Did the 2009 stimulus package work?

Congress passed the American Recovery and Reinvestment Act in February of 2009. If you will recall, at the time it was touted by its proponents as an economic treatment for the recession caused by the housing financial crisis that preceded it. But did it do its job?

In trying to determine if the stimulus package worked, we first need to define what “worked” means in this context. What is success for a stimulus package?

When the stimulus was being considered in 2009 the Council of Economic Advisors released a report predicting what would happen to the U.S. economy if the package was passed and contrasted this with what would happen if it didn’t pass. The unemployment rate was predicted to peak at just over 9 percent during 2010 and run between 8 percent and 8.5 percent in early 2011 if the stimulus package did not pass. With the stimulus, the unemployment rate was supposed to peak at just under 8 percent at the end of 2009 and would run between 6.5 percent and 7 percent for the first half of 2011.

In reality, the unemployment rate peaked at 10.1 percent in October of 2009 and is currently inching down slowly to its most recently reported values 8.3 pecent. Clearly, by the standards of the CEA report, the stimulus did not “work.” In fact, one might argue it worked negatively, since the actual outcome was worse than what was predicted if congress did not act.

But the problem with such comparisons is that it is difficult or impossible to establish a counterfactual. What really would have happened if the stimulus had not passed?

Proponents argue that the stimulus helped because things would have been even worse without it. Unfortunately, the validity of that argument cannot be tested, since we can't rerun the economy starting February of 2009 without the stimulus.

Over any lengthy span of time a great number of random, unpredictable events will occur that effect the economy. If we could rerun the economy a thousand times we would get a thousand different sets of random outcomes. If these events really are random and unpredictable before they happen, it is pointless to argue that something different should have been done.

Rather than ask if a policy was appropriate given what we know now, a more important question is if the policy was appropriate given what was known when it was put in place.