Exercise

Compute parametric VaR

In this exercise, you will practice estimating dynamic 5% daily VaRs with a parametric approach.

Recall there are three steps to perform a forward VaR estimation. Step 1 is to use a GARCH model to make variance forecasts. Step 2 is to obtain the GARCH forward-looking mean and volatility. And Step 3 is to compute the quantile according to a given confidence level. The parametric approach estimates quantiles from an assumed distribution assumption.

A GARCH model has been fitted with historical Bitcoin return data up to 1/1/2019, then it has generated mean and variance forecasts, saved in mean_forecast and variance_forecast respectively. The GARCH model assumes a Student's t-distribution, and its \(\nu\) (degree of freedom) is saved in nu.

Instructions

100 XP
  • Compute 0.05 quantile from the assumed Student's t-distribution.
  • Calculate VaR using mean_forecast, variance_forecast from the GARCH model and the quantile from the previous step.