I'm usually not one to say to refuse a certainty for a possibility, but IMO, UCLA all the way. I mean put it this way, despite the fact that time series was one of my most enjoyable courses at Rutgers (same one for stats and msmf masters), the professor didn't cover cointegration which I later found out about by reading part of Vidyamurthy's book (which I couldn't finish because it was on a pdf at work and the desk went under far faster than I expected it to, but that's another story).
In this case, I recommend you take the evil I don't know, rather than the evil I do.