Does anybody has a clear view concerning the distribution assumption CreditMetrics employs for calculating credit VaR. Do they go through the calculation applying the normality assumption or do they use the actual, skewed distribution of credit returns?