Are Reinhart and Rogoff Right Anyway?

By James Kwak

One more thought: In their response, Reinhart and Rogoff make much of the fact that Herndon et al. end up with apparently similar results, at least to the medians reported in the original paper:

Screen shot 2013-04-18 at 4.20.55 PM

So the relationship between debt and GDP growth seems to be somewhat downward-sloping. But look at this, from Herndon et al.:

Screen shot 2013-04-18 at 4.18.02 PM

Yes, we still have that downward slope. But two things:

1. There’s no cliff at 90 percent, which was the central finding of the original paper. This is the second sentence of that paper:

“Our main result is that whereas the link between growth and debt seems relatively weak at ‘normal’ debt levels, median growth rates for countries with public debt over roughly 90 percent of GDP are about one percent lower than otherwise; average (mean) growth rates are several percent lower.”

2. Aren’t we only supposed to be interested in empirical results that are significant? What the figure from Herndon et al. says, in their words, is this:

“Between public debt/GDP ratios of 38 percent and 117 percent, we cannot reject a null hypothesis that average real GDP growth is 3 percent.”

I find it hard to agree with Reinhart and Rogoff when they say, “We do not, however, believe this regrettable slip affects in any significant way the central message of the paper.”

17 responses to “Are Reinhart and Rogoff Right Anyway?

  1. Finally — a sensible analysis. Thank you.

  2. I am still mystified why one would conclude debt should be kept low to support growth, rather than, trying to keep debt low will result in slow growth.

  3. “We do not, however, believe this regrettable slip affects in any significant way the central message of the paper.” Even though economic growth at 90% debt-to-gdp goes from negative to positive when correcting that error. M’kay . . .

  4. The plot above shows that growth depends much more on factors other than government debt. I think Herndon et al. are correct that high debt to GDP doesn’t reduce growth based on the data given. I’d call the line drawn through the plot above meaningless — mean and median values for data with the degree of variation shown is pretty much meaningless.

  5. James Taylor

    I have many, many years of experience doing business related research and I have always found that when you are faced with incomplete data sets, you have two and only two choices. You can “create” data for the missing cells, or you can omit the the incomplete cells.

    If you chose to “create” data, your results will ALWAYS be suspect so I almost always use only the complete sets. If you do that with R&R’s data, you get this result:

    Debt/GDP Ratio GDP Growth Rate
    30% or less 4.6%
    30-60% 2.8%
    60-90% 2.7%
    90%+ -0.7%

    Now what were were talking about??

  6. The Herndon et al. paper finding flaws in the Rogoff and Reinhart work is _exactly_ how science is supposed to work. And, indeed, Rogoff and Reinhart’s ad hoc modifications to defend their position is a common feature of standard scientific process. Part of what is troubling in the lead-up to this tempest is that the ‘argument’ for Rogoff and Reinhart includes the ‘famous’ economists at a ‘great’ institution line — obviously not in any paper or formal discourse, but in the environment in which debate over economics and politics takes place in the US. Doing far more work than it should in the debate over the impact of debt on GDP (or any other major economic argument with significant policy implications today) is an argument from authority and the intellectual monoculture of mainstream economics and political science.

    Economists like Rogoff and Reinhart want us to believe they can be as rigorous as physicists (the golden idol of the sciences). _No_ department of physics or biology or chemistry is as dominated by a single school of political sentiment or by graduates of just five or six graduate programs in the way that economics and political science and business and law all are.

    If a physics department looked like the alumni club that leading economics departments do, alarm bells would go off about the quality of research. It is no accident that the critique of Rogoff and Reinhart came out of UMass Amherst. I’d be astonished if even _one_ member of Harvard’s economics department (or any other ‘leading’ department) dreamt of closely critiquing the Rogoff and Reinhart analysis.

  7. Moses Herzog

    Being the Economics addict that I am, I got the R&R book. I’m basically a Democrat, but I’ll purchase Republicans’ books if I feel they’re at least making the attempt to be neutral politically. I skimmed the graphs but have hardly touched the thing, sitting in a drawer looking brand new. Now I’m wondering if I should even bother reading the damned thing if they are twisting the numbers to get a HeadlineCatching/ideologyRationalizing result.

    R&R are trying to plead innocence, but it doesn’t come out clean in the wash. Ever notice how when your bank “f*cks up” and makes a “mistake” it’s always in the bank’s favor??? And all of these R&R f*ck ups “just happen to” fit their ideology perfectly and create a storyline that sells books. So all we have here is a replay of the Levitt and Dubner Superfreakonomics LIES and academic dishonesty of Chapter 5.
    http://delong.typepad.com/sdj/2009/10/yet-more-superfreakonomics-blogging-yes-i-know-i-know.html

    http://grist.org/article/2009-11-03-superfreakonomics-chapter-climate-change/
    Unfortunately for R&R, they have now lumped themselves together with Levitt&Dubner style trash to be consumed by the same ilk watching “Dancing With The Stars” now as they salivate over the ads telling them how they can “consolidate” the 8 credit cards they maxed out buying deep-fried sticks of butter at the state fair.
    http://i.dailymail.co.uk/i/pix/2011/08/11/article-0-0D64A14500000578-703_224x300.jpg

  8. Moses Herzog

    Uh oh…..Miss FreakyEyes just got a little freakier…..
    http://www.startribune.com/politics/statelocal/203515841.html

    Let’s go to the video replay, shall we?? From this:
    http://media.washingtonpost.com/wp-srv/politics/congress/members/photos/228/B001256.jpg
    to this:
    http://i.huffpost.com/gen/323987/MICHELE-BACHMANN-NEWSWEEK.jpg
    And if this goes to court and criminal charges of bribery, eventually this: http://www.theblaze.com/wp-content/uploads/2011/10/michele_bachmann_011.jpg

  9. Moses Herzog

    Can someone remind me, I got confused and lost my train of thought after seeing all the crowds of toothless rednecks getting a woody in their pants watching Sarah Palin discuss deep issues such as if her children could rot their teeth on soda pop….. was Michele Bachmann the darling favorite of the virtuous, conscientious, honorable, and morally sound Teabaggers’ Party?????
    If so, this story must either be the corroborated work of the “evil liberal commies from Kenya” (I know their existence for fact, as martian voices told me as I was listening to Alex Jones on the radio) or Satan: http://minnesota.cbslocal.com/2013/04/18/second-bachmann-aide-to-say-iowa-aide-wrongly-paid/
    I’d say it’s 55% chance the evil liberal commies from Kenya created this story and 45% chance it’s Satan’s creation. You decide.

    This message was Sponsored by former Texas Senator Phil Gramm and UBS bank’s “Derivatives and Swaps Smell like Lilacs” Joint Venture

  10. Moses Herzog

    Just try to remember folks, sometimes, there is poetic justice in this world:

    I’m just wondering if the dumb redneck obstetrician could ” ‘splain ” falling in a fountain like a jackA$$ as tactfully as he condescending jokes to a hispanic with 5 times his level of education.

    I guess it was about the same level of tactfulness.

  11. I would even caution using complete sets, with the underground economy’s numbers not added in. The true results could be off by more than “created data”. Without proper laws, the tax payer always foots the bill as we try not to make the same mistake twice.

  12. Moses, you just proved that speedy does kill.

  13. And this: “The shaded region indicat[es] the 95 percent confidence interval for mean real GDP growth.” How about the variation around the estimated mean? More or less your point #2, but the change in the estimated mean from Debt/GDP=40 to Debt/GDP=120 is minor compared to the variation about the estimated means. Arguing that the estimated change is meaningful is laughable. “Cannot reject the null hypothesis…” is far to diplomatic. “You’ve gotta be kidding.” is more like it.

  14. Re: Reinhart and Rogoff “This Time Is Different” Slamfest

    [Their book was published in July 2011, the criticisms arrive in 2013] “….Reinhart and Rogoff were in the spotlight after researchers discovered that their co-authored book This Time is Different had significant basic computation errors that, when corrected, undermined the central claim of the book that too much debt causes recession. Rogoff and Reinhardt defended their numbers and claimed that their fundamental conclusions were accurate, despite the errors.”

  15. Economics is not a science. You can skew data to fit whatever narrative you want. This is disgraceful. The Nobel prize in Economics should be dropped, it was not part of the original Nobel prizes left in the will of Alfred \Nobel. This was added later by the bankers in Sweden and has nothing to do with the actual Nobel prizes. This gives credibility to so called Nobel laureates in Economics and raises the profile of Economists. The findings in a paper should not be accepted as gospel until it has been replicated and or peer reviewed. How many lives have been damaged by politicians implementing this flawed research?

  16. Originally would not release the spread sheet? Excel spread sheet errors at this level (could it be they wanted to support their lies)? And the current leaders of the republicans and GOP have been took it, hook, line and sinkers! WOW! that will the excuse be for this, their results are still good because of magic? Hey, they love that line.

    http://www.colbertnation.com/the-colbert-report-videos/425748/april-23-2013/austerity-s-spreadsheet-error