Reason Foundation

Reason Foundation

Created or Saved or Estimated or Assumed

The Congressional Budget Office's predetermined stimulus reports

Peter Suderman
March 3, 2010

In selling the American Recovery and Reinvestment Act—otherwise known as the economic stimulus—to the American public last year, the Obama administration promised that the massive spending package would serve as a sort of Keynesian Red Bull, allowing the tired economy to keep partying hard by pumping up GDP and trapping unemployment in single digits. Or, as the administration put it in January 2009, the bill was to create or save three to four million jobs over the next two years, with over 90 percent of those jobs in the private sector.

Instead, the economy reacted like it just downed a glass of whiskey and warm milk: Private sector output fell sharply, and last fall, the unemployment rate hit 10.2 percent.

Yet the Obama administration continues to defend the stimulus, aided in no small part by legally required reports issued by the Congressional Budget Office. But those reports rely on assumption-packed models that effectively predetermine their outcomes; what they say, in essence, is that the stimulus worked because we assume it did. 

Here's what the Congressional Budget Office's (CBO) most recent report on the matter estimates the stimulus' effects were in the fourth quarter of 2009: Thanks to the stimulus, America is somewhere between 1 and 2.1 million jobs richer than it would have been with no government intervention. Federal dollars have fattened up our GDP as well, adding somewhere between 1.5 and 3.5 percent to the GDP. 

Naturally, the Obama administration is keen to take credit. And in touting the CBO's stimulus figures, the White House repeatedly employed the phrase "created or saved." After widespread eye-rolling at such an obvious rhetorical gimmick—not to mention significant evidence that many of the jobs it was claiming credit for were not, in fact, created or saved—the administration altered its lingo and started referring to jobs "funded." But this too is not as accurate as it could be, at least in the context of the CBO's reports; a better phrase might have been "created or saved or estimated or assumed." 

That's because CBO's estimates are generated using models that significantly boost the figures provided by existing measurements (measurements which themselves have been called into serious question). To some extent, that's understandable—determining how many jobs would exist in the absence of a policy is impossible to do with any certainty; no matter how good your models, building a counterfactual is always a guessing game. But it's also a game that's awfully easy to rig, or at least tilt in your preferred direction. 

That's especially true when estimating government spending's productive effects, which is accomplished by plugging numbers into a formula that assumes that government spending produces a multiplier—an increased return for every government dollar spent. In other words, it extrapolates from how much money is put in rather than from what has actually come out. And it does so using a formula that dictates that if money is put in, even more money will come out. According to the CBO's estimates, depending on how the money is spent, one dollar of government spending can produce total economic activity of up to $2.50. What a deal! 

It's certainly a good bargain for the administration, because this methodology grants them a certain amount of predictability in what the outcome of the CBO's stimulus reports will be. Since the stimulus was enacted, the CBO explains that it has only made "small revisions" to its formula (for the previous report, in December, it made no revisions). The CBO's multipliers are estimated from sources similar to those used by the Council of Economic Advisers when it first projected what the stimulus' effects on job creation would be. So for all practical purposes, the same multipliers that were used to predict how many jobs would be created are being used to estimate how many jobs have been created.

That still leaves us with a question: How many jobs did the stimulus actually create? The best answer to that question is not 1 million or 2.1 million or any of the other figures that have been batted around in recent months by the administration and its defenders. It's not even a figure at all; instead it's another question: Who knows? 

But don't take my word for it; take the CBO's. Unlike the administration, the CBO is a nonpartisan entity without a particular interest in strengthening its claims further than they should. All the numbers it produces are estimates, and the agency devotes plenty of ink to explaining its methodology and the uncertainties it entails. Last month's report cautioned that "considerable uncertainty exists about many of these economic relationships that are important in the modeling," which is why many of its estimates come in rather wide ranges. And its December report noted that "it is impossible to determine how many of the reported jobs would have existed in the absence of the stimulus package."

In other words, don't blame the CBO, which is merely doing its lawful duty to produce compliant estimates (a fact which it dryly makes clear in the introduction). Instead, blame the administration, the government-spending enthusiasts, the liberal pundits, and anyone else who treats these pre-cooked estimates as settled fact.

Peter Suderman is an associate editor at Reason magazine. This column first appeared at

Peter Suderman is Associate Editor

Print This