One in four had no import-export wording in its registered scope: here is what that can and cannot tell you.

The numbers

1. A quarter are not drafted for direct export

These are nationally recognised manufacturers, and 67 of them (25.4%) carry no import-export wording at all. Since the foreign-trade operator filing was abolished on 30 December 2022, that wording is no longer a legal wall. It remains a drafting signal. A scope without it usually describes a company drafted for the domestic market, the pattern behind the factory-makes, affiliate-invoices structure buyers keep meeting.

2. More than half have been renamed

If you check a Chinese supplier by a name from an old invoice, a catalogue or a colleague’s notes, the odds are meaningful. You may be checking a name the company no longer uses. The registered record carries former names. A storefront does not. This is why the legal-name chain starts from the current registered name and the Unified Social Credit Code, not from whatever name you met first.

3. Paid-in capital lags subscribed capital in 37% of records

262 of 264 records carried a paid-in (实缴) figure alongside the subscribed (认缴) one. In 97 of them (37.0%) the paid-in figure was below the subscribed figure: among China’s most vetted manufacturers. That is not a distress signal. Under the Company Law in force since 1 July 2024, subscribed capital must be paid in within five years. Companies are trimming and paying in inflated old subscriptions on a legal timetable. It does mean a big subscribed number on a licence was never evidence of money in the bank.

4. Wording separates roles only at the margin

Nearly every real manufacturer carries manufacturing verbs. The signal is sensitive. But nearly every one also carries trading wording, so the presence of 销售 or 批发 in a scope tells you nothing by itself. The discriminating case is the scope with trade-only wording and no manufacturing verbs at all. That is exactly how the factory-or-trader read is defined, and what the free in-browser check looks for.

5. The boring number that should not be boring

All 264 companies resolved as live registrations (存续 or 开业) under the exact Chinese names printed on the official lists: zero misses, zero revoked entries. Two working conclusions follow. An official government list is a reliable anchor for the exact registered name. When a supplier-supplied name does not resolve cleanly, that mismatch deserves attention instead of a shrug.

Method: a census of three official excerpts

• The frame. China’s Ministry of Industry and Information Technology publicises each batch of “Little Giant” (专精特新“小巨人”) companies. We used the sixth-batch announcement list as republished on the official websites of three provincial industry departments: Anhui (113 companies), Shandong (136, from gxt.shandong.gov.cn) and Yunnan (15). Government-published PDFs, company names only. The Yunnan PDF was withdrawn after we read it on 8 August 2026. That link now points at the department itself. The 15 company names we read are in our dataset.
• This is a census. It is not a sample. Every company on the three excerpts was queried: all 264. Within this frame there is no sampling error. The frame itself is the limitation, stated below.
• The lookups. Each exact list name was queried once through a licensed commercial registry-data API on 8 August 2026, from our China-side desk. We recorded scope wording flags, former-name counts, capital fields, status and establishment date. Wording counts are substring matches on the Chinese scope text (进出口; 生产/制造/加工; 批发/零售/销售).
• What we did not do. No marketplace or platform pages were scraped. No official portal was queried, proxied or automated. No CAPTCHA was touched. Person-name fields returned by the API were discarded at collection. The dataset holds no individual’s name, consistent with China’s personal-information rules and our editorial policy.

What these numbers do not say

• Not a picture of marketplace suppliers. This frame is an elite, state-vetted manufacturer list. Suppliers you meet on a marketplace are a different population. If anything, we would expect every caution here to apply more strongly there, but this study does not measure that.
• Not national. Three provinces, one coastal exporter heavyweight, two inland, out of thirty-one. The province-level spread (80.1% vs 68.1%) suggests geography matters, which is precisely why we do not project a national figure.
• Not a legal judgment. Missing import-export wording has not been a legal bar to exporting since 30 December 2022. The number describes how companies drafted their registrations, not what they may lawfully do.
• One day, one source. All lookups are dated 8 August 2026 through one licensed data channel. Registrations change; re-run before relying on any single company’s facts. Our advice is the same: read the live record, dated on the day you check it.

What a buyer does with this

1. Get the supplier’s exact registered Chinese name and Unified Social Credit Code in writing: over half of even elite companies have former names. The current registered identity is therefore the only safe anchor. Validate the code structure free, in your browser.
2. Read the scope for the trade-only pattern, not for the presence of trading words: 94% of real manufacturers carry both. The free check reads the wording for you.
3. Treat subscribed capital as drafting, not solvency, and a current, dated registry read as the only version of any of these facts that applies to your deal. That dated read is what this desk does.

All 264 records in that count were read from the GSXT, the National Enterprise Credit Information Publicity System, the China company registry every one of these facts lives in.

A second frame, measured later, and what it does and does not confirm

This study counts scope wording across 264 “Little Giant” manufacturers. On 22 August 2026 I measured something adjacent on a different group: whether a company holds an actual customs credit record. Different question, different frame, and the two are not a before-and-after.

The higher figure is a selection effect. It is not an improvement. The NHTSA group is defined by having already reached the United States, so a high customs-record rate is exactly what you would expect from it. Reading 74.6% → 91.1% as change over time would be wrong in both directions: different companies, different question, two weeks apart.

What the second frame does add is a limit on this study’s own finding. Scope wording records only how import-export activity appears in the registered business scope. It does not prove export qualifications, capacity or execution. A customs record says the company has conducted that activity under its own name. Four of those 45 hold no customs record at all and their products are on US roads regardless. So neither field, on its own, establishes that a given manufacturer will be the exporter on your declaration.

Counts for the second frame: per-company records kept on file, available on request. That group is used throughout this site along different dimensions and is described in the NHTSA manufacturer study. It is one sample read many ways. It is not a separate study per page.

Citing this study

Quote or reproduce these figures freely, including commercially. Keep the query date (8 August 2026), the frame (264 companies, three provincial excerpts of MIIT’s sixth “Little Giant” batch) and the stated limits travel with them. The numbers describe that frame on that date: nothing broader.

Archived copies, each with its own DOI, resolving independently of this site: Zenodo · Harvard Dataverse · figshare.

Machine-readable copy: the published aggregate figures as CSV: aggregate counts and shares only, one row per published measure, with the frame and query date on every row. No per-company records are published.

Currawong, “Import-export wording in Chinese manufacturer records: 1 in 4 had none”, census of three official provincial excerpts of MIIT’s sixth ‘Little Giant’ batch, registry records queried 8 August 2026. https://currawongweb.com/verify/china-manufacturer-registration-study/ Dataset: https://doi.org/10.5281/zenodo.21858946

@dataset{currawong_verify_china_manufacturer_registration_study_2026,
  author    = {Bao L. Zhou},
  title     = {{Registration facts of 264 'Little Giant' manufacturers}},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.21858946},
  url       = {https://doi.org/10.5281/zenodo.21858946}
}

If you re-run this on another batch, another province or another date and get different numbers, we want to hear it. Corrections that survive checking get published here with attribution, including ones that contradict us.

Browse all measured studies and methods in the research index.

This page reports aggregate registration statistics. It is not legal advice, and it says nothing about any individual company’s honesty, capability or suitability.