Description
Revisiting problem 1 from homework 3
We will consider additional variations on how to compute VaR for the return series and how to judge accuracy of VaR.
- Go back to the previous homework. For each day in the sample, computethe 95% confidence intervals of the historical VaR and the exponential weighted VaR you obtained in Question 1 of homework 3, using both parametric (for the historical VaR) and bootstrap methods (for the two measures). For the parametric method, assume the gains are normally distributed.
- Compute volatility using the EWMA with λ = 0. Compute the corresponding measure of VaR.
- Use maximum likelihood estimation to estimate a GARCH model forvolatility. Compute the corresponding measure of VaR.
- Revisit your answer to question 5 last week in light of these new results.
- Case Study: Implementing Quantitative Risk Management and VaR in a Chinese Investment Bank
You should buy the case study material using this link: https://hbsp.harvard. edu/import/722095.
1
- Explain the objectives and priorities of each player: Jasper Wang, JianguoLu, and Charles Pan. What is motivating the different players? What tensions existed among their different objectives?
- Why does Jasper choose to make the VaR model the first step towardsrationalizing the trading function? What is the appeal of the VaR model generally?
- Why do Jianguo and the traders resist the VaR model? Do you think theirpattern of resistance to risk management is unique to China, or might it be found elsewhere too?
- Using the spreadsheet provided, run backtests of the VaR predictionsagainst actual daily gains or losses for both the S&P 500 index and the Shanghai Composite index.
- Starting with a lookback period of three months, observe the numberof exceptions in all years for both the Shanghai and S&P indexes. How do they compare?
- Try different lookback periods (say, 3, 6 and 9 months) to see if thelength of the period changes your conclusions.
- Given that Jasper’s VaR model assumes a 95% confidence level, howwell does the backtest validate the model?
- How might Jasper use the backtest results to bolster his case for introducing the VaR model?
- How successful do you think Jasper will be in his attempt to implementWestern risk management practices? What advice would you give to someone in a role similar to his?
- What is the current regulation environment of risk followed by Chinesebanks and how has it evolved since the crisis? (Find information beyond the case study material)
2