Chongqing Rural Commercial Bank: The Deposit Castle of Western China

I. Introduction & Episode Roadmap

Drive two hours west out of Chongqing's neon-lit downtown, past the container terminals on the Yangtze and up into the terraced hills of Yongchuan or Fuling, and the built environment thins out fast. The high-rises give way to low concrete townships, roadside noodle stalls, and mustard-tuber fields. What does not thin out is one particular shopfront. Green signage, a bronze-coloured logo, an ATM vestibule that stays lit all night. In some of these townships it is the only bank branch for twenty kilometres. In many of them it has been the only bank branch for seventy years.

By the end of 2025, 重庆农村商业银行 Chongqing Rural Commercial Bank — CQRCB, or 渝农商行 to domestic investors — operated 1,727 outlets across Chongqing municipality, covering every one of the 37 administrative districts and counties, with 1,422 of those outlets sitting outside the eight urban districts.¹ It had issued 31.45 million debit cards in a municipality of roughly 32 million people.

It held RMB 1.03 trillion of customer deposits against a municipality-wide RMB deposit pool of about RMB 6.04 trillion — a share in the high teens that no other institution in Chongqing comes close to matching.¹² Total assets crossed RMB 1.666 trillion, roughly USD 234 billion, making it the largest rural commercial bank in China and, per The Banker's global ranking, the 112th largest bank in the world.¹

That is the fortress. And the fortress is real. But the interesting question — the one worth two hours of anyone's attention — is not whether CQRCB dominates its home territory. It obviously does. The question is what that dominance is actually worth in an era when Chinese loan yields fall faster than deposit costs, when the borrowers most eager for credit are the local-government-linked entities that also sit on the bank's shareholder register, and when the bank's own five-year fee-income line has been cut in half.

There is a second number that frames the whole story, and it is the one that never makes the headlines. In 2025 the bank earned a return on net assets of 9.15% — respectable for a Chinese regional bank in a compressing rate environment, and the fifth consecutive annual decline.¹ A franchise this dominant, in a municipality growing faster than the national economy, generates a single-digit return on the equity its owners have committed. Understanding why is the analytical work of this piece, and the answer is not incompetence.

The moat paradox, and where the consensus story breaks. The standard bull framing goes like this: a bank with the densest rural branch network in western China must enjoy the cheapest funding in western China, and cheap funding is the closest thing to a permanent advantage that commercial banking offers. The first half of that sentence is true. The second half does not survive contact with the filings.

CQRCB's 2025 cost of customer deposits was 1.45%, down 28 basis points year on year — a genuine achievement in a repricing environment.¹ But 中国邮政储蓄银行 Postal Savings Bank of China, which runs the only rural distribution network in China denser than CQRCB's, funded itself at roughly 1.15% in the same year, and several of the mega state banks also printed deposit costs below 1.5%.³

CQRCB is cheap relative to a city commercial bank. It is not cheaper than the national giants. Its funding advantage is regional, not absolute — and the reason is structural: 74.27% of its customer deposits are time deposits, because rural savers in Chongqing are yield-seeking savers, not transaction-account users.¹ A deposit castle built out of one-year term deposits is a very different asset from one built out of current accounts.

Nor is the market-share claim quite what it appears. CQRCB is unambiguously the largest deposit-taker in Chongqing. But its share of the municipality's loans is visibly smaller than its share of its deposits, and that gap is not a rounding artefact — it is the defining economic feature of the business, and it means the institution is better described as western China's largest savings aggregator than as its largest lender.¹²

That single correction reframes everything downstream, so it is worth stating the spine of this piece up front. Why CQRCB wins from here: it holds an irreplaceable physical and social distribution position in a municipality with a national-priority growth mandate, it has been demonstrably better than its peer group at repricing liabilities downward, and its earnings are cushioned by one of the largest provision stocks in Chinese banking.

What could break the case: its pre-provision operating profit has been shrinking for four years while reported profit rose, its asset growth is increasingly funded by capital it is consuming faster than it earns, its retail loan book is deteriorating even as headline asset quality improves, and its regulator has twice in four years fined it — including for concealing non-performing loans and misstating the very provision coverage ratio the bull case leans on.

One more framing note before the story starts, because it governs how everything below should be read. Chinese bank disclosure is unusually rich — CQRCB's 2025 annual report runs to hundreds of pages of segment, industry, maturity and geography breakdowns, audited under both Chinese Accounting Standards and IFRS, with KPMG Huazhen and KPMG each issuing unqualified opinions.¹ The data is there. What is frequently missing from the commentary around these banks is the willingness to read the tables against the narrative. Almost every uncomfortable fact in this piece comes from the bank's own filings; the work is in noticing which numbers contradict which sentences.

Here is the roadmap. First, the origin: how a network of 1951-vintage credit cooperatives, crushed under legacy bad debt, was swept into a single legal entity in 2008. Second, the capital-markets story: the 2010 Hong Kong breakthrough and the 2019 Shanghai listing that went from a 27% first-day pop to breaking issue price inside ten trading days. Third, the economics of the deposit engine — what it actually earns, and where the money goes when there are not enough good loans to make.

Fourth, the sidecars: a financial leasing arm that quietly out-earns the wealth management arm five to one, and a 30% stake in a Xiaomi-branded consumer finance venture. Fifth, a governance audit of a bank with no controlling shareholder, a fourteen-month chairman vacancy, and a habit of exchanging executives with its own largest holders. Sixth, the strategy frameworks — Helmer and Porter, applied honestly rather than flatteringly. Then the bear and bull cases, stress-tested against the bank's own record, and the two or three numbers that will actually tell you whether the thesis is working.

Start where the bank started: with nothing, and with a lot of bad loans.

↑ Back to Top

II. History Part I: From Mao-Era Cooperatives to the 2008 Big Bang

In 1951, in the villages surrounding what was then a wartime-scarred river port, the new government began organising 农村信用社 rural credit cooperatives — small, member-owned pools of agricultural savings intended to keep farmers out of the hands of informal moneylenders and to funnel household surpluses toward collective production.⁴ They were less banks than administrative instruments. There was no credit committee in any meaningful sense, no risk pricing, no capital in the modern regulatory sense. There was a cashier, a ledger, and a local Party secretary who had views about who deserved a loan.

It is worth pausing on what a rural credit cooperative actually was, because the term flatters the institution. A modern bank has three separations built into it: between the people who originate a loan, the people who approve it, and the people who own the downside if it goes wrong. A 1950s Chinese credit cooperative had none of them. Deposits came from villagers who were, nominally, member-owners, and loans went to those same villagers and to collective enterprises, on terms set by whoever held local authority.

There was another absence, and it was the decisive one: no capital buffer in any recognisable sense, because "capital" in a member cooperative is just the members' money under a different name. When a loan went bad, the loss was socialised across savers who had no idea it had happened. The system worked as a savings mobilisation device and failed as a credit institution, which is precisely what one would expect from its design.

For four decades this arrangement was tolerable because it was not really being asked to work as finance. It became intolerable in the 1990s. As China's rural economy commercialised, the cooperatives were pushed to lend like banks while retaining the governance of village committees. The results were predictable and, across the country, catastrophic. Loans went to township-and-village enterprises on political instruction and did not come back.

Each county's cooperative was a separate legal entity with its own balance sheet, meaning a bad harvest or a failed local factory could impair an entire institution with no group-level capital to absorb it. Chongqing, which became a directly administered municipality in 1997 and inherited 38 districts and counties of this fragmented machinery, was a particularly acute case: an urban core with real industry, ringed by a vast rural hinterland whose credit institutions were individually undercapitalised and collectively unmanageable.

Chongqing's specific version of the problem had a geographic dimension that most provinces did not share. The municipality is enormous — comparable in area to a mid-sized European country — and vertically extreme, with mountain counties whose economies had almost nothing in common with the Yangtze-side industrial districts. A credit institution serving Wushan in the far northeast and one serving Jiangjin in the southwest were exposed to entirely different economies, entirely different political establishments, and entirely different collateral. Consolidating them would later prove to be a genuine diversification benefit. In the 1990s it just meant the problem was 39 problems.

The fix came from the top. In 2003 the State Council launched its pilot programme for deepening rural credit cooperative reform, and 中国人民银行 the People's Bank of China began issuing special central bank bills to swap out accumulated losses and non-performing loans at participating institutions — RMB 170 billion of them nationally over the life of the programme.⁵ Chongqing was among the first batch of pilot provinces.⁴ The mechanics were, in effect, a state-funded balance sheet cleanup: the central bank absorbed historical losses in exchange for institutions agreeing to restructure into modern joint-stock companies with real capital, real boards, and real prudential supervision.

There is a useful analogy for what the special bills did. Imagine a chain of village pawnshops, each with a drawer full of pledges that will never be redeemed, and each individually insolvent once you mark those pledges honestly. The central bank's offer amounted to this: hand over the bad drawer, and receive a claim on the central bank in exchange — but only on condition of becoming a real company, with a real board, real auditors, and a regulator empowered to shut it down. It was recapitalisation conditional on institutional reform, and the conditionality was the point. Institutions that took the money and did not reform lost the benefit.

What Chongqing did with that opening was more radical than most. Rather than converting county cooperatives into dozens of small independent rural commercial banks — the approach taken across much of eastern China, and one that left provinces such as Jiangsu and Shandong with a long tail of subscale lenders — Chongqing consolidated everything.

On 27 June 2008 the Chongqing Rural Credit Cooperative Union and 39 district and county rural credit unions were combined into a single joint-stock company holding municipality-wide legal personality — one of the first province-level rural commercial banks in China, and the first in the west.¹⁴

The strategic logic of that choice deserves emphasis, because it was contrarian at the time and it explains most of what CQRCB became. Fragmenting the county cooperatives into dozens of independent rural commercial banks would have preserved local autonomy and local political relationships — the path of least resistance, and the one that most provinces took.

It also would have produced dozens of institutions individually too small to build a risk management department, too small to issue bonds, and far too small to list. By consolidating, Chongqing traded local control for scale, and scale is what made every subsequent option available: the Hong Kong listing, the Shanghai listing, the leasing licence, the wealth management licence, and the ability to absorb the losses of any single county.

The accounting treatment of that moment is a small detail that says a great deal. Under IFRS, the restructuring could not be booked as a merger of entities under common control, because no single party controlled the 39 cooperatives, the union, and the new bank both before and after. It had to be accounted for as an acquisition of a business by the new entity.¹ Legally and economically, CQRCB did not evolve out of the cooperatives. It bought them.

That distinction mattered because it forced fair-value recognition of what was being acquired, and what was being acquired was a portfolio with substantial embedded losses. The bank has never disclosed a clean pre-restructuring non-performing loan ratio for the Chongqing system, and the honest answer is that the figure is not in the public record. What is in the public record is the shape of the sector: China's rural commercial banks collectively still reported a 3.94% non-performing ratio at the end of 2008, years into the cleanup, at a time when the large state banks had already been recapitalised down to low single digits.⁶ The starting point was ugly.

What the 2008 consolidation actually bought. Three things, and it is worth being precise about which of them still matter.

The first was diversification of a fundamentally undiversifiable risk. A single county's cooperative lending to a single county's pig farmers and small foundries has no way to survive a local shock. A municipality-wide bank pooling 38 counties plus an urban core can. This is the least glamorous and most durable benefit of the restructuring, and it is why CQRCB's balance sheet has been able to absorb a 47-basis-point deterioration in retail asset quality during 2025 without the headline ratio moving in the wrong direction.¹

The second was a single credit-approval architecture, which took the loan decision out of the village and put it in a head office. This is also the thing that has repeatedly failed. The bank's regulatory record, examined later in this piece, shows enforcement actions in both 2022 and 2025 centred on exactly this: negligent loan review and approval, inadequate post-lending management, failure to implement unified credit management.⁷⁸ Building the architecture and making it work turned out to be different projects.

The third was a licence to raise external capital, which a cooperative structure could never have supported. That is what made everything after 2008 possible — and it is where the story turns from institutional plumbing to capital markets.

↑ Back to Top

III. History Part II: Capital Markets Trailblazer & A+H Dual Listing

In late 2010, a delegation from a bank most international investors had never heard of walked into fund manager meetings in Hong Kong, Singapore and London carrying a pitch that sounded, on its face, like a stress test of investor credulity: buy shares in a two-year-old institution assembled from Chinese village credit cooperatives, in an inland municipality, whose loan book was concentrated in agriculture and small enterprise, at a moment when global investors were already nervous about Chinese bank asset quality following the 2009 stimulus lending boom.

The timing was not accidental. 2010 was the tail end of a window in which Chinese banks could raise substantial equity from international investors on the strength of a growth story rather than a returns story. The 2009 credit stimulus had inflated bank balance sheets across China by roughly a third in a single year, and every institution that had participated needed capital to support the risk-weighted assets it had just created. Hong Kong absorbed a queue of Chinese bank offerings in that period. CQRCB's pitch was differentiated within that queue only by being stranger.

It worked, though not spectacularly. On 13 December 2010 CQRCB priced its Hong Kong IPO at HK$5.25 per share, the midpoint of a HK$4.50 to HK$6.00 range, selling 2 billion new shares for HK$10.5 billion, about USD 1.35 billion, with roughly USD 240 million of that taken by cornerstone investors including Nexus Capital, Value Partners, Chow Tai Fook and 富邦金融控股 Fubon Financial Holding.⁹ Morgan Stanley and Nomura ran the books. The shares began trading on the Main Board on 16 December 2010, making CQRCB the first rural commercial bank in China to list anywhere, and the first bank from western China to list in Hong Kong.⁴

The mid-point pricing is the tell. This was not a deal that international institutions fought over; it was a deal that got done. Nearly a fifth of the book was pre-committed by cornerstones, a structure that Hong Kong issuers use precisely when organic institutional demand is uncertain. The scepticism was rational — rural Chinese credit was a genuinely opaque asset class in 2010 — and it was not entirely misplaced.

What the cornerstone structure and the mid-point pricing also bought was a set of shareholders who were not natural long-term holders of a Chinese rural bank. Cornerstone investors accept a lock-up in exchange for allocation certainty; when the lock-up expires, many sell. For CQRCB, the practical consequence was that its Hong Kong shareholder base never developed the kind of concentrated, informed long-only ownership that supports a premium rating. Fifteen years later, the H-shares still trade at a persistent discount to the A-shares, and the great majority of the H-line sits behind a nominee account rather than with identifiable institutional holders.¹

Nine years to the second door. CQRCB filed for a Shanghai listing and then waited. By the time the application reached its decisive stage in late 2018, the China Securities Regulatory Commission had questions, and they were the right questions. The bank's non-performing loan ratio had climbed in three consecutive steps through 2018 — 1.18% at the first quarter, 1.23% at the half, 1.34% at the third quarter — after sitting near 0.98% through the 2015 to 2017 period.

More pointedly, overdue loans at the 2018 half-year stood at RMB 5.92 billion against classified non-performing loans of RMB 4.46 billion, and the regulator wanted to know why loans that were past due were not being classified as bad. Small and micro enterprise non-performing rates had reached 1.88%, above the bank-wide average, and asset impairment charges in the first half of 2018 had jumped 91% year on year.¹⁰

The bank got its approval anyway. On 29 October 2019 it sold 1.357 billion A-shares at RMB 7.36, raising RMB 9.988 billion — at the time the largest IPO ever by a Chongqing-headquartered company — and became the first rural commercial bank in China, and the first bank in western China, with a dual A+H listing.¹¹⁴

Then the market delivered its verdict. The shares closed the first day at RMB 9.35, up 27.04%, valuing the bank at RMB 106.2 billion.¹¹ The pop lasted almost exactly one session. The stock hit limit-down on day two, and on the tenth trading day it broke through the issue price, touching RMB 7.18 intraday and closing at RMB 7.19.[^12] For a Chinese A-share IPO in 2019 — an era when regulated issue pricing made first-day gains close to automatic and breaking issue within a month was almost unheard of — this was an unmistakable signal. Domestic investors, who could actually read the Chinese-language prospectus and knew the Chongqing market intimately, were not buying the growth story at 0.9 times book.

Why did it break? Not because of anything CQRCB did in its first fortnight as a listed A-share. The Chinese banking sector was already trading below book value across the board in late 2019 as investors priced in margin compression and asset quality uncertainty, and a bank IPO priced at roughly one times book — the regulatory convention for Chinese bank offerings, which are effectively barred from pricing below net asset value — was mechanically expensive relative to every comparable already trading.

The regulated pricing floor for bank IPOs meant CQRCB had to come to market at a valuation the secondary market had already rejected for its peers. The break was a valuation event, not a company event. But it was an unusually clear early signal about how domestic capital would value this franchise, and the signal has largely held: the stock has spent most of its listed life below book.

What the two listings actually did. It is tempting to narrate the dual listing as a triumph of institutional maturity. The more useful reading is mechanical: the two IPOs, plus retained earnings, supplied the capital that let total assets compound from a few hundred billion renminbi to RMB 1.666 trillion, and the equity base to grow to RMB 137.6 billion, or RMB 11.58 per share of book value by the end of 2025.¹ Access to permanent capital is what separates a bank that can grow from one that cannot, and CQRCB has not returned to the equity market for fresh common shares since 2019.

But the outline framing that listing capital "insulated" the bank's core capital position needs revising against the most recent evidence, and this is a case where the disconfirming data sits in the same table as the claim. Core Tier 1 capital adequacy fell from 14.24% at the end of 2024 to 12.67% at the end of 2025 — a 157 basis point decline in a single year — while total capital adequacy dropped from 16.12% to 14.46%.¹ The cause is arithmetic: risk-weighted assets grew 16.5% to RMB 1,034.8 billion while total assets grew 9.95%, meaning the bank was not merely growing but growing into denser, higher-risk-weight exposures — corporate loans rather than mortgages and bills.¹

Management's framing at the April 2026 results briefing was that core Tier 1 still sits 517 basis points above the regulatory minimum and that replenishment will prioritise internal generation, with tier-two and perpetual issuance used opportunistically.¹² That is a fair characterisation of the current position. It is not evidence that the buffer is self-sustaining: at 2025's growth rate, the buffer supports a few more years, not an indefinite runway. Whether the bank can fund its stated ambition of a RMB 2 trillion balance sheet without touching common equity is an open empirical question, not a settled one.

There is a related characterisation worth stating precisely rather than loosely. CQRCB has not issued fresh common equity since the 2019 A-share offering. It has, however, continued to raise capital in other forms: debt securities issued grew 9.80% during 2025, and the group's tier-one and total capital include instruments beyond common equity — the gap between the 12.67% core tier-one ratio and the 14.46% total capital adequacy ratio is exactly that.¹ So the accurate statement is that the bank has avoided dilution, not that it has avoided external capital. Those are different disciplines, and only the first one has been demonstrated.

Which brings us to the engine that generates the internal capital in the first place.

↑ Back to Top

IV. Core Business Deep-Dive & Segment Economics

Consider a specific person: a 58-year-old resident of a county town two hours from central Chongqing whose son works in a Guangdong electronics plant and sends money home. She receives her pension through a social security card. She has held an account at the same green-signed branch since it was a credit cooperative and her father banked there. When her son's remittance arrives, she walks it down the street and puts most of it into a one-year term deposit, because a one-year term deposit is what her mother did and because the alternative — an app-based money market fund run by a company in Hangzhou — is not something she is inclined to trust with her savings.

She is the business model. In 2025, CQRCB had issued 7.75 million social security cards, the most of any bank in Chongqing, and made 57.8 million pension and injury-benefit payments totalling RMB 44.4 billion to more than 4.6 million customers.¹ Multiply her by several million and you get RMB 880.4 billion of personal deposits, 85.58% of the bank's total customer deposits, funding a balance sheet in which customer deposits supply roughly two-thirds of all liabilities.¹

What the deposits actually cost, and why the structure matters more than the level. The aggregate cost of customer deposits was 1.45% in 2025. Underneath that number sits a striking bifurcation. Personal demand deposits cost the bank 0.06% — essentially free money, the closest thing to a pure economic rent in retail banking. Personal time deposits cost 1.83%. Corporate demand cost 0.49% and corporate time cost 2.20%.¹

The problem is the mix. Personal demand balances averaged RMB 155.1 billion during the year against RMB 705.5 billion of personal time deposits. Free money is 18% of the retail book; the other 82% is bought at close to two percent.¹ Overall, demand deposits were 25.30% of customer deposits at year-end and time deposits 74.27%, and the time balance grew 10.42% year on year while demand grew only 5.69%.¹ The mix is getting worse, not better, which is exactly what one would expect from a customer base of rural savers responding to falling rates by locking in duration.

The analytical conclusion is uncomfortable for the simple moat story. CQRCB's genuine skill in 2025 was not possessing cheap deposits; it was cutting the price of expensive ones — driving personal deposit costs down 29 basis points by capping high-rate product volumes and pushing term guidance.¹ That is a real management achievement and it is the single largest reason net interest income rose 7.85% in a year when interest income fell.¹

But it is a repricing benefit, and repricing benefits are finite. Once the back book has rolled onto today's rates, the tailwind stops. Management effectively conceded the point in the annual report's investor-concerns section, warning that "the interest rate centre will continue to move down" and that the low-rate environment "will bring great pressure" on margin in 2026.¹

For readers who do not spend their days in bank filings, net interest margin is worth translating. Think of a bank as a shop that buys money wholesale and sells it retail. The cost of deposits is the wholesale purchase price; the yield on loans and securities is the retail sale price; net interest margin is the gross margin on the whole inventory, expressed as a percentage of the money the bank has deployed.

CQRCB's 2025 margin of 1.60% means that for every hundred yuan of earning assets, the bank captured one yuan and sixty cents of gross spread before paying a single salary, writing a single loan loss provision, or opening a single branch.¹ Out of that 1.60 it must fund a branch network of 1,727 outlets. This is why bank margins that sound trivially small are in fact the entire game, and why a 10 basis point move — a tenth of one percent — is the difference between a good year and a bad one.

Where the money goes — and the awkward fact that half of it does not become a loan. At the end of 2025, loans and advances were 47.85% of total assets.¹ Slightly more than half the balance sheet is bonds, interbank placements and other financial investments. This is the single most underappreciated fact about CQRCB: it is a bank in name and, in economic substance, a large captive bond portfolio wearing a branch network.

That is not incompetence; it is arithmetic. County-area deposits are 73.30% of the deposit base, but county-area loans are 50.78% of the loan book, producing a county loan-to-deposit ratio of just 53.68% against 142.91% in the urban districts.¹ The rural franchise raises far more money than the rural economy can productively absorb. The surplus has to go somewhere, and it goes into securities and into the urban corporate book.

The segment disclosure makes the consequence visible. In 2025, retail banking generated RMB 12.14 billion of operating income (42.42%), corporate banking RMB 9.35 billion (32.67%), and financial markets RMB 7.03 billion (24.55%).¹ A quarter of revenue comes from what is effectively proprietary asset management. When bond markets cooperate, that flatters results. When they do not, it does the opposite: other net non-interest income fell 25.69% to RMB 3.06 billion in 2025 on weaker trading gains from fair-value instruments.¹ Investors should treat roughly a quarter of CQRCB's revenue as market-sensitive rather than franchise-driven, and this is not a small caveat.

The digital channel numbers complicate the "rural bank left behind by technology" caricature, and deserve a fair hearing. In the county areas alone, 12.43 million users had opened mobile banking with CQRCB by the end of 2025, a net increase of 488,100 during the year and 79.22% of the bank's mobile banking base.¹

The bank ran 6,276 pieces of self-service equipment, a ratio of 3.63 machines per outlet, and 301 rural convenience financial service centres extending basic services beyond the branch footprint.¹ Digital consumer lending is functioning: the flagship "Yukuai Zhenhao Loan" added RMB 3.602 billion of net balance during 2025 and grid marketing added more than 80,000 new credit customers, with digital operation converting more than 35% of identified potential customers.¹

What that evidence supports is a narrower claim than management's framing: CQRCB has successfully digitised its existing customer relationships. What it does not yet demonstrate is the ability to acquire customers it could not have reached physically — which is the thing that would make digital a growth vector rather than a cost-efficiency programme. The distinction matters because only the second version does anything about the demographic problem discussed later.

The loan book is quietly changing shape. Corporate loans grew 21.5% during 2025 to RMB 435.0 billion while retail loans grew 2.9% to RMB 301.0 billion, shifting the mix from 50.14% corporate to 54.55% corporate in twelve months.¹ A bank whose stated strategy has been 零售立行 — "build the bank on retail" — grew its corporate book more than seven times faster than its retail book.

Look at where the corporate growth went and the picture sharpens further. The two largest corporate industry exposures are leasing and commercial services, at RMB 112.1 billion or 14.06% of total loans, and water conservancy, environmental and public utility management, at RMB 82.3 billion or 10.32%.¹ In Chinese bank disclosure these two categories are the standard housing for 地方政府融资平台 local government financing vehicle exposure — infrastructure and utility construction entities that are commercially structured but whose ultimate repayment capacity depends on municipal fiscal capacity.

Together they are nearly a quarter of the loan book, and together they grew RMB 34.9 billion during 2025, accounting for the bulk of net loan growth.¹ By contrast, direct real estate developer exposure is tiny — RMB 7.2 billion, 0.91% of loans — which is genuinely reassuring about the property cycle and genuinely irrelevant to the LGFV question.¹

Yields tell you who has the pricing power. The average yield on total loans fell from 3.92% to 3.64% in 2025.¹ Within that, corporate medium- and long-term loans yielded 4.02% while retail medium- and long-term loans — mostly mortgages — yielded 3.67%.¹ Read that twice. The bank earns more lending to Chongqing infrastructure entities than to its own mortgage customers. The conventional wisdom that retail lending is the high-margin business and corporate lending the commoditised one has inverted, at least here, because successive 贷款市场报价利率 Loan Prime Rate cuts and the 2023–2024 nationwide mortgage repricing hit household loans hardest.

The inclusive finance book is where the policy mandate and the commercial book meet, and the scale is substantial: nearly 210,000 inclusive small and micro enterprise loan customers with balances above RMB 144.0 billion, agricultural loans of RMB 265.5 billion, and first place in Chongqing on the scale of both inclusive small and micro lending and agricultural lending.¹ CQRCB describes its support for the real economy as a "54321" pattern — roughly a fifth of Chongqing's manufacturing loans, a quarter of its inclusive small and micro loans, a third of its agricultural loans, and half of its farm household loans.¹

Read that as a statement about obligation as much as opportunity. A bank that provides half of a municipality's farm household credit is not choosing that exposure opportunistically each year; it is the designated provider, and the designation comes with expectations about pricing and continuity that a purely commercial lender would not accept. That is the real cost of the franchise, and it does not appear as a line item anywhere in the income statement.

Competitive position, stated honestly. CQRCB reports that it ranks first among banks in Chongqing on both the stock and the flow of deposits and loans, and first on manufacturing loans, inclusive small and micro loans, agricultural loans, farm household loans, social security card issuance, and consumer credit growth.¹ It does not disclose a precise market share percentage, and any specific figure circulating in secondary sources should be treated as an estimate.

Against the municipality's RMB 6.04 trillion RMB deposit pool and RMB 6.39 trillion loan pool at end-2025, the bank's RMB 1.03 trillion of deposits and RMB 797 billion of loans imply a materially larger share of local savings than of local credit — which is precisely the funding-surplus problem described above, expressed as a market-share gap.²¹

Its nearest local listed rival, 重庆银行 Bank of Chongqing, is roughly half the market capitalisation and competes primarily in the urban districts where CQRCB's own loan-to-deposit ratio already exceeds 140% — meaning the urban battleground is one CQRCB fights as a lender using rural money, not as a deposit incumbent. The national giants — 中国工商银行 ICBC, 中国建设银行 CCB, 中国农业银行 ABC — have the corporate relationships and the cheaper funding, but not 1,422 county outlets.

So the competitive picture is genuinely favourable on distribution and genuinely mixed on economics. Which raises the question of what else the bank owns.

↑ Back to Top

V. Hidden Assets & Capital Allocation: Leasing & Wealth Management

Every large Chinese bank accumulates subsidiaries. Most of them are regulatory artefacts — entities created because a licence became available or because a peer had one. Occasionally one turns out to matter. CQRCB has three worth examining, and the ranking of their contributions is the opposite of what the strategic narrative would predict.

The leasing arm, which is the real profit centre. 渝农商金融租赁 CQRC Financial Leasing was established in December 2014 with registered capital of RMB 2.5 billion; the bank owns 80%.¹ By the end of 2025 it had grown to RMB 77.3 billion of total assets and RMB 9.5 billion of net assets, and it earned RMB 1.436 billion of net profit during the year.¹

Put that in proportion. The leasing subsidiary alone generated net profit equal to roughly 11.6% of the group's RMB 12.42 billion, on a balance sheet under 5% of the group's size. Whatever else is true, this is the highest-return deployment of capital the bank has made outside its core lending book.

Why does it work? Because financial leasing does something a rural bank cannot easily do through a loan: it takes title to the asset. When a county government's cultural tourism project or a mid-sized agricultural processor needs equipment finance, a lease gives the lender legal ownership of the machinery rather than a security interest that must be enforced through a county court.

In a jurisdiction where collateral enforcement is slow and politically textured, ownership is a materially better position than a lien. It also allows longer-dated, higher-yielding structures than the LPR-benchmarked loan book. The unit's Chongqing lease balance stood at RMB 20.0 billion with 78.12% in county areas, and it deployed RMB 9.1 billion of new Chongqing projects during 2025.¹

The honest caveat: leasing books are where credit problems surface last. A lease that is being restructured looks current until the asset is repossessed. The subsidiary does not disclose a separate non-performing ratio in the group annual report, and investors should treat the absence of that disclosure as a gap rather than as reassurance.

The wealth management arm, and the thesis that has not converted. 渝农商理财 CQRC Wealth Management was established in June 2020 with RMB 2 billion of registered capital, wholly owned, and holds the distinction of being the first wealth management subsidiary set up by a rural commercial bank in China and the first by any corporate bank in western China.¹

The strategic logic is elegant and widely repeated: the bank sits on RMB 764 billion of time deposits paying savers under 2%; if even a modest slice migrates into wealth management products, the bank converts an interest expense into a fee income stream, defends against disintermediation by Alipay and its peers, and improves return on capital because managed products consume almost no balance sheet.

Now test that thesis against the record, because this is where certification and commercialisation diverge. The subsidiary's product balance grew 25.36% during 2025 to RMB 172.8 billion, with more than 139 external sales agencies distributing its products, and inclusive-oriented products exceeding RMB 69 billion, about 40% of the total.¹ Impressive growth. And in the same year, group net fee and commission income fell 19.71% to RMB 1.294 billion, with agency and fiduciary service fees down RMB 107 million.¹ Fee income is now 4.52% of operating income, down from 5.71% a year earlier.¹

Widen the lens and it gets starker. Net fee and commission income was RMB 2.724 billion in 2021. By 2025 it was RMB 1.294 billion — a decline of more than half over four years, during which the wealth management subsidiary was launched, scaled, and celebrated.¹ The subsidiary itself earned RMB 302 million of net profit on RMB 3.155 billion of total assets in 2025 — respectable in isolation, less than a quarter of what the leasing arm earned, and nowhere near enough to offset the erosion elsewhere.¹

The verdict has to be calibrated rather than dismissive. The claim that CQRCB's wealth management platform is a proven engine for converting rural deposits into fee income is rejected by the record to date. A narrower claim survives: the platform has demonstrated it can gather assets, and asset gathering is a necessary precondition for fee generation even if it has not yet been sufficient.

Part of the shortfall is industry-wide — the 2018 asset management rules eliminated guaranteed-return products, and falling rates have compressed management fees across every Chinese bank WMP subsidiary. But an industry-wide headwind is an explanation, not a defence, and it does not make the optionality more valuable. The falsifiable test going forward is simple: net fee and commission income must inflect upward while WMP balances grow. Balance growth alone tells you nothing.

One more observation on the leasing arm that bears on capital allocation quality. The bank owns 80% of it, not 100%, meaning roughly a fifth of that RMB 1.436 billion of profit accrues to minority holders — visible in the group's RMB 2.169 billion of non-controlling interests, which grew 13.26% during 2025, faster than equity attributable to the bank's own shareholders grew at 4.11%.¹ The bank's best-returning business is the one where its shareholders own the least. That is not a scandal; the minority partner presumably brought something at formation. It is, however, a small structural leakage worth knowing about when comparing group return on equity to the underlying economics.

The venture nobody talks about. CQRCB holds 30% of 重庆小米消费金融 Chongqing Xiaomi Consumer Finance, the second licensed consumer finance company in Chongqing, established in May 2020 with RMB 1.5 billion of registered capital. At the end of 2025 it had RMB 20.8 billion of total assets and earned RMB 131 million of net profit on an unaudited basis.¹

This is a genuinely interesting position: an equity-accounted stake giving a rural bank exposure to nationwide app-based consumer lending distributed through 小米 Xiaomi's device ecosystem — precisely the demographic and channel its branch network cannot reach. It is also, at RMB 131 million of profit against a 30% share, immaterial to earnings today. The right way to hold it analytically is as a cheap option with a real underlying, not as a growth driver. It is worth watching for a different reason: consumer finance companies are where retail credit stress shows up early, and the bank's own retail non-performing ratio is already rising.

The out-of-province experiment. The third subsidiary group is the one that tests whether CQRCB's model travels. The bank has established 12 village and township banks across 12 counties and districts in five provinces, holding at least 51% of each, with aggregate registered capital of RMB 1.662 billion.¹ At the end of 2025 they held RMB 4.924 billion of total assets, RMB 2.489 billion of deposits and RMB 4.194 billion of loans, and earned RMB 24 million of net profit with a 1.55% non-performing ratio and 325.98% provision coverage.¹

The numbers tell the story clearly enough. After more than a decade of operation, the entire out-of-province network is worth about two-tenths of one percent of group net profit, holds a non-performing ratio nearly 50% higher than the group's, and lends substantially more than it gathers in deposits — the exact inverse of the parent's funding profile, and a reminder that the deposit franchise does not transplant. The bank's genuinely successful geographic extension has been a single off-site branch in neighbouring Yunnan, not the village bank network.¹

The investor takeaway is not that these entities are dangerous — they are far too small for that — but that they falsify a specific growth story. CQRCB's advantage is not a replicable operating model that can be exported county by county. It is a seventy-year-old position in one specific place. Any thesis that assumes geographic expansion as a growth vector should be discounted heavily on this evidence.

Capital returned, honestly measured. For 2025 the bank paid an interim dividend of RMB 2.0336 per 10 shares and proposed a final of RMB 1.1755 per 10 shares, RMB 3.2091 per 10 shares in total — RMB 0.32091 per share, RMB 3.645 billion in aggregate, a 30.05% payout of profit attributable to shareholders.¹¹³ Over five years the bank has distributed roughly RMB 15 billion with payout consistently at or above 30%, and it moved to a twice-yearly distribution schedule beginning in 2024.¹¹²

On the yield, precision matters more than enthusiasm. At the closing price around the March 2026 results date, the A-share dividend yield was 4.49%.¹³ With A-shares near RMB 6.78 and H-shares near HK$6.365 in early September 2026, the A-share yield sits under 5% and the H-share yield — reflecting the persistent discount at which the H-line trades — somewhat above it, before withholding tax.¹⁴¹⁵ These are solid income characteristics for a bank trading below book. They are not the 6% to 7.5% figures that circulate in older secondary commentary, and an investor underwriting this as a high-yield instrument should use the current numbers, not the legacy ones.

The deeper capital-allocation question is not the payout ratio but its interaction with the capital consumption described earlier. Paying out 30% while risk-weighted assets grow 16.5% is arithmetically sustainable only if either growth slows or the payout does. Management has committed to maintaining both.¹² One of those commitments will eventually have to give, and which one gives will tell you a great deal about whose interests the board actually optimises for.

↑ Back to Top

VI. Strategy, Management & Governance Audit

For fourteen months, China's largest rural commercial bank did not have a chairman.

谢文辉 Xie Wenhui departed in October 2024, transferred out to become party secretary and chairman of Chongqing Yufu Holding Group — the parent of 重庆渝富资本运营集团 Chongqing Yufu Capital Operation Group, which happens to be the bank's largest domestic shareholder at 8.70%.¹⁶¹ President 隋军 Sui Jun absorbed the chairman's duties and the legal representative role on top of running the bank. In April 2025 the board nominated 刘小军 Liu Xiaojun as party secretary and executive director candidate. Regulatory approval from the 国家金融监督管理总局 National Financial Regulatory Administration Chongqing office did not come through until January 2026.¹⁶¹

An interregnum of that length at a systemically important regional bank is a governance fact, not a footnote. It is also unremarkable by the standards of Chinese state-linked financial institutions, where senior appointments require Party organisational approval and regulatory qualification review in sequence. The relevant investor question is whether it produced drift, and the operating record suggests it largely did not — 2025 delivered the bank's fastest asset growth in four years and its fifth consecutive year of declining non-performing ratio.¹ Sui Jun ran the institution through the gap competently.

Who runs it now. Liu Xiaojun is 49, born in December 1976, holding a master's in economics from 复旦大学 Fudan University earned in 2002. His career reads unusually for a rural bank chairman: four years at China Construction Bank's head office in international business and real estate finance, then eighteen years at 中信信托 CITIC Trust rising to vice president, then — in January 2024 — chairman of 重庆发展投资 Chongqing Development Investment Co., Ltd.¹⁶¹

Two observations follow. First, this is a capital markets and structured finance executive, not a branch banker. Someone who spent eighteen years in Chinese trust products understands non-standard credit, off-balance-sheet structuring, and how local government financing actually gets done. Given that a quarter of CQRCB's loan book sits in LGFV-adjacent categories and a quarter of its revenue comes from the financial markets desk, that skill set is arguably better matched to the bank's actual balance sheet than a career agricultural lender would be.

Second, and less comfortably: Chongqing Development Investment is itself a top-ten shareholder of the bank, holding 4.60%, and its acting-in-concert party Chongqing Development and Real Estate Management holds a further 5.19%.¹ The chairman arrived from a shareholder. His predecessor departed to the parent of the largest shareholder. Executive talent circulates within the Chongqing municipal state apparatus, and the bank is one node in that network rather than an entity standing apart from it.

Sui Jun, the president, provides the operating counterweight: a professor-level senior economist with a 2020 doctorate from 西南财经大学 Southwestern University of Finance and Economics, who served as vice president of Bank of Chongqing and chairman of Chongqing Automotive Finance before returning to CQRCB, and whose earlier career included running Jiangjin sub-branch and the Jiangjin rural credit union.¹

He knows the county network from the inside. Combined 2025 remuneration was modest by any international standard — RMB 420,700 for Liu and RMB 608,400 for Sui — which removes pay-driven risk-taking as a concern and, equally, removes equity-linked incentive alignment as a mechanism.¹ Sui held 75,400 shares personally.¹ Managers here are stewards of state assets, not owner-operators, and investors should price the incentives accordingly.

The control question, corrected. It is frequently asserted that Chongqing SASAC is CQRCB's ultimate controlling shareholder. The bank's own disclosure says otherwise, explicitly: "The shareholding structure of the Bank is diversified and no controlling shareholders and actual controller existed," with no shareholder holding voting rights sufficient to materially influence general meeting resolutions.¹

Technically accurate. Economically, the register tells a more textured story. Behind HKSCC Nominees' 22.08% custodial position for H-shares sit Yufu Capital at 8.70%, Chongqing City Construction Investment (Group) at 7.02%, Chongqing Development and Real Estate Management at 5.19%, Chongqing Water Conservancy Investment Group at 4.99%, Chongqing Development Investment at 4.60% and Chongqing Water Group at 1.10% — with Water Conservancy and Water Group disclosed as acting in concert with Yufu Capital, and Development and Real Estate acting in concert with Development Investment.¹ Aggregate the disclosed concert parties and the municipal state bloc is comfortably the dominant voting force even though no single entity controls.

Now overlay that on the asset side. The bank's largest corporate industry exposures are leasing and commercial services and water conservancy, environmental and public utility management.¹ Its shareholders include a city construction investment group, two water utilities and a water conservancy investment group. A skeptical investor is entitled to ask whether the largest customers and the largest owners are drawn from the same pool, and whether credit decisions involving them receive the arm's-length scrutiny that decisions involving unaffiliated borrowers would.

The bank states there was no misappropriation of funds by controlling shareholders or related parties for non-operating purposes, and KPMG Huazhen and KPMG each issued unqualified opinions on the Chinese-standards and IFRS financial statements respectively.¹ Those are meaningful assurances. They are not the same as an assurance that related-party credit is priced and underwritten identically to third-party credit, and no such assurance exists in the disclosure.

Credibility, tested against outcomes. Two enforcement actions bear directly on how much weight to put on management's asset-quality narrative, and they belong here rather than in a distant risk section, because they test the specific claim that CQRCB's reported credit metrics are reliable.

In November 2022 the Chongqing Banking and Insurance Regulatory Bureau fined the bank RMB 12.85 million for nine violations. The list included negligent review and approval producing excess working capital lending; concealing non-performing loans; a false provision coverage ratio indicator and insufficient loan loss reserves; improper lending to government financing platforms without adequate collateral review; inadequate interbank credit due diligence; non-compliant interbank investment operations; poor post-lending management with misused credit funds; and failure to implement unified credit management. Six individuals received warnings, two with RMB 50,000 fines.⁷

In December 2025 the NFRA's Chongqing bureau fined the bank a further RMB 8.7 million for inadequate loan "three checks," insufficient investigation and accountability for major risk losses, off-site statistical data errors, failure to verify guarantee fund sources, and rolling billing. Former vice president 舒静 Shu Jing and former chief loan officer 封洪伟 Feng Hongwei were banned from the banking industry for life. Shu Jing had already been expelled from the Party and removed from office in August 2024 following an investigation into misuse of lending authority and illegal financial gains.⁸

Note that Feng Hongwei appears in both actions — warned in 2022, permanently barred in 2025. That is not two isolated incidents; it is one problem surfacing twice, three years apart.

The calibrated conclusion: the specific bull-case claim that CQRCB's 1.08% non-performing ratio and 367% provision coverage constitute empirical proof of credit discipline is narrowed, not rejected. Narrowed, because a regulator has formally found that this institution previously understated non-performing loans and misstated its provision coverage ratio, which means these particular metrics carry documented institutional history of manipulation and cannot be taken at face value.

Not rejected, because the 2022 findings predate the current management regime, the loan book has since visibly de-risked on the corporate side, and — importantly — the bank is now disclosing deteriorating retail asset quality that it could have chosen to smooth, which is weak but genuine evidence of improved candour.¹ The confirming or falsifying evidence to watch is straightforward: whether overdue loans stay below classified non-performing loans, which is where the 2018 CSRC questions and the 2022 penalty both landed. At the end of 2025 the overdue ratio was 1.19% against a non-performing ratio of 1.08% — overdue still exceeds impaired, which is normal for Chinese banks but is precisely the gap regulators have twice probed here.¹

One further governance observation belongs here because it is easy to miss in the board disclosure. Three of the bank's five non-executive directors — Ma Bao, Dong Bin and Yuan Gang — took their seats in December 2025, all three drawn from shareholder entities: Yufu's private equity arm, Chongqing City Construction Investment (Group), and Chongqing Development Investment respectively.¹ Against seven directors in total, including two independents, that is a board where shareholder-nominated non-executives materially outnumber independent voices.¹

The practical consequence is not that decisions will be improper. It is that a board constituted this way is structurally better at representing the interests of the Chongqing municipal state than at representing the interests of a minority H-share holder in Hong Kong, and those interests diverge most sharply on exactly the questions that matter to an outside investor: growth versus returns, policy lending versus commercial lending, and dividend versus retention. Investors should price that as a permanent feature of the security, not as a remediable governance gap.

Narrative consistency. Comparing the 2025 annual report against the April 2026 briefing and the 2026 half-year, management's story has been unusually stable: "three new momentums" of digital, industrial-chain and scenario-driven growth, a pivot from scale-driven growth to balancing "volume, price and risk," and EVA-based internal assessment.¹¹² Guidance for the 15th Five-Year Plan period is specific and therefore testable: assets above RMB 2 trillion, net profit above RMB 15 billion, non-performing ratio around 1%, and return on equity above 10%.¹²

That last target deserves scrutiny. Return on net assets was 9.82% in 2021 and has declined every year since — 9.72%, 9.55%, 9.24%, and 9.15% in 2025.¹ Management is guiding to reverse a five-year deterioration in a falling-rate environment while maintaining a 30%-plus payout. It is not impossible; the 2026 half-year showed pre-tax profit up 11.50% and margin expanding.¹⁷ But it requires the recent margin inflection to be durable rather than a repricing artefact, and investors should treat the ROE target as the most aggressive of the four and the one most likely to be quietly dropped.

↑ Back to Top

VII. The Strategic Playbook: 7 Powers & 5 Forces

Strategy frameworks are most useful when applied adversarially — as a checklist of things to disprove rather than a scaffold for praise. Run CQRCB through both, and the interesting result is that the bank passes the distribution tests decisively and fails or barely passes most of the economic tests.

Hamilton Helmer's 7 Powers.

Cornered Resource — present, and the strongest of the seven. The 1,422 county-area outlets, 301 rural convenience financial service centres, and multi-generational household relationships are not replicable at any sane cost.¹ No commercial rival would build 1,400 branches in Chongqing's counties today; the unit economics of a township branch serving a shrinking, ageing population are terrible for a new entrant and merely mediocre for an incumbent that has already sunk the cost. The Yunnan Qujing branch — the first off-site branch operated by any Chinese rural commercial bank — is a reminder that even geographic expansion is licensed rather than bought.¹ This power is real and durable.

Scale Economies — present but weak, and weakening. The cost-to-income ratio was 31.25% in 2025, down 65 basis points, with operating expenses essentially flat at RMB 9.42 billion against 9.95% asset growth.¹ That is genuine operating leverage. But the base rate matters: 常熟银行 Changshu Rural Commercial Bank, a fraction of CQRCB's size, ran a cost-to-income ratio of 31.68% in the 2026 first half — statistically indistinguishable.¹⁷ If a bank one-quarter your size matches your efficiency ratio, your scale is not producing much of a cost advantage. Physical branch networks scale poorly; the fixed cost is per-branch, not per-bank.

Switching Costs — present, and the most underrated. Not because leaving is technically difficult, but because of what economists would call habit persistence and what everyone else would call inertia. A depositor whose pension arrives on a CQRCB social security card, whose village has one bank, and whose family has banked there for two generations faces a switching cost that has nothing to do with fees. The evidence is in the pricing: the bank cut personal deposit rates 29 basis points in 2025 and personal deposits still grew 8.88%.¹ That is the single cleanest empirical demonstration of pricing power in the entire filing — customers accepted a materially worse deal and added money anyway.

Counter-Positioning — largely absent. This power requires that incumbents cannot respond without damaging their existing business. Nothing prevents ICBC or ABC from underwriting small rural loans; ABC in particular has a national rural mandate and vastly cheaper funding. The reason they underweight the segment is unit economics, not structural constraint, and unit economics change when policy changes. Chinese regulators have repeatedly pushed the large banks down-market into inclusive small and micro lending, and when they do, they arrive with a funding cost roughly 30 basis points below CQRCB's. This is a preference, not a moat.

Process Power — unproven. Management describes a digital risk control three-year plan, an enterprise relationship knowledge graph, an industrial "super brain," seven business-technology integration centres and fifteen deployed intelligent agents.¹ These may become process advantages. Today they are inputs. The output metric — retail non-performing loans rising 47 basis points to 2.07% in the year all this was deployed — does not yet validate the claim.¹

Branding — present locally, worth little economically. The brand supports deposit gathering. It does not command a price premium on loans.

Network Economies — absent. The "Yukuai Hui" merchant ecosystem, with over two million merchants and RMB 275.4 billion of average daily merchant-linked AUM plus loans, is the closest thing.¹ But it is an acquisition channel, not a network with increasing returns; the marginal merchant does not make the platform more valuable to existing merchants in any structural way.

Stepping back from the individual powers, the aggregate reading is that CQRCB holds two genuine powers — cornered resource and switching costs — and both of them act on the same side of the balance sheet. Helmer's framework is explicit that a power must produce differential returns, not merely differential position. CQRCB's powers produce a differential funding position. Whether that translates into differential returns depends entirely on what the bank can earn with the funding, and that is determined by an asset market where it holds no power at all. A moat around the reservoir does not raise the price of water.

Porter's Five Forces.

Threat of new entrants: very low. Banking licences in China are issued by the NFRA and are not available on demand. Constructing a comparable county network would require regulatory approval no authority would grant. This is the most protective of the five forces.

Bargaining power of suppliers — depositors: low, and demonstrably so. Established above by the 2025 rate cut and simultaneous deposit growth. The caveat is that the composition of what depositors will accept is shifting toward term products, which raises the average cost even as headline rates fall.

Bargaining power of buyers — borrowers: moderate and rising, and this is where the pressure actually sits. The loan yield compression from 3.92% to 3.64%, the collapse of retail mortgage yields to 3.67%, and the 0.89% yield on discounted bills all point the same direction.¹ Large corporate and government-linked borrowers in a municipality where every bank wants infrastructure assets have real negotiating leverage, and the PBOC's LPR framework passes policy rate cuts through to borrowers mechanically.

Threat of substitutes: moderate, and older than the framework suggests. Alipay and WeChat Pay hollowed out transaction balances years ago; the residue is visible in demand deposits being only a quarter of the base. The live substitution threat now is not payments but savings products — money market funds, insurance savings products, and third-party wealth platforms competing for the same term deposits. The bank's counter is distribution: 430 new wealth products, 87 fund products, 89 precious metal products and 64 insurance products launched in 2025, plus eight new external product partners.¹ It is a credible defence of the relationship. As shown, it has not yet been a profitable one.

Competitive rivalry: bifurcated. Genuinely intense in urban Chongqing, where CQRCB is the challenger; genuinely mild in the counties, where it is often the only option. The strategic tension is that growth capital is being deployed where rivalry is intense, because that is where credit demand exists.

The synthesis: CQRCB's powers are concentrated on the liability side of the balance sheet and its vulnerabilities on the asset side. It is extraordinarily good at collecting money at a price it dictates, and structurally mediocre at converting that money into high-return assets. That is a real franchise — but it is a franchise whose value rises and falls with the spread between what it pays and what it can earn, and it controls only one of those two variables.

↑ Back to Top

VIII. Bear vs. Bull Case & Skeptical Investor Stress Test

Imagine an activist investor with a meaningful H-share position sitting across from the board. Not hostile — CQRCB is majority-owned by the Chongqing state and an activist knows a proxy fight is unwinnable — but sceptical, prepared, and holding the five-year financial summary. Here is the conversation that would follow.

"Your earnings growth is not coming from your business."

This is the strongest single bear argument and it is entirely visible in the bank's own five-year table. In 2021, operating income was RMB 30.85 billion and operating expenses RMB 8.80 billion, leaving roughly RMB 22.05 billion of pre-provision operating profit. In 2025, operating income was RMB 28.62 billion against RMB 9.42 billion of expenses — roughly RMB 19.20 billion of pre-provision profit.¹ Pre-provision earnings power has fallen about 13% over four years.

Over the same span, profit before tax rose from RMB 11.20 billion to RMB 13.71 billion and net profit from RMB 9.72 billion to RMB 12.42 billion.¹ The entire gap is credit impairment charges, which fell from RMB 10.85 billion in 2021 to RMB 5.44 billion in 2025 — a 50% reduction.¹

The conclusion is unavoidable: CQRCB's reported profit growth over the last four years has been funded almost entirely by declining provision charges, not by improving operations. This is not accounting impropriety — if credit costs genuinely normalise, lower charges are the correct accounting. But it is a finite source of earnings. Provision coverage is already 367%, roughly two and a half times the 150% regulatory floor, and the bank cannot release reserves indefinitely. The bull case requires that the 2026 half-year inflection — operating income up 7.81%, pre-tax profit up 11.50% — marks a genuine return to operating-driven growth.¹⁷ One half-year is a data point, not a trend.

"Your best asset-quality number is the one your regulator says you misstated."

Covered in the governance section, and the board's answer would presumably be that 2022 is history and the current disclosure is candid. The activist's rejoinder writes itself: the bank is reporting a five-year decline in non-performing ratio to 1.08% while its retail non-performing ratio jumped 47 basis points to 2.07% in a single year, corporate non-performing loans fell RMB 1.36 billion, and the aggregate improved.¹

Mix effects can produce that result honestly — corporate loans grew 21.5% and are the cleaner book. But a bank that has previously been penalised for concealing non-performing loans and misstating provision coverage does not get the benefit of the doubt on a favourable mix effect. It gets scrutiny.

The mitigants are genuine and should be weighed. Secured loans account for 77.92% of retail non-performing loans, collateralised and pledged loans 68.00%, with collateral valued at 1.68 times loan principal.¹ Direct real estate exposure is under 1% of loans.¹ The overdue ratio fell 13 basis points to 1.19% and special-mention loans fell to 1.42% — both leading indicators, both improving.¹ This is a book that looks better on the forward-looking metrics than the retail non-performing spike alone would suggest.

"A quarter of your loan book is lending to your own shareholders' sector."

The RMB 194.4 billion in leasing/commercial services plus water conservancy and public utilities, sitting alongside a shareholder register populated by city construction investment and water utility groups, is the structural concern.¹ Chinese local government debt restructuring — the ongoing programme of swapping LGFV debt into explicit municipal bonds — is broadly credit-positive for holders, because it converts implicit obligations into explicit ones.

But it is margin-negative, because swapped debt carries government bond yields rather than commercial loan rates. CQRCB's LGFV-adjacent book grew RMB 34.9 billion in 2025, at yields already compressing.¹ The bank is growing into an asset class where the credit risk is politically managed and the return is politically capped. That may be the safest available lending in Chongqing. It is not lending that will restore return on equity above 10%.

"Half your balance sheet is a bond fund, and you do not report it as one."

The financial markets segment generated a quarter of 2025 revenue, and the disclosure around it is materially thinner than around lending: investors get segment revenue, aggregate financial investment yields, and a maturity profile, but not the position-level detail that would let them assess duration risk or credit composition with any precision.¹ Financial assets measured at amortised cost yielded 3.33% in 2025, down from 3.50%, on an average balance that shrank by RMB 33.1 billion — the bank rotated within its portfolio rather than simply adding.¹

An activist would reasonably argue that a business generating a quarter of revenue and roughly half the asset base deserves the disclosure standard applied to an asset manager, not to a treasury function, and that the 25.69% decline in other non-interest income during 2025 is exactly the kind of volatility that thin disclosure makes impossible to underwrite.

"Your customers are getting older and fewer."

The demographic argument against rural Chinese banking is real and the bank's own disclosure inadvertently illustrates it: 78.67% of personal deposits sit in county areas, 78.97% of debit cards were issued there, and the bank's pension finance business — 4.6 million customers served, rated "excellent" by the PBOC's Chongqing branch — is one of its proudest franchises.¹

A bank whose deposit base skews toward pension recipients has a very stable funding profile for perhaps fifteen years and a structurally shrinking one thereafter. Rural-to-urban and inland-to-coastal migration compounds it. Nothing in management's strategy addresses this beyond the general hope that Chengdu-Chongqing integration keeps population in the region.

Now the bull case, tested with the same rigour.

The regional mandate is real, and it is unusual. Chongqing's GDP reached approximately RMB 3.38 trillion in 2025, growing 5.3% against a national 5.0%, and the municipality is the anchor of the 成渝地区双城经济圈 Chengdu-Chongqing Twin-City Economic Circle, a top-tier national development strategy.¹

CQRCB's exposure to that is concrete rather than rhetorical: RMB 33.3 billion of loans to 193 municipal major projects, RMB 71.1 billion of financing for the New International Land-Sea Corridor — the largest of any local corporate bank — RMB 65.9 billion into the "33618" manufacturing cluster programme, RMB 90.9 billion of technology-enterprise loans and RMB 82.8 billion of green loans.¹ A bank that is the designated local financing arm for a national strategic corridor has visible credit demand that most Chinese regional banks lack. The caveat: policy-directed lending is volume without pricing power, which is precisely the problem identified above.

The margin has inflected, and it inflected better than peers. Net interest margin was 1.60% in 2025, down just one basis point after an 11 basis point drop the prior year, and rose six basis points to 1.66% in the 2026 first half, driving net interest income up 15.20%.¹¹⁷ Context matters here: the average NIM across A-share listed banks in the 2026 first half was about 1.52%, with listed rural commercial banks averaging 1.54% and the large state banks 1.35%.¹⁷

CQRCB sits above both. It sits far below Changshu's 2.48%, which is the reminder that CQRCB is a low-margin, high-volume institution rather than a high-margin niche lender — but relative to the peer group that actually competes with it for capital, the margin performance is credible.¹⁷

Asset quality compares well within the peer set. The 2026 half-year non-performing ratio of 1.05% with 357.47% coverage places CQRCB better than 青岛农商银行 Qingdao Rural Commercial Bank at 1.74% and 紫金银行 Jiangsu Zijin Rural Commercial Bank at 1.33%, worse than 常熟银行 Changshu Rural Commercial Bank at 0.75% and 沪农商行 Shanghai Rural Commercial Bank at 0.94%.¹⁷ Middle of the pack, improving, with a very large reserve cushion.

The valuation embeds substantial scepticism. With book value per share at RMB 11.58 and A-shares near RMB 6.78 in early September 2026, the stock trades at roughly six-tenths of book; the H-shares, near HK$6.365, trade lower still.¹¹⁴¹⁵ Earnings per share were RMB 1.05 in 2025.¹ Whatever one concludes about the operating trajectory, the market is not pricing in a growth story. It is pricing in a persistent below-cost-of-equity return and some probability of credit deterioration. The bull case does not require the bank to be excellent; it requires the bank to be less bad than the price implies, while paying a 30%-plus payout ratio out of a 9% return on equity.

Second-layer items worth noting. The bank's shares were added to the CSI 300 Index and the MSCI China H Index during 2025, and it holds an "A" rating in MSCI's ESG assessment, both of which structurally broaden the passive shareholder base.¹ It obtained a securities investment fund custody licence in 2025 — the first Chongqing corporate bank to do so — which is a genuine new fee pool but one that has not yet generated revenue and should be treated as a milestone, not a business.¹

On the credit side, the group's total loan loss allowance stood at RMB 31.39 billion at the end of 2025 against RMB 30.44 billion a year earlier — reserves grew, but more slowly than the loan book, which is why the provision-to-loan ratio fell from 4.28% to 3.96% even as coverage of classified non-performing loans rose.¹ That divergence is the clearest single illustration of how the same reserve stock can be presented as strengthening or weakening depending on which denominator is chosen.

The calibrated verdict. The claim that CQRCB possesses a durable structural funding advantage is narrowed: it is a regional advantage against city commercial banks, not a national one against the mega-banks, and it derives more from customer inertia than from account structure. The claim that its provisioning and asset-quality metrics prove credit discipline is narrowed by documented regulatory findings and should be verified through overdue and special-mention trends rather than the headline ratio.

The claim that wealth management converts deposits into fee income is rejected on four years of evidence. The claim that the branch network is a cornered resource survives intact. And the claim that the bank can hit an ROE above 10% while growing to RMB 2 trillion and paying out 30%-plus remains unproven and internally tense.

↑ Back to Top

IX. 1-3 KPIs That Matter Most

Most bank dashboards contain twenty metrics, eighteen of which are noise. For CQRCB, three carry the analytical weight.

1. Net interest margin, read alongside the cost of customer deposits.

This is the spread between what the bank earns on assets and what it pays for funding, and for an institution deriving three-quarters of revenue from net interest income, it is the business. Both halves matter and they should be tracked as a pair, because they can move for opposite reasons. Falling deposit cost with stable margin means the repricing engine is working. Falling deposit cost with falling margin means asset yields are compressing faster than liabilities can follow — the condition that has defined Chinese banking since 2022. Rising margin with rising deposit cost would signal genuine asset-side pricing power, which CQRCB has not demonstrated.

The reason this KPI sits first is that it is the cleanest test of whether the 2026 inflection is durable or a one-year repricing artefact. Management has told investors that the interest rate centre will keep falling and that maintaining margin will require differentiated pricing and active liability management.¹ The bank has been better at that than its peer group. Whether it can be better at it once the back book has fully repriced is the open question.

2. Retail non-performing loan ratio, alongside the overdue ratio and special-mention ratio.

Not the headline non-performing ratio — that number can improve through corporate mix shift while the underlying household book deteriorates, which is exactly what happened in 2025.¹ The retail line is where genuine household credit stress in Chongqing shows up first, and it moved the wrong way by 47 basis points in a single year.

Pairing it with overdue and special-mention loans matters for a specific reason established earlier: these are the leading indicators, they sit ahead of classification decisions in the reporting chain, and they are precisely where regulators have twice questioned this institution's judgement. If overdue and special-mention ratios keep falling while the retail non-performing ratio rises, the deterioration is being recognised honestly and is likely near its peak. If overdue starts rising while the classified ratio stays flat, that is the pattern that drew regulatory attention in 2018 and 2022, and it deserves immediate scepticism.

A note on why the retail line rather than a broader credit metric: CQRCB's corporate book is increasingly weighted toward government-linked borrowers whose credit outcomes are determined by fiscal policy rather than by underwriting quality. Those exposures will not deteriorate in a way that a quarterly ratio captures; they will be restructured, extended, or swapped into municipal bonds. The retail book is the part of the balance sheet where ordinary credit dynamics still operate, and therefore the part where the bank's actual underwriting skill is observable.

3. Net fee and commission income.

The narrower third metric, and the one that will settle the biggest strategic argument. Every element of the bank's transformation story — wealth management, custody, the merchant ecosystem, the pivot from a "financial network operator" to a "scenario ecosystem operator" — is ultimately a claim that CQRCB can earn money without lending it. That claim has a single financial expression, and it has been falling for four years while every input metric improved.¹

Wealth management balances, AUM, merchant counts and card issuance are inputs. Fee income is the output. Until the output line inflects, the transformation is an expense, not a business. This is a low bar and it is the right bar.

A practical note on cadence. CQRCB reports on a quarterly basis for the A-share market and semi-annually in full detail for the Hong Kong market, which means the margin and cost-of-deposits pair is observable twice a year with full disclosure and roughly indicated quarterly, while retail-level asset quality granularity generally arrives only with the interim and annual reports. The two half-year data points per year are the ones that matter; quarterly profit prints for a bank with this much reserve flexibility carry less information than they appear to.

Deliberately excluded: total assets, which management is targeting and which measures ambition rather than value creation; provision coverage, which is a policy choice as much as a fact; and any single-quarter profit growth figure, which for a bank with a RMB 31.4 billion loan loss allowance is substantially a management decision.

↑ Back to Top

X. Epilogue & Key Takeaways

The Yunnan Qujing branch is a small thing to end on, but it is telling. It is the first off-site branch any Chinese rural commercial bank has been permitted to operate, sitting a few hundred kilometres outside Chongqing in a neighbouring province.¹ Everything else CQRCB owns — 1,727 outlets, 31 million debit cards, a trillion renminbi of deposits — sits inside one municipality's borders. The bank is, in the most literal sense, the financial infrastructure of a place.

Seventy-five years separate the first village credit cooperative from the institution that reported a RMB 1.79 trillion balance sheet at the 2026 half-year.¹⁷ For most of that span, the enterprise was not investable, not solvent on an honest marking, and not really a bank. The transformation into something that global investors could own took a state-funded balance sheet cleanup, a contrarian decision about legal structure, two public offerings separated by nine years, and a regulatory apparatus that has twice had to fine the institution to enforce standards it had committed to.

That produces the central investing lesson, and it cuts both ways. Localised distribution density and habit-based funding are among the most durable competitive assets in emerging market banking, because they are built out of physical presence and social trust rather than technology or price, and neither of those can be procured quickly. CQRCB proved this in the cleanest possible way in 2025: it cut what it paid rural savers by 29 basis points and they gave it 8.88% more money.¹ Very few businesses anywhere can raise their price and grow their volume simultaneously.

But distribution density determines only what a bank collects. What it earns is set by the economy it lends into, and there the picture is more constrained. Chongqing's credit demand runs toward infrastructure, utilities and government-linked projects — safe, policy-supported, and structurally low-yielding.

The bank collects deposits from a rural base that generates more savings than the rural economy can absorb, and deploys the surplus into an urban corporate market where it is a price-taker and a securities portfolio where it is a price-taker twice over. The moat is on the funding side; the returns are determined on the asset side. That mismatch is why an institution with a genuinely unassailable local franchise earns a nine percent return on equity and trades below book.

There is a second lesson embedded in the seventy-year arc, and it is about the relationship between institutional form and strategic possibility. Nothing about the credit cooperatives of 1951, or the fragmented and impaired county unions of the 1990s, made CQRCB inevitable. What made it possible was a single structural decision in 2008 to consolidate rather than fragment, taken at a moment when the opposite choice was easier and more locally popular.

Every subsequent advantage — the ability to list twice, to hold a leasing licence, to absorb a bad year in one county, to fund itself in the bond market — descends from that one decision about legal form. For investors examining regional financial institutions in emerging markets, the question of who consolidated and who did not is frequently more predictive than any operating metric.

A final note on what this business is for, since that shapes what it can reasonably be expected to deliver. CQRCB is not run primarily to maximise shareholder return, and it has never claimed otherwise. Its chairman's letter opens with Party guidance and closes with commitments to rural revitalisation, green transition and inclusive welfare.¹ It provides half of a municipality's farm household credit and the plurality of its pension payment infrastructure.

Those obligations are not decorative; they are the reason the licence exists and the reason the branch network was never rationalised into profitability. An investor buying this security is buying a claim on the residual returns of a public utility that also happens to be listed — which is a coherent thing to own at the right price, and an incoherent thing to own while expecting it to behave like a commercial bank optimising for return on equity.

The correct way to hold CQRCB analytically is as what it functionally is: a piece of essential regional financial infrastructure, extraordinarily well positioned in a place with a national growth mandate and a shrinking rural population, generating utility-like returns and paying out roughly a third of them, run by capable state-appointed stewards whose incentives are institutional rather than proprietary, and carrying a documented regulatory history that argues for verifying rather than accepting its most flattering numbers.

It is also worth naming what would genuinely change the picture in either direction, since the watchlist below is deliberately narrow. On the upside, the single most transformative development would be evidence that the county funding surplus can be deployed at commercial rather than policy yields — most plausibly through the leasing subsidiary, which has already demonstrated it can earn a materially higher return on the same deposits than the parent's lending book can.

On the downside, the development that would do the most damage is not a credit event in the LGFV book, which is politically managed and slow-moving, but a sustained rise in retail delinquency combined with any regulatory finding that reopens the question of classification integrity. The first would compress earnings; the second would compress the multiple, and the multiple is already the thinner cushion.

The bull and bear cases do not resolve into a verdict. They resolve into a watchlist: whether the margin inflection holds once repricing is exhausted, whether retail credit stress peaks or spreads, and whether four years of falling fee income finally turn. The 2026 half-year moved in the right direction on the first of those and gave no clear signal on the other two.¹⁷ The next several reporting periods will.

References

1. Results Announcement for the Year 2025 (containing the full 2025 Annual Report) — Chongqing Rural Commercial Bank / HKEXnews, 2026-03-25 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

2. Chongqing financial institutions' deposit and loan balances, end-2025 and January 2026 — Tencent News (腾讯新闻), 2026-03-02 ↩↩↩

3. Six large state banks' 2025 results: deposit cost rates compressed below 1.5%, Postal Savings Bank lowest at 1.15% — Securities Times (证券时报), 2026-03 ↩

4. Corporate Profile and Development History — Chongqing Rural Commercial Bank official website ↩↩↩↩↩

5. PBOC Reply to Suggestion No. 1424 of the Second Session of the 13th NPC (rural credit cooperative reform and special central bank bills) — People's Bank of China, 2020-02-27 ↩

6. Commercial banks' non-performing loan balances and ratios continued a "double decline" in 2008 — National Development and Reform Commission, 2009-02-18 ↩

7. Another eight-figure fine: concealing non-performing loans and false provision coverage indicators, CQRCB fined RMB 12.85 million — Jiemian News (界面新闻), 2022-11-21 ↩↩

8. Chongqing Rural Commercial Bank fined RMB 8.7 million; former vice president and chief loan officer banned for life — Jiemian News (界面新闻), 2025-12-26 ↩↩

9. Chongqing bank raises $1.35 billion ahead of HK listing — FinanceAsia, 2010-12-14 ↩

10. CQRCB's A-share bid: NPL ratio's "three-step jump" this year, overdue loans of RMB 5.9 billion — Sina Finance (新浪财经), 2018-11-08 ↩

11. CQRCB's A-share market capitalisation exceeded RMB 100 billion on its first trading day — Jiemian News (界面新闻), 2019-10-29 ↩↩

12. Inside the largest listed rural commercial bank's results briefing: assets expected to exceed RMB 2 trillion, twice-yearly dividends maintained — Sina Finance (新浪财经), 2026-04-15 ↩↩↩↩↩

13. The rural commercial bank leader earns RMB 33.22 million a day; cash dividends exceed RMB 3.6 billion, dividend yield 4.49% — Sina Finance (新浪财经), 2026-03-27 ↩↩

14. Chongqing Rural Commercial Bank Co Ltd (601077.SS) company profile and market data — Reuters ↩↩

15. Chongqing Rural Commercial Bank Co Ltd (3618:HK) stock price quote and financial profile — Bloomberg ↩↩

16. After 19 years away from CCB, a 49-year-old trust veteran returns to banking: CQRCB's new chairman — Time Weekly (时代周报), 2025-04-18 ↩↩↩

17. Margins warm for the first time in four years: ten A-share listed rural commercial banks earned RMB 26 billion in the first half — JRJ (金融界), 2026-09-07 ↩↩↩↩↩↩↩↩↩