Europe's Banks Q3 Review: After Margin Compression, the Provisioning Test

The margin story in European banking is over, and it ended the way it had to. Net interest income across the largest euro-area listed lenders fell roughly 3% year over year in the third quarter as deposit betas finally caught up with the rate cycle. That is a milder compression than the first half implied, and it is the reason most of these banks held full-year guidance. But the quarter was not decided by the revenue line, and the market did not trade it that way.

The mechanics are straightforward. Deposits repriced toward the policy rate with a lag of roughly four quarters, so the benefit of higher rates peaked for these banks somewhere in the first half of 2026 and has been fading since. With the ECB holding rather than cutting, that fade should flatten in the first quarter of 2027 rather than reverse, which is why guidance for flat net interest income next year was treated as credible. Hedge books maturing at lower yields are the swing factor nobody can see from outside.

Provisions are where the dispersion appeared. Cost of risk ran at roughly 28 basis points annualised against 22 basis points a year earlier, a modest increase that hides three very different buckets. French sovereign exposure is being marked against a spread that widened through October, commercial real estate in Germany and the Nordics is still grinding through refinancing walls, and small-business credit in southern Europe is deteriorating from a very low base. The banks that added to provisions in all three posted the weakest share price reactions of the reporting season.

Capital remains the strongest part of the story and the most fragile part of the equity case. Aggregate CET1 sits near 15.4%, comfortably above requirements, and buybacks have become the default use of excess capital with payout ratios around 50%. That is sustainable while cost of risk stays near 30 basis points. At 50 basis points, which is what a genuine European slowdown would produce, the buyback line is the first thing to go and the dividend is the second, and equity holders discover that a 15% capital ratio is a regulatory buffer rather than a shareholder asset.

Valuation already reflects some of this. Euro-area banks trade close to tangible book, up from roughly 0.6 times two years ago, and the re-rating has been driven almost entirely by returns on tangible equity rather than growth. The problem with buying a bank at tangible book is that it requires you to believe the through-the-cycle cost of risk, not the current one. French sovereign stress is the specific risk that could reset that assumption: a 20 basis point widening in OAT spreads is a mark-to-market event for domestic lenders before it is a credit event.

Three things decide the next leg. Fourth-quarter guidance on cost of risk, which most managements left unchanged and will have to revisit in January if French spreads stay wide; the 2027 net interest income trajectory once hedge rolls are disclosed in full; and the ECB supervisory review cycle, where commercial real estate classifications are still being debated bank by bank. None of those is a revenue story. All three are why this quarter should be read as a provisioning quarter rather than a margin one.