covariance stationary

CFA Level II Quant
#1

3 requirements for a time series to be covariance stationary.

  1. expected value or mean of time series must be constant and finite in all time periods??

Meaninging Y increase from 1 to 2 to 3 to 4 vs 1 to 2 to 4 to 8??

  1. variance must be constant and finite?

  2. covariance constand and finite??

#2

“Linear” regression supposes a well-behavied variables, if not, a linear regression is not appropriate.

If a variable is not convariance stationary, you can make it to be (applying log or differences for example).