In January, there was a huge debate about this paper by Romer and Bernstein, advisors to Obama. The paper showed simultations on how Obama plan (the fiscal stimulus plan, there are so many Obama plans) would help the economy.
I was going through this John Taylor paper where he along with other economists shows the projection by Romer/Bernstein is too aggressive and would not stimulate the economy as much as estimated. Taylor says Romer/Bernstein have used old keyensian model but they should have instead used the more advanced and used new keynesian model. The main difference between the two is latter uses expectations more in their models.
Now, this is interesting from two points. One, it is perplexing that why Romer/Bernstein do not show the forecasts using other models as well? They should have based their forecasts on the basis of all possible models to avoid criticism.
Two, do all these models really matter? I am reading this very good postby Buiter who has criticised all these macro models saying they aren’t good enough. New Keynesian models are used by all central banks and we all know how ineffective they have been. But then we need to understand somethings and move somewhere and need models. Confusing bit
Anyways, this is another criticism from Taylor on US Govt policies. (See this as well).