IRB Asset and Default Correlation: Rationale for the Macroprudential Mark-Ups to the IRB Risk-Weights

There is a vast amount of literature criticizing the Basel Committee approach to the credit risk regulation, more specifically, the Internal Ratings-Based (IRB), as an excessively conservative one. However, the novelty of the current paper is that we identify when the IRB approach is too lax, i.e., we are able to present cases with the material credit risk underestimation. We show that the portfolio default rate (DR) depends on two parameters: probability of default (PD) and default correlation. Inversely, we offer a reproducible approach on how to derive the default correlation from historical data. Then it also depends on two parameters: PD and the historical DR variance. However, the IRB approach previewed only PD (and asset class) as the correlation determinants, neglecting the second contributor (DR variance). Hence, we demonstrate that when the actual DR variance exceeds the mean DR value, IRB may result in the credit risk underestimation. The almost two-fold underestimation is found for the credit cards (qualified revolving retail loans) asset class. The paper offers a practical solution how to adjust the revealed credit risk underestimation. The macroprudential add-ons to IRB risk-weights might be a workable solution format. Opinions expressed in the paper are solely those of the author and may not necessarily reflect the official position of the affiliated institution. Bank of Russia neither assumes any responsibility for the publication.