When the National Institutes of Health announced on 7 February 2025 that it would cap indirect cost recovery at 15 percent for new and existing grants, the short-term fight was about laboratory salaries, animal facility contracts, the administrative machinery that keeps clinical trials running, and maintenance agreements for core instruments. The longer-term fight is now moving into quieter rooms: university libraries. Acquisitions librarians at several research institutions are preparing cancellation lists for the 2026 renewal cycle, and consortia that negotiate bulk journal packages are asking members to confirm which titles they can legally drop under existing multi-year contracts.
Three days after the NIH notice, a federal judge in Massachusetts temporarily blocked the policy while a lawsuit brought by 22 states moves forward. The order did not settle the budget arithmetic on campus. Finance officers at public research universities are treating the proposed cap as a contingent liability, and that contingency is large enough to change the renewal decisions libraries make in late 2025.
The mechanism is less direct than many faculty assume. Journal subscriptions are usually paid from the university's general operating budget, not from individual lab grants. But the general operating budget at a research university is partly built from indirect cost recovery, and that recovery has for years been used to cover the central costs a university cannot charge directly to a grant: library collections, research computing, compliance offices, and facility depreciation. When indirect recovery drops, the library does not lose a named grant line. It loses the top-up that makes its flat campus allocation enough to sustain a multi-million-dollar serials portfolio.
The path from an overhead cap to a serials list
Librarians describe the sequence in concrete terms. A college dean gets a revised budget memo in March or April. The dean protects staff and graduate student stipends first, then looks at recurring expenditures that are easy to defer. Journal packages usually fall into that second category because they renew annually and can be dropped, renegotiated, reduced, or postponed without immediate layoffs. The library then receives an instruction: prepare a cut list for the fall renewal meeting.
These cuts usually do not look like a single announcement. A chemistry department loses access to a specialized database. A nursing school's point-of-care reference tool is reduced to a subset of modules. A philosophy department's package of society journals is allowed to lapse, while a history department's primary-source database is moved to intermittent access. In isolation each item appears manageable. In aggregate, the collection itself begins to narrow along predictable lines: titles heavily used by professional schools survive, while journals used by smaller humanities and social science departments are approved for cancellation.
What gets cancelled first
The order of cancellation at most research libraries follows a pattern set in earlier budget squeezes, and it is not the pattern researchers expect. The most expensive commercial packages are often the hardest to cancel because they are bundled into multi-year read-and-publish deals. Instead, libraries cut around the bundles.
- Standalone society journals and small publisher titles, where per-use costs are high and the loss is felt by a single department.
- Secondary indexes and abstract databases that duplicate searchable content available in larger platforms.
- Clinical point-of-care tools and educational products licensed to a single professional school, often with per-student fees that rise each year.
- Print subscriptions that remain as add-ons to electronic packages, along with backfile access that is rarely used.
What remains is the bundle itself, paid on schedule, because leaving would trigger penalty pricing or eliminate access to high-demand titles. That means the cancellation pain is uneven: a medical school may keep its Elsevier package while comparative literature loses three society titles. The budget pressure is real everywhere, but the loss is concentrated in fields that had smaller leverage to begin with.
Equity inside the collections budget
The distinction between what a reform changes on paper and what it changes in practice is visible in the library budget. On paper, a cap on indirect recovery applies to every research institution with NIH funding. In practice, institutions with large endowments or state backstops can absorb the shock through internal transfers. Public universities in states that have already cut per-student funding are more likely to pass the loss into the collections budget, and the institutions with the thinnest endowments face the deepest cancellation list. That is the equity gap the litigation over the NIH cap has not yet addressed.
Community colleges and regional universities with little NIH activity may feel the least direct effect on their collection budgets, though they have their own state-funding pressures. The sharpest effect is concentrated in the research-intensive universities that the Carnegie Classification labels R1, the institutions that run medical schools, public health programs, and large federally funded research portfolios. Those are precisely the libraries that have long carried the biggest serials portfolios and the highest expectations from faculty.
Publishers and the shift to open access
Commercial publishers have not offered universal price relief. Many of the large read-and-publish agreements signed between 2021 and 2024 fixed annual increases into multi-year contracts, which locks institutions into paying both subscription fees and open access publication charges during the same period. A library facing an NIH-related budget cut cannot easily reduce that fixed cost; it can only trim the titles outside the contract. SPARC's public record of big deal cancellations by research libraries, accessible through the Big Deal Cancellation Tracking page, shows that institutions rarely exit the largest bundles in a crisis. They cancel periphery first.
The federal access rules add a separate stress. Under the public access updates described by the Nelson memo, agencies are pushing for immediate article deposition in repositories, but the subscription system still underpins the paywalled archive. AcademicJobs has reported on the Nelson memo's implementation; the two pressures now operate at once. The policy that says the public should read the research for free and the budget rule that makes it harder to buy the journals are both arriving at the same moment, and the library is where that contradiction gets resolved by spreadsheet.
Year two: what deans will confront
The NIH notice itself remains tied up in court. The indirect cost cap was announced in Notice NOT-OD-25-068 and blocked on an emergency basis in the District of Massachusetts, but the planning impact has already moved ahead of the litigation. Even where the cap is lifted, grant terminations and funding freezes under other executive actions have kept research budgets unpredictable. AcademicJobs reported earlier on the NIH grant terminations and funding freeze, and that uncertainty is driving the same decisions the cap itself would have forced.
The year-two question is not whether the litigation will succeed, but how many campuses will restore a cancelled journal once it is gone. Cancellations create usage data. A title that falls off a platform for a year may show lower demand simply because access was harder, and the reduction justifies the cut. Libraries that cancel small society bundles this fall will find it difficult to defend re-adding them in 2027, even if NIH funding stabilizes. The savings become permanent because the absence becomes evidence. That is the practical effect of policy on collections: the first cut is a budget line, the second cut is a research habit, the third cut is the argument that nobody was using the title anyway, and the fourth cut is the conference trip that no longer happens because the article was never read.
Photo by Alexander Grey on Unsplash