Scientific Isolation? The Consequences of Trump's China Initiative on Chinese Research

GAICIL Insights

Decoupling sounds like a two-way street.¹ In science, this paper finds, the traffic flows mostly one way. Scientific Isolation? The Consequences of Trump's China Initiative on Chinese Research, by Philippe Aghion, Céline Antonin, Luc Paluskiewicz, David Strömberg, Xueping Sun, Raphaël Wargon and Karolina Westin, measures what happens when cross-border collaboration, a driver of much scientific progress, is disrupted. The 2018 China Initiative complicated procedures and reduced funding for US-China research collaborations, creating a quasi-natural experiment.

A hard question for AI security policy: before cutting a research tie, ask which side loses more. In this evidence, the side with the stronger network held the leverage, and openness was part of its strength.

Using Scopus data to track researcher productivity, the authors analyze effects on the volume, quality and direction of Chinese research. The initiative significantly lowered the quality of both publications and co-authors of Chinese researchers with prior US collaborations, compared with those with prior European collaborations, and reduced yearly citations for affected researchers by 6%. The effect was stronger for high-productivity Chinese researchers in US-dominated fields. The stars paid the most.

The mechanism is reallocation that falls short. Chinese researchers with prior US collaborations moved away from US co-authors after the shock, and that reallocation did not let them find co-authors of the same quality as their previous US partners. The loss could not be fully replaced. The authors conclude that while China is at the forefront of the scientific frontier, it still depends more on the US than the reverse. The best collaborators have few substitutes.

The context makes the result striking. China's influence across research domains has surged, solidifying its global academic prominence to the extent that Chinese researchers can no longer be overlooked. The paper gauges how far Chinese innovation relies on foreign collaboration, particularly with the United States. Prominence and independence turn out to be different things.

The application to AI is direct. Artificial intelligence is where these findings land hardest. AI research is among the most internationally collaborative fields, and US-China research ties have been particularly significant. Policies restricting collaboration in AI and related technologies are increasingly common, driven by security and competitive concerns. This paper supplies rigorous evidence on the costs such restrictions impose on researchers whose collaborations are disrupted.

The asymmetry reshapes strategic debate. In this setting, Chinese researchers depended more on US collaboration than the reverse. That informs arguments about the strategic effects of restricting scientific exchange. It also highlights collaboration as a channel through which knowledge flows and quality is maintained. Openness, in this light, is a source of leverage as much as a vulnerability.

The concentration of harm among high-productivity researchers in US-dominated fields matters for the frontier. The most productive scientists hold the most valuable collaborations, and disruption hits them hardest. In AI, where top researchers drive progress, restrictions on collaboration may have outsized effects at the cutting edge. Blunt restrictions tax the very people who move the field.

For science policy, the costs are real. Collaboration enhances research quality, and disruptions are costly to those affected. Policymakers weighing restrictions should consider these costs alongside the security concerns motivating them. Targeted measures addressing specific risks while preserving open collaboration where possible may achieve security goals at lower cost to scientific progress. A scalpel beats a wall.

The method is reusable. Publication databases tracking productivity and collaboration, combined with policy shocks as quasi-natural experiments, provide a template for evaluating science policy. As governments adopt new measures affecting research in AI and other strategic fields, similar analyses can measure effects on the volume, quality and direction of research.

The European benchmark sharpens identification. Researchers with prior European ties were not affected in the same way, letting the authors isolate the initiative's effect and strengthening confidence in the findings.

Introduction

The 2018 China Initiative by the Trump administration complicated procedures and reduced funding for U.S.-China research collaborations. Using Scopus data, we analyze its impact on Chinese research. We find that it significantly lowered the quality of both publications and co-authors of Chinese researchers with prior U.S. collaborations compared to those with prior European collaborations, and reduced yearly citations for affected researchers by 6%. The effect was stronger for high-productivity Chinese researchers in U.S.-dominated fields. While China is at the forefront of the scientific frontier, it still depends more on the U.S. than the reverse.

Takeaway

In recent years, China’s influence across various research domains has surged, solidifying its academic prominence globally to the extent that Chinese researchers can no longer be overlooked. In this paper, we provide insights into the extent to which Chinese innovation relies on foreign collaborations, particularly with the United States. To achieve this, we use the Scopus database to track the productivity of researchers. We consider the China Initiative policy as a quasi-natural experiment to analyze how this shock affected the volume, quality, and direction of Chinese research. We first show that Chinese researchers with prior U.S. collaborations reallocated away from U.S. researchers after the shock. We also see that this reallocation does not allow them to find co-authors of the same quality as their previous U.S. co-authors. We believe that this is the reason why we find a negative effect of the Initiative on the average quality of both the publications of Chinese researchers with prior U.S. collaborations, whereas U.S. researchers with prior China collaborations do not appear to be affected. Moreover, we observe that this negative effect is stronger for Chinese researchers with higher research productivity and/or who work on U.S.-dominated fields prior to the shock. The lack of reallocation towards China or the rest of the world suggests that the main beneficiary of the policy might have been Europe.¹

Our analysis can be extended in several interesting directions. One direction would be to consider other dimensions of heterogeneity among Chinese researchers, for example, the extent to which they work on research topics that meet the strategic priority of the Chinese government: our conjecture is that the negative effect of the China Initiative on the quality of subsequent publications should be less pronounced for Chinese researchers who work on topics that are considered as priorities by the Chinese government, for example digital and face recognition, biotechnologies, and energy transition. The use of textual information could also provide more detailed information as to the exact role of China and U.S.-based researchers in common research projects. Furthermore, we focused on active researchers, but we leave to future research the subject of the long-term effects of the China Initiative on researchers entering academia, in particular PhD students. A second avenue for future research would be to investigate further the role of freedom and the mobility of Chinese researchers as determinants of the quality, nature, and direction of Chinese research: in particular, can Chinese research lead to Kuhnian discoveries and become truly frontier in the absence of both freedom at home and the ability to initiate collaborations with researchers abroad? A third avenue is to bridge the gap between the Scopus information on publications and the existing patenting information (see Bergeaud and Verluise (2022)) to better predict the technological fields where China is more likely to achieve frontier.²

These and other extensions of the analysis in this paper are left for future research.

Article

Notes

Philippe Aghion et al., Scientific Isolation? The Consequences of Trump's China Initiative on Chinese Research, Int'l Econ. Rev. iere.70107 (2026), https://doi.org/10.1111/iere.70107. Takeaway source: Conclusion, reproduced from the openly licensed version (Creative Commons open-access article). ↩
Bergeaud, A., and C. Verluise. 2022. “The Rise of China’s Technological Power: The Perspective From Frontier Technologies.” Discussion Paper No. 1876, Centre for Economic Performance, London School of Economics.. ↩

Cite as: Philippe Aghion et al., Scientific Isolation? The Consequences of Trump's China Initiative on Chinese Research, Int'l Econ. Rev. iere.70107 (2026), https://doi.org/10.1111/iere.70107

Header image: 3D-принтер в работе.jpg by Angelikagub, CC BY 4.0, via Wikimedia Commons.