Okay, I'm on a little shaky ground here because all we do in macro is consider coconut trees, determine how many coconuts should be consumed and how many planted, how the decision changes when global warming wipes out the island in finite time periods and when the villagers live forever, how different islands will trade coconuts, how many coconuts is a coconut tree worth, etc. Then you do it all over again when you don't know how many coconuts a coconut tree will bear, so on and so forth. Rochester fully embraces the ideal of macroeconomics as stochastic calculus porn.
Anyway, if I recall correctly, fiscal stimulus is only really effective when there's been a contraction in real GDP relative to potential GDP. If potential GDP is actually contracting, well, one just has to suck it up, the fiscal stimulus is simply an intergenerational transfer as opposed to a welfare-improving policy.
As such, there's relatively good reason, I think, for China's recently announced fiscal stimulus (even if most of it would have been spent anyway, and it's dragged out over several years) compared to the justification for American fiscal stimulus, where I think there's a real argument to be made that the economy is actually contracting, not just experiencing weakness, i.e. lots of capital in the financial and automotive sectors is not that useful as it was two years ago or twenty years ago, respectively.
Therefore I'd advocate Obama focus the little room for new spending he has on initiatives he finds personally compelling, rather than trying to spend where he's told it'll produce the biggest fiscal multiplier. Besides, this is at least in part a crisis of confidence and it would be a lot easier for him to give good speeches on health care initiatives, say, than auto bailouts.