A dynamic model of expected bond returns: A functional gradient descent approach

We propose a multivariate methodology based on Functional Gradient Descent to estimate and forecast time-varying expected bond returns. Backtesting our procedure on US monthly data, we collect empirical evidence of its strong forecasting potential in terms of the accuracy of the predictions, also in comparison to the classical univariate methodology used in the literature.

• Type: Journal
• Year: 2006
• Where: CSDA
• Authors: Francesco Audrino, Giovanni Barone-Adesi
• Keywords: Classical Univariate Methodology, CSDA 2006, Functional Gradient Descent, Strong Forecasting Potential