In: Statistics and Probability
Econometrics Question
(a) What properties must a time series {yt} have to satisfy in order for it to be a covariance stationary process?
Let {ut} be a sequence of i.i.d. N(0,1) variables.
Let yt= α+βt+ut, and yt = yt -yt-1.
(b) Is {yt} covariance stationary? Justify your
answer.
(c) Would your answer to part (b) change if {ut} were a
moving average of order 1?
(d) Is {yt} covariance stationary? Justify your
answer.
= change in, won't allow a triangle/delta.
Solution -
Ans part (a) - A time series {Xt} is need to satisfies the following properties for covariance stationary :