Chinese firms on the US Entity List cited 72.3% more science in their patents than matched peers

The two publication figures are not the same size, and the gap between them is the informative part. Listed firms published 33.3% more papers than their matched peers in CNKI, China's largest domestic index ^[7], and 85.2% more in Web of Science ^[8]. The gap in the internationally indexed record is about 2.6 times the domestic gap ^[1]. Firms shut out of licences and joint research projects ^[2] responded most in the literature that carries frontier work and foreign coauthors.

The design is a matched comparison. The authors drew on 2010-22 data from CNIPA, Web of Science and the China Stock Market and Accounting Research Database ^[4], then paired each listed firm with unlisted firms of similar industry, ownership and location ^[5]. The matching handles the possibility that listed firms were bigger and more technical to begin with. What it cannot handle is the possibility that whatever earned a firm its place on the list also predicted its turn toward science. Kenneth Huang of the National University of Singapore says the authors cannot be sure the listing caused the shift, or whether other factors were at play ^[16].

Dinsha Mistree, a political economist at Stanford, told Nature the analysis "uses some of the best data, as well as rigorous methods for such an analysis" ^[14]. He also said he would have expected Chinese firms to publish more and to cite more science whether or not they were sanctioned, given China's investment in science and technology over those years ^[18]. On the contrast between the two groups, he said: "What the authors show so convincingly is that the Chinese companies that were subject to US sanctions pursued an even more intensive science and technology approach than they would have otherwise" ^[15].

Set the 72.3% gap ^[6] against the trend it sits inside. CNIPA grants citing at least one scientific paper went from 5,225 in 2010 to 94,441 in 2022 ^[9], a factor of 18.1 ^[2], or roughly 27% a year compounded ^[3]. At that rate, an extra 72.3% is worth about 2.3 years of the sector's own growth ^[4]. The two numbers rest on different denominators: one is a national tally of grants, the other a firm-level difference against matched peers ^[9][6]. Over the same period, grants to Chinese firms from the US patent office rose from 452 to 6,071 ^[10].

Both outcomes count documents. A patent citation records where an inventor looked; a published paper records what a firm chose to disclose. On this evidence I'd call it redirection, that restriction changed where these firms went for knowledge. Whether a firm denied a licence rebuilt what the licence supplied is a different variable, and these counts do not track it.

Cong Cao, an innovation policy researcher at the University of Nottingham Ningbo China, says the result shows the "unintended consequences" of US export controls ^[12]. Mistree put the wider point plainly: "Whether or not one supports technology controls, it is important to recognize that they can come with costs" ^[13]. The Nature account adds that US inventors and firms still lead the world in the use of scientific knowledge in patents ^[11].