Forecasting the Greek yield curve

Abstract

Forecasting yield curve has been evolving into a major issue over the last years as theeconomic growth in advanced countries is based on expansionary monetary policy andespecially after the pandemic. Forecasting the bond yields is extremely important forcentral banks considering inflation as one of the first objectives along with economicgrowth. In this thesis is used the dynamic Nelson-Siegel model as introduced byDiebold and Li (2007). There are various Nelson-Siegel models demonstrated fordeeper understanding of concept around these models. The term structure used is theGreek bond yields which had soared after the eruption of financial crisis and remainedat these levels for several years. Also, In-sample forecast results are presented to checkout the efficiency of dynamic model. To better evaluate the efficiency of dynamicmodel TVAR(1) is used for checking the persistence of factors, and rolling OLS(100)for real time persistence. Out-of-sample forecast results are displayed to compare theefficiency between dynamic models compared by using Diebold-Mariano test forfurther analysis.Κύριος στόχος της εργασίας είναι η χρησιμοποίηση του δυναμικού Nelson-Siegel για την κατασκευή και πρόβλεψη της καμπύλης όπως στους Diebold και Li (2007).