HSBC's one degree of separation isn't saving shareholders from heartburn

Shareholders seldom like surprises. But HSBC Holdings plc seems to love springing them in its earnings.

The London-headquartered bank took a $400 million hit in the first quarter on a loan it didn't make. The bank financed the lender that did, and is now part of a growing list of European banks discovering what private-credit "back leverage" costs when the underlying borrower turns out to be a fraud.

Chief Financial Officer Pam Kaur called the loss "idiosyncratic," according to reporting by the Financial Times. Shareholders didn't agree. HSBC's shares fell 4.6% in Hong Kong and 5.5% in London on the day of the announcement.

Management would like investors to file the charge under a one-off. The numbers for the three-month period that ended in March, the bank's recent regulatory record, and the structure that produced the loss all point the other way.

HSBC's $400 million charge sat in its Corporate and Institutional Banking (CIB) division. The bank described the exposure in its earnings disclosure as "fraud-related, secondary, securitization," language that says HSBC didn't lend directly to the fraud without saying who it did lend to.

The FT has linked the loss to a credit line HSBC extended to Atlas SP, the asset-backed lending arm of Apollo, which had financed Market Financial Solutions (MFS), a U.K. specialist mortgage lender that collapsed earlier this year amid allegations it had pledged the same collateral to multiple creditors. Atlas SP wrote off £400 million of its own. Barclays, separately exposed to MFS, lost £228 million. HSBC's hit sits one rung up the chain, on the credit line that funded the lender that funded the fraud.

The structure isn't novel and the risk isn't new.

Bank lending to private-credit funds has been a recurring topic in financial-stability commentary on both sides of the Atlantic for at least the past two years. The appeal is that the bank earns spread on private credit's growth without underwriting any of the loans. The problem is that, when one of the loans turns out to be fraud, the bank finds out late and has nothing to repossess except the fund itself. Regulators flagged this risk before HSBC took the loss. The bank chose the structure anyway.

The rest of the quarter undermines the one-off framing.

Total expected credit losses (ECL) jumped 49% year-on-year to $1.3 billion. The annualized charge of 53 basis points sits well above HSBC's own 30-to-40 bps medium-term planning range. Management raised full-year ECL guidance from 40 to about 45 bps. Headline profit before tax of $9.4 billion came in below the roughly $9.6 billion analyst consensus. None of that pattern reads like a single bad week.

The recent record makes the framing harder still.

In January, HSBC paid €302.5 million to settle a French dividend-tax-fraud probe covering 2014 through 2019. In 2024 and 2025, the U.K. Competition and Markets Authority fined HSBC, alongside Citi and Morgan Stanley, for sharing pricing information on UK government bonds in proprietary chat rooms.

In 2021, the European Commission fined HSBC and other banks for running a foreign-exchange (FX) spot-trading cartel. In 2014, U.S. and U.K. regulators fined HSBC for foreign exchange manipulation and toxic-securities abuses. In 2012, the bank paid the U.S. government $1.92 billion to resolve charges that it laundered at least $881 million for the Sinaloa and Norte del Valle drug cartels and processed $660 million in transactions for sanctioned regimes including Iran, Cuba and Sudan.

HSBC keeps producing expensive surprises in different parts of its operation. Management keeps booking them as outside the normal course of business. The accumulation tells a different story than any single instance does.

The board's response, on paper, has been reorganization.

HSBC dissolved the Financial System Vulnerabilities Committee (FSVC), which was created after the 2012 settlement, and folded its mandate into the Group Risk Committee. Management framed the move as maturation: financial crime would no longer be a "special project" but part of normal risk management. The framing made sense if the underlying problem had been solved. Q1 2026 suggests it has shifted to a new version of the same problem -- fraud detection inside structured-finance vehicles funded by bank money. Either way, HSBC's regular risk apparatus didn't catch it in time.

Chief Executive Officer Georges Elhedery's pitch is for a "simple, more agile, growing HSBC." The simplification mandate is real, and it had to happen. The executive committee has been cut to 12 members. The bank has actioned $1.4 billion of $1.5 billion in targeted annual savings, six months ahead of schedule. The trade press has reported plans to cut 20,000 staff as artificial intelligence (AI) absorbs back-office work. None of that addresses the question Q1 raises.

A leaner organization with fewer people in non-client-facing roles is easier to manage. Credit quality in back-leverage portfolios needs the opposite: more credit analysts, more relationship coverage, more lawyers reading collateral documents at the borrower's borrower. That work is exactly what AI doesn't do well, and it's the work simplification programs target first.

The financials reflect the tension.

Reported return on tangible equity (RoTE) slipped to 17.3% from 17.9%. Strip out notable items (and the bank does want you to strip them out), and RoTE rises to 18.7%, above the 17% target Elhedery committed to. The trouble is what HSBC's quarterly disclosures have made notable items: $500 million of adverse revenue impact this quarter, $1.4 billion in 2024, and a near-permanent feature of the income statement going back several years. Items showing up every quarter are operating costs in disguise.

Banking net interest income guidance for 2026 went up to roughly $46 billion, a real positive driven by the reinvestment of the $604 billion structural hedge. Then ECL guidance also went up, by five basis points.

Whatever the bank earns on the asset side, it now expects to give back, in part, to provisions.

The Group's Common Equity Tier 1 (CET1) ratio fell to 14.0% from 14.9% at year-end, sitting at the floor of management's 14%-to-14.5% target range. The first interim dividend of $0.10 a share is unchanged from last year. Management called the dividend a sign of "financial strength and stability." A flat payout in a quarter when reported profit fell, ECL surged and capital sat at the floor of the target range suggests management was at the limit of what it could safely return, and knew it.

Hang Seng was supposed to be the bright spot.

The $13.7 billion privatization, completed in January, simplifies HSBC's Asian legal-entity structure and is expected to deliver $500 million in revenue and cost synergies by 2028. Hong Kong remains the Group's most profitable segment, with $2.6 billion in profit before tax for the quarter and 17% growth in wealth management fees.

The Asian pivot narrative has a gap.

The take-private also removed Hang Seng's deteriorating loan book from public scrutiny, at a moment when Hong Kong's banking sector -- having failed to diversify beyond the city's property market -- was sitting on an estimated $25 billion in bad loans.

At its half-year results in mid-2025, Hang Seng reported gross impaired loans of about $7 billion. Its non-performing loan (NPL) ratio stood at 6.69% of total gross loans, up from 6.12% at the end of 2024 and roughly 2.8% at the end of 2023. By year-end 2025, the NPL ratio had risen further to 7.04%.

Concentrating capital in Asia may improve allocation. But it does nothing to improve the controls in London.

The MFS exposure didn't land in the Hong Kong business or the International Wealth and Premier Banking (IWPB) segment. It landed in CIB, in the U.K., where the regulated bank meets the unregulated private-credit fund. That's where the next charge of this kind is most likely to land too, because the structure that produced this one hasn't changed.

The structural hedge will keep producing net interest income. The CIB book will keep extending credit to funds the bank doesn't fully see through.

Investors stripping out notable items to reach the 18.7% RoTE target are betting those two patterns won't intersect again.

That bet could come back to haunt them.

--The author is the Head of Research and Analysis at Icarus Asia, a Hong Kong-based risk and advisory business.