Trace the gift rather than the tax line. Nonprofit status is the thing that makes a donation deductible in the first place [7], which means the program that could cost a school its exemption is in many cases funded by gifts whose attractiveness to the donor depends on the exemption that program endangers, and gift agreements are being signed this quarter against a rule that bites in 2027.
Treasury did not price its own proposal, so put explicit assumptions on its ceiling: if one in a hundred of the institutions in scope actually lost the benefit, and the benefit were worth two million dollars a year to each, which sits inside the "millions" the century-old exemption is generally described as saving a university [6], that is 180 schools and roughly $360m a year [17]. Small against federal receipts, and roughly a year of operating surplus at a small private college.
The calendar is the part a finance office can actually act on. The proposal is dated September 3, 2026 [15] and, if finalized, applies after May 2027 [2], which is just under nine months [16] to absorb a final rule, rewrite catalog language and aid rubrics, and requalify pledges, all inside a single admissions and aid cycle.
The only completed precedent is Bob Jones University, which lost its exemption in the 1970s over a ban on interracial dating and marriage, saw the Supreme Court uphold the IRS, and did not regain the benefit until 2017 [10], a process that dragged through litigation for decades before it was finally reversed. The cheaper outcome for Treasury, and the more interesting version, is that it never revokes anything at all, because anticipatory compliance is already widespread and the department has said in advance that renaming a program is not a defense, so the deterrent costs the government nothing while the legal review costs the sector real money either way. The exit is that the rule never binds: Harvard's position, when its own exemption was threatened last year, was that there is no legal basis for revocation, and it argued the loss would force cuts to financial aid and medical research [11], while statute bars the IRS from targeting organizations for ideological reasons and bars federal officials from directing IRS investigations [13].
This is probably wrong, but I would weight the middle path highest, because it is the one that requires no litigation risk and no revenue estimate, and because the behavioural change it seeks has largely already happened. The thesis fails in two identifiable ways. If the final rule narrows to explicitly race-labelled programs, the rebrands hold and 2026's legal fees were the entire bill. If instead the IRS opens an examination of a named institution before the effective date, the exposure stops being an option and becomes a number on someone's audited statements.
One figure is still missing. The proposal as reported covers revocation of exempt status and the deductibility of gifts, but not how a revoked institution's endowment income would then be taxed [20], and until that is written down, no board can size the downside it is being asked to underwrite, which is a poor basis for approving next year's aid budget.
