CBO: Don’t Shoot the Messenger and Why They Face “Greenspan Risk”

March 2nd, 2014 at 12:45 pm

All of the sudden there’s a spate of pieces casting a jaundiced eye in the direction of those pristine nerds over at the Congressional Budget Office:

–Dean Baker riffs off of a Zach Karabell piece enumerating big mistakes the agency has made in the past and how they’ve constrained policy advances.

–Floyd Norris makes a reasonable point re an inconsistency in the budget agency’s current projections (i.e., they project slower growth but higher interest rates); Krugman agrees.

–I weighed in the other day with a critique of their minimum wage job loss estimate; not that they got it wrong, but that their weighting of the evidence looked off to me and others.

I think the problem is actually less with CBO, whose work represents state-of-the-art economic thinking, while largely maintaining a non-partisan rep in a town where that’s really hard to do, and more with the economic thinking itself.  There’s a risk of shooting the messenger here.  (If you insisted, I’d come up with a different critique: their reports are increasingly impenetrable and difficult to digest; assumptions are increasingly hard to follow–you have to track back through links like a Dan Brown mystery novel; the reports don’t hang together like they used to–at least to my eyes, they increasingly read like a movie script written by a large committee).

Take the finding to which Norris/Krugman object.  As far as I can tell, CBO’s just making the conventional assumption that once we’re back to full employment, interest rates will revert to more normal levels.  The fact that potential GDP will be lower doesn’t change that result.  Norris and Paul reasonably argue that if demand is weak then interest rates will be lower than CBO expects.  But CBO doesn’t expect demand to be weak.  Their markdown of growth comes from the economy’s supply side.  (BTW, I’ve got a piece on this coming out maybe tomorrow on the NYT Economix blog—a call for “reverse hysteresis!”—IKR!—exciting!)

Many other assumptions made by CBO that drive some, particularly on the left, nuts, are again standard issue, textbook economics: the idea that deficits crowd out investment and raise interest rates in expansions, the behavioral reactions to tax changes at the margins (as in their recent ACA analysis), the job loss effect from the minimum wage increase, the assertion of quantifiable outcomes decades away, the “straight-edged projections” that missed the bubbles, followed by the subsequent building of the now lower trend growth into future forecasts (the focus of my piece tomorrow)—all of these represent the current state of knowledge of mainstream economics, which CBO practices quite meticulously.

Thus, the problem is twofold: first, their assumptions (which is to say the state of contemporary economics), and second, their power.  It’s the same problem Alan Greenspan had when everyone thought he was the economy’s Maestro.  He too was operating from flawed assumptions (and in his case, ideologies) which led him—along with multitudes of others—to miss bubbles inflated by vastly under-priced risk while assuming markets would self-correct.

And he too was considered infallible.  You know the old DC joke: the tourist couple goes up to a vendor and asks, “how much does a hot dog cost?”  The reply: “whatever CBO says it costs.”  (I know—we’re such laugh riots down here…)

Again, not their fault—listen to CBO director Elmendorfer and he’s generally quite straightforward and realistically cautious about the uncertainty of their findings (though I’d encourage them to lean more on ranges and less on point estimates).  But once those findings get injected into the fractious politics of the moment, they become immutable truth, and that is both wrong and dangerous, as Karabell points out, for the advancement of useful policy.  That’s the Greenspan risk that CBO faces today: when policymakers assign too much credibility to highly uncertain forecasts and projections, the policy process will suffer.

What’s really needed here is a much more realistic skepticism about the ability of economics, as practiced by CBO or anyone else, to accurately forecast trends and behaviors next week, much less ten and twenty years from now (a point estimate of the labor supply reaction to the ACA in 2024??…really??).  I don’t expect CBO to change, other than to continue to slowly and carefully evolve along with the economics’ literature.  So it’s going to be up to the rest of to read their reports with a big saltshaker close at hand.

Print Friendly, PDF & Email

9 comments in reply to "CBO: Don’t Shoot the Messenger and Why They Face “Greenspan Risk”"

  1. Nhon Tran says:

    Thank you.
    It was reported widely in the press that Ms Janet Yellen agreed with the CBO’s conclusion on the minimum wage.

  2. Michael says:

    What I fail to understand is how some economists can actually use models and historical data to make semi-accurate predictions while other economists are essentially formulating and promulgating ‘flat Earth’ theories. The hallmark of a specific field is how much of its past information and analysis leads to future prediction and understanding.

    It would be as if there were two medical professions: one advocating for a microbial or biological root for diseases and the other advocating for a demonic, spiritual, personal root, and the one’s advocating for biological root causes keep being regulated to blog posts and newspaper/online columns. One of the major problems with economics is that it doesn’t explain how its own field is having this dialogue in the first place. Something fundamental within economics (in particular) and education (in general) is undermining its own usefullness, and the longer the discussion continues to revolve around CBO personel, job reports, GDP graphs, etc, the longer the fundamental problem will remain unaddressed. Graphs don’t have economic incentives to lie or misinform, people do.

    • mitakeet says:

      Hear hear!

      Related (at least to me) is this article:


      If science has degenerated to the point where computer generated nonsense is being published in respected journals, how can we rationally expect a field such as economics (perhaps not quite as fuzzy as psychology, but not too far either) to be immune from gibberish, even if that gibberish was produced by humans unintentionally. I was talking with a friend just yesterday about what I see as the steadily eroding (really, eroded) state of scientific comprehension of statistics. Since a couple of generations ago, nascent professors have been poorly taught/mentored in experimental design, yet got tenure anyway. Of course they are going to teach their students what they learned, obviously it worked, so the lack of comprehension feeds back on itself. In just a few decades no one really has any comprehension of good experimental design (meaning controls and proper statistical analysis to know you have collected meaningful data), so journals (whose reviewers, of course, are part and parcel of the same effect) publish meaningless papers. Not such a huge step, then, to publish computer generated gibberish, since so many are already human generated gibberish.

      Sometimes it amazes me that we, as a species, ever get anything useful done.

  3. Dave says:

    I would argue that the prediction of job losses/gains from a minimum wage hike by the CBO is an inappropriate use of the CBO.

    The CBO is supposed to deal with budgets and real numbers. To the extent they need to predict budgets, they need to do economic analysis. But the question about minimum wage wasn’t a budget question with an objective answer. If the bill under question doesn’t directly affect budgeting, the CBO should stay out of the issue and let the chips fall where they may.

    This was an inappropriate use of the CBO in my view.

  4. Robert Buttons says:

    “The curious task of economics is to demonstrate to men how little they really know about what they imagine the can design.” –Hayek

  5. Mark Jamison says:

    It may be that CBO is a casualty of the times. There is no such thing as nonpartisan any more, the Tea Party wing of the Republican Party and their penchant for “purity” creates an environment where no facts or methods can be agreed upon, everything must be fed through an ideological filter.
    In such an environment the only way CBO can be seen as nonpartisan is by becoming obtuse and impenetrable. There are some economists who do seem together more right than wrong. At the risk of being accused of being a sycophant I’d suggest that you, Dr. Bernstein, Krugman, Baker, and a few others seem to have some consistently accurate accounts and scenarios. On the other hand some economists on the Right like Mankiw and Holtz-Eakins seem to twist themselves in knots trying to be apologists for ideologies that are pretty clearly flawed.
    I’ve seen this sort of thing in other agencies that are primarily technocratic, the PRC for example. When you live in a world where one side refuses to stipulate any facts it makes it difficult if not impossible for an organization with an empirical and technocratic focus to remain an honest broker of information. The only answer from an institutional point of view is to hedge and obfuscate.
    Mr. Greenspan was a different problem. He was playing the role of a Grand Wizard. In order to appear omniscient he always needed to be cloaked in mystery, as if his gifts were beyond our ordinary mortal vision. He was elusive as a means to create the impression of being impressive – the Great and Powerful Oz (never mind the man behind the curtain).

  6. Andy Harless says:

    Either you or I appear to be misunderstanding the Norris/Krugman criticism. You seem to think it’s about the level of the output gap. It seemed pretty clear to me that it was about the growth rate once the output gap is closed. It is implicit in the Norris/Krugman argument (as I understand it) that this is a supply side issue, not a demand side issue. A more slowly growing labor force (along with more slowly growing productivity) means that the equilibrium real interest rate will be lower. I believe this result is typical even in models with full employment. If the CBO is assuming a lower output growth rate at full employment (compared to what we saw in the past), then it should be assuming a lower real interest rate at full employment (compared to what we saw in the past). I don’t think appealing to convention is an adequate defense here. Does conventional growth theory say that the real interest rate is independent of the growth rate?

    • Jared Bernstein says:

      Go back and look at prior versions of these same projections–you’ll see they’re predicting faster growth and higher interest rates. Like I said, this is what pops out of their model and it’s standard macrostuff assuming full employment.

  7. Kevin Rica says:

    This whole discussion reminds me of Major Major Major’s father in Catch-22 who disapproved of loose women who turned him down.

    Many people who dislike one particular CBO study complain only to embrace the next study that they happen to agree with.

    Let’s take that silly study on “Comprehensive Immigration Reform” (no immigrant left behind). It only compared legalizing the same number of immigrants to not legalizing them. It never compared the “Immigration Reform” bill to lower levels of immigration. It then assumed that lower wages caused by legalization would cause more investment that would restore wage rates (over the fullness of time) to their original level.

    So much to mock!