The Past Has the First Claim
China’s long-predicted financial crash has not arrived, and its absence has become part of the damage. Inherited claims consume the present while Beijing builds a narrower machinery of credit for the industries it has chosen to carry into the future.
China's financial system can produce money faster than it can produce the permission to lose it. Liquidity arrives whenever the money market tightens, in whatever volume the central bank decides. The capacity to absorb loss is a different substance: it sits in bank earnings and public balance sheets, and it depends finally on someone in authority agreeing that an earlier investment has failed. That capacity is finite, and most of it is spoken for by investments already made. An unrecognized loss does not wait its turn — it occupies the balance sheet until somebody is permitted to write it off, and new risk has to be underwritten out of whatever is left. The future is still financed, but the past has first claim.
The claims now doing the occupying were accumulated during an achievement without modern parallel. Between 2008 and 2017 Chinese banks added roughly $27 trillion in assets, financing an immense expansion of property and municipal works alongside an industrial system capable of supplying almost every stage of production. Not all of it was meant to pay for itself in the ordinary way. Projects with no credible cash flow were allowed to remain alive on the expectation that growth would eventually make yesterday's excess affordable, which meant that each year of expansion produced both the assets on the books and a quantity of deferred judgment about what they were worth.
For as long as the system expanded faster than the claims accumulating inside it, the deferral cost nothing. Property sales replenished local governments through land revenue, rising prices improved collateral, and banks could refinance weak borrowers while lending enough elsewhere to keep the economy moving. Solvency was repeatedly converted into growth. That conversion has now run backwards: instead of expansion absorbing the claims, the claims have begun to consume the expansion, because the income that was supposed to retire them has stopped arriving.
The figures describe the same reversal from the outside. Property investment fell by 17.5 percent in 2025 and by a further 19.2 percent during the first seven months of 2026, while annual growth in outstanding yuan loans slowed to a record-low 5.1 percent in July and new lending contracted by RMB340 billion during the month. The old engine is demanding support even as its ability to generate the income needed to sustain that support disappears.
Aggregates of that kind are made of individual obligations, and the obligations are more durable than the assets behind them. At Xiulan County Mansion in Guilin, the mortgage became binding long before the home became habitable. A buyer, Xu, moved into the unfinished apartment on which her family had spent its savings, sleeping beneath a mosquito net among unpainted walls and holes where electrical sockets should have been, carrying bottles of water up the stairs and sharing a makeshift outdoor toilet with around twenty other buyers. The developer's promise failed even while the family's liability did not.
The rule that bound her was changed in August 2026, when regulators required mortgages on new homes to be issued only after completion and instructed local governments to promote the sale of finished units. Separating a household obligation from a developer's unfulfilled promise may prevent the next Xu from inheriting the same risk. But the reform arrived more than five years into the property slump and governs only the creation of future claims. It does nothing about the stock already outstanding, which is where the weight sits.
Preserving that stock is the method by which the crisis long predicted for China has been stretched from an event into a condition. Loans are rolled over and maturities lengthened, local liabilities are exchanged for bonds carrying lower interest rates, and the institutions that created the old assets are enlisted to prevent their disorderly recognition as losses. Yet a crisis is also a form of resolution. It forces claims into the open, assigns the losses to somebody, and leaves the survivors free to move again — which means that preventing the crisis also prevents the clearing, and the absence of collapse becomes part of the decay rather than evidence against it. China's preservation economy allows obligations to consume income and administrative attention long after their original purpose has expired. There is no single year in which the losses are taken, and therefore none after which the system is free of them, which means the balance sheet remains upright only by surrendering lots of its freedom of movement.
That lost movement is hard to see in the price of credit, because the price looks so utterly unremarkable. A corporate loan a little above 3 percent is not obviously distressed lending. What it leaves the bank is another matter: once funding costs and expected losses have taken their share, only a fraction of that return survives, and it is in the margin rather than the rate that the constraint becomes visible.
Commercial banks' net interest margins reached a record low of around 1.4 percent in early 2026 and then steadied, which sounds like relief until one asks how. The stabilization came substantially from cutting what banks pay depositors rather than from any improvement in what their assets earn. A margin defended by paying savers less is not evidence of recovered capacity; it is the same shortage passed along to households. Where the shortage could not be passed along, the state has supplied capital directly. Four large state banks raised a combined RMB520 billion in 2025, most of it from the Ministry of Finance through special treasury bonds, with a further recapitalization in 2026. Presented as additions to the banks' capacity to support the real economy, which they are, the injections also reveal the rate at which low margins and weak assets consume that capacity.
None of this makes cheap credit proof of bad allocation. A safe borrower may sustain a low rate, and an innovative company may deserve finance before its profits become visible. But price loses much of its information once it can no longer distinguish patience with a productive uncertainty from the preservation of an asset that has already failed. A bank rolling over a local government financing vehicle carries yesterday's investment forward; one supporting an uncertain semiconductor process may be purchasing knowledge whose value appears only after several failures. Both transactions enter the aggregates as credit growth, although the claims they leave behind are different.
The same ambiguity runs through the local-government debt exchanges, on a far larger scale. Beijing authorized RMB10 trillion of relief, and the bonds issued under it have lowered average interest costs by more than two and a half percentage points. Provinces gain room to function and the risk of disorderly default recedes. What the exchange does not do is retire the underlying claims, since an asset whose income could never repay its debt is no better able to repay it at a lower rate over a longer term. The question of whether the investment was sound has been converted into a question of whether the interest can be met this year, which is a question the state can always answer. Financial time becomes a substitute for financial return.
The administrative record registers the same substitution. Between March 2023 and September 2025 the number of financing vehicles on the official list fell by 71 percent, a decline produced through refinancing and restructuring as much as through genuine transformation. A platform can disappear from the register while its obligations remain under supervision and its creditors continue to be paid. What has changed is the classification, not the liability, and the center has acquired greater sovereignty over how claims are described than power to extinguish them.
Claims that cannot be extinguished must coexist with the industries through which China expects to remain powerful, so new finance is being asked to preserve the institutional peace surrounding old debts while underwriting a different future. The two purposes once expanded comfortably together. Within a slower-growing balance sheet they have begun to compete, and the competition is visible in who still receives credit. Household borrowing contracted sharply in April and again in July 2026, while outstanding loans to technology-focused small and medium-sized firms had been growing at nearly 20 percent at the end of 2025, far ahead of overall lending. Real-estate loan balances were already falling as banks announced new targets for AI, semiconductors and advanced manufacturing. A shrinking total combined with a rising strategic share is the signature of allocation that has stopped being an outcome and started being a decision.
That decision is made more openly in fiscal policy, where China allocated nearly RMB1.3 trillion to science and technology in 2026. A relending facility for technology created in 2022 at RMB200 billion has been enlarged three times and stood at RMB1.2 trillion by early 2026, while bank executives came to be evaluated partly on the finance they directed toward it. A lending target that appears in a manager's appraisal has ceased to be a market signal and become an instruction, giving scarcity an administrative grammar of sorts.
The pattern admits a simpler explanation, since a government whose revenues have stopped growing may merely protect the sectors with the strongest political sponsors and confer the label "strategic" on whichever interests survive. The term offers no proof of discrimination, and the breadth of China's industrial catalogs gives the suspicion plenty to work with. More revealing are the institutions built before returns became visible, expanded through commercial disappointment and given horizons that ordinary bank lending cannot support. The National Venture Capital Guidance Fund has a planned life of two decades and is intended to channel capital through hundreds of subordinate funds into seed-stage and early-stage companies. A twenty-year fund is not a bet on any particular company; it is an admission that the returns are not expected within the period any commercial lender would tolerate. These arrangements cannot establish that Beijing has chosen the right industries, but they do show a state constructing financial instruments capable of keeping them alive while success remains uncertain.
Whatever the quality of its choices, the shift becomes more powerful as local finance weakens. Municipalities once converted ambition into bankable projects through land sales and financing vehicles. As those resources diminished, access to central facilities and nationally approved funds became more important. Political recentralization thereby acquired a financial counterpart, allowing Beijing to command a larger share of a narrowing spectrum of action even as the field itself lost vitality.
Greater central control may improve strategic discrimination because the claims produced by different investments are not economically equivalent. Another speculative property development mainly protects an old valuation, whereas a battery plant with poor margins may still train engineers and deepen a supply chain that foreign competitors struggle to reproduce. The difference is in where the return lands: some accrues to the company and some to the state's freedom of action, which is why a project that would fail any commercial test can still be worth financing to a government that is buying something other than profit. The strategic label can nevertheless shelter error as readily as it can recognize value, particularly when provincial governments learn to translate local interests into the vocabulary of national security. An official catalog containing integrated circuits, brain-computer interfaces, embodied artificial intelligence and 6G resembles abundance more than choice, and patient capital becomes dangerous once forbearance is demanded by every company that has exhausted more ordinary explanations for its survival.
Two things lie beyond the reach of any allocation system, however concentrated. The first is demand. During the first seven months of 2026 retail sales rose by only 1.2 percent and fixed investment contracted by 6.7 percent, even as investment in high-technology industries grew by 5 percent and exports by 14 percent. Selected capacity was expanding within an economy increasingly unable to purchase what it made, leaving foreign markets to clear part of the difference — and foreign markets that welcome a cheaper product eventually resist the dependence created by its success, which returns the strain to China as falling margins and forced consolidation.
The second is selection. Beijing can decide that a domestic battery or semiconductor ecosystem must exist without knowing in advance which company deserves to survive. It can choose the terrain and must then rely on competition to disclose what policy cannot. The same authority that concentrates resources can protect incumbents, delay exit and turn industrial policy into cartelization, leaving the state on the right battlefield with too many contestants still under finance.
These limits leave China's trajectory less settled than a simple account of decline would prefer. Growth accelerated to 5 percent in the first quarter of 2026 before slowing to 4.3 percent in the second; July then brought weaker industrial output, retail growth of only 0.6 percent and a deeper contraction in fixed investment, while high-technology investment continued to expand, exports remained strong and profit growth was led by electronics and communications equipment. The indicators have stopped moving together, which means a single national trend line no longer describes the economy. Deterioration has become a divergence among sectors as well as a movement through time.
From that divergence may emerge a state less financially flexible than it was five years ago and more formidable within the systems it has chosen to defend because the two judgments use different denominators. An economist comparing China with its former trajectory sees lost growth and declining returns, while a rival government must compare the concentrated remainder with the resources it can itself bring to bear. The narrowing of Chinese power does not make what remains less consequential.
Narrower power is also less forgiving. Every yuan committed to a failed process is unavailable to repair a provincial bank, and the industries chosen for survival still depend on a society able to finance and buy what they produce. China can trade breadth for hardness, concentrating diminished resources where they promise the greatest economic or geopolitical return. It cannot, however, indefinitely build the next model from the failing revenues of the last.
