Moment Estimators for Autocorrelated Time Series and Their Application to Default Correlations

In credit risk modelling, method-of-moment approaches are popular to estimate latent asset return correlations within and between rating buckets. However, the autocorrelation that is often present in time series of default rates leads to systematically too low estimations. We propose a new estimator that adjusts for the problems of this autocorrelation and the shortness of the time series, thus eliminating a significant portion of the bias observed in classical estimators. The adjustment is based on convergence and approximation results for general autocorrelated time series, and is easily implementable and nonparametric