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, autocorrelation that is often present in default rate time series leads to systematically too low estimations. Adjusting for autocorrelation and shortness of time series, we propose a new estimator. The adjustment is based on convergence and approximation results for general autocorrelated time series. Our adjusted estimator is easily implementable and nonparametric. The adjustment removes a big part of the bias observed in classical estimators due to autocorrelation in default rate time series.