As a market risk analyst, each day you calculate VaR from the available prior data. Then,...

Transcribed Image Text:

As a market risk analyst, each day you calculate VaR from the available prior data. Then, you wait ten days to compare your prediction value VaRt-10 to the realised return and check if the prediction about the worst loss was breached. You are given a dataset with Closing Prices. • Implement VaR backtesting by computing 99%/10day Value at Risk using the rolling window of 21 returns to compute o. (a) Report the percentage of VaR breaches and (b) number of consecutive breaches. (c) Provide a plot which clearly identifies breaches. VaR10D,t = Factor X ot x V /10 ● For comparison, implement backtesting using variance forecast equation below (recompute on each day). Rolling window of 21 remains for o? (past variance) computation. The equation is known as EWMA model, and you can check how variance forecast is done in the relevant lecture. o² +1 | t = λ 0² | t−1 + (1 − X) r² with X 0.72 value set to minimise out of sample forecasting error, and rt refers to a return. Provide the same deliverables (a), (b), and (c). As a market risk analyst, each day you calculate VaR from the available prior data. Then, you wait ten days to compare your prediction value VaRt-10 to the realised return and check if the prediction about the worst loss was breached. You are given a dataset with Closing Prices. • Implement VaR backtesting by computing 99%/10day Value at Risk using the rolling window of 21 returns to compute o. (a) Report the percentage of VaR breaches and (b) number of consecutive breaches. (c) Provide a plot which clearly identifies breaches. VaR10D,t = Factor X ot x V /10 ● For comparison, implement backtesting using variance forecast equation below (recompute on each day). Rolling window of 21 remains for o? (past variance) computation. The equation is known as EWMA model, and you can check how variance forecast is done in the relevant lecture. o² +1 | t = λ 0² | t−1 + (1 − X) r² with X 0.72 value set to minimise out of sample forecasting error, and rt refers to a return. Provide the same deliverables (a), (b), and (c).

Answer rating: 100% (QA)

Here are the steps to implement VaR backtesting and compare the results using the historical and EWM... View the full answer