Below the Poll's Floor on Earnings, Above Its Ceiling on Buyback: Upgrading ING to Outperform

Results vs. Consensus

ING is one of the few European banks that publishes the poll the market marks it against. The figures below are the company-compiled consensus of 15 contributing brokers, collected between 9 and 15 October, two weeks before the print. It is a median, so the lines do not sum. Its chief virtue for this quarter is that it carries minimum and maximum as well as median, and the quarter's character is visible only in the range.

3Q2026 Scorecard

Revenue

Total income of €6,227m rose 5.6% year on year and fell 0.9% sequentially, missing the poll median by €79m but landing comfortably inside the range. The composition is the story and it is favourable. Commercial net interest income, the line that represents the bank's actual business of lending and taking deposits, rose 13.4% year on year to €4,335m and 3.9% sequentially, driven by a €135m increase in liability income as the replicating portfolio continued to roll into higher yields. Fee income of €1,273m rose 9.3% year on year and held within €5m of the record set in the second quarter. Together these two lines, which are 90% of total income, missed the poll by €14m combined.

The entire income shortfall sits in the third line. All other income, which is total income less commercial net interest income and fees, fell to €619m from €832m in the second quarter and €910m a year ago. It printed below the poll's €624m minimum. Within it, hedge ineffectiveness swung from positive €91m to negative €72m, a €163m move that accounts for the majority of the sequential decline on its own. Treasury income fell a further €44m, which the bank attributes to short-term funding and liquidity management including the effect of higher rates on non-remunerated volumes under the minimum reserve requirement. Financial Markets gave back €51m after a strong second quarter. The Corporate Line offset part of this with a €110m increase in income from financial stakes, including an €89m annual dividend from the Bank of Beijing holding and a €57m gain on the partial sale of the TMBThanachart stake.

Assessment: This is the third consecutive quarter in which the volatile income line has moved the headline by more than a hundred million euros, and the second of the three in which it moved against the bank. That is precisely the pattern our standing thesis flagged. What has changed is the ratio: a year ago all other income was 15.4% of total income, this quarter it is 9.9%, because the commercial lines have grown into it. The line that the bank does not control is becoming a smaller share of the income it reports, which is the right direction even when the quarter itself is bad.

Costs and Provisions

Operating expenses of €3,119m rose 3.6% year on year and 1.1% sequentially, missing the poll median by €42m. Of that miss, €39m is simply more restructuring than the poll assumed: incidental items came in at €69m against a €30m median. Strip out regulatory costs and incidental items and underlying expenses were €2,975m, up 3.6% year on year and broadly flat sequentially, against a poll median of €2,957m. That is a €18m difference on a €3bn cost base, and the sequential performance absorbed a 3.5% collective-labour-agreement salary increase in the Netherlands that took effect on 1 July.

Loan loss provisions of €403m, or 21bp of average customer lending, are the larger half of the pre-tax shortfall. They printed €3m above the poll's maximum estimate. Inside the number sits a €111m management overlay taken in anticipation of a model methodology update in the Wholesale Banking project finance portfolio, worth about 6bp, and roughly 1.5bp of charges related to risk transfers and credit insurance. Excluding both, risk costs ran near 13.5bp, well under the bank's stated through-the-cycle average of about 20bp.

Assessment: Our July checklist asked whether third-quarter underlying cost growth would stay near the second quarter's 4.2%. It did not. It decelerated to 3.6%, with year-to-date growth running at 3% against a full-year guide the bank reiterated. The cost pillar that we marked at risk three months ago performed. The provision line is the one that deteriorated, and its single largest component is an accounting change the bank has gone out of its way to characterise as unrelated to asset quality.

Earnings

Net result of €1,850m rose 3.5% year on year and fell 5.0% sequentially, €10m below the floor of the poll. Earnings per share of €0.65 rose 8.3% year on year, a faster rate than the net result because the average share count fell 3.7% over the year. Return on tangible equity was 16.3% in the quarter against 15.4% a year ago, and 14.8% on a four-quarter rolling basis against 12.9%. The effective tax rate of 29.2% sits inside the guided 29–31% band, and was the one below-the-line item that helped.

Two different earnings growth rates were quoted on the call and they belong to different bases. Quarterly earnings per share rose 8.3%. Four-quarter rolling earnings per share, at €2.35, rose 17%, and the four-quarter rolling net result of €6,764m rose 11%. The rolling figures are the ones management used to describe the trajectory, and they are the better measure of it, but they are not the quarter.

Assessment: A 6.2% miss on net result against a quarter in which the two commercial income lines grew 13.4% and 9.3% is a mix problem, not an earnings problem. The question for the rating is whether the items that caused it recur, and on the evidence of the bank's own disclosure two of the three do not and the third is self-funding.

Quality of the Miss

ING publishes a reconciliation of reported results to results excluding volatile and incidental items. It is the single most useful table in the deck this quarter.

Applying the same treatment to the comparatives changes the direction of the quarter. The third quarter of 2025 carried a positive €4m of volatile items and the second quarter of 2026 a positive €42m. On that basis the gross result went from €2,882m to €3,229m year on year, growth of 12.0% against the 7.7% reported, and from €3,156m to €3,229m sequentially, growth of 2.3% against the 2.8% decline reported. The quarter accelerated underneath, in both comparisons, and decelerated on the surface in both.

Note what the reconciliation does not adjust. The €111m provision overlay is left in. The bank has not classified it as a volatile item even though it describes the cause as a methodology change rather than credit deterioration. On a basis that also excluded it, pre-tax profit would have been €2,937m against a €2,892m poll median, which is the arithmetic the market will do for itself.

Assessment: Over the last five quarters volatile items have contributed between negative €121m and positive €42m to the gross result, a swing of €163m on a line that averages around €3.1bn. The hedge ineffectiveness component alone has gone minus 81, plus 91, minus 72 in three consecutive quarters. Management's position is that this is accounting rather than economics and reverses over time, and the second quarter is direct evidence for that claim. The investor's position should be that a line with that variance deserves to be modelled at zero and watched for drift, which is what the full-year guide of €2.5–2.7bn for all other income effectively does.

Year-Over-Year Comparison

Quarter-Over-Quarter Comparison

Nine-Month View

Nine months in, every full-year guide is within reach on simple arithmetic. Year-to-date return on tangible equity of 15.6% sits above the "more than 15%" guide with a weaker fourth quarter already signalled. Year-to-date cost growth of 1.9% reported, and 3% on the underlying basis management cited, is tracking the full-year range. The nine-month picture is of a bank beating its own plan and spending part of the beat on things that pay back later.

Segment Performance

The group number conceals a wide divergence this quarter. Retail Banking grew pre-tax profit 17.9% year on year. Wholesale Banking shrank it 16.5%. Both outcomes are driven by the same two forces running in opposite directions: the replication tailwind, which lands almost entirely in Retail, and the Treasury and Financial Markets weakness plus the project finance overlay, which land almost entirely in Wholesale.

Retail Banking

Retail total income of €4,244m rose 9.4% year on year, with commercial net interest income up 14.4% to €3,248m and fees up 9.7% to €859m. Operating expenses rose 2.9%, producing 650 basis points of positive jaws and a cost/income ratio of 48.0% against 51.1% a year ago. Risk costs of €192m were flat in absolute terms year on year and 14bp of average lending. Return on equity at a 13% CET1 allocation reached 24.3% against 22.5%. Pre-tax profit of €2,013m is 74% of group pre-tax profit, against 67% a year ago.

Net core lending grew €8.5bn, of which €6.9bn was mortgage production concentrated in the Netherlands, Germany, Australia, Italy and Spain. Deposits grew €2.7bn despite summer holiday spending, the conclusion of several gathering campaigns, and continued migration of customer money into investment products. Assets under management and e-brokerage rose €6.7bn in the quarter to €328bn, up 22% year on year. The division completed its first risk-transfer transaction, on a Dutch business banking exposure, releasing €1.0bn of risk-weighted assets.

"Commercial net interest income rose 14% year-on-year and 4.2% sequentially, supported by sustained growth in customer balances and a higher average liability margin, which further strengthened during the quarter." — ING 3Q2026 results release, Retail Banking

Assessment: This is the engine working exactly as the thesis described. A 24.3% return on allocated equity in a division that is now three quarters of group profit, with the margin tailwind still building rather than fading, is the reason the group return on tangible equity has moved from 14.0% to 15.6% year to date. The sequential income decline of 2.9% is entirely the other-income line, which fell from €293m to €115m. Commercial net interest income rose 4.2% sequentially, while fee income slipped 2.4%.

Retail Netherlands

Total income of €1,402m rose 13.9% year on year. Commercial net interest income of €1,111m rose 20.0%, the fastest of any unit, which is what a home market with the largest replicating deposit book looks like when the portfolio rolls. Pre-tax profit of €881m rose 25.7%. The cost/income ratio of 37.0% improved from 41.7%, and return on allocated equity reached 35.9%. Risk costs were effectively zero at 0bp, with €2m of provisions on a €140bn mortgage book whose Stage 2 ratio improved to 13.4% from 14.0%. Net core lending grew €3.7bn; deposits were flat at €0.0bn as business banking inflows offset the holiday-season decline.

Assessment: The Dutch business is generating a mid-thirties return on allocated capital with essentially no credit cost. The flat deposit quarter is the thing to watch rather than the thing to worry about: it reflects seasonality and the deliberate migration of customer balances into investment products, where ING earns a fee rather than a spread. That migration is thesis-positive even though it costs deposits.

Retail Germany

Total income of €794m rose 5.3% year on year but fell 0.9% sequentially. Commercial net interest income of €625m rose 6.8%, the slowest of the four retail units. Fee income of €126m fell 6.7% year on year and 16.0% sequentially, which the bank attributes to reduced customer trading activity. Pre-tax profit of €392m rose 10.4%. Deposits declined €2.1bn, against a €7.8bn increase in the second quarter, including a €2.0bn shift from deposits into investment products. Excluding that shift, the bank says deposits were broadly stable, as the end of a promotional savings campaign and holiday spending were largely offset by new inflows.

Assessment: Germany is the unit to watch for the deposit-pricing trade-off. Almost all of the €2.1bn outflow after a €7.8bn inflow is customer money moving into investment products, and excluding that shift the end of a promotional campaign was largely offset by new inflows, so the cost of protecting the liability margin is not yet visible in volumes. Germany is also the only retail unit with fee income falling on both comparisons, which makes it the natural test of whether protecting the spread today costs the fee tomorrow. That exact question was put to management on the call and did not get an answer.

Retail Belgium

Total income of €687m was flat year on year at plus 0.3% and fell 6.9% sequentially. The story is cost: operating expenses of €373m fell 12.4% year on year, taking the cost/income ratio to 54.3% from 62.1% and lifting pre-tax profit 12.1% to €251m despite flat income. Return on allocated equity improved to 16.0% from 14.0%. Risk costs rose to 26bp from 14bp, with €63m of provisions, the highest of any retail unit relative to its lending book, and the Belgian business lending Stage 3 ratio rose to 3.9% from 3.6%. Net core lending fell €0.8bn, which the bank attributes to lower balances for one large client.

Assessment: Belgium remains the restructuring story rather than the growth story, and on that basis it is working. The credit trend is the one genuinely uncomfortable data point in Retail: business lending there carries the highest Stage 3 ratio of the group's business-lending books and the fastest Stage 3 deterioration in its retail book.

Retail Other

Total income of €1,361m rose 12.5% year on year, second only to the Netherlands' 13.9% among the four units, with fee income of €235m up 29.8%. Pre-tax profit of €490m rose 14.8%. This is also where the cost growth sits: expenses excluding regulatory costs of €743m rose 15.2% year on year, an increase of €98m which is 95% of the group's entire €103m underlying cost increase. Net core lending grew €2.7bn and deposits €4.2bn.

Assessment: The whole of the group's underlying cost acceleration is a deliberate build-out of business banking, private banking, the investment platform, payments and insurance distribution in Spain, Italy, Romania, Poland, Australia and Türkiye. Income in those markets is growing 12.5% against expenses growing 15.2%, so the jaws are currently negative by about 270 basis points. Management declined on the call to give a return hurdle for any individual investment, saying only that business cases exist. That is the single largest unquantified item in the cost base and it will stay that way until the Capital Markets Day.

Wholesale Banking

Wholesale total income of €1,802m fell 0.4% year on year and 4.1% sequentially, the only division to shrink on both comparisons. The underlying businesses did not. Lending income of €917m rose 3.7% year on year and 6.6% sequentially on syndication fees. Daily Banking and Trade Finance income of €522m rose 16.3%. Fee income of €419m rose 9.4% year on year and 4.5% sequentially. Deposits grew €9.4bn on Payments and Cash Management mandates.

Treasury and Other contributed nothing at all this quarter, against €86m a year ago and €97m in the second quarter. Financial Markets gave back €55m sequentially on credit and rate trading. Risk costs of €208m, or 39bp, rose from 27bp in the second quarter and 28bp a year ago, with the €111m project finance overlay inside them. Pre-tax profit of €646m fell 16.5%. Risk-weighted assets of €140.8bn fell 5.7% year on year and 2.2% sequentially, including €2.0bn of relief from a project finance risk transfer, so return on allocated equity held at 12.5% despite the profit decline.

Assessment: Two of the four Wholesale product lines grew and two collapsed, and the two that collapsed are the two the bank does not steer. Income over average risk-weighted assets improved to 506bp from 486bp a year ago, which says the division is earning more per unit of capital consumed even as its reported profit falls. That is the capital-velocity thesis working and the mark-to-market thesis hurting, in the same division, in the same quarter. The 39bp risk-cost print is the headline number that will be quoted; roughly 21bp of it is the overlay.

Corporate Line

Corporate Line income of €181m fell 13.0% year on year but rose from €36m in the second quarter. Financial stakes contributed €169m against €59m, carrying the €89m Bank of Beijing annual dividend and the €57m gain on the partial TMBThanachart sale. Foreign currency hedging contributed €96m against €77m. Pre-tax profit was €46m.

Assessment: The Corporate Line added €70m of other income sequentially, and a €145m swing in its total income, and still could not offset a €293m combined fall in Retail and Wholesale other income. Two of its three contributors this quarter are non-recurring by construction: an annual dividend and a disposal gain.

Key Performance Indicators

One note on the assets-under-management figure, because 22% growth on a deposit-funded balance sheet is the kind of number that deserves decomposition. Of the increase, roughly €20bn is net customer inflow, about 8% of the opening base. Roughly €11bn is the consolidation of the Polish fund management business acquired earlier in the year, about 4 points of the 22%. The remainder is market effect. About four fifths of the growth is organic, inflows plus markets, and the inorganic fifth was disclosed without being asked.

Key Topics & Management Commentary

Overall Management Tone: Management spent the prepared remarks on structure rather than on the quarter, devoting most of the deck to the three strategic levers, a wholesale reorganisation and a Capital Markets Day date, and reaching the third-quarter numbers late. The posture was the most forward-leaning of the past four quarters and noticeably willing to absorb near-term pain, to the point that the chief executive accepted an analyst's framing that this year's estimates need to come down. Conviction weakened in exactly one place, on the questions that asked for a number behind a judgement, where specifics were declined three separate times.

1. The Miss Is Three Items, and the Bank Volunteered the Bridge

The pre-tax shortfall against the poll was €187m. Three identifiable items account for more than all of it: €111m of provision overlay, €72m of negative hedge ineffectiveness, and €39m of restructuring beyond what the poll assumed. The notable thing is not the arithmetic, which anyone could do, but that management led with it and characterised each item as temporary before being asked.

"Total risk costs were EUR 403 million in the third quarter included EUR 111 million related to an IFRS model update. Excluding this model update and SRT related charges, Risk costs were well below our through-the-cycle average of 20 basis points, reflecting our prudent risk management and the quality and strength of our loan book." — Ida Lerner, CFO

On the income side the framing was the same, and it has a precedent. In the first quarter hedge ineffectiveness cost €81m and management said the majority would come back over time. In the second quarter it delivered positive €91m. This quarter it cost €72m and the same sentence was used again.

"All other income was affected by interest rate volatility, resulting in negative results from hedge ineffectiveness. However, this effect is expected to reverse over time." — Ida Lerner, CFO

Assessment: The claim that hedge ineffectiveness reverses is not a management assertion the market has to take on trust. It was made two quarters ago and then paid off in full. That is a materially stronger position than the same words would carry from a bank making them for the first time, and it is the main reason we treat a below-the-floor print as a non-event for the medium-term earnings power.

2. The Liability Margin Reached 113bp, Above Its Own Long-Term Range

The liability margin stepped 6bp to 113bp, after moves of 5bp and 3bp in the first two quarters. The sequence is now 99, 99, 104, 107, 113 across five quarters. That takes it above the 100–110bp band the bank calls its expected longer-term range, and the full-year guide moved with it, from the upper-middle of 100–110bp at the half-year to around 110bp now.

Liability net interest income has grown 21.1% in four quarters while deposit balances grew 5.9%. That gap is the replication effect: the portfolio rolls off at old, low fixed rates and reinvests at current ones, independently of what the policy rate does next. Lending net interest income grew 7.4% over the same period on 9.1% balance growth, so lending is a volume story and deposits are a margin story.

The second disclosure behind the margin is less flattering and the bank publishes it anyway. Blended deposit cost rose to about 92bp from about 86bp, while the pass-through ratio fell to about 40% from about 42%. On savings and term deposits alone, cost rose to about 115bp with pass-through at about 50%, down from about 53%. The margin widened partly because reference rates rose faster than the bank passed them on.

Assessment: Our first bull pillar is that this tailwind arrives largely regardless of policy rates, and it is delivering ahead of schedule. The guide upgrade to around 110bp for the full year is the second consecutive quarter of upward revision on this line. The arithmetic of that guide is awkward, though, and an analyst did it on air: a full-year average of about 110bp against 108bp for nine months implies roughly 115bp in the fourth quarter, another 2bp step on top of this quarter's 6bp. Management confirmed the numbers and did not dispute the implication.

3. The 2027–28 Replication Upgrade Rests on a Curve That Has Already Moved

The deck carries a slide titled "Short to medium term liability margin outlook positively impacted by changes in forward curve". It compares two vintages of the three-month Euribor forward curve, June 2026 and September 2026, and shows what each implies for gross replicating income on retail eurozone deposits.

Three months of curve movement added €1.7bn of projected 2027 replicating income and €2.6bn of 2028. For scale, group total income in 2027 is guided above €26bn. Then, in the Q&A, the CFO was asked why the commercial net interest income guidance had not moved given the better curve, and included this in her answer:

"But there, you also need to be mindful that we have seen quite a significant volatility in the interest rate outlook over the past months and quarters. And just comparing the outlook today compared to what we saw end of September, you already dare see that it's coming down quite a bit from the heightened levels in September." — Ida Lerner, CFO

Assessment: The slide and the answer are in tension and the bank let both stand. The replication chart is dated to 30 September and the curve has fallen since, so the €1.7bn and €2.6bn figures were stale on the morning they were published. No sensitivity was offered, and the bank separately declines to discuss pass-through, which is the second input to the same calculation. An investor cannot rebuild this number. The honest conclusion is that the long-dated portion of the replication tailwind is a directional claim rather than a quantified one, and it should not be capitalised into a valuation at the rates the slide implies. The near-dated portion, the part already locked into the 113bp print and the around-110bp full-year guide, is a different matter and is real.

4. Commercial Net Interest Income: Range Unchanged on Paper, Upper End on the Call

The full-year commercial net interest income guide of €16.8–17.0bn was not changed. It was also quietly removed from the consolidated outlook slide, which now carries mobile primary customers, fee income, total income, operating expenses, CET1 and return on tangible equity, but not commercial net interest income. The number survives verbally in the prepared remarks and on the net interest income slide. Then, under questioning, the CFO narrowed it.

"The NII range is still valid. But however, on the basis of what we know today, we are more looking at the upper end of that range of EUR 16.8 billion to EUR 17 billion for 2026." — Ida Lerner, CFO

Nine months delivered €12,568m. The upper end of the range implies €4,432m in the fourth quarter, growth of 2.2% sequentially against the 3.9% just printed. Consensus already sits at €17,020m for the year, marginally above the top of the guide.

Assessment: A narrowing to the upper end delivered in a Q&A answer rather than on the outlook slide is a raise that does not look like one, which is a pattern this management has used before. The more interesting signal is the removal of the line from the outlook table. Commercial net interest income is the single largest driver of the thesis and the bank has moved it from the summary page to the body. That may be housekeeping before a Capital Markets Day that will reset the whole target set. It may also be a bank that does not want a hard number on its most rate-sensitive line going into a year when it has just told the market the curve is falling.

5. Fee Income Held the Record, With the Mix Rotating Underneath

Fee income of €1,273m rose 9.3% year on year and was within €5m of the second quarter's record. The full-year expectation of about €5.0bn, pulled forward a year at the half-year, was confirmed, and implies about €1,214m in the fourth quarter.

Category figures are the bank's own and do not sum exactly to the total because of rounding. The sequential flatness conceals a real rotation: investment product fees fell €38m on reduced customer trading activity, particularly in Germany, and against an unusually strong Belgian second quarter, while lending fees rose €44m on wholesale syndication. Fee growth decelerated from 13.9% year on year in the second quarter to 9.3%, though part of that is the base effect of the Polish fund management consolidation annualising.

Assessment: Our second bull pillar holds. What is now visible is that the fee line has its own volatility: investment product fees track customer trading activity, which tracks markets. A quarter in which market-linked fees fell 10% sequentially and the total still held within €5m of a record is evidence of genuine diversification across five categories, four of which grew year on year by between 9% and 18%.

6. Underlying Costs Decelerated to 3.6%, and the Full-Year Guide Is Now Comfortable

Underlying operating expenses, excluding regulatory costs and incidental items, were €2,975m, up 3.6% year on year against 4.2% in the second quarter, and broadly flat sequentially against the €2,961m then. The sequential performance absorbed a 3.5% Dutch collective-labour-agreement salary increase effective 1 July, offset by lower internal and external headcount and lower marketing spend. Headcount fell 1.5% year on year.

The CFO decomposed the €103m year-on-year increase: €98m in the Retail Other countries and €36m in Wholesale Banking, partly offset by lower costs in the Netherlands, Belgium and Germany. In other words every euro of the group's underlying cost growth, and then some, is investment spend in new markets and in wholesale transaction banking, while the three mature franchises shrank in aggregate.

On the full-year arithmetic: the guide of €12.6–12.8bn excludes incidental items booked after the first quarter. Nine months on that basis is €9,309m, leaving €3,291–3,491m for the fourth quarter against €3,050m this quarter on the same basis. Most of that step is the fourth-quarter regulatory charge, which includes the annual Dutch bank tax and ran at €361m in the comparable quarter last year.

Assessment: Three months ago we moved the operating-leverage pillar to at risk because underlying cost growth had accelerated from 1.1% to 4.2% and no analyst had challenged it. This quarter it decelerated, year-to-date growth is running at 3%, the guide is reiterated and the step into the fourth quarter is mostly a known tax. The pillar goes back on track. The residual concern is not the rate of cost growth but its concentration: €98m of €103m landing in markets whose aggregate jaws are currently negative, with no disclosed return hurdle.

7. The €111m Project Finance Overlay Is a Model Change, Not a Credit Event

Stage 1 and 2 risk costs of €165m were almost entirely one item: a €111m management overlay taken in anticipation of an IFRS model update in the Wholesale Banking project finance portfolio. The release is unusually explicit about causation.

"This model update reflects revised model methodology and does not stem from a deterioration in the underlying credit quality of the portfolio." — ING 3Q2026 results release

The supporting evidence is consistent with that claim. Stage 2 outstandings fell €2.2bn in the quarter and the Stage 2 ratio improved to 7.8% from 8.2%, driven by Wholesale Banking Lending where it fell to 7.7% from 8.7%. The Stage 3 ratio was flat at 1.6%. Stage 3 additions of €239m were described as mainly collective provisions in consumer lending reflecting portfolio growth and model refinement, with individual Stage 3 risk costs modest. Stage 3 coverage fell to 33.1% from a restated 33.5%, the restatement being the inclusion of purchased or originated credit-impaired assets in the ratio from this quarter, disclosed in a footnote.

Assessment: An improving Stage 2 ratio, a flat Stage 3 ratio and modest individual charges are not the signature of a deteriorating book. This looks like what the bank says it is. Two cautions remain. First, a management overlay booked in anticipation of a model update is a discretionary estimate about a future accounting change, and the actual update has not landed yet. Second, this is the second consecutive quarter in which the discretionary overlay has been the entire Stage 1 and 2 story, after the second quarter released one overlay and immediately opened another. The group's provisioning is now being steered by overlay decisions rather than by the macro scenario, and the bank does not size the stock of them.

8. Wholesale Risk Costs at 39bp, and the Running Cost of Capital Velocity

Wholesale Banking risk costs of €208m, or 39bp of average customer lending, rose from 27bp in the second quarter and 28bp a year ago. The €111m overlay is about 21bp of that 39bp. The remainder includes charges related to risk transfers and credit insurance, which the group quantifies at roughly 1.5bp at the group level.

This is the second quarter in which the bank has flagged risk-transfer and credit-insurance charges as a recurring item inside risk costs. It is the accounting cost of the capital-velocity strategy: the transactions that release risk-weighted assets also carry a premium that lands in provisions.

Assessment: Our second bear point held that credit drift was moving between divisions rather than stopping, and that remains the right description. Retail rose to 14bp from an unusually low 10bp base and Wholesale deteriorated to 39bp. What has changed is the quality of the explanation: last quarter the Wholesale increase was attributed to Stage 3 files and a weaker economic outlook, which are genuine credit signals. This quarter it is a model overlay plus the running cost of a strategy the bank has chosen. Those are different things and the second is more comfortable, provided the group number stays near the mid-teens excluding overlays, which it did at roughly 13.5bp.

9. Risk Transfers Delivered 12bp and Met the Full-Year Target

Two risk-transfer transactions closed in the quarter, one in Wholesale Banking on a project finance portfolio releasing €2.0bn of risk-weighted assets, and the first ever in Retail Banking, on a Dutch business banking exposure, releasing €1.0bn. Together they delivered 12bp of CET1 relief, taking the year-to-date total to 16bp and the cumulative total including 2025 to 28bp.

That matters because at the half-year this programme was conspicuously behind. Only 4bp of the 15–20bp full-year target had been delivered by 30 June, and we flagged it as the weakest item on the checklist. One quarter delivered three times what the first half had.

Assessment: Target met, at the bottom of the band, with a quarter to spare. The strategic significance is larger than the 16bp: completing the first retail transaction proves the technique extends beyond wholesale corporate exposures to the retail book, which is where two thirds of the group's risk-weighted assets sit. That widens the runway materially. The bank deferred its 2027 risk-transfer target to the fourth-quarter release, and the Dutch mortgage risk-weight floor, which expires on 1 December 2026 and was worth roughly 15bp on previous management commentary, went unmentioned in the release, the deck and the entire Q&A.

10. The €1.9bn Distribution and the Capital It Did Not Cost

ING announced a €1.9bn distribution: a €1.65bn buyback starting the day of the print, plus €250m in cash payable on 14 January 2027. The poll's median expectation for the buyback was €1,250m with a €1,500m maximum. The April programme of €1.0bn completed on 23 October with almost 35 million shares repurchased.

The CET1 walk is the part that deserves attention.

Net profit added 50 basis points and the 100% reservation removed exactly 50. The entire 26 basis point improvement came from risk-weighted assets, which fell €4.3bn, and from other capital movements. CET1 capital rose €322m while risk-weighted assets fell €4.3bn. The announced distribution is funded from profits already reserved outside CET1 and therefore does not touch the reported ratio.

Assessment: Our first bear point said capital return depends on risk-weighted-asset relief rather than retained earnings, and this quarter is the cleanest illustration of it yet: the earnings contribution to the ratio was mathematically zero. The point stands, but its character has improved. Last quarter the relief came from a €2.8bn model update management would neither decompose nor forecast and warned could reverse. This quarter €3.0bn of the €4.3bn came from two named risk-transfer transactions with disclosed sizes, disclosed basis-point effects and a stated multi-year runway. A denominator-driven capital build is a weaker thing than an earnings-driven one. A denominator-driven build from repeatable, disclosed, contractual transactions is a considerably better thing than one from an opaque model change.

11. Refocusing Wholesale Banking, and a Date for the Strategy Reset

ING announced the restructuring of Wholesale Banking into a unit renamed Corporate and Institutional Banking. The coverage model moves from sector-focused to country-focused while retaining sector expertise, specialists consolidate into global product teams, and about 400 roles go globally. A pre-tax provision of about €175m lands in the fourth quarter, delivering about €75m of annual savings fully effective from 2028. The changes take effect on 1 March 2027.

"I previously said, look, we -- and we have said during the previous Capital Markets Day that the target for 2027 was 12% -- that is what we have said then. And in '27 in the Capital Markets Day, we will talk about ambition going forward. And I've previously also said during one of the conferences that I thought that where we currently are and even with the 12% target, that is not enough, and we will continue on improving the performance." — Steven van Rijswijk, CEO

That is a more pointed statement than it first reads. Wholesale Banking returned 12.5% on allocated equity this quarter and 11.9% year to date. The 2027 target set at the 2024 Capital Markets Day was 12%. The division has already reached the target it was given, eighteen months early, and the chief executive used the earnings call to say the target was never good enough.

Separately, ING dated its next Capital Markets Day for 14 June 2027, where it will publish a longer-term financial plan including targets for 2030. Management also framed the current 2027 outlook against the original 2024 targets: total income to exceed the original 2027 target by at least €1bn, operating expenses to land about €300m below it, and return on tangible equity to exceed it by two percentage points.

Assessment: Our July checklist asked whether a refreshed medium-term ambition would be announced and whether a Capital Markets Day would be dated. Both arrived. The gap we flagged, that the 2024 capital-allocation target had been met with no published destination beyond 2027, now has a closing date. On the restructuring itself, a €175m charge for €75m of annual savings is a 2.3-year payback, which is unremarkable, and the strategic logic of moving from sector coverage to country coverage in a business whose stated advantage is being a gateway to Europe is at least coherent. The thing to weigh is that this charge lands in a quarter the market will already be marking down.

Guidance & Outlook

Every line of the 2026 and 2027 outlook was confirmed. Nothing was raised and nothing was cut on the published slide, which is itself notable after two consecutive quarters of upgrades.

The footnote beneath the expense guide carries real content: the €115m of incidental cost items booked in the second and third quarters, and those to come in later quarters, are incremental to the guide rather than absorbed within it. The €175m fourth-quarter provision therefore sits outside the €12.6–12.8bn range, as does any further restructuring.

Implied fourth-quarter ramp: commercial net interest income of €4,432m at the upper end of the guide, up 2.2% sequentially. Fee income of about €1,214m, down 4.6% sequentially. All other income of €520–720m against €619m this quarter. Operating expenses of €3,291–3,491m on the guide basis, against €3,050m this quarter on the same basis, with most of the step being the annual Dutch bank tax. Reported expenses will be higher still by the €175m provision and whatever further restructuring follows. The liability margin at roughly 115bp.

Street at: consensus before the print carried €17,020m of 2026 commercial net interest income, which is marginally above the top of the guide, €24,777m of total income, €7,199m of net result and 15.1% return on tangible equity. For 2027 it carried €26,987m of income, nearly €1bn above the guide floor, and 17.0% return on tangible equity, a full point above it. Consensus is running ahead of guidance on both years, which is the normal condition for this stock and is why a guide reiteration lands as a non-event rather than as support.

Guidance style: this management guides to floors and then beats them, and it has raised something in each of the last two quarters. The decision to confirm rather than raise, in a quarter when the liability margin ran above the long-term range and risk transfers met their target early, is the first sign of restraint in a year. The most plausible reading is that the fourth-quarter restructuring charge and a falling forward curve absorbed the room that would otherwise have gone into an upgrade, and that the Capital Markets Day in June is now the designated venue for the next reset.

Analyst Q&A Highlights

Nine questioners, and unusually for this stock the hardest questions were about cost and capital rather than about margin. Two separate lines of questioning pressed the restructuring charge, one from the cadence angle and one from the payback ratio.

Whether the Full-Year Margin Guide Implies Another Fourth-Quarter Step

The call opened with the arithmetic of the margin guide done out loud. A full-year average of around 110 basis points against 108 for the nine months implies roughly 115 in the fourth quarter, a further step on top of the 6 basis points just delivered. The question then turned to whether deposit competition would allow it. Management confirmed both figures and declined to guide the pass-through assumption that would make the step possible, pointing instead to the structural reprice of the replicating portfolio and to the volatility of the rate curve.

Q: "Are you confident that this is under control at this point of time in your main markets?"
— Benoit Petrarque, Kepler Cheuvreux

A: "Looking ahead, we will continue to benefit from the structure of our replication portfolio with both the short and the long-term part being repriced upwards, but you also need to keep in mind that you can see the volatility in the curve of the interest rate."
— Ida Lerner, CFO

Assessment: Management did not resist the 115 basis point implication, which is as close to confirming it as a bank gets without printing it. Note which half of the answer carries the confidence. The replication reprice is described as structural and certain; the curve is described as volatile. That is the correct order of reliability and it is the same order we use in the valuation.

Why Commercial Net Interest Income Guidance Did Not Move With the Margin

The sharpest question of the call pointed at an inconsistency: if the liability margin guide went up and the forward curve is higher than it was in June, why is the income guide that depends on both unchanged? The answer narrowed the range without reprinting it, and then volunteered that the curve has already fallen back since quarter-end.

Q: "I'm just wondering, and the curve is also a little bit sort of higher compared to June. I'm just wondering why your guidance for commercial NII remains the same."
— Delphine Lee, J.P. Morgan

A: "The NII range is still valid. But however, on the basis of what we know today, we are more looking at the upper end of that range of EUR 16.8 billion to EUR 17 billion for 2026."
— Ida Lerner, CFO

Assessment: This is the most valuable sentence of the call and it was extracted rather than offered. A move to the upper end of a €200m range is worth about €100m of pre-tax profit, delivered as a Q&A aside on a day when the headline was a miss. The caution that followed matters just as much: the bank is telling the market that the curve supporting its own replication slide has already moved against it.

Whether the Fourth-Quarter Provision Is Front-Loading or a New Run-Rate

A recurring line of questioning treated the €175m provision as a surprise and asked whether 2026, at roughly €300m of restructuring in total, is the new normal or a pull-forward of actions that would otherwise fall in 2027. Management gave the 2025 comparison and the year-to-date figures, declined to guide the level, and reiterated the 2027 outlook unchanged.

Q: "So my first question is on restructuring costs. The EUR 175 million Q4 is quite a surprise. And looking forward, so we got used to having EUR 200 million roughly of restructuring costs per year over the past few years. This year will be higher, will be around EUR 300 million. How should we think about the restructuring costs going forward?"
— Giulia Miotto, Morgan Stanley

A: "If you look at the restructuring provisions for 2025, for example, they were at around EUR 300 million, and they expect to generate EUR 170 million in -- excluding wholesale banking. For now, we've taken EUR 130 million in provisions year-to-date that would lead to annualized savings of about EUR 100 million and we will continue to look for opportunities."
— Steven van Rijswijk, CEO

Assessment: The answer is better than it sounds. Restructuring at this bank is running at roughly €300m a year and generating roughly €170m of annual savings, a payback under two years, repeated annually. That is not a one-off charge pattern, it is a permanent efficiency programme funded out of the income statement every year, and it is one of the reasons the three mature franchises can shrink their costs while the group invests. The honest conclusion is that investors should model a recurring restructuring charge rather than treat each one as exceptional.

Whether the Restructuring Payback Ratio Is Competitive

A pointed follow-up compared the €175m charge against €75m of annual savings with programmes at other banks and judged the payoff low. Management did not engage with the comparison on its own terms, arguing that the cost of exiting these particular roles is not comparable to operational restructuring elsewhere.

Q: "My second question is on the restructuring cost again. So EUR 175 million cost to EUR 75 million run rate cost saves. It seems like a bit of a lower payoff versus some of the other types of restructuring costs we've seen from other banks."
— Shrey Srivastava, Citi

A: "But expected provision is a logical outcome of where we're active and the type of roles that we are changing. And therefore, we cannot simply compare it with operations or other banks that are different in scope."
— Steven van Rijswijk, CEO

Assessment: The answer is almost certainly correct on the economics, because 400 senior wholesale bankers across 36 countries carry higher severance per head than back-office consolidation, and it is also entirely unquantified. No split between severance and other costs was given, and no comparison basis was offered. This was the clearest instance of the pattern that ran through the whole Q&A: where a judgement could be defended with a number, the number was withheld.

How Much Risk-Transfer Capacity Remains

A questioner framed the risk-transfer programme as potentially approaching its limit, noting that some European peers have already optimised corporate risk-weight density heavily. The answer gave the fullest disclosure of the programme so far and explicitly positioned ING as behind, not ahead, of peers.

Q: "Could you please update us on how you're thinking about SRT's efficiencies over the next few years. Could you increase those given that you're starting from a low starting point?"
— Giulia Miotto, Morgan Stanley

A: "The transactions that we did this quarter contributed 12 basis points to the core equity Tier 1, bringing the capital relief for 2026 to 16 basis points. But if you also add what we did in 2025, we have now a total relief of 28 basis points. If you compare that to some of our peers, we're still significantly lower than that."
— Ida Lerner, CFO

Assessment: Being behind peers is the bullish answer here, and management knows it. Twenty-eight basis points of cumulative relief against peers who have done materially more means the runway is long rather than exhausted. The first retail transaction extends the technique to the two thirds of risk-weighted assets that sit outside wholesale. The missing piece is a 2027 target, deferred to the fourth-quarter release.

Whether the Headline Cut Is an Earnings Downgrade or a Timing Shift

The most analytically complete question of the call built the whole bridge and then asked management to confirm it: this year's estimates need to come down by roughly 4%, but each of the three headwinds either reverses or buys a permanent saving, so the net effect across the forecast period is positive. Management agreed without qualification.

Q: "So have I got that all right because on the incidentals and the Wholesale Bank division, you're basically buying in earnings at 2x PE, which is a bit of a headwind for this year, but obviously a fairly logical thing to do."
— Chris Hallam, Goldman Sachs

A: "You said a lot in the first question. But I think generally, you got it right. So in terms of that, we create a bit of short-term pain for long-term benefit."
— Steven van Rijswijk, CEO

Assessment: A chief executive agreeing on an earnings call that current-year estimates need to fall 4% is rare and is the single most useful thing said on this call. Our own arithmetic supports the figure: the third-quarter shortfall to the median plus the fourth-quarter provision after tax is about 3.4% of the full-year consensus, and management flagged further fourth-quarter restructuring beyond the €175m. The consequence for the rating is that the downgrade cycle is now a known, bounded and pre-announced event rather than a risk.

Whether the Lending Margin Has Found Its Level

The lending margin has now fallen in two consecutive quarters, from 126 to 124 to 123 basis points, and a questioner asked whether it is heading to 120. The answer attributed the decline entirely to mix rather than to price competition, and gave a specific guide for the remainder of the year.

Q: "And then my second question, the lending margin trickled down by 1 basis point this quarter. I was just wondering if 120 basis points is what we should expect going forward?"
— Namita Samtani, Barclays

A: "But for the remainder of the year, we expect lending margin to remain broadly stable at 123 basis points."
— Ida Lerner, CFO

Assessment: The mix explanation holds up against the disclosure. Growth skewed to mortgages and investment-grade wholesale lending, both of which carry lower margins and lower risk density, which is also why risk-weighted assets fell while the loan book grew. A one basis point margin give-up that buys a lower capital charge and a lower expected loss is a good trade, and the bank is steering on new-production return on equity rather than on margin. Two basis points of lending margin erosion over four quarters against fourteen of liability margin gain is not a competitive problem.

Whether Mortgage Demand Is About to Normalise

With lending growing at a 6% annualised pace against a 5% outlook, and mortgages driving most of it, a question asked whether higher rates would pull volume growth back toward target. The answer was the only genuinely cautious forward-looking statement management made all morning, and it was volunteered rather than conceded.

A: "We haven't seen that yet. So we still see, especially on the mortgage side, a significant increase in many markets. But we also -- we do also see the first signs, and we would expect the signs coming in of new house mortgages or mortgages for new houses to decline a little bit. So that could well be the case, but it's a bit too early to call."
— Steven van Rijswijk, CEO

Assessment: Worth noting because it is the second consecutive quarter in which management has pre-flagged a slowdown that has not arrived. In July the caution was about wholesale lending demand and it was not borne out, with the book growing €2.5bn this quarter. The pattern suggests a management that prefers to set a low bar on volume. Volume matters more than usual right now, because the commercial net interest income guide requires balance growth to carry it while the lending margin is flat.

What They're NOT Saying

1. The replication slide was not re-based to the current curve. The 2027 and 2028 replicating income figures rest on the 30 September forward curve, and the CFO said on the same call that the curve has come down "quite a bit" since. No sensitivity was published, so the single largest forward-looking disclosure in the deck cannot be adjusted by anyone outside the bank.
2. Pass-through expectations remain withheld. "We are not commenting on our expectation of pass-through." The bank discloses that every 10 basis points of pass-through on savings and term deposits is worth about €0.4bn of commercial net interest income, which makes this the second input, alongside the curve, that would let an investor rebuild the margin guide. Both are withheld.
3. The share of rising deposit costs attributable to campaigns was refused. Total deposit costs rose about 6 basis points to about 92 in the quarter, and the bank declined to say how much came from promotional campaigns versus repricing of the back book. That is the difference between a cost that stops when the campaign stops and one that does not.
4. Commercial net interest income left the outlook slide. The number still exists verbally and on the net interest income page, but the consolidated outlook table that carries fee income, total income, costs, CET1 and return on tangible equity no longer carries the largest driver of all of them.
5. No 2027 risk-transfer target, and no mention of the Dutch mortgage floor. The 2026 target was met at 16 basis points and the 2027 figure was deferred to the fourth-quarter release. The Dutch mortgage risk-weight floor expires on 1 December 2026, was previously described as worth roughly 15 basis points, and appears nowhere in the release, the deck or the Q&A.
6. No return hurdle for the Retail Other investment programme. Asked directly what return a given investment is underwritten to, management answered that business cases exist for each one. This programme is €98m of the group's €103m underlying year-on-year cost increase and currently runs negative jaws of roughly 270 basis points.
7. Sovereign portfolio duration and country mix were refused. "We are not displaying or announcing the duration or the country mix per se," on a €125bn portfolio, with the assurance that it is diversified, hedged for rate risk and exposed only to credit spreads.
8. The fee-versus-margin trade-off question was not answered. Asked whether a more selective approach to deposit campaigns risks fewer new customer relationships and therefore slower fee growth, the chief executive said he was not sure he had understood the question and answered a different one about headcount. Germany, where fee income fell on both comparisons and a promotional campaign ended in the quarter, is the live test case.
9. The €175m provision was not decomposed. No split between severance and other restructuring costs, and no basis for comparing the 2.3-year payback with peer programmes, despite being asked for exactly that.
10. Two live items from the July call simply vanished. The Singular Bank acquisition in Spain, due to close in the first quarter of 2027, and the global subscription model with 17 million customers migrated and fee benefits promised for later in 2026, appear nowhere in the release, the deck or the Q&A. Neither was asked about either.
11. The Russia exit remains a footnote. The outlook continues to exclude the potential impact of the intended exit, with no update on timing or on the roughly €0.8bn post-tax impact previously indicated.

Market Reaction

• Pre-print setup: the Amsterdam line closed at €29.445 on 28 October, up 22.6% year to date and 40.6% over twelve months, against a +13.1% year-to-date S&P 500. But momentum had already turned: the stock was down 7.2% over the trailing 30 days from €31.73, and entered the print 8.9% below its 52-week closing high of €32.31, which was set during the third quarter itself. The quarter ended at €31.49, so of the 10.7% decline from that level to today, roughly three fifths happened before the results were published.
• Reaction session (29 October, results before the open): opened at €28.66, a 2.7% gap down, traded a range of €28.22 to €29.075, and closed at €28.85, down 2.02% or €0.59. Volume of 8.0 million shares against a 6.6 million 30-day average, 1.2 times normal.
• Subsequent sessions: €28.255 on 30 October, down a further 2.06%, and €28.135 on 2 November. Cumulatively the shares are down 4.4% from the pre-print close across three sessions.
• Peer and index context on the reaction day: the AEX rose 0.59%. The two closest Benelux peers were effectively unchanged, at −0.25% and −0.30%. A large French peer rose 0.52%, a large Italian peer was flat, and a large German peer rose 3.80%. ING underperformed its index by 2.6 percentage points and its nearest comparables by about 1.75.
• New York line: the ADR fell 4.07% in premarket before recovering to close down 1.98% at $32.24, and stood at $31.38 on 2 November.

The move was entirely idiosyncratic. European banks were flat to higher on a session when ING fell 2%, so nothing in the sector tape explains it. What explains it is the shape of the miss against a poll that the market knew in advance: a print that lands below the minimum estimate on four income-statement lines is a different event from one that lands below the median, because it tells every holder that their own number was too high rather than that the average was.

The buyback failed to offset it, and the reason is instructive. A €1.65bn programme against a €1.25bn median is a €400m positive surprise on an €82bn company, worth about 0.5% of market value. The earnings shortfall and the pre-announced fourth-quarter provision together imply roughly 4% off this year's consensus, worth considerably more on any multiple. The market did the arithmetic correctly on the day: the capital surprise is real and smaller than the earnings surprise.

What the market appears not to have done is distinguish between an earnings cut and an earnings deferral. Three sessions of selling have taken price-to-tangible-book from 1.74x to 1.66x, and from 1.78x at the second-quarter print, a 7% de-rating over three months, on a quarter in which the commercial engine grew 13.4% and 9.3%, the capital target was met early, and the full-year guide was confirmed. That gap between what moved the price and what moved the business is the opportunity.

Street Perspective

Debate: Is the Miss an Earnings Downgrade or a Timing Shift?

Bull view: the bull case on the Street is that none of the three drivers survives into next year. Hedge ineffectiveness is accounting and reversed within one quarter the last time it happened. The provision overlay anticipates a methodology change explicitly unrelated to asset quality. The restructuring charges buy permanent savings at a payback under two years. Current-year estimates come down, out-year estimates go up.

Bear view: the bear camp contends that a bank which has missed on its volatile income line in two of three quarters is not experiencing noise, it is experiencing a structural mismatch it cannot hedge, and that a "one-off" overlay arriving in the same quarter as the second "one-off" overlay in a row is a pattern rather than an exception. On this reading the base rate of surprises is itself the problem, and a quarter that needs three adjustments to look good is a quarter that was not good.

Our take: the bull case has the better of it on evidence rather than on assertion, because the reversal claim was tested once already and paid. The bear case lands one genuine hit: the provisioning is now steered by discretionary overlays whose aggregate stock the bank will not size, which is a disclosure gap rather than a credit problem. We model the volatile income line at the middle of its guided range and the overlay as non-recurring, and we would change that view if a third consecutive quarter arrives with a new overlay replacing the last one.

Debate: Can the Liability Margin Hold Above Its Own Long-Term Range?

Bull view: the argument gaining traction is that the replicating portfolio makes the next two years nearly mechanical. Around 55% of the retail eurozone book has one to fifteen years of remaining maturity and is still rolling off vintages priced near zero, so the reprice continues whatever the central bank does. The bank itself now says the margin may stay above the 100 to 110 band through 2027 and 2028.

Bear view: the skeptics point out that the margin widened this quarter partly because pass-through fell to about 40% from about 42%, which is a competitive choice rather than a structural gain, and that Germany lost €2.1bn of deposits in the same quarter, although the bank attributes €2.0bn of that to customers moving money into investment products. On this view the bank is choosing margin over volume, that choice has a limit, and the published 2027 and 2028 uplift is anchored to a curve management has already said is falling.

Our take: both halves are right about different horizons. The near-dated reprice is contractual and we underwrite it. The 2027 and 2028 figures on the replication slide are a curve forecast dressed as a disclosure and we do not capitalise them. German deposits are the number to watch, not because of this quarter's €2.1bn, which was almost entirely a shift into investment products, but because Germany is where the price of the pass-through discipline would show first and the question of what it costs in fees went unanswered.

Debate: Is the Capital Build Repeatable Now That Risk Transfers Are Doing the Work?

Bull view: some desks argue the capital story just got considerably more credible. Three billion euros of risk-weighted asset relief from two named transactions with a disclosed 12 basis point effect is auditable in a way that last quarter's model update was not, the first retail transaction opens up two thirds of the balance sheet that was previously untouched, and management says it remains well behind peers on cumulative relief.

Bear view: the bear camp contends that this is the point: a bank whose CET1 ratio improves only when it sells risk is a bank whose organic capital generation does not cover its distribution ambition. Net profit added 50 basis points this quarter and the reservation removed exactly 50. Risk transfers also carry a premium that now sits permanently inside risk costs, so the relief is not free, and when the easy portfolios are done the ratio stops improving.

Our take: the bear framing is arithmetically correct and has been for three quarters, which is why this remains a live risk point in our thesis. But the conclusion drawn from it is too strong. A bank distributing 100% of profit by policy will always show zero net earnings contribution to its ratio; that is the design, not a defect. The real question is whether the denominator work is repeatable, and the evidence this quarter moved toward yes: named transactions, disclosed sizes, a new asset class proven, and a stated gap to peers. We would want the 2027 target at the fourth-quarter release before upgrading this from a risk to a pillar.

Debate: Is 1.66x Tangible Book the Right Price?

Bull view: the bull case is that the de-rating is mechanical and the earnings base is not. The shares have fallen 7% from the second-quarter print while consensus 2027 earnings per share rose 6% over the same period, so the forward multiple has compressed from both ends to roughly 9.3 times. A distribution yield near 7.7% on confirmed guidance and a buyback above the poll's ceiling is support that was not there three months ago.

Bear view: the bear camp notes the Amsterdam shares are still up about 34% from their late-October 2025 close, after 40.6% into the print, and trade at 1.66x tangible book against a European sector that mostly does not, that the entire re-rating since April rests on a margin tailwind the bank itself calls temporary, and that a guide which has been raised twice and then merely confirmed is a guide that has stopped giving.

Our take: the price is now below our fair value band rather than inside it, which is the whole of the rating change. We are not arguing that 1.66x tangible book is cheap in the abstract; it is a full multiple for a European bank. We are arguing that it is the wrong multiple for this bank's delivered return, that the fair-value arithmetic is unchanged from July while the price has fallen 7%, and that the distribution yield has returned to where it was at initiation. The bear point about the margin being temporary is correct and is already in our sustainable-return assumption of 16%, which is a full point below what the Street carries for 2027.

Model Framework & Valuation

The revisions below take the full-year guide at the upper end where management pointed to it, take the fourth-quarter restructuring at the announced size plus an allowance for the additional charges flagged on the call, and leave the sustainable through-cycle return unchanged. The last of those is deliberate: the operating case improved this quarter, but it improved in ways the Capital Markets Day in June is the proper venue to capitalise.

The Valuation Arithmetic

Tangible book value per share at 30 September is shareholders' equity of €49,980m less intangible assets of €1,783m over 2,843.7m shares, or €16.95. That is effectively unchanged from €16.96 at 30 June, because the €1,850m net result was offset by the €1,142m interim dividend paid in August and a €1,164m reduction in the cash flow hedge reserve. Reported shareholders' equity per share actually rose, from €17.57 to €17.58, because the share count fell faster than the equity did.

A residual-income frame values a bank at a justified price-to-tangible-book of return on tangible equity less growth, divided by cost of equity less growth. At a 10.5% cost of equity and 3.0% terminal growth, a 16% return justifies 1.73x and a 17% return justifies 1.87x. Against €16.95 of tangible book those are €29.38 and €31.64, which we round to a €29.40 to €31.65 band.

One thing changed in the inputs without changing the output, and it is a quality improvement. In July the band was bracketed by the 2027 guided floor of 16% and the 2028 consensus return of 17%. The pre-print poll now carries 17.0% for 2027 and 18.1% for 2028. The same fair value is now underwritten by a Street estimate one year nearer.

Tangible book per share is held flat throughout for comparability, so every scenario is conservative on a two-year view by roughly the retained share of a mid-teens return.

Yield support. At the 2 November close the shares carry a market value of about €80.0bn. The pre-print poll's full-year ordinary dividend of €1.28 is a 4.6% yield, and its €2.5bn of 2026 structural excess distribution adds a further 3.1%, for a total distribution yield near 7.7%. That compares with 6.4% at the second-quarter print and roughly 7.6% at initiation in April. The entire improvement is price and a larger distribution, not leverage.

On the New York line. At the euro-dollar rate implied by the two reaction-session closes, the €29.40 to €31.65 band is approximately $32.85 to $35.35, against the $31.38 close on 2 November.

Thesis Scorecard

The scorecard grades the standing thesis established at initiation on 1 May and carried through the second quarter, unchanged in its pillars, against what this quarter's print and call revealed.

Overall: the thesis is strengthened. Three of nine pillars changed tag, all favourably, and a fourth moved against us inside its tag. The operating-leverage pillar that slipped in July came back. The cost bear point that opened in July closed. The valuation bear point that drove the July downgrade has been substantially resolved by the market itself. The one pillar that moved against us, the volatile income line, moved in a way that is already inside the guidance and already inside our model. The capital bear point holds on arithmetic but its underlying mechanism became auditable this quarter in a way it was not last quarter.

Action: upgrade. In July we wrote that we would revisit the rating upward if the October distribution review converted the risk-transfer pipeline into a materially larger buyback, or if third-quarter underlying cost growth slowed from 4.2%. Both conditions were met, in the same quarter, and the shares fell 4.4% anyway on a miss built from a model overlay, an accounting swing and a restructuring charge that buys a two-year payback. The fair-value band is €29.40 to €31.65 against a €28.135 close, giving 4.4% to 12.5% of upside against 3.6% of downside to the 2026 guided floor, with a 7.7% distribution yield while waiting and a Capital Markets Day on 14 June 2027 as a dated catalyst. We are raising the rating to Outperform with conviction 6. We would move to a higher conviction on a fourth-quarter print that confirms the fourth consecutive step in the liability margin and publishes a 2027 risk-transfer target, and we would reconsider the upgrade if a third consecutive discretionary overlay appears or if the forward curve retreat visibly reaches the commercial net interest income guide.