A recent LinkedIn post by Dr. Jillian Goldfarb called attention to an invoice most of us never see. She had been asked to peer review a paper for a Nature-family journal. The article processing charge, paid by the author to make that single paper open access, was $12,850.
That figure deserves a moment. It is not a typo. It is more than the summer stipend for a Ph.D. student. It is two undergraduate research fellowships. It is real scientific labor, priced out of existence so that one paper can be read without a paywall. Goldfarb’s point was not that the number was shocking. It was that the number was routine.
What makes the charge worth examining is exactly that. Publishers list these fees openly. Grant offices factor them into budgets. No one, anywhere in the chain, treats them as an emergency. A system that has trained its participants to shrug at a five-figure price for a single PDF has done something to itself that is worth looking at carefully.
From Access Problem to Allocation Problem
Open access was sold as a fix. Research paid for by taxpayers should be readable by taxpayers. Locking publicly funded knowledge behind subscription paywalls was a scandal, and the movement to break those paywalls had the moral high ground.
The fix worked. And then it did something else. The paywall moved.
Under the dominant open access model today, readers no longer pay. Authors do. That change sounds neutral. It is not. Paywalls that face readers block access to knowledge. Paywalls that face authors block access to visibility. The barrier did not disappear, it relocated, from the end of the pipeline to the beginning, where it is much harder to see and much easier to ignore.
Visibility in modern science is a function of where you publish, and where you publish is increasingly a function of what you can pay. The question is no longer whether knowledge can be read. It is whose knowledge gets to be read at all.
Who Absorbs the Cost
Here is how that plays out. A principal investigator at a well-resourced R1 has layers of cushion. Central library funds, read-and-publish agreements negotiated at the consortium level, discretionary pools inside the dean’s office, indirect cost recovery flush enough to absorb a $12,000 APC without flinching. The lab treats the fee as friction, not an obstacle. The paper goes out.
A principal investigator at a small regional university, a teaching-focused institution, or a minority-serving school has none of that. The APC comes out of grant direct costs, which means it competes with the student who would have been hired for the summer. The fee is not friction. It is a trade-off, and the trade-off has a face. A name. A stipend not paid. A project not launched.
Both PIs are writing the same kind of paper. Both are following the same peer review process. Only one of them is making a decision that costs a student a research position. That is not market neutrality. That is a system that has allocated its costs in a way that presses hardest on the people already carrying the heaviest loads.
HBCUs as the Stress Test
HBCUs are where the system shows what it actually is. Not because they are an exception to how American higher education works, but because they are what the rest of the system looks like once you remove the cushioning.
The constraints are layered. HBCUs operate under a documented history of inequitable state and federal funding, the consequences of which compound across decades. Private HBCUs draw close to 45 percent of their revenue from tuition and fees, compared to 37 percent at peer private institutions, leaving little discretionary margin for research subsidies. Their mission to serve students with limited financial resources rules out the tuition increases other schools use to offset state divestment. Their grant portfolios are smaller, their indirect cost rates lower, their administrative research support leaner.
Now add a $12,000 APC on top of all of that, paper after paper. The cost does not land on a well-padded central budget. It lands on a single grant, held by a single faculty member, who is already stretching to fund students. Community colleges, where applied research and workforce-adjacent scholarship genuinely happen, face a compressed version of the same pressure with even less institutional infrastructure to absorb it. The system behaves differently under constraint. Open access, at its current price, functions as a tax the constrained cannot afford to pay.
Gatekeeping Without Gatekeepers
Nobody has to be told they cannot publish. No journal hangs a sign on the door saying faculty from under-resourced institutions are not welcome. The filtering is quieter than that.
It works through the invoice. It works through the grant budget narrative, where the PI at a small school writes an APC line and watches the rest of the project shrink to accommodate it. It works through the tenure file, where publication venue matters and the highest-impact venues charge the highest fees. It works through a thousand individual decisions by individual faculty members who look at a $12,850 price tag and pick a different journal, not because their work is weaker but because their budget is smaller.
The result is gatekeeping without gatekeepers. No villain, no explicit exclusion. Just a price structure that sorts, paper by paper, invoice by invoice, into those who can pay to be seen and those who cannot.
The Triple-Dipping Asymmetry
The imbalance runs deeper once you look at how institutions interact with publishers. Large universities negotiate transformative agreements, sometimes called read-and-publish deals, where the institution pays a bundled fee that covers both subscription access and APCs for its faculty. Faculty at those institutions publish open access at no direct cost to the lab. The fee is absorbed centrally and invisibly.
Smaller institutions, and especially under-resourced ones, rarely have the leverage, the consortium membership, or the legal staff to negotiate these arrangements. Their libraries pay subscriptions. Their faculty pay full-price APCs. And in some cases their institutions also contribute to bundled agreements whose economics favor the larger consortium members. The same publisher collects at three different points in the transaction, and the distribution of those payments is not even.
This is not a market failure. This is a market working exactly as designed, for the participants who had the power to design it.
The Reshaping of the Record
What does a system like this produce over time? A scholarly record weighted toward the institutions that can afford to be in it.
Citations flow to the papers people can find. People find the papers that are open access in the highest-impact venues. Highest-impact venues charge the highest APCs. The institutions that can afford those APCs publish more, get cited more, win more grants, and negotiate better read-and-publish deals, which lowers their costs further and widens the gap. The feedback loop is tight and self-reinforcing.
A faculty member at an HBCU publishing equally strong work in a lower-APC venue, or in a subscription journal, does not disappear from the literature. She just becomes harder to find, less frequently cited, and therefore, over a career, less visible to the funders and committees who decide what research matters. The scholarly record is not simply a record of what was discovered. It is a record of what could afford to be broadcast.
The Data Void
One more feature of this system deserves attention. There does not appear to be a systematic empirical study quantifying what HBCU and community college faculty actually spend on APCs, or how their publication patterns differ as a result. The numbers that would let policymakers act on the disparity have not been collected.
The absence is convenient. You cannot regulate what you cannot measure. You cannot redirect funds to fix a problem that has not been documented. A missing dataset is not a neutral gap, it is a structural feature of a system that has no incentive to make its own asymmetries legible. The research opportunity here is real, and it is waiting for someone with the institutional standing to ask the question in a form that cannot be ignored.
Where the System Could Be Pressed
None of this points to a single fix, but it does point to places where the system could be pressed.
Federal funders could clarify whether APCs are an allowable cost, and if so, whether they should be capped, scaled to institution size, or shifted out of direct costs entirely. Transformative agreements could be required to demonstrate equity across consortium members, rather than across total volume alone. Diamond open access models, where neither readers nor authors pay and the costs are carried by societies or public funders, deserve to be studied seriously rather than dismissed as unrealistic. Consortia could be built specifically for under-resourced institutions, giving them collective negotiating power they lack individually.
None of these ideas are new. Each one has advocates and pilots already in motion. What has been missing is the political pressure to treat the current arrangement as a problem rather than a default.
The Question Beneath the Model
Back to the number. $12,850 to make one paper readable.
The debate over open access has long been framed as a debate about access to reading. That framing is increasingly beside the point. The more consequential question is access to authorship. Who gets to put their name on a paper in the venues that shape fields, win grants, and earn tenure? Who gets to be cited, taught, built upon?
Right now the answer is tilting sharply toward institutions that can pay. HBCUs, community colleges, small regional universities, and the researchers working inside them are being sorted out of the highest-visibility tiers of American science, not because their work is weaker, but because the price of admission has quietly moved beyond their reach. Good ideas can come from anywhere. That has always been the premise. A pricing structure that contradicts the premise, paper by paper, is not a billing practice. It is a statement about whose knowledge is valued enough to be seen. The answer that statement gives, one invoice at a time, is one the rest of us should refuse to accept.