Methodology and Implementation of Value-at-Risk Measures in Emerging Fixed-Income Markets with Infrequent Trading.

This paper deals with the issue of calculating daily Value-at-Risk (VaR) measures within an environment of thin trading. Our approach focuses on fixed income portfolios with low frequency of transactions in which the missing data problem makes VaR measures difficult to calculate. We propose and implement a methodology to calculate VaR measures with an incomplete panel of prices. The methodology is composed of three phases: Phase I, generates a complete panel of prices, using a term-structure dynamic model of interest rates. Phase II, calculates portfolio VaR measures with several alternative methods using the complete panel data generated in phase I. Phase III, shows how to back-test the VaR measures obtained in phase II using the original incomplete panel of prices. We provide an empirical implementation of the methodology for the Chilean fixed income market. The proposed methodology seems to provide reliable VaR measures for thinly traded markets addressing an important issue for financial risk management in emerging markets.