The PRA is refining market-risk internal models rather than reopening FRTB
CP9/26 is a targeted consultation on the internal model approach for market risk under Basel 3.1. The important point is that the PRA is not revisiting the whole market-risk package. It is consulting on a narrow set of adjustments to the internal model framework after reviewing firm feedback, Basel Committee impact data, and the practical challenges firms have faced in preparing for approval.
That matters because the PRA clearly still wants a credible internal model regime in the UK. In the consultation it notes that fewer firms than expected had been planning to adopt internal models and signals that some parts of the current framework may be more burdensome or less proportionate than intended. The consultation is therefore about keeping prudential standards robust while making the modelled approach more usable for internationally active trading firms.
What the consultation actually proposes
The package is focused on seven changes, but four do most of the work. First, the PRA proposes to extend the monitoring period for the profit and loss attribution test from one year to three years, with the test remaining non-binding during that period. Second, it proposes targeted adjustments to the risk factor eligibility test, including reducing the minimum verifiable price threshold from 24 to 16 for risk factors with liquidity horizons above 20 days and allowing a proportionate approach for new issuances.
Third, the PRA proposes a new two-tier treatment for non-modellable risk factors. Type 1 NMRFs would stay within expected shortfall where they pass qualitative standards but fail the quantitative price test, while still attracting an add-on. Type 2 NMRFs would remain outside the model. Fourth, the PRA wants to make gradual IMA adoption easier by recognising diversification between internal-model and advanced standardised portfolios and by replacing the current partial caps with a permission-based cap at the full ASA level.
There are also practical operational changes around collective investment undertakings, internal hedges, alternative tests for reduced sets of risk factors, and reporting and disclosure. Taken together, the direction is clear: less artificial friction, better operational proportionality, and more room for firms to adopt IMA incrementally.
Why this matters for trading firms now
The headline implementation date for IMA remains 1 January 2028, so the consultation does not change the broad timetable. What it does change is the likely shape of firms' mobilisation plans between now and implementation. Trading firms with existing model approvals, or those considering whether to apply, now have a live consultation suggesting that some of the hardest edge issues in the framework may soften in measured ways.
Three points stand out. The first is that the PRA is willing to revisit parts of the framework where evidence suggests double-counting, unnecessary operational burden or barriers to gradual adoption. The second is that it still wants real discipline around model performance, data quality and supervisory approval. The third is that this is explicitly linked to international consistency and competitiveness, particularly for UK firms operating across borders. So this is not just a technical recalibration exercise. It is also about whether the UK ends up with a market-risk internal model regime that firms will actually use.
What firms should do before the consultation closes
The response deadline is Friday 18 September 2026. Firms that care about internal model economics, approval strategy or future trading-book capital should use that window carefully. The most useful exercise now is not simply to agree or disagree with the proposals in principle, but to map where the current framework is driving disproportionate cost, where the proposed changes would materially alter adoption decisions, and where operational simplifications would still leave hidden implementation pain.
In practice, that means market risk, front office, regulatory policy and finance teams should review the PLAT assumptions, RFET data availability, NMRF treatment, and the interaction between standardised and modelled desks as one joined-up programme. The firms that respond best will be the ones that can explain, with evidence, where the framework improves risk sensitivity and where it still risks discouraging adoption without prudential benefit.