Greg Clark says it is a good bet:
Picking sides (sort of) - The Atlantic Business Channel: Brad DeLong asks that I not just bemoan the exquisite irrelevance of most academic economists that the current crisis has revealed. Instead he asks that I take sides in the debate between Keynesians like him and Krugman and neoclassicals like Eugene Fama and John Cochrane on the efficacy of fiscal policy. I am no macroeconomist. And I am reluctant to get trampled in a clash of macroeconomic titans. But I find Brad's various arguments, particularly his reductio ad absurdum, convincing. And I think Obama is right to try a fiscal stimulus from the following simple chain of argument:
There is no logical reason why a stimulus cannot work -- it is a matter of empirics whether it will work or not, depending on the reactions of various economic actors.
Because of the absence of controlled experiments over the last 80 years we do not know whether a stimulus will actually increase output (though Christina and David Romer have some decent evidence from the US that it will).
Even if the stimulus package does not produce one extra job, the social cost of the stimulus is a fraction of the $789 billion being spent, since the tax reductions, unemployment benefits, aid to states and educational and health investments all have some value. The true social cost, absent any output gains, is likely $100 billion or less.
The potential gains are huge. If the multiplier really is as high as the 1.9 sometimes found in Econ 1 texts such as Greg Mankiw's, then the social gain from the $100 billion expense could be as much as $1,400 billion. It is a risk, for sure, but those seem like attractive odds at which to gamble.
(1) - (4) amount to a version of Pascal's Wager -- on why the prudent person should believe in God -- but applied to more mundane economic concerns.