Scientific papers cited in intergovernmental organization (IGO) documents concentrate among a small group of researchers: Mapping how scientific knowledge flows from academia to IGOs

Key Findings

• Intergovernmental organizations (IGOs) use vast amounts of scientific evidence to formulate policies on issues ranging from infectious-disease control to climate change. Drawing on an integrated database of global academic papers and policy documents (230,000 papers), this study quantitatively elucidated the pathways and mechanisms by which scientific knowledge reaches IGOs.

• The authors of papers cited in IGO policy documents are concentrated among a small group of researchers, and the concentration is particularly pronounced in mature fields such as climate change, compared with evolving fields such as artificial intelligence. This concentration pattern has persisted despite IGOs’ efforts to broaden their evidence base.

• The study revealed how IGO citations come to concentrate on a particular set of scientists (Highly IGO-Cited Scientists: HIC-Sci): through dense coauthorship among a limited group of scientists and through their overlapping memberships on IGO advisory bodies.

Overview

Reports issued by intergovernmental organizations (IGOs)^*1 such as the United Nations, the World Health Organization (WHO), the World Bank, and the Intergovernmental Panel on Climate Change (IPCC) form the core of science advice that underpins national policies and international agreements. The documents produced by IGOs cite papers spanning fields as diverse as climate change, energy, artificial intelligence, and COVID-19. Yet the pathways and mechanisms by which this academic knowledge reaches IGOs and is selected for use have remained unclear. A research group led by Project Associate Professor Kimitaka Asatani, then-graduate students Yurie Iwata and Yuta Tomokiyo, and Professor Ichiro Sakata at the Graduate School of Engineering, The University of Tokyo, together with Research Fellow Basil Mahfouz at University College London and Professor Masaru Yarime at the Hong Kong University of Science and Technology, analyzed 230,737 papers cited in IGO policy documents between 2015 and 2023, and addressed this question.

The findings can be summarized as follows. First, the transfer of knowledge between academia and a wide range of IGOs is mediated by a remarkably small number of scientists, and this concentration has persisted over the long term despite IGOs’ efforts to diversify their evidence base; the pattern is especially pronounced in mature fields such as climate modeling. Second, the concentration is sustained through dense collaboration among a limited set of scientists and through their overlapping appointments to IGO advisory bodies. Third, the findings of papers by these Highly IGO-Cited Scientists (HIC-Sci)^*2 are taken up more rapidly and across a broader range of IGOs. Fourth, when examined at the country and institution level, IGO citation rates are high in Western Europe and North America but low in Asia (with exceptions such as Australia and Singapore) and in the Global South including Africa, indicating limited diversity in the knowledge that is utilized.

These findings suggest that IGOs need to continue their efforts to diversify the knowledge they draw on as evidence. Japan’s scientific community, in turn, needs to expand its participation in IGO activities and joint research in order to make a greater contribution to addressing global challenges. It is also necessary to strengthen mechanisms bridging the scientific frontier and IGOs, and to enhance international brain circulation so researchers can participate in the central communities.

The findings were published on April 22, 2026 in Proceedings of the National Academy of Sciences (PNAS).

Details

[Background]

IGOs typically do not produce scientific knowledge themselves. Their selection of which findings to draw on—from the vast volume of knowledge produced by researchers—directly shapes the direction of international policy. Prior research has examined paper-level citations and country-level geographic biases, but which researchers are selected at the individual level has not been systematically addressed. This study used large-scale data to examine whose research is taken up by IGOs.

[Methods and Results]

The research group linked the policy-document database Overton with the academic publication database Scopus^*3, classified more than 230,000 papers into 23 fields via citation-network clustering, and analyzed the author-level concentration structure overall and within each field.

By country, IGO citation rates are high in Western Europe and North America, low in Asia (with exceptions such as Australia and Singapore), and low in the Global South including Africa (Figure A). Japan’s IGO citation rate is less than half the global average. Among the top 200 institutions by number of IGO-cited papers, the only Japanese institution is The University of Tokyo (ranked 115th)—a position substantially lower than the country’s research volume and academic impact would predict.

Most of the cross-national differences in IGO citation rate can be explained by accounting for the publication journal, the IGO-cited references, and the coauthors’ prior IGO-citation history (Figure A). The author-level history is particularly striking: when even a single coauthor has 16 or more prior IGO-cited papers, the probability that a new paper is cited in an IGO document exceeds about 15%, rising to roughly 35% at 64 or more coauthor-citations (Figure B). This pattern can be interpreted as a Matthew effect, in which credibility from IGOs accumulates around specific authors.

The concentration of IGO citations on specific researchers is common to all 23 fields, but the level of concentration varies. Mature domains—such as climate change and environmental conservation—show strong concentration on a few authors, whereas emerging domains—such as COVID-19, the circular economy, and data science & AI—are more dispersed (Figure C). The smallest set of authors whose cumulative output accounts for 30% of IGO-cited papers was defined as “Highly IGO-Cited Scientists (HIC-Sci).”

Figure 1: (A) Ratio of observed to predicted IGO-cited papers by country. As predictors are added sequentially—publication journal (J), IGO-cited references (R), and coauthors’ prior IGO-cited papers (A)—cross-national disparities narrow. (B) Probability of citation in IGO documents (blue line) as a function of the maximum number of coauthors’ prior IGO-cited papers; the probability reaches about 35% at 64 or more. (C) Complementary cumulative distribution of IGO-cited papers per scientist by domain (log-log). Mature fields such as climate modeling show strong concentration on a few researchers, whereas emerging fields such as data science & AI are more dispersed.

Underlying the concentration on HIC-Sci is a network structure. In all 23 fields, central scientists preferentially coauthored with one another. In climate modeling, 31% of HIC-Sci also served as IPCC report authors, indicating substantial overlap between research networks and IGO advisory bodies. The findings of HIC-Sci-authored papers are also taken up more rapidly and across a wider range of IGOs as evidence.

The study further showed that the structure in which multiple IGOs cite the same papers—and in which large institutions such as the UN, the World Bank, and the WHO cite a paper first, with smaller specialized agencies following—reinforces the concentration on HIC-Sci.

The level of concentration was largely unchanged when the study period was divided into three sub-periods. IGOs have repeatedly pursued initiatives to broaden participation, including IPCC governance reform and WHO conflict-of-interest guidelines. Even so, the speed at which procedures change appears to differ from the speed at which citation structures change.

[Significance]

Using large-scale bibliometric data, this study revealed whose knowledge is being selected for IGO policy and clarified the mechanisms underlying the observed concentration. Concentration is not in itself a problem; in mature fields, the existence of a tightly knit core community plays an important role in synthesizing complex evidence into actionable policy advice. At the same time, this concentration raises questions about the legitimacy of advisory processes. For example, the United Nations General Assembly established an Independent International Scientific Panel on Artificial Intelligence (AI) in August 2025—but in the AI domain, the advisory body is being established before scientific consensus has formed, the reverse of what occurred in climate modeling. Procedures that make visible whose knowledge is selected, and whose is not, will become increasingly important. The author-level analytical framework presented in this study is expected to provide a foundation for accountability at the science–policy interface.

Researchers

The University of Tokyo, Graduate School of Engineering

　Kimitaka Asatani, Project Associate Professor

　Yurie Iwata, Graduate Student (at the time of research)

　Yuta Tomokiyo, Graduate Student (at the time of research)

　Ichiro Sakata, Professor

University College London

　Basil Mahfouz, Research Fellow

The Hong Kong University of Science and Technology

　Masaru Yarime, Professor

Publication Details

Journal: Proceedings of the National Academy of Sciences (PNAS)

Title: Structure of scientific knowledge flows to intergovernmental organizations

Authors: Kimitaka Asatani, Yurie Iwata, Yuta Tomokiyo, Basil Mahfouz, Masaru Yarime, Ichiro Sakata

DOI: 10.1073/pnas.2514861123

URL: https://doi.org/10.1073/pnas.2514861123

Funding

This research was supported by the Japan Society for the Promotion of Science (JSPS) Grants-in-Aid for Scientific Research (Grant Number: JP21K19817). Some of the coauthors received research funding from Elsevier.

Notes

*1 Intergovernmental Organizations (IGOs)

International institutions established by treaty among multiple states. Examples include the IPCC, WHO, World Bank, UNESCO, and UNEP. Through assessment reports and guidelines based on scientific evidence, IGOs provide guidance that underpins national policies and international agreements.

*2 HIC-Sci (Highly IGO-Cited Scientists)

A concept introduced in this study. For each research field, HIC-Sci is the smallest set of authors—ranked in descending order by the number of their papers cited in IGO documents—whose cumulative output accounts for 30% of all IGO-cited papers in that field. They can be understood as a small core group of researchers whose work is particularly likely to be incorporated into IGO documents.

*3 Overton / Scopus

Overton is a policy-bibliometrics database that comprehensively records policy documents and the academic papers they cite. Scopus is an academic database known for its bibliographic coverage and high-precision author identification. This study linked the two using DOIs.