The phrase “China brain drain” encourages a simple story: talented people leave, the country loses, and stronger efforts to keep them must be the answer. The available evidence is more complicated. The principle should be simpler: a country can compete for a person’s work. It does not own that person’s future.
My argument is not that China has already lost a measurable, one-way contest for talent. It is that treating international mobility as a loyalty test is a poor way to build an environment in which talented people choose to work.
The data do not support a one-way exodus story.
China’s Ministry of Education reported that 535,600 people returned from overseas study in 2025, compared with 495,000 in 2024. It also reported 570,600 departures for overseas study in 2025.[1] Those figures concern different cohorts and stages of education. Subtracting one from the other would not produce a valid estimate of net brain drain.
Returnees are not all scientists. Departing students are not all permanent emigrants. A move overseas can lead to a return, a collaboration or a longer stay. The categories do not become interchangeable merely because a dramatic headline needs a single number.
A different dataset shows the strength of U.S. retention in a specific group. The U.S. National Center for Science and Engineering Statistics estimated that 83 percent of science and engineering doctorate recipients from China who had held temporary visas and graduated from U.S. institutions in 2017–2019 were residing in the United States in 2023. Its China category includes Hong Kong.[2]
That is not 83 percent of all Chinese graduates, not a permanent-emigration rate and not a measurement of why those individuals stayed. It does show why aggregate returnee counts cannot answer every question about highly specialized talent.
America can repel talent too.
Any account in which the United States automatically represents freedom and China automatically represents flight is too convenient to be credible.
A 2023 PNAS study surveyed 1,304 scientists of Chinese descent employed in tenured or tenure-track positions at U.S. universities. In the survey, conducted from December 2021 to March 2022, 61 percent of respondents had considered leaving the United States. The authors’ accompanying explanation identifies convenience sampling and possible selection and response biases. Thinking about leaving is not the same as leaving, and the result is not a population-wide departure rate.[3]
The implication I draw is narrower, and more useful, than a declaration that either country has won. Talent policy can be undermined by distrust in the receiving country as well as by problems in the country of origin. A scientist choosing a laboratory is not necessarily voting for an entire political system.
A recruitment bonus cannot answer every institutional question.
Money, equipment and ambitious research programs matter. So do the terms under which people can use them. Can researchers challenge an influential colleague’s conclusion? Can a grant decision be explained? Can they collaborate across borders under clear rules? Can they change employers without being treated as disloyal?
These are the standards I propose for evaluating talent policy. The migration statistics above do not establish the answers for every institution in China or the United States, and they cannot quantify the causal effect of academic freedom on each person’s decision.
But that uncertainty is not a reason to substitute patriotic pressure for institutional reform. If the goal is voluntary retention, policymakers should improve the offer. A demand for gratitude is not an improvement to the offer.
The objection that public investment creates obligations deserves an answer. A defined scholarship or employment contract may contain legitimate commitments. Those commitments should be transparent, specific and proportionate. They should not quietly expand into an indefinite moral claim over where a person may build a career.
The state has an interest in the benefits of education. The person who spent years acquiring that education has an interest in a life.
Measure the quality of staying, not just the act of returning.
A serious talent strategy should ask what returnees can do after arrival: whether they remain productively employed, build teams, pursue credible research and maintain useful international relationships. A recruitment ceremony is an input. A functioning research environment is an outcome.
The same scrutiny should apply to U.S. institutions. Attracting international students and then leaving them uncertain about whether they belong is not a durable competitive advantage. Security rules should identify specific risks and conduct rather than turning ancestry or ordinary scholarly contact into a substitute for evidence.
This is not a demand to abandon safeguards. It is a demand to make safeguards intelligible and bounded. Researchers should know which actions are prohibited, how decisions are reviewed and what avenues exist to challenge an error.
China should want people to return because the work is compelling and the terms are credible. The United States should want people to stay without requiring them to sever every meaningful connection to their origins. Neither ambition is strengthened by treating a human being as a national asset that must be possessed.
A country that wants loyalty should first offer reasons for trust. Talent retention is not a headcount to be secured at any cost. The strongest result is not that people cannot leave. It is that they are free to leave and still choose to build something there.
nn
Sources
n
- Ministry of Education of China, 2025年我国留学回国人数达53.56万 [Overseas-study returnees reached 535,600 in 2025], 11 April 2026. Official report; returnees and departures are not a matched cohort or net migration measure.
- National Center for Science and Engineering Statistics, Most U.S.-Trained Science and Engineering Doctorate Recipients on Temporary Visas Remain in the United States, NSF 25-325, 24 February 2025. 2017–2019 degree cohorts, residence in 2023; Table 1 notes that China includes Hong Kong.
- Yu Xie and coauthors, Caught in the crossfire: Fears of Chinese–American scientists, PNAS 120(27), e2216248120, 2023. Coauthor Junming Huang’s study explanation, survey dates, findings and caveats; bibliographic record and abstract. DOI: 10.1073/pnas.2216248120. Percentages refer to respondents, not all Chinese-descent scientists.