PRA Basel 3.1 Market Risk IMA Adjustments: What UK Banks Must Update in Capital and Pillar 3 Reporting

ByOfficer

Last updated: June 2026

A trading desk that loses internal model treatment can face an immediate change in market risk capital. Under the PRA Basel 3.1 market risk rules, a desk that fails the model tests falls back to the standardised approach, and for many portfolios the standardised number is materially higher than the model number the desk was built around. That is the operational stakes behind the PRA’s June 2026 consultation. The regulator is changing the tests a desk has to pass to keep its internal model approach, and it is changing the timetable on which those tests bite. If your capital plan, PRA regulatory reporting returns, and Pillar 3 market risk disclosures still assume the rules as they stood in early 2025, they now need a CP9/26 change log covering PLAT, RFET, NMRF, ASA/IMA aggregation, IMA treatment of CIUs and consequential reporting/disclosure changes.

On 19 June 2026 the PRA published CP9/26, Basel 3.1: Adjustments to the internal model approach (IMA) for market risk. It is a further substantive PRA consultation on the UK’s Basel 3.1 market risk framework, and it sits on top of an already-confirmed implementation date split: the bulk of Basel 3.1 takes effect on 1 January 2027, but the internal model approach for market risk has been pushed back a full year to 1 January 2028. The consultation closes on 18 September 2026. This article walks through what actually changed, who it touches, and what reporting teams need to map before the interim year begins.

One framing point before going further. This is a UK item, not an EU one. The PRA implements the Fundamental Review of the Trading Book through its own near-final rules and supervisory statements, not through the EU’s Capital Requirements Regulation. The article numbers, the implementation dates, and the consultation references are all British. Reading EU FRTB guidance and assuming it maps cleanly onto a UK trading book is one of the easiest and most expensive mistakes a cross-border group can make right now.

Related reading: EU Basel III market risk: FRTB and trading book capital

PRA Basel 3.1 Market Risk IMA: What CP9/26 Changes for Capital

CP9/26 is a consultation, not a final rule. That distinction governs everything below. Nothing in CP9/26 is binding yet, and a firm building to it today is building to a proposal that may move before the PRA confirms it. The consultation period runs to 18 September 2026, and the PRA proposes that the implementation date for the IMA, including the changes that come out of this consultation, will remain 1 January 2028.

The headline adjustment is to the profit and loss attribution test. Under the final PRA Basel 3.1 rules in PS1/26, the PLAT would have had a one-year monitoring period after FRTB-IMA implementation. CP9/26 proposes to extend that monitoring period to three years. During that monitoring period, the PLAT would not be binding: a failed Spearman correlation test or Kolmogorov-Smirnov test would not automatically make the trading desk ineligible for IMA capitalisation, although the desk would remain subject to supervisory review and possible PRA action if model-performance or risk-management concerns arise.

The second adjustment is to RFET and NMRF treatment for risk factors with limited trading data. CP9/26 proposes to reduce the minimum verifiable-price requirement from 24 to 16 for risk factors with a liquidity horizon greater than 20 days, while keeping the other quantitative RFET requirements unchanged. It also proposes pro-rating the verifiable-price requirement for new issuances during the first 12 months after issuance. For NMRFs, CP9/26 proposes a Type 1 category for risk factors that fail the quantitative RFET but pass the qualitative data standards: those risk factors would remain in the expected shortfall model and also attract an NMRF capital add-on. Type 2 NMRFs, which fail both quantitative and qualitative requirements, would continue to be removed from the model and capitalised under the NMRF framework. The PRA also proposes operational simplifications, including aligning the NMRF stress period with the ES stress period, reducing Type 2 NMRF calculation frequency to monthly, and removing the idiosyncratic/non-idiosyncratic distinction.

The third capital-calculation strand is ASA/IMA aggregation. CP9/26 proposes to recognise diversification between ASA and IMA portfolios through a marginal ASA adjustment and to replace the existing partial caps on IMA capital with a permission-based cap at the full ASA level. The fourth operational strand is IMA treatment of collective investment undertakings: CP9/26 proposes a 90% de minimis look-through threshold for IMA inclusion and extends the ASA treatment of index-tracking funds to IMA. The fifth strand covers smaller IMA clarifications, including reduced-risk-factor coverage tests, GIRR internal hedge desk treatment and listed closed-ended investment funds. The final strand is reporting and disclosure, but the source needs to be separated. PS1/26 already implemented the CP17/25 interim-template approach: existing IMA reporting and disclosure templates are retained to 1 January 2028, FRTB-IMA reporting and disclosure templates are delayed to 1 January 2028, and cross-referring templates are aligned during the interim year. CP9/26 then proposes further consequential reporting and disclosure amendments to reflect the new IMA proposals, including draft changes to OF 24.02 / IMA1B for the Type 1 and Type 2 NMRF split and related NMRF stress-period fields. For a reporting officer, the key point is to treat the PS1/26 interim-template decision and the CP9/26 proposed IMA1B amendments as two separate change-log items.

The implementation date split is the thing to get right first

If you take one operational fact away, take this one. The PRA confirmed in PS1/26, its final Basel 3.1 rules published on 20 January 2026, that the Basel 3.1 policy, rules, supervisory statements and statements of policy take effect on 1 January 2027. The internal model approach for market risk is the exception. In CP17/25 the PRA proposed to delay the FRTB-IMA by one year to 1 January 2028, and that delay is now baked into the timetable. So the UK runs a one-year gap where most of the framework is live but the market risk internal model approach is not yet on its final footing.

Where teams get this wrong is by treating “Basel 3.1 goes live on 1 January 2027” as a single switch. It is not. Credit risk, operational risk, the output floor, the standardised market risk approaches and the bulk of the reporting and disclosure architecture move on 1 January 2027. The FRTB-IMA, its dedicated reporting templates and its dedicated disclosure templates move on 1 January 2028. A capital plan that flips the entire trading book to the new world on day one of 2027 has mis-modelled the market risk leg by a year.

The reason for the staggered date matters for how you read the rest of the framework. The PRA tied the IMA delay to international coordination. The internal model approach is predominantly used by major trading firms, including international groups engaged in cross-border trading activity, so the PRA wanted the UK’s IMA rules to settle in step with the larger jurisdictions rather than ahead of them. That is also why CP9/26 frames its adjustments partly around increased clarity on approaches in other major jurisdictions. The UK is deliberately not first-mover on the model approach.

What happens during the interim year, 1 January to 31 December 2027

The interim year is where most reporting teams will spend their effort, because it is a hybrid state. From 1 January 2027, firms retain their existing IMA model permissions for the interim year. A desk that holds an internal model permission today does not lose it at the start of 2027 simply because Basel 3.1 went live. It carries that permission forward only until the end of the interim period. Existing IMA permissions cease to apply automatically at the end of that period, and firms that want to use IMA under the Basel 3.1 FRTB-IMA framework from 1 January 2028 need a new PRA IMA permission.

Positions that are not in scope of existing IMA permissions are the ones that move. During the interim period, those positions move to the new standardised approaches, either the advanced standardised approach (ASA) or the simplified standardised approach (SSA), under the requirements set out in PS17/23 and PS9/24. So a trading book in 2027 can be a mix: legacy IMA permissions running on the existing model framework and existing IMA templates, sitting next to in-scope-of-standardised positions running on the new ASA or SSA. Two capital engines, one book, one reporting cycle.

The reporting consequence is specific. PS1/26 already implemented the interim-template approach: existing IMA reporting and disclosure templates are kept alive through the interim year, precisely so that legacy IMA permissions have somewhere to land in the returns while the new FRTB-IMA templates wait until 2028. This is the opposite of what an unwary reader might assume. The instinct is that 1 January 2027 brings a clean new template set for everything. For the model approach, it does not. The existing templates are deliberately retained for another year.

A common misreading here is to assume firms are forced onto the new standardised approaches for their whole book in 2027. They are not. During the interim period, firms may retain existing IMA models for positions within existing permissions, and the PRA has confirmed that firms also retain the option to implement the ASA for their entire trading book from 1 January 2027. However, moving completely or partly out of existing IMA permissions is not a free position-by-position switch: the existing revocation process applies for a complete move to ASA, and a partial revocation is treated as a change to the permission.

How the internal model approach is built, in practitioner terms

To map the changes, it helps to be precise about what the FRTB-IMA actually is, because the adjustments target specific components rather than the whole edifice. The internal model approach for market risk replaces the older value-at-risk measure with an expected shortfall measure, calibrated to a period of stress and calculated at a 97.5% confidence level, with liquidity horizons that vary by risk factor. It also replaces the old incremental risk charge with a default risk charge, which captures the loss from an issuer or obligor defaulting rather than the loss from spread moves.

Three gates control whether a desk can use the model approach at all. The first is the risk factor eligibility test, which asks whether each risk factor has enough real, observable price data to be modelled. Risk factors that fail are pulled out of the expected shortfall calculation and capitalised separately as non-modellable risk factors, which is where the limited trading data adjustments in CP9/26 land. The second gate is desk-level backtesting, which reconciles the model’s forecast losses against what actually happened. The third gate is the profit and loss attribution test, which asks whether the desk’s risk-management model explains the desk’s actual daily profit and loss well enough, and that is the test whose monitoring period CP9/26 proposes to extend from one year to three.

The practical point for a reporting team is that the IMA is a permission you can lose at the desk level, not a status you hold at the firm level. A single desk can fall out of the model approach while the rest of the book stays in. When a desk fails its tests, its positions get capitalised under the standardised approach instead, and the capital number and the reporting treatment both change for that desk alone. The whole architecture of CP9/26 is about how harsh and how fast that desk-level fallback should be.

The standardised approach is already settled, and it frames the IMA

The internal model approach does not stand on its own. It sits above the standardised market risk approaches, and the standardised number is the floor, comparator and fallback where a desk cannot use IMA. The PRA consulted on standardised-side market risk adjustments in CP17/25, published on 15 July 2025, and finalised those points in PS1/26. The final package includes the one-year FRTB-IMA delay, operational simplifications for collective investment undertakings in the trading book boundary and ASA, a permissions regime for proportionate capitalisation of residual risks in the ASA, and reporting and disclosure updates aligned with those changes.

The residual risk add-on permission is worth flagging because it is a genuine relief mechanism, not just a calculation. The residual risk add-on captures complex and exotic risks that the main standardised charges do not, computed simply as gross notional times a fixed percentage. The PRA accepted that for a small number of instruments or business lines the add-on could be disproportionate to the actual risk, and proposed a permissions regime: where a firm can demonstrate the add-on is disproportionate, it may apply to the PRA to capitalise those risks using an alternative methodology. That is an application process a reporting and capital team needs to know exists, because it changes the standardised number that a fallen-out desk would otherwise carry.

For collective investment undertakings under the ASA, PS1/26 finalised two different de minimis thresholds. Firms must allocate CIU positions to the trading book where at least 90% of the CIU’s underlying holdings by value would be allocated to the trading book, with the residual position capitalised under the ASA fall-back approach. For market risk look-through approach eligibility, the PRA lowered the final threshold to 50% of the CIU’s holdings by value, not 90%. If your book holds fund positions, the boundary treatment and look-through mechanics decide whether those positions enter the market risk framework and which ASA treatment applies. Our guide to CRR3 FRTB market risk reporting walks through how the EU handles equivalent boundary and reporting questions as a contrast for cross-regime books.

What changes in the market risk reporting returns and the Bank of England taxonomy

The reporting plumbing moves on the main 1 January 2027 date for everything except the model approach. The technical implementation of PS1/26 is delivered through version 4.0.0 of the Bank of England Banking Taxonomy, which is intended to take effect from 1 January 2027. That is the version a reporting team builds its 2027 submissions against. The market risk disclosure instructions that accompany PS1/26 are also effective from 1 January 2027.

The model approach templates are the carve-out. The FRTB-IMA reporting and disclosure templates are deliberately held back to 1 January 2028, the existing IMA templates are retained through the interim year, and the templates that cross-refer to the FRTB-IMA are updated so they do not point at a template that is not yet live. A reporting officer who maps the taxonomy change for 2027 and assumes the IMA templates come with it will find a mismatch between the rule timetable and the template timetable. The mismatch is intentional, and reading it as an error wastes a change cycle.

For Pillar 3, the PRA’s disclosure requirements set different frequencies for different firm categories. Firms should check the disclosure frequency that applies to their own categorisation under PS1/26 before finalising their Pillar 3 build. For cross-regime context, our Pillar 3 disclosure explainer shows how disclosure frequency tiers and template logic are handled in an EU/Luxembourg setting, but UK firms should build to the PRA disclosure requirements in PS1/26.

Who is actually in scope, and who can mostly ignore this

The internal model approach is concentrated. It is predominantly used by major trading firms, including the large international groups that run cross-border trading desks. A smaller deposit-taker with a modest trading book and no existing or planned IMA permission is more likely to treat CP9/26 as a horizon-scanning item than as a primary implementation workstream. The standardised adjustments finalised in PS1/26 and the main 1 January 2027 implementation date are what bite for the standardised population.

Where this gets misjudged is at firms that hold IMA permission on a narrow set of desks and standardised treatment everywhere else. Those firms sometimes treat the IMA changes as someone else’s problem because the model book is small. The opposite logic applies. The smaller and more concentrated the model book, the more a single desk’s failure of the extended PLA test moves the firm’s reported capital, because there is no diversified portfolio of model desks to absorb it. A two-desk model book is more exposed to a test change than a twenty-desk model book.

The competition and proportionality framing also matters for how a smaller firm reads the whole package. The PRA expects that, once Basel 3.1 is fully implemented, average Pillar 2A requirements for major UK lenders will fall from around 2.5% to around 2% of risk-weighted assets, reflecting the recalibration of Pillar 1 risk measurement. That is a system-level expectation about the major banks, not a promise to any individual firm, and the actual outcome depends on how banks respond to the changed risk measurement. It is useful context for a capital plan, but it is not a number to put into a firm-specific projection.

The mistakes reporting teams are most likely to make

Three errors recur when teams plan for this. The first is collapsing the two implementation dates into one. The standardised approaches and the bulk of the reporting architecture go live on 1 January 2027; the FRTB-IMA and its dedicated templates go live on 1 January 2028. Build the interim year as a distinct state, not as a soft launch of the end state.

The second is treating CP9/26 as settled rules. It is a consultation open to 18 September 2026. The three-year PLA monitoring period, the RFET and NMRF adjustments, the ASA/IMA aggregation changes and the consequential IMA1B template changes are proposals. A firm should plan around them, but it should not hard-code them into a system before the PRA confirms the final position, and it should track the response to the consultation rather than assume the draft survives intact.

The third is importing EU FRTB assumptions wholesale. The UK and the EU both implement the Fundamental Review of the Trading Book, but the instruments, the dates and the article references diverge. The EU runs its market risk framework through CRR3 and a separate set of Commission delegated acts and implementing standards; the PRA runs its version through PS17/23, PS9/24, PS1/26 and the CP9/26 adjustments. A cross-border group that lets its EU teams set the UK assumptions, or vice versa, will mis-state at least the dates and quite possibly the capital treatment. Our coverage of supervisory findings on internal approaches shows how closely regulators now scrutinise model permissions on both sides of the Channel, which is exactly why the divergence in the rules needs to be tracked deliberately.

Frequently Asked Questions

When does the PRA’s Basel 3.1 market risk internal model approach actually take effect?

The internal model approach for market risk takes effect on 1 January 2028. The rest of Basel 3.1, including the standardised market risk approaches and most reporting and disclosure requirements, takes effect on 1 January 2027. The PRA proposed the one-year IMA delay in CP17/25 and confirmed the wider 1 January 2027 date in PS1/26, published on 20 January 2026.

What is the single biggest change in CP9/26?

The most significant proposed change is to the profit and loss attribution test. CP9/26 proposes to extend the test’s monitoring period from one year to three years. The PLA test is the gate that decides whether a trading desk can keep its internal model permission, so lengthening the window changes how poor model performance translates into a desk falling back to the standardised approach.

Do firms lose their existing model permissions on 1 January 2027?

No. From 1 January 2027, firms retain their existing IMA model permissions through the interim year. Positions that are not in scope of those existing permissions move to the new standardised approaches, either the advanced standardised approach or the simplified standardised approach, under PS17/23 and PS9/24. Firms can also choose to implement the ASA for their entire trading book during the interim period, subject to the existing revocation process.

What happens to the market risk reporting templates during 2027?

PS1/26 finalised the interim-template approach: existing IMA reporting and disclosure templates are retained for the period to 1 January 2028, and dedicated FRTB-IMA templates are delayed until 1 January 2028. Templates that cross-refer to FRTB-IMA are updated so they stay consistent during the interim year. Separately, CP9/26 proposes consequential amendments to reporting and disclosure, including IMA1B changes for the Type 1/Type 2 NMRF split. Standardised reporting moves with the main 1 January 2027 date through version 4.0.0 of the Bank of England Banking Taxonomy.

How is the limited trading data issue being addressed?

CP9/26 proposes RFET and NMRF adjustments for risk factors with limited trading data. For risk factors with a liquidity horizon greater than 20 days, the minimum verifiable-price count would fall from 24 to 16; for new issuances the verifiable-price requirement would be pro-rated during the first 12 months. CP9/26 also proposes Type 1 NMRFs, which fail the quantitative RFET but pass qualitative standards and therefore remain in the ES model with an NMRF add-on. Type 2 NMRFs, which fail both tests, would continue to be removed from the model and capitalised under the NMRF framework.

Does this UK consultation change anything for an EU trading book?

No. CP9/26 is a PRA consultation and applies to the UK framework only. The EU implements the Fundamental Review of the Trading Book through CRR3 and its own delegated and implementing measures, with its own dates and article references. A group running a book across both regimes has to track the two timetables separately and should not assume the UK and EU positions on the internal model approach align.

When does the CP9/26 consultation close?

The consultation closes on 18 September 2026. Until the PRA publishes its final position, the proposed adjustments, including the three-year PLA monitoring period and the RFET and NMRF changes, remain draft and may change.

Where do residual risks fit if a desk falls back to the standardised approach?

A desk on the standardised approach capitalises complex and exotic risks through the residual risk add-on, calculated as gross notional times a fixed percentage. The PRA introduced a permissions regime in PS1/26 so that, where a firm can demonstrate the add-on is disproportionate, it may apply to use an alternative methodology. That application route is part of the standardised treatment a fallen-out desk would carry.

Related Articles

Key Takeaways

  • The PRA published CP9/26 on 19 June 2026, proposing adjustments to the Basel 3.1 market risk internal model approach. The consultation closes on 18 September 2026 and the adjustments are proposals, not final rules.
  • The headline proposed change extends the profit and loss attribution test monitoring period from one year to three years; during this period the PLAT would not be binding, though desks remain subject to supervisory review.
  • CP9/26 also proposes RFET and NMRF adjustments, ASA/IMA aggregation changes and IMA CIU simplifications, including reducing the verifiable-price count from 24 to 16 for risk factors with a liquidity horizon greater than 20 days, introducing a Type 1/Type 2 NMRF split, recognising diversification between ASA and IMA portfolios through a marginal ASA adjustment, and introducing a 90% de minimis look-through threshold for IMA inclusion of CIU positions.
  • Implementation is split: most of Basel 3.1 takes effect on 1 January 2027, but the FRTB-IMA is delayed to 1 January 2028. Build the interim year as a distinct hybrid state.
  • From 1 January 2027, firms keep existing IMA permissions for the interim year only; those permissions cease automatically at the end of the interim period, and firms wanting FRTB-IMA treatment from 2028 need a new PRA IMA permission.
  • PS1/26 already finalised the interim-template approach: existing IMA templates are retained to 1 January 2028. CP9/26 separately proposes consequential IMA1B amendments for the Type 1/Type 2 NMRF split.
  • The standardised market risk side is already settled in PS1/26, including CIU thresholds of 90% for trading book allocation and 50% for ASA look-through eligibility, and a residual risk add-on permissions regime.
  • This is a UK framework. Do not import EU CRR3 FRTB dates, article references or capital treatment onto a UK trading book.

Sources and References

  • PRA, CP9/26 – Basel 3.1: Adjustments to the internal model approach (IMA) for market risk (19 June 2026): bankofengland.co.uk
  • PRA, PS1/26 – Implementation of Basel 3.1: Final rules (20 January 2026): bankofengland.co.uk
  • PRA, CP17/25 – Basel 3.1: Adjustments to the market risk framework (15 July 2025): bankofengland.co.uk
  • PRA, PS17/23 – Implementation of the Basel 3.1 standards near-final policy statement part 1 (December 2023): bankofengland.co.uk
  • PRA, PS9/24 – Implementation of the Basel 3.1 standards near-final policy statement part 2 (September 2024): bankofengland.co.uk
  • PRA, Implementation of the Basel 3.1 standards – Chapter 6, Market risk: bankofengland.co.uk
  • PRA, Annex – Instructions for the disclosure of market risk templates (PS1/26): bankofengland.co.uk
  • Bank of England, Financial Stability in Focus: The FPC’s assessment of bank capital requirements (4 December 2025): bankofengland.co.uk

What to put on the change board before 2027

Treat the next eighteen months as two distinct mapping jobs, not one. The first job is the 1 January 2027 cutover for the standardised approaches and the wider reporting and disclosure architecture, built against version 4.0.0 of the Bank of England Banking Taxonomy and the PS1/26 final rules. The second job is the 1 January 2028 cutover for the FRTB-IMA, its dedicated templates and whatever CP9/26 finally lands as after the consultation closes on 18 September 2026. The interim year between them is a hybrid you have to report through, with legacy IMA permissions on retained templates sitting beside standardised positions on the new ones. Get the two dates, the two template sets, the revocation process for any intended switch, and the consultation status onto the change board now, and the model book will not surprise you when the calendar turns.

Disclaimer: The information on RegReportingDesk.com is for educational and informational purposes only. It does not constitute legal, regulatory, tax, or compliance advice. Always consult your compliance officer, legal counsel, or the relevant supervisory authority for guidance specific to your institution.

Similar Posts