In this morning’s Gray Lady, Bill Keller wrings his hands over the state of economic policy discourse in an age of hyperbole and half-truths. His complaints aren’t without merit. But he falls victim to the very behavior he seeks to condemn.
About the American Recovery and Reinvestment Act (ARRA), Keller writes:
The Web site PolitiFact, the Pulitzer-winning fact-checking service, recently did a thorough debunking of Republican claims that Obama’s 2009 stimulus program created, quote, “zero jobs.” In fact, the checkers established, using still-trustworthy sources like the Congressional Budget Office, that the stimulus created or saved a couple of million jobs. Case closed? No, the Republicans just went on repeating the claim.
“The talking points drive the discourse,” said Bill Adair, the editor of PolitiFact. “They repeat the talking points so often I think they start actually believing them.”
But Keller is fighting a talking point with a talking point. And while his may be less untrue than the one he seeks to debunk, it’s still not totally accurate.
As Peter Suderman has ably reported, the CBO’s estimates of job creation are just that—estimates. All that the CBO has really done in their quarterly reports on ARRA (the most recent of which came out last week) is to re-run their pre-existing models, tweaked slightly, using new data about actual spending. But the underlying assumptions about stimulus remain basically the same. In other words, before ARRA passed they said it would create X jobs per Y dollars spent. And now they’re saying that since we’ve spent X dollars, we must have created Y jobs.
This isn’t to fault the CBO, and they have been abundantly clear about the limitations on their ability to accurately assess the impact of ARRA. But it’s misleading to suggest that the CBO has done empirical, ex post analysis of actual job creation. As CBO chief Douglas Elmendorfer put it in 2010, “if the stimulus bill did not do what it was originally forecast to do, then that would not have been detected by the subsequent analysis.”
Finally, the CBO has never given a precise number of jobs created, but rather a range. The high estimates range from 900,000 jobs created in 2009 to a high of 3.2 million jobs in 2010. But the lower end estimates run from 200,000 jobs to 700,000 jobs. That’s quite a range, and only citing the favorable estimates is misleading.
It is, in other words, a talking point.