HALF OF EINSTEIN'S PAPERS GET 1.5% OF HIS CITATIONS
How unequal is a scientific career? Klaus M. Frahm and Dima L. Shepelyansky, physicists at the Laboratoire de Physique Théorique in Toulouse (CNRS), borrowed the economists’ toolbox to find out.
Wealth, but in citations
Economists describe inequality with a Lorenz curve: line up households from poorest to richest and plot what share of the total wealth each fraction of the population holds. The Gini coefficient summarises the curve: 0 means perfect equality, 1 means one household owns everything. The paper quotes the whole world in 2021: a Gini of 0.843, with the poorer half owning 2% of wealth and the top 10% owning 75%.
The authors swap wealth for citations. For a scientist, each paper is a household and its citations are its wealth. For a university, each affiliated researcher is a household and their total citations are their wealth. The data come from Google Scholar, collected between 11 and 17 September 2026.
Twenty-five careers
They picked 25 scientists: 15 Nobel laureates, from Albert Einstein and R. Feynman to M. Devoret, physics laureate in 2025, along with leaders of network science, chaos theory, quantum computing and machine learning. For contrast, they also included a colleague from their lab and themselves.
- Einstein: 498 cited papers, 200,742 citations, Gini 0.883. The less-cited half of his papers gathers only 1.5% of his citations; the top 10% gathers 84%. His most cited paper is the one on Einstein-Podolsky-Rosen pairs, with 28,657 citations, a count the authors say has been boosted by the rise of quantum information.
- Feynman: 126 papers, Gini 0.850.
- The most unequal careers belong to Reka Albert, K. Novoselov and Y. LeCun (Ginis of 0.91 to 0.93). For LeCun, the three most cited papers alone account for about half of all his citations.
- Across the 15 Nobel laureates, the average Gini is 0.81: the bottom half of their papers collects 2.7% of citations, the top 10% collects 71% and the top 1% collects 31%. That is, the authors note, quite comparable to world wealth inequality.
The two authors and their colleague, with a few thousand to fifteen thousand citations, have lower Ginis, around 0.6. Overall, though, the paper finds no direct link between how many citations someone has and how unequal they are.
Universities as small countries
For eight universities, the authors took every researcher listing the institution on Google Scholar with at least 500 citations, to leave out students and postdocs who arrive with citations earned elsewhere.
Princeton is the most unequal, with a Gini of 0.802: half of its 2,454 researchers produce only 2.6% of its citations, and the top 10% produce 69%. Its most cited name is Daniel Kahneman, known for prospect theory, with 666,949 citations; Einstein comes ninth. Cambridge and ETH Zurich follow at 0.775, while the two Paris schools in the sample, ENS and ESPCI, and Israel’s Technion sit around 0.68. On average, a tenth of a university’s researchers produce about 61% of its citations.
Condensation, from physics to society
The authors fit all these curves with a model they previously proposed for wealth, based on the Rayleigh-Jeans distribution from the physics of waves and heat. At low total energy, most of the probability “condenses” into the lowest states: a large poor majority and a small, oligarchic top. Their model matches not only the Lorenz curves but also the long tails of the most cited papers, over several orders of magnitude, which they see as strong support.
They acknowledge that their arguments for why papers or researchers should interact this way remain heuristic, and that Google Scholar relies on affiliations declared by the researchers themselves. A more rigorous citation database, they suggest, could sharpen the picture, and they hope the same condensation idea will find uses in other corners of social physics.