number is the cleanest place to start. The largest educational endowment in the United States — larger than Harvard’s, larger than Yale’s, roughly twice the size of New Mexico’s school land trust — sits in a state that never received a single acre from the federal government for the support of its schools. The Texas Permanent School Fund holds investment assets of approximately $56.9 billion as of FY2024; recent valuations including mineral and surface acreage place the corpus in the $67 billion range, with a record $4.8 billion biennial distribution authorized for 2026–27.¹ None of it traces to a federal admission-act grant. There was never a federal Section 16 in Texas. The United States acquired no public domain inside Texas’s borders at statehood, and so it had nothing to grant.

I belabor that point because it does most of the analytic work of this chapter. The story Chapter 1 told — of original states that had no federal grant because the federal government did not yet have land to give — was true but partial. There is a second cohort, smaller and stranger, that came into the Union after the federal land system was already up and running but that left the federal floor underneath them anyway. They got there by six different doors. None of those doors opened onto a Section 16. One of the six built the largest school endowment in the country. One built nothing of consequence. The other four sit in between, and their cases test, as cleanly as American history allows, the proposition that a federal land grant is neither necessary nor sufficient to build a school trust.

I want to be precise about what this chapter shows. It does not show that federal grants are unimportant — Chapter 3 is going to show that, where they were given, they did real work in real states. It shows the narrower and more interesting thing: that the architecture of a school trust — corpus, beneficiary, fiduciary, perpetuity, segregation from general revenue — is what carries the weight, and that architecture can be built either federally or out of the state’s own materials. The federal grant is one path to the architecture. It is not the only path, and where it has been given, it has not always been honored. Where it has not been given, it has sometimes still been built. The federal grant was an opportunity. What each state did with the opportunity, or with its absence, is the chapter.

Six doors, no federal floor

Let me lay the six down quickly so the reader can see the shape of the cohort.

Vermont came in first, in 1791, having spent the previous fourteen years as an independent republic.² Kentucky followed in 1792, carved out of Virginia’s western district under Virginia’s own act of cession.³ Tennessee arrived in 1796, on land North Carolina had ceded to the federal government — but seven years before Ohio, the state whose admission act fixed the federal school-grant template.⁴ Maine separated from Massachusetts in 1820 as the free-state half of the Missouri Compromise.⁵ Texas was annexed by joint resolution of Congress in 1845 — a former independent republic, like Vermont, but on a continental scale.⁶ West Virginia was wrenched out of Virginia in 1863, by the consent of the Wheeling Unionist government and a Lincoln proclamation.⁷

What unites them is what is missing. None of these admissions came with a school-land grant. The mechanisms of the Land Ordinance of 1785 and the Northwest Ordinance of 1787 — the survey grid, the reservation of Section 16 in every township — did not reach them. Vermont, Kentucky, Maine, and West Virginia carved themselves out of land that had never been federal. Tennessee was admitted before the federal template existed in mature form, and what federal land lay inside its borders had been so encumbered by prior North Carolina warrants that almost nothing was left to grant. Texas kept its own public domain as the price of joining the Union: the Joint Resolution of March 1, 1845 expressly preserved to the new state “all the vacant and unappropriated lands lying within its limits.”⁸

Each of the six had to figure out, on its own, what to do about schools. The answers vary considerably — and the answer that breaks every assumption is the one this chapter ends with.

Vermont, 1791: the school-clause that came before the federal template

I want to slow down on Vermont, because Vermont matters here for a reason that is easy to miss if you are reading the chapter for its scale. The reason is the date on Vermont’s constitution.

The Vermont Constitution of 1777 — drafted at Windsor while Vermont was still an independent republic, with no expectation of admission to the Union it would not join for another fourteen years — directed that “a school or schools shall be established in each town, by the legislature, for the convenient instruction of youth.”⁹ That language was on the books eight years before the federal Land Ordinance of 1785 reserved its first Section 16, and ten years before the Northwest Ordinance enshrined the school clause as federal policy. The text traces forward through the 1786 Constitution’s § XXXVIII to the current Chapter II § 68.¹⁰

The mechanics ran ahead of the federal template too. From 1749 onward, New Hampshire’s colonial governor Benning Wentworth had been issuing town charters in what would become Vermont. The standard Wentworth charter divided each township into roughly seventy shares of about 330 acres each and reserved several for public uses, including a “school right” to support a school in the town.¹¹ When Vermont declared itself a republic in 1777, it kept the school-right reservation in its own town charters. By the time Vermont entered the Union in 1791, the school-rights pattern was in place across roughly 250 towns.

I belabor this point because the chronology is the argument. Vermont is the proof case that the school-clause civic instinct did not require federal sponsorship. It was already in the air. The federal templates of 1785 and 1787 codified an idea Americans had already arrived at on their own. The framers of the Land Ordinance were not innovating on a blank slate; they were ratifying a pattern New England had been working out for more than a generation. Vermont implemented that pattern through three different sovereigns in succession — colonial governor, independent republic, fourteenth state — without ever needing Congress to bless it. The 1791 admission act admitting Vermont contained no school-land provisions, because there was no federal land in Vermont to grant.¹²

What Vermont did not build is also worth naming. There is no centralized fiduciary architecture. Each town administers its own school-rights under state enabling statutes — currently Title 24, Chapter 65, of the Vermont Statutes Annotated — and rents flow to the towns rather than to a state corpus.¹³ The Vermont Supreme Court has construed the lease-land substrate town by town in Mikell v. Town of Williston (1971), Broughton v. Town of Charlotte (1976), and Galkin v. Town of Chester (1998), the last confirming that towns hold fee title to school lots and lessees do not own minerals absent town authorization.¹⁴ These cases enforce the substrate. They do not consolidate it. There is no Vermont equivalent of Lassen v. Arizona Highway Department, the federal-trust-enforcement case that disciplines public-land state trusts, because there is no federal compact for Vermont to enforce.¹⁵

I want to slow down on what that decentralization buys and what it costs, because the trade-off recurs in different forms across the chapter. The doctrinal upside of the Vermont model is its resilience against single-point failure: there is no statewide trust to capture. The downside is the absence of unified accounting or enforcement. Act 152 of 2018, codified at 24 V.S.A. § 2409, illustrates the cost: lease lands not affirmatively retained by their towns before January 1, 2020 passed in fee simple to the lessees of record on that date, and a January 2026 Vermont Superior Court decision applied § 2409 to vest fee title to a former glebe in a lessee where the town had not voted to retain.¹⁶ Whether Act 152 amounts to a structural seizure of school-origin lease lands is, as of this writing, an empirical question that depends on town-by-town records — which is itself the point. A decentralized architecture can be quietly drained without anyone in a single statewide office noticing in time to act.

Vermont’s modern K-12 funding architecture sits separately atop a state-equalization regime: the Education Fund created after Brigham v. State (1997), which held under Chapter II § 68 and the Common Benefits Clause that local-property-tax mechanics produced constitutionally intolerable inequality.¹⁷ Acts 60 (1997) and 68 (2003) built a uniform statewide property tax with a per-pupil distribution formula, with annual flows now in the two-billion-dollar range. The Education Fund is current-revenue equalization, not a trust corpus. The school-rights, where they survive, sit alongside it as a quieter inheritance from 1777 — older than the United States in any modern sense, and at this point doing less of the funding work than the post-Brigham equalization regime that wraps around them.

Kentucky, 1792: federal cash, not federal land

Kentucky’s case rewards careful reading, because the surface story misleads. Kentucky did establish a Common School Fund. It did receive federal money to capitalize that fund. Those two facts can produce the impression that Kentucky belongs in the federal-grant category. It does not, and the distinction matters.

Kentucky was admitted to the Union as the fifteenth state on June 1, 1792, by an Act of Congress consenting to Kentucky’s separation from Virginia.¹⁸ The admission act ran only a few paragraphs and contained no school-land grant. The lands inside Kentucky’s borders had been claimed, granted, or reserved under Virginia’s prior land system; Virginia’s 1789 Compact with Kentucky preserved those Virginia-era arrangements as a condition of separation.¹⁹ The federal government had nothing to grant inside Kentucky because the federal government had never owned anything inside Kentucky.

The state-derived alternative began to take shape in the 1830s. The Surplus Revenue Distribution Act of June 23, 1836 — Andrew Jackson’s mechanism for distributing the federal Treasury surplus to the states by deposit — gave Kentucky a substantial cash deposit in 1837, and the General Assembly set apart a portion for educational purposes.²⁰ On February 23, 1838, Kentucky created the State Board of Education, the office of Superintendent of Public Instruction, and the Common School Fund. Contemporary commentary reports that $1,000,000 was initially set aside for the fund, later reduced to $850,000 as fiscal pressure mounted, with bond-interest defaults accumulating through the 1840s.²¹

I want to stop here because this is where the careful reading earns its keep. Kentucky’s seed money was federal cash, not federal land. As someone whose career has been spent inside the technical legal architecture of school trusts, I want to be precise about why that distinction is more than semantic. The Surplus Revenue Distribution did not impose the kind of compact-trust obligation that Cooper v. Roberts (1855) and Lassen would later derive from enabling-act school-land grants. The federal cash, once deposited with the state, became state money. Whatever trust obligations Kentucky assumed by setting it apart for schools were state-law obligations, not federal compact obligations. Kentucky has a school fund seeded by federal money but governed entirely by state constitutional and statutory law — a hybrid with no exact parallel in the federal-grant cohort.

The 1891 Constitution supplied the load-bearing text. Section 183 imposes the affirmative duty: “The General Assembly shall, by appropriate legislation, provide for an efficient system of common schools throughout the State.”²² Section 184 establishes and protects the Fund, declaring the corpus and any sum produced by taxation for common-school education “shall be appropriated to the common schools, and to no other purpose.” Section 184 closes with a voter-approval clause: no sum shall be raised or collected for education other than in common schools without majority voter approval.²³ Section 184 has not been amended since 1891.

That § 183 phrase did its real work almost a century later. In Rose v. Council for Better Education (1989), the Kentucky Supreme Court held the entire Kentucky common-school system unconstitutional under § 183 — a ruling that became the seminal national precedent for the modern adequacy wave of school-funding litigation.²⁴ The Kentucky Education Reform Act of 1990 and the SEEK funding formula were the legislative responses.

It is important to be honest about what Rose is and is not. Rose is a state-constitutional ruling on a state legislature’s positive duty under § 183. It is not a school-trust enforcement case in the federal-grant sense, because Kentucky has no federal-grant trust to enforce. Rose tells the General Assembly to fund schools adequately. It does not tell anyone they hold land in trust for children and have breached that trust. The two kinds of cases live in different rooms of the law — and one of the things I want this chapter to leave the reader able to do is to tell them apart at first read.

Kentucky’s modern doctrinal record on § 184 is the counterpart to Rose’s § 183 work: Fannin v. Williams (1983) struck down a textbook appropriation for nonpublic schools by looking through the labels to the substance; Commonwealth ex rel. Cameron v. Johnson (2022) struck down a tax-credit scholarship program on the same anti-evasion principle; Commonwealth ex rel. Coleman v. Council for Better Education (February 19, 2026) extended the principle to charter schools.²⁵ When Kentucky voters defeated Amendment 2 in November 2024, declining to authorize funding for students outside the common-school system, the lockbox the 1891 framers built held.

I want to leave a marker on this case. The absence of a federal school-land grant did not prevent Kentucky from building a school-funding lockbox. Section 184’s “inviolate” clause and voter-approval mechanism function as a robust anti-diversion architecture — arguably more robust, in modern enforcement, than the federal-grant trust architecture in many public-land states. State-derived architecture can hold; holding it is continuous work, and the work runs through the 1891 framers, the 1989 court, the 2024 voters, and every legislator and parent who has had to decide, on the record, what to do with the duty.

[BOXED-INSERT placeholder: chapter 2 — see Volume_I_Boxed_Inserts_2026-05-10.md]

Tennessee, 1796: the awkwardest case, and a cautionary one

Tennessee is the awkwardest member of the cohort, and the awkwardness is worth naming. Tennessee was admitted June 1, 1796, on land North Carolina had ceded to the United States — making it, in one sense, a public-domain admission like Ohio. But Tennessee predates the Ohio template by seven years, and the federal land system never fully took hold inside Tennessee’s borders to begin with. Outstanding North Carolina warrants and settlement claims encumbered most of the federal domain inside the new state’s borders before Congress could grant anything. The 1796 admission act is a one-paragraph instrument that contains no school-land provision.²⁶

What Tennessee got instead was the Compact of April 18, 1806 — a tripartite settlement among the United States, North Carolina, and Tennessee, ratified by Congress as 2 Stat. 381.²⁷ Under the Compact, Tennessee assumed two obligations: honor outstanding North Carolina warrants, and set aside two tracts of one hundred thousand acres each, one in East Tennessee and one in West, “for the use of two colleges, one in each grand division,” together with academy acreage in each county. The Compact also directed that 640 acres in every six-mile square be reserved for common schools “where existing claims allowed.” But the prior North Carolina warrants had so thoroughly exhausted the valuable land that very few common-school tracts were ever actually laid off.²⁸ The two college tracts were sold during the nineteenth century to capitalize specific institutions rather than to build a permanent fund. The University of Tennessee at Knoxville is the historical heir to the East Tennessee tract.

What Tennessee built instead, and built well for a time, was a state-constitutional school-fund clause. The 1835 Constitution, in Article XI, § X, made the common school fund and “all the lands which have heretofore been appropriated by the General Assembly of this State for the use of common schools, and all such as shall hereafter be appropriated, [shall] remain a perpetual fund, the principal of which shall never be diminished by legislative appropriation; and the interest thereof shall be inviolably appropriated to the support and encouragement of common schools throughout the State.”²⁹ The language is classically fiduciary — perpetual fund, principal not diminished, inviolably appropriated. The 1870 Constitution retained the architecture in Article XI, § 12.³⁰

It did not survive the 1978 rewrite. On March 7, 1978, Tennessee voters approved a constitutional revision that replaced the entire Article XI, § 12 text. The new clause reads, in full: “The state of Tennessee recognizes the inherent value of education and encourages its support. The General Assembly shall provide for the maintenance, support and eligibility standards of a system of free public schools.”³¹ The change is fundamental. The 1978 clause is a service-duty clause. It is not a fund-protection clause. The perpetual-fund language, the principal-protection language, the inviolable-interest language all simply disappeared from the constitution. Whether the residual common-school-fund corpus was liquidated, absorbed into general revenue, or quietly run down is a question this chapter cannot fully answer; the financial-history record is incomplete.³²

I want to slow down on this, because Tennessee is the project’s clearest cautionary case and the lesson it carries is the one the next century of school-trust politics is most likely to need. Tennessee had built a school-fund trust out of state-derived materials. It had kept it intact for over a century and a half. Then, in a single revision cycle, it dismantled the constitutional scaffolding around it, and the architecture has not reassembled itself in the forty-eight years since. The doctrinal life of the new § 12 has been adequacy litigation, not trust enforcement. Tennessee Small School Systems v. McWherter (1993, 1995, 2002) held that the state’s school-funding scheme failed to provide substantially equal educational opportunities and ordered legislative remedies, producing first the Basic Education Program and ultimately, in 2022, the Tennessee Investment in Student Achievement Act.³³ Real work, in its own register. But adequacy litigation cannot reach back through a century of fund management. It is a different tool. The fiduciary architecture, once removed from the constitution, is not something the courts can put back.

I have come to believe two things about Tennessee, simultaneously, with equal weight. The first is that the framers of 1835 knew what they were doing — that the perpetual-fund clause, the principal-protection clause, the inviolable-interest clause were the load-bearing pieces of a trust they meant to bind across generations. The second is that the voters of 1978 were not malicious. They were operating on an assumption a great many Americans share: that the constitutional architecture inherited from earlier generations is rhetorical, that what really matters is what the legislature does this year for schools this year. The architecture sits in the document doing nothing the voter can see. Until it is gone, and then what is gone is also invisible, until the next directed-seizure moment arrives and there is no wall to hold against it. That is the cost of treating constitutional fiduciary architecture as decorative. The longer warning Tennessee offers is that constitutional fiduciary architecture, once removed, does not reassemble itself.

Maine, 1820: a Massachusetts inheritance carried forward, then drifted

Maine’s case is the simplest of the cohort below Texas. When Maine separated from Massachusetts in 1820 under the Articles of Separation — ratified by Maine voters in July 1819 and embedded in the new state’s framework as Article X, § 5 — the unsold “wild lands” of the District of Maine were divided in half. One undivided moiety remained with Massachusetts, the other passed to the new state.³⁴ The Articles required that all future grants of unlocated public land reserve, for the use of inhabitants of any town later organized, lots equivalent to the long-standing colonial pattern: typically 320 acres in each township for the support of schools and the ministry — the so-called “minister and school lots.” This was the same template Massachusetts had used since the seventeenth century, now carried forward into Maine’s land base as a state-derived (not federal-grant) school endowment. In 1853, Massachusetts sold its remaining undivided half-share to Maine for approximately $362,500, consolidating the regime under a single sovereign.³⁵

Maine’s 1820 Constitution supplied the second piece. Article VIII, Part First, § 1 — drawn substantially intact from the Massachusetts Constitution of 1780, drafted primarily by John Adams — set out the public duty: it shall be the Legislature’s duty to require “the several towns to make suitable provision, at their own expense, for the support and maintenance of public schools.”³⁶ The Adams-line civic-republican framing — education as the precondition of liberty rather than as the income from a corpus — placed Maine architecturally upstream of the 1785 Land Ordinance, in the same constitutional family as Pennsylvania 1776 and New Hampshire 1784.

The early operational history of Maine’s public lots is one of administrative drift. Many lots were never located; the unincorporated North Woods townships were thinly settled and unsurveyed. The 1850 statute authorized the Land Agent to sell rights to cut and carry away timber and grass from public reserved lots; between 1850 and 1857 alone, more than 230,000 acres were reported sold for cutting rights.³⁷ The doctrinal turn came late. In 1973, Lee M. Schepps of the Maine Attorney General’s office authored Maine’s Public Lots: The Emergence of a Public Trust, arguing that nineteenth-century cutting-right conveyances may have covered only timber then existing, not all subsequent growth.³⁸ In Cushing v. State (1981), the Maine Supreme Judicial Court adopted that theory, holding that nineteenth-century cutting-right conveyances did not include trees not in existence when the Land Agent made the conveyances.³⁹ Cushing restored future-growth value to the State’s public reserved lots and is, in the school-trust frame, Maine’s closest analogue to a corpus-recovery decision.

But the school-benefit thread had largely been severed by then, and I want to be honest about how. In an advisory Opinion of the Justices (1973), the Court construed “schools” and “ministry” in the original Articles as illustrative rather than exclusive limits.⁴⁰ The Legislature reoriented the public-lot regime accordingly: lots would be used for the benefit of the State as a whole. Maine today manages approximately 600,000 acres of Public Reserved Lands under the Bureau of Parks and Lands.⁴¹ Earnings flow to a Public Reserved Lands Management Fund dedicated to land management, not to schools. A narrow 2017 logging-and-forestry-education grant program, capped at $300,000 per biennium, is the only standing channel through which Public Reserved Lands revenue reaches a school-coded purpose. K-12 funding flows through a separate architecture: heavy reliance on local property taxation under the Adams-derived constitutional duty, layered with the Essential Programs and Services formula and a 55-percent state-share target frequently not met. The architecture remembers; the revenue stream has moved on.

Maine is the cohort’s quiet drift case. No directed seizure, no constitutional rewrite. Just the slow legal reinterpretation of “schools” as one purpose among several, the slow administrative absorption of the lands into a multi-use management regime, and the slow disappearance of the school-benefit purpose from the place in the budget where it used to be. Drift, of the patient kind. Drift is harder to point at than seizure precisely because there is no single moment to point to. That is part of what makes it dangerous.

West Virginia, 1863: the war improvisation that built fiduciary architecture out of nothing

West Virginia is the most architecturally ambitious of the cohort below Texas. Its statehood was an artifact of the Civil War — the only state ever carved out of an existing one over the active objection of the parent state’s Confederate government — and the lands within West Virginia had never been part of the federal public domain. They had belonged to Virginia. Congress could not grant the new state a sixteenth-section endowment because Congress did not own a sixteenth section to grant.⁴² The December 31, 1862 consent act and Lincoln’s April 20, 1863 proclamation contained no school-land grant; the conditions concerned slavery and constitutional ratification, not federal land.

At separation, an obvious question was whether West Virginia would inherit a pro-rata share of Virginia’s 1810 Literary Fund corpus — the state-derived permanent fund Chapter 1 walked through. The answer, eventually, was no. The post-separation debt-and-assets settlement became multi-decade litigation, culminating in the U.S. Supreme Court’s decisions in Virginia v. West Virginia, 220 U.S. 1 (1911), and 246 U.S. 565 (1918), which apportioned debt obligations but produced no Literary Fund transfer.⁴³ West Virginia’s School Fund had to be built, not inherited.

The 1872 Constitution, in Article XII, did the building. Section 1 imposed the duty: “The Legislature shall provide, by general law, for a thorough and efficient system of free schools.”⁴⁴ Section 4 carried forward a separate School Fund — capitalized from forfeited and delinquent lands, escheats, grants and bequests for school purposes, and other state-derived revenue streams — with the principal “inviolate” and the interest applied to free schools “and to no other purpose.”⁴⁵ Read together, Article XII gave West Virginia a hybrid: a legislative-duty floor under § 1, and an irreducible state-derived corpus under § 4.

The corpus did not, however, accumulate the way the framers may have imagined. In 1902, voters ratified the Irreducible School Fund Amendment, which capped the permanent corpus at $1 million and directed excess receipts into the general school fund for current support.⁴⁶ That redirection still flowed to schools, but it ensured West Virginia’s constitutional permanent fund would remain small. The Supreme Court of Appeals confirmed the consequences in Board of Education of Wyoming County v. Board of Public Works (1959) and reaffirmed the boundary in State ex rel. Marockie v. Wagoner (1993): what the legislature could later alter was, by definition, not what Article XII had constitutionalized.⁴⁷

The doctrinal earthquake came in 1979. In Pauley v. Kelly, the Supreme Court of Appeals held that Article XII § 1 was judicially enforceable, that education was a fundamental constitutional right, and that the state was required to provide a high-quality and substantially equal system of free schools across all counties.⁴⁸ Judge Arthur M. Recht’s May 1982 decree on remand built a comprehensive, court-supervised remedial framework addressing facilities, curriculum, teacher quality, and finance equity. Pauley v. Bailey (1984) affirmed.⁴⁹ Pauley is among the most cited adequacy decisions in American state constitutional law, and along with Kentucky’s Rose it gave the broader adequacy movement one of its founding decisions. State v. Beaver (Nov. 17, 2022) recently held that West Virginia’s Hope Scholarship Act did not facially violate Article XII §§ 1, 2, 4, or 5, emphasizing that the Hope Scholarship was funded from general revenue rather than from the constitutionally walled-off § 4 School Fund.⁵⁰

I want to put the West Virginia takeaway plainly. Even where no federal land grant was on offer, the school-clause civic instinct produced fund architecture. The framers of 1872, with no federal sections to draw on and no Virginia Literary Fund inheritance, built a constitutionally-dedicated permanent fund out of state-derived revenue streams — escheats, forfeitures, delinquent lands. The fund stayed small. But it stayed real, and it stayed walled off, and the courts have enforced the wall. West Virginia is not a public-land trust state. But it is, in its own state-derived way, a school-trust state. As someone who has spent my career inside this architecture, I find the West Virginia case quietly heartening. It says that even a state coming into the Union in the middle of a war, with no federal grant, no parent-state inheritance, and the wreckage of a contested separation to negotiate, can still write fiduciary architecture into its founding document if it chooses. The instinct does not require federal underwriting. It requires the choice.

Texas, 1845: what a sovereign builds when it keeps its own land

Texas is the case that breaks the pattern, and the case worth lingering on.

The Republic of Texas, before it joined the Union, had already done some of the work. As early as 1839 — three years after independence from Mexico — the Republic’s Education Act dedicated three leagues of land per county to public school support, expanded to four leagues in 1840.⁵¹ When annexation came in 1845, it came not by ordinary admission-act path but by Joint Resolution of Congress, and the resolution carried an unusual condition. Texas would enter the Union as a state, but Texas would keep its own public domain. The federal government would acquire no public lands inside Texas’s borders. The operative paragraph of the Joint Resolution of March 1, 1845 — 5 Stat. 797 — provides:

said State, when admitted into the Union, after ceding to the United States, all public edifices, fortifications, barracks, ports and harbors, navy and navy-yards, docks, magazines, arms, armaments, and all other property and means pertaining to the public defence belonging to said Republic of Texas, shall retain all the public funds, debts, taxes, and claims of every kind, which may belong to or be due and owing said Republic; and shall also retain all the vacant and unappropriated lands lying within its limits, to be applied to the payment of the debts and liabilities of said Republic of Texas, and the residue of said lands, after discharging said debts and liabilities, to be disposed of as said State may direct.⁵²

A vast retained public domain — historical record places it at roughly 175 million acres at annexation — disposable as the state may direct.⁵³ That is the foundation Texas built on.

The 1854 capitalization came from an unexpected direction. The Compromise of 1850 had settled the boundary disputes that followed Texas’s admission, particularly Texas’s expansive claims to territory in present-day New Mexico, Colorado, Wyoming, Kansas, and Oklahoma. Under the Compromise, Texas relinquished those claims in exchange for $10 million in United States bonds. Approximately $7 million was used to retire the remaining debt of the Republic, satisfying the condition the 1845 Joint Resolution had attached to the retained lands. That left $3 million in U.S. bonds in the state’s hands.⁵⁴ In January 1854, the Texas Legislature appropriated $2 million of those bonds to a “Special School Fund” — the direct legal precursor of the Permanent School Fund. The first distributions reached Texas common schools in 1855.⁵⁵ The seed corpus of the Texas school endowment was federal bonds, transferred to Texas as consideration for the cession of territorial claims, with the rest of the architecture supplied by the state. From its inception, the Texas school endowment was a hybrid of real and financial assets — and that hybrid character would, more than 150 years later, become the architectural pivot for the dual-trustee structure the 1876 Constitution institutionalized.

The Civil War subjected the young Special School Fund to its first directed-seizure pressure. Confederate Secretary of War Judah P. Benjamin proposed cashing the school fund’s U.S. bonds and replacing them with Confederate bonds — a transaction that, had it been completed, would have annihilated the corpus when the Confederacy collapsed. The bond substitution was never executed, but Texas war boards did access portions of the fund for military finance, and the corpus emerged from the war reduced to approximately $1.8 million.⁵⁶ I want to register that detail because it is one of the cleanest period-source illustrations in the substrate of what directed seizure looks like as a live political proposition. The corpus survived because a particular transaction was not completed. It nearly did not.

The recovery began under Reconstruction Governor Edmund J. Davis. On March 18, 1873, Davis signed House Bill 283, which dedicated half of the remaining unallocated public domain — approximately 88 million acres at the time — to the school fund.⁵⁷

The 1876 Texas Constitution — the seventh Texas constitution, still the current one — set the building’s foundations into the bedrock. I want to walk through the relevant sections deliberately, because together they are the closest thing in American state constitutional law to a complete fiduciary architecture for a school trust written without any federal-compact backstop. Article VII, § 1, declared that “a general diffusion of knowledge being essential to the preservation of the liberties and rights of the people, it shall be the duty of the Legislature of the State to establish and make suitable provision for the support and maintenance of an efficient system of public free schools.”⁵⁸ Article VII, § 2, dedicated to the Permanent School Fund “all funds, lands and other property heretofore set apart and appropriated for the support of public schools; all the alternate sections of land reserved by the State out of grants heretofore made or that may hereafter be made to railroads or other corporations of any nature whatsoever; one half of the public domain of the State; and all sums of money that may come to the State from the sale of any portion of the same.”⁵⁹ Half. The single largest dedication of public lands to schools by any state in American history, expressed in one sentence of a state constitution. Article VII, § 4, governed sale and investment of PSF lands, with a structural anti-rescission provision: “the Legislature shall not have power to grant any relief to purchasers thereof.”⁶⁰ Article VII, § 5, segregated the Available School Fund — the distributable income — from the Permanent School Fund corpus and governed distribution.⁶¹

The architecture also built in something most public-land states never did, and this is the piece I want to leave the reader thinking about hardest. Texas built in two trustees, both constitutional, both elected statewide. The Commissioner of the General Land Office runs the school land portfolio — approximately 768,000 surface acres and roughly 13 million mineral acres, with the Permian Basin oil and gas at the heart of recent revenue. The State Board of Education runs the financial-asset portfolio. Each is accountable to Texas voters directly, in their own elections.⁶² That is unusual. Most public-land states delegate trust management to ex-officio boards of executive officers (Oregon among them) or to a single elected commissioner. Texas put two separately-elected fiduciaries in the room.

I belabor this point because the dual-trustee design is the part of the Texas architecture that the AI-era trust architects are most likely to need in their own toolkits. The standard governance design for a long-horizon trust — at the Founding and in our own moment — is to consolidate trustee authority in one office. The Texas design splits it. The land-trustee and the financial-trustee answer to the same beneficiary class but operate under different incentives, on different time horizons, with different professional cultures. They are not redundant; they are corrective of one another. When the financial trustee wants to be aggressive, the land trustee’s slower cadence holds the corpus back; when the land trustee wants to liquidate at speed, the financial trustee’s investment discipline pushes back. Two professionally-distinct fiduciaries in the room, answerable directly to the people whose children the trust exists to serve, is not the only design that works. It is, on the evidence the next century and a half produced in Texas, a design that worked unusually well.

The growth that followed is its own kind of evidence. The Spindletop discovery of 1901 opened the Texas petroleum era. House Bill 9 of 1939 dedicated the mineral estate in all riverbeds, lakes, channels, and tidelands to the Permanent School Fund and created the School Land Board.⁶³ The Tidelands controversy of 1950–60 — United States v. Texas (1950), the Submerged Lands Act of 1953, United States v. Louisiana (1960) — produced, on the whole, a recovery of submerged-mineral title that flowed directly to the school fund, with offshore oil and gas royalties contributing more than $9 billion to the corpus by 2010.⁶⁴ The 1983 amendment authorized the PSF Bond Guarantee Program, which now stands behind approximately $125.8 billion in school-district and charter-district bond debt and produces an estimated $425 million per year in taxpayer interest savings — a use of the corpus that no other state’s school trust replicates at scale.⁶⁵ The 2003 amendment (HJR 68) authorized total-return investing.⁶⁶ Senate Bill 1232 in 2021 created the Texas Permanent School Fund Corporation to unify management of the fund’s investment assets.⁶⁷ The corpus crossed fifty billion dollars. The annual distribution to the Available School Fund crossed a billion. The Available School Fund flows out as a per-capita payment to every Texas school district and as the funding source for the state’s Instructional Materials Allotment, the textbook line for every public school in Texas.

What Article VII accomplishes, taken as a whole, is the construction in state constitutional law of every architectural feature that the Northwest Ordinance, the 1850 Oregon Donation Act, the 1875 Colorado Enabling Act, and the 1910 New Mexico–Arizona Enabling Act constructed in federal law for other states. The corpus is named, dedicated to a beneficiary class (“the children attending the public free schools of this State”), declared perpetual, and segregated from general revenue. The fiduciary is named — the State Board of Education for investment, the Land Commissioner for land. Investment is constrained. The Legislature is barred from rescuing defaulting purchasers. Modern Texas Attorney General opinions read all of this as creating a fund that is constitutionally “nonspendable” except where Article VII expressly authorizes payments — most clearly in Opinion KP-0229 (2018), with a substantial supporting body of opinions including GA-0407 (2006) blocking uncompensated transfer of artificially reclaimed PSF submerged lands and GA-0707 (2009) governing the total-return distribution methodology.⁶⁸ The federal compact-trust analogue is absent. The architectural function is fully discharged in state text.

A federal grant is sometimes described, in the project’s anchor phrase that I have used for years, as “a forever gift to forever schools for a forever democracy.” Texas built one without any federal grant at all.

What it means

Set Chapter 1 next to this chapter and a finding falls out that the rest of the book leans on hard.

The original thirteen states had no federal school grant because the federal government did not yet exist as a landholder when they joined the Union. The six states in this chapter had the same option, with later starts and in some cases — Texas above all — with vastly more land to work with. Five of the six built modest state-derived architecture, or none, or, in Tennessee’s case, built a real one and then dismantled it. One built the largest educational endowment in the country.

The federal grant, on the evidence of these eighteen states alone, is neither necessary nor sufficient. Texas had no grant and built the largest fund in the United States. Vermont had no grant and wrote, in 1777, the school-clause sentence that the federal templates of 1785 and 1787 would later codify. Kentucky had no grant, was capitalized partly by federal cash, and produced Rose v. Council for Better Education — the seminal national precedent for school-funding adequacy litigation. West Virginia had no grant, no Virginia Literary Fund inheritance, and a war on, and still managed to write Pauley v. Kelly. Maine had no grant and built nothing much. Tennessee had no usable grant, built constitutional fund-protection architecture out of state-derived materials, kept it for 143 years, and then voted it out in 1978. The presence of a federal grant did not, in itself, rescue any state from its own choices, and the absence of one did not prevent any state from making the architectural moves that produce a real endowment.

I want to be precise about what did matter, across these cases, because the chapter exists to leave the reader with that. What seems to have mattered is something more basic and harder to legislate from outside. A constituency — people who know what the trust is and care that it survive. A constitution that does the architectural work — corpus, beneficiary class, fiduciary, segregation, perpetuity. Trustees who are accountable, on the record, to the people whose children the trust exists to serve. Texas had all three early and locked them in. Kentucky and West Virginia had two of the three, embedded in 1891 and 1872 respectively, and the architecture has held. Vermont built decentralized parallel architecture and protected it for 250 years across 250 towns. Maine inherited a colonial template, drifted away from it, and recovered part of it through Cushing. Tennessee built the architecture, held it for over a century, and then voluntarily took it out — and the architecture has not reassembled itself.

I have come to believe two things, simultaneously, with equal weight, after a career inside this architecture. The first is that the federal grant, where it was given, did real work — it scaled an idea that was already in the air, attached a corpus to a duty the states had been articulating for themselves, and gave the states a fiduciary frame whose vocabulary the courts could later reach for. The second is that the federal grant was never the load-bearing piece of the trust. The load-bearing pieces were the ones the states themselves had to build: the constitutional architecture, the elected trustees, the constituency that would notice when the architecture started to drift or when a directed seizure was being attempted. Where those three were built and held, as in Texas, the trust has produced extraordinary results. Where any of the three was missing, or quietly removed, the trust has produced what the missing pieces left behind.

Chapter 3 turns to the states the Land Ordinance template did reach — the federally surveyed states from Ohio onward, where Section 16 actually showed up on the plat. Those states received the federal floor this chapter’s six did not. What they did with it varies more than the federal template alone would predict, in directions worth following.

— Margaret Bird, Salt Lake City, 2026.

Chapter 2 Footnotes