Is the CBO Cheerleading for the Stimulus?


Obviously, the Congressional Budget Office has been telling us all along that the $800 billion stimulus package signed by President Obama in 2009 was a rousing success to be praised and emulated. Indeed, I heard a rumor recently that CBO director Douglas Elmendorf is producing a Broadway musical titled More Stimulus Now, a tragicomedy about a cross-ideological romance between a Keynesian and a Hayekian in a time of weak aggregate demand. Washington Post columnist E.J. Dionne certainly seems to be singing and dancing already:

Yet the drumbeat of propaganda against government has made it impossible for the plain truth about the stimulus to break through. It was thus salutary that Douglas Elmendorf, the widely respected director of the Congressional Budget Office, told a congressional hearing last week that 80 percent of economic experts surveyed by the University of Chicago's Booth School of Business agreed that the stimulus got the unemployment rate lower at the end of 2010 than it would have been otherwise. Only 4 percent disagreed. The stimulus, CBO concluded, added as many as 3.3 million jobs during the second quarter of 2010, and it may have kept us from lapsing back into recession.

Dionne links to a Post report on Elmendorf's recent congressional testimony that starts by asking "Did the stimulus work" and answers by saying that even though Republicans insist the stimulus failed, "the director of the nonpartisan (and widely respected) Congressional Budget Office was emphatic about the value of the 2009 stimulus"—the value being that most economists in the survey Dionne notes believe that the stimulus package increased employment in 2010, an assessment that the CBO's official reports on the stimulus agree with.

So Elmendorf and the analysts at the CBO might as well be a bunch of Paul Krugman clones, right? Or maybe not.

The first mistake here is to assume or imply that Elmendorf's statements constitute an endorsement of the stimulus. The CBO as a rule does not make policy recommendations; Elmendorf, in particular, has been extremely careful not to endorse any particular policy agenda. Instead, the CBO's business is describing the potential budgetary effects of policies past and future for Congress to the best of the agency's abilities.

But with big-ticket macroeconomic policies like the stimulus, there are serious limitations on its capabilities—not due to competence, but due to the fact that the effects of stimulus spending are notoriously hard to measure, as the CBO notes frequently. ("A key disadvantage of the model-based approach is the considerable uncertainty about many of the economic relationships that are important in the modeling.")

Indeed, they are so hard to measure that when reporting on the stimulus, the CBO doesn't actually measure the outputs at all: Instead, the agency uses contemporary economic research to create models of how it believes stimulus spending is most likely to work. It then plugs actual spending numbers into those models and uses the results to report on the law's job creation. The CBO plugs the spending into a model that assumes, as a matter of course, that many types of federal stimulus spending saves or creates jobs. The agency then reports its results…which, not surprisingly, always say that federal spending saves or creates jobs.

These reports don't measure real-world results, and don't provide any check on the real real-world effects of stimulus, as Elmendorf himself confirmed when some of the early reports were released. If it were in fact the case that no jobs had been created, or only a very small number of jobs, the CBO's reports wouldn't reflect that.

What these reports show is that having looked at the available research, the CBO's analysts think that in the most plausible macroeconomic model, federal stimulus spending creates some jobs for some people for some amount of time. This does not strike me as particularly objectionable: Spending roughly $800 billion, much of it on projects that require some type of hiring, is likely to fund employment for some number of people, at least temporarily.  

But how many people? Well, that's less clear. The CBO certainly doesn't seem very confident in its ability to answer that question. The Dionnes of the world will trumpet figures like "as many as 3.3 million jobs," but what they rarely add is that 3.3 million is the high end of a very large range. On the other side of the estimate, the CBO estimates that the stimulus may have created as few as 600,000 jobs—a low-end estimate that has actually gotten lower over time; the previous low end was about 1.2 million jobs.

This tells us two things: First, the size of the range suggests that CBO's confidence that it can accurately predict or understand the employment effects of stimulus spending is not terribly high. Second, the fact that CBO has significantly revised its low-end estimate suggests that the agency's confidence in the job-creating power of stimulus spending has actually decreased over time.

Time, of course, is another relevant factor: The survey that Elmendorf cites asks only whether the stimulus increased employment in 2010 but doesn't say anything about the years that follow. And as The Post's report notes, but Dionne doesn't, there's far more debate about the long-term effects of stimulus. For one thing, there's reason to believe that many of those jobs are temporary: A Mercatus center survey of employers receiving stimulus funds found that many jobs paid for with stimulus money lasted only a few weeks or months but were counted as full-time equivalent jobs.

And the long-term burden of increased debt may ultimately be a net drag on the economy. In fact, that's exactly what Elmendorf and the CBO have reported is most likely: Last November, Elmendorf told legislators that the stimulus would be a "net negative effect on the growth of GDP over 10 years" and that barring additional action, "it would represent a drag on the economy."

So there you have it—the CBO, giving us what Dionne calls "the plain truth about the stimulus." In the long run, borrowing and spending $800 billion in order to temporarily juice the economy will end up dragging the economy down over the next decade. On second thought, maybe the singing and dancing can wait?