Prescribed LGD values instead of internal estimation
- Under the F-IRB approach, firms would continue to model Probability of Default (PD) but would apply PRA-prescribed LGD values rather than estimating them internally.
- This would reduce the data and modelling burden for firms lacking extensive historical default data.
- It is noted that the F-IRB approach is already applicable for non-retail exposures, but no such approach currently exists for retail exposures.
Loan-to-Value (LTV) as primary risk driver
- LTV is identified as the most significant determinant of LGD for residential mortgages.
- The PRA is considering whether additional risk drivers should be incorporated, though it acknowledges that doing so may increase complexity.
Reassessment of probability of possession given default (PPGD)
- The PRA is proposing to withdraw the existing PPGD reference points, which were intended to guide firms with limited possession data.
- Feedback suggests the reference points are overly conservative and have not effectively facilitated IRB adoption.
Collateral valuation standards
- The default valuation method would be value at origination, aligning with the revised Standardised Approach (SA) valuation requirements.
- Revaluation is permitted only if property modifications clearly increase value, ensuring consistency and avoiding cyclical distortions.
Segmentation by exposure type
- The F-IRB framework may differentiate between:
- Buy-to-Let (BTL) and Owner-Occupied (OO) mortgages, given their differing risk profiles.
- Potential further segmentation (e.g., flats vs. houses) is under consideration but may be limited to avoid excessive complexity.
LGD implementation options
- Two possible methods for applying LGD values:
- A formula-based approach, where firms input risk drivers and receive LGD outputs based on segmentation.
- A lookup table, offering prescribed LGD values for specific LTV ranges; simpler but less granular.
Calibration philosophy
- LGD values under F-IRB would be conservatively calibrated to reflect downturn conditions.
- The calibration would be less conservative than the SA, but more conservative than the Advanced IRB (AIRB) approach.
- This ensures prudence while enabling broader IRB access.
PD estimation
In response to the challenges faced by medium-sized firms in developing compliant PD models under the current hybrid modelling framework, the PRA is considering a suite of policy options aimed at reducing complexity while maintaining prudential integrity. Key features discussed in the paper include:
Retention of the 30% cyclicality calibration assumption cap
- Under current policy, the cyclicality calibration assumption cap is set at 30%. The cap is used when firms back-cast historical default rates due to lack of granular data, especially from the early 1990s. The cap assumes that no more than 30% of changes in portfolio default rates are due to grade migration, with the rest reflected in within-grade default rate changes.
- Recent model submissions to the PRA indicate higher-than-anticipated cyclicality, prompting a reassessment of the appropriateness of the 30% calibration cap.
- PRA is hence considering increasing the cap (e.g., to 50%) or removing the cap entirely, allowing firms to estimate cyclicality themselves.
- PRA leans toward retaining the cap of 30% due to the following reasons:
- Provides a conservative buffer against uncertainty in downturn behaviour and mitigates risk of under-calibrated PDs.
- Might increase modelling complexity of smaller firms.
- Firms may struggle to robustly estimate cyclicality with limited data. These changes (if implemented) would have resource implications for both the firm and PRA to review the model changes.
Use of the Global Financial Crisis (GFC) as a proxy for downturn conditions
- The PRA is considering allowing firms to use the 2007–2009 GFC as the ‘bad’ period in their long-run PD calibration.
- Firms would need to demonstrate that the GFC period is representative of a downturn for their specific portfolio.
- Where the GFC data does not fully capture downturn dynamics, adjustments or supervisory overlays may be required to ensure conservatism.
Simplified PD estimation for firms with limited data
- For firms lacking sufficient historical data, the PRA may permit a streamlined approach to computing long-run average default rates.
- This could involve using recent representative data (e.g., five years) and applying PRA-prescribed uplifts to approximate downturn conditions.
Shift in emphasis from cyclicality to discriminatory power
- The PRA is exploring whether medium-sized firms could place less emphasis on managing model cyclicality and instead focus on ensuring strong discriminatory power (the ability of models to differentiate between risk levels).
Permitting Through-the-Cycle (TtC) modelling approaches
- As an alternative to hybrid models, the PRA is considering allowing medium-sized firms to adopt TtC models, which are generally less cyclical and simpler to implement.
- While TtC models may lack granularity in capturing cyclical risk, they could offer a pragmatic solution for firms with limited data and modelling resources.
What’s ahead for medium-sized banks?
- Review and respond to the Consultation
- Carefully assess the proposals in DP1/25.
- Submit feedback to the PRA by the consultation deadline (i.e., 31st October 2025) to help shape the final policy.
- Evaluate IRB readiness
- Consider whether the proposed Foundation IRB and simplified PD modelling options make IRB adoption feasible.
- Assess internal data availability, modelling capabilities, and resource needs.
- Strategic planning
- Explore how IRB adoption could impact capital requirements, pricing strategies, and competitive positioning.
- Consider whether IRB could support growth in specific mortgage segments (e.g., low-LTV or owner-occupied loans).
- Monitor policy developments
- Stay updated as the PRA refines its proposals based on industry feedback.
- Prepare for potential implementation timelines and transitional arrangements.
Get in touch with our quantitative solutions experts
|
National contacts
Xavier Larrieu Partner - Co-Head of Quantitative Solutions UK - London
Andreas Mavrocordatos Director - Banking Risk Consulting - London
This website uses cookies.
Some of these cookies are necessary, while others help us analyse our traffic, serve advertising and deliver customised experiences for you.
For more information on the cookies we use, please refer to our Privacy Policy.
This website cannot function properly without these cookies.
Analytical cookies help us enhance our website by collecting information on its usage.
We use marketing cookies to increase the relevancy of our advertising campaigns.