Some people really hate their own birthday — they hide it from friends and they eschew any suggestion that a dinner be put together in their honor. So it might be fitting that last week came and went with little fanfare for the American Recovery and Reinvestment Act, more widely known as the stimulus.
February 17 was the four-year anniversary of the stimulus, signed into law in 2009 with much pageantry by a newly minted Obama administration to “save the economy.” And while a fourth birthday may not be as much of a milestone as entering the terrible twos or reaching double digits, it is worth taking stock of where the economy is today, some 1,460+ days into the life of the stimulus.
Since the Recovery Act began to distribute money, GDP has gone from -0.3 percent in early 2009 to -0.1 percent at the end of 2012, though it has averaged 1.9 percent in the fifteen quarters between then and now. Over the same time period headline unemployment has gone from 8.7 percent up to 10 percent and back down to 7.9 percent. (A broader measure of unemployment shows a change of 15.7 percent unemployed in early 2009 to 14.4 percent today.)
What story these numbers tell depends on the reader’s interpretation. We could imagine both an “it could have been worse” counterfactual and an “it should be better” response with the same data. To better understand this complication, we took a look at what standard Keynesian theory says should have happened with stimulus spending, as well as a brief look at what actually did happen. Both are shown in the chart below.
While the Recovery Act was passed in February 2009, the money it designated for spending didn’t really start rolling out the door until April 2009. By then the economy was already bouncing back on its own from a deep recession that began in late 2007 and ultimately ended in June 2009 (as can be seen by the orange line on the chart). The dark blue shows the cumulative federal spending as authorized by the Recovery Act, and the light blue the additional tax benefits of the stimulus that reduced federal revenues, and essentially act like spending. As of this writing most of the stimulus money has been earmarked for specific projects, but $55.7 billion of the $840 billion has not yet been spent.
Ironically, this pattern breaks from standard Keynesian theory — the guiding intellectual framework for those that designed the stimulus. The basic concept of stimulus spending is to fill in the gaps of output and consumption in the economy during a recession to boost the economy back to a point of recovery. According to textbook Keynesian expansionary and contractionary rules, stimulus spending should begin at the start of an economic downturn and cease when GDP is once again growing at a positive rate.
If Congress were actually following this theory, then federal spending — aka stimulus — would have increased in the first quarter of 2008, peaked in the third quarter of 2008 and returned to pre-recession levels as the economy began to show positive GDP growth in mid-2009. However, tax benefits from the Recovery Act were not made available until the trough of the recession — when the economy was darkest. Further, most grants, contracts, and loans to businesses made available under the Recovery Act did not begin to be disbursed until the economy was expanding again.
Theoretically, by this measure it could be argued the stimulus didn’t “work” or didn’t have anything to do with the recovery. The Recovery Act was at best faux-Keynesianism.
Still, there are plenty of other measures that suggest the stimulus “worked” or “failed” — it all depends on what benchmark you want to use to measure success.
For instance, the White House claimed that the Recovery Act would prevent unemployment from passing 8 percent, when in fact it peaked at 10 percent. Therefore, many center-right commenters have declared the stimulus a failure because unemployment remains at 7.9 percent today (and the White House also claimed that by January 2013 unemployment would be down to 5.2 percent because of stimulus spending). The second chart below reflects this.
Center-left commenters counter that these unemployment projections did not fully account for how bad the economy was going to get. Moreover, they argue that without the stimulus the economy would be much worse off and unemployment would be higher, therefore it was a success. But by this logic, the stimulus was destined to be a success because no matter how bad unemployment got, even to 25 percent, stimulus supporters could just claim job losses would have been even greater without the economic support. Therefore, success by any result.
Meanwhile, the libertarian view is that stimulus spending crowded out private sector spending and prevented the recession from clearing away asset bubbles and reallocating resources via creative destruction. But just like the progressive claim, we can’t know this for sure either because this counterfactual is as unknowable as the “it could have been worse” theory.
It is also clear from observing anecdotal stories that there are conflicting views on the affects of the stimulus on employment in America — whether positive or negative.
- In January 2010, Charleston, SC Mayor Joe Riley credited the Recovery Act with job creation in his city, highlighting $16 million received from the federal government to help “pay for affordable housing renovations, new police officers, a new community center and more.”
- However, across the country Seattle was not as successful in turning Recovery Act funds into job opportunities. In August 2011, the Seattle Post Intelligencer reported that a $20 million grant designed to create 2,000 jobs was used to hire a mere 14 workers to weatherize three homes. This was a far cry from the 2,000 homes that were intended to be retrofitted.
This ultimately means there is no perfect benchmark that we can measure the Recovery Act against to determine whether it was a good program that “saved” the economy, or an awful one that “made things worse.” That’s what drives the conflict. Every methodology that has attempted to measure how many jobs were created by the stimulus has to deal with a host of exogenous factors that cloud the final judgment. Many use multipliers to estimate the affects of particular types of spending and tax breaks, but these are estimates built from macroeconomic models that are by their very nature of full of assumptions, simplified metrics of the economy, and estimates themselves. As a result we get a lot of dubious estimates built from models of estimates.