Basel 3.1 / CRR3 and the capital squeeze on CRE lending

Basel Endgame — divergence across jurisdictions

United Kingdom. The PRA has delayed Basel 3.1 twice. Final rules were published in PS1/26; the main capital regime now applies from 1 January 2027, with the transitional period running to 31 December 2029.

European Union. CRR3 / CRD VI came into force on 1 January 2025, but the FRTB market-risk package has been postponed to 1 January 2027. The EU has also used its discretion to extend the output-floor transition to 2032 (rather than 2028).

United States. No final rule. A revised Fed proposal is expected in 2026, reportedly aimed at easing the calibration for large U.S. banks. Non-banks — outside the Basel perimeter altogether — continue to take share.

For cross-border lenders, the divergence itself is the problem. A German Pfandbrief bank, a London branch of a U.S. bulge-bracket and a Paris insurer now face materially different capital costs for the same underlying real-estate exposure. That asymmetry is a primary driver of the six products covered on this site.

Mechanics of the squeeze on CRE balance sheets

Three mechanics are doing the work on banks' CRE books.

Output floor. IRB RWAs cannot fall below 72.5% of the standardised calculation. German, Dutch and Nordic lenders whose internal models priced low-LTV income-producing CRE at sub-30% risk weights lose that advantage. Because CRE standardised risk weights remain relatively high, the floor bites hardest where it hurts most.

Revised standardised approach for real estate. The new SA is granular and LTV-sensitive. It removes the old 100% blanket floor for exposures not materially dependent on property cashflow — which is directionally helpful for investment-grade lending — but layers on stricter valuation rules, a revaluation cap, and more punitive treatment of transitional / development lending.

Operational and concentration overlays. CRR3's new operational-risk framework (SA-OR) and the ongoing EBA focus on CRE concentration mean that, even before the output floor, CRE loans consume more of the bank's scarce capital budget than they did under Basel II.

The €315 bn refinancing wall

Against this tighter capital backdrop, roughly €130 bn of European CRE loans matured in 2025, with a further €185 bn due in 2026. Capital values have partially recovered from their 2023 trough but are still materially below origination LTVs for vintages struck in 2020–2022. The result is a structural funding gap that no single lender, and no single product, can close alone.

Three bank playbooks

Faced with Basel 3.1 and the refinancing wall, banks are gravitating towards three playbooks, all visible today in the European CRE market.

Raise capital. Politically expensive, particularly for institutions with state or Länder shareholders. Rarely sufficient on its own.

Retrench. Scale back from capital-inefficient business lines — development lending, high-LTV transitional deals, certain geographies. This is already creating a funding vacuum that private credit is filling.

Re-engineer. Keep the customer, move the capital. CMBS (distribute the asset), synthetic securitisation — the product by which Significant Risk Transfer (SRT) is most commonly pursued — (distribute the risk), back-leverage (fund the non-bank lender that takes the asset) and fund finance (finance the fund that holds the loan) are four tools in a single re-engineering playbook. CFO technology and bespoke structures give the stack further reach.

EU Taxonomy and sustainable-finance overlays

Under Regulation (EU) 2020/852 and Delegated Regulation (EU) 2021/2139, real-estate activities and the financing that supports them fall within the scope of taxonomy-eligible and, where the technical screening criteria are met, taxonomy-aligned activities. For CRE structured credit the practical consequence is that lenders subject to the SFDR, to the EU Green Bond Standard or to internal sustainable-finance frameworks increasingly test the financed asset against the activity description and the Do-No-Significant-Harm (DNSH) criteria before pricing.

For practice, the two items that matter in the financing structure are (i) a clean Taxonomy-alignment statement from the borrower at funding (with supporting technical evidence, not merely a legal assurance), and (ii) an undertaking to maintain alignment over the financing term, including DNSH compliance where retrofits or repowerings would otherwise shift the asset out of alignment. Taxonomy misalignment is not in itself a default under most current facility agreements but increasingly appears as a draw-stop or margin-ratchet trigger.

Circularity and the supervisory agenda

The single largest current concern on the supervisory agenda is the circularity between junior-tranche investors in synthetic securitisations achieving SRT recognition and the banks that finance those investors through back-leverage facilities — with exposures stacked across fund finance and potentially topped with a CFO layer. The PRA, ECB and FSB have all published on this risk. Expect supervisory attention to tighten over 2026–2027, with particular focus on:

• disclosure expectations for investors in SRT tranches that are themselves levered (including where the financing is cross-border);
• stress-testing of the underlying CRE portfolio under scenarios that move multiple layers of the stack concurrently;
• the robustness of enforcement rights under the back-leverage documentation against a concurrent enforcement by the protection-seller under the synthetic-securitisation documentation; and
• the cross-institution analysis of how a single CRE credit event would propagate through the chain.

Documentation should anticipate heightened disclosure and stress-testing expectations over the life of the deal, rather than positioning them as post-closing supervisory hygiene.

Liquidity: the LCR changes

The capital debate gets the attention; the liquidity treatment decides who is allowed to hold the paper. Under the CRR amendments in trilogue the LCR haircut on securitisation positions falls from 35% to 25%, and as far as 15% for larger tranches of the new “resilient” category. Credit quality steps 1 to 7 become eligible as Level 2B high-quality liquid assets, and the five-year cap on weighted average life disappears, with EBA monitoring over four years.

For CRE that last point is the significant one. A cap on weighted average life sits badly with the tenor of commercial real-estate paper, and its removal widens the set of bank treasuries that can take the asset at all — which reaches secondary liquidity, and therefore pricing, before it reaches anyone’s capital ratio.

The capital side of the same instrument is less uniformly favourable. New minimum risk-weight floors under Article 259 CRR run at 5% for an STS originator, 10% for a non-STS originator and 12% for a non-STS investor, with 7% and 12% for resilient positions. The SEC-SA p-factor is set at 0.3 and 0.5 for STS and 0.6 and 1 for non-STS — and, as Brandt notes, some of those values sit above the transitional treatment of Article 465(13) CRR. CRE portfolios rarely reach STS. For a good part of this market the reform is a tightening.

The other half of the file

Basel 3.1 governs the cost of holding CRE risk. The parallel review of Regulation (EU) 2017/2402 governs the cost of moving it — due diligence, transparency, the definition of a private transaction, the SRT test, and the eligibility of insurers as protection sellers. It is in trilogue, with the next session on 29 September 2026 and no date of application in sight.

The two files are one policy. The capital relief in the CRR amendment is the price of the process burden in the Regulation, and neither reads correctly alone. The state of play, the open points and the calibration are set out on Securitisation Regulation 2.0.

The honest summary of the whole exercise is Huertas’s: without a recalibration of the insurance treatment under Solvency II, the reform leaves the market safer and clearer, but not materially larger. Against a EUR 315 bn refinancing wall, that is the sentence on which the rest of this page depends.

Last reviewed: 13 September 2026. Data and primary sources: Resources. Short definitions of the technical terms used on this page: Glossary.