Peer review — the system that stands between a manuscript and publication — is performed almost entirely without payment. A typical manuscript receives two to four referee reports, each taking several hours to a day of expert work, and the referee receives, in the standard arrangement, nothing: no fee, no course release, rarely even formal credit. Estimates of the aggregate effort are striking. One widely cited analysis by Balázs Aczel and colleagues calculated that US-based researchers alone contribute on the order of a hundred million hours per year to peer review — work whose implied value runs into the billions of dollars if priced even modestly. Journals receive that labor free and, in many cases, sell the resulting product back to university libraries through subscriptions.
This is an explainer on the economics of academic publishing's labor, not commentary on any specific journal.
What does a review actually involve?
The labor is easy to underestimate. A competent referee does not merely read: she re-derives the central estimate where possible, checks whether the sample supports the claims, hunts the literature for contradictory results the authors missed, assesses whether methods sections hide analytic flexibility, and writes a report that must be technically precise, constructive in tone, and delivered within weeks — often for a paper adjacent to, but not inside, her own specialty. Editors' guidelines routinely assume four to eight hours; surveys of reviewers' self-reports commonly run higher. Multiply a conservative half-day across several million manuscripts a year in a single large field, and the aggregate number stops being surprising. The expertise being donated is also precisely the expertise that the market prices highest — a mid-career specialist's judgment — which sharpens the contrast with the zero price tag.
Why is the labor unpaid in the first place?
Because peer review was designed as reciprocity among peers, not as employment. The system matured inside scientific societies, where the community that writes papers is the same community that reviews them: you referee others' manuscripts so that yours get refereed. Payment would create awkwardities beyond cost — paying reviewers gives journals a claim on their output, invites volume-for-pay schemes that corrode judgment, and would have to be funded, ultimately, out of the same research budgets. The norm also serves referees' self-interest in a subtle way: reviewing grants early sight of the field's findings and influence over what gets published, which is status income even when it is not cash. The unpaid norm is therefore not an oversight — it is the design — and the current strain comes from the design outgrowing the community that was supposed to sustain it.
Related stories: The Postdoc Squeeze: Years of Expertise, Priced Below the Market · Pay to Present: Who Actually Affords an Academic Conference.
What evidence shows the strain?
Several converging streams. Editor surveys and journal reports document rising difficulty finding referees, with invitation acceptance rates falling for years. Bibliometric studies show reviewing loads are unevenly concentrated: analyses of editorial databases find that a small fraction of academics supplies a large share of all reviews, while a majority contribute rarely or never. And submission volumes keep growing — global scholarly output rises annually, multiplying review demand at a rate the referee pool's growth does not match. Each strand is individually documented in the publishing-research literature; together they sketch a commons problem, in which the system's benefits are universal and its costs are carried by a shrinking minority of the willing.
How strong are the numbers?
Cautiously strong on direction, soft on precision. The Aczel estimate — around 94 million hours in the US, published in a peer-reviewed venue — rests on assumptions about articles per researcher and hours per review that honest readers should treat as an order-of-magnitude argument, and the authors said so themselves. The concentration findings come from journal-side data, which publishers rarely release comprehensively, so field-wide generalizations lean on the journals that do share. What survives any reasonable sensitivity analysis: the hours are enormous, the price is zero, the growth is real, and the distribution is skewed. Disputes concern exact digits, not the picture.
Who profits from the free labor?
The uncomfortable answer is that the labor props up a publishing market worth billions in annual revenue, dominated by a handful of large commercial publishers with margins that rival technology firms — a comparison analysts have made from public financial reports for years. University faculty generate the content and referee it without compensation, then institutions pay again for access. Open-access models changed who pays to read without changing who works: article-processing charges replaced subscriptions, and the referee remained unpaid in both worlds. This inversion is why faculty senates at several universities have formally debated whether their institutions should demand better terms from publishers, and why publishing costs have joined indirect costs in the broader fight over research finance. For individual researchers, the honest framing is that peer review is a public good they already pay for twice — once in hours and once in subscriptions — and the only open question is whether its recognition as scholarship ever arrives.
What reforms are actually running?
Payments remain rare, but the ledger is changing at the edges. Publons and similar services let researchers document reviews for credit in hiring, and some institutions now count review work in promotion files. Journals experiment with paid fast-track review, stipends for editorial boards, and AI-assisted triage that aims to save referee hours — with the obvious risks. Open peer review publishes the reports, converting hidden labor into citable contribution. And some journals pay reviewers directly, a practice still marginal but no longer hypothetical. The signal to watch is institutional recognition: when promotion committees count refereeing, the labor has acquired a price inside academia even if it never acquires a fee.
