The Ezelman Regulatory Thesis 2026–2030.

The four structural calls we make to our mandate roster, reduced to four readable sections. Published in the belief that the regulatory agenda that matters for the next five years can be stated in 2,000 words and reconciled to the published rulebook in four footnotes.

Call 1 — CRR3 productionisation

The CRR3 regulation entered application on 1 January 2025. Every bank that needed to be compliant on the day was compliant on the day. That is not the interesting part. The interesting part is what happens over the three following years, during which the banks move from meeting the deadline to running CRR3 as business-as-usual.

Productionisation is not a technical footnote. It is the workstream in which the bank's RWA calculators stabilise, the data-lineage defects surface, the Pillar 3 disclosures harden, and the Output Floor binding becomes measurable rather than hypothetical. We observe three recurring patterns across the 2025 go-live cohort:

• Data-lineage remediation — the 2025 go-live was met using tactical data feeds in roughly 70% of G-SIBs we worked with. The strategic remediation runs into 2027–2028.
• Output Floor misreading — banks that declared themselves non-binding on 1 January 2025 are, in several cases, already crossing the 50% threshold by mid-2026, as the phase-in ratchet bites.
• Pillar 3 drift — the first post-CRR3 Pillar 3 disclosures (H1 2025) showed large presentational and methodological inconsistencies across peer banks. Convergence will take 18–24 months.

The Ezelman call: between 30 and 60 bps of CET1 are available to each G-SIB in the productionisation window, and that capital is lost to the banks that treat productionisation as an IT workstream instead of a capital workstream.

Call 2 — Pillar 2 dispersion

Over the 2022–2025 SREP cycles, the range of Pillar 2 Guidance (P2G) add-ons across large European banks widened from a peer band of 20–60 bps to an observed range of 0–120 bps. We expect this dispersion to continue widening through 2030.

The mechanism is not mysterious. The ECB has signalled repeatedly that it will use Pillar 2 as the primary lever for differentiated supervision, and that banks that invest in the governance quality of their stress-testing and ICAAP engines will be rewarded with lower add-ons. Banks that do not will be penalised asymmetrically.

The Ezelman call: three structural levers drive a bank's position inside the dispersion band.

• Stress-testing engine quality — top-down/bottom-up reconciliation discipline, reverse stress testing governance, board-level narrative coherence.
• ICAAP-SREP dialogue maturity — whether the bank's submission engages the supervisor's concerns or merely rebuts them.
• OSI remediation track record — whether findings from the last three inspections have translated into visible governance change.

Banks in the top quartile of these three dimensions have consistently exited 2025 with zero P2G add-on. Banks in the bottom quartile have consistently exited with 60–120 bps. The dispersion is legible; it is also addressable.

Call 3 — GCC parity with the EU

The GCC regulators — SAMA in Saudi Arabia, CBUAE in the UAE, and QCB in Qatar — have signalled closer alignment with the EU's CRR3 than with the US's Basel III Endgame, on timing and on substance. This is not obvious. It is a structural call the thesis makes on the basis of three observations.

• Standards imported — the Output Floor, revised CVA, operational risk SA, and FRTB rule sets are being adopted largely as-is from the EBA rulebook, with local calibrations on concentration limits and Shariah-compliant products.
• Timing imported — the implementation dates track the EU's 2025–2030 phase-in schedule, not the US's post-Endgame schedule.
• Supervisory practice imported — SAMA's evolving on-site supervisory model draws explicitly on SSM and PRA practice rather than US inter-agency practice.

The Ezelman call: for Tier-1 GCC banks, the EU playbook is the playbook. Programme templates, methodological choices, governance patterns and even the language of board-level engagement translate with only local calibration. This is the structural reason Ezelman operates a GCC desk on the same methodology as the Paris/Frankfurt desk.

Call 4 — US-UK drift from CRR3

The US Basel III Endgame proposal (FRB / OCC / FDIC, 2023–2024) and the UK PRA's Strong & Simple / Basel 3.1 implementation diverge from CRR3 in ways that matter operationally, not just academically. Three divergence points are material for banks with EU, UK and US subsidiaries:

• Operational risk SA calibration — the US Endgame proposes a materially different internal loss multiplier treatment than the EBA rule set.
• Output Floor binding — the PRA's framing of the Output Floor differs in application date and calibration from the CRR3 phase-in.
• FRTB desk structure — US Endgame and EU CRR3 differ on internal model approval scope and on the boundary between the trading book and the banking book.

The Ezelman call: banks with EU-UK-US footprints will need to run three separate programme streams, not one translated programme. The cost of assuming translation is a 6–9-month delay at the subsidiary end, plus a material supervisory trust deficit that compounds over the next SREP cycle.

Five-year regulatory milestones (as we see them)