Many time series used in finance are described as non-stationary series.

a. What do we mean by non-stationary? Why is it important to test for non-stationarity and account for it in estimation?

b. Three forms of standard Dickey-Fuller test are given below: A. = 1 + B. = +1 + C. = + + 1 + Which of these forms would you use to test for non-stationarity in each of the following variables and why? i. Stock prices ii. Stock returns iii. The residuals of an Engle-Granger test

c. Consider the following data generating process for a series : t t ut y = +1.5y 1 + What most accurately describes the process for ? How would you work with this model for ?

d. What difficulties can arise when applying and interpreting Dickey-Fuller tests? Describe in detail another test which could be applied to test for non-stationarity.