Working Paper: CEPR ID: DP4636
Authors: M Hashem Pesaran; Davide Pettenuzzo; Allan G Timmermann
Abstract: This Paper provides a novel approach to forecasting time series subject to discrete structural breaks. We propose a Bayesian estimation and prediction procedure that allows for the possibility of new breaks over the forecast horizon, taking account of the size and duration of past breaks (if any) by means of a hierarchical hidden Markov chain model. Predictions are formed by integrating over the hyper parameters from the meta distributions that characterize the stochastic break point process. In an application to US Treasury bill rates, we find that the method leads to better out-of-sample forecasts than alternative methods that ignore breaks, particularly at long horizons.
Keywords: Bayesian Model Averaging; Forecasting; Hierarchical Hidden Markov Chain Model; Structural Breaks
JEL Codes: C11; C15; C53
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| breaks complicate forecasting (F37) | predictions can be biased (C52) |
| Bayesian hierarchical hidden Markov chain approach (C11) | better out-of-sample forecasts (C53) |
| historical break data (Y10) | refine future predictions (C53) |
| Bayesian approach yields more accurate forecasts (C53) | traditional methods (C90) |