Consider the following ARMA models. Suppose that all the estimated parameters are statistically significant. Which of the models will you prefer for the forecasting and WHY? Use the conditional and/or unconditional expectations and unconditional variance to justify your answer. (a) rt = 0.12 + 0.4rt1 0.1t1 + t ,

Where 2 = 0.05

(b) rt = 0.1 + 0.95 rt1 + 0.8 rt2 0.5 t1 + t Where 2 = 0.1

(c) rt = 0.2 + 1.7 rt1 0.7 rt2 0.8 t1 0.5 t2 + t Where 2 = 0.12

Suppose rt1 = 0.1, rt2 = 0.2, t1 = 0.2, t2 = 0.2