Exercise 4.2 Assume that the returns of a portfolio loss X at any time t are correctly modeled by a process of the form μ+σW(t), where W is a Brownian motion.

Using n = 506 daily returns, the mean and the standard deviation are respectively 400 and 250. Compute the estimation of the 99% VaR for the value of the portfolio in one year. Also compute the estimation error corresponding to a 95% confidence level. How does this compare to the nonparametric estimation error?