On this page

Outlet assessment

Stanford Center on China's Economy and Institutions (SCCEI)

85/ 100

Reliability

Assessment confidence
medium
Check
full assessment · provisional
Assessed
2026-11-15
Reevaluation
due after 14 May 2027 (180 days; only if the outlet is still in use. How)
Kind
think-tank, US
Leaning
market-oriented, pro-engagement

How scores are set: outlet assessment.

Kind: think-tank. SCCEI is a university research center at Stanford (USA), founded 2021 as a joint effort of the Freeman Spogli Institute (FSI) and the Stanford Institute for Economic Policy Research (SIEPR). It publishes two-page China Briefs that summarize recent papers (about 90 issues by 2025, also on Substack) and publication pages that reproduce abstracts of its affiliates’ journal articles and working papers. It runs REAP, CLIP, PEG and eSET programs, the Big Data China partnership with CSIS and a partnership with the Asia Society Policy Institute.

Summary#

  • Reliability overall is 85 (70-89 band). 23 of 24 checkable first statements were accurate or mostly accurate: rate 95.8%, Wilson 95% interval 79.8-99.3 (n = 24; 19 accurate, 4 mostly accurate, 1 misleading, 0 false).
  • Strongest: publication pages (12 of 12). They copy the published abstract verbatim, so the check is fidelity, not whether the research holds.
  • Weakest: China Briefs (11 of 12). One brief misstated a PNAS paper’s corpus and shares. Others carried small slips: an unsupported “34% peak”, a figure absent from the cited version, misspelled author names.
  • The score is below the sample rate for three reasons: the briefs’ own statements err more than the copied abstracts, no corrections policy was found, and one sampled paper’s data are publicly disputed.
  • Orientation: market-oriented, skeptical of industrial policy, pro-engagement on US-China ties.
  • Blind second grading (2026-11-15): 6 items regraded by a separate grader from masked packets; 5 of 6 agreed (83.3%). The one disagreement (the $912B government VC fund total) went to a third blind look, which graded it mostly accurate (minor); the majority grade stands. Agreement is above 80%, so confidence is not lowered.
  • Assessment confidence is medium. The sample met its stop rule (24 checkable items), but the wording comparisons are headline-level (6 of 8 events), which limits what they show. Blocked platforms are queued for retry and are not counted (see Pending checks).

Ownership#

  • Chain: SCCEI → Stanford University (FSI and SIEPR) → Board of Trustees. Stanford is a private nonprofit without shareholders. Gifts go to “Stanford University” with an SCCEI designation (giving page).
  • Funding FY2024-25: endowment payout $2,124,610 and expendable gifts $1,976,028, total $4,100,638 (annual report). Donors are not named.
  • Leadership: co-directors Hongbin Li and Scott Rozelle. Li’s page says he was founding co-director “from 2021 to 2026”, so leadership changed in 2026; no successor was found. A faculty steering committee also serves: Li, Miller, Pan, Rozelle, Zhou.
  • Stanford University has no group profile and was not staged. No evidence was found that the university directs SCCEI content; the center’s news items are cross-posted on fsi.stanford.edu. Editorial independence is unknown.
  • University-level ties: Section 117 disclosures reported by the Stanford Review show Chinese state-linked gifts to Stanford, for example at least $3 million in 2025 from “Chen Yuan” for Hoover Institution research. The report does not name SCCEI.
  • REAP has long worked with Chinese local governments on rural education and health trials. Its outputs are therefore in part evaluations of programs it helped design.

Accuracy#

  • Research findings: 10 of 10, all verbatim abstracts checked against the journal (Science, Child Development, World Development, J. Law & Economics and others).
  • Statistics: 11 of 12. The one misleading item is “Speaking with One Voice”:
    • The brief claimed “11 million articles in 700 … newspapers” with scripting “grown from 6% to 8% … and 4% to 6%”.
    • The paper has “10 million articles from 46 local newspapers” and commercial papers at “typically 2 to 4%” against “6 to 8% in party papers”, which are levels, not growth.
    • The 700 figure comes from a separate COVID-19 corpus of 708 papers.
  • Mostly accurate (4):
    • The school feeding program is dated 2012; the State Council launched it in autumn 2011.
    • “Overbuilt” gives a 34% 2015 peak that is not in the cited version of the paper (table peak 32.9%, 2018).
    • The gaokao is called “three-day”; it takes 2-4 days by province.
    • The fertility brief’s 744,000 births figure is not in the paper version that could be read.
  • Corrections: briefs show “Updated <date>” without change notes; no corrections policy page exists.
  • References: 3 grades rest only on unprofiled sources (Wikipedia and a Sohu repost of MoE data). The full texts on SSRN and OUP were blocked, so the number of officials in the Zeren Li study and the published Rogoff-Yang tables could not be checked.

Outside record#

  • The sampled Science paper on solar policy and birds (co-author Shanjun Li, SCCEI eSET director) is disputed:
    • The China Ornithological Society found “flaws” in the crowdsourced bird data, for example a retained record of “over 2 billion little egrets at a single location” (Sixth Tone, 2026-08-28; SCMP, 2026-09-06).
    • China Daily called the “kill birds” claims a “disguised attack on solar power industry”.
    • No retraction or correction was found. SCCEI’s page repeats the abstract without noting the dispute.
  • OpenAlex lists no retracted works for Scott Rozelle (1,276 works). All 24 sampled source papers are listed as not retracted.
  • No media-rating group rating of SCCEI was found.

Orientation#

  • Economic policy (moderate, medium confidence). Brief titles credit markets and doubt state direction:
    • “Industrial Policy Had Little to Do with It” (drug innovation).
    • “China’s Financial System Boosts Production but Bypasses Demand”.
    • “Behind China’s Local Protectionism: Political Rivalry”.
    • Subsidies are framed as misallocation (“Subsidies for Sale”).
  • Chinese government (mild, medium confidence). Story selection often covers governance problems:
    • Problem topics: scripted propaganda, the revolving door, rising public pessimism, the neglected rural poor.
    • Chinese strength is also credited: “China’s Innovative Capacity Is Underestimated”, “Global Drug Innovation Powerhouse”.
    • Official claims are weighed against data, not repeated.
  • US-China relations (mild, medium confidence). The briefs stress costs of rivalry for the US:
    • “The China Penalty Is Quietly Reshaping American Science”.
    • Chinese students “expanded graduate programs … for Americans”.
    • “U.S.-China Economic Interdependence Has Shifted, Not Disappeared”.
    • Conference speakers range from engagement voices (US-China Business Council) to competition voices (Rush Doshi, Elizabeth Economy).

Wording#

  • Brief titles are more causal and assertive than the source paper titles. Examples:
    • “China’s Fertility Decline Is Driven by What Women Want, Not What the Government Permits”; paper: “Desired Fertility, Actual Fertility and the Effects of China’s Universal Two-Child Policy”.
    • “China Is Now a Global Drug Innovation Powerhouse”; paper: “From Free Rider to Innovator”.
  • On the solar study, SCCEI used the journal title “Reduces Bird Diversity”. Other outlets wrote:
    • Phys.org: “linked to”.
    • Scientific American: “could have unintended ecological harms”.
    • Sixth Tone: “Draws Scrutiny”.
    • China Daily: “disguised attack”.
  • On China’s clinical-trial rise, SCCEI wrote “powerhouse” and placed the cause in insurance reform. US trade outlets framed the same rise as “security risk” (BioSpace) and “China threat” (The Hill).
  • On scripted news, SCCEI says “scripted propaganda” and “tighter central control”; the author’s university wrote “propaganda masquerading as news”. The vocabulary is close to the source paper’s.

Perception#

  • Hacker News (6 posts): REAP and Rozelle are cited as evidence-based and trusted, for example “evidence-based approach … thanks to guidance from Stanford’s REAP group” and “not a China hater”. One post notes “Rozelle’s group at Stanford has consulted at the prefecture level with the CCP”. No post discusses the China Briefs.
  • Not readable: Reddit and X (blocked on this network), YouTube comments, review sites and forums, media-rating groups.

By audience (platform profiles, 2026-11-15): center left audiences, 6 posts (Hacker News 6): 5 trusts, 1 mixed; mean trust 91.7, gap +6.7 points against measured reliability 85. All posts come from left-of-center audiences (one audience side), so perception is capped at medium and no cross-audience reading is possible yet; the next top-up samples center and right-leaning venues (Free Republic, GETTR, patriots.win). Platform pages: Hacker News.

How it is weighed#

  • Trust more: publication pages, as faithful copies of peer-reviewed abstracts. Weigh them as the underlying paper, not as an SCCEI judgment.
  • Trust less: exact figures in China Briefs. Recheck them against the source paper before citing; one in twelve misstated a source.
  • Likely bias:
    • Skeptical framing of Chinese industrial policy and state intervention.
    • Engagement-leaning framing of US-China science and education ties.
    • REAP evaluations of its own interventions in rural China.
  • Counting: a China Brief and the paper it summarizes are one source, not two. FSI and SCCEI copies of the same item are one source.

Used in posts#