(a) Consider the following AR(1) model Design a simulation experiment to determine the effect of increasing the value of ϕ from 0 to 1 on the distribution of the t-ratios.

(b) Consider again the AR(1) model from part

(a) of this question. As stated in Chapter 4, the explanatory variables in a regression model are assumed to be non-stochastic, and yet yt−1 is stochastic. The result is that the estimator for ϕ will be biased in small samples. Design a simulation experiment to investigate the effect of the value of ϕ and the sample size on the extent of the bias.