Hey all - looking at Jorion chapter 9, pg. 230 (The RiskMetrics Approach). The following formula is given for h_t, today's conditional variance:
h_t = lambda * h_t-1 + (1 - lambda) * (r_t-1)^2
Exercise #11 of the same chapter has you using this formula to get an estimate of volatility. The problem gives the latest volatility (h_t-1) as 1% and the latest return (r_t-1) as 2%. lambda = 0.94
Looking at the solution, h_t, h_t-1, and r_t-1 are scaled up. The author solves the problem like this:
h_t = 0.94 * (1)^2 + (1-0.94) * (2)^2 = 1.18
So in these formulas, 1% = 1, not 0.01, and 2% = 2, not 0.02. I don't see this convention explained in the text. So....is this the way we should be solving conditional variance problems if they appear on the exam?
Thx
h_t = lambda * h_t-1 + (1 - lambda) * (r_t-1)^2
Exercise #11 of the same chapter has you using this formula to get an estimate of volatility. The problem gives the latest volatility (h_t-1) as 1% and the latest return (r_t-1) as 2%. lambda = 0.94
Looking at the solution, h_t, h_t-1, and r_t-1 are scaled up. The author solves the problem like this:
h_t = 0.94 * (1)^2 + (1-0.94) * (2)^2 = 1.18
So in these formulas, 1% = 1, not 0.01, and 2% = 2, not 0.02. I don't see this convention explained in the text. So....is this the way we should be solving conditional variance problems if they appear on the exam?
Thx
Exponentially Weighted Moving Average
