Sinosure: Buyer Credit Limits, the Cost of Payment Terms and Subrogation on Default

Concept

Sinosure is China's state export credit insurer, and its place in a transaction is described wrongly almost everywhere. The policyholder and the claimant is the supplier, but the credit assessment is run on the buyer: the insurer approves a limit on a named importer (买方信用限额), and only inside that limit does the factory ship on open account. That gives the direct answer to the question this topic turns on — payment terms are not won by persuasion or by order size, they are won because a Sinosure limit is open on your company. The limit is an asset the importer manages although it is held in someone else's name, and a profile built once works across every purchase in China.

The asset has a second side that appears neither in sales material nor in most English-language coverage. On non-payment the supplier files a loss, Sinosure indemnifies it up to 90% and by subrogation takes over collection from the buyer — now as a Chinese state institution, with a database of 490 million enterprises and banks (SINOSURE Profile) and a network of overseas agents. The counterparty across the table changes within two to three weeks of the missed date, and the new one has no authority to discount principal.

What follows is the insurer's own domain. Documentary instruments and renminbi settlement sit in trade finance; the price of money and credit arbitrage in corporate lending.

Sinosure insures the Chinese supplier, not you; your gain is deferred payment, your risk is a subrogated claim from a state insurer. Both sides of the trade hang on one object: the credit limit approved on your company.

The frame of the trade in parameters:

The institution: who ends up as your creditor

Sinosure is the sole authorised operator of China's policy export credit insurance, incorporated on 18 December 2001 as a state-owned insurance company with its own legal personality.

The sharpest self-description sits not in the charter but in the preambles to the investment policies: a "state-funded policy insurance company holding independent legal personality, designated as the implementing institution of policy export credit insurance", with "the Ministry of Finance of the PRC, in its capacity as a shareholder of the insurer, acting as the department in charge of policy export credit insurance" (海外投资股权保险, 2025 revision, clause 1.1, PDF). Prudential supervision sits elsewhere: it is exercised by the NFRA under a dedicated instrument (NFRA supervisory measures on Sinosure).

Insured volume across all lines reached $1,021.4 billion in 2024, up 10% year on year, across 227,000 clients (SINOSURE); in the first half of 2025 the insurer covered 27.4% of Chinese exports (Sina Finance, 23 March 2026); on Berne Union data its insured amount has ranked first among export credit agencies since 2015 (Trade Finance Global). Practically every fourth shipment out of China moves with a state insurer standing behind the seller.

Origins and ownership

Sinosure was assembled out of two previously separated state functions — the export credit insurance department of PICC and the corresponding unit of China Exim Bank — and its Berne Union membership is three years older than the company itself.

Before 2001. Export credit insurance began in China in 1988, when the People's Insurance Company of China (PICC, 中国人民保险公司) set up an export credit insurance department to support exports of machinery and electronic products. China Exim Bank was founded in 1994, and the function split: the bank ran guarantee operations and lending for machinery exports, PICC developed medium and long-term insurance products. Chinese industry literature calls that period "dual management" (王稳, 周依晴, 《中国金融》, news.qq.com, 30 October 2024). The institutional continuity is documented: Berne Union membership is dated 1998, which means Sinosure inherited it from its predecessor (Berne Union, member profile).

Incorporation. The State Council decision to form the company was taken in May 2001 (the instrument 《国务院关于组建中国出口信用保险公司的通知》). Operations began on 18 December 2001; the legal form is 国家出资设立的政策性保险公司, a wholly state-funded policy insurance company with standalone legal personality (Belt and Road Portal). Start-up capital was 4 billion yuan from the budget (Sina Finance, 16 June 2026).

Huijin discloses the percentages itself. On its holdings page (generation stamp — 16 April 2026) the line 中国出口信用保险公司 carries registered capital of 461.00 亿元 (46.1 billion yuan), a Huijin contribution of 340.00 亿元 (34.0 billion) and a published participation ratio 出资比例 of 73.63% (huijin-inv.cn, Shareholding). The arithmetic of the amounts on that same page gives a different number: 34.0 / 46.1 = 73.75%; against the 46.1254 billion shown by the enterprise register, 73.71%. The source does not explain the gap. One observation helps to read it, though it is a reconstruction rather than a statement by Huijin: 73.63% is exactly the pre-recapitalisation share (20.0 of 27.1611 billion), so the percentage appears not to have been recomputed when the amounts were updated. What follows uses the published 73.63%; the computation from the disclosed amounts gives 73.7–73.8%, and either figure supports the same conclusion — control sits with Huijin, not with the Ministry of Finance.

What follows from this. The standard formula in English-language reference material — "a state-owned enterprise under the Ministry of Finance" — is wrong on control: since 2012 the controlling shareholder has been Central Huijin, a vehicle of the sovereign fund CIC, and the Ministry of Finance holds a blocking minority. There is indirect confirmation: when stakes in three distressed-asset management companies were transferred from the Ministry of Finance to Huijin in February 2025, Sinosure was not on the list — there was nothing to transfer, control having sat with Huijin for thirteen years already.

The intermediate step remains open. If capital was 4 billion yuan in 2001 and the total after Huijin's 20 billion injection was 27.161 billion, then the Ministry of Finance stake stood at 7.161 billion by 2012 — meaning the budget contribution nearly doubled between 2001 and 2012. No available source describes or dates that step.

The Ministry of Finance plays three roles at once: minority shareholder; the department administering the state export credit insurance risk reserve fund (国家出口信用保险风险准备金); and the authority that admits other players to the market. The 2013 pilot allowing PICC P&C into short-term cover was approved by the Ministry of Finance, not by the insurance regulator (PICC, 27 April 2013). The fiscal wiring is written into the rules: article 64 of the NFRA measures exempts policy business whose ultimate risk is carried by the state budget from contributions to the Insurance Security Fund (full text of 金规〔2024〕21号).

The practical consequence of the ownership structure: on policy business the insured has no protection from the industry Insurance Security Fund — a direct budget backstop stands in its place. For a counterparty this means Sinosure's negotiating position is set by its policy mandate rather than by its solvency: capacity does not run out, and under article 57 of the NFRA measures the policy-function targets carry more weight in the KPI system than commercial ones.

Timeline: four phases

Sinosure's history divides into four distinguishable phases — build-out under monopoly, counter-cyclical expansion with partial demonopolisation, Belt and Road expansion, and the current period of regulatory formalisation.

2001–2008. Compound growth above 40% a year: from $27.5 billion in 2002 to more than $500 billion by November 2018 (Wuhan Financial Bureau, 17 December 2019). The crisis year of 2008 produced the first visible loss experience: claims tripled, rising 150% on American buyers and 206% on European ones (People's Daily / 中国经济周刊, 30 March 2009). The phase ended in a criminal case: on 18 April 2008 the general manager Tang Ruoxin (唐若昕), who had run the company since incorporation, was removed for "unlawful and non-compliant conduct" (Sina Finance, 24 April 2008); in February 2010 he was sentenced to 14 years for bribery and extortion of more than 3 million yuan, and the appeal was dismissed by the Beijing High Court in September 2010 (Chinanews, 12 February 2010).

2009–2013. The line "China demonopolised export credit insurance in 2013" misleads without two qualifications. First, admission was authorised by the Ministry of Finance, not by the insurance regulator CIRC — the question was settled through the fiscal channel rather than through insurance law, because what was at stake was access to a budget-supported activity. Second, only the short-term segment was liberalised; Sinosure retained exclusivity over medium and long-term export credit insurance and over overseas investment insurance (Huize, 25 August 2017) — precisely the lines where the risk is sovereign and political and the backstop is the treasury. CPIC, Ping An P&C and CITIC Pacific Insurance later obtained the same short-term access.

2013–2020. Export credit insurance has appeared in the Government Work Report for ten consecutive years since 2015; the 2015 wording was "expand the scale of export credit insurance, achieve full cover for exports of large complete sets of equipment" (Wuhan Financial Bureau, 13 March 2024). Its role in infrastructure lending is usually put as "Sinosure's involvement is most often what gives a project the green light": cover on medium and long-term products reaches 95% of project value, so the insurer carries almost the whole default risk (Belt and Road Initiative portal; IISD). On an ODI estimate built on AidData, ≈$230 billion — at least 12% of all Chinese overseas lending in 2000–2021 — was made possible by Sinosure cover (ODI, Chen & Herscowitz, 4 April 2024); that is a research-centre calculation, not a company disclosure.

2021–2026. The 14th Five-Year Plan totals: $4.8 trillion insured cumulatively (+65% on the preceding plan), more than $12 billion paid out to enterprises, over 320,000 clients, with micro-enterprise clients rising from 111,000 in 2020 to 217,000 in 2025 and $208.6 billion insured for them in 2025 (人民日报海外版 via 163.com, 26 June 2026). The recapitalisation raised registered capital by 69.8%, to 46.1254 billion yuan, but the exact date of the decision is not disclosed: the 16 June 2026 piece cites "the latest enterprise-register data" and gives no date of change (Sina Finance / 机构之家, 16 June 2026), while Central Huijin's holdings register, generated 16 April 2026, already shows the new capital — so capital was raised no later than mid-April 2026.

Organisation: network, leadership, subsidiaries

Sinosure is built as a single-tier state company with a domestic branch network, seven overseas representative offices and its own credit information infrastructure, rather than as a group.

Network. 26 branches inside China and one subsidiary; overseas representative offices in the United Kingdom, the UAE, South Africa, Russia, Brazil, Indonesia and Egypt (2025 recruitment notice). The first overseas office was London, in 2004. There is no reliable public headcount: the figure of 576 that circulates in industry directories is incompatible with a 26-branch network and probably refers to a single unit.

Sources for the sequence: Tencent News / 《国际金融报》, 3 December 2024 and Sina Finance, 12 February 2025. The president (总经理) since 2025 is Sheng Hetai (盛和泰) (SINOSURE, annual work conference of 21–22 January 2025).

The company's own English-language site now shows the new leadership. Its Senior Management page lists Wang Hao as chairman and Sheng Hetai as vice chairman and president (SINOSURE, Senior Management), but gives no dates of appointment, so a change of executives is dated by the NFRA approvals of executive qualifications, not by the "About us" pages.

Subsidiary and research bodies. 中国信保资信有限公司 (SinoRating) was established in Beijing on 18 May 2023 as a wholly owned subsidiary — the first company approved by the State Council to handle foreign-trade business credit data (The Paper, 18 May 2023). It is its identifier and its credit report that appear in the buyer's file on a limit application. The Country and Regional Research Institute (国别研究所) publishes an annual Country Risk Analysis Report; the 2025 edition was the 21st, published on 20 October 2025 in two volumes covering 40 key countries and global investment, sector and insolvency risk (Wuhan Financial Bureau, 20 October 2025). The country risk scale described in the report itself runs to nine grades (1 lowest, 9 highest) and covers more than 190 states (Chinanews / 每日经济新闻, 15 December 2010).

The information base. 490 million enterprises and banks as at June 2025, against 320 million at the end of 2021, drawn from over 400 collection channels; client-facing products are the 信步天下 app and the 全球查, 资信红绿灯 and 资信导航 services (Sinosure Profile). This is the pool an importer lands in, and the reason a default on one transaction closes limits at every Chinese supplier at once.

The legal nature of the policy

Export credit insurance is not governed by the PRC Insurance Law at all: its provisions apply to a Sinosure policy only by reference (参照适用), and only where the policy itself is silent.

A check against the full text of the Insurance Law (2015 revision) returns nothing: the terms 出口信用保险 and 政策性保险 do not appear once; article 184 refers out to marine insurance, article 186 to agricultural and compulsory insurance (text of the Law). The gap is closed by a judicial rather than a legislative instrument — the Supreme People's Court reply 法释〔2013〕13号, adopted on 15 April 2013 at the 1575th session of the Judicial Committee in answer to a request from the Guangdong High Court 粤高法〔2012〕442号, published on 2 May 2013 and in force from 8 May 2013 (SPC Gazette). The formula: given the special nature of export credit insurance, the people's courts may apply the relevant provisions of the Insurance Law by reference (参照适用); where the export credit insurance contract provides otherwise, the contract governs.

The mechanism is finer than "insurance law applies". 参照适用 is neither direct application nor analogy in the continental sense: the Insurance Law is converted from a mandatory statute into a default reserve that fills gaps but yields to any policy term. The price of that is measurable in deadlines.

What that means for the parties. For the supplier, the procedural protections a policyholder expects from a commercial policy are absent. For the buyer, the assumption that "the factory will be paid out and go away" is unreliable: loss verification can take four months, and collection runs in parallel with it rather than waiting for payment. The SPC reply is tied to the type of contract, not to the identity of the insurer: a commercial insurer selling export credit insurance can invoke it too, and the courts allow that — in (2018)浙0602民初12543号 and (2021)川0104民初4444号 the defendant was PICC and the same analysis was applied.

Regulatory evolution

The Sinosure regime was not built out of insurance legislation but out of discrete instruments: a 2013 judicial reply, a dedicated 2024 NFRA supervisory measure and annual policy circulars from the Ministry of Commerce.

What matters in the 2025 supervisory measures. Articles 3 and 6 fix the status of a 政策性金融机构 and the requirement to "adhere to the functional positioning of a policy financial institution". Article 50 sets the prudential floors: core solvency of at least 50%, comprehensive solvency of at least 100%, and an integrated risk rating no lower than class B. Article 53 directs capital to be allocated in priority to the policy function rather than to return. Article 57 requires the weight of policy-function delivery, risk management and compliance in the KPI system to be higher than that of other indicators. Article 59 introduces deferred remuneration for senior management over at least three years with clawback. Article 31 obliges the company to build a country risk assessment system proportionate to its country exposure and to reserve fully against it — this is the regulatory anchor of the non-public country policy that makes a refused limit look unreasoned. Article 63 permits cross-provincial business and umbrella policies for corporate groups.

The product range: whose loss gets paid

Sinosure's products differ less in the risks they list than in who receives the money — the exporter, the lending bank or the investor.

Indemnity percentages come from Sinosure's own product pages: short-term, medium and long-term, investment.

Short-term comprehensive export credit insurance (短期出口信用保险) is the policy your shipment moves under. One asymmetry matters: if the buyer, or the issuing bank under a documentary credit, is registered in China, political risks are not covered at all. Excluded are exchange-rate losses, the exporter's own breach, exports to affiliated companies, documentary discrepancies under letters of credit, and continued shipment after the exporter knew of the risk (wording 4.0, PDF). Four provisions reach the buyer directly. Where part of the export is deliberately left undeclared, the insurer is entitled to cut the indemnity percentage or decline liability across every shipment under the policy (article 6). Article 20 bars parallel cover of the same shipments with Coface or Allianz Trade without the insurer's written consent. Article 24 imposes confidentiality on buyer information — which is why the factory will not tell you what limit you carry. Article 22 requires the insurer's prior written consent to any change in the payment date, and moving a payment "by email" entitles the insurer to cut the indemnity or decline it.

The "provincial editions" of the wording are a registration artefact, not a substantive variation. A byte-level comparison of the published 4.0 texts shows the editions for different provinces differ by exactly one line — the province name in the title; the product specifications differ only in the heading and the sale start date (19 March 2026, 30 March 2026, 31 March 2026). The regional subdomains of sinosure.com.cn are full mirrors of the corporate disclosure section, not repositories of local documents. The real customisation goes into two unpublished documents: 《保险单明细表》 (the policy schedule) and 《批单》 (the endorsement). Looking for differences between provinces is pointless; the question to ask is about the schedule.

What changed between 3.0 and 4.0. Three items were deleted from the exclusion list, among them "unauthorised release of documents by a bank, or of cargo by a forwarder or carrier" — meaning the classic telex-release-without-payment scenario moved from uncovered to covered. Export of services was introduced (3.0 knew only goods), and refusal to confirm a completed service was added to the risk list. A definition of "trade dispute" appeared for the first time. And the central contested clause — the trade dispute condition precedent — was softened (analysed below). Wording 3.0 was withdrawn from sale on 31 August 2026 but continues to apply to policies already in force.

The 2026 generation. 普惠定额保险 and 普惠随心保险, published 14 April 2026, are sold only through the self-service portal, with automated underwriting and no buyer credit limit (PDF); the claim filing window is two years instead of four months, assessment takes 60 days, payment 10 days; collection costs sit with the insurer in full (clause 6.7.3), while court, arbitration and legal fees are advanced by the exporter itself (clause 6.5.2) and reimbursed only where it wins and the loss is covered. Payment terms here are possible without a credit check on the buyer — but the sanctions clause is broad.

Cover of costs, not of debt. 出口前保险 rev. 2026 (11 December 2025) reimburses production cost where the contract becomes impossible to perform, and liability starts from the date the trade contract takes effect (PDF). 特定合同保险 rev. 2025 covers both costs and receivable and adds the buyer's breach of contract and termination for convenience (PDF) — a cancelled order after production has started becomes an insured event.

Buyer's credit insurance explains the cheapness of Chinese project finance: the insured is the bank, principal and interest are covered with penalty interest excluded, and there are only two exclusions — the bank's own fault and risks outside the schedule. There is no trade-dispute clause and no documentary-discrepancy clause, and the waiting period is 90 days (出口买方信贷保险单, 2022 rev, PDF). Supplier's credit insurance (March 2018) covers the exporter or contractor, separating advance from deferred payments, with a 90% maximum (PDF).

Investment, domestic trade, guarantees. 海外投资保险 rev. 2025 (27 August 2025) exists in an equity form and two debt forms; the breach-of-undertaking risk requires a final and binding court or arbitral decision for a specified amount, and clause 7.1.2 fixes the premium rate for the life of the policy regardless of any deterioration in the host country's risk (PDF). It also carries the only express ESG clause in the range: clause 4.2.6 withdraws cover where the host country's environmental or anti-corruption legislation is breached. 国内贸易信用保险 rev. 2024 covers domestic Chinese receivables up to one year, but only against insolvency and protracted default.

The buyer credit limit

A credit limit is Sinosure's decision on the maximum single exposure to a named buyer, outside which no cover exists.

The rev. 4.0 rule is hard: with no limit in force on the shipment date, the insurer does not answer for that export at all — not "in part". The wording of article 4(2)(6) is unconditional: it is not a right to reduce the ratio, it is an absence of liability. This is the source of the familiar scene in which a factory ready to ship on credit yesterday demands prepayment today.

Procedure. The supplier files the application, separately for each buyer. Sinosure runs an investigation on its database of 490 million enterprises and banks, its credit bureau subsidiary SinoRating and 400-plus information channels. The outcome is recorded in a Credit Limit Approval Notice (信用限额审批单) — and here sits the detail missing from every survey: the indemnity percentage set at limit approval prevails over the percentage in the policy schedule. The published "90%" is a ceiling; the operative rate is set buyer by buyer, and only for shipments made after the limit takes effect.

From the buyer's side the steps are: move to open account under cover; register with the SinoRating bureau and obtain an identifier with a credit report; give the factory constitutional documents, two years of accounts and a payment history — the point at which audit inside the Chinese perimeter stops being a formality. Sinosure publishes no buyer-eligibility thresholds; the brokers who file these applications name the same benchmarks — turnover above $1 million in the last financial year, at least one completed financial year of trading, profitability and no arrears to suppliers; a credit investigation of up to 21 days; a typical opening limit of $100,000 (Axton Global, Sinosure FAQ).

The limit revolves — it frees up again as the receivable is repaid, so it expresses single exposure rather than annual turnover. Four things cancel it: a long dormant period with no declarations; the filing of a Notice of Possible Loss on that buyer, automatically; a material change in risk; and adjustment to actual utilisation.

The hidden lever. Approving a limit is not a purely credit decision: provincial agreements set a target "limit satisfaction ratio" (限额满足率), and the notice issued by the Tibet AR Department of Commerce and the Sichuan branch on 16 June 2026 (published 22 June 2026) requires it to be held at no less than 85% (swt.xizang.gov.cn), while MOFCOM's notice of 27 March 2026 speaks of "optimising the supply of limit resources". The policy provides no formal appeal. The working route after a refusal is different: refile with an expanded file and a SinoRating report, and on a systemic refusal, escalation by the supplier through the local commerce department that carries the 满足率 target.

The premium: who pays and why tariffs are not published

Sinosure publishes premium rates for its credit products in no document at all — the rate card exists as an internal instrument and is disclosed only as the approved rate on a specific policy.

There are exactly four public tariff documents, and none of them concerns credit risk insurance: the fee for a letter of intent under project insurance, the guarantee business tariffs (financial guarantee 1% per annum, bid guarantee 0.2%, payment guarantee 0.6%), the tariffs for credit information products and those for agency credit information products, the last updated on 3 September 2026. In every set of policy wordings the rate is referred out to 《保险单明细表》, the individual schedule.

How the premium is computed. For short-term cover the industry formula is invoice amount × the rate from the tariff table × an adjustment coefficient (CPIC, 31 May 2021). For overseas investment the formula is primary and written into the wording: sum insured × rate × (months of liability / 12). For buyer's credit the base is principal plus computed interest over the whole tenor, with the premium paid in one instalment. The rate goes up with: country grade on the nine-point scale, the length of the deferral ("the longer the terms, the higher the rate"), the exporter's own credit and claims history, and the type of goods. Payment method feeds in directly: open-account risk on a contract written for a letter of credit or D/P can be covered only if a limit has been granted on OA terms and the premium paid at the OA rate (article 4(7) of rev. 4.0) — meaning open account is priced higher and paid up front.

The only rate documented by Sinosure itself relates to the 小微企业信保易 product in the company's 2012 material: annual premium = the prior year's customs export value × a rate, "as low as 0.13%", with a worked example — $1 million of exports, a $50,000 cap per buyer, a 70% indemnity percentage → a rate of 0.2% → a premium of $2,000, with the maximum annual indemnity set at 60 premiums, that is $120,000 (js.sinosure.com.cn, 2012). The current edition of that tariff is not published, so the figure cannot be carried into 2026: it sets an order of magnitude, not a current price.

Who actually pays. For a small or medium exporter, often not the exporter. The 政府统保平台 ("government umbrella platform") model has the local government buying cover in bulk and the enterprise paying nothing. Shenzhen exempts exporters with annual exports up to $8 million (2025/26 programme); Panyu district in Guangzhou raised that to $10 million in 2026 (163.com); Sichuan subsidises premiums up to 2 million yuan per enterprise. In Shenzhen the 大统保 model grew from 10,000 firms in 2020 to 24,000 (Shenzhen Bureau of Commerce via 163.com, 23 April 2026). MOFCOM's 2026 circular notes that local support is to be built "subject to compliance with WTO rules" — a qualification that did not appear by accident (see the section on criticism).

The practical conclusion for negotiation: the factory's argument that "insurance is expensive, hence the mark-up" is testable. The premium is measured in tenths of a percent of the amount, and for a small exporter in an umbrella scheme it is zero; the mark-up for payment terms is measured in whole percent. These are different orders of magnitude.

What deferred payment actually costs

The real interest rate on the transaction is the supplier's mark-up for the delay, not the Sinosure premium: the premium is measured in fractions of a percent, the mark-up in percent.

The formula. If the contract price on prepayment is P and, for D days of credit, the factory asks P × (1 + m), then you are borrowing P and paying P × m for it. The effective annual rate is r = m × 360 / D.

The benchmarks: the weighted average rate on newly issued corporate loans in China is 3.04% per annum (June 2026, PBoC Q2 2026 monetary policy report, PDF) with the LPR at 3.00% at the fixing of 20 August 2026 (NIFC); dollar trade finance runs at 5.2–6.2% (estimate); rouble working capital at 18–25% (estimate; the analysis sits in Chinese corporate lending).

Arithmetic on a live transaction. A $500,000 contract, 120 days of credit, a 2% mark-up: that is $10,000, or 2% × 360 / 120 = 6.0% per annum. A yuan loan at 3.04% over the same 120 days would cost $5,067. The $4,933 difference buys you the absence of a Chinese entity, of collateral and of a bank credit file.

What the delay costs the factory. Its own cost of money is the same 3% per annum, around 1.0% of the contract over 120 days, plus a premium of a few tenths of a percent — zero under an umbrella scheme. An honest cost of a 120-day deferral is 1.15–1.25% of the contract (estimate); everything above that is margin, not recovery of expense.

The policy as collateral: how banks plug in

A Sinosure policy becomes bank collateral only through a tripartite structure — a bilateral assignment of policy rights does not bind the insurer.

Assignment of insurance proceeds (转让赔款权益): the exporter remains the insured, but a tripartite agreement is signed between exporter, bank and Sinosure under which the insurer pays the financing institution directly and in full (Shanghai Financial Regulatory Bureau). The bank as the insured: bank policy 2.0 covers purchased receivables up to 180 days — unencumbered, with notice of assignment served on the buyer, and undisputed — or forfaiting, a non-recourse purchase of the claim with a participation up to 95%.

The legal point that is usually reported inaccurately. Article 28 of rev. 4.0 does not declare the assignment void — it says that dealing with policy rights without the insurer's written consent "has no binding force on the insurer". The difference is decisive: article 545 of the PRC Civil Code prevents a contractual ban on assigning a monetary claim from being asserted against a third party, and a "void" formulation would have broken on that rule. By writing "does not bind the insurer", Sinosure moved the question from validity to opposability: the assignment is valid between exporter and bank, but the insurer may ignore it and pay the policyholder. A further layer is article 22, which separately prohibits assignment of rights and transfer of obligations under the trade contract without consent: factoring the export receivable without Sinosure's consent breaches it independently of article 28.

Three different documents are routinely conflated in practice: 《赔款转让协议》, an assignment of the payment (the insurer remits the indemnity to the bank); 《代理索赔协议》, an agency mandate to present the claim, which by itself gives the bank no right to the indemnity; and 《索赔权转让协议》, an assignment of the claim right, effective only on signature by both parties and written notice to Sinosure (analysis in 《金融法苑》, Peking University, finlaw.pku.edu.cn; the leading case is (2012)皖民二终字第00101号, Anhui High Court).

By the end of 2021 some 300 banks had financed more than 4 trillion yuan against Sinosure policies (Trade Finance Global); credit enhancement under short-term insurance came to 262.46 billion yuan in 2024 (People's Daily, 6 March 2025). The most active lenders are CCB with its 信保贷 product, ICBC and Ping An Bank (the six-step process). No bank publishes an advance rate against the invoice; the ceiling is set by the indemnity percentage. For the buyer this yields a concrete argument: a factory that has assigned its policy rights to a bank is paid at shipment, not 120 days later, so a mark-up "for tying up working capital" has no basis. Bank appetite is set out in the survey of Chinese banks, payment routes in paying Chinese suppliers.

Claim, subrogation, collection

Subrogation is the transfer to Sinosure of the claim against the buyer to the extent of the indemnity paid; after payment your creditor is a Chinese state insurer, not the factory.

The average time to close a case for micro and small enterprises is 13.4 days, and the "48 hours" quoted in marketing material is a 2025 record rather than a service standard (People's Daily Overseas Edition, 26 June 2026).

How the indemnity is computed

The base is the lowest of three figures: the loss as determined, the credit limit on that buyer, and the declared invoice amount; the indemnity is that base × the indemnity ratio, capped in aggregate by the policy limit (article 11). The three ceilings apply in sequence, which is why under-declaration cuts the payout directly. Deducted from the base (article 13) are: amounts paid or set off by the buyer; unilateral discounts and waivers granted without the insurer's written consent; admitted counterclaims; proceeds of resale of the goods and of enforcement of security; payments by a guarantor; and interest, penalties and liquidated damages received under the contract. There is no deductible in the policy at all — the indemnity ratio does that job.

The legal basis for collection is contractual transfer, not statutory subrogation. Article 60 of the Insurance Law is drafted for damage caused to the insured object by a third party, and a buyer's non-payment fits that formula badly. Article 17 of the policy therefore obliges the exporter, after payment, to transfer to the insurer its rights under the trade contract and the letter of credit, to continue assisting collection, and to treat any sums received from the buyer before allocation as held for the insurer — a construction close to a trust — remitting the insurer's share within 10 working days. A mandatory order-of-application rule also operates (article 18): once the risk has materialised and before indemnity is paid, the buyer's payments discharge debt in order of payment date, oldest first, whatever the parties may have agreed otherwise. That is a direct departure from article 560 of the PRC Civil Code and it is what shuts down the practice of booking a payment against a new shipment: paying for fresh goods legally discharges the old insured debt.

Who runs the collection

After the notice is filed the exporter must either mandate the insurer to collect or collect on the insurer's instructions (article 15); otherwise the ratio is cut or the claim declined. Costs are shared in proportion to participation where the loss is covered, and fall entirely on the exporter if Sinosure finds the case uncovered (article 16). The definition of costs (article 31(8)) expressly includes the fees of collection agencies and law firms — confirmation of the outsourced model of overseas recovery; Sinosure does not publish the list of contractors.

Clawback (article 19). The indemnity is repayable with interest within 10 working days if the insurer discovers an undisclosed payment guarantee or a trade dispute; if the exporter accepted a return of goods on its own initiative, agreed a discount, waived the claim or concluded a settlement with the buyer behind the insurer's back; if the possibility of subrogation was lost through the exporter's fault; or if the cause of loss turns out to be outside cover. The contract sets no time limit on clawback — the grounds run from the moment of discovery. Hence the consequence usually missed: once the loss notice has been filed, the factory has no legal standing to agree a write-off with you — any deal behind the insurer's back wipes out its own indemnity.

A separate trap is the goods themselves. If the buyer has refused delivery, the exporter must first agree a disposal plan with Sinosure and realise the goods, and only then is the claim considered (article 14). For perishable or made-to-order goods this paralyses settlement for months.

How collection runs. The forums are the Chinese courts, including attachment and preservation measures, and arbitration, primarily CIETAC; abroad, Sinosure relies on the New York Convention. The pressure points inside China are the debtor's trademarks and patents, its accounts and receivables, inventory, moulds and tooling, and future shipments; an exit ban on the debtor's representatives is also reported (Harris Sliwoski / Mondaq, 27 July 2026). The insurer does not reduce principal — concessions are available on timing and on the removal of unjustified charges, which is consistent with the policy mechanics, since an unauthorised reduction of the claim triggers clawback and the negotiator simply has no authority to discount principal. Separately there is the double payment trap: where Sinosure has already paid the supplier, a payment made directly to the factory will most likely not discharge the insurer's claim, and a settlement requires a bilingual PRC-law agreement executed under company chop with releases from every holder of the claim (Harris Sliwoski).

The machinery works on the investment line too: in April 2026 Sinosure disclosed its first full recovery under an overseas investment policy — $11.38 million paid out, more than $10 million recovered, and the instrument was not litigation but an emergency payout followed by cross-border engagement with the host country's regulator (Sina Finance, 28 April 2026).

A default on one shipment hits every purchase in China at once: the loss notice cancels the limit automatically, the record is consolidated in a single database, and the 2026 products provide for joint handling of claims from different policyholders against the same risk subject. Restoring access begins only with payment in full.

Case law: trade disputes, refusals and forum

The main battleground under Sinosure policies is not the fact of non-payment but the trade dispute condition precedent, and the courts systematically narrow it, requiring the insurer to prove the dispute is material.

Why shipments with a probable goods dispute do not belong under cover. The policy contains no "quality dispute" exclusion, but the 贸易纠纷 construct operates as a deferred exclusion with a broad definition — in rev. 4.0, any dispute "sufficient to affect" the establishment and enforcement of the claim against the buyer (article 31(13)). The lever of withholding payment until a defect is remedied does not work here: withholding qualifies as protracted default, triggers the loss notice, cancels the limit and moves the dispute into collection by an insurer indifferent to the quality of the goods. Acceptance has to be closed before the payment date, and disputed shipments and first shipments from a new factory belong under a letter of credit.

Three generations of the clause. In rev. 3.0 it was hard: on a dispute the insured had to first obtain a final and binding judgment or award in the buyer's country and file it for enforcement, and only then would the insurer proceed to assess the loss. Rev. 4.0 turns it into a three-step ladder: first negotiation with the buyer (where a settlement is reached, the insurer verifies it and pays); then the insurer must itself determine liability on the basis of the contract and its investigation; and only if the insured has neither settled nor accepted that determination does the foreign proceeding follow. Bank policy 2.0 keeps the hard version and adds a cut-off: if within two years of the payment date no final and binding decision has been obtained and filed for enforcement, cover on that receivable lapses. The materiality threshold reached the comprehensive policy from below — the 普惠 range introduced both the ladder and the "sufficient to affect" formula as early as 2016.

These cases are assembled from consistent publications by several independent law firms (Tsinglaw; AnJie Broad, 26 June 2025; Deheng; Jingtian & Gongcheng); the case numbers and holdings are given as reported there.

Case law on the adjacent ground — shipment after knowledge of the risk (知险后出运) — is inconsistent. The Qianhai court in (2018)粤0391民初319号 treated the buyer's first missed payment as sufficient; the Chongqing High Court in (2019)渝民申2769号 required evidence of deteriorating creditworthiness and held that on a simplified renewal of the policy the insurer had not discharged its duty to explain — meaning auto-renewal destroys reliance on the exclusions (AnJie Broad, 22 October 2024). The burden of proving that the exporter knew or ought to have known rests on the insurer.

Applied less often are non-disclosure (article 25, rescission within 90 days), non-payment of premium (article 7), unauthorised disposal of the goods (article 14) and refusal to mandate collection (article 15).

How exporters push back

Five lines of defence run through these cases.

1. Deny that the dispute is material, leaning on the "足以影响" formula.
2. Place the burden of proof on the insurer.
3. Attack the discharge of the duty to explain under article 17 of the Insurance Law, reinforced by article 9 of 法释〔2013〕14号, which classes proportional indemnity as a liability-exempting term.
4. Challenge the foreign enforcement requirement as excessive.
5. Require the insurer to work through the second rung of the rev. 4.0 ladder — to determine liability itself before sending the insured to a foreign court; the newest line and the least exploited.

Separately, the courts apply the three-year limitation period of article 188 of the PRC Civil Code rather than the two-year period of article 26 of the Insurance Law.

Forum and enforcement. The clause demands arbitration or litigation in the buyer's country, and a Chinese judgment does not satisfy it. The reason is enforceability: an arbitral award is enforceable in roughly 170 jurisdictions under the New York Convention (China acceded in 1987, with the commercial and reciprocity reservations), whereas China has no civil judicial assistance treaty with the United States, the United Kingdom, Germany, Japan, Canada, Australia or India; it has not signed the 2019 Hague Judgments Convention, and it signed but did not ratify the 2005 Hague Choice of Court Convention.

The mirror case shows the machinery running in both directions: in (2015)浙甬商外初字第181号 the Ningbo Intermediate Court on 24 March 2017 upheld a claim by the Hungarian export credit agency Magyar Exporthitel Biztosító against the shareholders of a liquidated Chinese buyer, splitting the applicable law — the sale contract under the CISG, the guarantee and the subrogation itself under Hungarian law, and shareholder liability after liquidation under PRC law (SPC International Commercial Court). Chinese courts recognise a foreign ECA's subrogation and pierce the veil of a liquidated debtor — exactly what Sinosure does abroad.

Sanctions perimeter and country restrictions

Sanctions protection under a Sinosure policy depends not on the company's stance but on which product was bought and in which year — there is no single regime.

There are three regimes, and conflating them is the standard error of surveys. The flagship rev. 4.0 wording contains no sanctions clause at all — nor do rev. 3.0, the 小微 range or bank policy 2.0.

The April 2026 products contain the widest possible one: payment must not breach UN resolutions or the sanctions or anti-sanctions rules of China or of any other jurisdiction, plus a separate exclusion of countries named in a list annexed to the policy.

Overseas investment insurance contains an inverted construction: sanctions against the host country that block currency conversion are a covered risk, and only certain sanctions are carved out — and the carve-out has been rewritten over six years. The 2019 edition excluded "sanctions other than those of the UN Security Council"; the 2025 edition excludes "sanctions imposed by the Chinese government, in which it participates or which it has undertaken to observe".

This is the mirror image of a Western sanctions exclusion: loss from unilateral US and EU measures stays covered because China does not implement them, while the carve-out has quietly widened from UN resolutions to any Chinese measure, counter-measures included.

Country exclusions. The documented primary list is the exclusion schedule of the 小微企业信保易 product as published by Sinosure in 2012: Cuba, Iran, Sudan, South Sudan, Libya, Côte d'Ivoire, Yemen, Syria, Afghanistan, Iraq, Somalia, Burundi, Chad, Comoros, DR Congo, Eritrea, Guinea-Bissau, Liberia, Rwanda, Sierra Leone, Zimbabwe and the State of Palestine (js.sinosure.com.cn, 2012); the current edition of that schedule is not published and it cannot be read as a live 2026 list. The "Sinosure country blacklist" of roughly 35 states that circulates online is described as speculative by its own author (Harris Sliwoski). In law there is no blacklist as a public institution: country access is set by the 短期险国家（地区）分类表, the non-public country classification table supplied with the policy, and by the internal country underwriting policy that article 31 of the NFRA measures makes mandatory.

The Russian perimeter, factually. Russia does not appear on the documented exclusion list, no public act terminating cover of Russian risk exists, and the reports of refusals are confined to the Russian business press of 2024, which has not returned to the subject since. The restricting mechanism needs no public act: it runs through the non-public country classification, underwriting policy under article 31 of the NFRA measures and an unreasoned decision on the limit — which presents as a refusal without stated grounds and a factory demanding prepayment or a confirmed letter of credit. The second layer is the sanctions clause in the 2026 products, which refers to the rules of any jurisdiction and allows a claim to be declined after the goods have shipped. The third is counter-sanctions: article 24 of the State Council regulation on outbound investment, in force from 1 July 2026, refers to the PRC Anti-Foreign Sanctions Law (Sinosure). Sanctions screening of the chain belongs before the limit application rather than after the refusal; the wider country picture sits in the map of sanctions regimes.

Two non-obvious conclusions follow from what the policy does not contain. There is no general force majeure clause in short-term cover in any edition — so an epidemic, a natural disaster or a port closure that stops the buyer paying is not automatically excluded: it simply has to fit the definition of protracted default or the list of political risks. Currency restrictions are likewise not an exclusion but a covered political risk; only exchange-rate loss is excluded.

The numbers year by year

Sinosure's public reporting consists of exactly three quantities — insured volume, claims paid and client count; the company discloses neither premiums, nor loss ratios, nor solvency.

Sources by row: 2008 — People's Daily, 30 March 2009; 2020 — Belt and Road Portal, 2 March 2021; 2021 — Sina Finance, 19 January 2022; 2022 — Belt and Road Portal, 17 January 2023; 2023 — People's Daily Overseas Edition, 24 January 2024; 2024 — 中证网, 20 January 2025 and People's Daily, 6 March 2025.

Annual and cumulative client counts are different measures and are regularly conflated: 227,000 for 2024 means clients served during the year, 380,000 at June 2025 and 400,000 at the end of 2025 (SINOSURE Profile) are cumulative since inception, and 320,000 in the 14th Five-Year Plan totals covers five years. Insurance penetration of exports ran at 22% in 2020, 23.1% in 2023 and 27.4% in the first half of 2025. The ratio of claims paid to insured amount for 2024 is 0.25%; since the insured amount is a revolving flow rather than an outstanding balance, that figure sets the scale but does not by itself distinguish a low-loss portfolio from a hard claims-handling practice.

What Sinosure does not publish: premium income (保费收入), loss ratio, actual solvency, total assets and net profit. Unlike every large Chinese insurer, it issues no publicly available quarterly solvency reports. That is a fact in its own right rather than a gap in the sources: it is corroborated both by the OECD's 2015 findings and by the European Commission's record of non-provision of premium and profitability data in an anti-subsidy investigation.

Place in the world: what compares and what does not

Sinosure is the world's largest export credit agency by insured volume, but comparing its $1 trillion directly with the figures of Western ECAs is a category error rather than a statement about the size of the gap.

The reason is structural: ≈84% of Sinosure's volume (860 of 1,021 billion in 2024) is short-term revolving cover of trade flows — the segment in which the West is served by the private insurers Allianz Trade, Atradius and Coface rather than by state agencies. Western ECAs mostly publish medium and long-term commitments. Setting "$1 trillion for Sinosure against $4.7 billion for US EXIM" compares a revolving annual turnover of shipment insurance with a flow of new project loans.

The source is US EXIM, Report to Congress on Global Export Credit Competition, 2024, PDF. The Chinese line in this table is not reported data but EXIM's own estimate: the agency keeps an internal database and draws on academic publications because Chinese transaction data is "typically not publicly available", and it writes that in 2023 the Chinese ECAs "are estimated to have authorized approximately $15.3 billion". The other lines are agency-reported figures. In that comparable category Italy held the top spot in 2023, as it had the year before, and China was second. The trend runs both ways: the Chinese figure averaged $35 billion a year in 2016–2019 and more than halved by 2023, yet within 2023 the same report records a surge of more than 41% in Chinese official medium- and long-term export credit activity — against a 2022 base that EXIM calls artificially low because of the pandemic. The mix shifted from megaprojects to smaller transactions with creditworthy counterparties; at the same time China was scaling back its use of policy banks in favour of state-owned commercial banks, which makes measurement harder.

The correct comparison by portfolio rather than flow reverses the picture: Sinosure's exposure in 2022 was almost $900 billion against $55.6 billion for Japan's NEXI (ODI, 11 April 2024). Here the gap is real and it is an order of magnitude.

What makes Sinosure singular is that it holds both roles on one state balance sheet: simultaneously the world's largest short-term trade credit insurer and a policy agency for sovereign risk. For scale: the entire global private trade credit insurance market in 2024 was $12.8 billion of premium at a combined ratio of 76%, with Allianz Trade at a 31% share and Coface turnover of $1.85 billion (Atlas Magazine, 26 March 2026). Berne Union data for 2025: new short-term trade supported $3,714 billion (+12%), new medium and long-term commitments a record $230.9 billion (+17.7%), total claims paid $11,107 million, and the four largest providers accounting for 82% of short-term cover (Berne Union, State of the Industry 2025).

China never signed the rulebook. The PRC is not a Participant in the OECD Arrangement on Officially Supported Export Credits. The International Working Group on Export Credits (IWG), set up in 2012 on a US initiative precisely to extend common rules on minimum rates, premiums and tenors to China, was suspended in December 2020: the suspension statement was signed by the EU, the United States, Japan, Australia, Brazil, Canada, Korea, New Zealand, Norway, Switzerland and Turkey — and not signed by China, India, Indonesia, Israel, Malaysia, Russia and South Africa (GTR, 2 December 2020). It failed over the refusal of non-OECD agencies to disclose transaction terms and over disagreement on scope and debt sustainability. The response was to modernise the Arrangement itself in 2023: maximum repayment tenor was raised from 8 to 15 years (up to 22 for green projects) and premiums on long-term guarantees were cut — the OECD adjusted its own rules to Chinese competition rather than the other way round.

The paradox of industry weight. Holding first place by volume since 2015, Sinosure appears once in the Berne Union's table of leaders since 1934: HUANG Zhiqiang served as vice-president in 2018/19, and no president has ever come from Sinosure (Berne Union, Former Leaders, as at 4 September 2026). Nor is the company in the published Management Committee — at the same date the page shows the 2024/2025 committee (president Yuichiro Akita, NEXI; vice-president Rafael Docavo-Malvezzi, AXA XL), whose institutional members are Atradius, Bpifrance, Coface, EDC, EIFO, EKN, Everest, KredEx, K-SURE, PwC, Swiss Re, UKEF and US EXIM (Berne Union, Leadership). Institutional influence in the trade association is inversely proportional to market share.

How to compare correctly: by flow of new medium and long-term commitments, China sits in the same league as Italy and Germany; by portfolio, an order of magnitude above any other ECA; in the short-term segment, the comparators are not US EXIM but Allianz Trade and Atradius, and Sinosure alone covers a volume that is a multiple of the whole private market's premium. Any figure quoted without its category is useless.

Criticism and disputes

The principal charges against Sinosure are subsidy (cover priced below the market price of risk), transparency (non-disclosure of tariffs and results) and debt (its position in sovereign debt restructurings).

Recognition as a "public body" in EU anti-subsidy cases

In the electric bicycles regulation (Commission Implementing Regulation (EU) 2019/72 of 17 January 2019) the European Commission classified Sinosure as a public body. The company is called "a fully State-owned entity" in recital 160 and confirmed as much at the verification visit to the Chinese government (recital 352); on its own questionnaire reply it is "100 % owned by the State", it has a board of directors and a board of supervisors, and the government has the power to appoint and dismiss its senior managers (recital 354).

The conclusion itself sits in recital 360: the legal framework is implemented by Sinosure "in the exercise of governmental functions", so it acts as a public body in the sense of Article 2(b) of the basic Regulation read in conjunction with Article 3(1)(a)(i); the WTO Appellate Body test from WT/DS379/AB/R was applied by the Commission in the adjacent section on the state-owned banks (recital 176), and not to Sinosure.

How it was proved matters: Sinosure replied only in part, produced no corporate governance documentation — annual report and articles of association, withheld as confidential (recital 157) — and "did not give any specific information about the export credit insurance provided to the electric bicycle industry, the level of its premiums or detailed figures relating to the profitability" (recital 158), after which the Commission applied Article 28(1) of the basic Regulation — the "facts available" rule (recitals 160–162) (EUR-Lex, 32019R0072). The refusal to disclose became evidence against the company.

The United States and the WTO. On 18 December 2008 the USTR filed a request for consultations at the WTO against the Chinese "Famous Export Brand" and "China World Top Brand" programmes — more than 70 measures, with "payments to reduce the cost of export credit insurance" expressly named among the challenged subsidies, though Sinosure was not named (USTR, 18 December 2008). There has been no separate WTO dispute against Sinosure as such: enforcement runs through national countervailing duties rather than through the WTO.

Pricing below OECD norms

US EXIM documents specific transactions: on Ghana (2021) the spread was roughly 25% below the fee produced by the OECD risk model, and on Serbia (2023) nearly 40% below (EXIM, Report to Congress 2024, PDF). A separate scale estimate from the 2019 EXIM report: Chinese medium and long-term export credit activity in 2015–2019 ran at no less than 90% of the volume of all G7 countries combined (GTR, 2 December 2020).

Transparency. The primary document is the OECD study TAD/ECG(2015)3, "Chinese Export Credit Policies and Programmes", of 16 March 2015 (one.oecd.org). Of China's three external finance institutions, Exim Bank and CDB did not respond at all; only Sinosure engaged — its reputation for opacity is relative. The OECD's 2013 estimate: Sinosure covered ≈15% of all Chinese exports against ≈2% for Exim Bank and 3–5% for CDB, a share materially above that of OECD-country agencies. What is not disclosed: premium rates, interest rates and fees, country breakdowns, environmental standards and results by line of business. A separate charge concerns the non-publication of country refusal criteria, which turns a commercial decision into an opaque trade barrier.

Sovereign debt restructurings

In the Zambian case Sinosure holds subrogation rights that give it standing in inter-creditor negotiations, yet its participation has been called "curiously muted" — China's lead representative was Exim Bank. The substantive position matters more: Sinosure, with Exim Bank, resisted classifying the loans it insured as "official" debt subject to restructuring — a departure from Paris Club norms which, on the Boston University GDP Center's assessment, took $4.1 billion of CDB loans outside the perimeter and left Chinese commercial creditors worse off than bondholders, undermining comparability of treatment (ODI, 4 April 2024). The June 2023 agreement, under the co-chairmanship of China and France in the Official Creditor Committee, covered $6.3 billion. The side effect on the market is a contraction of risk appetite: portfolio data show a marked reduction in medium and long-term cover after 2018, and cover on Pakistan was cut from 95% to 70%.

Libya, 2011. Industry press reporting put the claims produced by the uprising at more than $1 billion from 13 Chinese state enterprises invested in the country. Chinese reporting of the time put total losses of Chinese companies at around $20 billion against insurance indemnities of less than $400 million — meaning only a small fraction of the exposure was insured (Jinhua Bureau of Commerce, 26 April 2011). It was this episode that preceded the capital injection of 2011–2012.

Local premium subsidies. The 大统保 model, under which a local government buys the policy centrally and enterprises receive cover free of charge, is exactly the practice of paying premiums out of the budget that the EU and the United States treat as a countervailable subsidy. Hence the "subject to compliance with WTO rules" qualification in MOFCOM's 2026 circular.

Internal discipline. The company has a corruption episode of its own: Tang Ruoxin, general manager since incorporation, was sentenced in 2010 to 14 years for bribery — the case that closed the first phase of its history.

A rule that confirms the charge. Article 64 of the NFRA measures exempts policy business from Insurance Security Fund contributions precisely on the ground that the ultimate risk is carried by the state budget. That is a direct regulatory acknowledgement of the budget backstop — the very proposition on which the anti-subsidy cases in Brussels and Washington are built and which the Chinese side contests in those cases.

What changed in 2025–⁠2026

Three things changed over two years: the supervisory regime, the insurer's capacity and the rules of access to cover for small business.

One gap is worth recording: national full-year figures for 2025 and first-half figures for 2026 have not been published so far, and the $565.6 billion figure that circulates in several write-ups relates to the first half of 2025, not to a full year. Only provincial cuts are available — the Hubei branch, for instance, reported $7.1 billion and 4,600-plus enterprises for the first half of 2026.