The Price Reached Our Upgrade Level, the Credit Did Not: Maintaining Hold on BBVA

This report grades BBVA against the standing thesis set in our first-quarter 2026 initiation and revised at our second-quarter downgrade to Hold. Where a pillar's status changed this quarter, the change is stated in the Thesis Scorecard. The rating triggers published in July are graded verbatim in the Model Framework section.

Results vs. Consensus

Q3 2026 Scorecard

BBVA reports in euros under IFRS before the Madrid open. The third-quarter print landed on the morning of October 29 and the call followed the same morning. Two independent polls were struck on the euro lines: a 19-analyst poll dated October 14 with a full income-statement set, and a wire poll on profit, net interest income and provisions. They agree to within €3M on profit. We use the first as the benchmark and the second as the range check.

The table hides a mix that the profit line flattens. The 2.9% net interest income beat was worth about €230M of revenue. Gross income beat by only €148M, because fees missed by roughly 1% and trading income by roughly 8% on the line-by-line comparisons circulating after the print. Provisions then took €87M more than the wire poll expected. Profit beat by €8M. The ADR earnings line is the same quarter translated into dollars, and providers disagree on whether it beat or missed by a cent; we score it in line.

Year-Over-Year Comparison

The right-hand column is management's constant-currency presentation, which is how the bank frames every operating comparison on the call. The reported column is what hit the P&L. The gap is about five points at the net interest income line again this quarter. The polls above are struck on reported euros.

Constant-currency rates are shown only where management or the release quoted them for the standalone quarter. The bank publishes full constant-currency detail on the cumulative nine-month statement, and we have not manufactured the missing quarterly cells.

Quarter-Over-Quarter Comparison

This is the table that explains the share-price reaction. Net interest income rose €452M sequentially and gross income rose €407M. Then expenses took €217M, impairments €270M, and profit fell 1.5%.

Segment Performance

BBVA reports five geographic areas plus a Corporate Center, with a pro-forma Corporate & Investment Banking view laid across them. The 2025 comparatives were restated in the first quarter for a resegmentation into CIB; group figures were unaffected. The full area income statements are published only on a cumulative basis, so the first table is nine-month.

The gap between the areas at +15.4% and the group at +13.6% narrowed from four and a half points at the half to under two points, because the Corporate Center's third-quarter loss of €278M was smaller than the €333M of a year ago on hedging gains. Mexico's share of area profit rose to 45.0% from 44.1% at the half.

Quarterly Contribution and Area Ratios

Ratio columns in the filing's area blocks compare against December 31, 2025. The sequential cost-of-risk and spread moves below come from the presentation's three-point series and the filing's narrative.

Spain

Spain earned €934M in the quarter, down 5.5% year on year and 13.3% sequentially, and €3,106M over nine months, flat on the year. Management's preferred view excludes the redundancy charge and the value-added-tax items and gets to +2.3% for the nine months. The quarter was below what the Street had pencilled in for the area by about a tenth, and Spain was the most-cited disappointment in the reaction.

The line we were grading moved the right way. The customer spread widened 2bp to 2.87% after 3bp in the second quarter, because the yield on loans rose 14bp while the cost of deposits rose 13bp. Net interest income grew 1.2% sequentially and 2.3% year on year to €1,709M. Every other revenue line went backwards: trading income fell 42.1% year on year to €78M, fees fell 2.1% sequentially on lower CIB fees, and impairments rose 41.0% sequentially as the second quarter's portfolio-sale benefit fell away.

"Customer spread improved slightly in the quarter, benefiting from higher lending yields due to the repricing of our new loan production. Deposit costs also increased, mainly reflecting the immediate repricing of wholesale deposits..." — Gonzalo Rodríguez, Group CFO

Management's account of the deposit line was detailed. Ten of the 13bp increase came from non-retail balances, which reprice immediately with Euribor. In the €150bn retail and SME book the cost of deposits rose about 4bp, and about 2bp excluding a small segment that reprices automatically. About 80% of the €4bn year-on-year growth in demand deposits came from retail. Lending grew 7.3% year on year and only 0.2% in the quarter, held back by the seasonal unwind of public-sector advances, with larger companies up 3.9% and SMEs up 2.5% sequentially.

Assessment: Spain met the spread commitment and missed on everything around it. Two consecutive quarters of widening make the rate-floor argument unambiguous for this franchise, which is what we said would happen if the spread moved twice. The cost is that the ECB's second hike reached the corporate deposit book a quarter before it reaches the mortgage book, so year-on-year net interest income growth slowed to 2.3% in the quarter from 4.1% in the half even as the spread improved. Management's repricing schedule says the asset side catches up from the fourth quarter. That is credible arithmetic and it is not in the numbers yet.

Mexico

Mexico earned €1,517M in the quarter at constant FX, 0.7% below the second quarter and 8.7% above a year ago, and €4,522M over nine months, up 8.4%. Lending accelerated to 12.0% year on year and 3.0% sequentially, led by commercial lending up 4.2% in the quarter, with SME market share now at 34%. Net interest income rose 2.6% sequentially. Efficiency of 30.7% is still the best in the group.

"Regarding margins, we do think we have reached the bottom on the margins. It's very much related to the interest rate, obviously, of the Central Bank." — Onur Genç, CEO

The customer spread did not confirm that yet. It fell 7bp to 10.68%, a third consecutive decline, with loan yields down 8bp and deposit costs down only 2bp. Management attributes part of the compression to mix, because enterprise lending is growing fastest and carries a lower spread, and says the repricing of loans and deposits to Banxico's 6.50% is complete. Credit metrics moved the wrong way on the surface: the NPL ratio rose 9bp to 2.86% on retail inflows and coverage fell 7 points to 111%. The reported cost of risk improved 7bp to 3.19%, but the filing and the call both attribute that to the annual recalibration of model parameters. Underlying cost of risk was given as around 330bp, inside the guide of below 335bp.

Assessment: The volume half of the Mexican story is better than guided and management said loan growth will most likely beat the 10% guide. The spread half is still unproven for a third quarter, and this quarter's improvement in the cost-of-risk ratio was a model change, not a credit improvement. BBVA Research raised its 2026 Mexican growth forecast to 1.4% from 1.2%, which helps. We give the area a pass on volume and an incomplete on spread.

Türkiye

Türkiye earned €258M in the quarter, 4.3% below the second quarter, and €789M over nine months against a full-year guide of around €1bn with a downward bias. Net interest income rose 16.8% sequentially to €1,205M as the lira customer spread recovered from 0.38% to 1.08%, a 70bp widening, on normalizing funding conditions. Operating income rose 6.6%. Impairments rose 35.8% sequentially to €452M.

"As a result, cost of risk in Turkey could come slightly above our guidance of 220 basis points for the year, reflecting a more prolonged period of elevated rates than we initially anticipated. Nevertheless, we remain confident in delivering our net profit guidance in Turkey of around EUR 1 billion with a downward bias." — Gonzalo Rodríguez, Group CFO

The presentation slide still reads "FY26 CoR expected at around 220 bps." The arithmetic does not support it. Year-to-date cost of risk is 255bp, up from 236bp at the half. On simple arithmetic that ignores book growth, the third quarter alone ran near 290bp, and holding the year at 220bp would need a fourth quarter near 115bp. The macro assumptions behind the original guide have all moved against the bank: year-end inflation of about 30% against 25%, and a year-end policy rate of 35% to 36% against 32%. The central bank resumed easing on October 22 with a 100bp cut to 36%.

Assessment: This is the second consecutive quarter in which the Turkish credit guide has failed, and this time management did not move the published number; the CFO said it could be exceeded and the CEO said it will be. We treat that as the event we flagged in July. The offset is real: a 70bp spread recovery is the rate sensitivity working, and management held the profit guide. But a fourth-quarter profit of about €211M is what that guide now requires, below the €258M just earned, and the profit guide's downward bias is no longer described as slight. The variance in this group lives here.

South America

South America earned €265M in the quarter, 14.0% below the second quarter, and €821M over nine months, up 42.1% in reported euros. Colombia contributed €278M over nine months, more than double a year ago at constant FX, Peru €276M and Argentina €102M. Lending grew 3.3% in the quarter. The efficiency ratio held at 41.5%.

"For the full year, in South America, cost of risk could come slightly above our guidance of below 250 basis points, mainly depending on asset quality developments in Argentina. However, a stronger gross income performance should offset this impact in the region." — Gonzalo Rodríguez, Group CFO

Cost of risk rose to 2.72% from 2.69% at the half and 2.76% at the first quarter. Argentina's ratio eased 19bp to 7.52%, with lower NPL inflows, and the country NPL ratio fell 39bp to 6.0%. But Peru's rose 18bp as the second quarter's unusually low impairments normalized, and the hyperinflation charge reached €384M over nine months against €277M a year ago. To reach 250bp for the year, the fourth quarter would need roughly 184bp on simple arithmetic.

Assessment: We called this the most exposed commitment at the first quarter and gave it a pass at the second on the promise of second-half improvement. The second half has not delivered it, and management has now said the guide could be missed. The profit offset is genuine, because Colombia and Peru are compounding. But three quarters of deferral is a pattern, and we grade it as failed.

Rest of Business

Rest of Business earned €291M in the quarter at constant FX, up 7.0% sequentially and 55.7% year on year, and €801M over nine months. Loans grew 15.4% in the quarter and 68.5% year on year to €99.7bn, with the US CIB book up 77.5% to €39.6bn and Asia up 94.9%. Gross income rose 11.8% sequentially and efficiency improved to 42.0% from 49.0% at the end of 2025.

"When you look into those transactions, 95% of those project financings, they are backed by an investment-grade hyperscaler." — Onur Genç, CEO

Non-performing loans in the area rose to €427M from €392M at the half and €153M at the end of 2025. Coverage fell 4 points to 84%, which the filing explains as "the impairment of two wholesale customers," and nine-month impairments of €64M are "mainly explained by higher provisions linked to specific exposures in the United States." Cost of risk is still only 11bp. The Italian and German digital banks, which sit in this area, reported €10.6bn of deposits against €11.9bn given at the second quarter.

Assessment: The economics are still excellent and the disclosure on AI-related project finance is better than most peers give. The seasoning question we raised in July got its first answer, and it is a small but real one: two named-name impairments in a book that grew 15.4% in a quarter. An 11bp cost of risk on a book this size is not a problem. A book still growing at this pace while its first problem loans appear is the profile that deserves a lower multiple than its current return implies.

Corporate Center and CIB

The Corporate Center lost €278M in the quarter against €333M a year ago, helped by trading gains on hedges. Nine-month personnel expense was €748M against €626M. The presentation answers the question we left open in July: excluding the redundancies and the value-added-tax items, Corporate Center expenses grew 1.1% year on year. On the pro-forma CIB view, the wholesale franchise earned €1,044M in the quarter, up 7.7% sequentially at constant FX, and €3,082M over nine months, up 19.0%, with a return on capital of about 23% excluding Argentina and Türkiye.

Assessment: The holding-company cost line is a one-off, not a trend, and we close it. CIB now earns about a third of group profit and is growing faster than the group, which raises the importance of the Rest of Business credit question above its 8% weight in the area table.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Confident on revenue and more qualified on everything else than at any point this year. The guidance vocabulary moved from fixed numbers to "could come slightly above" on two credit lines, and the capital discussion was explanatory rather than announcing. Management was most specific where the news was good (deposit costs by segment, mortgage repricing lags, risk-transfer pipeline) and least specific on the number the market wanted, which was the next distribution.

1. Net Interest Income Accelerated, and the Beat Stopped There

Net interest income rose 5.7% sequentially at constant FX and beat the 19-analyst poll by 2.9%. Management's framing is that interest-rate headwinds are fading in Mexico and Spain, the two markets where net interest income correlates positively with rates, so volume growth now converts into earnings.

"This means that looking forward, volume growth should increasingly translate into earnings growth. While part of our lending growth has so far in the past few years has so far helped offset margin compression, we expect to capture more of this positive contribution to our P&L, the contribution of volume in the future." — Onur Genç, CEO

The composition matters. Türkiye supplied roughly 38% of the sequential increase on our arithmetic and Rest of Business another tenth. Spain and Mexico, which are 76% of area profit, supplied a smaller share than their weight.

Assessment: The core revenue line is doing what the thesis says it should do at the group level. The problem is that a 2.9% net interest income beat converted into a 0.3% profit beat, because fees, trading and provisions all missed. A bank can carry one weak line in a quarter. Three is a pattern worth watching for one more quarter.

2. Spain's Spread Widened Again; Spain's Profit Did Not Follow

Our July checklist asked for one thing from Spain: a second quarter of customer-spread widening. It got 2bp. It did not get revenue growth, because the cost of the ECB's September hike reached corporate deposits immediately and the mortgage book reprices on a lag.

"In our mortgage book, 2/3 of our mortgage book is repriced every 6 months and 1/3 every year. And you get the Euribor typically 12 months of 2 months before. So actually, it's 6 plus 2 and 12 plus 2 repricing. So the loan repricings have not come into effect yet from the latest increases that ECB had done." — Onur Genç, CEO

Management restated the sensitivity of about 4% of Spanish net interest income per 100bp parallel move, said the real exposure is to three- and six-month Euribor, and said the fourth quarter's spread will be better than the third. In Spain the bank is gaining share in consumer (+6bp) and commercial (+51bp) lending and losing it in mortgages (−31bp), with new-mortgage origination share at or below 10% in recent months.

Assessment: The commitment was met and the thesis mechanism is confirmed for Spain. We said in July that convergence of Spanish net interest income toward loan growth is a 2027 event; nothing this quarter brings it forward, and the gap widened to five points in the quarter. The deliberate retreat from mortgages is the right pricing decision in a rising-rate market. It also slows the volume side of the conversion.

3. The Credit Flow Did Not Fall Back

Gross NPL entries were €4,077M against €4,094M in the second quarter and €3,600M a year ago. As a share of credit risk, entries fell to 0.65% per quarter from 0.67%, because the book grew 3.3%. Write-offs rose 10.0% to €1,393M, the highest of the five quarters disclosed. Coverage fell 77bp to 84%. Within provisions, Stage 2 allowances rose 18.7% while Stage 1 allowances fell 19.0%, consistent with the model recalibration the bank disclosed for Mexico.

"But starting with the cost of risk on the bottom left, it remained broadly stable at 144 basis points over the first 9 months of the year, in line with our expectations. The quarterly evolution of absolute impairments, it shows an increase due to the impact of a large portfolio sale that we have done, if you remember, in the second quarter." — Onur Genç, CEO

Assessment: The test we set in July was that NPL entries must fall back from €4,094M. They did not; they held. Our harder test, two consecutive quarters of accelerating entries with falling coverage, did not fire either, because entries fell by €17M. The honest reading is that €4bn a quarter is the new run-rate of a book growing at 20%, not a one-quarter spike. The ratio is stable because the denominator is growing. The two areas whose guides are now overshooting tell us where the flow is concentrated.

4. Türkiye: The Spread Came Back and the Credit Guide Broke Again

The lira customer spread recovered 70bp in the quarter and net interest income rose 16.8% sequentially. That is the positive rate sensitivity management described in July. It came with impairments up 35.8% sequentially and a year-to-date cost of risk of 255bp against a guide of around 220bp that management now expects to exceed.

"Asset quality will come out a bit higher than the 220, as Gonzalo mentioned in the presentation, so it will come a bit higher. But that asset quality, higher asset quality figures, higher cost of risk will be compensated for other factors, especially NII." — Onur Genç, CEO

The guide was built on 25% year-end inflation and a 32% policy rate. The bank now expects about 30% and 35% to 36%. Management said it delivered roughly what the original guide implied despite missing both assumptions, and disclosed that it had provided sensitivities on exactly these variables.

Assessment: Management's disclosure here remains excellent and its macro calls have been wrong twice. A raised credit guide followed one quarter later by a statement that the raised guide will be exceeded is precisely the sequence our July trigger described. We grade it as fired, and the Model Framework section sets out what that means for the rating.

5. South America: The Commitment Failed in the Opposite Direction

Cost of risk rose to 272bp from 269bp, and the CFO said the year could come in slightly above the guide of below 250bp. Argentina improved at the margin. Peru normalized from a low second quarter. The hyperinflation charge rose.

"In the case of Argentina, impairments remained broadly stable quarter-on-quarter, though still elevated. Following our proactive adjustment of risk appetite, we are beginning to see signs of stabilization in asset quality metrics in Argentina and a gradual improvement expected throughout 2027." — Gonzalo Rodríguez, Group CFO

Assessment: The improvement timeline moved from "second half" in the first quarter to "throughout 2027" this quarter. Our July upgrade trigger asked for South American cost of risk to clear the guide. It moved away from it.

6. Capital: Earnings Generated 71bp, Growth and Markets Took 53bp, and Nothing New Was Announced

CET1 fell to 12.21% from a 12.41% pro-forma. Earnings added 71bp. Dividend accrual and AT1 coupons took 38bp. Risk-weighted asset growth took 44bp, net of risk transfers that contributed only 3bp in a slow August. Other items took 9bp, including a 22bp hit from rate curves and currencies.

"Actually, we are expecting around 12 to 15 -- 12 to 14 basis points SRT impact in the fourth quarter because deals are already in process." — Onur Genç, CEO

The 2027 SREP sets the CET1 requirement at 9.36% from January, up from 8.97%. Management said it will not raise its 11.5% to 12% target range, because the 264bp buffer between the requirement and 12% still exceeds the 258bp average of large European peers. The prudential-provisioning add-on leaves the requirement and becomes a year-end capital deduction that management called marginal. Risk transfers have contributed 21bp so far this year against a 30bp to 40bp guide that management describes as "towards the upper end"; adding the 12bp to 14bp expected in the fourth quarter gives 33bp to 35bp, which is the middle of the range.

Assessment: The quarter is weaker than it looks in one respect and stronger in another. Weaker: the commitment to distribute everything above 12% now has about €0.9bn to work with, and the market had expected roughly 13bp more. Stronger: most of the shortfall is a slow risk-transfer quarter and a market move, both of which can reverse, and management restated the €36bn four-year distribution figure without hedging it. We downgrade the capital pillar for one quarter and set a test for December.

7. Rest of Business: The Fastest Book in the Group Recorded Its First Named Impairments

The dominant growth question on the call concerned project finance tied to the artificial-intelligence build-out, after press reports named BBVA among the lenders to a troubled data-centre project. Management would not discuss the client. It said the exposure was a broad syndicate position, that net exposure must account for post-origination sales and hedges, and that 95% of its data-centre project finance is backed by an investment-grade hyperscaler.

"Our return on capital for the CIB business is now 23% if we exclude Argentina and Turkey, which is very high return. So you have to exclude them to have a more conservative figure, and that's 23%. In the rest of business, depending on where you count the revenues, it's between 20% to 24% return on capital." — Onur Genç, CEO

Assessment: The returns case is intact and the AI-specific question is reasonably answered. The filing's own language, "the impairment of two wholesale customers" and "specific exposures in the United States," is the first evidence of loss content in a book that has grown 68.5% in a year. It is small. It is also the first time this question has an answer that is not zero.

8. Mexico: Volume Beat, Spread Still Falling, Credit Helped by a Model

Loans grew 12.0% year on year with growth across every product, and management signalled it expects to beat its 10% loan-growth guide. The customer spread fell 7bp for a third consecutive decline, which management attributed partly to enterprise mix and said is now at its floor. Reported cost of risk of 319bp benefited from a model parameter update; underlying was about 330bp.

"So on lending growth, you see 12% growth. Because of the base effect, this 12% might modulate a little bit. But as you know, our guidance was around 10%. So most likely, we will beat our guidance, but very positive lending growth, very positive." — Onur Genç, CEO

Assessment: Mexico is now 45% of area profit, growing volume ahead of guide, and has not yet shown a single quarter of spread stabilization. Banxico is expected to hold 6.50% through 2027, which supports management's floor call. We need one quarter of a flat or wider spread to accept it.

9. The October Strategic Talks: Reaffirmation, Not Revision

The October 6 event covered artificial intelligence, technology and the Spanish franchise. It did not change the 2025-2028 financial targets and did not put a euro figure on artificial intelligence. The third-quarter deck marks all four group goals "aligned with plan": ROTE averaging c.22% (20.6% delivered so far), tangible book plus dividends compounding at mid-teens (14.2%), a c.35% cost-income ratio by 2028 (37.9%), and c.€48bn of cumulative profit (€19.6bn delivered).

"To be able to get to 22% on average, we have to do better in 2027 and 2028. In short, we expect the return on tangible equity to further improve in 2027 and in 2028." — Onur Genç, CEO

Assessment: We wrote in July that if the plan was only reaffirmed, "the multiple has nothing to work with until 2027." That is what happened. The new and useful disclosure is the forward commitment that ROTE rises in each of the next two years, because the arithmetic requires it: averaging 22% from 20.6% so far means roughly 23% in 2027 and 2028. That is now a gradeable line.

10. Rates Rose Again, and the Asset Side Has Not Caught Up

The ECB hiked again in September, taking the deposit rate to 2.50%, and the Federal Reserve raised to 4.00%. BBVA Research now expects further hikes to 2.75% and 4.25% by year end, holding through 2027. Twelve-month Euribor averaged 3.25% in September against 2.80% in June. Spanish inflation reached 4.9% in September. BBVA Research raised its 2026 Spanish growth forecast to 2.6% and its eurozone forecast to 0.9%.

"Interest rate curves, it seems like they're going to be going up further. So there will be even more improvement in the coming year in our view, with that change." — Onur Genç, CEO

Assessment: The rate environment is still better than the one we underwrote at initiation. It also cost the bank twice this quarter outside net interest income: the ALCO portfolio and primary-dealer holdings marked down as yields rose, hitting Spanish trading income and contributing to the 22bp market-related capital charge. The rate thesis pays through net interest income on a lag and costs through marks immediately.

11. The New CFO's First Call: Disclosure Continuity Held

This was Gonzalo Rodríguez's first results presentation as Group CFO. The disclosure set was unchanged: spreads by country, cost of deposits by segment, CET1 sensitivity to currencies, published macro assumptions and the risk-transfer pipeline. Two guides moved, and both moved the wrong way, which is the behaviour we credited to the outgoing CFO.

"I believe most of you have already had the opportunity to meet him at different events since he took on the role. But I would like to welcome Gonzalo to his first earnings webcast as the CFO of BBVA Group." — Onur Genç, CEO

Assessment: We said in July that disclosure quality is person-dependent until proven otherwise. One quarter is not proof, but it is the right first data point: bad news was published in the prepared remarks, not extracted in Q&A. We close this watch item.

Guidance & Outlook

Nothing was raised formally. Two credit guides are now expected to be missed, the group return guide was maintained with an upside bias, and the four-year targets were reaffirmed.

Implied fourth-quarter path: Türkiye's roughly €1bn profit guide against €789M delivered implies about €211M in the fourth quarter, below the €258M just earned, and "downward bias" allows less. Group ROTE of 22.0% to date against an around-21% guide implies a fourth quarter below the run-rate, though management now speaks of upside to the guide. The fourth quarter is seasonally heavy on costs, and the year-end capital deduction for prudential provisioning lands in it. The CEO briefly misstated the group guide on the call as "above 20%" before correcting it to "about 21."

Street at: the 19-analyst poll was at €3,007M for the quarter and the print came in at €3,015M. The poll had CET1 at 12.34%, so the capital line is where expectations will be cut, along with Spanish revenue and Türkiye's provisions.

Guidance style: unchanged in substance. Management still publishes the assumptions and moves the guide when they move. The change this quarter is form: on two credit lines management kept the published number and said it would be exceeded, instead of moving it. That is a smaller step than the Türkiye revision in July and it lands in the same place.

Analyst Q&A Highlights

Whether a Higher Supervisory Requirement Forces a Higher Capital Target

The 2027 supervisory decision raises the CET1 requirement by about 40bp. A question asked whether the 11.5% to 12% management range survives that, and whether return on tangible equity rises from 22% in the next two years. Management refused the first and committed to the second.

Q: "The bank's SREP requirement will go up 40 basis points from January. Given this movement, management still comfortable with running the business at 11.5% to 12% CET1 ratio range? Could we see the range move up to a circa 12% target ratio?"
— Benjamin Toms, RBC

A: "The difference between the requirement and the upper end of our target range, which is 12%, it's 264 basis points, 264 basis points. How does that compare the 264 basis points with our competitors? When you look into that large bank group that we also list in our presentations in the footnotes who they are, the large European banks, the average is 258 basis points."
— Onur Genç, CEO

Assessment: The answer is right on the arithmetic and it narrows the margin. The buffer above the peer average is now 6bp. Holding the range is the shareholder-friendly choice, and it means the distribution engine now runs with less slack than it did at the half.

Whether Corporate Deposits Are Squeezing the Spanish Margin

The first question asked how much of Spain's deposit growth was wholesale and whether a loan-to-deposit ratio near 100% will force the bank to buy more expensive funding. Management decomposed the 13bp rise in deposit costs by segment.

Q: "You mentioned that cost of deposits went up due to wholesale. I was wondering how much of the growth in deposits was driven by wholesale. And given that you are at almost 100% loan-to-deposit ratio, do you think loan growth may force you to attract more wholesale deposits pressuring the overall deposit costs going forward?"
— Maksym Mishyn, JB Capital

A: "The 13 basis points that you see as the cost of deposit increase in Spain, 13 basis points basically 10 of that increase has come from non-retail, which is mainly corporate. If you look into -- if you isolate for that impact and if you only look into the -- what we call Banca Comercial, which is EUR 150 billion of the roughly EUR 200 billion of the deposits ... the cost of deposit increase for that group is around 4 basis points."
— Onur Genç, CEO

Assessment: A strong, specific answer. The retail franchise is not paying up for balances. Later in the call management added that the 100% ratio is the euro book including CIB funding, and that the Spanish unit runs at about 80%. The deposit pressure is a timing effect of corporate repricing, not a funding constraint.

Whether the Mexican Spread Has Bottomed

A recurring line of questioning asked whether the third quarter marked the low for Mexico's customer spread after another decline. Management said yes, with a caveat on mix.

Q: "On the customer spread, I think the guidance was that it would bottom out in the second half of the year. Is this now the bottom of the customer spread? Could it dip a bit more?"
— Multiple analysts incl. Alvaro Serrano, Morgan Stanley; Miruna Chirea, Jefferies

A: "We do think we reached the bottom. It obviously depends on the mix a bit. If you grow much more on the enterprise side, it has a negative impact on the average spread. But beyond the mix effect, we have seen all the repricing of the loans and the deposits done already."
— Onur Genç, CEO

Assessment: The caveat matters more than the claim. Enterprise lending is the fastest-growing Mexican segment, so the mix effect will keep pressing the average spread even if pricing has stabilized. A flat spread next quarter would confirm the call; another decline would mean the floor is a pricing floor, not a spread floor.

Whether Türkiye's Worst Credit Quarter Is Behind It

A question observed that the bank took the best of Türkiye's margin recovery and the worst of its asset quality, and asked whether the second quarter marked the peak in NPL inflows. The answer conceded the credit guide will be exceeded and placed the cause in macro assumptions.

Q: "While you've taken the better part of NII growth this quarter, you also seem to have taken the worst part when it comes to asset quality. So I wonder if embedded in your guidance for cost of risk in Turkey to be around 220 bps, you're also really telling us that the worst with Q2 is now behind us in terms of NPL inflows and cost of risk in the region."
— Antonio Reale, Bank of America

A: "At the beginning of the year, we were expecting 25% inflation for the year and 32% end of year interest rate. We are very sensitive to those 2 parameters. And 25% inflation turned out to be -- it's going to be around 30%. And the 32% end of year interest rate, it's going to be 36%, maybe 35% ... Asset quality will come out a bit higher than the 220, as Gonzalo mentioned in the presentation, so it will come a bit higher."
— Onur Genç, CEO

Assessment: The question asked whether the worst is over; the answer did not say so. Management explained the miss well and offset it with net interest income, but declined the one forward statement the question was looking for. We read that as no.

How the CIB Book Is Underwritten as It Grows

With Rest of Business loans up 15% in the quarter, a question asked for sector exposure and specifically for comfort on AI-related lending after the press reports, and whether the parent's loan-to-deposit ratio constrains growth.

Q: "BBVA has appeared as a lender to one of the AI build-out in which the market believes there could be some trouble. So how can you reassure about risk taking in this fast-growing loan book?"
— Francisco Riquel, Alantra

A: "It's mainly project finance that we do on the AI topic, mainly project finance, data centers. When you look into those transactions, 95% of those project financings, they are backed by an investment-grade hyperscaler. And the construction risk that you are looking into, a majority of them are in the middle of their construction period and so on. We see very little risk in that."
— Onur Genç, CEO

Assessment: A credible answer on the AI exposure. It did not address the two wholesale impairments the filing discloses in the same area, and nobody asked. The strategic framing, corporate lending that follows existing clients across borders with 90% of CIB revenue client-driven, remains the best defence of the growth rate.

Whether 12.2% Is a Temporary Low, and What It Means for the €36bn

A question set the four-year distribution figure against a CET1 ratio that just fell, and asked whether more risk transfers can rebuild it. Management gave the quarterly risk-transfer sequence and restated the distribution total.

Q: "How should we think about kind of your core equity Tier 1 ratio? What can you do to improve it? Is the 12.2% that you reported today a little bit like a transitionary and you can do more SRTs going forward?"
— Sofie Peterzens, Goldman Sachs

A: "We still stick to that EUR 36 billion. We just looked into the numbers very recently on where we are. There might be more than EUR 13 billion dedicated to growth, which is very good. Once again, as long as it's profitable, as long as it's about cost of equity, with a nice margin, we do think growing should be the priority."
— Onur Genç, CEO

Assessment: The most important exchange on the call, and it says two things at once. The €36bn holds. Growth now has first call on capital, and the growth allocation may exceed plan. Both can be true only if earnings run ahead of plan, which so far they do. If they stop running ahead, growth wins and distribution waits.

Whether Spain's Standalone Cost of Risk Is the New Level

A question computed Spain's standalone third-quarter cost of risk at about 38bp, above the low-30s guide, and asked whether macro inputs drove it. Management attributed it to the reversal of the second quarter's portfolio sale and a small IFRS 9 macro recalibration.

Q: "So the 38 basis points that I get to in the quarter stand-alone, would that be a reference for the future? Was there any reviews of macro inputs here that trigger this?"
— Carlos Peixoto, CaixaBank

A: "House prices are going very strong, but they're going to be growing a bit slower than what we have originally thought, which created a slight negative impact on IFRS 9 provisioning. And then in the second quarter, again, there was a very big one-off, positive one-off on that number. But at the business as usual level, stability and in line with our guidance, we said low 30s in our guidance."
— Onur Genç, CEO

Assessment: Plausible and small. Spain's NPL ratio rose only 4bp and coverage rose 116bp, so this is not a credit turn. It does confirm that the second quarter's record-low Spanish ratio was flattered, as we said in July.

How the New Securitization Rules Change the Capital Arithmetic

A question asked about the European securitization reform approved the week before. Management framed it as disproportionately helpful to BBVA because of its high risk-weight density.

Q: "I wanted to know your expectation on the impact of the reform of securitization that was approved by the European Commission last week."
— Andrea Filtri, Mediobanca

A: "Our RWA density, even after the full program, the model review and everything else that we did, which helped us a bit is now 44%. So RWA density, we have 44%. And the average of our peers is around 27%. So any SRT would be helping us more because it's -- at the moment, it's basically the risk weight of how much capital would an outside investor would be dedicating to this versus us."
— Onur Genç, CEO

Assessment: This is the most constructive capital disclosure on the call. A 44% risk-weight density against 27% for peers is the arbitrage that makes risk transfers unusually valuable for this bank, and management signalled a higher 2027 allocation. It is a medium-term offset to a short-term capital squeeze.

What They're NOT Saying

1. No new distribution and no date for one. The second €1bn tranche of the existing programme is running. With about 21bp above the 12% ceiling, the call restated the €36bn four-year figure and said nothing about when the next programme comes.
2. The Türkiye credit guide was not formally changed. The slide says around 220bp; the CFO said it could be exceeded and the CEO said it will be. No new number was offered, so the size of the overshoot is the market's to estimate. Our simple arithmetic puts the third quarter alone near 290bp.
3. No fourth-quarter Turkish profit figure. The roughly €1bn guide implies about €211M, below the third quarter's €258M, and the profit guide's downward bias is no longer described as slight.
4. The two impaired wholesale customers in Rest of Business were not discussed. The filing discloses them; the call did not mention them, and the AI-lending answer was framed around project finance in general.
5. Still no euro figure for artificial intelligence. The October event set out frameworks, pilots and adoption figures. The third consecutive quarter of AI narrative passed without a cost, revenue or headcount number, while headcount rose 0.2% and expenses 16.0%.
6. The €48bn plan was not raised at the event built for it. Management now speaks of "full confidence to beat" without quantifying by how much.
7. Digital bank economics remain undisclosed. One million clients in Italy and 100,000 in Germany were given. Deposits fell to €10.6bn from €11.9bn at the half, consistent with the promotional German balances running off as signalled. Still no revenue, loss or break-even date.
8. Romania's gain and capital effect are still unquantified. The filing now expects completion in November 2026, inside the fourth quarter. The consideration is "subject to certain adjustments" and no figure was given.
9. USMCA was not raised for a second consecutive call. The filing says the review has "already adversely affected" Mexico's economy and prospects. Nobody asked, and management did not volunteer an update.
10. Spanish general elections appear in the filing's risk factors and not on the call. The filing flags political and regulatory uncertainty "arising, in part, from the upcoming general elections in Spain." For a bank whose Spanish profit carries a special tax on net interest income and fees, that is relevant and was not discussed.

Market Reaction

• Pre-print setup: The Madrid line closed at €22.45 on October 28, up 12.0% year to date against the IBEX 35 up 9.1%, down 10.8% over the trailing thirty days and up 27.7% over twelve months. The 52-week closing range was €17.30 to €25.41, so the stock entered 11.6% below its closing high. The New York ADR closed at $25.01, up 7.3% year to date against the S&P 500 up 13.1%, and down 12.7% over thirty days. Over the two weeks to October 10, European bank stocks fell about 8% on rising sovereign yields, renewed French debt worries and profit-taking after a three-year rally.
• Reaction session (October 29, before-the-open print): The Madrid line opened at €21.23, a 5.4% gap down, traded €21.07 to €21.89 (down 6.1% to down 2.5%), and closed at €21.87, down 2.6% or €0.58, near the top of its range. Volume was 14.4M against an 8.2M thirty-day average, or 1.8x. The IBEX fell 0.1%.
• New York listing: The ADR opened at $24.22, a 3.2% gap down, traded $24.14 to $24.50, and closed at $24.45, down 2.2%. Volume was 2.0M against a 1.6M thirty-day average, or 1.3x. The S&P 500 fell 0.3%.
• Peer reaction: European banks were mixed and mostly lower. CaixaBank fell 1.7%, Bankinter 1.3%, Santander 1.2%, Intesa Sanpaolo 0.5%, HSBC 0.3% and Sabadell 0.3%. UniCredit was flat and BNP Paribas rose 0.5%. The Euro Stoxx 50 rose 0.4%. BBVA was the weakest of the group by about a point.
• Next session (October 30): The Madrid line fell a further 3.7% to €21.06 and the ADR 4.1% to $23.44, in a broad risk-off session in which the S&P 500 fell 1.8%, the Euro Stoxx 50 1.3% and the IBEX 1.1%. Santander also fell 3.7%. We found no BBVA-specific news for the day.
• Since our July note: From the July 30 close of €23.96 to October 30, the Madrid line fell 12.1%, or 10.6% including the €0.37 interim dividend, against the IBEX down 5.6% and the S&P 500 up 1.9%.

The intraday path is the most informative feature of the reaction. The opening gap of 5.4% priced the headline concerns that circulated before the open: provisions above expectations, CET1 below them, and a weak Spanish quarter. Over the session the stock recovered more than half of the gap to close down 2.6% on heavy volume. The second session's 3.7% fall came with the whole market, so we do not read it as a second verdict on the print.

The setup explains why the reaction was not larger. The shares had already fallen 10.8% in thirty days, most of it in the sector sell-off, and entered the print at 2.26 times tangible book against 2.36 times after the July print. In July the stock was at a record and the marginal holder was paying up; in October the marginal holder had already de-risked. A print that was in line on profit and soft on credit and capital produced a modest relative underperformance of about 2.5 points against the home index.

What the market sold is consistent across the commentary we reviewed: the capital ratio and what it implies for the next distribution. That is the right line to focus on. Profit was in line and net interest income beat, but the bank's equity story for two years has been the combination of a 20%-plus return and a hard 12% ceiling with everything above it returned. This quarter the return held and the excess above the ceiling shrank to about €0.9bn.

Street Perspective

Debate: Is Türkiye's Margin Recovery Worth More Than Its Credit Bill?

Bull view: The bull case being made on the Street is that the lira spread recovered 70bp in one quarter with the central bank only now resuming cuts after holding at 37% from January to October, so net interest income should keep compounding through 2027 as the easing cycle continues. Credit costs are elevated but bounded, and management held the profit guide.

Bear view: The bear camp contends that the credit guide has now failed twice, the macro assumptions behind it have been wrong twice, and the area's profit guide implies a fourth quarter below the third. A margin recovery that depends on a central bank cutting into 30% inflation is the least reliable earnings stream in the group.

Our take: The bears have the better argument for the next two quarters and the bulls for 2027. The size is what keeps it from deciding the rating: Türkiye is 7.9% of area profit. What it does decide is our confidence in management's credit guidance elsewhere, because South America showed the same pattern this quarter.

Debate: Has the Capital Engine Slowed, or Did One Quarter Just Look That Way?

Bull view: The constructive reading is that the 20bp fall is a slow risk-transfer quarter in August and a 22bp market move, both reversible, while earnings still generated 71bp. Management has 12bp to 14bp of transfers already in process for the fourth quarter, a new securitization framework that favours high-density banks, and a €36bn distribution figure it restated without qualification.

Bear view: The sceptical reading is that growth is now absorbing more capital than planned, management said so, and the supervisor just raised the requirement by about 40bp while the bank declined to raise its target. Excess capital above 12% is about €0.9bn, the ratio missed the poll by 13bp, and no new programme was announced.

Our take: The bulls are right about the mechanics and the bears are right about the priority. The 22bp market charge and the slow August are temporary. The statement that growth may take more than €13bn is not, and it changes the mix of shareholder return from distribution toward reinvestment at a time when the stock's appeal had been the distribution. We want to see December's CET1 after the risk transfers before restoring the pillar.

Debate: Is Spain Broken or Just Late?

Bull view: The optimistic view is that the spread widened for a second quarter, the mortgage book has not yet repriced to two ECB hikes, retail deposit costs rose only 2bp to 4bp, and the bank is gaining share in the most profitable segments. The weak lines this quarter were trading income marked down by rising yields and a provision comparison distorted by the second quarter's sale.

Bear view: The sceptical view is that Spain's profit fell 5.5% year on year in a quarter when rates rose, net interest income growth slowed to 2.3%, fees fell, and the area came in about a tenth below expectations. A rate-sensitive bank that cannot grow its home-market profit in a hiking cycle has a deposit-pricing problem, whatever the segment detail says.

Our take: Late, not broken. The segment detail on deposit costs is specific and checkable, and the mortgage repricing schedule is contractual. But late has a cost: the Spanish convergence we expected in 2027 now has to come through a deposit book that reprices faster than we assumed. We leave the pillar on track and push our expectation for Spanish net interest income acceleration into the second half of 2027.

Debate: At 2.1 Times Tangible Book, Is This the Entry?

Bull view: The argument on the bull side is that the shares have fallen 12% since July on a sector sell-off while the bank earned 22% on tangible equity, raised net interest income faster than expected, and restated every four-year target. At 2.1 times tangible book and about 10.6 times trailing earnings, a 22% return business is priced for 20%.

Bear view: The bear framing is that the de-rating is not only sector: two credit guides are breaking, capital is tighter, Spain is soft, and the four-year plan relies on returns rising in 2027 and 2028. A sustainable 20% return is a reasonable central case for a bank with this emerging-market mix, which makes 2.1 times fair rather than cheap.

Our take: Both sides are describing the same price from different return assumptions. The stock now trades at the bottom of our fair-value range rather than the top. That removes the valuation objection that drove our July downgrade. It does not create a margin of safety, because the quarter gave us reasons to lower our confidence in the top of the range, not raise it.

Model Framework & Valuation

The table marks our July framework to the nine-month outturn. Net interest income runs at the top of our range; costs, credit and capital move against us. The net effect on our ROTE estimate is small and slightly positive; the effect on our capital and distribution path is negative.

Valuation. At the October 30 Madrid close of €21.06 the shares trade at 2.12 times the €9.95 tangible book per share reported at September 30 and 2.00 times the €10.54 book value. Trailing four-quarter earnings per share of €1.98 put the multiple at 10.6 times. Market capitalization on 5,511M shares is about €116bn. At our July note the same arithmetic read 2.36 times tangible book and 12.7 times trailing earnings. The September tangible book is struck after deducting the €0.37 interim dividend, and the October price is ex that dividend, so the two are on the same basis.

On our residual-income framework, the price now equals the value of a 20% sustainable return, the bottom of the range we published in July. A 21% return supports €22.39 per share and 22% supports €23.63. The shares therefore price the low end of management's own range: the 2026 guide is around 21%, the nine-month delivery is 22.0%, and the four-year commitment requires about 23% in 2027 and 2028. That is no longer a stock paying for the top of the range. It is a stock fairly priced for the bottom of it.

Distribution. Cash distributions against 2025 results totalled €0.92 per share, a 4.4% yield at €21.06. The 2026 interim of €0.37 is 15.6% above the prior year's €0.32. Buybacks executed or running since December total €5.96bn, of which €1bn is the tranche now in progress.

The July Triggers, Graded Verbatim

Our July note published five conditions that would take us back to Outperform and four that would take us to Underperform. We grade each as written.

One downgrade trigger fired and no upgrade trigger fired. The only upgrade trigger whose price condition was met carried an operating condition that was not. Our July note said a second Turkish raise would take us to Underperform. It attached no price condition to that trigger and did not say how it would net against a lower share price. We should have written both in advance.

We are not moving to Underperform, and the reason is a price condition we are applying after the fact, so we state it plainly. An Underperform rating means we expect the stock to trail the S&P 500 over twelve months. The shares now trade at 2.12 times tangible book, which on our framework is the value of a 20% sustainable return. That low case is now our central case. At that return, with a 4.4% cash yield, the shares are fairly priced, not overpriced, and fairly priced is a Hold. For scale only: our arithmetic puts a 30bp to 40bp Turkish overshoot at about €0.1bn of annual profit after tax and minorities, roughly €1bn capitalized at the current earnings multiple, against a fall of about €14bn in market value since July excluding the dividend. But the trigger was not built to measure profit. It was built to test whether management's credit guidance can be trusted, and that test failed. We act on it in three ways: conviction falls to 5 from 6, our central return assumption moves to the bottom of the range, and the Turkish trigger stays on the board for the fourth quarter.

Triggers for the fourth quarter (full-year results, late January 2027). The July triggers above that did not fire stay in force as written. We add three on each side. Our own fourth-quarter estimates are NPL entries of €3.9bn to €4.3bn, December CET1 of 12.2% to 12.4% (roughly €0.9bn to €1.7bn above 12%), and full-year Türkiye profit of €950M to €1bn. Each new trigger sits outside those ranges.

• Toward Outperform: fourth-quarter NPL entries below €3.6bn with coverage stable or higher; a new distribution of at least €2bn announced with full-year results, beyond the €1bn tranche now running; or December CET1 at or above 12.5% after the risk transfers. Each applies only with the shares at or below 2.2 times tangible book.
• Toward Underperform: fourth-quarter NPL entries above €4.5bn; full-year Türkiye profit below €900M; or December CET1 below 12.0%.
• Netting, stated in advance: we count every trigger that fires on each side at the fourth quarter, July's and these, and the Turkish trigger that fired this quarter counts as one downgrade already on the board. If downgrade triggers outnumber upgrade triggers by two or more, we move to Underperform. If upgrade triggers outnumber downgrade triggers by two or more, we move to Outperform. Any other count leaves the rating at Hold. On that count, one further downgrade trigger with no upgrade trigger moves the rating to Underperform.

Thesis Scorecard

Second-Quarter Commitments, Graded

Standing Thesis, Graded

The pillars below are the ones established at our initiation and carried in our thesis of record, with the valuation bear added in July. Where a status tag moved, the move is stated.

Overall: The thesis is unchanged in direction and lower in confidence. The valuation bear that drove the July downgrade has receded. The capital pillar weakened, and the credit bear did not resolve; it spread to a second area. Net interest income did more than we expected and is the reason the earnings path is intact.

Action: Hold. The price came down to the level we said would justify an upgrade, and the quarter did not supply the operating evidence that the same trigger required. A Turkish credit trigger fired, and we carry it into the fourth quarter instead of executing it at a price that already equals our low-case value. That leaves the stock fairly priced for a 20% return business, and our confidence in anything better than 20% is lower than it was in July. Holders should stay. New money has a better case here than at any point since our initiation, but it should wait for December's capital ratio and the fourth quarter's NPL entries, which are the two numbers that decide which side of the range this bank belongs on.