We are excited to share our new report Asset Correlations in Credit Risk: An Empirical Study with GCD Data. Drawing on our large PD & rating dataset of corporate obligors in Europe and North America (2008–2023), the study estimates default-based asset correlations—offering a practical benchmark for economic capital, stress testing and IFRS 9 modeling. The results indicate that correlations do not simply decline with higher PDs: after an initial drop at low PDs, a minimum is reached and correlations rise again at higher PDs. We compare our computed values with findings from the literature—particularly the European Banking Authority’s (EBA) “The Calibration of the IRB Supervisory Formula”—and find broadly similar patterns in the PD–correlation relationship.
This work benefited from close industry collaboration with Nordea, whose team engaged with us during the study and provided practitioner perspective on how banks apply correlation benchmarks.
“Nordea pursued this collaboration with GCD because of the value of their independent analysis and dataset. The levels of their new asset-correlation benchmark have the potential to impact how banks approach default correlations and economic capital,” said Nicolás Gutierrez, Lead Quantitative Analyst, Nordea.
