“The White House estimates of 3.6 million new jobs is based on an ‘Old Keynesian’ model on the impact of government spending, while the new models adjust for the rational behavioral response to the stimulus by businesses and consumers. The White House figures, by economists Christina Romer and Jared Bernstein, also assume zero interest rates for a minimum of four years. The alternative assumes, more reasonably, that as growth returns interest rates will also rise What the four economists [Cogan, Cwik, Taylor, and Weiland] found is that the Administration’s estimates for stimulus growth were six times as high as they could produce under a modern Keynesian simulation. By their estimates, the stimulus would produce, at most, 600,000 jobs and add perhaps 0.6% to GDP at its peak. That’s nowhere near a multiplier of 1.5 and suggests the $800 billion would have been better devoted to business tax cuts or fixing the financial system. That’s $1.3 million in spending per job, for those keeping score at home.”
The paper on which much of the article is written is “New Keynesian versus Old Keynesian Government Spending Multipliers” which is by John F. Cogan, Tobias Cwik, John B. Taylor, Volker Wieland.