Interest rate risk management - calculating Value at Risk using EWMA and GARCH models

Value at risk assesses financial risk by evaluating the probability of loss that results from stochastic variation of the rate of return. The methodology is based on historical data reflecting this variation, usually as an estimated probability of default function. The fact that return distributions are not constant over time poses exceptional challenges in the estimation. In order to remedy this problem we can estimate the volatility of the financial variables using EWMA and GARCH models, that are robust to fat-tailedness in the conditional distribution of returns. The assessment of the models’ performance is based on a range of measures that address the conservativeness, accuracy and efficiency of each one.