Estimation of the Term Structure of Interest Rates: Methodology and Applications

In macroeconomics and finance, it is extremely useful to have knowledge of the Term Structure of Interest Rates (TSIR) and to be able to interpret the related data. However, independently of its latest particular application, the TSIR is not observable directly in the market and a previous estimation of the yield curve is needed. There are two distinct approaches to modelling the term structure of interest rates. The first is essenciallly to measure the term structure using statistical techniques by appliyng interpolation or curve-fitting methods to construct yields. The second approach is based on models, known as dynamic models, which make explicit assumptions about the evolution of state variables and asset pricing models using either equilibrium or arbitrage arguments. There is no consensus on any particular methodology and the choice between alternative curve models is, in part, subjective. Nevertheless, the interpolation or curve-fitting methods have showed good properties and are those used nowadays by the vast majority of central banks. The objective of this article is to make a concise study to enhance knowledge of the Term Structure of Interest Rates and of the principal methodologies based on statistical techniques that have been proposed for its correct estimation. The most relevant empirical work on comparing the principal methodologies is also presented. A better understanding of the principal methodologies and their limitations will provide the researchers a panoramic vision of the problematic of estimation of the yield curve and will enable them to chose the best model according to their objectives.