Title: Forecasting the yield curve dynamics: A just-identified no-arbitrage FAVAR approach with tractable economic factors
Authors: Alena Audzeyeva - Keele University (United Kingdom) [presenting]
Robin Bladen-Hovell - Keele University (United Kingdom)
Sadeeptha Jayathilaka - Keele University (United Kingdom)
Abstract: A term structure model formulation is proposed that links the term structure dynamics to evolution in underlying economic factors within a data-rich setting. Using a factor-augmented vector autoregression, we examine the influence of dynamic factors that exploit large information sets tracking real activity, inflation, money and market asset pricing and also a latent factor during different economic policy regimes. In contrast to many popular canonical formulations of Gaussian affine term-structure models, our model is just-identified which ensures an improved robustness of the parameter estimation, based on no-arbitrage restrictions, by means of minimum-chi-square estimation. While economically and practically appealing, our term-structure formulation affords more accurate out-of-sample forecasts than a number of previously suggested affine term-structure modeling approaches during both the period of ``great stability'' and a recent period of near zero short-term interest rates in the UK.