Exposé dans le cadre de la Summer School of the Groupe Consultatif Actuariel Européen: Enterprise Risk Management (ERM) and Solvency II. Petite réflexion à partir des déclarations du rapport technique QIS3, en particulier la déclaration qui mériterait des heures de commentaires,
In view of the insufficiency of currently available data, the setting of these correlation coefficients will necessarily include a certain degree of judgement. This is also true because, when selecting correlation coefficients, allowance should be made for non-linear tail correlation, which is not captured under a “pure” linear correlation approach. To allow for this, the correlations used should be higher than simple analysis of relevant data would indicate.