Three years after the 2023 turmoil, the four major IRRBB regimes are not converging. They are answering the same question in increasingly different ways, and the 2026-2028 pipeline will widen the gap, not close it.
That is not the story most people expected. A common crisis usually produces common rules. Instead, each jurisdiction took from March 2023 exactly the lesson that confirmed its own philosophy, and each is now proceeding along its own track on its own clock. Here is the map, regime by regime, of what actually moves between now and 2028.
The recalibrated shock scenarios (d578) have been in force at Basel level since 1 January 2026: currency-specific shock factors, a 99.9th percentile calibration and finer rounding, with the six scenarios themselves unchanged. The structural event to watch was always different: the Committee's assessment of whether the 2016 IRRBB standard is achieving its intended objectives, the workstream that keeps the Pillar 1 question formally open.
On 1 October 2026 that assessment produced its first output. The Committee reported that it had empirically assessed the standard, identified specific shortcomings in banks' management of IRRBB, and agreed to consult on additional Pillar 2 guidance for banks and supervisors, with the consultation due in November. Read the direction carefully. The verdict lands on how banks manage the risk, not on the architecture: Pillar 2 guidance, not a Pillar 1 charge, the supervisory rather than legislative path Basel has followed since 2016. But it is now a live consultation, and every other regime prices its own choices off the Basel baseline. How each jurisdiction absorbs the new guidance, in binding rules, in supervisory expectations or not at all, becomes one more axis of divergence.
The machinery is built; the pipeline is about tuning it. The EBA still has to adopt the recalibrated d578 shock table into the supervisory outlier test RTS (Delegated Regulation (EU) 2024/856), and until it does, EU SOT numbers sit on the original calibration. The EBA has said it will assess the impact on EU institutions before considering a review of the RTS; as of October 2026, no amending draft has been published.
CSRBB is regulation ahead of methodology. The EBA's heatmap work has already flagged divergent CSRBB perimeters, and if measurement dispersion persists, escalation to technical standards is the live option; banks should assume those requirements ratchet only one way. The 2026 reporting consultation (EBA/CP/2026/07) opens the first door to simplifying the J-template burden, though the likely landing zone is proportionality refinement rather than retreat: assumption-level data now powers EU supervision and will not be handed back.
And the NII SOT threshold remains quietly contested. The EBA drafted the test at 2.5% of Tier 1. When the Commission sought to amend it with a ranking approach that the EBA warned would make the NII test "factually less stringent", the EBA instead raised the threshold to 5% as a pragmatic fix for the post-2022 rate environment, and committed to reconsider the level, and possibly the methodology, as conditions change. That reconsideration has not yet happened. Recalibration remains possible.
The UK closed its shock gap quietly. Through a low-impact rule amendment, consulted on in July 2025 and finalized in October 2025 without a single response, the PRA updated the currency shock values in the Internal Capital Adequacy Part of its Rulebook to the Basel recalibration, applying from 1 July 2026. UK banks now run the recalibrated table, six months after the Basel date and ahead of the EU.
The bigger story is Pillar 2A review Phase 2: the more in-depth review of individual Pillar 2A methodologies that follows the Phase 1 policy statement (PS15/26, May 2026). The PRA has said the Phase 2 consultation will come in 2027. If it reaches IRRBB, it will be the first methodological reopening of the PRA's IRRBB capital approach, essentially unchanged in its statistical design for a decade, since it was built. Add the Future Banking Data redesign's eventual verdict on FSA017, the gap report whose vintage shows. None of this threatens the architecture: the UK's model of binding rules plus supervisor-computed capital emerged from 2023 looking vindicated, and the PRA is in no hurry.
One consequential action, one open question. The consequential action is finalization of the March 2026 capital re-proposals, which would remove the AOCI opt-out for Category III-IV banks and phase unrealized securities gains and losses into CET1 over five years. Comments closed on 18 June 2026; as of mid-October, no final rule has been issued. When it lands, it will be the single most consequential US action on banking-book rate risk, and partial Pillar 1 capital in everything but name, aimed squarely at the unrealized-loss channel at the heart of SVB's failure.
The open question is durability: whether the 2025-26 supervisory refocus on material financial risks holds, and whether the interagency guidance triad is eventually consolidated. There is still no prescribed calculation, no outlier test and no dedicated IRR report in the US framework. The re-proposals change capital arithmetic, not supervisory philosophy.
The common item on all four lists is the first supervisory verdicts on deposit models recalibrated with post-2022 data, and the first real test of the recalibrated shocks against a genuinely new rate environment. Nobody yet has through-the-cycle data from the digital-deposit era; the 2026-2028 window is when supervisors first grade the models built without it.
Step back and the pattern is unmistakable. The EU is codifying further. The UK is recalibrating its own distinctive capital machinery. The US is moving capital arithmetic while leaving measurement unprescribed. Basel is reaching for supervisory guidance rather than new rules. These are four trajectories with different directions, not one trajectory at different speeds.
Divergence is not a transition phase on the way to convergence; for IRRBB, it is the operating environment. Even the one nominally shared object, the six prescribed shocks, now sits at different calibrations in different rulebooks: recalibrated at Basel and, since July 2026, in the UK; still on the original calibration in the EU, pending the RTS update; and never prescribed at all in the US. A banking group spanning the jurisdictions already produces several legitimately different ΔEVE figures from one balance sheet; the pipeline above adds moving parts to each.
For practitioners, that dictates the strategy. Waiting for harmonization is not a plan, because harmonization is not coming in this window. The workable approach is a single control framework built to satisfy the strictest applicable version of each requirement:
Banks that build to the strictest version pay once; banks that build jurisdiction by jurisdiction pay every time one of the four regimes moves.
The map will need redrawing as the Basel consultation, the EU RTS update, the PRA's Phase 2 consultation and the US finalization land. The direction of travel, though, is already legible. Four regimes, four answers, and a widening premium on knowing all of them.
For the complete picture, from EVE and NII sensitivities and the six supervisory shock scenarios to outlier tests and how Basel, the EU, the UK and the US differ, see Mirai RiskTech's guide IRRBB regulation across Basel, the EU, the UK and the US: a complete reference.