I really like the idea of doing Econometric work with R. Currently, I've been looking for a more hands on way to run some regressions and test data, this guide looks to prove immensely useful.
(2) It has a huge repositoriy of libraries. There's hardly anything that you can't do with R. In my special case it was about hazard-rate models, which are barely implemented in other tools that I tried (namely EViews, SPSS)
(3) It has a BIG community behind it, which you can ask for help once you're stuck. I haven't found this kind of "service" in other software (namely, again, EViews).
Almost the only reason you'd need to use R. I cringe when professors ask their students to use non-free software when equivalent or better free alternative exist.
In a mostly irrelevant personal note Grant and I graduated in Numerical Mathematics from BYU the same year; there were only a handful of us so we spent a lot of time together in the top floor of the Math/CS Building. He was working on this during a big chunk of our undergrad and I think some after as well. I'm embarrassed to say I haven't taken the R plunge yet (SAS and STATA right now) but when I do I'm sure this will be my starting point.
This paper covers a lot of the basics in addition to the complex stuff I don't understand yet. I'll definitely be keeping it saved as a general R reference as I do more work in R.