Here's how fiscal stimulus is supposed to work: The federal government injects a few hundred billion dollars into a sluggish economy through federal spending. That spending sparks additional consumer demand. And that additional demand drives new economic activity that results in a multiplier effect, in which a dollar of initial government spending creates more than a dollar of economic activity, spurring economic growth and job creation.
That was the basic stimulus theory, as explained by Lawrence Summers, who chairs President Obama's National Economic Council, at the tail end of 2008. The story was convincing enough: In 2009, President Obama signed the American Recovery and Reinvestment Act (ARRA)—an $830 billion stimulus package—into law, promising "unprecedented transparency" in tracking how the money would be used and how many jobs it would create.
So how did the stimulus work in practice? Daniel Rothschild, a former researcher at George Mason University's Mercatus Center who now works for the American Enterprise Institute, wanted to find out. Working with Mercatus Center economist Garett Jones between August and November of last year, he oversaw 50 hours of interviews with businesses and contractors that received and applied for stimulus funding. And what he found was that the on-the-ground reality of the stimulus was far messier than the simple theory behind it.
In his initial report, Rothschild relays an illustrative story about a contractor with 25 years of construction experience, much of it laying tile in government buildings. Heading into an otherwise typical job that he expected to account for about two percent of his annual income, the tile-layer made plans to install standard blocks of four-inch white tiles—the same tiles he usually installed, the same tiles found in other parts of the same office complex, and the exact materials called for in the architectural plans.
Then he got updated specs. The large white tiles were out. Tiny, colored tiles that needed to be laid in an intricate pattern were in. Did it matter that the smaller tiles would cost the government 50 percent more than the larger white tiles? Not at all. In fact, the higher cost may have been the point. The tile-layer told Rothschild's interview team that "the only reason he could see for using the smaller tiles was to move the money out the door on the ARRA schedule." So in exchange for their stimulus dollars, taxpayers got a government building with fancier floor tiling.
At other times, they got even less. Rothschild tells another story of a truck salesman who placed a stimulus-funded order for an expensive big-rig. Delivery times on those vehicles range from six to nine months. But stimulus recipients are required to report how many jobs they've created every three months. "There was no work that he had actually generated. He was just a salesman," says Rothschild. And the lag complicated the reporting process as well. "There was no real way to report what would be a fairly normal business transaction."
Rothschild's survey results suggest that the law's reporting requirements, despite having been billed as unprecedented transparency measures, often created more confusion than clarity. In some cases, for example, firms that received stimulus money were instructed by their federal overseers to count jobs created by taking the total amount of money they were given, divide it by some predetermined per-job salary figure, then report the result as the number of jobs created.
Essentially, they were told to assume in their reports that if they got the money, they created jobs—regardless of whether or not any jobs were actually created.
Many firms, however, were given no reporting instructions at all, and thus had to figure out how to count jobs created all on their own. That proved particularly challenging for blue-collar contractors. "The things that were much easier to report were the things that government typically contracts for," Rothschild says. "And the things that the government typically contracts for are knowledge and managerial positions—things that require a college degree or higher. The entire expense and reporting system was based around the idea that you're hiring white collar people to do white collar jobs."
Part of the complexity comes from determining exactly what counts as a job created or saved. An ongoing full-time position certainly counts. But what about the temporary work required by most labor-intensive construction projects? Rothschild describes a local housing authority he interviewed that reported creating 25 jobs with stimulus funds. Seven of the jobs, however, were for house painters who worked for just a few weeks. "But they were still reported as being newly created jobs," he says. The reports did not distinguish between short-term project work and long-term new positions.
Stimulus money also appears to have funded projects that worked at cross-purposes: Rothschild recalls interviewing a manager at a non-profit that trains non-skilled workers to become skilled craftsmen and construction workers. But wasn't the stimulus supposed to help create jobs for the existing construction workers who'd been laid off as the economy slumped? In theory, the stimulus was supposed to solve the problem of too little demand for construction labor. But it also ended up funding projects designed to increase its supply.
It's clear that the stimulus created some jobs in some instances. But the self-reporting is inconsistent enough that the true number is hard to know with much confidence. And many of the projects that the stimulus funded may have been little more than busywork—like the tiler ordered to lay smaller tiles.
"Modern economics has become divorced from micro-economic realities," Rothschild says. "Things that don't fit neatly into a model tend to be discarded." But in any system as expansive as the one created by ARRA, there's bound to be an awful lot that doesn't fit the model. Inevitably, then, complications arise. "No matter how well you design something, the micro level details become very difficult," Rothschild says. The way stimulus is supposed to work and the way it does, in other words, are rarely the same.
Peter Suderman is an associate editor at Reason magazine.