1. (Wooldridge, 12.4) True of false: "If the errors in a regression model contain autoregressive conditional heteroskedasticity, they must be serially correlated."
2. (Wooldridge, 12.5) In a particular statistical exercise, a regression of the OLS residuals on the lagged residuals produces ρˆ = 0.841 and se(ρˆ) = 0.053. What implications does this have for OLS? If you want to use OLS but also want to obtain a valid standard error for the coefficient, what would you do?
3. (Wooldridge, 18.3) Suppose that {yt} and {zt} are I(1) series (series that need to be first differenced to produce a stationary and weakly dependent process), but yt -βzt is I(0) for some β? =/= 0.
Show that for any δ? =/= β, yt-δzt must be I(1).