The Guidance Raise Landed and the Re-Rate Landed With It: Downgrading HSBC to Hold at 1.95x Tangible Book

Results vs. Consensus

HSBC's annual results are a different kind of event from its quarterly trading updates. The Group Chief Executive presents, the three-year targets get re-set, the Annual Report and Accounts publishes the same morning, and the fourth quarter itself is not disclosed as a standalone income statement. Everything below that describes a fourth quarter is either the company's own quarterly summary from the results presentation or is derived as the full year less the nine months published in October, and the two agree to the presentation's rounding on every line.

The benchmark for the print is HSBC's own compiled analyst consensus, which is published at group level only. There is no company-compiled fourth-quarter or segment consensus, so the scorecard below measures the full year against consensus and the fourth quarter against the guidance in force and the prior-year quarter.

FY 2025 Scorecard

Three mechanical notes before the tables that follow. First, HSBC's headline measures are constant currency excluding notable items, and its statutory measures are reported currency. Where a row is labelled constant currency it is compared against a constant-currency comparative. Second, each ADS represents five ordinary shares, so the $1.51 of basic earnings excluding material notable items becomes $7.55 per ADS and the $9.64 of tangible net asset value becomes $48.20 per ADS. Third, the quarterly return series in the results presentation is stated on an updated methodology: the third quarter, reported at 16.4% in October, is now shown at 16.6%, and the second quarter has moved from 17.7% to 17.8%. The revisions are small but they are revisions, and the sequential comparison below uses the restated series.

Full Year: 2025 vs. 2024

Two lines explain the whole shape of the year. Reported operating expenses rose 9.9% and target-basis operating expenses rose 3.0%; the difference is $2,964M of notable items inside the cost line, chiefly $1.4bn of legal provisions and $1.0bn of restructuring. And reported pre-tax profit fell 7.4% while the constant-currency measure excluding notable items rose 7.1% to a record. The gap between those two numbers is $6.7bn of notable items in 2025 against $1.8bn in 2024, a $4.9bn year-on-year swing, of which $2.1bn is the Bank of Communications dilution and impairment, $1.5bn is the reserve recycling on the completed French home-loan sale, $1.4bn is legal provisions and $1.0bn is restructuring. Statutory net interest income of $34,794M was $2.1bn higher than 2024, but $1.7bn of that increase is lower trading-book funding costs that reverse inside trading revenue. Banking net interest income, which strips those costs out, grew $0.5bn.

Fourth Quarter: 4Q25 vs. 4Q24

The 39% revenue increase and the near-tripling of reported pre-tax profit are almost entirely a comparative artefact: the fourth quarter of 2024 carried $5.2bn of recycled foreign currency losses on the completion of the Argentina sale, and the fourth quarter of 2025 carried $1.5bn on the French home-loan portfolio. Strip both and revenue grew 6.0% and pre-tax profit 16.2%. That underlying 16.2% is the number worth holding, because it is the highest year-on-year growth rate in any quarter of 2025 and it comes from a quarter in which the cost line was seasonally heaviest.

Fourth Quarter vs. Third Quarter

The sequential view says the revenue engine accelerated and the cost line absorbed it. Banking net interest income rose 6.1% in a single quarter, of which $0.4bn came from Hong Kong on the HIBOR recovery and roughly $0.1bn is a benefit management told the market not to repeat. Against that, target-basis costs rose 6.3% sequentially to $8,905M on the fourth quarter's usual freight: the UK bank levy, performance-related pay and the technology spend that runs heaviest at year end. The net effect is that pre-tax profit excluding notable items stepped down 5.5% and the annualised return excluding notable items stepped down 70 basis points to 15.9%, which is below the 17% floor the bank has just adopted for 2026.

Revenue assessment. A 1.4% revenue beat against the bank's own compiled consensus is thin, and the pre-tax beat of 3.6% is the more useful figure because it captures the credit and cost lines as well. What matters more than either is the shape: revenue excluding notable items grew 5.1% for the year and 6.0% in the fourth quarter, and management has now set a target that requires growth in every one of the next three years rising to 5%. The bank delivered 5% in 2025 and 4% in 2024, so the 2028 target is the level it has just achieved rather than an acceleration. On the call the Group Chief Executive said as much when pressed on what is special about 2028.

Margin and cost assessment. Net interest margin of 1.59% is three basis points above 2024 and the structural hedge has grown to $593bn with an average life of roughly 3.1 years. The reinvestment schedule is the most durable thing in the release: approximately $110bn matures in 2026 and is replaced at around 2.7%, $110bn in 2027 at around 3.4%, and $100bn in 2028 at around 3.7%. That is a rising yield on a growing balance for three consecutive years, and it is the mechanism that lets management guide banking net interest income up in a year with two to three rate cuts assumed. The offsetting sensitivity is real and disclosed: a 100 basis point parallel down-shock costs $3,389M of banking net interest income in year one on a static balance sheet.

Earnings assessment. Reported earnings per share fell 3.2% and reported return on tangible equity fell 130 basis points, and neither is the number this bank is run to or valued on. The number that matters for the dividend is basic earnings per share excluding material notable items, which is the base for the 50% payout, and it grew 15.3% to $1.51 per ordinary share or $7.55 per ADS. On that base the declared dividend of $0.75 is a 49.7% payout. The uncomfortable part is that a fifth of the growth came from a share count that no longer shrinks, and the fourth quarter's annualised return excluding notable items of 15.9% sits below the new 17% floor.

Segment Performance

HSBC assesses its segments on a constant-currency basis, so the table below is constant currency on both sides and the 2024 column uses constant-currency comparatives, which is why the group total reads $66,009M rather than the $65,854M reported. Every row reconciles: revenue less expected credit losses less operating expenses plus the associate line equals segment pre-tax profit.

The associate line sits below the rows shown: Corporate Centre carries $1,886M, chiefly the Bank of Communications and Saudi Awwal Bank stakes net of the $2.1bn dilution and impairment, IWPB carries $24M and CIB $1M. The group reconciles at $68,274M less $3,850M less $36,428M plus $1,911M equals $29,907M.

Returns are the more useful capital-allocation view, and the table below adds the nine-month figures published in October so the fourth-quarter direction is visible.

Deposits grew in all four businesses: Hong Kong to $543bn on growth of $37bn, CIB to $598bn on $10bn, UK to $364bn on $11bn and IWPB to $281bn on $9bn, taking the group to $1,787bn. The $78bn of group growth includes $16bn of held-for-sale balances at the fourth quarter of 2025 and $6bn at the fourth quarter of 2024, which are not in the $1,787bn absolute figure. Note that three of the four businesses show a lower full-year return than they showed at nine months, because the fourth quarter is the seasonally heaviest cost quarter. Only Hong Kong improved.

Hong Kong

Still the highest-returning business in the group by a wide margin and the only one whose return improved through the fourth quarter, to 35.5% excluding notable items from 34.5% at nine months. Revenue rose 5.5% to $15,878M and pre-tax profit 4.9% to $9,576M. Deposits grew $37bn, the largest increase of the four businesses, to $543bn, and management put the deposit market share at 25%. The expected credit loss charge of $1,476M is the largest of the four segments and reflects the Hong Kong commercial real estate book.

The customer-acquisition disclosure that dominated the third quarter looks different at the full year. HSBC added 1.1 million new-to-bank customers in Hong Kong across 2025, taking the total past 7 million. It had already reported more than 900,000 in the first nine months, so the fourth quarter contributed roughly 200,000 against a nine-month run rate near 300,000 a quarter. Management addressed the deceleration directly and pre-announced that it will continue.

"We have added a fee for the new to bank customers who have a balance less than HKD10,000, and because of that, we expect that there would be slightly slower acquisition in 2026." — Pam Kaur, Group Chief Financial Officer

Assessment: A deliberate quality-over-quantity pivot, disclosed before the market could infer it, and the right trade if the affluent mix improves. It also removes the single most quotable growth statistic the bank had. From here Hong Kong has to be underwritten on deposit share, wealth conversion and the 35.5% return, not on customer counts.

Corporate & Institutional Banking

The segment we flagged in October as the least defensible line in the group, and nothing in this release answers it. Revenue grew 3.2% to $27,637M and pre-tax profit 0.9% to $11,386M, on operating expenses of $15,556M and a cost efficiency ratio of 56.3%, the worst of the four businesses. The full-year return excluding notable items is 16.2%, down from 16.8% at nine months and the lowest in the group, on a segment that carries $305bn of customer loans and $598bn of deposits. Fee and other income of $13.1bn grew 7%, and wholesale transaction banking fee and other income of $10.9bn grew 4% for the year, but in the fourth quarter that line was flat at $2.6bn with global trade solutions down 5%.

No analyst asked about the CIB cost line at this results event, as none did in October.

Assessment: A business absorbing more than a third of the group's tangible equity at a 16.2% return, with the return falling through the fourth quarter and a cost ratio 13 percentage points worse than the UK's, is the structural drag on the new 17% target. Management has now committed to a return floor that CIB does not currently clear. Either the segment improves or the mix has to shift further towards Hong Kong and wealth, and the second of those is what the Hang Seng transaction is for.

International Wealth & Premier Banking

The fastest-growing segment on profit, up 10.0% to $4,367M on revenue up 5.1% to $14,520M, and the purest read on the wealth strategy outside Hong Kong. Return excluding notable items is 19.0% for the year, second-highest in the group but down from 19.7% at nine months. The bank opened a further 20 wealth centres in the segment during 2025, focused on Asia and the Middle East, and reported double-digit invested-asset growth in mainland China, India, Singapore and the UAE. The Group Chief Executive put the whole-bank total at 64 wealth centres after 26 or 27 openings across 2025.

The insurance manufacturing line is the machine underneath. Wealth total insurance revenue grew 36% to $2,599M, of which the contract service margin release was $1,593M, up 19%. New business contract service margin grew 35% to $3,405M and the balance ended at $14.6bn, up 21% from $12.1bn, after $(0.4)bn of release in the fourth quarter and a $(0.1)bn reclassification of UK balances to held for sale.

Assessment: The contract service margin is the reason to believe the wealth revenue line persists: it is a stock of unearned insurance profit that releases into revenue over years, and it grew 21% even after a disposal removed part of it. That is a genuinely compounding asset and it is the strongest single item in the release. The offset is that the same segment's return fell 70 basis points through the fourth quarter, and that the sale of HSBC Life Singapore remains under strategic review with no disclosure of what share of the $14.6bn balance would go with it.

UK

The only segment whose pre-tax profit fell, down 1.7% to $6,705M on revenue up 4.8% to $12,938M. Costs of $5,537M and a credit charge of $696M absorbed the revenue growth. The return excluding notable items of 22.9% is down from 23.3% at nine months but remains second-best in the group, and this is where the loan growth is: customer loans grew 6% to more than $300bn, business banking lending grew 13% excluding the run-off of pandemic-era loans, and the bank added roughly 60,000 new Premier customers, up around 70% year on year, taking the UK Premier base past one million.

Assessment: A segment being deliberately spent down to buy growth, and so far the growth is showing up. The UK is the only place in the group where the loan book is genuinely expanding, which matters more than it looks: the group loan-to-deposit ratio fell again to 55.3%, so incremental deposits are going into the structural hedge rather than into lending spread. If the UK book keeps compounding at 6%, that is the one lever that improves the mix of the balance sheet rather than just its size.

Corporate Centre

Swung from a $1,180M constant-currency profit to a $2,127M loss, carrying $2,664M of the group's revenue-line notable items and the Bank of Communications dilution and impairment inside a $1,886M associate contribution. Revenue excluding notable items in this segment is $(35)M against $(375)M a year ago.

Assessment: Corporate Centre remains where the noise lives, and in 2025 it lived there loudly. The structural point from October has not changed: a meaningful block of the group's tangible equity sits outside the four operating businesses, and management has not addressed it in either the results or the new targets.

Wealth Franchise KPIs

Assessment: Read the two flow rows together, because they are the most consequential disclosure in the release. On the definition the bank has used throughout our coverage, net new invested assets were $7bn in the fourth quarter against $29bn in the third, and Asia was negative $3bn. On the definition that replaces it in the first quarter of 2026, net new money was $26bn with Asia at positive $19bn. The difference is deposits. The new measure adds $608bn of Premier and Private Bank deposits and removes $580bn of asset-management third-party distribution, and on that basis the full-year net new invested-asset figure is $46bn rather than the $80bn headline. The bank is not hiding this: the $46bn is printed in a footnote and the reconciliation runs across three appendix slides. But the effect is that 42.5% of the flow number that carried the bull case for four quarters turns out to have been third-party distribution rather than wealth-customer money, and the metric is being retired in the quarter that fact became visible.

Key Topics & Management Commentary

Overall Management Tone: Confident and unusually specific, with the Group Chief Executive present for the first time in our coverage and the strategic argument delivered as arithmetic rather than assertion. The contrast with October is sharp: the Hang Seng transaction, which management was legally barred from valuing four months ago, was walked through line by line against an explicit buyback threshold. Where management was least forthcoming was on trajectory, refusing three separate attempts to say whether 17% is a floor that rises or a level that holds, and on the wealth disclosure change, which was presented as a comparability improvement and drew no question.

1. The New Targets: Raised, Extended, and Set at the Level Just Delivered

The mid-teens return target that covered 2025 to 2027 has been replaced by 17% or better for each of 2026, 2027 and 2028, alongside a new revenue target of year-on-year growth every year rising to 5% in 2028, and a reaffirmed 50% payout. This is the guidance raise we said in October would decide the case, and it arrived on schedule.

"We will target revenues growing year on year every year, rising to 5% in 2028, excluding notable items. We will target return on tangible equity of 17% or better in each year from 2026 to 2028, excluding notable items, and a dividend payout ratio of 50% for each year, excluding material notable items." — Georges Elhedery, Group Chief Executive

Two features of the construction matter. The return target is set at 17% against a delivered 17.2%, so it codifies the run rate rather than extending it, and the fourth quarter annualised at 15.9%, below the new floor. The revenue target reaches 5% in 2028 against 5% delivered in 2025 and 4% in 2024, which prompted a direct question about what changes in 2028.

"I would look at 2028 not specifically as the year 2028, but as what we believe our long-term structural opportunity to grow. It's our guidance for 2028, but we've delivered 5% revenue growth in 2025. We've delivered 4% in 2024. So it's just the footprint we're in, and the capabilities for us to capture the growth is there." — Georges Elhedery, Group Chief Executive

Assessment: The honest reading is that these are floors set at levels management is confident it can clear, which is exactly the guidance style we described in October and which the bank then beat on every 2025 line. The information content is in the duration, not the level: committing to 17% in each of three years, including a 2026 in which two to three rate cuts are assumed, says more about the durability of the deposit franchise than a higher one-year number would. What it does not do is give the market anything to upgrade towards.

2. Banking Net Interest Income: $44.1bn Delivered, and a 2026 Floor Management Priced Below Its Own Run Rate

Full-year banking net interest income of $44,084M cleared the "$43bn or better" guide by $1.1bn. The fourth quarter printed $11,722M against the $10.6bn floor management named in October, and the finance chief flagged approximately $100M inside it as non-repeating.

"On banking NII, you're right; there was a $100 million non-repeat items. If you take that out for modelling purposes, you come to $11.6 billion. There are no other one-offs." — Pam Kaur, Group Chief Financial Officer

A clean $11.6bn annualises to $46.4bn against a 2026 guide of at least $45bn. Pressed on that gap at the end of the call, management conceded the arithmetic without moving the guide.

"If you underpin that just based on the fourth quarter, obviously, it will give you a number of 46, but we are not calling that out because we are very cognisant of the headwinds on interest rates." — Pam Kaur, Group Chief Financial Officer

The disclosed offsets are specific: a $300M first-quarter headwind from two fewer days, US rates already 50 basis points lower year to date and sterling 25 basis points lower, with two to three further cuts assumed on the end-January forward curve. Against those, the structural hedge at $593bn reinvests approximately $110bn during 2026 at around 2.7%.

Assessment: Treat $45bn as a floor, as $43bn was. We carry $46.0bn. The more important disclosure is the hedge schedule, which converts a rate-cut cycle into a three-year tailwind: the 2027 and 2028 reinvestment tranches roll on at 3.4% and 3.7% respectively, both above the 2026 rate. That is the mechanism that makes a 17% return plausible in years two and three of the target, and it is contractual rather than forecast.

3. Costs: 3% Delivered, 1% Guided, and Where the Room Comes From

Target-basis costs grew 3.0% to $33,464M, exactly on guidance, in a year when reported costs grew 9.9%. The 2026 guide of approximately 1% on a $33.5bn baseline implies roughly $33.8bn, an increase of about $335M. That is a demanding number in a year of accelerating technology investment, and management explained where the room comes from.

"I would note that we are ahead on our simplification saves because our plan and the actions we have taken have come through sooner than what we had originally outlined. This gives us incremental saves of $700 million in full-year 2026." — Pam Kaur, Group Chief Financial Officer

The programme is running well ahead of the plan disclosed a year ago: $1.2bn of annualised savings actioned against $1.0bn guided by year-end 2025, $0.6bn realised in the 2025 income statement against $0.3bn guided, and the remaining $0.3bn to be actioned in the first half of 2026, six months early. Expected realisation is approximately $1.3bn in 2026 and $1.5bn in 2027. Underneath, headcount excluding technology fell by approximately 4,400 in 2025 while technology headcount rose by approximately 1,800, managing director positions fell approximately 15%, and 1,165 non-strategic applications were retired.

"Half of that additional cost will go towards payroll inflation, but the other half will go towards additional investments." — Georges Elhedery, Group Chief Executive

Assessment: Add the $0.7bn back and underlying target-basis cost growth in 2026 is roughly 3.1%, the same rate as 2025. The 1% guide is therefore a savings guide, not a spending guide, and its credibility rests entirely on a programme that has so far over-delivered on every disclosed milestone. That is a reasonable bet, and it is also the single line where a miss would be most visible against the new 17% floor, because the fourth quarter already annualises $1.8bn above the guided level before any seasonal unwind.

4. Wealth: 24% Fee Growth, and a Disclosure Change That Resets the Flow Metric

Wealth fee and other income grew 24% for the year to $9.4bn and 20% in the fourth quarter to $2.1bn, with all four income lines growing. On the same morning, the bank announced that from the first quarter of 2026 it will replace invested assets with wealth balances, and net new invested assets with net new money.

"We have added $608 billion of Premier and Private Bank deposits to the Invested assets. That is offset by taking out $580 billion of Asset Management third-party distribution assets. This is a good business, but it does not reflect our Wealth customers. Adjusting our disclosure in this way also means our Wealth business is more easily comparable to the broader peer group." — Pam Kaur, Group Chief Financial Officer

The stated rationale is sound and the reconciliation is fully disclosed across three appendix slides. The consequence is also arithmetic: on the new definition, full-year net new invested assets were $46bn and net new deposits $40bn, against a headline $80bn of net new invested assets on the old definition. The $34bn difference is asset-management third-party distribution. In the fourth quarter specifically the old measure showed $7bn with Asia at negative $3bn, while the new measure showed $26bn with Asia at positive $19bn.

Management also removed the standing "double-digit percentage CAGR" wealth fee guidance and withdrew any assumption that 2025's market conditions repeat.

"Last year, there was a great advantage of a tailwind coming from markets, which gave us a lot of uplift on the transactional-based fee income. We can't assume that for this year." — Pam Kaur, Group Chief Financial Officer

Assessment: The revenue is real, the contract service margin build is real, and the deposits being added to the measure are genuinely wealth-customer money. What has changed is our confidence in the specific claim we underwrote in October, which was that the wealth build is flow-driven rather than market-driven. Market moves supplied $125bn of the $242bn increase in invested assets across the year, the wealth-customer flow figure is 42.5% smaller than the headline, and Asia's quarterly flow on the old basis went negative. This does not break the bull case; it removes the part of it that was doing the most work.

5. Hang Seng: The Arithmetic Arrives, Four Months Late and One Quarter Early

The transaction we criticised in October as an unpriced assertion completed on 26 January, ahead of the first-half expectation, and this release gives the number we said was missing. The $13.7bn purchase price less $3.8bn of disallowed minority capital gives $9.9bn of common equity tier 1 consumption, which management sized against a buyback for the first time.

"The $9.9 billion CET1 consumption is equivalent to buying back 4% of group shares at the point of announcement. Next, we show the $0.8 billion minority interest in the P&L and the $0.5 billion of pre-tax synergies from the privatisation. Together, the minority interest and the synergies contribute more than 4% to our profit, beating the buyback threshold." — Pam Kaur, Group Chief Financial Officer

Above the $0.5bn of accounting-recognised synergies sits $0.4bn of further ambition, giving $0.9bn of gross benefits by the end of 2028, split broadly evenly between revenue and cost, against a $0.6bn restructuring charge that will be booked as a material notable item. The split between the two buckets was explained on the call.

"The $500 million is what I would call the reported synergies following accounting rules. The $400 million synergies are depending to some extent on markets and customer behaviour, so there is some degree of uncertainty, and they don't strictly fall between what is considered to be accounting reported synergies." — Pam Kaur, Group Chief Financial Officer

Hang Seng's own numbers are now disclosed: full-year 2025 revenue of $5,336M, an expected credit loss charge of $1,032M against $612M in 2024, and pre-tax profit of $2,297M, down from $2,693M. Its loan book shrank to $101,148M from $105,475M.

Assessment: This is the disclosure we asked for and it is a good one. The buyback threshold framing is falsifiable, the restructuring cost is quantified, and the accounting and ambition buckets are separated rather than blended. The case against has narrowed to two points: the $0.4bn is management ambition rather than a recognised synergy, and the business being consolidated saw its credit charge rise 69% in 2025 while its loan book shrank 4%. Management's answer to the asset-quality question was that the cycle will normalise, which is a view rather than a plan.

6. Credit: 39 Basis Points, and a Hong Kong Property Book That Stopped Migrating

The full-year charge of $3,850M was 39 basis points of average gross loans, inside the approximately 40 basis point guide, and the fourth quarter ran at 36 basis points. Hong Kong commercial real estate cost $0.7bn for the year against $0.1bn in 2024, and mainland China commercial real estate $0.2bn against $0.4bn. Management guided 2026 to approximately 40 basis points and described that as "at the higher end of our typical range", retaining a 30 to 40 basis point medium-term planning range.

Inside the Hong Kong book, the migration we flagged in October paused. Credit-impaired balances were $6.3bn at year end against $6,357M at the third quarter, on a book that shrank to $30.6bn from $31,033M, so the impaired share was essentially unchanged at 20.6%. Sub-standard balances fell to $2.4bn from $3.7bn a year earlier, the secured share of the portfolio rose to 57% from 54%, and the total expected credit loss allowance rose to $1.1bn from $0.4bn. The company's own slide says the pace of migration "slowed in the 4th quarter".

The uncomfortable line is collateral. Average loan-to-value on credit-impaired exposures rose to 71% from 58%, and credit-impaired exposure above 70% loan-to-value grew to $1.9bn from $1.2bn.

"Office is a sector we are watching very carefully because recently we have seen, both in terms of rental demand and in transactions for the best properties with the best spec in Central, in the best areas, there are some green shoots. The downside is, at this point of time, vacancy rates are still around 17%, so that's what we will be managing through and then following up very carefully." — Pam Kaur, Group Chief Financial Officer

On the residential component management declared the problem essentially over, citing the Hong Kong price index up 5% year on year with sales volumes up 10% and rentals stabilised on both demand and pricing. Retail sales turned positive and are running 6.6% higher year on year, though management was explicit that this alone will not fix a sector whose oversupply and changed shopping patterns are structural.

Assessment: Better than we modelled. We carried 2026 group credit cost at 42 to 45 basis points on the assumption that fourth-quarter migration would continue; it did not, and the guide is approximately 40. We revise to 40 to 42 basis points. The residual worry has changed shape rather than disappeared: it is no longer the rate at which loans migrate into stage 3, it is the collateral behind the ones already there, where average loan-to-value has moved 13 percentage points in a year on a market with 17% office vacancy.

7. Capital: 14.9%, Then Roughly 13.8%, and a Buyback With a Condition Rather Than a Date

CET1 rose 40 basis points in the quarter to 14.9%, on approximately 80 basis points of regulatory profit less approximately 40 basis points of dividend, with risk-weighted assets up $10bn to $888.6bn including an approximately $12bn operational-risk increase driven by higher average income. In January the Hang Seng privatisation took a net 110 basis points, comprising an approximately 120 basis point day-one impact partly offset by a 10 basis point release of hedging-related risk-weighted assets, leaving the ratio around 13.8% against a 14.0% to 14.5% target range. The tangible book impact is $6.7bn.

The buyback framing changed in a way that is easy to miss. In October the suspension was time-bound, "up to three quarters". It is now condition-bound.

"We expect to restore our CET1 capital ratio within our target range through a combination of organic capital generation and not initiating any further buy-backs until CET1 capital is back within, or above, this range." — from the results announcement

Asked whether a first-quarter return to the range could bring the buyback forward, management restated the process without moving the timing, but did make one notable statement of principle.

"Share buyback remains for us a useful tool to deploy surplus capital, irrespective of where the share price is." — Pam Kaur, Group Chief Financial Officer

Assessment: At approximately 80 basis points of regulatory profit per quarter against approximately 40 basis points of dividend, the bank rebuilds roughly 40 basis points a quarter and needs about three to refill a 110 basis point hole, which is the three-quarter framing arriving from the other direction. That points to a resumption no earlier than the second quarter and probably modest when it comes. The "irrespective of where the share price is" line deserves attention from a shareholder buying at 1.95x tangible book: it is an explicit statement that the bank will not treat its own valuation as an input to the repurchase decision.

8. The Disposal Drag Finally Gets a Number

In October we said the disclosure we would most like to see and expected least was what the eleven divested businesses were earning. It arrived. Completed and announced exits will generate $0.7bn of annualised cost savings with approximately $1bn of associated revenue, and exits in active execution, including businesses under strategic review, are expected to generate a further $0.6bn of cost. Four of the eleven have completed. Australia, Indonesia and Egypt retail, plus the Singapore insurance manufacturing business, remain under review.

The 2026 material notable items from these transactions are also now scheduled: Malta $(0.4)bn in the first half, UK life insurance $(0.2)bn in the first quarter, South Africa $(0.1)bn in the second half, and Germany custody $0.1bn during the year.

Assessment: A modelling gap closed. Roughly $1bn of revenue and $0.7bn of cost leaving implies a pre-tax drag near $0.3bn on the 2026 base, small enough to be absorbed inside a $36.6bn number and large enough that its absence would have been arguable in both directions all year. The reallocation commitment was raised from $1.5bn to $1.8bn on the back of the Hang Seng cost synergies, which is the same capital being counted once as a synergy and once as investment capacity, and readers should not add the two.

9. Tangible Book Compounds 12% and the Share Count Stops Falling

Tangible net asset value per ordinary share rose to $9.64 from $8.61, a 12.0% increase, and $0.42 of that came in the fourth quarter alone. The quarterly walk is instructive: $0.31 from attributable profit, $(0.10) from the dividend, $0.03 from currency, $0.05 from cash flow hedge reserves, $0.11 from fair value movements through other comprehensive income and $0.02 from other items. Roughly two fifths of the quarter's book accretion came from reserve movements rather than from earnings net of the dividend. Share buybacks contributed nothing, and the period-end count fell only 43m shares.

Assessment: This is where the twelve-month return actually comes from, and it now works differently than it did in October. With no buyback running, retained earnings accrete to book rather than to share count, so tangible book per share compounds faster and earnings per share compounds slower. Both halves of that trade are already visible: book per share up 12.0% for the year, attributable profit excluding material notable items up 9.1% against earnings per share on the same basis up 15.3%. The January Hang Seng completion takes $6.7bn, or roughly $0.39 per ordinary share, straight out of that book.

10. The Deposit Franchise Is the Thing Being Sold

Customer accounts rose $131.9bn as reported and $67.6bn on a constant currency basis to $1,786,828M, with growth in all four businesses and $50bn in the fourth quarter alone. Loans grew $17.6bn on a constant currency basis, almost entirely in the UK, so the loan-to-deposit ratio fell again to 55.3% from 56.2%. Hong Kong time deposits held at 36% of Hong Kong customer deposits, down from 39% a year ago.

"We have seen deposit growth, as you've seen in our new disclosures and wealth, across the spectrum of our customer base, Premier, Private Bank, Retail, in every market, in every jurisdiction, even when there isn't a home market. And that really underpins the growth that we are seeing in our banking NII. We have taken a very conservative trajectory on loan growth." — Pam Kaur, Group Chief Financial Officer

Assessment: The falling time-deposit share is the underrated number here. A three percentage point shift out of time deposits into current and savings accounts lowers the cost of the deposit base and is exactly what the finance chief means by low betas in Hong Kong. Combined with a $593bn hedge rolling onto higher yields for three more years, it is why a bank guiding for two to three rate cuts can also guide banking net interest income up. It is also why loan growth matters less to this thesis than the deposit mix does.

11. Technology Spend Becomes an Explicit Line in the Cost Argument

Management put technology at approximately 20% of the cost base and framed the 2026 investment envelope around three generative-AI workstreams: colleague enablement, with approximately 85% of staff on the productivity suite and 31,000 engineers on an AI coding assistant; end-to-end process re-engineering across more than 50 processes; and customer-facing tools, with more than 1,000 relationship managers on a wealth AI platform. Application retirement is the funding mechanism: 1,165 non-strategic applications were demised in 2025, approximately 36% of the 2025 to 2028 programme, out of roughly 3,000 flagged from a base near 9,000.

Assessment: The disclosure is better than most European banks give and the funding logic is coherent: retire applications, release run-the-bank cost, redeploy into new capability. The claim that deserves scepticism is that all of this fits inside 1% cost growth while headcount is being added in technology and a $0.6bn Hang Seng technology-led restructuring runs alongside. It is achievable on the numbers management has disclosed. It leaves no margin for a savings programme that slips.

Guidance & Outlook

One raise, one new multi-year target, five reiterations, one withdrawal and one conversion of a timetable into a condition. The withdrawal is worth pausing on: the double-digit wealth fee growth commitment has been removed, not missed, and management's stated reason is that the wealth ambition now sits inside the group revenue target. That is defensible framing, and it also happens to remove a standing commitment on the line whose flow metric is simultaneously being redefined.

Implied path to the 2026 numbers. On banking net interest income, the $45bn floor requires an average of $11,250M a quarter against $11,722M just printed and $11.6bn clean. Management named the offsets: a $300M first-quarter day-count headwind and further rate cuts. On costs, approximately 1% growth on $33,464M implies roughly $33,799M, and the fourth quarter's $8,905M annualises to $35.6bn, so the guide requires the quarterly run rate to fall back from the seasonal peak as it did in 2025, when the quarterly average was $8,366M. On credit, approximately 40 basis points against 39 delivered implies a modestly higher charge on a modestly larger book. On returns, a 17% floor against 17.2% delivered and 15.9% in the fourth quarter implies the first quarter has to do most of the year's work, as it did in 2025 when it printed 18.4%.

Where the market sits. The full-year print beat the company-compiled consensus by 3.6% on pre-tax profit and 1.4% on revenue, and the 2026 targets are above where the mid-teens framing had the Street modelling returns. The specific line where consensus looks most likely to be low is banking net interest income, where the guide sits roughly $1.4bn below the clean fourth-quarter run rate annualised, and management said so on the call rather than leaving it to be inferred.

Guidance style. Unchanged and now with a full year of evidence behind it. Every 2025 line came in at or better than guided: banking net interest income $1.1bn above the floor, returns 20 basis points above the raised target, credit one basis point inside, costs exactly on. The corollary is that the 2026 and 2028 targets should be read the same way, as floors set where the bank is confident, which argues for modelling above them and against paying for the difference in advance.

Analyst Q&A Highlights

Whether 17% Is a Floor That Rises or a Level That Holds

The dominant line of questioning on the call, raised by two participants in succession and never fully answered. The first framing asked whether the target is all-weather beyond the three-year horizon; the second asked whether the return improves year on year within it, and noted that the long-term incentive plan's lower performance boundary had moved up to 16.5% from 14% while the 18% upper boundary was unchanged. Management declined to give a trajectory in either direction.

Q: "The first one, just coming back on the upgraded ROTE targets, the 17% plus, I just want to check, in terms of how you're thinking about that on a year-on-year-on-year basis for 2027/28, are you thinking that ROTE continues to improve, or are you thinking more 17% is a very acceptable and a good level, and so any additional upside you would look to reinvest? Within that, I note that, on the LTIP, you've brought up the lower end of the boundary for performance, I think to 16.5% from 14%, but the 18% at the top end hasn't changed."
— Amit Goel, Mediobanca

A: "So, firstly, yes, we have ambitious targets and the target is '17% plus' each year. We are not giving a trajectory, whether it's the same or progressive, but, of course, we continue to grow our business and invest in it diligently, but the target is just '17% plus' each year."
— Pam Kaur, Group Chief Financial Officer

Assessment: The incentive plan is the more informative disclosure. Moving the threshold from 14% to 16.5% while leaving the maximum at 18% means the remuneration committee has compressed the payout band around the new target, which tells you management is being paid to hold 17% rather than to beat it. That is consistent with a bank that thinks 17% is defensible and does not think 18% is routine, and it caps what an investor should extrapolate.

Whether the Fourth-Quarter Run Rate Makes the 2026 Guide Too Low

The most valuable exchange for a 2026 model, and the one where management gave the most away. The question tried to establish a clean jumping-off point by asking what else besides the flagged $100M should be stripped out of the fourth quarter, and what deposit growth assumption sits behind the guide. The answer named the clean number, ruled out any other one-offs, and then conceded what that number annualises to.

Q: "Then I just wanted to ask around banking NII, please. Clearly, that Q4 jumping off point is flattered by $100 million. I don't know if there's anything else that you'd direct us to, to strip out of that number in terms of deposit catch up or the impact of HIBOR. I was particularly interested in what your deposit growth assumption is that sits behind the guidance you give in 2026, please, because I think that could be a sensitivity around the ultimate outturn of banking NII in 2026."
— Aman Rakkar, Barclays

A: "On banking NII, you're right; there was a $100 million non-repeat items. If you take that out for modelling purposes, you come to $11.6 billion. There are no other one-offs. There was a higher HIBOR quarter-on-quarter, and HIBOR also stabilised... We also saw low betas on savings accounts in Hong Kong. Very importantly, we saw strong deposit growth."
— Pam Kaur, Group Chief Financial Officer

Assessment: Management answered, precisely, and in the shareholder's favour. "There are no other one-offs" is an unusually definite statement about a quarter that includes a HIBOR recovery, and it makes $11.6bn a defensible base. The residual uncertainty is not the starting point but the rate path, and management put the sensitivity on a slide rather than in prose: a 100 basis point down-shock costs $3,389M in year one.

Whether the Buyback Can Return Before the Three Quarters Are Up

A direct attempt to pull the resumption forward, built on the observation that capital generation could put the ratio back inside the range during the first quarter, paired with a second question about whether inorganic uses now beat repurchases at the current share price. Management held the line on process and answered the second question with a principle rather than a preference.

Q: "There looks like a decent chance that you'll be within your target CET1 range in Q1, based on historical and perhaps projected capital generation estimates. Obviously, it raises the prospect as to whether you might be able to reintroduce a buyback earlier than planned. I don't know if you're able to comment on whether that's a realistic or plausible scenario. I guess, more specifically, just interested in the capital allocation thought process from here. Clearly, your stock's trading at a level now where the return on investment around buyback might be beaten by alternative uses of capital."
— Aman Rakkar, Barclays

A: "As you know, we look at our share buyback decisions on a quarterly basis, and that will be a quarterly process. The starting point is clearly our target operating range, which we are working hard on, so that we can replenish the capital that has been deployed in the Hang Seng Bank privatisation... Our distribution priorities, hence, have not changed, and share buyback remains for us a useful tool to deploy surplus capital, irrespective of where the share price is."
— Pam Kaur, Group Chief Financial Officer

Assessment: The questioner made the sharper point and did not get it addressed. At 1.95x tangible book a repurchase is a materially worse use of capital than it was at 1.5x, and the answer that buybacks are useful "irrespective of where the share price is" is a statement that valuation does not enter the decision. The Group Chief Executive's separate framing was more disciplined: four acquisition criteria that are a high bar, and a preference for returning capital when any one of them fails. Investors should hold management to the second formulation.

What Changes in 2028 That Cannot Happen in 2026 or 2027

The most probing challenge to the new revenue target, and a recurring one: two participants separately observed that if fee income keeps growing at anything like its 2025 rate, the 5% target would be cleared well before 2028. The question asked what management can see arriving in 2028 that the market cannot. The answer removed 2028 as a specific event.

Q: "I'm not quite sure what is so specific about 2028 that means you can get 5% revenue growth there, but I assume you don't feel you can before then. I know that at the moment in 2026 we've got some headwinds from interest rates, but you've also got very strong tailwinds around things like wealth management... And 2027, I guess, should be a reasonably normal year. So I'm just trying to think, is there something that I need to think about that you can see that is happening from 2028 and beyond that will make your revenue growth better that we're not seeing today?"
— Ed Firth, KBW

A: "I would look at 2028 not specifically as the year 2028, but as what we believe our long-term structural opportunity to grow. It's our guidance for 2028, but we've delivered 5% revenue growth in 2025. We've delivered 4% in 2024. So it's just the footprint we're in, and the capabilities for us to capture the growth is there."
— Georges Elhedery, Group Chief Executive

Assessment: A candid answer that reframes the target as a structural statement rather than a forecast of acceleration. It also concedes the questioner's premise: nothing specific happens in 2028. The target says the bank can grow at 5% through a cycle, which is useful for a terminal-value assumption and close to worthless as a 2026 estimate. The finance chief's follow-on, that the bank considers downside scenarios and does not take the best case, is the more honest description of how all three targets were set.

Hong Kong Property Coverage Ratios, and the Question That Went Unanswered

The sharpest challenge of the morning, and the only one to attack the disclosure directly rather than the outlook. The questioner computed stage 3 coverage from the bank's own table at 14% for Hong Kong commercial real estate and 18% for mainland China, said plainly that the area remained uncomfortable, and asked where management intends to take those ratios. The reply was a detailed and genuinely useful walk through the Hong Kong property market by sub-sector, ending in a different metric entirely.

Q: "And then, CRE, this is still an area where I'm a little bit uncomfortable. Stage 3 CRE China, 18% coverage; 14% on Hong Kong stage 3 – can you talk a little bit where you want to drive that to? And clearly, I heard Pam's remarks about provisions, and CRE was mentioned."
— Kian Abouhossein, JP Morgan

A: "The one metric which I personally follow, though my job has changed now, is what happens to credit impaired names, and the exposure that you need to focus on is credit impaired names which have an LTV over 70%. Now, this number has grown, and it now stands at $1.9 billion but the ECLs against it have also grown, to around 900 million and if you go back quarter on quarter, that differential between the ECL number and the exposure sits around a $1 billion number. That's where you think your risk is."
— Pam Kaur, Group Chief Financial Officer

Assessment: The answer is substantive and the market colour inside it is the best given all morning, but it does not respond to the question asked. Coverage was not addressed, no target was offered, and the metric management prefers is a narrower slice than the one the questioner computed. There is also a live ambiguity in the numbers: the release's table shows a $0.9bn credit-impaired allowance against a $6.3bn credit-impaired balance, which is where the 14% came from, while this answer describes approximately $0.9bn as sitting against the $1.9bn of exposure above 70% loan-to-value. Those two descriptions cannot both be complete, and neither was reconciled.

Why the Hang Seng Benefit Is Split Into Two Buckets

A structural question about the disclosure rather than the deal, and one that produced a useful answer about what management is willing to book versus what it is willing to promise. The line of enquiry asked why $0.5bn is labelled a synergy and $0.4bn an upside, whether the difference is certainty, and whether the $0.6bn restructuring charge covers both. A second participant returned to the same split later and asked what the restructuring money is actually being spent on given management had previously said there would be little severance.

Q: "Then secondly, on the Hang Seng synergies on slide 13, do you mind just talking a little bit around why you've adopted two buckets that you've labelled 'synergies' and 'upside'? Is the upside bucket basically where there's a lower degree of certainty over the synergies? Some colour on what type of synergies fall into each bucket would be useful. Presumably, there's no incremental restructuring costs associated with the upside bucket on top of the $0.6 billion."
— Multiple analysts incl. Benjamin Toms, RBC; Amit Goel, Mediobanca

A: "Classically, what you would see in terms of cost synergies and all the restructuring is actually severance costs. That is not the case here. Because there is so much on the revenue as well, a lot of the restructuring will be in terms of investment from a technology perspective... think of it as a $900 million benefit in those two buckets, with different degrees of accounting rules and different probability of expectations, and then an overall restructuring cost of $600 million to achieve that total."
— Pam Kaur, Group Chief Financial Officer

Assessment: The split is a credit to the disclosure rather than a criticism of it: management separated what accounting standards let it recognise from what it hopes to achieve, instead of blending the two into a single headline. The consequence for a model is that only $0.5bn belongs in a base case, against $0.6bn of certain restructuring cost, so the recognised synergies barely cover the charge and the entire net benefit by 2028 rests on the $0.4bn of ambition plus the $0.8bn minority pickup.

The Slowdown in Hong Kong Customer Acquisition

The only question of the morning to test the wealth franchise's flow engine, and it came from arithmetic rather than from the disclosure. The questioner noted that 1.1 million new-to-bank customers for the year against a roughly 300,000 quarterly run rate implied a materially slower fourth quarter, and asked whether that was seasonal or structural. Management confirmed it is structural, deliberate and continuing.

Q: "Then just on the wealth side for a second question, looks like there's a bit of a slowdown in the new account opening in fourth quarter if you've got 1.1 million customers for the year and you were running at about 300,000 a quarter. Is that just seasonal, or anything's changing there that we should be aware of?"
— Alastair Warr, Autonomous

A: "We are trying to be now a little bit more selective on the acquisition, because we have now had a three-year trend on how the acquisition has happened in between the lower end of the customer base and the more Premier. We have added a fee for the new to bank customers who have a balance less than HKD10,000, and because of that, we expect that there would be slightly slower acquisition in 2026."
— Pam Kaur, Group Chief Financial Officer

Assessment: An honest answer and a defensible strategy, and it retires a metric that had been presented as evidence of the wealth flywheel one quarter earlier. Taken with the negative Asian net new invested assets in the same quarter and the redefinition of the flow measure, three separate wealth-flow indicators moved the wrong way or changed shape in a single results event, and only this one was put to management.

What They're NOT Saying

1. Why the flow metric is being replaced now. The bank is retiring net new invested assets in the quarter that measure printed $7bn with Asia at negative $3bn, and replacing it with one that shows $26bn and positive $19bn. The stated rationale, peer comparability and a fuller view of the customer relationship, is reasonable on its own terms. The timing was not addressed, no analyst asked, and the release does not restate the new measure for the quarters before 2024.
2. What the $80bn of net new invested assets actually was. A footnote discloses that the 2025 figure on the new definition would be $46bn of net new invested assets plus $40bn of net new deposits. The $34bn gap is asset-management third-party distribution, which is 42.5% of the number the bank has quoted all year as its wealth flow.
3. Whether Hong Kong commercial real estate coverage is where management wants it. The question was asked directly, using the bank's own table, and produced a market commentary and a different metric rather than an answer. Coverage of the credit-impaired balance is not presented in the disclosure and was not discussed.
4. Which allowance sits against which exposure. The release shows $0.9bn of credit-impaired allowance against a $6.3bn credit-impaired balance; the call describes approximately $0.9bn as sitting against the $1.9bn of that balance above 70% loan-to-value. Both figures are stated as fact by the same company on the same morning and they cannot both be complete.
5. The CIB cost line, for a second consecutive results event. The segment holds the largest share of group tangible equity, runs a 56.3% cost efficiency ratio, and earns the lowest return of the four businesses at 16.2%, down from 16.8% at nine months. Management did not volunteer an explanation and no analyst asked. Against a newly adopted 17% group floor, this is the one segment that does not clear it.
6. What share of the insurance contract service margin leaves with HSBC Life Singapore. The Singapore insurance manufacturing business remains under strategic review. The $14.6bn contract service margin balance is the mechanism the entire wealth revenue argument rests on, and no disaggregation of it by market has been given.
7. Whether the Madoff matter is settled. Legal provisions in the fourth quarter were immaterial and the release outlook makes no reference to the outstanding Luxembourg proceedings. The $1.1bn provision taken in the third quarter was described then as a best judgment on which the eventual impact could be significantly different, with a second appeal pending. Silence for one quarter is not resolution, and no question was asked.
8. What the buyback will be sized at when it returns. The timetable has become a condition, the condition is a capital ratio, and management explicitly declined to link the decision to the share price. A shareholder therefore knows when the constraint clears and has no basis at all for estimating the amount.

Market Reaction

• Pre-print setup: the ADR closed at $87.88 on 24 February, up 11.7% year to date against 0.7% for the S&P 500, up 52.1% over twelve months and up 3.3% over the trailing thirty days. The 52-week closing range entering the print was $46.68 to $90.72, so the stock sat roughly 3% below its high. The company put 2025 total shareholder return at more than 57%, including a more than 49% share price gain.
• Reaction day, 25 February: the ADR opened at $92.60, a 5.4% gap up, traded a $92.36 to $94.78 range, and closed at $94.20, up 7.2% or $6.32. Volume was 3.1M against a 2.3M thirty-day average. The S&P 500 rose 0.8% on the same session.
• Primary listing: the London line opened at 1,343.4p and closed at 1,394.0p, up 7.95%, on 58.5M shares against roughly 15M to 20M on the preceding sessions. It closed at the high of the day.
• Hong Kong: the release landed at 07.00 GMT, which is 15.00 Hong Kong time, so the Hong Kong session had almost closed before the numbers were public and is not a read on the print.

This was a re-rate, not a relief rally, and the pattern of the session says so. The London line opened up 4% and closed at its high on triple normal volume, which is what accumulation looks like rather than a gap that fades. The setup was not depressed going in: the stock entered the print near a 52-week high after a 52% twelve-month run, so there was no discount to close. What the market bought was duration. A bank that had guided mid-teens returns for 2026 and 2027 now guides 17% or better for 2026, 2027 and 2028, and it attached a revenue growth commitment and a reaffirmed 50% payout to the same three years.

The valuation consequence is the whole story for a new buyer. At $94.20 the ADR trades at 1.95x the $48.20 of tangible net asset value per ADS reported that morning, against 1.52x at the October print. Put differently, the implied cost of equity on the delivered 17.2% return has compressed from 10.8% to 8.8% in four months. The ADR is up 34.5% from the third-quarter reaction close while tangible book per ADS has risen 4.6%. Essentially all of the return since we initiated has come from the multiple.

Street Perspective

Debate: Is "17% or Better" a Step-Change or a Restatement of the Run Rate?

Bull view: The target moved from mid-teens to 17% and gained a year, which is a two to three point uplift at the floor across a three-year window that includes an assumed rate-cutting cycle. A bank willing to commit to that level in 2028, with a hedge that reinvests at 3.7% in that year, is signalling that the deposit franchise now produces a structurally higher return than the market has been paying for.

Bear view: The bank delivered 17.2% and set the floor at 17%. The fourth quarter annualised at 15.9%, the incentive plan's payout band has been compressed around the same number, and management refused three times to say whether the return improves inside the window. That is a target that codifies the present rather than promising the future.

Our take: The bulls have the better of the substance and the bears have the better of the price. Committing to 17% through a cutting cycle is genuine information, and this management team has beaten every number it has guided. But the market moved 7.2% in a session and 34.5% in four months on that information, which is more than the target's incremental content is worth. The right response to a credible floor is to model above it, not to capitalise the difference in advance.

Debate: Has the Wealth Flow Story Been Proven or Repackaged?

Bull view: Wealth fee income grew 24% for the year with all four lines contributing, the contract service margin balance grew 21% to $14.6bn on 35% new business growth, and net new money of $26bn in the quarter on the measure that will govern from 2026 shows the franchise gathering assets and deposits together. The old measure understated the relationship by ignoring wealth customers' deposits, which is exactly what a bank with a 55% loan-to-deposit ratio should be counting.

Bear view: The measure changed in the quarter the old one deteriorated. Net new invested assets were $7bn with Asia negative, market moves supplied 52% of the year's invested-asset growth, the wealth-customer flow figure is $46bn against an $80bn headline, the double-digit fee growth guidance has been withdrawn, and Hong Kong customer acquisition is being deliberately slowed. The revenue is real; the flywheel is less proven than it looked in October.

Our take: Closer to the bears, and this is the change that moves our thesis. The new measure is genuinely better and the disclosure is genuinely complete, so this is not a governance criticism. It is an analytical one: the specific pillar we underwrote, that the wealth build is flow-driven rather than market-driven, is weaker than the reported numbers suggested, and management itself said 2025's market tailwind cannot be assumed for 2026. The insurance contract service margin remains the strongest part of the case, because it is contractual.

Debate: Does 1.95x Tangible Book Already Capitalise the New Targets?

Bull view: A bank compounding tangible book at 12%, earning 17.2% on it, guiding to 17% or better for three years with a 50% payout and a buyback that returns when the capital hole from a value-accretive acquisition is refilled, is worth close to two times book. The implied cost of equity of 8.8% is not obviously wrong for a franchise with 25% deposit share in Hong Kong, a $593bn hedge and two home markets.

Bear view: Two times book prices a bank for delivering its own best case with no allowance for the office vacancy rate in Hong Kong at 17%, an unresolved Luxembourg legal proceeding, a segment holding the largest block of tangible equity that earns below the new group floor, and a 2026 in which the share count stops shrinking and $6.7bn of tangible book leaves in January. The stock has done 34.5% in four months on 4.6% of book growth.

Our take: With the bears, which is why the rating moves. The bull argument is not wrong about the franchise; it is a statement about what the business is worth, and the market has now paid it. Our estimate of exit-2026 tangible book of $50.10 per ADS puts the stock at 1.88x forward book today, so even holding the multiple flat for a year gives a total return of roughly 5% including a 4.4% dividend. That is a Hold, and it would take a de-rate rather than a stumble to make it a mistake.

Model Update & Valuation Framework

We initiated in October with a full set of estimates for 2025, so the first table grades our own work before setting new numbers. We were conservative on every operating line, which is a bias worth naming rather than explaining away.

New estimates. The single largest change is that the fourth-quarter migration in Hong Kong commercial real estate did not continue, which removes the reason we carried 2026 credit above the company guide.

Valuation. We continue to value HSBC on tangible book and a normalised return rather than on earnings, because the reported earnings line has been distorted by notable items in every quarter of our coverage and carries a further $0.8bn of scheduled distortion in 2026. At the 25 February close of $94.20 the ADR trades at 1.954x the $48.20 of tangible net asset value per ADS reported that morning. Against the delivered 17.2% return excluding notable items that implies a cost of equity of 8.8%; against the new 17% target it implies 8.7%. The equivalent figures at the October close were 10.8% and 9.9%.

Our twelve-month framework runs off an estimated exit-2026 tangible book of $50.10 per ADS. That build is deliberately conservative in one respect and generous in another: it assumes no contribution at all from currency or other comprehensive income, which supplied roughly a third of the fourth quarter's book accretion, and it charges the full $6.7bn Hang Seng tangible book impact in January.

The dividend contribution of 4.4% is our estimated 2026 payout of $4.11 per ADS measured against the $94.20 close, and it is the same in all three scenarios because the 50% payout applies to earnings excluding material notable items and is insulated from the disposal and restructuring charges. On a trailing basis the declared 2025 dividend of $3.75 per ADS gives 3.98%.

The skew is what decides the rating, and it has inverted since October. Then we underwrote roughly three points of downside against thirty-two of upside with a low-teens base case. Now it is roughly eleven points of downside against nineteen of upside with a base case of five. The business is better than it was four months ago on every measure the bank reports. The distribution of outcomes for a buyer at this price is worse.

Thesis Scorecard Post-Earnings

The pillars below are the ones established at initiation in October and carried in our thesis of record. They are graded here against what this results event disclosed, not replaced.

Overall: The operating thesis is stronger than it was in October on five of six original pillars. Both credit and capital-allocation bear points weakened, the two engine pillars were confirmed with raised guidance behind them, and the one bull pillar that weakened did so on the flow mechanism rather than on the revenue. What changed is the price, which is why a better business gets a lower rating.

Action: Hold. The bull trigger we published in October, a guidance raise at the February results, fired precisely as specified, and the market settled it in a single 7.2% session that took the stock 34.5% above our initiation print. At 1.95x tangible book the base case is a 5.5% twelve-month total return against roughly 10% of downside. Existing holders should stay; new money has better entry points ahead of it.

Bottom Line

HSBC did in February exactly what we said in October it needed to do, and did it more completely than we expected. The return target went from mid-teens to 17% or better and gained a year. The Hang Seng transaction, which we criticised as an unfalsifiable assertion, arrived with an explicit buyback-threshold arithmetic and completed a quarter early. The disposal drag we said would never be quantified was quantified. Banking net interest income beat its own raised guide by $1.1bn, costs landed exactly on target, credit came in a basis point inside, and the Hong Kong property book stopped migrating. Of the eleven commitments we said in October we would grade at these results, eight were met or beaten, one was not addressed, and two are not yet due.

Two things pull the other way. The wealth flow story is weaker than the revenue line implies, and the bank chose this results day to retire the metric that showed it. And the market did all the work in one morning: at $94.20 the ADR carries a 1.95x multiple on tangible book and an 8.8% implied cost of equity, against 1.52x and 10.8% four months ago. We initiated on a distribution of roughly minus three to plus thirty-two. That distribution is now roughly minus eleven to plus nineteen with a base case in the mid single digits, and a mid-single-digit base case against the S&P 500 is the definition of a Hold. The three things that would move us back up are specific: a buyback restarting at scale rather than as a token, a 2027 revenue path that clears 5% early, and a wealth flow number on the new definition that holds above $20bn a quarter with Asia positive. The one that would move us down is a Hong Kong property charge that shows the collateral, not the migration, was the problem all along.