Modeling and Forecasting the term structure of interest rates on listed Federal securities

This research paper investigates the economic rationale and the econometric fit of three modeling, estimation, forecasting methods of the term structure of interest rates, which are devised to address as many shortcomings in the literature. First, to evaluate the extent to which the contraction-expansion regimes impact the yield curve, we design a Dynamic Nelson-Siegel model incorporating a hidden Markov-Chain motion. This framework significantly contributes to predicting the prospective dynamics of the medium-end of the yield curve. Second, we adapt the Dynamic Nelson-Siegel Macro-Finance model to a controlled Dynamic Linear Model to recreate a data-rich environment where the yield curve is controlled by the policy-rate and is influenced by production, inflation, volatility. The accuracy of the size and sign of the forecasts, particularly of the short-end of the yield curve, are enhanced by a model specification where economic factors are inputs for the Nelson-Siegel factors and where synergies between the Nelson-Siegel factors are extracted. Third, a Vector Autoregression model is brought to the Nelson-Siegel factors and to the economic factors to learn about and to learn from their mutual relationships. Whether or not the latter vary depending on the business-cycle alternation is studied through a Markov-Switching extension. In the single-regime economy, the policy-rate is found to be the most important driver of the slope and the level of the yield curve. In the two-regime economy, the slope is a powerful predictor of the recession to come and the level is the most sensitive factor to the alternation of the business-cycle phases.