Measuring and evaluating factors of dynamic term structure models for value-at-risk estimation in the Kenyan market

Abstract

The growth of the Kenyan fixed income market and the growing need for investors to diversify the risk in the portfolios has driven the need to quantify the risk associated with the fixed income assets. The measures that are commonly used to estimate bond risk are duration and convexity. However, these measures do not sufficiently assess the risk in fixed incomes. The risk associated with the yield curve illustrates how a portfolio will react to different exposures based on how the yield curve shifts. In this research, we will seek to model the yield curves for the Kenyan market using the dynamic factor models, namely, Nelson-Siegel and Svensson models. We will estimate the factors for both models and seek to establish the distributions of the estimated factors. We will then seek to use the estimated factors from both models to generate the vector of expected bond yields and the covariance matrix that will be used to measure the Value-at-Risk. The results of this research will be used to seek a parametric method of measuring risk in a fixed income in an illiquid market and check whether the estimated factors are good fits to be used in the parametric model.