Modelling the Yield Curve in a Low Interest Rate Environment with a Markov-switching Dynamic Nelson-Siegel Model
This thesis introduces Markov-switching in the dynamic Nelson-Siegel model to fit and forecast the yield curve in a low interest rate environment. I allow for regime-switching in the mean of the Nelson-Siegel slope factor to distinguish between a period where the yield curve is at and one where the curve is steep. I extend the regime-switching model by linking the yield curve to the macro-economy in three ways: by adding the macro-economic indicators as state variables; by letting the transition probabilities depend on the indicators; and a combination of both. The regime-switching model significantly improves the in-sample fit and out-of-sample forecasting performance relative to the regular dynamic Nelson-Siegel model. The addition of macro-economic indicators marginally increases the model fit. Out of all models, a regime-switching model that allows the transition probabilities to depend on the macro-economic indicators produces the most superior forecasts, especially at the short end of the yield curve.
|Keywords||Nelson-Siegel yield curve model, State-space representation, Markov-switching, Zero lower bound, Kalman filter, Kim filter, Macro-finance|
|Thesis Advisor||Wel, M. van der|
Noteboom, M.S.E. (2019, December 12). Modelling the Yield Curve in a Low Interest Rate Environment with a Markov-switching Dynamic Nelson-Siegel Model. Econometrie. Retrieved from http://hdl.handle.net/2105/50596