The Perils of Wonkery

April 16th, 2013  |  Published in Politics  |  30 Comments

The economics blogosphere is buzzing about the errors that were recently exposed in an influential paper by Carmen Reinhart, Vincent Reinhart, and Kenneth Rogoff, which claimed that countries with high levels of debt tend to have slower economic growth. See Mike Konczal for the summary or here for the full paper by Thomas Herndon, Michael Ash, and Robert Pollin. In short, the original Reinhart-Rogoff paper had three significant problems, ranging from cherry-picking data, to dubious weighting schemes, to—most embarrassing of all—an Excel spreadsheet error that accidentally left out several crucial data points.

The reaction of the left-wing peanut gallery (at least to judge by my Twitter feed) has been to ridicule liberals for caring about this at all. Obsessing over the analytical missteps in this paper reeks of the preoccupation with having correct and empirically supported arguments, while ignoring the importance of power and ideology. For while this new critique of Reinhart-Rogoff just now became possible because they finally made their original data available, plenty of people pointed out earlier that the whole analysis rested on shaky conceptual foundations. It used a correlation to assert that high debt to GDP ratios lead to slower growth, ignoring the much more plausible theory that the causal order was the opposite, with slow growth leading to increasing debt loads. If the political elite in Washington failed to heed these criticisms, it wasn’t because they were unaware of them, but because the claim that debt leads to slow growth fit a deficit hysteria that was already entrenched. In other words, Reinhart-Rogoff was being used as rhetorical cover for a pre-existing position, not as an actual empirical aid to decision-making.

But rather than dismiss Excel-gate as much ado about nothing, maybe we can use it as a cudgel against the pernicious rise of the “policy wonk” as a model for journalism. As Bhaskar Sunkara noted in a recent article for In These Times, the wonk is a new iteration of American journalism’s obsession with “objectivity”, in this case filtered through the predilections of the “technocrat, obsessed with policy details, bereft of politics, earnestly searching for solutions to the world’s problems through the dialectic of an Excel spreadsheet.” The Reinhart-Rogoff revelations do more than just reveal the folly of relying on the wrong spreadsheets—they expose the shallowness and dishonesty that pervades much of the wonk-journalist milieu.

To return to a familiar whipping boy, let’s review the initial reaction to Reinhart-Rogoff’s paper last summer, over at Ezra Klein’s Wonkblog at the Washington Post website. A post by Suzy Khimm was entitled “Study: Long-term deficits are linked to 24 percent lower growth”, and it simply repeated the study’s claims without critique. For added truthiness, the post is embellished with a graph reproduced from the paper, demonstrating the difference in GDP growth between a group of low-debt and high-debt countries.

As we now know, that graph was based on erroneous data marshaled in support of a logically flimsy premise. But while the data errors wouldn’t be revealed for months, not everyone was fooled. Matt Yglesias—a writer often lumped in with wonks like Klein— dismissed the Reinhart-Rogoff paper as “confused correlation-mongering”, on the grounds that the reverse causal story about debt ratios and growth was far more plausible. (Incidentally, the way Yglesias approached Reinhart and Rogoff’s claims demonstrates how poorly he fits the mode of the Ezra Klein-style wonk-journalist. In contrast to the wonky preoccupation with empirical studies and pretty graphs, Yglesias has argued that “evidence is overrated”, and he often offers positions based on his own ideological predilections and reasoning from first principles.)

This is an approach that can get you into trouble in other ways, but it does sidestep one of the big problems confronting the wonk. The function of the wonk is to translate the empirical findings of experts for the general public. And he is supposed to be distinguished by an immersion in the details of studies and policy papers. But if the wonk wants to cover a wide range of subjects, they will necessarily have far less expertise than the people whose findings are being conveyed. Hence it becomes necessary to make a concealed argument from authority. When Wonkblog presents the findings of Reinhart and Rogoff without comment, they are implicitly telling us, “trust these people—they’re famous academic economists”. This is because they don’t have the ability to do what people like Paul Krugman did, and actually assess the correctness of the famous economists’ claims.

Performing this con on the public is dangerous enough. But insofar as the wonk gets high on his own supply, and starts to trust the findings of congenial academics without verifying, the temptation to take shortcuts can be overpowering. It’s easy to read the abstract and the conclusion of a paper and trumpet its findings, without looking too closely at whatever equations or models lie in between. This isn’t actually any more hardheaded than relying on one’s feelings, but it’s an appealing way to give one’s prejudices a fact-like veneer. That’s what seems to have happened to Ezra Klein’s understudy Dylan Matthews, who uncritically accepted some claims about the effect of teachers’ strikes on student achievement, which Doug Henwood was able to easily pick apart by actually reading the studies he was referring to.

This wouldn’t be so aggravating if the wonks were more open about their ideological orientation. If Yglesias promoted a study finding a relationship between strict occupational licensing and slow economic growth, I’d know to look carefully into the details, since his pre-existing views on that subject are well-known. With the wonks, though, close reading and an understanding of ruling class ideology are required to extract the political orientations that are guiding their judgment.

As the policy wonk has risen in prestige, we seem to have reached the point where this entire class of commentators is highly susceptible to what I’ll call “Charlie Rose disease”. It’s a malady named for the host of the eponymous TV show, who has always impressed me with his ability to convey an impression of knowledge and gravitas to his viewer. If you watch his show and actually listen to him talk, you’ll quickly notice that Rose is a shallow thinker even by television standards, and generally quite ignorant about the things he interviews people about. But everything about him—from his face to his cadence to his posture to his austere black-background set to having his show on public television—works together to produce the image of intellectual seriousness, even more than for most TV news hosts.

And so it is with the wonk—he needs to appear to be deeply knowledgeable about a wide range of obscure and technical subjects. But this entails concealing both one’s ideological biases and one’s substantive lack of knowledge, and relying on the borrowed prestige of academics and experts. In doing so, the wonk becomes the conduit for the experts, or more exactly a crucial means by which their authority is reproduced. The wonk takes the expert’s pronouncements at face value because they are serious, mainstream figures, and the fact that journalists do this reinforces their seriousness and mainstream-ness. One could hardly devise a better way of policing ideological boundaries and maintaining the illusion that the ruling ideology is merely bi-partisan common sense.

The unraveling of the Reinhart-Rogoff “fact” about debt and growth was only unusual because the supporting research was unusually sloppy. In that sense, critics are correct that there’s nothing particularly special about this one case. But the very absurdity of the episode makes it useful as a means of unmasking the entire corrupt enterprise of policy wonk journalism and its “just the facts, ma’am” pretensions.

  • http://profiles.google.com/lumpoflabor Tom Walker

    Back in the days of C. Wright,Mills, we called these technocratic critters, “crackpot realists.” Nothing has changed..

  • Pingback: Peter Frase on Reinhart-Rogoff debacle | MattBruenig | Politics

  • http://ralphhaygood.com/ Ralph Haygood

    As a biologist, I’ve occasionally been interviewed by journalists for stories about my research. It has usually been at least mildly uncomfortable, because the journalist has usually been at least one of two things: (1) extremely ignorant of biology (e.g., the newspaper reporter who didn’t know that pollination is a form of sex); (2) eager to turn my work into a sound bite (e.g., transgenic crops: GOOD or BAD?). However, it usually isn’t too hard to be patient with such journalists, because most of them know they don’t know much about biology, and they’re willing to learn if you’re willing to teach them. I imagine the situation is much worse for economists and political scientists, because so many people, including journalists, believe they know so much about economics and politics.

    Of course, it may well be that “wonks” like Ezra Klein don’t even bother to talk with the people whose research they’re discussing or, even more importantly, with other people in the field who might point out its flaws.

    • Glib Fighter

      As an economist, my take is that lots of economic reporting is hideous, e.g., “The stock market went up today because ____.” As I filter out the nonsense in such journalism, one of my serial worries is that I can’t do such filtering wrt to science writing. Knowing how bad so many economics journalists are at, e.g., the NYT, my running assumption is that the science journalists at such media outlets are of the same low quality.

      The work they do is akin to lots of university/college-level instruction, which far too often is done by faculty who teach the text they’re assigned but are thoroughly incapable of finding mistakes/problems in such textbooks. The non-thinking parroting of textbooks done by these PhD-holding instructors is strongly correlated with their failing to carry out any sort of serious peer-reviewed research.

      The Ezra Kleins and Matt Ys are the journalistic analogues of these PhDs. The record also shows that Paul Krugman has been a major abettor of EK. This was especially so during the buildup to the passage of the ‘Patient Protection and Affordable Care Act’ of 2010.

      • F.S.J. Ledgister

        The big problem is that journalists (and I was one a quarter century ago) need a nice simple monocausal explanation that can fit into a two-sentence lede. The world doesn’t work that way, as all social scientists know.

  • http://www.facebook.com/andreapastor13 Andrea Pastor

    This jives so entirely with Nate Silver in ‘The Signal and the Noise’ that if you haven’t read it (doubtful, I’m sure) you should pick it up immediately. Nice article, btw.

  • Freddie_deBoer

    “In contrast to the wonky preoccupation with empirical studies and pretty graphs, Yglesias has argued that “evidence is overrated”, and he often offers positions based on his own ideological predilections and reasoning from first principles.”

    Which, of course, he uses to generate whatever particular conclusion suits his given whim.

  • Pingback: In defense of wonk-journalism - beyond the times » beyond the times

  • JW Mason

    The reaction of the left-wing peanut gallery (at least to judge by my Twitter feed) has been to ridicule liberals for caring about this at all. Obsessing over the analytical missteps in this paper reeks of the preoccupation with having correct and empirically supported arguments, while ignoring the importance of power and ideology.”

    Really? People are saying this? Bizarre. Mind sharing some examples?

  • Charles Henry Clarke Reinhardt

    One thing that still eludes me however, is the idea that ideological and philosophical arguments can’t or shouldn’t serve as a basis for a tightly argued form of socialist wonkery that simply takes into account a different set of quantifiable variables. I’m not sure I buy that that’s just “playing into the neoliberal mindset.”

  • MarshallG

    Case(s) in point. To be fair, the Atlantic story is a bit more tempered than the Slate one, but those atrocious bar graphs are completely misleading and, actually, don’t even support the author’s claims. The “sample” for the original study, which underpins the claim that “No Big Deal, but This Researcher’s Theory Explains Everything About How Americans Parent”, is described thusly:

    “In each cultural site, we recruited a sample of 60 families with target children divided evenly into five age-groups balanced for birth order and sex: 6 months, 18 months, 3 years, 4.5 years, and 7 to 8 years. The sample families, recruited mostly through community networks, were broadly middle-class, with one or both parents employed and no major health problems; most of them were nuclear families with both parents present in the home; and parents in each sample were all native-born to that culture.”

    http://www.slate.com/blogs/how_babies_work/2013/04/10/parental_ethnotheories_and_how_parents_in_america_differ_from_parents_everywhere.html

    http://www.theatlantic.com/international/archive/2013/04/how-parents-around-the-world-describe-their-children-in-charts/274955/

  • Tim D.

    To defend Ezra Klein a tiny bit, this would be a stronger argument if he were actually pro-austerity. But he’s not. Substantively he seems to have stayed more toward the Krugman end of things. He has clearly rebranded himself as a WaPo centrist, but the content has not shifted as much as the tone has. (His underbloggers are a different story tho.)

  • Pingback: Rogoff, Reinhart y la hoja de cálculo que destruyó el mundo

  • Pingback: First Links — 4.17.13 » First Thoughts | A First Things Blog

  • Pingback: Wednesday, Oh, Wednesday | Gerry Canavan

  • Pingback: “The Perils of Wonkery” | Minjae Park

  • Pingback: Links 4/17/13 | Mike the Mad Biologist

  • Pingback: There’s a reason TIE is so focused | The Incidental Economist

  • Pingback: Counterparties: R-squared regression analysis | Felix Salmon

  • Jake H

    It’s a bit strange to blame Klein (or journalism, or blogging) for reporting on a paper that was published in the American Economic Review. I agree that Klein and his fellow bloggers are probably not equipped to assess the validity of findings like this, but the people (peer reviewers, editors) who SHOULD be equipped to do so dropped the ball.

    We’ve all seen plenty of examples of reporters trumpeting research that hasn’t been properly vetted, in part because it’s easier to find interesting or newsworthy results if you look at such research, and in part because many of them probably don’t know what the peer review process is or why it should matter. That’s a problem. This is a different problem – a major peer reviewed journal published work that turned out to have been carelessly done if not willfully fudged. Academia can’t blame journalism for that.

    • adam3smith

      this is a misunderstanding of AER. R&R was published in the papers&proceedings section of the journal, which is not peer reviewed – it contains selected papers presented at the January AEA meeting. Given recent-ish trends in economics which reifies identification (i.e. establishing actual causality) to a fault this would never have been published in AER via peer-review.

  • JK

    To be fair, I noticed last night, at the bottom of a post linking some excellent Colbert videos on the subject, that Wonkblog linked to a bunch of critical reporting they did on R-R in the past: http://www.washingtonpost.com/blogs/wonkblog/wp/2013/04/24/even-stephen-colbert-is-piling-on-reinhart-rogoff/

  • http://twitter.com/PoliticsDebunk Politics Debunked

    re: “As we now know, that graph was based on erroneous data marshaled in support of a logically flimsy premise”

    It is unfortunate that many people writing about this don’t bother to notice that the paper critiquing R&R also shows a correlation between higher debt and slower growth.. it just isn’t as large. There are also a number of other papers that independently show the same thing.

    re: “on the grounds that the reverse causal story about debt ratios and growth was far more plausible” . It is true that yes, politicians often engage in poor policies and don’t cut back on spending when growth doesn’t bring int the revenue increase they planned on. However there is a lot of work on “crowding out” since it is very plausible to those who actually know something about the topic that higher debt leads to slower growth.

    Regardless of the attempts for people to try to obfuscate simple realities, and investor can only do one of a limited set of things with each $: lend it to government, invest in the private sector within the country, invest it outside the country, or stuff it in a mattress. If the government doesn’t borrow it they are most likely to invest it privately within the country, and usually in advanced economies private investments in growing companies lead to more growth than giving the money to government to say build a “bridge to nowhere”. (an economy lacking basic infrastructure might be a different story).

    • Alf Hickey

      I’d like to see some of that empirical evidence for crowding out.

  • Alf Hickey

    It seems like you believe a wonk is a jounalist who blogs graphs, This is a very strange definition. You are simply complaining about journalists not wonks. A wonk just means a policy expert, it doesn’t desperate acedrmics from journalist, but it does imply a tedious punt of technical competence. Every liberal wonk blog has been complaining about r and r for years.

    Doug Henwood could be considered a wonk because he is very knowledgable about the inner workings of the financial system. For this reason he is one of the most respected leftists in liberal wonk circles. Contrast him with graeber who was rejected for getting details wrong.

  • Anthony Juan Bautista

    excellent. really great writing….

  • Pingback: Somewhere else, part 52 | Freakonometrics

  • Pingback: Rapide retour sur Reinhart et Rogoff | Polit’bistro : des politiques, du café