Macroeconomics in interest rate term structure modelling

This thesis uses a literature review to describe the latest progress in the models used to describe interest rate term structure (TS). The thesis emphasises macroeconomic variables and models, as these elements have become an essential part of TS modelling in the 21 st century. Although this type of literature review has not been conducted before, the rapid development of numerous kinds of models to describe the TS has highlighted the need for a comprehensive summary. The macroeconomic explanation of TS is quite different than what it used to be. In the past, TS had no real explanation for the driving forces behind the term structure of interest rates. Current research has changed this setup significantly for the better, as the empirical results show that these new models outperform the ones used before the use of macro variables as explanatory variables of TS. There has been a huge amount of literature on TS, both past and present. Accordingly, the selection of source material for this literature review has not been self-evident. Hördahl et al. (2006), De Graeve et al. (2009) and De Grauwe (2008) are the most interesting recent studies on TS. The core analysis of this literature review has been divided into three subsections: the VAR model, three structural models and one DSGE model. Almost all of these models contain empirical sections and these have been compared with each other. This comparison showed that only few of these models were better than the random walk hypothesis. Although the DSGE model did not beat the random walk hypothesis, it did give a valid theoretical model that will probably become the starting point for TS in the future. The result of this literature review is that, although many obstacles have been overcome, several still exist, such as linear dependence between the yields and state variables. The undeniable fact, proven by several recent studies (some of which have been presented in this thesis), is that macro variables have a significant effect on TS. Accordingly, these macro variables should be taken as explanatory variables in the determination of TS.