Paul Krugman provides an analysis showing arithmetically that the fiscal stimulus will have a declining contribution to GDP growth, even as stimulus spending increases.
What is interesting to me about this analysis is that the concept of a multiplier has completely disappeared. If you believe in a multiplier, then as far as stimulus goes, the sooner the better. The more you spend this quarter, the more people are employed this quarter, the more they spend next quarter, and so on. If you do not believe in a multiplier, then any time you slow the rate of government spending growth you slow the rate of overall GDP growth.
If there is no multiplier, and we follow the Krugman logic, then we have no choice but to keep increasing government spending until...what point? when it is 100 percent of GDP?
On the other hand, if there is no multiplier, then I think we should question the whole concept of stimulus. With no multiplier, its benefits are almost entirely transitory and artificial.