School Voucher Programs and State Constitutional Limits
A dollar that reaches a religious school through Arizona's tax credit program has never, in the eyes of that state's own supreme court, touched the public treasury at all. A dollar that reached the same kind of school through Florida's Opportunity Scholarship Program was treated, by Florida's, as exactly the public money its constitution forbids diverting outside a uniform system. Neither ruling turned on anything the United States Constitution says. The difference was architecture, and architecture is now doing most of the work in this area of law.
Start with what the federal cases actually closed off, because it is less than voucher opponents once hoped and less than voucher proponents now claim. Zelman v. Simmons-Harris, 536 U.S. 639 (2002), ended the Establishment Clause objection to the basic mechanism nearly a quarter century ago, holding that a scholarship program is religiously neutral so long as aid reaches religious schools only through the independent choices of parents rather than by government design. Nobody credibly argues anymore that a voucher program is unconstitutional merely because parents are free to choose a parochial school with it. That closed off the one line of attack that would have worked the same way in every state at once, since the Establishment Clause binds all fifty of them identically. Everything that has happened in this area of law since has instead played out state by state, because nothing left in federal doctrine forces a uniform answer.
The Free Exercise fight that replaced it is narrower than it sounds. Espinoza v. Montana Department of Revenue and Carson v. Makin hold that once a state opens a scholarship or tuition program to private schools generally, it cannot carve religious ones back out. Neither case requires a state to open such a program in the first place, and the Court took care in Espinoza to distinguish rather than overrule Locke v. Davey, 540 U.S. 712 (2004), where Washington's exclusion of devotional theology majors from an otherwise available scholarship survived because it turned on what a student proposed to study, not on what he was. A state that funds nothing but its own public schools violates no one's free exercise rights by that choice alone.
That leaves state law to decide whether a program may exist in the first place, and courts reading their own state's version of a similar guarantee have split down the middle. Indiana let its voucher program stand under an education clause dating to 1851. Florida did not let a strikingly similar one stand, under a uniformity clause the Florida Supreme Court read through the lens of a 1998 constitutional revision the majority leaned on heavily. Both provisions commit the legislature to a common school system in roughly the same words; where they were amended and how recently turned out to matter to how much room each court found for spending outside that system. A legislature drafting a new program has, from the case law alone, no way to know in advance which reading its own state's courts, and its own state's amendment history, will produce.
Arizona shows a third way to reach the same economic outcome without ever asking the uniformity question at all. Its tax credit program lets a taxpayer redirect up to a set amount of what would otherwise be owed to the state treasury into a nonprofit that funds private school tuition, including at religious schools. When that program reached the state's own supreme court, the majority held that no public money was ever at issue, because the funds are privately donated at the donor's own choice and never pass through the treasury, no matter how large a hole the aggregate credits leave in state revenue. A no-aid clause written to stop the state from spending public money on religious education has nothing obvious to say about money the state never collected from a particular taxpayer in the first place.
The same state proves the point a second time, on the opposite side of the ledger, only a decade later. Arizona had also built a program that funded private tuition for disabled and foster children directly out of the state treasury, no tax credit involved, and when that program reached the same court, it did not survive. Cain v. Horne, 202 P.3d 1178 (Ariz. 2009), held unanimously that direct appropriations of this kind violate the state constitution's own Aid Clause, barring public money spent in aid of private or religious education, without the court needing to reach a second constitutional provision the challengers had also raised. Nothing about the underlying policy shifted between the two rulings, decided a decade apart by two entirely different sets of justices, Kotterman by a bare three-to-two majority and Cain unanimously. What changed was only the mechanism: one program spent money the state had already collected, the other let a taxpayer redirect money before the state ever collected it, and Arizona's highest court, in each of the two eras it sat, treated that difference as decisive on the same underlying question of who gets to attend a religious school on the public's dime.
The practical consequence is that a state's Blaine-era text predicts less than it appears to. A legislature facing a hostile uniformity clause of the Bush v. Holmes variety is not necessarily blocked from getting a voucher program's economic effect past that same court; it may only be blocked from getting there through a direct appropriation. Structure the same transfer as a credit against taxes never collected, and the constitutional objection that would have doomed a check from the state treasury may have nothing left to attach to, because the theory that money belongs to the state until it spends it does not obviously reach money the state agreed in advance not to collect. A no-aid clause written a century ago to keep public money out of religious education has, against that kind of architecture, no public money left to keep out.