Lars Christensen writes: "...monetary policy determines NGDP (Remember MV=PY!)."
Not quite right. And this sentence should send up red flags for you just like the sentence "fiscal policy determines NGDP (Remember MV=P(C+I+G)!)" should send up red flags for you.
What would be OK to say is "monetary policy is a determinant of NGDP", and then cite whatever accounting identities float your boat.
Why is this worth pointing out?
For the same reason that Brad DeLong has been harping on the problem of expectations and monetary policy. M only matters in the quantity theory if we assume V, P, and Y are playing nicely, just like G only matters in the national income equation if we assume C and I sit idly by.
Generally speaking, these are bad assumptions.
Elsewhere Christensen writes that "there is no such thing as fiscal policy", following much the same logic. In response to that, we have to move out of nitpicking territory and firmly into "he's just wrong" territory.
Let's make a pact: you market monetarists promise not to tragically abuse the quantity theory if we Keynesians promise not to tragically abuse the national income equation.
UPDATE: Oh! And it's also always worth remembering the excellent point that regular commenter Current is always good about bringing up: MV=PY is actually a lie we tell ourselves anyway! Money and prices are exchanged in transactions, not for output. Assets, used couches, etc. MV=PY is a convenient lie, but a lie nevertheless.