Two of Three Upgrade Criteria Met, and We Are Not Upgrading: HSBC Stays at Hold

Results vs. Consensus

Unlike the first quarter, this was a full results event. HSBC reports half-yearly under UK convention, so the second quarter arrives inside the Interim Results with a chief executive presentation, a chief financial officer walk-through and a nine-questioner call. That matters relative to the first quarter, where the two items that moved the stock were credit judgements and no one above the finance function took questions. This time the chief executive opened, and the questions that drew the least satisfying answers were about costs rather than credit.

2Q 2026 Scorecard

Two mechanical notes on the scorecard. The consensus HSBC's results are judged against is the one the bank compiles and publishes itself, which is the convention for UK banks and the figure the wires quoted; an independent aggregator carried $18.51bn for revenue rather than $18.57bn, so the revenue beat is between 3.0% and 3.3% depending on which book you use. On earnings per share, each ADS represents five ordinary shares, so the $0.45 of basic earnings excluding material notable items translates to $2.25 per ADS against aggregator consensus of $2.23 to $2.24. Earnings per share is the least informative line in the release: the Street models this bank on pre-tax profit and return on tangible equity, and the pre-tax beat is six times the earnings-per-share beat.

Year-Over-Year Comparison

The two profit lines at the bottom of that table are the whole quarter in miniature. Attributable profit rose 68%, and attributable profit excluding material notable items rose 14%. The gap is the year-ago base: the second quarter of 2025 carried a $2.1bn dilution and impairment charge on the Bank of Communications associate stake plus $0.5bn of restructuring, so the comparative was depressed by items that will not recur. Fourteen percent is the number to hold in mind. It is a good number and it is not a sixty percent number.

Quarter-Over-Quarter Comparison

Unlike the first quarter, the sequential column here is clean and can be read at face value. The first quarter of 2026 was not distorted by a large notable-item load, so the 8% sequential rise in pre-tax profit is genuine progression rather than a base effect. Three sequential facts carry forward: banking net interest income recovered $0.4bn of the first quarter's decline, the credit charge fell 11 basis points, and tangible net asset value per share fell for a second consecutive quarter while the share count rose.

First-Half Comparison

Where the prior-period column reads "n/a," the company disclosed the growth rate rather than the base on the management basis, and we do not reverse a rounded percentage to manufacture a comparative.

Quality of Beat

Revenue. Reported revenue of $19,118M beat the company-compiled $18,570M by $548M. Two things sit inside the 16% headline growth rate and neither survives contact with the underlying figure. Revenue notable items were a positive $75M this quarter against a negative $1,245M a year ago, a $1.32bn favourable swing that is nearly the entire difference between 16% reported growth and 7% constant-currency growth excluding notable items. What is left after that adjustment is the number that matters, and it is genuinely better than the first quarter: 7% against 6%, with each of the four revenue drivers management names, banking net interest income, wealth fee income, wholesale transaction banking fee income and loan and deposit growth, stronger than in the first quarter. That is the first acceleration this franchise has printed since the reorganisation.

Credit and costs. This is the cleanest part of the print and the direct answer to the first quarter's problem. The $1.1bn charge equals 41 basis points annualised, eleven basis points better than the first quarter and four better than the year-ago quarter, and it contains no repeat of the fraud item that defined the first quarter. The Hong Kong commercial real estate component was around $0.2bn, of which the top-up on already-impaired positions was just over $150M, and it reflects valuation marks on existing Stage 3 exposures rather than new impairments. On costs, target-basis expenses grew 1.1% year over year against a full-year guide of around 1%, which is the first quarter this year that the run rate has matched the guide. Reported costs actually fell 3%, but only because notable items inside costs shrank from $664M to $269M; the underlying comparison is the 1.1%.

Returns and earnings per share. The first quarter's central criticism was that the record return was flattered by a shrinking equity base while attributable profit was flat. That criticism does not survive this quarter. Attributable profit excluding material notable items grew 14% year over year to $7,701M, and the reported and excluding-notable figures converged to within $13M because notable items were only $0.2bn adverse at the profit level. Reported first-half return on tangible equity of 18.2% against 14.7% is flattered by the year-ago base, but the 19.5% quarterly and 19.1% first-half returns excluding notable items are on a consistent basis and are comfortably above the 17%-plus target. One caution remains and it is on the denominator rather than the numerator: the average share count rose 34M sequentially with no buyback running in the quarter, and tangible book per share fell.

Segment Performance

The table below shows the reported revenue line, the notable items removed from it, and the constant-currency figure excluding notable items that HSBC's chief operating decision maker uses, with the prior-year quarter restated on the same basis. Return on tangible equity and deposits are year-to-date figures on the first-half annualised basis, which is how the company discloses them.

Two features of that table deserve emphasis before the segment detail. First, the breadth: every operating segment grew revenue and every operating segment returned better than 18% on tangible equity, which is the first quarter since the four-business structure took effect where that has been true. Second, the Corporate Centre swing. The year-ago quarter carried $1,202M of negative notable items in the Corporate Centre, the Bank of Communications charge, which is why that column shows a positive $51M on the underlying basis against negative $80M this quarter, and why the group's underlying growth rate is a point below the sum of the four operating businesses.

The deposit and loan bridges are the cleanest evidence of the quarter's balance-sheet momentum. Deposits grew 8% year over year to $1,828bn and $46bn in the quarter alone, and loans grew $20bn sequentially with contributions from all four businesses.

Loans went from $1,002bn to $1,022bn and deposits from $1,781bn to $1,828bn. The deposit growth needs a qualifier that management supplied unprompted rather than under questioning.

"I will highlight that about half of the CIB deposit growth in the quarter was large and short term. These balances come and go."
— Manveen Kaur, Group Chief Financial Officer

Hong Kong

The best segment in the group and the one that most directly validates the $13.7bn committed to the territory in January. Underlying revenue grew 9.5%, the fastest of the four businesses, and first-half return on tangible equity excluding notable items reached 44.3%. Deposits grew $28bn year over year to $537bn, on a base management describes as almost twice the size of the second-largest competitor in the market, and wealth balances in Hong Kong reached $0.5tn, up 10% year over year. The bank added 640,000 new personal banking customers and 24,000 business banking customers in the half. Most importantly for the forward model, lending returned to growth after many quarters of contraction, and it is converting to net growth rather than being offset.

"in Hong Kong, we are not seeing so much of those early repayments on some of the Hong Kong commercial real estate because that sort of stabilized the gross growth in Hong Kong lending comes straight to the bottom line as a net growth. There's no sort of repayments taking that off."
— Manveen Kaur, Group Chief Financial Officer

Assessment: That quote is the most under-appreciated disclosure on the call. For several years Hong Kong gross lending growth was being cancelled by early commercial real estate repayments, so origination did not show up in balances. If repayments have genuinely stopped, the same origination effort now compounds, and the segment's earnings leverage to a recovering Hong Kong economy is higher than the recent balance history suggests. The offsetting caution is the same one as last quarter: a 44.3% segment return is measured against a denominator that shrank when the Hang Seng minority came out, so the level overstates the underlying improvement even though the direction does not.

Corporate & Institutional Banking

The segment that absorbed the first quarter's damage reaccelerated hard. Underlying revenue grew 8.9%, client deposits grew 16% year over year, and the transaction-banking core that HSBC actually sells improved on every component: wholesale transaction banking fee and other income grew 7% against 2% in the first quarter, with securities services up 16% on new mandates, trade up 7% on volumes, payments up 6% and foreign exchange returning to growth at 5%. Trade balances grew 29%. Cross-border client revenue grew 15%, with 85% of client revenue coming from multi-jurisdictional clients and mainland Chinese clients contributing 10% of that, three quarters of it booked outside China. First-half return on tangible equity was 18.9%.

"Trade business delivered this quarter with balances up 29% year-on-year. Trade is at the heart of HSBC, and we are seeing the trust our customers place in us to help them navigate and invest."
— Manveen Kaur, Group Chief Financial Officer

Assessment: The reacceleration is real but the arithmetic deserves care, because wholesale transaction banking fee income fell sequentially from $3.1bn to $3.0bn even as the year-over-year growth rate went from 2% to 7%. The acceleration is therefore a comparative-base effect, not a sequential step-up: the first quarter of 2025 was a strong $3.0bn and the second quarter of 2025 was a weaker $2.8bn. What is unambiguous is the composition. Four of four streams growing, with the two stickiest, securities services and trade, growing fastest, is the mix a network bank wants. Separately, the 26% decline in debt and equity markets we flagged last quarter as the line to watch was not broken out in the interim disclosure, so that watch item is unresolved rather than closed.

International Wealth & Premier Banking

Underlying revenue grew 8.2% and first-half return on tangible equity excluding notable items was 27.4%, the second-highest in the group. Wealth fee and other income grew 21% in the quarter to $2.8bn with all four streams up: investment distribution 26% on higher mutual fund and structured product sales, insurance 21% from an already strong base on higher contractual service margin release, private banking 22% on client trading activity and recurring fees, and asset management 7% on assets under management growth. Net new money was $25bn in the quarter, of which $22bn came from Asia. Deposits fell $2bn year over year and $7bn sequentially, which management attributed to held-for-sale transfers and to private bank deposits moving into investments.

Assessment: The insurance line is where the first quarter's central bull argument started paying. We wrote in May that a $15.2bn contractual service margin balance growing 19%, of which roughly a tenth reaches the profit and loss account in any year, meant the wealth business was writing materially more economics than it was reporting. Insurance income up 21% "on higher CSM release" is that mechanism converting to revenue. The deposit outflow is the item to keep honest about: a wealth franchise whose clients move deposits into investments is doing the right thing for fee income and the wrong thing for net interest income, and this quarter it did both.

UK

Steady, and slowest of the four, which for this segment is the point. Underlying revenue grew 4.7% and first-half return on tangible equity excluding notable items was 20.8%. The lending performance was the strongest part: commercial lending grew $10bn in the half, up 10%, and mortgages grew $10bn, up 5%, against a retail mortgage stock of roughly $190bn and an 8.1% market share. Active Premier customers grew 7% year over year and new-to-bank business banking customers grew 48%. First-half UK revenue was $10bn, split $7bn in the ring-fenced bank and $3bn outside it, on deposits of $578bn.

Assessment: Double-digit commercial lending growth in a UK market that has been demand-constrained for two years is the most surprising number in the segment disclosures, and management tied it explicitly to the UK's own growth agenda rather than to share gain. The structural hedge reinvestment tailwind has further to run here. The complication is that 4.7% is the group's slowest growth rate at a moment when the group is guiding to 5% by 2028, which means the outer-year target leans on Hong Kong, wealth and trade rather than on the largest developed-market franchise.

Corporate Centre

Underlying revenue of negative $80M against a positive $51M a year ago, and a first-half return on tangible equity of 4.6%. The reported line carried a negative $81M against negative $1,024M in the year-ago quarter, the difference being the $1,202M of Bank of Communications notable items in the base.

Assessment: The associate line remains a structural decliner following the dilution of the Bank of Communications stake through 2025, and the Corporate Centre is now the only part of the group returning below the cost of equity. Neither fact is new and neither is large enough to matter to the thesis. What matters is that the year-ago Corporate Centre charge is doing most of the work in the group's 60% reported pre-tax growth, which is why we anchor on the underlying 13% instead.

Wealth Franchise KPIs

One definitional caution, because it is the kind of thing that gets mis-transcribed into models. The $1.1tn figure is Asia wealth balances, not group wealth balances. Group balances were disclosed at $1.6tn on a new definition in the first quarter and were not restated in the interim materials, so the two numbers are not comparable and $1.1tn is not a decline.

Key Topics & Management Commentary

Overall Management Tone: Confident and forward-leaning, and inverted from the first quarter. Where the previous call was pre-armed and over-disclosed against a credit problem, this one spent most of its airtime defending upside: why targets are being reiterated rather than raised when the first half beat them, why costs may go up because revenue is going up, and why the buyback is small because loan growth is better. Analyst pushback concentrated almost entirely on the cost and investment line rather than on credit, and it was there that management was least willing to be specific, declining to engage the published 2027 cost consensus while confirming the direction of travel. Tone on the guidance itself thinned in one identifiable place: the chief financial officer conceded, unprompted, that the second-half net interest income arithmetic implies a larger number than the guide.

1. The Buyback Restarts, at Less Than Half the Expected Size

HSBC announced a buyback of up to $1bn, its first since pausing repurchases three quarters ago to rebuild capital after the Hang Seng Bank privatisation. The Street had been positioned for something closer to $2.2bn. The restart itself satisfies the specific commitment management made in May, when it said the decision would be looked at again from the second quarter, and the size is exactly what the framework it set out at the same time predicted.

"our priority in the way we use our capital generation 100 basis points, first is to deliver the 50 basis points accrual for our dividend. That's to meet our 50% dividend payout ratio on our earnings adjusted for material notable items. The second priority use of our capital is to support business organic growth. This is how we can generate the highest value for our shareholders when we do it, of course, within risk appetite and at the right returns. And then the leftover excess capital preferred mechanism to distribute it is through share buybacks"
— Georges Elhedery, Group Chief Executive

The quarter's capital bridge shows that framework operating exactly as described. Regulatory profits generated 100 basis points of CET1, the dividend accrual consumed 50, balance-sheet growth consumed 30, and other items 10, leaving the ratio 10 basis points higher at 14.1%. The $20bn of loan growth that consumed those 30 basis points is the same growth the chief executive ranked ahead of buybacks. Risk-weighted assets rose $22.5bn to $906.4bn while CET1 capital rose $3.7bn to $127.7bn. A $1bn buyback is roughly 11 basis points of that risk-weighted base, and management guided it to cost around 10 basis points in the third quarter. Inorganic opportunity still sits ahead of buybacks in the stack, behind an explicit hurdle.

"any acquisition should be accretive to a share buyback, should be fully aligned to strategy, should be enhancing our scale or capabilities and should be easy to integrate and not distracting us from delivering organic growth."
— Georges Elhedery, Group Chief Executive

Assessment: The $1bn is not a disappointment against the framework; it is a disappointment against an expectation the framework never supported. In May management told the market that buybacks were the residual after the dividend, loan growth and inorganic opportunity, and then loan growth accelerated to $20bn in a single quarter. Anyone modelling $2.2bn was modelling against what they wanted rather than what they had been told. The genuine negative is narrower and it is about pace rather than principle: with quarterly decisions, no annual envelope and no capital threshold disclosed, a shareholder cannot underwrite the repurchase line, and at roughly $354bn of market capitalisation a $1bn quarterly pace is about 1.1% a year.

2. Banking Net Interest Income Raised to "At Least $46bn," and Conceded as a Floor

Guidance moved from "around $46bn" to "at least $46bn" on a favourable rate outlook, deposit and loan growth, and the reinvestment of roughly $50bn of maturing structural hedge assets currently yielding 2.8% during the second half. Second-quarter banking net interest income was $11.6bn, up $0.8bn year over year and $0.4bn sequentially, of which $0.1bn was the non-recurrence of a first-quarter Hong Kong one-off. The quarterly series now reads $10.8bn, $11.1bn, $11.8bn, $11.2bn and $11.6bn from the second quarter of 2025 forward, so the current run rate has still not recovered the fourth-quarter 2025 peak.

"So, all in all, really comfortable with the guidance update, and I would just stress on at least $46 billion. And of course, mathematically, you're quite right with the run rate, you could pretty much get a larger number."
— Manveen Kaur, Group Chief Financial Officer

Assessment: The arithmetic makes the concession unavoidable. First-half banking net interest income was $22.9bn, so $46bn for the year requires only $23.1bn in the second half, about $11.55bn a quarter against a second-quarter print of $11.6bn. The guide therefore requires no growth at all from the current run rate, before the $50bn of hedge reinvestment at 2.8% yields, before day count, and before the loan and deposit growth management just described as accelerating. If the second half simply matches the fourth quarter of 2025, the year lands at roughly $46.4bn. Model $46.0bn to $46.5bn and treat the guide as the floor of the range rather than the centre of it. The named risks management set against that, HIBOR volatility at 2.62% and under recent downward pressure, and a weaker sterling translating into a dollar-reported line, are real but second-order.

3. Credit: 41 Basis Points Printed, and the Guide Left Untouched

The expected credit loss charge was $1.1bn, or 41 basis points annualised, against 52 in the first quarter and 45 in the year-ago quarter. The quarterly series reads 45, 41, 36, 52 and 41 basis points from the second quarter of 2025 forward, so this quarter returns the charge to the level that prevailed before the first quarter's two unforecast items. Around $0.2bn related to Hong Kong commercial real estate Stage 3, of which the top-up on positions already impaired was just over $150M. The full-year guide of around 45 basis points was reiterated rather than cut, against a stated medium-term planning range of 30 to 40.

"The top-up on the ECL line was just over $150 million for the quarter. Hong Kong residential is recovering very well, very stabilized now, house price index was up 18% year-on-year, and volumes are also up 36% in the first half year-on-year."
— Manveen Kaur, Group Chief Financial Officer

The Hong Kong property picture management described is genuinely two-sided rather than uniformly improving. Residential prices are up 18% year over year with transaction volumes up 36% in the half, retail sales have grown for 13 consecutive months at 11%, and Central office vacancy has fallen to around 10% with rents up 6% year to date. Against that, East Kowloon office vacancy remains at 20% with rents down a further 4%, and management flagged small pockets of mid-market credit pressure outside Hong Kong.

"Overall, of course, what we are very mindful is that pockets of increase in defaults globally, particularly in mid-market, where our exposures are small. Small exposures can also add up and that's contributed to some of the ECL charge for this quarter."
— Manveen Kaur, Group Chief Financial Officer

Assessment: This is the single most important item on the print for the standing thesis, and it resolves the bear point favourably. In May we wrote that reaching around 45 basis points for the year after a 52 basis point first quarter required roughly 43 across the remaining three quarters, above the planning range, and that this was the number to hold management to. The second quarter came in at 41, better than required. The guide was nonetheless reiterated rather than cut, which after a first half of 52 then 41 implies a second half in the low-to-mid forties, above the 41 just printed. Management is holding buffer rather than releasing it, which is prudent and also means the guide now embeds cushion in the same way the net interest income guide does. We move our central case to 44 basis points with 47 as the live downside, and we note that a bank choosing not to cut a credit guide it is beating is telling you something about the mid-market pockets it named.

4. The Middle East Overlay Held, Then Folded Into the Base Case

The $300M precautionary reserve built in the first quarter after the onset of the Middle East conflict was neither released nor topped up. Separately, the standalone Middle East IFRS 9 scenario introduced in the first quarter at a 30% weight was removed and its substance folded into the general downside economic guidance. Actual credit experience in Middle East exposures was described as benign, with charges well below $100M.

"the $300 million reserve that we built in Q1 still holds we have not released that reserve. That's in line with the policy. We like to see full two clear quarters of stability"
— Manveen Kaur, Group Chief Financial Officer

Assessment: Held rather than topped up is the outcome we said to watch for, and it is the better of the two available outcomes. The asymmetry we flagged in May, that a forward-looking overlay tops up readily and releases slowly, remains structurally true but did not cost anything this quarter. The technical change deserves a note of its own: removing the standalone scenario and folding it into the base downside keeps the provision while reducing the granularity at which an outsider can test it. Management also set a two-clear-quarters-of-stability policy for releasing reserves, which given benign observed experience since February makes the third or fourth quarter the earliest plausible release window, and a release would be a small tailwind the guide does not currently assume.

5. No Second Sponsor-Intermediated Charge

The $0.4bn Stage 3 charge on a fraud-related, sponsor-intermediated securitisation exposure that defined the first quarter did not recur. No comparable item was disclosed, and the Stage 3 additions management named were Hong Kong commercial real estate and small mid-market positions. In May, management said it had completed a full review of its highest-risk exposures and found no comparable fraud concerns; one quarter later, that statement has survived its first test.

Assessment: This is the pillar where the quarter delivers the most and proves the least. A clean quarter is exactly the evidence a shareholder wanted, and it moves the near-term probability materially. It does not address the structural point from May, which was that reliance on a financial sponsor's diligence cannot be reviewed away, only tightened. Two things keep this from being fully resolved. The roughly $6bn of total private-credit exposure and roughly $3bn securitisation-financing bucket disclosed at the first quarter were not refreshed in the interim materials, so an outsider cannot see whether the book grew, shrank or stayed flat. And no analyst asked, which means the market has moved on faster than the disclosure has.

6. Simplification Raised to $2.0bn on an Unchanged Cost Envelope

The most verifiable item on the call, and the clearest instance of management beating a commitment it made in public. In May the target was $1.5bn of annualised savings fully actioned by the end of June, six months ahead of the original plan, with $1.4bn actioned at that point. It has now been exceeded, and the target raised.

"We originally set out to deliver USD 1.5 billion of annualized saves. We have now exceeded this target, reaching USD 1.7 billion of action saves. We have achieved this with associated restructuring costs of USD 1.4 billion lower than originally expected. We're now planning to use the full USD 1.8 billion of restructuring costs as per our initial commitment to increase total organizational simplification saves to USD 2 billion to be actioned before the end of the year."
— Georges Elhedery, Group Chief Executive

The chief executive attributed the overachievement to three sources: conservatism at inception in areas where the programme's scope was uncertain, headcount reduction delivered through attrition rather than severance, and additional opportunities identified once the cost envelope proved underspent. Roughly half of the 2025 to 2028 non-strategic application decommissioning plan is now complete, with a further 20% delivered in the half, and around 50 processes are under reengineering.

Assessment: The economics of the raise are better than the headline and worth doing precisely. The first $1.7bn of annualised savings cost $1.4bn of one-time restructuring, a payback of roughly ten months. The incremental $0.3bn that takes the target to $2.0bn will consume the remaining $0.4bn of the envelope, a payback of roughly sixteen months. Both are excellent, and the second is worse than the first, which is what you would expect as a programme extends into harder ground. The important point for the model is that these are actioned decisions with completion dates rather than forecast savings, and they are the reason the full-year cost guide deserves the benefit of the doubt despite what comes next.

7. The Cost Line Becomes a 2027 Problem

The most consequential new information on the call, and the least quantified. Management pre-announced two upward pressures on costs, then declined to size either. First, additional performance-related pay in 2026 if business momentum continues, against a variable pay pool that was $3.9bn last year, roughly 11% to 12% of group costs. Second, accelerated investment in 2027, pulled forward from an existing plan because revenue growth is running ahead of expectations.

"just in terms of variable pay as a context, it was $3.9 billion last year, which is 11% to 12% of our group costs. So any increase which we will consider provided revenue growth momentum continues, and that's a decision, as Georges said, at the end of the year, they'll only have a modest change in total cost for the year."
— Manveen Kaur, Group Chief Financial Officer

"we are not commenting on consensus or giving a cost guidance for 2027."
— Manveen Kaur, Group Chief Financial Officer

The offsets management named are real but explicitly partial: the incremental $0.3bn of simplification savings carries into 2027, and roughly $0.4bn of cost reallocated from divested businesses becomes available, against $0.3bn reallocated in 2026. Inflation and investment added about 5% to first-half costs before simplification savings offset 3% of it.

Assessment: The 2026 arithmetic is tighter than the reiterated guide implies. First-half target-basis costs were $16,979M against a restated 2025 baseline of $33,936M, so around 1% full-year growth implies roughly $34.3bn, which requires the second half to land near $17.3bn against $17.3bn a year earlier. In other words, the guide now requires the second half to be flat year over year, at the same moment management has flagged two reasons it might not be. The 2027 position is worse for a different reason: the direction has been disclosed and the magnitude has not, and a bank whose chief financial officer says next year's investment number "would be a bit higher" while refusing to engage a published 3% consensus is inviting the market to mark that consensus up without giving it a number to mark to. Three of the nine analysts who got questions in spent both of theirs on this, which tells you where the doubt is.

8. Wealth: $25bn of Net New Money and a 21% Fee Acceleration

Net new money was $25bn in the quarter, of which $22bn came from Asia, against $39bn and $34bn in the first quarter. First-half net new money was $64bn with $57bn from Asia, up 32% year over year. Wealth fee and other income grew 21% to $2.8bn with all four streams growing.

"in terms of the net new money, I would just say $25 billion is still a strong number for the quarter and is broadly aligned with our recent quarterly run rates, 8% to 9% annualized."
— Manveen Kaur, Group Chief Financial Officer

Assessment: We have to grade ourselves here rather than management. In May we set an upgrade criterion of net new money holding above $30bn per quarter into a seasonally weaker period, and $25bn misses it. On reflection the bar was badly set: it anchored on $39bn, which management flagged at the time as carrying favourable first-quarter seasonality against a strong prior-year base, so we built a test around the outlier rather than the trend. The trend is 8% to 9% annualised on Asia balances of $1.1tn, and the more informative line accelerated: fee income up 21% against 18% for the half, with insurance up 21% specifically on higher contractual service margin release. That release is the mechanism the first-quarter note argued was under-appreciated, and this is the first quarter it visibly converted into reported revenue. The pillar is intact; our test for it was not well constructed.

9. Hong Kong Loan Growth Returns and Hang Seng Synergies Go Live

Hong Kong lending grew $6bn sequentially and resumed growth after an extended period of contraction or flatness. On the Hang Seng privatisation, management reiterated $500M of audited reported synergies plus roughly $400M of related benefits, about $900M in total over roughly three years, and gave the first operational evidence.

"we're also very pleased to see finally Hong Kong resume loan growth after many quarters of contraction or flat. So this is very encouraging for the future."
— Georges Elhedery, Group Chief Executive

"We're more than 80% now live in execution on the various execution work streams of these synergies, therefore, have a high level of confidence in our ability to drive those synergies."
— Georges Elhedery, Group Chief Executive

The specific evidence offered was that Hang Seng Bank nearly doubled new customer acquisition sequentially, to around 60,000 from around 30,000, after adopting HSBC's digital onboarding. Single leadership has been put in place across both banks in infrastructure, back office, technology and manufacturing. Hong Kong capital markets activity grew nearly 50%, with 40 live IPOs in Hong Kong and 70 across Asia.

Assessment: In May management said Hang Seng synergies were "very little, if any," and expected to appear from the second half of 2026 and mainly in 2027 and 2028. The first operational proof point arrived a quarter early, and it is the right kind: a capability transfer that doubles customer acquisition costs nothing incremental and compounds. The financial synergies remain absent from the numbers, and the $500M of reported synergies is still broadly offset by $600M of associated restructuring through 2028, so the deal's near-term case continues to rest on the minority buy-in economics and Hong Kong's own trajectory. Both are now working.

10. Tangible Book Fell Again, and Roughly $4bn of Dividend Sits Undeclared

Tangible net asset value per ordinary share fell $0.10 sequentially to $9.36, a second consecutive quarterly decline. The bridge is straightforward: $8.2bn of profit accretion to tangible equity, worth $0.48 per share, against $9.4bn of ordinary dividends paid in the quarter, worth $0.55 per share, that being the fourth-quarter 2025 final dividend of $0.45. Foreign exchange added $0.03 and fair value movements through other comprehensive income subtracted $0.04. Tangible equity fell to $160.7bn from $162.3bn. The average share count rose 34M sequentially with no buyback running.

Against that, the dividend entitlement is running well ahead of the declaration. HSBC accrued $0.44 per share of dividend in the first half under its 50% payout target but has declared only $0.20, two interim dividends of $0.10, the second payable on 25 September. The roughly $0.24 per share difference is equivalent to about $4bn at the June share count and is deferred to the final declaration.

Assessment: These two facts pull in opposite directions and both are load-bearing. On the negative side, two consecutive declines in tangible book per share is a genuine problem for a stock the market values on a multiple of tangible book, and it explains most of why the multiple expanded this quarter even though the price rose less than earnings did. On the positive side, the forward yield is materially better than the declared run rate suggests: 50% of roughly $1.78 of full-year earnings excluding notable items is about $0.89 per ordinary share, or $4.45 per ADS, which is roughly 4.3% at the current price rather than the 1.9% the two declared interims annualise to. The second-half book pattern should also invert, because no large final dividend is paid in the second half; on our estimates tangible book per share exits 2026 near $10.00, which is the anchor that matters more than the spot figure.

11. The Perimeter Closes: Fifteen Exits and the Singapore Insurance Sale

Fifteen business or market exits have now been announced since 2025, three of them in the ten days around this print. The Singapore insurance business goes to Allianz for S$2.7bn, generating a pre-tax gain of around $1.8bn and up to 15 basis points of CET1 on completion in the first half of 2027. Australia retail exits. Egypt retail banking goes to Emirates NBD Egypt for a gain of around $0.3bn, completing in the second half of 2027. Malta was classified as held for sale at a $0.3bn loss. Roughly $1.1bn of costs associated with the exited businesses becomes available for redeployment.

"We don't have this commanding market share and leadership role in Retail. And that's the trade-off we have made."
— Georges Elhedery, Group Chief Executive

Assessment: All three strategic reviews we flagged in May as unresolved capital, Australia retail, Egypt retail and HSBC Life Singapore, closed in a single quarter. That removes a specific uncertainty from the 2028 revenue target's perimeter and is worth more than the gains themselves. The doctrine is consistent: keep the country as a network market, sell the sub-scale retail footprint. The one that deserves harder thought is Singapore insurance, because HSBC is selling an insurance manufacturer inside the wealth franchise it calls its primary growth engine, and the interim materials do not disclose what share of the $15.2bn contractual service margin balance goes with it. That is a gap rather than a red flag, but the contractual service margin is the mechanism the bull case rests on and it should be sized.

12. China's Cross-Border Wealth Clarifications

The newest identifiable risk to the wealth pillar is regulatory: China's State Council Decree 837 and related clarifications issued in May and June set out guidelines for individual outbound investment, which bears directly on the mainland-sourced flows that make Hong Kong the cross-border wealth hub. Management welcomed the clarifications and answered with observed data rather than assurance.

"when we look at our client onboarding in May, in June and so far in July, we have seen account opening not affected by these additional onboarding declarations and therefore, remain confident in the medium- to long-term outlook for Hong Kong to be the leading cross-border wealth hub for the planet"
— Georges Elhedery, Group Chief Executive

Management confirmed constant engagement with regulators in both Hong Kong and mainland China, characterised as business as usual, and separately disclosed that there has been no slowdown in Hong Kong non-resident new-to-bank customer acquisition in recent months. It also noted that new-to-bank customers arrive with relatively low balances that grow substantially as the relationship matures, which is the mechanism by which today's acquisition becomes tomorrow's net new money.

Assessment: Three months of observed onboarding volumes is the highest-quality answer available to this question and better than the reassurance most banks would offer. It is also, by construction, a lagging indicator of a rule change that has been live for two months. The risk here is not that volumes have fallen; it is that the mechanism of risk is regulatory intent, which expresses itself in enforcement rather than in the text of a decree. This belongs on the watch list at low probability and high consequence, since roughly a quarter of group revenue comes from wealth relationships and 88% of the quarter's net new money came from Asia.

Guidance & Outlook

Three of twelve lines moved and all three moved favourably, which is the opposite of the first quarter, where two lines moved in opposite directions. Nothing was guided down. The reiteration of the medium-term targets rather than a raise is the item the market focused on, and management addressed it directly.

"if we see opportunities to accelerate and if we see supporting market conditions as we have seen in the first half of this year, you should expect us to be able to beat the targets, including starting in 2026. But we're not changing the foundational baseline target for how we run the bank across a range of scenarios."
— Georges Elhedery, Group Chief Executive

Implied ramp. On banking net interest income, $46bn for the year against a $22.9bn first half needs only about $11.55bn a quarter in the second half, against $11.6bn just printed, so the guide requires no growth from the current run rate. On credit, around 45 basis points for the year after a first half of 52 then 41 implies a second half in the low-to-mid forties, above the second quarter's 41, so buffer is being held. On costs, roughly 1% full-year growth on a restated $33,936M baseline implies about $34.3bn, so after $16,979M in the first half the second half must land near $17.3bn against $17.3bn a year earlier, which is flat year over year at a moment when two upward pressures have just been flagged.

Street position. The net interest income raise was expected on the rate outlook and the debate was about size rather than direction. The simplification raise was not widely anticipated. The buyback restart was expected in timing and undershot in size against roughly $2.2bn of expectation. The reiterated medium-term targets disappointed a market that had begun to price a raise off a first half that beat them, and the 2027 cost consensus of roughly 3% growth to about $35.4bn was put to management directly and left unaddressed.

Guidance style. Conservative by construction, by admission and now by pattern. The net interest income guide has been conceded as arithmetically beatable in two consecutive quarters. The credit guide is being held after a quarter that beat it. The cost guide is the exception and runs the other way: it is the one line where the arithmetic is demanding rather than cushioned, and it is the one line management pre-warned about. Model the revenue and credit guides as floors and ceilings respectively, and model the cost guide at the top of its range.

Analyst Q&A Highlights

Whether Capital Deployment Now Outranks Buybacks

The opening question of the call, and the one that framed everything that followed. It read the day's messaging as a shift in priority toward deploying capital into the business rather than returning it, and asked what metrics govern that decision and whether proceeds from upcoming disposals would also be deployed rather than returned. The answer set out the priority stack in sequence and put a discipline condition on the growth.

Q: "It does sound like part of the message today is that there's more opportunities out there to deploy capital into the business. So now it's more of a priority, perhaps over buybacks from here. So two questions sort of falling off that. What sort of metrics are you looking at when making that decision?"
— Guy Stebbings, BNP Paribas

A: "one thing to reassure you about is that we will only grow at the right returns and within our risk appetite. That is paramount in the way we drive our growth."
— Georges Elhedery, Group Chief Executive

Assessment: The question correctly identified the trade the market had not yet priced. Management did not resist the framing, which is the tell: capital deployment does now outrank buybacks, and it does so because loan demand returned in Hong Kong and stayed strong in the UK. For a shareholder the read is uncomfortable but coherent. Deploying at a return above the cost of equity beats repurchasing at 2.2 times tangible book on arithmetic alone, and a bank that repurchased stock at this multiple instead of funding 10% commercial loan growth would deserve criticism. The buyback line is the residual, and this quarter the residual got smaller because the alternative got better.

Why Second-Half Net Interest Income Does Not Grow Off the Second-Quarter Run Rate

The most quantitatively specific challenge of the call, and a direct repeat of the line of questioning that produced the same concession three months ago. The argument was that a benign rate environment, hedge support in the second half and an improving volume message are inconsistent with a guide that implies no growth from the second-quarter level. Management conceded the arithmetic and defended the guide on the words "at least."

Q: "I'm struggling a little bit with the guidance for the second half of the year if we're expecting that to work through. What is it that means that we don't see growth in the second half versus that Q2 run rate?"
— Guy Stebbings, BNP Paribas

A: "as always, in any guidance, we take into account a range of plausible outcomes when setting out the guidance. So there's a certain buffer in it, and that includes consideration of volatility, particularly in HIBOR. You've seen some of that recently. As well as FX rates. And as you can imagine, if sterling is weaker. That has an impact for us for on a dollar-based Banking NII."
— Manveen Kaur, Group Chief Financial Officer

Assessment: Two consecutive quarters in which management has confirmed in plain language that its net interest income guide contains buffer. That converts a habit into a pattern and a pattern into a modelling assumption. The named risks are specific and testable rather than generic: HIBOR at 2.62% and under downward pressure, and sterling translation. Neither is large enough to consume the roughly $0.4bn of headroom between the guide and a flat run rate, let alone the additional contribution from reinvesting $50bn of hedge assets currently yielding 2.8%. This remains the highest-confidence upward revision available from the print.

Where the Additional Simplification Savings Came From

A short, precise question about the provenance of a $0.5bn increase in a savings target, asked because a raised cost-savings number with no change in the cost to achieve it invites the suspicion that the original target was soft. The answer conceded exactly that, in the most useful way available.

Q: "If you could just give a little bit more color on where the additional simplification saves are coming from, what's changed there?"
— Alastair Warr, Autonomous

A: "We have seen strong progress on our cost delivery, part of it refers to areas of cost where we had a little bit more uncertainty at the inception of this program, and we felt more authority now 1.5 years into it. So you can argue this is a kind of 1.5 years ago, some conservatism in our approach, which we are comfortable to deliver. Some of it has been delivered because we managed through attrition, some of the headcount adjustments that may have been required as opposed to using severance."
— Georges Elhedery, Group Chief Executive

Assessment: An executive volunteering that his own original target embedded conservatism is worth more than the $0.5bn. It calibrates how to read every other number this team publishes, and it is consistent with the net interest income guide being conceded as beatable twice in a row. The attrition point is the substantive part and cuts both ways: savings delivered by not replacing leavers are cheaper than severance and are also a form of capacity reduction that does not show up as a restructuring charge. If revenue keeps accelerating, some of those roles come back as investment spend in 2027, which is precisely the conversation the next three questioners had.

The 2027 Cost Consensus and the Payback on Accelerated Investment

The pivotal exchange of the call. It put the published 2027 cost consensus to management as a number, then asked why accelerating investment does not come with any change to the revenue growth targets it is supposed to produce. Management declined the first half outright and answered the second at the level of principle rather than quantum.

Q: "So just a very simple question, I guess, can I ask you to comment on current consensus, which is the 3% cost growth to $35.4 billion next year? And then the second question related to that is that you talked about accelerating investment spend, but there's no change in your revenue growth targets. So what is the payback on this incremental investment? And what's the time frame?"
— Andrew Coombs, Citi

A: "firstly, we are not commenting on consensus or giving a cost guidance for 2027. But let me just unbundle a few things. As Georges said, that targets are the baseline to which we manage the bank. We continue to work hard as we have done this year to exceed our targets and we see no change in doing that."
— Manveen Kaur, Group Chief Financial Officer

Assessment: The question was constructed so that either answer was informative, and the refusal is the answer. A management team comfortable with a 3% consensus would have found a way to say so, particularly having just volunteered that investment would be higher. What the market now holds is a disclosed direction with no magnitude, which is the worst combination for a stock trading at a full multiple, because it forces the marginal buyer to underwrite a cost number the company will not stand behind. The second half of the question went unanswered in substance across the whole call: accelerated investment was justified by revenue momentum, and the revenue targets it is meant to accelerate were reiterated unchanged.

Whether the Accelerated Spend Exceeds the Incremental Savings

A follow-up that tried to bound the same question numerically by asking whether the acceleration is larger than, or similar to, the incremental simplification savings, and when it pays back. The response separated the payback question from the sizing question and answered only the first.

Q: "just on the cost piece, just to make sure I understand it correctly in terms of the incremental or the acceleration in spend. So is that -- in terms of the accelerated spend, is that more than the kind of $500 million incremental simplification savings or around the same amount?"
— Amit Goel, Mediobanca

A: "the payoffs of the investment we do sometimes comes very quickly, actually intra-year. And I'll draw your attention to some of the additional acceleration of investment we did in our Security Services business last year, which pretty much gave us payoffs and new mandates, both last year and into this year. So the payoffs don't have to wait for 1 or 2 years."
— Manveen Kaur, Group Chief Financial Officer

Assessment: The securities services precedent is a genuinely strong piece of evidence and it is checkable: that line grew 16% this quarter on new mandates, which is what an intra-year payback looks like. But the question asked for a magnitude and received an example. Management said elsewhere that simplification savings will "partly offset" the acceleration, with the operative word emphasised, which is as close to a sizing as the call produced and which implies the acceleration is the larger number. Combined with the refusal to engage the 3% consensus, the reasonable modelling response is to carry 2027 target-basis cost growth above consensus rather than at it.

Hong Kong Commercial Real Estate and the Middle East Overlay

Two credit questions in one, both aimed at whether the improvement is durable and whether the reserve built in the first quarter is now the ceiling. The answer was the most granular of the call and separated stabilisation from resolution by naming the specific pockets that are still deteriorating.

Q: "So if we look at the Hong Kong CRE, it does seem like the asset quality trend has stabilized. Do you expect this trend to change or to further improve in the upcoming quarters? And also that we see, it seems like there is no notable overlay related to the Middle East situation that's been taken this quarter, how should we look at the situations with some escalation of the conflict again in the third quarter?"
— Katherine Lei, JPMorgan

A: "where we see some stress continuing is in pockets like in East Kowloon, where vacancy rates remain elevated at 20% and rents have further slipped by 4%. And that's in the office space and also some in the non-core retail space."
— Manveen Kaur, Group Chief Financial Officer

Assessment: The answer is credible precisely because it refuses to declare victory. Prime Central office at 10% vacancy with rents up 6% year to date and East Kowloon at 20% with rents down 4% is a bifurcating market, not a recovering one, and a bank that names its weak pocket is more trustworthy on its strong one. The Middle East half of the answer was the more consequential disclosure: the $300M reserve is held under a stated policy of two clear quarters of stability before release, while the standalone scenario has been folded into the base downside. That combination sets up a possible release in the second half that the current guide does not assume, and it also removes the disclosure through which an outsider could have tested the provision.

The Hong Kong Dollar Hedge and the HIBOR Sensitivity

A technical question about why disclosed Hong Kong dollar rate sensitivity halved sequentially, asked at a moment when HIBOR had been falling. The answer named three mechanisms rather than describing a single hedge.

Q: "Just wondering on the HIBOR moves that we have seen, which is slightly towards the end of the second half of the second quarter. And if you look at the sensitivities that you have put in your Hong Kong dollar sensitivity has fallen by about half since the first quarter. Just can you give some color on what hedges have you taken? And what have you done there?"
— Melissa Kuang, Goldman Sachs

A: "in terms of Hong Kong hedge in the first half of '26. We have increased our Hong Kong hedge. We have looked at new products. We've had some new fixed rate customer lending, so that has helped. And that is again, reduced our sensitivity, as you can see to any rate moves. And of course, HIBOR has moved down to 2.62%."
— Manveen Kaur, Group Chief Financial Officer

Assessment: This is the most consequential piece of risk management disclosed on the call and it got the least attention. Halving Hong Kong dollar rate sensitivity while HIBOR falls is straightforwardly good execution, and doing it partly through fixed-rate customer lending rather than purely through derivatives means the protection came with a revenue relationship attached. The read-through for the model is that the largest single-currency threat to the net interest income guide has been materially reduced, which is a further reason to treat "at least $46bn" as a floor rather than a target.

The Disposal Doctrine Against a Peer That Is Expanding

The most strategically direct question of the call, contrasting HSBC's steady programme of exits with a large domestic competitor opening in the same markets HSBC is leaving, and asking how far the narrowing goes. The response set out the doctrine in full and then applied it to the specific market raised.

Q: "I'm just trying to understand the sort of logic and how far that goes? Because it just seems to be that much like Egypt is about 120 million people. I mean Surely, HSBC can add value there over time with your expertise, in your banking, and numerous skills. And I'm just wondering sort of how far you want to take this sort of focusing strategy."
— Edward Firth, KBW

A: "We are very supportive of Egypt wholesale business Egypt is a major network market. Many of our -- vast majority of our international clients have operations in Egypt, and it's very important for us to support them in this space. We don't have this commanding market share and leadership role in Retail. And that's the trade-off we have made."
— Georges Elhedery, Group Chief Executive

Assessment: The clearest articulation of the strategy this management has given, and it answers the question that matters more than the individual disposals: the exits are not a retreat from geographies, they are a retreat from sub-scale retail inside geographies the bank continues to serve wholesale. That distinction is what makes fifteen exits a focusing programme rather than a shrinking one, and it is why the 2028 revenue target can survive a narrowing perimeter. The question the answer does not settle is where the programme stops. Fifteen exits in eighteen months with three announced in ten days is a cadence, and a cadence eventually reaches businesses a shareholder would rather keep.

What Is Driving Asian Corporate Loan and Deposit Growth

A composition question on the quarter's headline balance-sheet growth: which industries, and whether the deposit inflow is durable corporate liquidity. The answer drew a distinction with the pattern visible at US banks that is worth more than the sector detail.

Q: "Just on the Corporate growth, both in deposits and loans that you're seeing in Asia Pacific, could you comment on both sides of the balance sheet, the nature of the industries that you're seeing is this broad based? Is it around certain industries? And on the deposit side, is this liquidity, corporate liquidity and are there any particular industries that stand up?"
— Joseph Dickerson, Jefferies

A: "the overarching message, I would say, is we are more broad-based compared to some of the activity you've seen in the U.S., which is quite focused on a few sectors relating to AI."
— Georges Elhedery, Group Chief Executive

Assessment: A useful piece of differentiation delivered almost in passing. Loan growth concentrated in a handful of borrowers financing a single technology cycle carries correlation risk that broad-based trade and corporate lending does not, and HSBC's growth this quarter was led by trade, up $6bn in the corporate bank and $4bn in Hong Kong, with additional demand in technology and institutional sectors. The deposit side came with the honest qualifier that roughly half of the corporate bank's $42bn quarterly inflow was large and short-term. That matters for the net interest income model: a deposit base growing 8% year over year of which a slice is transient supports the guide but not a raise, which is consistent with management declining to raise it further.

What They're NOT Saying

1. Any 2027 cost number: the direction was volunteered and the magnitude withheld, and a published consensus of roughly 3% growth to about $35.4bn was put to management directly and left unaddressed. Disclosing that investment will be higher while refusing to size it is the least helpful combination available.
2. Whether the accelerated investment is larger than the incremental savings: asked explicitly, answered with a precedent rather than a figure. The word "partly" in "partly offset" is the only sizing the call produced.
3. Updated private-credit and securitisation exposure: the first-quarter release quantified roughly $6bn of total private-credit exposure and a roughly $3bn securitisation-financing bucket after the fraud charge. Neither figure was refreshed in the interim materials, and no analyst asked, so the market cannot see whether the book grew or shrank.
4. Any buyback framework beyond one quarter: repurchase decisions remain quarterly with no annual envelope, no capital threshold that triggers a larger tranche, and no statement of what "up to $1bn" implies for the run rate. The line the market most wants to model is the line least specified.
5. Debt and equity markets: we flagged the first quarter's 26% decline in this line as the thing to watch. The interim disclosure does not break it out, so a stated watch item is unresolved rather than closed.
6. Singapore Life's share of the insurance contractual service margin: HSBC is selling an insurance manufacturer inside the wealth franchise it calls its primary growth engine, and the $15.2bn contractual service margin balance disclosed in the first quarter was neither restated nor adjusted for the sale.
7. When the Middle East reserve becomes releasable: management set a policy of two clear quarters of stability before releasing reserves and simultaneously removed the standalone Middle East scenario by folding it into the base downside. Those two statements point in different directions on the timing, and a release is a second-half tailwind nobody has been told to expect.

Market Reaction

• Pre-print setup: the ADR closed at $107.86 on 3 August, up 37.1% year to date, up 74.6% over the trailing twelve months and up 11.4% over the trailing 30 days, at the very top of a $61.30 to $107.86 52-week closing range. The S&P 500 was up 11.0% year to date entering the print. There was no valuation cushion for a surprise in either direction.
• Reaction session (4 August, results released before the US open): the ADR opened at $106.70, a 1.1% gap down, traded a $105.91 to $107.85 range and closed at $106.04, down 1.7% or $1.82, on 2.1M shares against a 1.3M 30-day average, roughly 1.6 times normal.
• Relative: the S&P 500 rose 1.8% on the session, making the ADR's move a 3.5 point relative underperformance.
• Following session (5 August): the ADR closed at $102.68, a further 3.2% decline, on 2.8M shares, heavier volume than the reaction day itself. Cumulative decline from the pre-print close: 4.8%.
• Since: the ADR closed at $103.21 on 11 August, 4.3% below the pre-print close and 2.7% below the reaction-day close.
• Primary listing: the Hong Kong line traded broadly flat on 4 August after setting an intraday record of HK$169.50 earlier in the same session.

The two-session shape carries the information, and it is the reverse of the first quarter's. In May a 4.1% opening gap closed at down 3.0%, meaning the market recovered a third of its initial reaction while it read the release. This time a modest 1.7% first-day decline was followed by a heavier 3.2% second-day decline on more volume. A one-day fade on a 6% pre-tax beat is a shrug at a stretched price. A larger second-day fade on larger volume, after the sell side has had a night with the transcript, is a decision.

What changed overnight was not the print. Day one traded the visible items: the beat, the buyback restart, the net interest income raise, credit improving. Day two traded the read-through, and there were three things to read. The buyback landed at less than half the expected size. The medium-term targets were reiterated rather than raised off a first half that beat them. And management pre-announced that 2027 costs go up because revenue is going up, while declining to say by how much. That third item is the underrated mover. A bank telling you it will spend more because the growth is coming is asking you to underwrite operating leverage you cannot yet model, into a multiple with no room for the exercise.

The setup amplified all of it. Entering the print up 74.6% over twelve months, at the top of its 52-week closing range, with the Hong Kong line printing a record intraday the same morning, this was a stock positioned for a raise rather than a reiteration. Read that way, the 1.7% first-day move understates the disappointment in the guide and overstates it in the print.

Street Perspective

Debate: Is the $1bn Buyback Discipline or a Downgrade of the Capital-Return Story?

Bull view: Capital return resumed three quarters after a $13.7bn acquisition, from a CET1 ratio that rose 10 basis points in the same quarter the bank funded $20bn of loan growth. Buybacks were always described as the residual behind the dividend and organic growth, and the residual is smaller because organic growth got better, which is the outcome a long-term shareholder should prefer. The declared dividend also understates the entitlement: about $4bn of first-half accrual has not yet been declared.

Bear view: Half the expected size, with quarterly decisions, no envelope and no threshold, from a bank whose tangible book per share has now fallen two quarters running and whose share count rose in a quarter with no repurchases. Inorganic opportunity still ranks ahead of buybacks in management's own stated stack, and the chief executive was invited on the call to discuss using the shares as acquisition currency and did not decline the premise.

Our take: Discipline, and the bears are right that it is not free. At 2.2 times tangible book, funding 10% commercial loan growth beats repurchasing stock on arithmetic, and a management team that did the opposite would deserve the criticism instead. But an unspecified repurchase line is an unmodellable one, and the practical consequence is that total shareholder yield here is roughly 4.3% of dividend plus about 1.1% of buyback if the quarterly pace repeats, which is a good yield and not a re-rating catalyst.

Debate: Does "At Least $46bn" Mean $46bn or $47bn?

Bull view: The first half delivered $22.9bn, so the guide asks for about $11.55bn a quarter against $11.6bn just printed, which means it embeds no growth at all. Layer in the reinvestment of $50bn of hedge assets yielding 2.8%, day count, 8% deposit growth and 6% loan growth, and the second half should exceed the first. Management conceded the arithmetic on the call for the second consecutive quarter.

Bear view: HIBOR at 2.62% and falling matters more for this bank than for any peer, the second quarter benefited $0.1bn from a first-quarter one-off not recurring, and a weaker sterling translates directly into a dollar-reported net interest income line. The current run rate has still not recovered the fourth-quarter 2025 level of $11.8bn.

Our take: The bulls, and the halving of Hong Kong dollar rate sensitivity settles it. The single largest named risk to the guide was HIBOR, and management has materially hedged it, partly through fixed-rate customer lending. We carry $46.0bn to $46.5bn and treat the guide as a floor. This is the most reliable upgrade available from the print and it is also, at roughly 1% of group revenue, not large enough to change the rating on its own.

Debate: Is Accelerating 2027 Investment a Growth Signal or the End of the Cost Story?

Bull view: A management team that raised its savings target 33% on an unchanged cost envelope, delivered its original target early, and is pulling forward previously earmarked investment because revenue is running ahead has earned the benefit of the doubt. The securities services precedent shows intra-year payback is achievable. Spending into demonstrated demand is what a growth bank should do, and the first-half revenue outperformance raises the base from which 2027 is measured.

Bear view: The cost programme was the most verifiable part of the equity story, and it has just been reframed from a savings narrative into a reinvestment narrative with no number attached. Management refused to engage a published 3% consensus, said the offset would be only partial, and separately flagged additional 2026 variable pay against a $3.9bn pool. The 2026 guide already requires flat second-half costs year over year.

Our take: The bears have the better of the near term and the bulls may well be right about 2028. The specific problem is sequencing: the incremental savings and the reallocated divestment costs arrive in tranches of roughly $0.3bn and $0.4bn, while the investment acceleration is undefined and management's own word for the offset is "partly." We carry 2027 target-basis cost growth at 3.5% to 4.5%, above consensus, and treat a guide of above 4% at the full-year results as a downgrade trigger.

Debate: Does 2.21x Tangible Book Leave Anything for the Shareholder?

Bull view: A bank returning 19.1% on tangible equity in the first half against a 17%-plus target, with all four businesses above 18%, revenue accelerating, credit improving, a raised savings target and Hang Seng synergies still absent from the numbers, should trade at a large premium to tangible book. On tangible book per share that exits 2026 near $10.00 rather than today's $9.36, the multiple is closer to 2.06 times, and a 4.3% forward dividend yield does much of the work.

Bear view: The multiple went from 1.85 times to 2.21 times in three months while tangible book per share fell twice, which means the entire move was price rather than value. At 2.21 times with 3% growth, the market requires a sustainable 18.4% return at a 10% cost of equity, above the company's own target and roughly at its first-half peak. On the 17%-plus target the implied cost of equity is about 9.3%, which is not a defensible number for a bank with Hong Kong property, Middle East and mid-market credit tails.

Our take: The bears, and this is the debate that sets the rating. Every operating criticism we made in May has been answered, and the market answered it first and charged us for it. The honest counterweight is the bull point on exit book: crediting second-half tangible book accretion takes the multiple to roughly 2.06 times and makes the skew far less adverse than it was at initiation, which is why this is a Hold rather than an Underperform. But a stock cannot be an Outperform when the entire operating improvement has already been paid for and the residual return is a 5.4% shareholder yield against a multiple that must simply hold.

Model Update & Valuation Framework

Valuation. The ADR closed at $103.21 on 11 August. Each ADR represents five ordinary shares, and tangible net asset value was $9.36 per ordinary share at 30 June, so tangible book is $46.80 per ADS and the stock trades at approximately 2.21 times. Holding growth at 3%, that multiple requires a sustainable 18.4% return on tangible equity at a 10% cost of equity, or 20.6% at 11%. Inverted against the company's own 17%-plus target, 2.21 times implies a cost of equity of roughly 9.3%, against roughly 10.6% at the first quarter's 1.85 times. Annualising the second quarter's $0.45 of earnings excluding notable items gives $9.00 per ADS and a multiple of 11.5 times.

That comparison is the single most important number in this note. Over three months the operating case improved on every measure we said to watch, and the price paid for the 17% target went up by more than the improvement, because the ADR rose 18.1% from the first-quarter reaction close of $87.40 while tangible book per ADS fell from $47.30 to $46.80.

Framework rather than a target. We continue not to publish a price target. On spot tangible book of $46.80 per ADS, a return to the 1.85 times multiple of three months ago implies $86.58, about 16.1% below the current price; 2.0 times implies $93.60, about 9.3% below; and 2.5 times implies $117.00, about 13.4% above. On our estimated exit-2026 tangible book of roughly $50.00 per ADS the same bands read very differently: 2.0 times implies $100.00, about 3.1% below the current price, 2.25 times implies $112.50, about 9.0% above, and 2.5 times implies $125.00, about 21.1% above. The gap between those two sets of bands is the entire bull case, and it rests on tangible book resuming its build in a second half that pays no large final dividend.

Total return arithmetic for the Hold. A 4.3% forward dividend yield, plus roughly 1.1% of buyback if the quarterly pace repeats, gives about 5.4% of shareholder yield. Add high-single-digit growth in earnings excluding notable items and hold the multiple at roughly 2.06 times exit book, and the twelve-month total return lands in the low double digits. That is a respectable outcome and it is approximately in line with what the index should deliver, which is the definition of the rating.

What would change it. An upgrade requires two of the following three: a third-quarter buyback at or above $1.5bn, which would show the residual growing rather than tokenising; second-half target-basis costs landing flat year over year, which would validate the guide against the two pressures management flagged; and the credit charge holding at or below 42 basis points with the Middle East reserve released. A de-rate toward 2.0 times spot tangible book without operational deterioration would also do it. A downgrade requires any of: 2027 target-basis cost growth guided above 4%, a second sponsor-intermediated or private-credit charge, the credit guide raised again, or quarterly net new money below $20bn.

Thesis Scorecard Post-Earnings

The pillars below are the ones established at initiation in May and carried in the standing thesis, graded against what this quarter's print and call showed. Status tags follow the thesis convention: bull pillars run on track, at risk or broken; bear points run contained, emerging or materialising.

Overall: The thesis is stronger than it was at initiation on five of six pillars and weaker on the one that sets the rating. Two bear points de-escalated, one bull pillar upgraded from at risk to on track, and the operating uncertainty that made May's Hold a genuinely balanced call has largely resolved. What replaced it is a single, cleaner constraint: price.

Action: Hold, with conviction raised to 7 from 6. Own the franchise. The conviction increase is in the call itself rather than in the upside, because a Hold justified purely by valuation is a more robust position than a Hold justified by unresolved operating questions.

Bottom Line

In May we wrote that the second quarter would settle the two questions that mattered, the buyback and the credit trajectory, and that both would resolve in a direction that was knowable rather than speculative. Both resolved favourably. Credit came in at 41 basis points against the roughly 43 the guide required, with no repeat of the fraud charge. The buyback restarted. Revenue growth accelerated to 7% with all four businesses growing and each returning better than 18%. The simplification target went up by a third on the same cost envelope, and all three open strategic reviews closed. We published three upgrade criteria and two were met.

We are not upgrading, and the reason is worth stating plainly rather than dressing up. The variable we did not put in the criteria was the multiple, and it moved further than the fundamentals did. Three months ago the ADR was $87.40 on tangible book of $47.30 per ADS, or 1.85 times, which capitalised the 17%-plus target at roughly a 10.6% cost of equity. Today it is $103.21 on $46.80, or 2.21 times, which capitalises the same target at roughly 9.3%. The improvement is real and it has been paid for in advance, by a market that bid the stock 18% higher across a quarter in which tangible book per share fell.

The forward case is not bad, and the exit-2026 book anchor of roughly $10.00 per ordinary share, which takes the multiple to about 2.06 times, is why this note reads as a Hold and not an Underperform. A 4.3% dividend yield with a buyback restarting and high-single-digit earnings growth is a perfectly reasonable thing to own. It is also, arithmetically, an in-line return that depends on a demanding multiple simply holding, at a bank that has just told the market its costs go up next year and declined to say by how much. We maintain Hold and shift the watch list from credit to costs.