View Full Version : Why use DW to test for random walk?
11-02-2003, 05:23 PM
11-02-2003, 05:45 PM
If the variable is a random walk (w/o drift?), I think yt minus yt-1 (in the numerator of the DW stat) should equal zero. I think the Econometrics textbook discusses this in the section on co-integration........
11-02-2003, 05:57 PM
11-02-2003, 06:49 PM
and don't forget that DW ~ 2(1-rho)
so that if DW = 0, rho = -1
if DW = 2, rho = 0
and if DW = 4, rho = +1
01-23-2004, 08:19 PM
DW knows about random walk. I've walked randomly the few times I was at their happy hours.
01-23-2004, 08:20 PM
I hear ya, brother. David Ueki sent me a bottle of wine that knocked me on my ass.
vBulletin® v3.7.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.