This thesis focuses on modelling and forecasting the yield curves and economic factors using vine, factor copula models and factor models. To remove serial dependencies, time series models were employed. The autoregressive moving average model was initially considered, which was then extended by the GARCH model with skew Student’s t innovation. The standardized residuals were then formed and transformed into copula data using the probability integral transform. The dependencies were then modelled using the vine copula models, the regular, canonical and drawable vine copula, the one- and two-factor copula model, the Student’s t copula and finally the one- and two-factor Gaussian model. The resulting models were compared to determine the best one, which was then used to simulate the forecasts. The modelling and forecasting were performed on two case studies. The first case study considered German yield curves with maturities of 1, 5, 10, 15 and 20 years. The second case study considered the same yield curves as in the first case and in addition the German inflation rate. The vine copula models produced the most accurate results when modelling and forecasting yield curves and the inflation rate.
«
This thesis focuses on modelling and forecasting the yield curves and economic factors using vine, factor copula models and factor models. To remove serial dependencies, time series models were employed. The autoregressive moving average model was initially considered, which was then extended by the GARCH model with skew Student’s t innovation. The standardized residuals were then formed and transformed into copula data using the probability integral transform. The dependencies were then modelle...
»