As of: 13 October 2026
Geographic scope: People's Republic of China (mainland financial system unless stated otherwise)
Reader question: How is China’s financial system organised, where do its principal transmission risks sit, and what limits the international role of the renminbi (RMB)?
Main Content
China’s financial-sector transformation is not a linear story of state control giving way to a globally integrated market. A more useful reading is that the country has built a very large, bank-led system in which the state remains deeply involved in institutions, policy coordination and risk management, while market infrastructure and selected channels to foreign participants have expanded. The practical framework is four questions:
- Who allocates credit and absorbs losses? Start with banks and the public-sector links around them.
- How can a property or local-government shock travel? Trace it through borrowers, banks, bondholders and asset-management products rather than treating sectors in isolation.
- Which authority owns which risk? Separate monetary and macroprudential functions from prudential oversight, securities supervision and foreign-exchange administration.
- What can cross-border RMB use achieve without full capital-account openness? Distinguish trade settlement, market access and reserve use from unrestricted convertibility.
This framework matters because the system’s scale and structure shape the way vulnerabilities are managed. The IMF’s 2025 Financial System Stability Assessment (FSSA) described China’s financial system as having grown to more than 450% of GDP, with banks dominant; it also described the system as largely government-controlled and highly interconnected, with limited direct connections to global financial markets. Those features make simple labels such as “state-led” or “liberalising” incomplete. 1
The system is bank-led, but not bank-only
Banks are the central channel for deposits, credit and policy transmission. They sit alongside large bond and equity markets, insurers, trusts, funds and bank wealth-management products (WMPs). The World Bank’s 2025 financial-sector assessment likewise characterises banks, including globally systemic institutions, as dominant while recognising sizeable capital-market and asset-management sectors. 2
The distinction is important. A bank-led system is not one in which all risk stays on bank balance sheets. Asset managers and other non-bank financial intermediaries can hold bonds and other credit claims, compete for household and corporate savings, and be connected to banks by ownership, common assets and funding. The IMF reports that non-bank risks have been reduced since the 2017 assessment through regulatory reforms, but stresses that remaining linkages can transmit a redemption shock across sectors. 1
State involvement also operates through more than ownership. Policy institutions, regulatory direction, public-sector borrowers, local authorities and expectations of official support all influence the incentives facing lenders and investors. The IMF cautions that implicit support and forbearance can defer loss recognition and weaken risk pricing. That is an analytical point about incentives and balance sheets, not a claim that support is automatic in every case. 1
The reader should therefore resist two opposite shortcuts. It is inaccurate to treat China’s institutions as ordinary private-market intermediaries operating in a fully open capital market. It is equally inaccurate to treat market prices, non-bank products and corporate-bond funding as irrelevant. The policy challenge is the interaction: how formal rules and market instruments work alongside state-linked incentives.
Property and LGFVs are connected credit problems
The property downturn and local-government financing vehicles (LGFVs) are often discussed as separate issues. In a financial-stability framework, they are connected through cash flow, collateral, local fiscal capacity, bank credit, bond markets and confidence.
A property-market adjustment can weaken developers, reduce construction activity and affect households, suppliers and local land-related revenue. It can therefore change loan performance, demand for credit and the value of collateral. The IMF’s 2025 FSSA identifies the property downturn as a material vulnerability that is already affecting financial institutions, and notes that slower growth can impair credit quality. 1 The IMF’s subsequent 2025 Article IV consultation similarly identifies ongoing property correction and the LGFV debt overhang as significant vulnerabilities. 3
LGFVs are entities used to finance local development and investment. Their financial position matters because they borrow from, and issue securities held by, parts of the financial system. The relevant question is not only whether an individual issuer can refinance, but whether weaker debt-servicing capacity could affect lenders, holders of bonds, asset managers and local funding conditions. The FSSA’s illustrative work finds channels from an LGFV restructuring to bank earnings and capital, as well as potential valuation and redemption pressure in some asset-management products. These are stress-analysis mechanisms, not forecasts of a specific event or loss. 1
That distinction is essential. The IMF assessment says that aggregate bank capital and liquidity appeared adequate, and its stress testing found resilience at the system level under its adverse scenario. It also found meaningful variation across banks, with smaller institutions and some business models more exposed to weaker profitability, higher funding costs or concentrated risks. 1 The Article IV assessment likewise says smaller and mid-sized banks have relatively greater credit-quality, funding-diversification and profitability concerns than large state-owned banks. 3
For analysis, the transmission chain can be read in five steps. First, a property or local-government cash-flow problem affects a borrower. Second, that exposure can affect a bank loan, a bond or an asset-management product. Third, losses or lower interest income can reduce an intermediary’s capacity to lend or meet redemptions. Fourth, confidence and funding costs can shift, especially where exposures or fiscal positions are concentrated. Fifth, the policy response can alter the distribution and timing of losses without making the underlying cash-flow issue disappear. This is why headline non-performing-loan measures alone are an incomplete map of risk.
There are evidence limits. The IMF reports that its system-wide assessment faced gaps in access to granular balance-sheet and exposure data, including the distribution of LGFV exposures across banks. It explicitly recommends more granular analysis using non-public data. 1 Public commentary should therefore not infer precise exposures, loss rates or rankings for individual institutions unless current, underlying supervisory disclosures support them.
Regulation has become more consolidated, but the architecture remains multi-agency
China’s framework assigns different layers of financial risk to different authorities. The People’s Bank of China (PBOC) is the central bank. Its published functions include monetary and credit policy, macroprudential management, systemic-risk prevention, lender-of-last-resort functions, payment-system oversight, supervision of specified financial markets, RMB exchange-rate policy and monitoring of cross-border capital flows. The PBOC also administers the State Administration of Foreign Exchange (SAFE). 4
The National Financial Regulatory Administration (NFRA) is responsible for unified supervision and regulation of the financial industry other than securities, according to its published mandate. Its responsibilities include banking, insurance and financial-holding-company oversight; authorisation; corporate-governance, risk-management and capital or solvency supervision; consumer protection; resolution mechanisms; and coordination with local governments in handling relevant financial risks. 5
Securities and futures activity is outside that NFRA mandate. The result is not a single-regulator model. It is an architecture in which macroprudential oversight, prudential supervision, securities-market supervision and foreign-exchange administration must work together. The FSSA recognises progress in supervisory reform, including the creation of the NFRA, but calls for further supervisory resources, analytical capacity and inter-agency cooperation. 1
A concrete example is systemically important banks. In February 2026, the PBOC and NFRA reported that their 2025 assessment identified 21 domestic systemically important banks and said they would combine macroprudential management with microprudential supervision. 6 This does not by itself establish that every potential systemic risk is resolved. It does show the intended division of labour: system-wide calibration and institution-specific supervision are designed to be complementary.
Crisis management is the remaining structural test. The IMF says the current framework does not yet support the full range of options needed to manage systemic distress, and recommends a strengthened Financial Stability Law, an independent and adequately resourced lead resolution authority, and an effective emergency-liquidity-assistance framework. 1 These are IMF recommendations, not a statement that a particular law has been enacted or that a future crisis will occur.
Non-bank finance has changed the location, not necessarily the existence, of risk
China’s non-bank sector is particularly relevant where savers’ funds are invested through WMPs, trust products, funds or other asset-management arrangements. Regulatory reforms reduced some earlier risks, according to the IMF. But non-bank intermediaries remain linked to banks through ownership, funding and common holdings, including credit instruments. 1
This means the key issue is liquidity and interconnectedness as much as credit quality. A loss in a bond or other asset may not be confined to the original borrower if it prompts redemptions, valuation pressure or a need to raise cash. Conversely, a system with more market-based instruments can make risk more visible and diversify funding. The balance depends on product structure, investor expectations, liquidity management and the credibility of loss allocation. Those facts need product-level evidence; this article does not make claims about the safety or suitability of any individual product.
Capital-account management is a constraint and a policy choice
China has widened selected routes for cross-border investment and financing, but that is different from full capital-account convertibility. SAFE’s own mandate includes administration under the capital account and making gradual advances in RMB capital-account convertibility. Its 2025 work programme describes a goal of a foreign-exchange system that is “more convenient, more open, and secure,” while also emphasising macroprudential management, counter-cyclical adjustment and stronger monitoring and early warning for cross-border capital flows. 7 8
This combination helps explain a central feature of the system: opening is managed by channel, eligibility and purpose. It can include facilitation pilots, foreign-investor access arrangements, overseas securities investment channels or reforms to foreign-debt administration, while retaining tools to manage capital flows and foreign-exchange risk. The presence of a channel should not be read as permission for unrestricted movement of all capital.
The constraint affects the domestic system as well as foreign participation. The IMF observes that strict capital controls channel domestic financial savings primarily into domestic assets. 1 It may reduce some immediate exposure to volatile cross-border flows, while also limiting the ease with which domestic savers and foreign investors can freely reallocate portfolios. Neither effect is inherently good or bad in isolation; both are part of the policy trade-off between openness, exchange-rate management and financial stability.
RMB internationalisation should be measured by use, access and liquidity—not a single ambition label
The RMB can become more useful internationally without becoming a fully convertible, dominant global reserve currency. Trade settlement, cross-border payments, offshore liquidity, foreign access to onshore assets, official reserve holdings and foreign-exchange trading are separate measures. Each has a different user base and different requirements.
BIS data offer one clear, dated reference point: the 2022 Triennial Survey found that the RMB was on one side of 7% of global foreign-exchange trades in April 2022, making it the fifth most traded currency in that survey. 9 That statistic is a historical market snapshot, not a current ranking and not a measure of trade invoicing or reserve use.
Reserve data need the same discipline. The IMF’s COFER database separately identifies the Chinese RMB and explains that it is a quarterly, voluntary survey of currency composition; it also notes a methodological revision beginning with 2025Q3. 10 Any article that reports a current RMB reserve share should specify the exact COFER vintage and whether it uses the revised series. This draft intentionally does not state one without locking a publication-date data extraction.
Internationalisation also depends on whether users can acquire, hedge, hold and exit RMB positions with sufficient confidence and liquidity. A Federal Reserve research note concludes that China’s capital controls and managed exchange-rate system can weigh on international use by increasing the risk or cost of divesting RMB assets, even as formal access channels exist. 11 That is an institutional analysis, not an assertion that capital controls make international RMB use impossible.
The most defensible conclusion is therefore narrow: China has developed more routes for RMB use and for cross-border engagement with its financial markets, but the policy objective of controlled opening places a structural limit on equating those developments with unrestricted capital mobility. A serious assessment should track distinct indicators and their definitions, not make a binary claim that the RMB has either succeeded or failed internationally.
What this transformation means—and what it does not establish
China’s financial system is changing through regulatory consolidation, market development, non-bank reform and selective external opening. Yet its central constraints remain visible: a bank-dominant, state-shaped allocation system; property and LGFV balance-sheet links; interconnected non-bank finance; incomplete public data on some exposures; and an external-finance regime designed to combine opening with management of cross-border risk. 1 7
This is a framework for understanding institutional structure and transmission channels. It is not a prediction of financial stress, a judgement on particular issuers or banks, or investment, legal, tax, accounting or financial advice. Readers assessing a specific exposure should use current primary disclosures, applicable law and appropriately qualified professional advice.
Research Desk Notes
Methodology. Aurex Intelligence Research Desk rebuilt this article from regulator and institutional sources read on or before 13 October 2026. Priority was given to the PBOC, NFRA, SAFE, IMF, BIS and World Bank. The analysis distinguishes official mandates and data from IMF or other institutional assessments, and labels stress-test mechanisms as illustrative rather than predictive. It does not use market-price targets, investment recommendations, rankings, sponsored placements or affiliate links.
Evidence limitations. The most detailed system-wide risk findings cited here come from the IMF’s 2025 FSSA and its 2026 publication of the 2025 Article IV consultation. They are not a substitute for confidential supervisory data or for institution-level due diligence. Several PBOC, NFRA and SAFE English pages state that the Chinese text controls if interpretations differ; this draft relies on the English materials as accessible reference translations. 4 5 7
Corrections. To suggest a correction or source update, contact info@AurexIntel.com with the article URL, the relevant passage and a primary or institutional source.