1. Suppose you have estimated GARCH (1, 1) model from daily returns of a stock. The annualized representation (such that the volatility, i.e., square root of variance, is in percentages) of the estimated model is(σt)^2=11.25+0.05*(ut-1)^2+0.90*(σt-1)^2
a.)What is the unconditional volatility (unconditional standard deviation) of the series?
b.)Find predicted volatility two days ahead (i.e., for t+2), if current volatility, σt=18%, and ut= - 0.4%.
2. Suppose that Yt~WN(0,σ^2) (white noise)
a.)What is the first order autocorrelation of Yt?
b.)Compute the first order autocorrelation of the differenced series ΔYt.
1a. Usually it is used the unconditional variance (E(h(t)=E(h(t-1)=v). But, actually any value will do as the variance process settles down after a couple of iterations.
1b. Plug σt=18% and ut= - 0.4% into the formula twice.
2. R1=E[(Y(t)-Mu(t)) (Y(t-1)-Mu(t-1))] / σ^2