White House's Job Creation Figures

PrintPrint

The Council of Economic Advisors’ latest analysis of the $862 billion stimulus bill figures that it created 2.5 to 3.6 million jobs. How good is this estimate? The same CEA told us that the stimulus bill would keep the unemployment rate below 8 percent. Almost a year and a half later unemployment is still above 9 percent. 

The CEA uses two types of analysis to calculate its estimates. The first is a model built on the assumption that government spending will have a positive multiplier effect on the economy. In other words, the outcome of job creation is predetermined. 
Economist Greg Mankiw explains: 
First, there are model simulations. That is, the CEA took a conventional Keynesian-style macroeconomic model and used those set of equations to estimate the effect the stimulus should have had. Essentially, the model offers an estimate of the policy’s effect, conditional on the model being a correct description of the world. But notice that this exercise is not really a measurement based on what actually occurred. Rather, the exercise is premised on the belief that the model is true, so no matter how bad the economy got, the inference is that it would have been even worse without the stimulus. Why? Because that is what the model says. The validity of the model itself is never questioned. 
The second method tries to estimate the stimulus bill’s effects by comparing real changes in GDP and employment against a baseline forecast. However, even the CEA admits that this approach is subject to “considerable margins of error,” and that “the comparison will reflect not just the impact of fiscal policy, but all other unusual influences on the economy following passage of the Act.”
 
Mankiw again throws cold water on the analysis: 
Second, the CEA offers some statistical evidence that things got better after the stimulus passed. Some of this evidence comes early in the document in the form of simple graphs.  Some comes later by examining deviations from forecasts based on a two-variable vector autoregression. But the nature of the evidence is basically the same: Post hoc ergo propter hoc.
 
Of course, there were a lot of other things going on in the economy at this time.  Monetary policy, for example, has gone to extraordinary measures to get the economy going. TARP was also an unusual intervention that seems to have done its job of returning the economy to some degree of financial normalcy (even if leaving the bad taste of increased moral hazard). Giving credit for the economic improvement to the fiscal stimulus is a large leap. 
Does anyone think the White House economic team was going to produce estimates that didn’t show substantial job creation resulting from the stimulus?
 
This isn’t to suggest that the stimulus bill didn’t create any jobs. It did. The real question is whether it created any net jobs after all the negative effects of the spending and debt are taken into account. How many private sector jobs were lost or not created in the first place because of the resources diverted to the government for its job creation? How many jobs were lost or not created because of increased uncertainty about future tax increases and worries about rising debt creating another financial crisis? Don’t expect the administration’s economists to tackle these questions anytime soon.