Unsecured IRB Models: PiT Plus Buffer – but not as we know it!

As the latest round of residential secured IRB model remediation draws to an end, and we see the majority of lenders submit their revised approaches to the PRA, focus within retail banking has moved towards the unsecured – Fixed Term, Revolving (Current Accounts, Credit Cards), Motor Finance and SME lending – IRB model development suites. Given the resources consumed on the secured developments, firms will no doubt be looking for synergies that can be applied to the unsecured model builds, and avoid a recurrence of the extended secured rework position. There are however additional challenges to consider when it comes to unsecured portfolios.

Focusing on PD models, the development of a long run average (LRA) approach on an unsecured portfolio, and the requirement to align to an economic cycle (as with the secured developments) may prove to be more complicated to identify than the typical Mortgage Portfolio. Significant changes to consumer attitudes and product offerings over time, such as lengthy promotional deals, rewarding cash back offers and the rise and changing use of comparison sites, makes the argument of the identification of a comparable ‘representative’ cycle over time more challenging.

Therefore, given the challenge of accurately modelling a compliant LRA average model on historical data, one that will react to current portfolio and future portfolio trends, this does however leave the option of PiT + Buffer on the table. CP 16/22 outlines the requirement for firms to use a discrete rating scale solution and in particular outlaws the use of a variable scalar approach, therefore, approaches used today that use a PiT + Buffer methodology which deliver a continuous rating system is not appropriate. However, considering this approach within the framework of a discrete rating solution still holds merit. This approach together with the creation of suitable risk grades is where synergies with a firms Hybrid methodology can be exploited.

From a modelling perspective, considering alignment to the portfolio mix and quality, the PiT approach seems to give a simpler development. However, for some firms, wider considerations should be given when considering the two approaches, LRA vs PiT, the European interpretation of the Basel 3.1 framework guides against the use of a PiT + Buffer approach, so firms with dual reporting requirements are likely to consider two separate modelling approaches arduous and exorbitant.

A high-level view of Pros and Cons of both approaches is outlined below:

Pit+ Buffer

Pros:

• Aligns with industry standard approach.
• Potentially less historical data required for model development.
• Easier to incorporate Evolving Landscape of Data (i.e. new data types)
• Easier to Stress as direct relationship with Actual Default Rates.
• Aligns with Consumer Attitude to Unsecured Credit evolution.
• Easier to monitor.

Cons:

• Requires regular realignment to Actual Default Rate.
• More cyclical capital measure – potential to be holding more capital in downturns and Pillar 2.
• Potential for more volatile capital movements.

Long Run

Pros:

• Aligns recent Mortgage Hybrid PD developments.
• Requires less ongoing model maintenance.
• Less pro-cyclical-hold less capital in downturns.
• Prudent approach against EBA guidelines

Cons:

• Approach not adopted within UK industry for unsecured products and therefore no methodology approved by the PRA.
• Wide range of historical data required to accurately code scorecard/RG through time.
• No long-term standard view of an economic cycle for unsecured – paucity of data Harder to monitor.

On release of CP 16/22, 4most spoke to some organisations that highlighted there was some initial confusion within the industry that a PiT plus Buffer approach was not deemed to be compliant. This was not something that we necessarily felt to be the case, as highlighted above, and has since been superseded by comments made to firms that the approach would be deemed acceptable – if approached in the right way. This does not ignore the fact that one of the key concerns of this approach is that it will still lead to procyclical capital requirements. However, the proportion of RWAs that arise from unsecured lending is much lower than secured and therefore there it is less likely to result in tightening of lending due to capital constraints – something that is a big concern with secured given what was observed during the Global Finance Crisis.

Please get in touch if you would like to learn more about how 4most can support your organisation – info@4-most.co.uk.