Assessing IRB Model Fit for Acquired Credit Portfolios

Summary

The document frames a regulatory validation problem: when a bank using the internal ratings based approach acquires another bank’s loan portfolio, how can it show that its existing probability of default model and risk classes are suitable for the new exposures? It asks what analysis and evidence would support integrating the acquired portfolio into the bank’s IRB framework and satisfying supervisory expectations.

The text does not propose a validation method, report results, or cite examples. It is a question seeking best-practice and regulatory guidance, so it leaves key details open, including portfolio type, borrower mix, data availability, and the criteria a regulator would apply. Its value is in identifying model portability and regulatory approval as connected issues; it does not establish that an existing model can be reused without further assessment.

Key ideas

• An acquiring bank must assess whether its IRB risk model is suitable for the acquired exposures.
• The question focuses on probability of default estimates and assignment to internal risk classes.
• The document seeks quantitative validation and regulatory evidence for portfolio integration.
• It provides no proposed analysis, published example, or empirical result.

Tags

# Quantitative and regulatory aspects of portfolio integration in IRB credit portfolios


# Quantitative and regulatory aspects of portfolio integration in IRB credit portfolios












Say bank A buys a credit portfolio "B" (e.g. corporate loans or retail mortgage, ...) from bank B.

Bank A fulfills the the requirements of CRR (capital requirement regulation) for its existing portfolio and applies the IRB (internal ratings based) approach. Thus (among other things) it has a risk model that estimates the probability of default and that can assign customers to one of $N$ risk classes.

How can A prove to the regulator (in this case ECB) that it can integrate the new portfolio "B" into its existing one and treat it in IRB? What analysis should be done to show that the risk model in bank A fits for "B" too?

This is not a pure brainstorming question but rather a best pratice/regulatory one. I would be happy about any references where this was done (and some report was published) or any opinion how this could be done taking into account the regulatory framework and quantitative analyses.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.