Simulations of Mark Zandi’s economic model, which are reported in the press to show that a new temporary stimulus package will create 1.9 million jobs, are being touted as evidence that it will work. This is the same type of model simulation that predicted the very similar 2009 stimulus package would create millions of jobs, and the same type of simulation that claimed that that package worked. Andrew Ferguson reviews the predictions in a recent article in Commentary.
But simulations of such models do not provide such evidence, as I have explained on this blog before, for example here and here. They are wrong because they assume “multipliers” for temporary one-time payments or tax changes far in excess of the basic “permanent income” or “life cycle” models (which we teach in Economics 1). They are wrong because they assume that state and local government infrastructure and other purchases respond to federal stimulus grants in a mechanical way, unlike what we have seen practice, as I explained in this article with John Cogan. And they are wrong because they do not take account of the negative growth effects of expected future permanent increases in tax rates. I have debated Mark Zandi on these topics many times before, for example on the NewsHour and in congressional testimony.