Non-performing loans (NPLs) are loans 90 days past due or judged unlikely to be repaid. The NPL ratio is NPLs divided by total loans. Common equity tier 1 (CET1) is the core capital that absorbs losses; the CET1 ratio is CET1 capital divided by risk-weighted assets (RWA). This piece outlines how NPL trends and the final Basel III package will interact to pressure capital ratios, how that flows into 2025-2026 pricing and strategy, and where the practical opportunities lie for banks and investors.
Why rising NPL risk collides with Basel III at a tough moment
Rising NPL ratios in parts of Europe are meeting Basel III finalization at the wrong time for CET1. Provisions and prudential deductions pull down the numerator, while output floors and model changes push up the denominator. That double squeeze will vary by bank, but the direction is clear: less headroom and higher marginal capital costs through 2025-2030.
The current readings and the tells that matter
Headlines look steady, but the internals are moving. The European Banking Authority (EBA) pegs the EU aggregate NPL ratio at 1.8% as of Q2 2024, flat quarter on quarter. Stage 2 exposures sit at 9.5%, well above pre-pandemic norms. For ECB-supervised significant institutions, the NPL ratio runs at 2.27%. Different samples, same takeaway: stable reported NPLs with a lingering Stage 2 overhang that is a key near-term watchpoint.
Why the portfolio mix is the problem
Commercial real estate values have fallen, with the ECB noting broad price declines and rising refinancing stress. MSCI estimates pan-European all-property capital values down roughly 18% peak to trough by Q1 2024, driven by yield expansion. Lower collateral values lift loss-given-default assumptions and provisions even before defaults rise, which drags on capital and earnings.
Retail and SME books show early arrears consistent with higher rates. The ECB’s deposit facility rate at 3.75% as of October 2024 still bites, and many mortgages reprice into 2025. Stage 1 to Stage 2 migration continues across retail and SME, pulling forward lifetime expected credit losses that pressure earnings over the next 12-18 months.
Country mix matters. Germany and the Nordics have large CRE exposure and shorter debt maturities. Italy and Spain carry SME and consumer tail risk. Greece and Cyprus have lower on-balance sheet NPL ratios after major securitizations and sales, but recoveries now hinge on servicer execution and secondary trades. Expect CET1 pressure where CRE and short-dated refinancing coincide, and where SME and consumer books are large.
How NPLs hit CET1 and inflate RWA
Provisions are the first-order hit. Under IFRS 9, Stage 2 loans require lifetime expected losses, and Stage 3 loans recognize interest income net of provisions. Provision expense reduces profit and CET1. The EU’s transitional IFRS 9 add-back that softened the 2020-2022 impact has largely phased out by end 2024, removing a cushion. With Stage 2 elevated, a small default uptick can produce a disproportionate provision step-up, which leaves earnings more sensitive. For mechanics on staging, see IFRS 9 staging rules.
The NPE backstop adds a second channel. Under CRR Article 47c and Delegated Regulation 2019/630, banks must deduct from CET1 any shortfall to minimum coverage levels on newly originated NPEs – 100% for unsecured after 3 years, and for secured after 7 to 9 years depending on collateral. Thin collateral or long recoveries accelerate deductions as vintages age. The practical effect is to nudge sales or restructurings sooner to reset the clock and preserve capital.
RWA inflation arrives via several routes:
- Output floor: The EU starts phasing in the 72.5% standardized floor on January 1, 2025 through 2030. The EBA estimates a 9% average rise in minimum required capital at full implementation across large EU banks, with the floor a big driver where internal ratings based risk weights are low on mortgages and corporates.
- IRB parameter drift: Post IRB repair and ongoing validations are pushing probability of default and loss-given-default higher in segments like CRE and SME, reflecting downturn and long-run averages. That inflates IRB RWAs even before the floor binds.
- CCF and operational risk: CRR3 raises credit conversion factors on some off-balance sheet items and moves operational risk to a standardized approach based on business indicators. Non-credit RWAs rise for some banks.
- Market risk and CVA: Banks with trading books will see higher market risk and CVA RWAs. Credit portfolios drive the NPL story, but total RWAs still go up.
Put the pieces together: a bank with €100 of CET1 on €1,000 of RWA (10% ratio) that takes €5 of provisions and backstop deductions, and sees RWA rise 10% to €1,100 from floors and parameters, ends at €95/€1,100 = 8.6%. The operating buffer shrinks. When the buffer shrinks, the price of every marginal RWA goes up. Banks feel that in loan pricing and in portfolio choices starting in 2025.
What 2025-2026 will likely bring
Aggregate capital looks firm. The EBA shows a fully loaded CET1 ratio of 15.6% as of Q2 2024. But buffers are uneven by bank and country. Some mid-sized lenders carry concentrated CRE or SME exposure; some face active countercyclical capital buffers. Stack composition matters too: AT1 and Tier 2 constraints can make leverage or TLAC the binding metric instead of RWA for certain issuers.
Stage 2 persistence points to lagged losses. If growth slows or rates stay higher for longer than modeled, impairment will drift up. The ECB notes a larger share of borrowers with high debt service to income ratios among variable-rate loans. If unemployment moves up, those tails get fatter and earnings sensitivity rises over the next 6-12 months.
CRE remains the main accelerant. Valuations lag transactions in thin markets. Loans underwritten in 2021-2022 with low margins and tight exit yields face either equity cures or restructurings at refinancing. Both outcomes weigh on capital unless banks move assets off balance sheet.
Risk-weight inflation is largely a 2025 story. CRR3 and CRD6 apply from January 1, 2025, with transitional relief but mechanical RWA rises for IRB-heavy banks. Credit officers are already baking higher capital charges into grids and origination thresholds.
How banks will respond – and where investors fit
Expect balance sheet actions that maximize capital velocity in the face of higher marginal RWA costs.
- Tighter underwriting and repricing: Marginal lending slows in capital-intensive segments. Pricing widens in unsecured consumer, SME, CRE, and for undrawn commitments facing higher credit conversion factors.
- Significant risk transfer: Synthetic securitizations and other credit risk transfer trades are the fastest way to reduce RWAs without shrinking assets. Issuance and demand were robust in H1 2024 and should build into 2025. Mezzanine spreads will widen with default expectations, but economics are driven by RWA relief against scarce CET1. For a primer, see Significant risk transfer.
- NPL disposals and restructurings: Primary NPL sales remain below the 2015-2019 clean-up. The mix is tilting to secondary trades of legacy securitizations, granular unsecured pools, and reperforming loans with forbearance. CRE and SME portfolios will feature more as backstops ramp and workout teams hit bandwidth limits.
- Servicing capacity: Banks and buyers will favor servicers with proven collections, litigation, restructuring skills, and clean reporting. Capacity is scarce, so governance and transparency command a premium.
For private capital, three practical openings stand out:
- CRT mezzanine: Underwrite mezzanine tranches on performing portfolios with thick first-loss protection and strong originator alignment. Focus diligence on asset selection, historical loss dispersion, trigger design, excess spread mechanics, and commutation rights. SRT rules require quantitative and qualitative tests, risk retention, and loss distribution that keep structures disciplined.
- NPL portfolios with capital-light financing: Target granular unsecured consumer and SME pools, and reperforming mortgages with forbearance. Price to realistic recovery curves, updated collateral values, and higher legal costs. Execution wins with clean data, limited conditionality on reps and warranties, and short timelines.
- CRE structured solutions: Provide rescue capital, preferred equity, or loan-on-loan to transitional assets, often alongside banks seeking to reduce exposure while keeping clients. Build cash sweeps, interest reserves, and step-up margins. Underwrite to current rents and cap rates; add capex cushions for energy performance rules. Where appropriate, use mezzanine financing structures to improve downside protection.
Edge cases and governance pitfalls to avoid
- Provision overlays: Many banks still carry top-down IFRS 9 overlays from recent shocks. Supervisors want overlays justified and released in step with risk. Early release flatters near-term earnings but leaves less protection when defaults arrive.
- Backstop timing: The NPE backstop clock starts when a loan becomes non-performing. Late recognition shortens the runway to required coverage, creating larger CET1 deductions later.
- Model risk and floors: IRB changes need approval. Supervisors can impose floors or add-ons. Banks counting on favorable recalibration may not get it, so assume conservative supervisory stances in 2025-2026.
- Valuation lag: External appraisals can be stale in thin CRE markets. Rely on independent refreshes, rent rolls, tenant credit, and market-based cap rates.
- CRT structure risk: Scrutinize excess spread, alignment of interest, counterparty strength of protection providers, and commutation triggers that could strand mezzanine. Stress correlated tails, not just averages.
- NPL data and legal frictions: Errors in borrower IDs, liens, or court status still show up in tapes. Bid with contingencies, escrow, and price adjustment mechanics for defects.
Regulatory overlay and usable buffers
Banks start from strong reported capital and liquidity. Liquidity coverage and stable funding ratios sit well above minimums. Several countries maintain positive countercyclical buffers, which can be released if conditions worsen. That said, buffers are slow to adjust, and management teams avoid running them down because markets notice and funding costs can rise.
CRR3 and CRD6 timing matters. Output floors phase in from 2025 to 2030. Revised standardized approaches raise RWA density for off-balance items and specialized lending. CRE and income-producing real estate get more risk-sensitive treatment. CVA and market risk reforms add RWAs for banks with market activity. Implementation is complex, so smaller banks will triage RWA-heavy lines first for remediation or de-risking.
Supervisory focus is tightening. EBA guidelines on origination and monitoring are raising data and underwriting standards. Supervisors are watching leveraged finance, interest rate risk in the banking book, and concentration risk. Pillar 2 Guidance can move with supervisory judgment, affecting usable buffers and dividends even when headline CET1 looks robust.
Valuation and pricing implications banks and investors should model
- Performing loans: Banks’ all-in capital cost is higher. Pricing grids now reflect higher RWAs and more volatile expected loss. Borrowers with thin coverage see repricing or restructuring. Private credit steps in where banks want club solutions or capital-light exposure.
- CRTs: Spreads price both expected losses and the value of capital relief. If a bank can free €1 of CET1 by paying €0.30 in present value premiums, it will do the trade while CET1 is the binding constraint. As output floors bite, protection value falls for portfolios already floor-bound; banks will target portfolios where IRB still binds.
- NPLs: Bid-ask depends on coverage, collateral realism, and servicer credibility. EU NPL coverage ratios around 44% as of Q2 2024 give room to clear sales where recoveries align. Where gaps exist, expect deferred consideration, profit-sharing, or servicing-as-a-service to bridge valuation. Insist on granular tapes, reps on title and litigation status, and robust indemnities.
Practical diligence kill tests
- Stage migration: If Stage 2 exceeds 12% of gross loans at a retail-heavy bank with no clear de-risking plan, model elevated provisioning over 12-18 months unless overlays are large and well supported.
- CRE concentration: If CRE loans exceed 10% of total and valuations refresh less than annually, raise loss-given-default assumptions and widen mezzanine CRT spreads.
- Refinancing cliffs: If more than 25% of SME or CRE books mature within 24 months, assume more forbearance and rescheduling, with NPE backstop pressure building over time.
- Output floor sensitivity: Model RWA with and without floors by exposure class. If over 25% of RWA will be floor-bound at full phase-in, steer CRT to portfolios where IRB still binds or rotate to assets already at standardized weights.
- Servicer bandwidth: If expected annual case throughput requires a 30% plus increase in headcount versus the last two years, haircut collection curves unless automation is proven.
- Enforcement times: If collateral enforcement averages more than four years in a jurisdiction, haircut collateral values and bias to consensual restructurings.
What to watch next
- EBA Risk Dashboard: Quarterly updates on Stage 2, coverage, and sectoral NPLs. A move above 10% Stage 2 EU-wide would be a clear deterioration sign.
- ECB supervisory statistics: Country and sector splits, especially CRE NPLs and provisioning.
- CRE valuations and volumes: MSCI and national series. Further declines from Q1 2024 troughs would push loss-given-defaults higher.
- CRR3/CRD6 implementation: National transpositions, output floor phase-in, and any early floors.
- CRT supply and pricing: AFME issuance volumes and mezzanine spreads. A supply jump with softer demand improves investor terms.
- NPL pipeline trackers: Secondary sales of reperforming mortgages and unsecured pools indicate monetization over in-house workout.
Conclusion
NPL ratios have not spiked, but the mix of risks and the regulatory calendar work against stable CET1 at a subset of banks. Persistent Stage 2 shares, CRE valuation resets, and refinancing pressures will lift impairments through 2025. Basel III finalization raises RWAs for many IRB-reliant lenders, compressing CET1 even with modest losses. Banks will respond with credit risk transfer, sharper underwriting, and targeted portfolio sales. For investors, scarce risk-weighted capital is creating willing sellers of capital relief and assets. The most attractive targets are capital-efficient for banks and straightforward to analyze for buyers: mezzanine CRTs on diversified portfolios with strong performance data; unsecured NPLs with short legal cycles and reliable tapes; and CRE structured deals with transparent collateral and credible business plans. Underwrite to today’s cash flows and capital rules, build governance protections into structures, and assume supervisors prefer conservative treatment over leniency.