Double default modelling: a risk-sensitive view of guarantees under IFRS 9

Our new white paper explores double default modelling, a quantitative framework that explicitly recognises that losses on guaranteed exposures generally arise only when both the borrower defaults and the guarantor fails to perform.

While prudential regulation has moved away from recognising double-default benefits for regulatory capital purposes, IFRS 9 has a different objective. The standard requires expected credit losses to be measured on an unbiased and probability-weighted basis, creating scope for institutions to consider whether more risk-sensitive approaches may better reflect the economics of guaranteed exposures.

Why does this matter?

Guarantees are usually treated using simplified approaches

Many methodologies recognise credit protection by applying the guarantor's risk characteristics to the protected portion of an exposure. While operationally straightforward, such approaches do not explicitly model the joint occurrence of borrower default and guarantor non-performance.

Dependence between borrower and guarantor is critical

A guarantee provided by a financially independent third party behaves differently from one provided by a parent company operating in the same economic environment. The strength of this relationship can materially influence the probability of loss and, therefore, ECL. Understanding and modelling dependence becomes a key component of any double default framework.

Correlation can materially affect outcomes

The paper explores how quantitative techniques such as copulas can be used to combine separate probability of default estimates into a joint probability framework. It also examines the practical challenge of estimating default correlation, one of the most important and most difficult assumptions in any double default model.

The impact on ECL can be significant

Using an illustrative corporate loan portfolio, in our whitepaper we demonstrate how different dependence assumptions can materially affect modelled expected credit losses. This analysis highlights the importance of understanding not only the credit quality of the borrower and guarantor individually, but also the relationship between them.

Download the white paper to explore the quantitative foundations of double default modelling, including regulatory considerations, copula methodologies, correlation estimation techniques, and practical IFRS 9 implications for guaranteed exposures.

Download the white paper