The Prudential Regulation Authority (PRA) has confirmed that it will publish a Discussion Paper (DP) this summer, aimed at improving access to Internal
Ratings Based (IRB) model permissions for residential mortgages among mid-tier banks.
The IRB approach, which allows firms to use their own credit risk models to determine capital requirements, remains largely the domain of larger institutions.
These banks typically benefit from lower capital charges and models more closely aligned to their risk management practices.
While the PRA’s introduction of the modular application process was designed to encourage uptake by simplifying and clarifying requirements, mid-tier firms
have still faced a long and uncertain path to approval. The new Discussion Paper is expected to explore proportional adjustments that could make IRB more accessible, while maintaining prudential standards.
This article considers the main obstacles mid-sized firms encounter and where proportionate adjustments might be considered.
The case for change
The IRB framework was originally designed with sophisticated firms in mind. For mid-tier institutions, the cost, complexity and uncertainty involved in
applying for IRB permission can be prohibitive.
Although IFRS 9 programmes and the introduction of Basel 3.1 have spurred many firms to begin investing in modelling infrastructure, several have found
the IRB bar remains too high. Common challenges include gaps in historic data, difficulties embedding models across business processes, and the need for governance structures that rival those of the largest banks.
The PRA’s latest signals suggest a willingness to address these issues without compromising on core standards. That could open the door to greater competition
and broader adoption of internal models in the residential mortgage space.
Where the barriers lie
Across the mid-tier segment, several key pain points frequently emerge:
1. Data availability and representativeness
Many mid-tier firms lack the volume and depth of historical data needed to meet IRB requirements. In particular, datasets with material changes in default
rates over time are difficult to obtain, making it harder to model default probability (PD), downturn loss given default (LGD), and cyclicality effectively.
This often leads to:
-
The use of conservative assumptions
-
Higher Margins of Conservatism
-
Regulatory default rates being assumed (e.g. 100% probability of possession given default)
The consequence is that some firms struggle to demonstrate a capital benefit under IRB, undermining the business case for investment.
2. Model development, governance and validation
Meeting IRB expectations requires:
-
Deep technical expertise across model development, maintenance and monitoring
-
Independent validation frameworks
-
Responsiveness to evolving regulatory expectations
Firms must often balance these demands alongside other competing priorities, such as SS1/23 implementation, IFRS 9 model reviews, and economic scenario
testing. Resource constraints can slow progress or dilute quality.
3. Model use and integration
IRB models must be embedded into key areas of decision-making for at least three years. These include:
This level of integration demands system investment, process redesign and, in some cases, cultural change. Firms often underestimate the level of governance
and MI required to demonstrate genuine use.
4. Governance and documentation
A clear understanding of the models and their role is essential, supported by:
-
Defined roles and responsibilities
-
Effective oversight mechanisms
-
Detailed documentation and control frameworks
These elements must be evident throughout the “use and experience” period, requiring significant resource even before permissions are granted.
What might the PRA adjust?
Any adjustments the PRA makes are likely to be proportionate. Basel 3.1 and IRB are international frameworks, and the UK regulator will be keen to avoid
lowering standards or creating a perception of preferential treatment.
That said, several pragmatic options may be on the table:
-
Data flexibility: Accepting increased use of proxy or pooled data where firms lack a long operating history
-
Staged roll-out: Allowing phased implementation within asset classes could help firms with diverse residential mortgage books
-
Regulatory engagement: Continued dialogue through the model lifecycle, building on the roundtables already held, could help address
red flags early -
Clarity on expectations: Publishing common pitfalls and good practices from thematic reviews could give firms greater confidence
when designing their programmes
These changes would not reduce the level of rigour required but could make the path to compliance more achievable.
Closing thoughts
The upcoming Discussion Paper signals a constructive step by the PRA to support the ambitions of mid-tier mortgage lenders. While challenges remain, even
modest adjustments could materially improve the feasibility of IRB adoption for this group.
Mid-sized firms preparing for this shift will benefit from taking stock now: identifying current blockers, understanding where proportional adjustments
would have the greatest impact, and evaluating their readiness across data, modelling, governance and integration.
