Stima e previsione della curva dei rendimenti italiana con i GARCH multivariati

Abstract

In this paper we propose an econometric model to predict the daily returns of the italian government bonds with 3, 5 and 10 years maturity. The movement of daily returns is described by a vector autoregression model with predetermined variables (VARX model). The distinctive characteristics of this approach is that the conditional second moments of the daily returns are modeled as multivariate GARCH processes. The dynamic specification of the conditional second moments allows to calculate confidence bounds for point forecasts and to forecast the future returns volatility.