Two new studies take a crack at stubborn puzzles: pay gaps in academia and the social logic of a beehive

A new study in the Proceedings of the National Academy of Sciences turns a familiar headline into a harder question. Gender pay gaps in academia are well documented across countries, disciplines and career stages. The harder question, and the one the PNAS paper tries to answer, is why they survive decades of reform.

The findings land at a moment when universities are under simultaneous pressure to defend their budgets, justify their diversity commitments, and publish the kind of evidence that survives peer review. A separate paper, from biologists at Heinrich Heine University Düsseldorf and partner institutions in Cologne and Frankfurt/Main, looks at a different kind of institution, the honeybee colony, and finds that even without a manager, work gets distributed with remarkable regularity. Read together, the two studies say something about how labour gets routed through any organisation, and who ends up rewarded for it.

What the PNAS paper actually tested

The PNAS study took the standard explanations for academic pay gaps and stress-tested them against the data. The candidates were familiar: differences in publication rates, citation counts, field of specialisation, seniority, family leave, and the slow accumulation of prestige that turns into salary over a career. Each explanation has its partisans, and each tends to survive because the others can be waved away.

The authors looked at whether the gap shrank once these variables were controlled for, and whether the patterns held across institutions. The narrower finding is that the gap does narrow when productivity and field are accounted for, but it does not close. The wider finding is more uncomfortable: a meaningful slice of the gap is not explained by the variables universities typically cite when they explain the gap away. Productivity, citation impact and seniority explain some of it. The residual is the part that does not respond to the usual levers, and it is large enough that the authors treat it as a structural feature of the institution rather than a noise term.

That conclusion is uncomfortable for two reasons. It implies that the gap is not the sum of individual decisions a dean can correct with a memo. And it implies that the variables academics themselves are most likely to invoke, output, prestige, field, are partly a screen for the variable that actually moves pay: institutional habit, and the patterns of who gets offered what when no one is watching.

Where the bee study enters the picture

The honeybee paper, led by teams at Heinrich Heine University Düsseldorf with collaborators at the universities of Cologne and Frankfurt/Main, asks a parallel question in a setting with no managers, no HR department and no salary bands. What determines which tasks a bee performs in its colony? The intuitive answer is age: older bees forage, younger bees nurse the brood. That is roughly true, and has been for about a century.

The new work traces the assignment to something more granular, the neural circuitry that makes individual bees more or less likely to pick up a particular task on a particular day. The point is not that bees have careers. The point is that even in a system with no central planner, labour allocation is shaped by something downstream of effort: the wiring that makes one worker more responsive to one stimulus than another. Translate that carefully into institutional language, and what you get is a reminder that organisations distribute work along lines that are set long before anyone fills out a form.

The bee paper does not argue that humans are bees. It does, however, undercut the comforting story that pay and assignment are simply the reward for effort and output. In a colony with no bias, no negotiation and no history, work still sorts along lines that are baked in early.

Why the gap survives every reform

The academic pay story is not new, and most readers have heard the headline version. The PNAS contribution is to treat the headline as a research question rather than a slogan, and to ask which of the standard explanations actually carry weight.

The variables universities are most willing to count, publications, citations, grants, h-indices, do some of the work. They do not do all of it. Once they are controlled for, the gap narrows but does not close. The rest sits inside the institution: in the salary a department chair offers a candidate who has just been poached from a rival; in the start-up package attached to a hire; in the slow drift of merit raises that compound over decades; in the workloads assigned to people who are not in the room when the workloads are decided.

This is the part that does not show up in a regression. It is also the part that survives every reform, because each reform is typically aimed at the variables universities can already see. What the PNAS paper is implicitly arguing, without putting it that way, is that the gap will continue to survive reform as long as reform is aimed at the symptoms the institution is already comfortable counting.

What this leaves open

The PNAS paper is a single study in a literature that has produced, by now, several generations of conflicting results. The honest reading is that the gap is real, partially explained by the variables everyone names, and partially not. The partially not is where the harder argument lives, and where the harder data will need to be collected.

The bee study is more directly falsifiable, because the underlying circuits can be mapped and the behaviour manipulated. Its relevance to the academic question is suggestive rather than proof. What it does establish is that labour allocation in any system is shaped by something that operates before the work is done, and that something is not always effort.

For universities, the practical question is whether the residual gap is treated as a research finding or as a budget line. For the rest of us, the question is whether the institutions we work in are willing to look at the parts of their own logic that the usual explanations cannot reach.

Desk note: Monexus treats both papers as primary research and reads them together rather than as competing items. The PNAS study is the load-bearing claim; the bee study is structural context for how organisations route work, not a metaphor.