I had this in mock 1, some of you might’ve had it in mock 2 the question was about beloit suspecting the model to have unit root. the model was something like (P/E)t = B0 + B1 (P/E)t-1 + e now if you had to test whether this model had unit root the appropriate model would be: (P/E)t - (P/E)t-1 = B0 + B1 (P/E)t-1 + e Shouldn’t it be… (P/E)t - (P/E)t-1 = B0 + B1 (P/E)t -1 - (P/E)t-1 + e
to calculate unit roo it would be (P/E)t - (P/E)t-1 = B0 + G(P/E)t-1 + e -G is b1-1…then you run the dicker fuller test to see if g=0 (if so–>unit root) To fit unit root you would then do -(P/E)t - (P/E)t-1 = B0 + B1 [(P/E)t -1 - (P/E)t-2] + e
ughhh…not cool… you’re right…they use G in the damn equation - not B…can’t believe i missed that… thanks…that make sense…