Hua Xia Bank 华夏银行: The Bank a Steelmaker Built — Can a New Team Turn Around China's Weakest Joint-Stock Lender?

I. Introduction & Episode Roadmap

On March 27, 2026, sixteen directors of 华夏银行 Hua Xia Bank sat down in Beijing and signed off on the bank's 2025 annual report. Every seat at the table was filled; the vote was unanimous.¹ It was an unremarkable procedural moment, except for one thing: most of the people in that room had not been in those jobs two years earlier. The chairman who signed the document, 杨书剑 Yang Shujian, had been in the chair for barely a year. The president sitting beside him had been approved by regulators seventeen months before. The chief risk officer, the chief information officer, the chief financial officer and the board secretary were all new. And the man who had chaired the board for the previous eight years was, by then, publicly named by Beijing's discipline authorities as being under investigation for "serious violations of discipline and law."²

That is a lot of turnover for a bank with ¥4.74 trillion of assets — roughly $660 billion — a balance sheet that ranks it 47th in the world by Tier 1 capital, according to The Banker's 2025 league table.¹³ Hua Xia is not a small institution. It runs 943 outlets across more than 120 Chinese cities, employs about 36,200 people, and holds deposits from 2.4 trillion yuan of Chinese savers and companies.³

And yet the equity market prices it like a distressed asset. In August 2026 the A-shares changed hands around ¥6.74, giving the bank a market capitalisation near ¥107 billion — about $15 billion — against reported net assets per common share of ¥19.84 at the end of 2025.⁴¹ The market, in other words, values Hua Xia's equity at roughly a third of what the accountants say it is worth. That gap is the whole story. It is not a rounding error or a liquidity discount. It is the market saying, in the bluntest available language, that it does not fully believe the carrying value of the loan book — or that whatever profit the bank generates will never reliably belong to minority shareholders.

Here is the hook that makes Hua Xia genuinely different from the other Chinese banks you might read about. It was not founded by the People's Bank of China, or by a provincial government, or by an entrepreneur with a licence and a dream. It was founded by a steelmaker. 首钢集团 Shougang Group — the vast Beijing steel complex that once symbolised Chinese heavy industry — set the bank up in October 1992 as a wholly owned subsidiary, essentially an in-house treasury that grew into a national commercial bank.³ Then a German bank bought in, spent a decade as the second-largest shareholder, and quietly sold out at a very large profit. Then a state insurer, a state power utility's investment arm, and Beijing's municipal infrastructure financier became the owners of record. And then, in the space of about eighteen months, almost the entire executive suite was replaced with people from a different bank across town.

The framing question for what follows is simple and unsentimental. Between 2024 and 2026 Hua Xia executed one of the most complete leadership resets in recent Chinese banking history, and management has been telling investors it is producing a "V-shaped" recovery.⁵ Is that a genuine turnaround — new discipline finally cutting into a stubborn credit problem — or is it a fresh coat of paint on a structurally sub-scale bank whose recent headline numbers happen to have been flattered by bond-market marks and lighter provisioning?

The road ahead runs like this. First, the origin story, kept short — it matters mainly because it explains who owns the bank today. Then the Deutsche Bank decade, which is the single most instructive episode in Hua Xia's history for anyone trying to evaluate "strategic partnership" claims anywhere in emerging-market finance. Then the cap table as it actually stands, and the extraordinary 2022 share placement that tells you more about Chinese bank capital regulation than any textbook. Then the economics: how this bank makes money, and what its revenue mix reveals about the franchise it does and does not have. Then the competitive landscape — where Hua Xia sits among China's nine national joint-stock banks, and why that position has been so hard to move. Then the credit reckoning, the governance record, the reboot itself, and the forensic question of what the 2025–2026 numbers really show. Finally, the bull and bear cases, the handful of metrics that actually matter from here, and where this story stands as of August 2026.

↑ Back to Top

II. Origins: A Steelmaker's Bank (1992–2003)

Picture Beijing in 1992. 邓小平 Deng Xiaoping had just completed his southern tour, the ideological brake on market reform had been released, and every large Chinese state enterprise was suddenly being encouraged to think of itself as a business rather than a production unit. Shougang — the Capital Iron and Steel Works — was among the most aggressive. It was the flagship of the state industrial pilot programme, it had unusual autonomy over its own capital, and it wanted to become a conglomerate rather than a steel mill.

So in October 1992 it did what large industrial groups in that era were briefly permitted to do: it started a bank. Hua Xia Bank was established in Beijing as a nationwide commercial bank wholly owned by Shougang.³ The logic was not mysterious. A steelmaker of that scale sat at the centre of an enormous web of suppliers, hauliers, contractors and downstream fabricators, all of whom needed working capital. Owning the bank that financed the supply chain meant capturing the spread on credit the group was effectively already extending, and it gave Shougang a route into financial services at precisely the moment Chinese policymakers were tolerating — briefly — non-financial state enterprises moving into finance.

That founding fact has echoed down thirty-four years. Hua Xia's institutional DNA was never retail deposits or wealth management. It was corporate lending to industrial borrowers, sourced through relationships, priced on relationships, and underwritten with the risk culture of a manufacturer rather than a credit institution. Every one of the bank's later difficulties — the concentration in corporate credit, the thin fee income, the recurring failures of loan "three-checks" discipline that regulators would fine it for decades later — traces back in some form to that origin.

From Steel Subsidiary to Listed Bank

The corporate form caught up with the reality in March 1995, when the bank went through joint-stock restructuring and stopped being a wholly owned Shougang subsidiary.³ Ownership diversified across a group of state and quasi-state entities, though Shougang remained the anchor. Then, in September 2003, Hua Xia listed on the Shanghai Stock Exchange under the code 600015, becoming only the fifth listed bank in China.³ That was early. 招商银行 China Merchants Bank had listed in 2002; the giant state banks — 中国工商银行 ICBC, 中国建设银行 China Construction Bank — would not go public until 2005–2006. For a few years, Hua Xia was one of the very small number of Chinese banks whose shares a public investor could actually buy.

Being early conferred a genuine advantage at the time: access to public equity capital before most competitors had it, and a public listing that forced a level of disclosure discipline unusual for the era. What is striking, looking back from 2026, is how little that head start compounded. Among the nine national joint-stock commercial banks — the tier below the Big Four state giants — Hua Xia has ended up the smallest by assets, and for years the one with the highest bad-loan ratio. Being first to the capital markets did not translate into being first to a franchise.

Shougang, for its part, never left. Thirty-four years after founding the bank it remains the largest shareholder, holding 21.68% of the shares at the end of 2025.¹ That continuity is unusual, and it turns out to matter enormously — because the identity of who actually controls Shougang today is the key that unlocks the modern ownership story. And before we get there, there is a decade-long detour through Frankfurt.

↑ Back to Top

III. The Deutsche Bank Experiment (2005–2016)

In the mid-2000s, there was a trade that seemed to every global bank CEO like free money. China's banking system was being recapitalised and opened up; Chinese regulators wanted foreign "strategic investors" to bring governance, risk systems and product expertise; and Chinese bank equity was available at prices that, in hindsight, look absurd. Bank of America took a stake in China Construction Bank. Goldman Sachs led a consortium into ICBC. Royal Bank of Scotland went into Bank of China. The pitch was always the same: we will not merely own shares, we will transform our partner.

Deutsche Bank's entry point was Hua Xia. In May 2006, Deutsche Bank and affiliated vehicles including Deutsche Bank Luxembourg bought 4.16 billion shares at ¥4.5 apiece, taking a 9.9% position, wrapped in a comprehensive strategic cooperation agreement that included technology transfer for the credit-card business.⁶ It was, on paper, exactly the model the regulator wanted: patient foreign capital plus operational know-how, in exchange for a seat at the table of a fast-growing bank in the world's fastest-growing large economy.

Averaging Up Into the Thesis

Deutsche kept buying. In March 2008 it added 2.67 billion shares at ¥14.62 — more than three times its original entry price — lifting its stake to 13.7%. In 2009 it picked up indirect holdings by acquiring a parent entity, reaching 17.12%. In April 2011 it subscribed for a further 5.15 billion shares at ¥11.17, settling at 19.99%, the regulatory ceiling for a single foreign investor and enough to make it the second-largest shareholder in the bank.⁶ Cumulative outlay across the decade came to roughly ¥12.3 billion.⁶ This was not a toe in the water. Deutsche Bank was averaging up, repeatedly, into a Chinese mid-tier lender at prices well above its initial cost. That is the behaviour of an investor who believed the strategic thesis, not merely the trade.

What the Partnership Actually Delivered

So what did the partnership actually produce? This is where the story becomes genuinely useful, because the honest answer is: remarkably little that outlived it. The credit-card joint venture launched in 2007 — the marquee deliverable of the technology-transfer agreement — did not reach profitability before Deutsche exited the arrangement in 2014.⁶ A 2008 commodity-linked wealth-management product, built with Deutsche input, generated customer losses and litigation.⁶ Deutsche's influence, as the Economic Observer put it in its post-mortem, remained largely technical rather than strategic: it could second specialists and share manuals, but it could not reshape how Hua Xia originated credit, priced risk, or competed for customers.⁶

Meanwhile, Deutsche Bank's own world changed. By 2013, when the two sides ceremonially re-signed their strategic cooperation agreement, the German bank was in the middle of a grinding post-crisis rebuild — capital ratios under pressure from Basel III, litigation provisions mounting, and a strategic review that would ultimately conclude the bank had to shrink almost everywhere. Minority stakes in foreign banks, however strategically attractive, consumed capital under the new rules without conferring control. The cooperation agreement was a press release. The capital math was the reality.

The Exit, and What It Taught

The exit came in December 2015. Deutsche Bank sold its entire 2.136 billion-share, 19.99% position to 中国人民财产保险 PICC Property & Casualty for a price reported in the range of ¥23 billion to ¥25.7 billion — roughly €3.2–3.7 billion.⁶ Add the roughly ¥3.8 billion in cash dividends collected along the way, and total gains came to somewhere between ¥14.5 billion and ¥17.2 billion on ¥12.3 billion invested.⁶ Over a decade, Deutsche Bank had roughly doubled its money on a Chinese bank stake, in a period when its own share price was collapsing. From a treasury perspective, it was one of the better decisions the institution made in the 2000s.

Here is the part investors should sit with. Deutsche Bank made an excellent investment and a failed partnership, and those two outcomes are not in tension — they are the same outcome viewed from different seats. For Deutsche, the position was always, at bottom, a balance-sheet trade on Chinese financial deepening. For Hua Xia, it was supposed to be a transformation. Only one party got what it came for.

The durable lesson generalises well beyond China. When a company announces a "strategic partnership" with a marquee foreign name, the announcement itself contains almost no information. What contains information is the operating evidence that arrives two, three and five years later: did the product win share, did the cost-to-serve fall, did the risk metrics improve, did the partner's people stay? By that standard, the Deutsche Bank decade left almost no trace on Hua Xia's financial statements. Its credit-card franchise never became a peer-leading business. Its fee income never approached the levels of banks that built genuine wealth-management platforms. The German bank left, the manuals stayed, and the bank went back to doing what it had always done: lending to Chinese companies.

What replaced Deutsche Bank on the register, though, changed the character of the institution permanently — because the buyer was not a strategic partner at all. It was a Chinese state insurer, and it arrived at the head of a queue.

↑ Back to Top

IV. Ownership Today: State Capital, an Insurer, and a Utility

Read Hua Xia's shareholder register and a curious thing happens: you keep expecting to find a controlling shareholder, and you never do.

At the end of 2025, the top of the register looked like this. Shougang Group held 21.68%. 国网英大国际控股集团 State Grid Yingda International Holdings — the financial holding arm of 国家电网 State Grid Corporation of China, the world's largest utility — held 19.33%. PICC Property & Casualty, holder of the old Deutsche Bank block, held 16.11%. 北京市基础设施投资有限公司 Beijing Infrastructure Investment, the municipal entity that finances the capital's subway system, held 10.86%. Below them sat 云南合和（集团） Yunnan Hehe Group at 3.52%, the Hong Kong Securities Clearing nominee account at 2.60%, 润华集团 Runhua Group at 1.65% with its entire stake pledged, and the two national "stabilisation fund" vehicles — 中国证券金融 China Securities Finance and 中央汇金资产管理 Central Huijin Asset Management — at 1.27% and 1.03%.¹

Four holders, therefore, control more than 68% of a bank with no controlling shareholder of record. And the annual report includes a small organisational chart that is worth more than several pages of narrative: Beijing's municipal State-owned Assets Supervision and Administration Commission owns 100% of 北京国有资本运营管理有限公司 Beijing State-owned Capital Operation and Management Co., which owns 100% of Shougang Group, which owns 21.68% of Hua Xia Bank.¹ The steelmaker that founded the bank is itself now a wholly owned instrument of Beijing's municipal state-capital apparatus.

Line the four blocks up and what emerges is not a corporate cap table so much as an allocation of a public utility. A municipally controlled industrial group, a nationally controlled power utility's finance arm, a nationally controlled insurer, and a municipal infrastructure financier. There is no private strategic owner, no founder, no activist, and — critically — nobody whose economic interest is served primarily by maximising the return on this particular pool of equity.

That structure produces two effects that pull in opposite directions, and the tension between them is the core of the investment debate.

The first effect is a capital floor. When a bank is owned in these proportions by entities that are themselves owned by the Chinese state, the probability of a disorderly failure is close to zero, and the probability of a supportive capital injection in stress is high. That is not speculation; Hua Xia has already run the experiment.

The ¥15.16 Placement: Buying Stock at Three Times the Market

Which brings us to October 2022, and one of the strangest capital-markets transactions in recent memory. Hua Xia completed a private placement of 528 million A-shares at ¥15.16 per share, raising ¥8 billion, of which Shougang subscribed ¥5 billion and Beijing Infrastructure Investment ¥3 billion, both accepting a five-year lockup.⁷ The proceeds went entirely to core Tier 1 capital.⁷

Now the detail that makes it remarkable. On the day the placement result was disclosed, Hua Xia's A-shares closed at ¥5.04.⁷ The two state shareholders paid ¥15.16 — a premium of more than 200% to the traded price. They voluntarily bought stock at three times the market price, and agreed not to sell it for five years.

Why would anyone do that? The answer is regulatory, and it is the single most important structural fact about Chinese bank equity. Chinese state-owned enterprises are generally barred from issuing new shares below net asset value per share, to prevent the dilution of state assets. Hua Xia's book value per share was in the high teens; its market price was around five. Therefore the only price at which Hua Xia could legally issue equity was a price at which no rational outside investor would buy. The only possible buyers were shareholders willing to treat the purchase as something other than an investment — that is, state entities under instruction. The original plan had been to raise ¥20 billion; what actually got done was ¥8 billion.⁷

Sit with the implications. First, Hua Xia's access to external equity is effectively closed. It cannot tap the market, because the market price is far below the legally permitted issue price. Its only equity sources are retained earnings and the goodwill of state shareholders — a point management itself would later make explicitly when defending the dividend policy. Second, the willingness of Shougang and Beijing Infrastructure Investment to write those cheques is genuine evidence of support, and a bear case that ignores it is incomplete. Third — and this is the part bulls tend to skip — a bank that needs insider capital injections priced at three times market is, by definition, a bank not generating enough capital organically to fund its own growth. Support is not the same thing as strength.

Support Without Accountability

The second effect of this ownership structure is subtler: it blunts the mechanism by which capital allocation normally gets disciplined. Shougang does not need Hua Xia's return on equity to rise; it needs Hua Xia to be stable, to serve Beijing's policy priorities, and not to embarrass anyone. State Grid's finance arm has its own strategic reasons for the stake. PICC inherited a large block from a foreign seller. None of these owners will run a proxy fight over a sub-9% return on equity or a below-peer payout ratio. The stock has traded below book value for well over a decade, and no owner of consequence has treated that as a problem requiring action.

One further thread is worth pulling, because it resurfaces later. Beijing's municipal financial family — the state-capital operator, the infrastructure financier, the municipal trust companies, 北京银行 Bank of Beijing — increasingly behaves as a connected group with shared personnel and overlapping mandates. In early 2024, an executive moved from Hua Xia to chair 北京农商银行 Beijing Rural Commercial Bank, while a Bank of Beijing executive moved the other way into a Hua Xia vice presidency.⁸ Read the 2025 management reshuffle against that backdrop and it looks less like a talent search and more like an internal reassignment within a single municipal system.

For an investor, this ownership picture sets the boundary conditions for everything that follows. Downside is cushioned; upside is capped by the absence of anyone whose job is to insist on shareholder returns. Which makes the actual operating economics — what this bank earns, and on what — the only place where value can be created or destroyed.

↑ Back to Top

V. How Hua Xia Actually Makes Money: Segments & Economics

Strip away the ownership drama and Hua Xia is, at its core, an almost startlingly plain business. It borrows money from depositors and bondholders at one rate, lends it to Chinese companies at a higher rate, and keeps the difference. That is roughly two-thirds of the story, and the remaining third is mostly the investment portfolio.

The 2025 numbers show the shape of it clearly. Total assets reached ¥4.7376 trillion at year-end, up 8.25%. Loans grew 8.47% to ¥2.5667 trillion. Deposits grew a faster 10.71% to ¥2.3817 trillion.¹ Those are respectable growth rates for a Chinese bank in a year when nominal GDP growth was subdued, and the fact that deposits outgrew loans is a quiet positive — it means the bank funded itself more cheaply and leaned less on interbank borrowing.

But revenue went the other way. Operating income fell 5.39% to ¥91.914 billion, and net profit attributable to shareholders slipped 1.72% to ¥27.2 billion.¹ A bank growing its balance sheet by 8% while its revenue shrinks by 5% is telling you something specific: the spread it earns on each yuan of assets is compressing faster than the asset base is expanding. Net interest margin came in at 1.56% for 2025, down three basis points from 1.59%.¹ That is the entire Chinese banking industry's problem in one line — successive policy rate cuts and mortgage repricing have squeezed asset yields faster than deposit costs have fallen — but it lands harder on a bank with nowhere else to earn.

The Six Percent Problem

And Hua Xia genuinely has nowhere else to earn. Net interest income was ¥62.948 billion, up 1.43%. Non-interest income was ¥28.966 billion, down 17.44%.¹ Within that non-interest bucket sits the number that defines the franchise: net fee and commission income of just ¥5.576 billion, up 2.44%, representing 6.07% of operating income.¹

Six percent. Take a moment with that, because it is the most diagnostically important figure in this entire article.

Fee income is what a bank earns for doing things other than lending — selling wealth-management products, managing assets, running custody, underwriting bonds, processing payments, servicing credit cards. It is the part of banking that does not consume capital, does not create credit risk, and does not evaporate when interest rates fall. It is, in the language of business models, the capital-light layer sitting on top of the capital-intensive one. Banks that have built it — through distribution reach, brand trust, or product manufacturing — earn higher returns on equity through the cycle and are structurally less exposed to margin compression.

Hua Xia has essentially not built it. After thirty-four years, more than 900 branches, and a decade-long technology partnership with a global investment bank, roughly ninety-four cents of every yuan of revenue still comes from taking credit or market risk with its own balance sheet. The bank does have a wealth-management subsidiary, 华夏理财 Hua Xia Wealth Management, and it grew: group wealth-management product balances rose 45.82% in 2025 and the subsidiary added 48 external distribution partners.¹ Custody scale rose 18.89%.¹ These are real improvements off a small base. They moved fee income's share of revenue by 0.47 percentage points.¹ That is the honest measure of how far there is to go.

Corporate Strength, Retail Stall, Treasury Swing

Three business lines carry the bank. Corporate banking is the largest and, in 2025, the most dynamic: corporate customers grew 6.72%, corporate deposits 11.43%, and corporate loans excluding discounted bills 13.88%.¹ Retail banking is the weak link, and the 2025 disclosure is unusually revealing about how weak. Personal deposits grew 9.01% — fine. But personal consumer loans grew 0.40%, and cumulative credit-card issuance grew 3.20%.¹ Read that carefully: the retail lending engine is, to a first approximation, not running. Consumer credit is where Chinese banks with strong retail franchises have been generating high-margin, granular, diversified income. Hua Xia's is flat.

The third line is financial markets and treasury — the investment portfolio, interbank business, bill trading and custody. Financial investments totalled ¥1.718 trillion, or 36.26% of assets, against loans and advances at 53.16%.¹ More than a third of Hua Xia's balance sheet is securities. That matters because it makes reported revenue hostage to bond and equity marks, a fact that will become central when we examine the 2026 numbers. The 2025 annual report is candid that the treasury desk was actively traded: management describes dynamically adjusting strategy, increasing trading frequency and investment intensity, and flexibly shifting position structure.¹ That is a description of a bank that increasingly needs its trading book to work.

Capital: Adequate, Not Abundant

Capital is adequate but not abundant. Core Tier 1 stood at 9.38% at end-2025 — down from 9.77% a year earlier — with Tier 1 at 11.75% and total capital adequacy at 13.16%.¹ All comfortably clear the minimums, including the additional buffer Hua Xia carries as a designated domestic systemically important bank.¹ But note the direction: risk-weighted assets grew 8.9% to ¥3.394 trillion while core Tier 1 capital grew only 4.6%, which is why the ratio fell.¹ The Tier 1 ratio held up only because the bank issued ¥20 billion of perpetual bonds during the year, on which it paid ¥1.43 billion of interest.¹ Perpetuals count as capital for regulators; for common shareholders they are a prior claim on earnings.

The synthesis is unglamorous but clear. This is a plain-vanilla commercial lender with a shrinking spread, a stalled retail business, a negligible fee franchise, and a large securities book doing more and more of the revenue work. There is no hidden crown jewel here — no undervalued asset-management arm, no fintech platform being carried at cost. Anyone building a case for Hua Xia has to build it on credit quality and cost discipline, because there is no second engine. Whether that is a survivable position depends entirely on who else is competing for the same borrowers.

↑ Back to Top

VI. The Competitive Landscape: Bottom of China's "Joint-Stock Nine"

Chinese banking is best understood as a pyramid with unusually rigid tiers. At the apex sit the Big Four state commercial banks — ICBC, China Construction Bank, 中国农业银行 Agricultural Bank of China and 中国银行 Bank of China — plus 交通银行 Bank of Communications and 中国邮政储蓄银行 Postal Savings Bank. They have the deposit franchise of a sovereign, the funding cost of a government, and the policy mandate to match.

Below them sits the tier that Hua Xia belongs to: the national joint-stock commercial banks, the 股份制银行. There are nine of them, and they are worth naming because they define the competitive set: China Merchants Bank, 兴业银行 Industrial Bank, 浦发银行 Shanghai Pudong Development Bank, 中信银行 CITIC Bank, 中国民生银行 China Minsheng Bank, 中国光大银行 China Everbright Bank, 平安银行 Ping An Bank, 浙商银行 China Zheshang Bank, and Hua Xia. Below them come the city and rural commercial banks, thousands of them, mostly local.

The joint-stock nine occupy an awkward strategic position by construction. They lack the state banks' funding advantage and policy protection, but they have national licences and are expected to compete nationally. The successful ones solved this by building something the state banks could not: a differentiated franchise. The unsuccessful ones simply became smaller, more expensive versions of the state banks, competing for the same corporate borrowers with worse funding costs.

Hua Xia is in the second group, and the evidence is not ambiguous.

The Sub-Scale Trap

Start with size. Among the nine, Hua Xia is the smallest by assets. That is not a fatal condition in banking — plenty of small banks earn superb returns — but it is fatal when combined with an undifferentiated business model, because scale is the only advantage available to a commodity lender. Fixed costs of compliance, technology and branch infrastructure get spread over a smaller revenue base. Hua Xia's cost-to-income ratio was 30.61% in 2025, up 0.81 percentage points, even though it cut operating and administrative expenses by 2.84% in absolute terms.¹ Costs fell and the ratio still worsened, because revenue fell faster. That is the signature of operating deleverage at a sub-scale player.

Now the contrast that makes the gap concrete. China Merchants Bank — CMB, the acknowledged retail and wealth-management leader among the nine — reported 2025 net fee and commission income of ¥75.258 billion, up 4.39%, with non-interest income accounting for 36.13% of total revenue against an industry average of 22.53%.⁹ Its retail assets under management reached ¥17.08 trillion, growing 14.44%.¹⁰ Its non-performing loan ratio was 0.94%, the lowest among the major national banks.¹¹

Hua Xia's fee income was ¥5.576 billion.¹ CMB's was more than thirteen times larger. CMB's retail AUM is roughly 3.6 times Hua Xia's entire balance sheet. And CMB earns more of its revenue from fees precisely because it has spent two decades building a retail deposit and wealth franchise that gives it both cheaper funding and a product shelf to sell across. That is a genuine competitive advantage in the Helmer sense — it derives from a resource (an installed base of affluent retail relationships) that a competitor cannot replicate without spending a decade and a great deal of money doing so.

Ping An Bank built a different edge: it sits inside 中国平安 Ping An Insurance's ecosystem, with access to an enormous insurance agent force and customer database that functions as a low-cost customer acquisition channel. Whether that model is currently performing is a separate question — its retail credit book has had its own problems — but it is at least a mechanism, an answer to the question "why would a customer choose you?"

Ask that question of Hua Xia and the honest answer is that, for most customers, they would not choose it for any product-level reason. They would bank with it because of a corporate relationship, a location, or an existing borrowing arrangement. Hua Xia's own strategy documents effectively concede this: the 2025 report leans heavily on serving Beijing — the "京华行动" Jinghua Action, aligning with the capital's development priorities, targeting Beijing state-owned enterprises and municipal projects, and reporting that its deposit and loan market-share rankings in Beijing each improved by one place among peers.¹ Being the preferred bank of Beijing's municipal apparatus is a real asset. It is also, definitionally, a local advantage held by a bank that carries a national cost structure.

Five Forces in Chinese Mid-Tier Banking

Run the five forces briefly, because they explain why this position is hard to escape. Rivalry among the joint-stock nine is intense and largely undifferentiated, which means price competition on loan spreads — hence the margin compression. Buyer power has risen sharply: Chinese depositors and payers now have 支付宝 Alipay and 微信支付 WeChat Pay-linked money-market and wealth products that pay more than bank deposits and are one tap away, which strips banks of exactly the low-cost, low-value transactional deposits that used to subsidise everything. Supplier power — the cost of funds — is set by policy and by competition, not by any bank's negotiating skill. New entry in the licensed sense is essentially closed, which helps, but functional entry by technology platforms into payments and distribution has already happened. And substitutes are proliferating: bond markets, trust products, and direct financing all disintermediate traditional lending. Regulatory constraint sits on top of all of it, capping pricing, mandating lending to designated sectors, and setting capital requirements.

The one force working in Hua Xia's favour is switching costs among corporate and state-owned clients. Once a company runs its payroll, settlement, cash pooling and bill discounting through a bank, moving is genuinely painful. That is why the 2025 corporate loan growth of nearly 14% is the healthiest operating number in the report, and why management's "客户倍增计划" customer-doubling plan is focused on settlement services as a way of making relationships stickier.¹ It is a real, if modest, source of franchise durability.

Bottom line: the "why does Hua Xia win from here" case is thin. It is not that the bank is badly run today — the evidence of the last eighteen months suggests it is being run more carefully than before. It is that the structural position offers no mechanism through which above-average returns can be earned. Which means the entire investment question collapses into a narrower one: can it stop losing money to bad credit? That question has a long and uncomfortable history.

↑ Back to Top

VII. The Credit Reckoning: NPLs, Real Estate, and Thinning Provisions (2021–2025)

In November 2024, Hua Xia's credit-card centre published four separate non-performing loan transfer announcements on a single day. One of the packages contained 14,821 individual assets with ¥1.247 billion of unpaid principal. It was put up for auction at ¥120.5 million — under ten cents on the yuan.¹²

That is what the end of a credit cycle actually looks like from the inside. Not a dramatic write-off announcement, but a steady procession of auction notices, batch sales, and repeated failed auctions at declining reserve prices. One restaurant-related debt went through three unsuccessful auctions with the starting price falling from ¥458 million to ¥366 million by November 2024.¹²

The headline numbers tell a story of slow, genuine, hard-won improvement from a bad starting point. Hua Xia's non-performing loan ratio was 1.77% in 2021 and 1.75% in 2022 — the highest among the nine A-share-listed joint-stock banks in both years.¹² It fell to 1.67% at end-2023, 1.60% at end-2024, and 1.55% at end-2025, the fifth consecutive year of decline.¹¹³ Progress. But at the end of the third quarter of 2025 the ratio of 1.58% still ranked highest among A-share joint-stock banks,¹⁴ and to put 1.55% in perspective, the median NPL ratio across 23 listed Chinese banks reporting 2025 results was 1.23%, with only three banks above 1.5%.¹¹

Five years of improvement, and Hua Xia is still in the bottom handful. That persistence is what makes the problem structural rather than cyclical. Cyclical problems mean-revert; structural ones require the underwriting culture that created them to change.

The Property Book

The property sector is where the damage concentrated. Hua Xia's real-estate-sector non-performing loan ratio reached 4.21% at the end of 2023, on a non-performing balance of ¥4.07 billion in that book.¹⁵ To be clear about the mechanism: this was not primarily mortgage lending to households. It was development finance — loans to property developers to acquire land and build. When 恒大 Evergrande defaulted in late 2021 and the presale-funded development model seized up across China, developers stopped being able to refinance, and banks that had lent aggressively into that model found themselves holding claims on half-built projects. A 4.21% bad-loan rate in a book is severe; the comparable figure across the whole loan portfolio at the time was 1.67%.

There has been repair. Management reported that the real-estate sector NPL ratio improved by 0.93 percentage points year on year in 2025, and that the group cleared and disposed of ¥43.3 billion of non-performing assets during the year, recovering ¥9.8 billion in cash.¹³ Those are large numbers relative to a ¥2.57 trillion loan book, and they represent genuine balance-sheet work rather than accounting cosmetics.

What the Cleanup Cost

But now look at what it cost, because this is where the analysis gets uncomfortable. Provision coverage — the ratio of loan-loss reserves to non-performing loans, a measure of how much cushion the bank holds against loans it has already identified as bad — fell to 143.30% at the end of 2025, down 18.59 percentage points from 161.89%.¹ The loan provision ratio fell from 2.59% to 2.23%.¹

Both directions are wrong. Coverage falling by nearly nineteen points in a single year while the NPL ratio improved by only five basis points means the reserve stock shrank considerably faster than the problem did. And 143.30% sits below the 150% level that has historically been treated as the regulatory benchmark for Chinese commercial banks. The annual report addresses this directly, noting that under a 2018 regulatory notice the joint-stock banks are subject to a differentiated, dynamically adjusted provisioning requirement, and stating that the group's provision coverage and loan provision ratio complied with regulatory requirements at period end.¹ That is a legitimate technical answer — the 150% figure has not been a hard universal floor for years. It is not a reassuring one.

Follow the money through the income statement and the mechanism becomes visible. Operating income fell ¥5.232 billion in 2025. Credit and other asset impairment losses fell ¥3.375 billion, from ¥28.791 billion to ¥25.416 billion — a 11.72% reduction.¹ Income tax fell ¥1.26 billion.¹ Add those together and you have accounted for almost the entire revenue shortfall. Net profit attributable to shareholders fell only 1.72% against a 5.39% revenue decline, and the arithmetic reason it did not fall further is that the bank charged less against earnings for credit losses in a year when it was actively disposing of bad assets.

One independent Chinese analysis put it bluntly, describing the coverage decline as reducing provisions to support profits — borrowing tomorrow's earnings for today's numbers.¹⁶ That framing is harsh but the arithmetic behind it is not in dispute. It is also worth stating the more charitable reading: if the bank genuinely wrote off and sold ¥43.3 billion of bad assets, the reserve stock legitimately declines as the assets it was reserved against leave the balance sheet, and coverage falls mechanically. Both explanations are partly true. What is not disputable is the resulting position: Hua Xia now enters any future credit shock with a thinner buffer than it had a year ago, and thinner than most peers.

There is a second-order signal in the disclosure worth flagging. The concentration of the loan book to its largest borrowers rose in 2025: the single-largest customer loan ratio went from 2.51% to 3.12%, and the top-ten-customer ratio from 14.08% to 15.67%.¹ Both remain well within regulatory caps, but the direction indicates growth is coming disproportionately from large-ticket exposures. In a bank pushing hard into corporate lending, that is worth watching.

The credit problem, then, is roughly half-solved and half-financed. The bank has been working through the property legacy with real effort and real cash. It has also been paying for that effort partly out of its reserve cushion. Which raises the obvious question: how did a bank with a national licence and thirty years of history get into this position in the first place? The regulatory file offers an answer.

↑ Back to Top

VIII. Governance Cracks: Fines, an Impersonation Scandal, and Revolving-Door Management

On May 21, 2021, the China Banking and Insurance Regulatory Commission announced penalties against five banks covering 142 separate business violations. Hua Xia's share was ¥98.3 million for 27 violations — the largest single bank fine of that half-year.¹⁷

The list of what it did wrong reads like a syllabus of everything a credit institution is supposed to prevent. Using proprietary funds and wealth-management funds to buy credit assets the bank itself had transferred out — the classic manoeuvre for moving loans off the balance sheet while retaining the risk. Inadequate pre-loan review and post-loan management, the failure of what Chinese regulators call the loan "three-checks." Lax pre- and post-investment management in interbank investments. Providing financing to enterprises through interbank investment channels for the purpose of acquiring bank equity. Providing financing to land-reserve projects in breach of rules. Opening interbank accounts improperly.¹⁷ Across 2021 as a whole, Hua Xia received 41 separate penalty notices totalling ¥115 million.¹⁷

If that had been a one-off, it would be a bad year. It was not. Between 2022 and 2024, Hua Xia branches accumulated 58 penalties totalling ¥42.43 million, with 2024 alone exceeding ¥25 million for violations including improper credit management and fraudulent accounting practices.¹²

2025: The Worst Year on the Record

And then 2025 was worse than any of them. On September 5, the National Financial Regulatory Administration fined the bank ¥87.25 million for imprudent management of loans, bills and interbank business, and for non-compliant submission of regulatory data; multiple employees were warned and four were fined a combined ¥200,000.¹⁸ It was the largest banking-sector fine in China that year.¹⁸ In November, the People's Bank of China added a further ¥13.81 million.¹⁴ By December 2025, Hua Xia had received more than twenty penalty notices during the year and cumulative fines and confiscations exceeding ¥118 million.¹⁴ Even the branch-level penalties told a consistent story: in January 2025 the Shenzhen branch was fined ¥5.6 million for five violations including improper non-performing asset transfers and insufficient internal-control effectiveness; in August the Wenzhou branch was fined ¥1.7 million for inadequate loan "three-checks."¹⁴

The bank's public response to the September fine was that it "earnestly accepts" the decision, had "rapidly implemented comprehensive remedial measures," and had held "serious accountability discussions with relevant responsible personnel."¹⁸ That language is worth noting precisely because it is the language institutions use when they are managing a disclosure rather than diagnosing a system. Compare it to what the same violations looked like four years earlier and the substantive overlap is close to total: loan three-checks, imprudent interbank business, improper asset transfers. The remedial measures announced in 2021 evidently did not prevent the violations penalised in 2025.

For an investor, the analytical content here is not the money. ¥118 million against ¥27.2 billion of net profit is immaterial in cash terms. The content is what recurrence reveals: that the control environment which produced the credit problems in Section VII was not an accident of a particular cycle but a persistent feature of how the bank operated, and that successive management teams either could not or did not fix it.

The Impersonation Episode

Then there is the brand problem. In late 2024 and early January 2025, fraudsters impersonated Hua Xia Bank branches online, posting fake advertisements that lured users to scan malicious QR codes. The bank issued public denials on December 7, 2024 and January 6, 2025 and reported the matter to police.¹² On one level this is simply a crime committed against the bank rather than by it, and it happens to institutions everywhere. On another level it is a data point about digital brand control: a bank with a strong, well-defended digital presence is a harder target, and one whose customers have a habit of interacting through official channels is less vulnerable to a fake advertisement. The episode landed at a bank whose consumer franchise was, as the numbers show, not growing.

Layered on top of all this was a management carousel. Hua Xia went through three presidents in three years between 2022 and 2024, a rate of turnover that makes multi-year strategy execution close to impossible.¹² Then 2024 itself brought departures of the chairman, the president, multiple vice presidents and the supervisory board chair.

The Chairman

And in December 2025 the most serious governance shoe dropped. 李民吉 Li Minji, born in January 1965, had been chairman and party secretary of Hua Xia Bank since March 2017. On January 27, 2025, the bank announced he had resigned as chairman, executive director and from his board committee roles, citing personal reasons.² In June 2025 his membership of the Beijing committee of the Chinese People's Political Consultative Conference was revoked.² On December 22, 2025, the eighth plenary session of Beijing's 13th municipal party committee reviewed and approved a report from the Beijing Commission for Discipline Inspection on Li Minji's serious violations of discipline and law.² Before joining Hua Xia, Li had been a party committee member and standing deputy general manager of Beijing State-owned Assets Management, and party secretary and chairman of 北京国际信托 Beijing International Trust.² The specific findings against him have not been publicly detailed.

An eight-year chairman formally designated as having committed serious disciplinary and legal violations is a governance event of the first order. It transforms how one should read the entire 2017–2024 period at this bank — the credit decisions, the internal-control failures, the strategic drift. It also, in fairness, reframes the leadership reset that followed: what looked like an unusually abrupt reshuffle starts to look like a necessary one.

↑ Back to Top

IX. The Reboot: A New Chairman and a Wholesale C-Suite Turnover (2024–2026)

The reset actually began before the chairman left. On October 28, 2024, after a seven-month regulatory approval process, 瞿纲 Qu Gang's qualification as director and president of Hua Xia Bank took effect.¹⁹

Qu was an unusual choice for a commercial bank presidency. Born in November 1974, a senior economist with a master's degree, he had spent close to twenty-eight years across state banks, a bank investment group, and — for most of the recent stretch — trust companies. He had been party committee member and deputy general manager of Beijing International Trust, then deputy general manager of 北京金融控股集团 Beijing Financial Holdings Group, then returned to Beijing International Trust as deputy party secretary, director and general manager.²⁰ Note the institution: Beijing International Trust is the same firm Li Minji had chaired before coming to Hua Xia. The Beijing municipal financial system does not have a deep bench so much as a rotating one.

Three months later Li Minji was gone, and Qu Gang took on the acting chairmanship at the age of fifty.²⁰ The bank ran with an acting chairman for roughly six weeks.

The Outsider from Bank of Beijing

Then, on March 17, 2025, Hua Xia's board elected Yang Shujian as chairman.⁸ This was the pivotal appointment, and it came from outside.

Yang was born in August 1969. He took an economics degree from 吉林大学 Jilin University in 1991 and went on to doctoral study at 中央财经大学 the Central University of Finance and Economics. Then he spent nearly twenty-eight years — essentially his entire career — at Bank of Beijing.⁸ The progression is worth tracing because it explains what kind of executive he is. He worked in the office and in human resources, ran the Xueyuanlu branch, returned to headquarters in 2005 as deputy board secretary, and became board secretary in 2007 when Bank of Beijing listed on the Shanghai exchange — meaning he personally ran the capital-markets and governance workstream of a bank IPO. By 2014 he was the founding president of the Shijiazhuang branch and was promoted to vice president. In December 2017 he became party secretary and president of Bank of Beijing, a role he held for roughly eight years, the longest tenure in that seat.⁸

Two things stand out. First, Yang is a capital-markets and governance person as much as a banker — an unusual and arguably well-matched profile for an institution whose central problems are internal control and investor credibility. Second, he had spent his entire professional life at a different Beijing bank. He arrived at Hua Xia with no loyalty to its incumbent structures and, crucially, with a network he could recruit from.

He used it. Within months of Yang's arrival the executive bench was rebuilt, and the pattern is unmistakable. The new chief risk officer, 方毅 Fang Yi, was born in 1980, holds a doctorate, and came from Bank of Beijing, where he had been president of the Shijiazhuang branch — the same branch Yang had founded — and chairman of Bank of Beijing's wealth-management subsidiary. The new chief information officer, 龚卫华 Gong Weihua, born 1971, also came from Bank of Beijing. The new board secretary, 刘艳蕾 Liu Yanlei, born 1977, likewise. The one significant internal promotion was the new chief financial officer, 刘悦 Liu Yue, born 1977, previously general manager of the planning and finance department.²¹

The handover completed on July 21, 2026, when 杨伟 Yang Wei — born January 1966, the last senior "old Hua Xia" veteran — retired from his roles as executive director and vice president, along with all his board committee positions.²¹²² Yang Wei had spent his entire career at Hua Xia, working up through regional branch leadership before becoming a head-office vice president in February 2019, and he simultaneously served as chief financial officer, board secretary, party committee standing committee member and union chairman.²¹ His departure at sixty was ordinary retirement, but it was also symbolic: the institutional memory of the pre-reset bank left the executive floor.

What emerged is a leadership cohort born in the 1970s and 1980s, blending Bank of Beijing alumni in the chief-officer roles with a small number of internal promotions.²¹ Add the earlier movements — a Hua Xia executive leaving to chair Beijing Rural Commercial Bank in early 2024, a Bank of Beijing executive arriving as a Hua Xia vice president⁸ — and the picture is of two banks in the same municipal system with increasingly shared leadership DNA.

What the New Team Has Actually Built

What has this team actually done? On the evidence of the 2025 disclosures, more than cosmetics. The governance architecture was rebuilt: the company charter was revised and the supervisory boards at head office and subsidiaries were abolished, consolidating oversight into the board committee structure.¹ Departments were systematically reorganised into four categories — party and mass organisations, governance, business, and support.¹ A compliance officer system was built from scratch and put into live operation. Risk management was overhauled with a credit-cost attribution mechanism, a model risk management framework, a rapid economic-penalty mechanism for non-performing assets, an operating-unit principal responsibility mechanism, and a stationed-risk-officer regime placing risk staff inside business units.¹ Yang's framing for all this was 刀刃向内 — turning the blade inward — and an instruction to leave the comfort zone.⁵

Read against the fine record in Section VIII, these are precisely the right interventions. Credit-cost attribution and stationed risk officers attack exactly the failure mode that produced years of loan three-check violations: relationship managers originating credit with no personal consequence for how it performed.

The Credibility Clock Reads Zero

Now the credibility test, stated plainly. This team has no track record at this institution. Yang Shujian has been chairman for under eighteen months, Qu Gang president for under two years, and the chief risk, information and financial officers for roughly a year. There is not yet a multi-year record of target-setting and delivery against which to judge them. There is no public disclosure establishing meaningful shareholding or long-dated incentive alignment for the incoming team. Pedigree from a better-performing bank is a reasonable prior, not evidence. Bank of Beijing is a well-regarded institution, but it is a city commercial bank with a fundamentally different business mix, and running it does not automatically transfer to fixing a national joint-stock bank's corporate credit culture.

One more consideration cuts both ways. With Bank of Beijing and Hua Xia now sharing leadership lineage and overlapping municipal shareholders, the implicit-support argument gets stronger — Beijing has visibly invested political capital in this bank's recovery. It also means the executive team's principal stakeholder is a municipal government, not the minority shareholders who own roughly a third of the register. When those interests diverge — on dividend policy, on policy-directed lending, on the pace of credit recognition — it is not difficult to predict which way the decision goes.

The reset is real. The question is what it has produced. And the numbers from the last four quarters are, on their surface, spectacular — which is exactly why they need taking apart.

↑ Back to Top

X. Reading the 2025–2026 Turnaround: Real Improvement or a Trading-Gains Mirage?

On April 30, 2026, Hua Xia released first-quarter results that would make any bank CEO's year. Operating income of ¥24.62 billion, up 35.3% year on year.²³ In an industry where mid-single-digit growth counts as strong, a 35% jump is the kind of number that gets a stock re-rated.

It did not. And working out why is the most useful analytical exercise this company offers.

Start with where the growth came from. Net interest income — the core banking business — was ¥17.669 billion, up 13.66%.²⁴ Non-interest income was ¥6.953 billion, up 162.5%.²³ Within that, gains from fair value changes were ¥2.207 billion, against a loss of ¥2.473 billion in the same quarter of 2025.²⁴

Do the arithmetic. The swing in that single line item — from a ¥2.473 billion loss to a ¥2.207 billion gain — is ¥4.68 billion. Total revenue rose by roughly ¥6.4 billion. So approximately three-quarters of the entire revenue increase came from the mark-to-market on the investment portfolio reversing direction.

This is not fraud, or even aggressive accounting. It is what a bank with over a third of its assets in securities looks like when bond yields move favourably after a quarter in which they moved unfavourably. But it is emphatically not recurring income. A fair-value gain in Q1 2026 versus a fair-value loss in Q1 2025 tells you about the Chinese bond market in two particular quarters. It tells you almost nothing about whether Hua Xia lends better than it used to. And crucially, base effects of this kind reverse: the same line that added ¥4.68 billion of year-on-year revenue this quarter will subtract from it as soon as the comparison base is a gain rather than a loss.

Where the Windfall Went

The tell is further down the income statement. Despite revenue rising 35.3%, net profit fell 1.50% to ¥4.987 billion.²⁴ Revenue up a third, profit down. The reconciling item is provisioning: asset impairment losses of ¥11.524 billion, up 101.9% year on year.²⁴ The bank took the windfall from the trading book and put it into reserves.

That, for what it is worth, is the most encouraging thing in the release. Provision coverage recovered to 146.37%, up 3.07 percentage points from year-end, and the loan provision ratio rose from 2.23% to 2.27%.²⁴ The NPL ratio held flat at 1.55%.²⁴ Management chose to rebuild the buffer it had spent in 2025 rather than book the gain as profit. Sell-side commentary framed this as front-loading risk disposal to create future earnings elasticity, with one broker suggesting credit costs around 1% may persist for roughly three years before improving.²⁴ Whether or not that path proves right, the behaviour itself — using a windfall to strengthen the balance sheet rather than flatter the earnings line — is the kind of thing that, repeated for several years, builds management credibility.

Two further things in the Q1 print deserve attention, and neither is comfortable.

The first is the margin. Net interest margin came in at 1.63% for the quarter, up 7 basis points on the 2025 full-year figure and 3 basis points sequentially, attributed to better liability cost management and improved asset composition.²⁴ If that holds, it matters more than the trading gain, because a stabilising margin at a bank whose revenue is two-thirds spread income is the single largest swing factor in normalised earnings. One quarter is not a trend. But after years of compression, the direction changed.

The second is the pace of lending. Total loans reached ¥2.7666 trillion at the end of Q1 2026, up 7.8% from the end of 2025 — in one quarter. Deposits rose 6.9% to ¥2.545 trillion, and total assets 3.0% to ¥4.8785 trillion.²³ Growing a loan book by nearly eight percent in three months is extraordinarily fast for a bank of this size, and Chinese banks do front-load lending into the first quarter. Still: this is the bank with the joint-stock sector's weakest recent asset quality, expanding credit at an annualised pace far above nominal GDP growth, while simultaneously doubling its impairment charges. Rapid loan growth is the mechanism by which future bad debt gets created. It deserves scepticism, not applause, until the vintages season.

The Shape of 2025

Zoom out to the full-year 2025 picture and the "V-shaped" framing gets more texture. Quarterly revenue ran ¥18.194 billion, ¥27.328 billion, ¥19.359 billion and ¥27.033 billion through the year; quarterly net profit attributable to shareholders ran ¥5.063 billion, ¥6.407 billion, ¥6.512 billion and ¥9.218 billion.¹ The fourth-quarter profit figure is 82% above the first quarter's. Full-year revenue declined 5.39%, but the rate of decline narrowed by 3.40 percentage points versus the first three quarters, and the profit decline narrowed by 1.14 percentage points.¹ Management's own summary language was "steady with progress, and improving within that progress."¹ The trajectory genuinely did improve through the year. Whether the improvement is durable is a different question from whether it happened.

The Dividend Is the Capital Constraint

Then there is the dividend, and here management has been unusually direct about the constraint. At the 2024 annual results briefing on April 18, 2025, Yang Wei — then vice president and finance chief — was asked why the payout ratio sat below 30%. His answer had three parts: the bank remained in a stage of continuous transformation and upgrading; capital regulation had tightened and the bank needed sufficient capital accumulation to support business growth against a narrowing net interest margin; and external capital raising had become increasingly difficult, so capital replenishment relied on internal accumulation for an extended period. He said the bank would work to enhance profitability and gradually raise dividend levels.²⁵

That is a candid and internally consistent answer, and it connects directly to the ¥15.16 placement. Hua Xia cannot issue equity to the market. Every yuan paid out is a yuan not available to support risk-weighted asset growth. So the payout ratio is the direct output of the capital constraint. The ratios themselves have crept upward: 23.18% for 2023, 25.04% for 2024, and 25.94% for 2025, the last comprising an interim dividend of ¥0.10 per share and a final of ¥0.32, totalling ¥6.684 billion.²⁵¹³²⁶ The gradual increase management promised has materialised, in small increments, which is a modest point in its favour on guidance discipline.

But the diagnosis cuts against the bulls. A payout below the roughly 30% typical of the large state banks is not, in this case, evidence of retained-earnings-fuelled compounding. It is evidence of a bank whose organic capital generation barely covers its own growth. At a 8.32% weighted average return on equity, retaining three-quarters of earnings adds roughly six points of internal capital growth annually — against risk-weighted assets that grew 8.9% in 2025.¹ The gap is why core Tier 1 fell 39 basis points during the year despite the profit.

An Accounting Note Worth Monitoring

One accounting item belongs on the record. Effective October 1, 2025, Hua Xia changed the accounting estimate for amortisation of software assets, extending the amortisation period from three years to five under the straight-line method. The change was applied prospectively, with no restatement, and the bank stated it had no material impact on net profit, total assets or net assets.¹ Extending an amortisation life reduces the annual charge and increases reported profit, so the direction of the change is earnings-favourable and it took effect in the quarter that produced the strongest profit of the year. Management's "no material impact" characterisation may well be accurate — software is a small asset class for a bank of this size — but the quantum was not separately disclosed, and it is the kind of item a sceptical investor notes and monitors. The financial statements were audited by 安永华明 Ernst & Young Hua Ming, which issued a standard unqualified opinion.¹

Looking forward, management has committed to a 2026–2030 five-year plan explicitly aligned with national policy priorities in green finance and inclusive finance.¹ That signals where growth will be directed. It does not, by itself, say anything about the economics of that growth — which brings us to the businesses management most wants investors to talk about.

↑ Back to Top

XI. Strategic Bets Worth Watching (Not Yet Core to the Thesis)

Every struggling bank has a slide deck full of fast-growing initiatives, and the analytical discipline is to size them honestly against the core.

Hua Xia's are organised around what Chinese policy calls the "五篇大文章" — the five major articles of technology finance, green finance, inclusive finance, pension finance and digital finance. These are national mandates, not proprietary strategies; every Chinese bank is executing the same list. The relevant question is whether Hua Xia's execution is producing anything differentiated.

Green finance is the most substantial. The green loan balance stood at ¥332.9 billion at the end of June 2025, up 16.79% year on year, backed by ¥20 billion of green financial bond issuance during 2025.⁵¹ By year-end the balance had reached ¥373.4 billion, growing 30.99% — a rate 22.27 percentage points faster than total loan growth.¹ The bank also partnered with the Asian Development Bank on sovereign-backed lending.⁵

That is genuinely fast growth on a meaningful base — roughly 15% of the loan book. It is aligned with the 双碳 dual carbon goals, which means policy support, cheaper funding through green bond channels, and regulatory credit. But investors should be clear about what green lending is and is not. It is a categorisation of loans, not a different business. A green loan to a solar developer carries credit risk, consumes capital, and earns a spread like any other loan — often a thinner spread, because policy-favoured sectors attract every bank simultaneously. Growing it fast is a way to grow the loan book in a direction regulators approve of. It is not, on current evidence, a source of superior economics.

Technology finance grew faster still: the technology-enterprise loan balance reached ¥244.6 billion at end-2025, an increase of ¥85.5 billion or 53.74% — 45.02 percentage points faster than overall loan growth — supported by ¥10 billion of technology innovation bond issuance.¹⁵ A 54% growth rate in lending to technology companies is arresting, and it deserves the same scrutiny as any rapid credit expansion: lending to early-stage and mid-stage technology firms is structurally higher-risk than lending to established industrial borrowers, and the credit quality of a book that has more than doubled in two years will not be knowable for several more.

Inclusive and small-business finance is the compliance-driven leg. The small-and-micro enterprise loan balance reached ¥643.28 billion at the end of June 2025, up 6.94% from the start of the year, with the bank giving small-enterprise lending a ring-fenced credit plan and 100% credit resource support, and cutting small-enterprise lending rates by 52 basis points.⁵ Chinese regulators set explicit inclusive-lending growth targets, and hitting them is a condition of good standing rather than a strategic choice. Cutting rates by 52 basis points into a segment with above-average credit risk is precisely what a mandate looks like when it collides with economics. This is a business the bank must be in, not one that will change its returns.

Digital transformation is the least quantifiable and, in the long run, possibly the most consequential. Loans directed to core digital-economy industries grew 29.89% in 2025, while the number of customers in those industries grew only 3.40% — meaning the growth came from lending more to existing relationships rather than winning new ones.¹ Operationally, the bank deployed robotic process automation across more than 2,000 online service scenarios, reporting savings of over 350,000 work hours.⁵ Internally it rebuilt its wealth and private banking systems, relaunched its mobile banking app, established an IT business partner mechanism, and built out an enterprise cloud platform.¹ Translated into plain terms: this is a bank replacing manual back-office processes with software, roughly a decade after the leaders in Chinese banking did the same. It should show up eventually in the cost-income ratio. It has not yet — that ratio worsened in 2025.¹

Two smaller items round out the picture. Hua Xia completed the restructuring of its remaining village banks — Kunming Chenggong and Sichuan Jiangyou — by buying out minority shareholders in July and November 2025, simplifying a legacy structure.¹ And the group operates one financial leasing company, with over ¥200 billion in assets, and one wealth-management subsidiary.³¹³ Neither is transformative at group scale.

The honest sizing: none of these initiatives is currently large enough or economically distinct enough to change the investment picture. Green and technology lending are growing fast off bases that matter, and if either eventually generates genuine sector expertise — better underwriting in solar, storage or semiconductors than a generalist bank can manage — that would be a real advantage. There is no evidence of that yet, only volume. Treat all of it as optionality to monitor, not as the case.

↑ Back to Top

XII. Bull vs. Bear

Here is where the arguments meet, and both sides have real material to work with.

The bull case rests on four legs.

First, the downside is genuinely cushioned. This is a designated domestic systemically important bank whose four largest shareholders are all arms of the Chinese state, and which has already demonstrated — at ¥15.16 against a ¥5.04 market price — that those shareholders will inject capital when needed.⁷¹ A bank that cannot fail and cannot be starved of capital is a different animal from one that can.

Second, the operating trajectory turned. Revenue decline narrowed through 2025, fourth-quarter profit was 82% above the first quarter's, the NPL ratio fell for a fifth consecutive year, real-estate asset quality improved, ¥43.3 billion of bad assets were disposed of with ¥9.8 billion recovered in cash, and net interest margin stabilised and then ticked up in the first quarter of 2026.¹¹³²⁴ Most importantly, when the first quarter of 2026 handed management a windfall, they spent it rebuilding provision coverage rather than booking profit.²⁴ That is the behaviour of a team managing for the balance sheet.

Third, the reset is substantive. The compliance officer function, the credit-cost attribution mechanism, the stationed risk officers, the governance restructuring — these attack the specific failure modes documented in the regulatory file.¹ And they are being implemented by people with no ownership of the prior regime's decisions.

Fourth, the starting valuation embeds a great deal of pessimism. The shares trade at roughly a quarter of reported net assets per share on the broader measure, and the ¥0.42 per share paid for 2025 represents a mid-single-digit yield at the August 2026 price.⁴¹³ If credit quality genuinely normalises and the payout ratio continues its gradual climb toward peer norms, the arithmetic of a re-rating does not require heroic assumptions.

The bear case is, on the current evidence, the stronger of the two.

Start with structure. Hua Xia is the smallest of the joint-stock nine with the weakest recent asset quality, and it has no identifiable competitive advantage. Run Hamilton Helmer's seven powers against it honestly and the result is close to a clean sweep of absences. Scale economies: it has the least scale in its tier. Network economies: banking barely has them, and what exists accrues to the payment platforms. Counter-positioning: Hua Xia is the incumbent being counter-positioned against, not the challenger. Switching costs: present in corporate cash management, and this is the one genuine power the bank holds — but it is a power CITIC, Industrial and the state banks hold equally. Branding: 6.07% fee income is the quantitative proof that Hua Xia cannot charge a premium for trust the way CMB can.¹⁹ Cornered resource: none disclosed. Process power: the regulatory record argues the opposite. A business with one weak power out of seven, in an industry with intensifying rivalry and rising buyer power, has no structural mechanism for earning above-average returns.

Then execution risk. A nearly complete C-suite replacement resets the credibility clock to zero. Multi-year consistency from a team assembled in twelve months is an assumption, not a fact, and the incentive alignment of that team is not established in public disclosure.

Then the quality of the recovery. Roughly three-quarters of the Q1 2026 revenue surge came from a fair-value swing that mechanically reverses.²⁴ Fee income moved 0.47 percentage points as a share of revenue in a year management described as strategically prioritising intermediary business.¹ Personal consumer loans grew 0.40%.¹ The core franchise did not improve much; the bond market did.

Then the balance sheet. Provision coverage at 143.30% at year-end sat below the traditional 150% benchmark and nineteen points below the prior year, and the Q1 rebuild recovered only three of those points.¹²⁴ Core Tier 1 fell during a year of positive earnings. Loans grew 7.8% in a single quarter at the bank with the sector's weakest asset quality. And the top-ten-borrower concentration rose.¹

Then governance. Cumulative fines exceeding ¥118 million in 2025 alone, on top of ¥115 million in 2021 and ¥42 million across 2022–2024, for overlapping violation categories.¹⁷¹²¹⁴ A former chairman formally designated by Beijing's discipline commission as having committed serious violations of discipline and law.² A public impersonation incident.¹² These are not independent events; they are outputs of the same control environment.

Then the macro overhang. Hua Xia remains exposed to China's property cycle and to local-government financing vehicle credit, to industry-wide margin compression in a persistently low-rate environment, and to the risk that consumer credit deterioration — visible across the sector in 2025 retail NPL data¹¹ — spreads into a retail book the bank has not been growing anyway.

Where this lands. An activist looking at Hua Xia would find plenty to attack — a decade below book value with no strategic review, a payout ratio defended by capital scarcity that is itself the product of an unfixed return problem, an executive team with no disclosed equity alignment, a securities book doing an outsized share of the revenue work — and would find no lever to pull, because the register is controlled by state entities that will not be pressured. That absence of accountability mechanism is itself part of the discount.

This is a "show me" situation. The reboot appears real and the balance sheet is state-supported. What does not yet exist is operating evidence that Hua Xia can close the structural gap with stronger peers, as opposed to merely stopping the bleeding. Those are very different outcomes, and only one of them justifies paying more than a distressed multiple.

↑ Back to Top

XIII. Durable Lessons, KPIs, and the Risk Radar

Three lessons travel well beyond this one bank.

The first is about strategic partnerships. Deutsche Bank spent a decade as Hua Xia's second-largest shareholder, signed two cooperation agreements, transferred credit-card technology, and left roughly ¥15 billion richer having changed almost nothing about how the bank operated.⁶ The generalisable rule: when a company announces a partnership with a prestigious name, treat the announcement as zero information and wait for operating evidence — product share gained, cost-to-serve reduced, risk metrics improved, people retained. If those never appear, the partnership was a financing transaction wearing strategic clothing.

The second is about state ownership. It is genuinely a backstop; it is also genuinely a brake. State shareholders will fund a bank through a crisis at prices no market investor would accept. They will not demand a return on that capital, will not force a strategic review after a decade below book value, and will not act as a constituency for minority shareholders. Investors in state-linked institutions are buying downside protection and selling upside discipline, and should price both.

The third is about management resets. A wholesale change of leadership is often the necessary precondition for improvement — and it sets the credibility clock to zero. Every promise the new team makes is a hypothesis until it has been tested across at least one full cycle of guidance and delivery. Pedigree from a better institution raises the prior; it does not substitute for the record. The right posture toward Yang Shujian's Hua Xia in 2026 is neither the scepticism appropriate to the old regime nor the confidence a proven team would earn. It is attentive neutrality.

The metrics that matter. Three, and only three, are worth tracking closely.

Provision coverage alongside the NPL ratio, together. Neither means much alone. An NPL ratio falling while coverage falls faster is balance-sheet consumption dressed as improvement, which is what 2025 delivered. An NPL ratio flat while coverage rises is genuine buffer rebuilding, which is what the first quarter of 2026 delivered. Watch whether the Q1 2026 direction persists — coverage back through 150% with a stable or falling NPL ratio would be the clearest single signal that the credit problem is actually behind the bank.

Net interest margin, versus the joint-stock peer group. This is the core earnings engine of a bank with no meaningful second revenue stream. The 2025 figure was 1.56% and the first quarter of 2026 came in at 1.63%.¹²⁴ Sustained stabilisation above the peer trend would validate the liability-cost work management has described; renewed compression would mean the operating improvement was a rate-cycle artefact.

Net fee and commission income as a share of operating revenue. At 6.07%, this is the cleanest available proxy for whether Hua Xia is building any franchise at all, as distinct from simply lending more.¹ It is the number that separates the joint-stock banks that solved their strategic problem from those that did not. Movement of a percentage point or more, sustained across several years, would be the first hard evidence of differentiation. Continued drift in the six-to-seven percent range would confirm that this remains a spread lender competing on price.

The risk radar, restricted to what is actually material here.

Credit risk remains the dominant exposure, concentrated in property development, local-government-linked infrastructure borrowers, and — increasingly — the rapidly expanded technology lending book whose vintages have not been tested. Margin risk is structural: with two-thirds of revenue from spread income and Chinese policy rates likely to stay low, any renewed compression flows almost undiluted into earnings.

Capital risk is the compounding concern. If organic profit generation does not improve, risk-weighted asset growth will keep outrunning capital accumulation, and the bank has no market access for equity. The options then are slower growth, further reliance on perpetual and Tier 2 instruments that add fixed charges ahead of common shareholders, or another insider placement — none of which are good for the equity.

Regulatory and conduct risk has demonstrated persistence across three distinct management teams. Fines are immaterial in cash but material as evidence, and further large penalties would suggest the new control architecture is not yet working. Cybersecurity and brand risk sit alongside it, as the impersonation episode showed.

Execution risk is the one that is genuinely new. A nearly all-new executive team, drawn heavily from one outside institution, must hold together and deliver consistent strategy across a five-year plan while the balance sheet, branch network, staff and competitive position all remain exactly what they were. Cultural change at a 36,200-person institution takes years, and personnel turnover among the new cohort itself would be a meaningful negative signal.

↑ Back to Top

XIV. Epilogue

As of August 2026, Hua Xia Bank sits in an unusual position: almost everything about its leadership has changed, and almost nothing about its business has.

The bank has now put up two consecutive quarters that management characterises as a "V-shaped" recovery.⁵ It has a 2026–2030 strategic plan aligned to national priorities in green and inclusive finance, still in its first months.¹ It has a leadership team whose combined tenure at the top of this institution amounts to less than two years. And it carries a ¥2.77 trillion loan book, a 943-branch network, 36,200 employees and a competitive position that is exactly where it was when 李民吉 Li Minji resigned in January 2025.²³³

The near-term calendar is unusually informative. First-half 2026 results are the immediate test, and the specific thing to look for is whether net interest income growth holds up without the fair-value tailwind that carried the first quarter — the base effect that made Q1 look spectacular begins to fade, and what remains is the actual operating trend. Whether provision coverage continues the recovery it began in Q1, or resumes eroding, will say more about the durability of the credit repair than any NPL headline. Further executive appointments will reveal whether the Bank of Beijing pipeline is a one-time transfusion or an ongoing channel. And the 2026 dividend decision will show whether the gradual payout increase management promised in April 2025 survives contact with continued capital pressure.²⁵

There is a broader question buried in this story, and it is the one that makes Hua Xia worth studying even for investors who will never own a Chinese bank. Most turnaround narratives are built around a bold move — a new product, an acquisition, an exit from a failing business, a repricing. Hua Xia's contains none of these. Nobody is proposing to sell the corporate bank, buy a wealth manager, exit a geography, or bet the institution on a new model. The entire thesis is institutional: replace the people, rebuild the control systems, tighten the risk culture, and let the same balance sheet perform better.

Sometimes that is enough. Banks are, more than most businesses, machines for making credit decisions, and who makes those decisions under what incentives genuinely determines outcomes over a decade. A bank that stops originating bad loans looks like a different company five years later even if nothing else changed.

But institutional reset does not create a franchise. It cannot manufacture the fee income Hua Xia does not earn, conjure the retail relationships it never built, or close the thirteen-fold gap in intermediary revenue against the leader in its own tier. The best plausible outcome from the current reforms is a bank that stops destroying value — clean credit, stable margin, adequate capital, a normal payout ratio. That would be a substantial improvement over the last decade, and it is very different from a bank that wins.

Thirty-four years ago, a steelmaker decided it wanted a bank. What it built has outlived the industrial logic that created it, survived a foreign partner's decade-long experiment, absorbed a property crisis, and cycled through leadership at a rate that would destabilise most institutions. Whether the newest set of leaders can do what none of their predecessors managed — turn a balance sheet into a business — is a question that will be answered slowly, in provision coverage ratios and fee-income percentages, over a period considerably longer than the eighteen months they have had so far.

References

1. 华夏银行股份有限公司2025年年度报告摘要 — Hua Xia Bank Annual Report Summary FY2025, 2026-03-30 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

2. 官宣！华夏银行原董事长李民吉，严重违纪违法 — Sina Finance, 2025-12-23 ↩↩↩↩↩↩

3. 华夏银行-公司概况 — Hua Xia Bank Corporate Profile ↩↩↩↩↩↩↩↩

4. 华夏银行(600015) 股票行情 — 中财网 quote page ↩↩

5. 新帅改革百日，华夏银行何以实现"V型"反弹 — 南方都市报, 2025-09-10 ↩↩↩↩↩↩↩↩

6. 10年赚走超150亿 德银结束与华夏银行不圆满的婚姻 — Economic Observer, 2016-01-01 ↩↩↩↩↩↩↩↩↩

7. 华夏银行80亿定增结果出炉，首钢集团、京投公司两大股东为何溢价200%认购？ — Sina Finance, 2022-10-14 ↩↩↩↩↩

8. 杨书剑掌舵，4万亿华夏银行换帅 — East Money, 2025-03-18 ↩↩↩↩↩

9. 招商银行年报解析：13万亿资产背后的"零售之王"韧性 — Sina Finance, 2026-04-24 ↩↩

10. 营收增速三年来首度转正！招行2025交卷：AUM突破17万亿 — 凤凰网 iFeng Finance, 2026 ↩

11. 23家上市银行不良资产扫描：零售贷款承压 — 财联社 Cailianshe, 2026 ↩↩↩

12. 华夏银行遭遇"冒用门"，不良资产频繁转让折射经营难题 — Sina Finance, 2025-01-14 ↩↩↩↩↩↩↩↩

13. 华夏银行2025年报：一份以系统性改革重塑经营韧性的答卷 — 钛媒体 TMTPost, 2026 ↩↩↩↩↩↩

14. 华夏银行原董事长落马：2025年被罚1.18亿 不良率居A股股份行之首 — East Money, 2025-12-24 ↩↩↩↩↩

15. 华夏银行2023年净利263.6亿元，房地产贷款不良率4.21% — 农村金融时报 ↩

16. 华夏银行2025：V型反转还是报表魔术？ — DoNews, 2026 ↩

17. 华夏银行因27条违规行为被罚9830万！相关责任人遭警告并被罚5万 — Sina Finance, 2021-05-21 ↩↩↩↩

18. 因业务管理不审慎等，华夏银行领罚8725万元 — Sina Finance, 2025-09-08 ↩↩↩

19. 历时7个月，原北京信托总经理瞿纲获准出任华夏银行行长 — 财联社 Cailianshe ↩

20. 华夏银行行长瞿纲新任代理董事长，年龄51岁 — Sina Finance, 2025-01-28 ↩↩

21. Huaxia Bank Accelerates Leadership Overhaul: First Vice President Yang Wei Retires — BigGo Finance ↩↩↩↩

22. Hua Xia Bank (600015.SH): Yang Wei, Executive Director and Deputy President, resigns — Futu News ↩

23. 华夏银行2026年一季报出炉：营收大增35.3%，科技绿色贷款双增近20% — China Economic Net, 2026-04-30 ↩↩↩↩

24. 【财经分析】华夏银行2026年Q1营收同比增长35.3% 积极处置风险有望增强业绩弹性 — Sina Finance, 2026-04-30 ↩↩↩↩↩↩↩↩↩↩↩↩↩

25. 华夏银行管理层回应分红比例低于30% — NetEase, 2025-04 ↩↩↩

26. Hua Xia Bank (600015.SH) 2025 Annual Dividend Distribution: RMB 0.32 per share — Futu News ↩