Forecasting of macro aggregates using yield curve information

This paper obtains the forecasts of Colombian macroeconomic variables and the yield curve by jointly modeling their dynamics. For this purpose, I use unrestricted Bayesian Vector Auto Regressive (VAR) models and the no-arbitrage state-space representation developed by Ang and Piazzesi [2003]. Both the Bayesian VAR and the no-arbitrage representations are used to estimate closed economy, small open economy ancl macro-latent factor models. The parameters of the models are estimated with Bayesian techniques for different horizons using the predictive likelihood function. Monthly data between 2006-2012 of the inflation, the overnight-interbank interest rate, an economic activity indicator, the 10-year treasury rate and the 5-year CDS was used The main finding is that the out-of-sample forecasts of the interbank overnight interest rate and the inflation consistently improve when the yield curve is incorporated. Moreover, the models thnt irnpose the no-arbitrage restriction consistently out-perform the unrestrict.ed VARs. On the Other hant, the model wit,h the best. performance in terms of both the RMSE and the standard deviation of the forecasts incorporates closed-economy variables and the short-term yield. Adding longer-term yields and small open economy variables does not appear to improve further the forecasts

• Autores: Rassa Robayo, Juan Sebastián
• Fecha de publicación: 2015
• Institución: Universidad de los Andes
• Repositorio: Séneca: repositorio Uniandes
• Idioma: eng
• OAI Identifier: oai:repositorio.uniandes.edu.co:1992/13349
• Acceso en línea: http://hdl.handle.net/1992/13349
• Palabra clave: Finanzas - Modelos econométricos; Macroeconomía - Modelos econométricos; Pronóstico de la economía - Investigaciones - Colombia; Economía
• Rights: openAccess
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/