Skip to main content

View Diary: Wall of Shame: Republicans Lose Intellectual Cover for Austerity/Spending-Cut Policies (87 comments)

Comment Preferences

  •  OK, I was focusing on the link in this paragraph (1+ / 0-)
    Recommended by:
    Sychotic1
    And now, we have this UMass study that highlights numerous issues with the Reinhart & Rogoff study, including, shockingly, the revelation that there was an Excel coding error in their data! Imagine that, the study that gives right-wingers in Europe and the US academic and intellectual cover to justify wildly anti-social policies, has essentially been DEBUNKED!
    that emphasizes the "coding error" but that turns out not really to be the major problem at all - can that paragraph be fixed / deleted?
    •  Multiple errors. (6+ / 0-)

      Try this diary. R&R also made inexplicable exclusions and illogically weight the data due to their method.

      “Birds…scream at the top of their lungs in horrified hellish rage every morning at daybreak to warn us all of the truth. They know the truth. Screaming bloody murder all over the world in our ears, but sadly we don’t speak bird.” Kurt Cobain

      by RadicalParrot on Wed Apr 17, 2013 at 05:29:17 AM PDT

      [ Parent ]

      •  I'm not an academic (4+ / 0-)
        Recommended by:
        Sandino, Sychotic1, Nailbanger, quill

        But is the preliminary assessment of actual academics and economists that their errors were of a type, nature and degree to constitute serious and perhaps even willful professional malpractice that could lead to their being brought up on ethical charges, answering to a review board, and even be fired? If there were so many errors, this seems unavoidable.

        "Liberty without virtue would be no blessing to us" - Benjamin Rush, 1777

        by kovie on Wed Apr 17, 2013 at 05:52:43 AM PDT

        [ Parent ]

        •  It's way too early to tell (5+ / 0-)
          Recommended by:
          kovie, ColoTim, Nailbanger, quill, ferg

          But, early indications fromDean Baker and Krugman here and here show that there is really no explaining their way out of this.

          Those were posts Krugman made AFTER R&R issued a "response" which was widely viewed as just pathetic, and if you check out the FT link in the Update I posted, it's the authors of the Amherst study really digging into R&R.

          You won't really find anyone coming out in public on a blog or anything demanding heads roll and calling them corrupt. That sort of aggression, especially in the public sphere, doesn't come from intellectuals/academics. We're talking about two esteemed tenured Harvard economists and a paper that has been really hailed for years. This will be a process that takes time to play out fully, not days, not weeks...

          The first stage is the civil, intellectual back and forth...and then as it begins to sink in, people will look into whether other motives are at play and whether this was really gross malpractice.  

          Give it time, let the back-and-forth play out, and see where it goes.

          Deficits don't matter, jobs do.

          by aguadito on Wed Apr 17, 2013 at 07:23:50 AM PDT

          [ Parent ]

          •  I understand that this is how academia works (1+ / 0-)
            Recommended by:
            aguadito

            and agree with it. But in the end if they're found to have engaged in egregious and willful misconduct, they must pay a proportionate and exemplary price for it, or else what's the point of even having academic standards?

            "Liberty without virtue would be no blessing to us" - Benjamin Rush, 1777

            by kovie on Wed Apr 17, 2013 at 07:41:05 AM PDT

            [ Parent ]

            •  they will suffer majorly (1+ / 0-)
              Recommended by:
              Nailbanger

              and their reputation has obviously taken a huge hit.

              but all those guys who pushed the housing bubble -- do you see any of their careers damaged? professionals in economics and finance don't always get held accountable, or even corporations for that matter. even totally failed CEO's like the Ferrino from HP goes on to have illustrious careers. A failure as a CEO just means you've at least been a CEO and that qualifies you to be all in all sorts of positions lol.

              check out the link I added in the diary from HBR that offers a little discussion about this:

              Reinhart, Rogoff, and How the Macroeconomic Sausage Is Made

              Deficits don't matter, jobs do.

              by aguadito on Wed Apr 17, 2013 at 08:06:19 AM PDT

              [ Parent ]

              •  Academia has a different set of standards (1+ / 0-)
                Recommended by:
                aguadito

                than do politics, business or finance, whose standards are set and enforced by the people at their top, whereas academia is still mostly peer-regulated. It's one of the few remaining institutions that still does this for the most part.

                "Liberty without virtue would be no blessing to us" - Benjamin Rush, 1777

                by kovie on Wed Apr 17, 2013 at 08:18:44 AM PDT

                [ Parent ]

                •  very true (1+ / 0-)
                  Recommended by:
                  New Rule

                  I just feel like  of all academic disciplines, Economics really is one of the more corrupt ones.

                  It really has a deep incestuous relationship with Business and modern corporatism.

                  We're talking about top academics that run off to start hedge funds and stuff.

                  and HAP (in the Financial Times op-ed) call the methodological makes "defensible" -- so i dont know. i don't really think this will blow up to be too much and I think the top Harvard economists will remain tenured and gainfully employed doing research for the private industry :>

                  Deficits don't matter, jobs do.

                  by aguadito on Wed Apr 17, 2013 at 08:23:29 AM PDT

                  [ Parent ]

    •  Lots more wrong (6+ / 0-)

      Go to the diaries here yesterday. (I'm too tired to dig out the links.) One key problem is often called "the unit of analysis.") As Konczal points out, they counted a few years for (maybe) NZ and averaged this with (maybe) 19 years of UK data. They also used "unweighted" means, where data from small countries - with small economies - counted the same as the larger countries. I believe the data they "missed" came from just post-WW2, where many countries with large debt/GDP fractions also had booming economies.  

      If there's fraud, in addition to honest mistakes, in their data analysis, it will show in where they chose the cutoff points.  It's long been argued that splitting data into high, medium, and low groups, especially with attention to cutoffs that amplify the differences authors want to claim, is bad practice.  If data points just below their 90% level, in addition to "Excel coding errors," show higher than 90%-level average GDP growth, it is entirely possible that they picked cutoffs opportunistically.  

      Your joke about Harvard students unable to count and MIT students unable to read is new to me and funny. But, if you can recall or research Cyril Burt, you'll see that there are data forensics that can be used to separate mistakes from misrepresentation. I have no doubt that Harvard will be taking a close look at what went on here.  

      Even if it turns out these were all honest mistakes, you have to be simultaneously saddened and also have a wry smile --- economists are supposed to be competent with data, right?

    •  It was more than just the coding error (5+ / 0-)
      Recommended by:
      Sychotic1, ColoTim, Nailbanger, quill, ferg

      And according to Konczal the Excel SNAFU was actually enough on its own to make the data bad, but there were other issues including "selection bias" and "unusual weighting" -- basic methodological problems.

      As others have pointed out, Konczal has the DailyKos full review of what was wrong with the study:

      http://www.dailykos.com/...

      Please go there and recommend his story.  He's definitely a Progressive Economist you want to keep your eye on.

      Deficits don't matter, jobs do.

      by aguadito on Wed Apr 17, 2013 at 07:04:25 AM PDT

      [ Parent ]

    •  I just double checked (1+ / 0-)
      Recommended by:
      ColoTim

      the link in that paragraph -- it works fine. Business Insider is actually a fairly progressive publication for Finance-world standards.

      And it's written by Konczal himself...goes into detail about the errors.

      The "Excel coding error" is just sort of the head-line point that grabs people's attention, but it was significant enough to fubar the whole shabang too.

      Deficits don't matter, jobs do.

      by aguadito on Wed Apr 17, 2013 at 07:13:09 AM PDT

      [ Parent ]

      •  I realize the link works, that was how I (1+ / 0-)
        Recommended by:
        aguadito

        was able to read it.

        My point was that the "coding" error was rather insubstantial in the larger scheme of things - there were at least two larger issues.

        Plus, is simply not being able to use excel competently really even a "coding" issue?

        •  i just saw (0+ / 0-)

          a link from the Harvard Business Review by Justin Foxthat goes into more detail if you're interested in something other than Konczal's review.

          The Excel coding error alone accounts for the negativity of the growth rate in the >90% part in the R&R study.

          The other issues are the "methodological" ones I reference and link to in the diary -- as mentioned above the selective exclusion and unconventional weighting of country data.

          Deficits don't matter, jobs do.

          by aguadito on Wed Apr 17, 2013 at 07:41:19 AM PDT

          [ Parent ]

          •  As I pointed out above, the "coding" (3+ / 0-)
            Recommended by:
            aguadito, Nailbanger, ferg

            error resulted in a change from -0.1 to +0.25%

            While the real value should have been +2.2%  - IOW it accounted for about 15% of the discrepancy (i.e,. 0.35 relative units of the 2.3 gap).

            The bigger point here is that in theory at least this could have been a simple, inadvertant error.  While the other problems (selective use of data, non-standard weighting) are much more serious and could not happen from a simple mistake - instead they were deliberate and smack of fraud.

            Just saying, they seem like much more worthy candidates to highlight and emphasize.

            •  good point (1+ / 0-)
              Recommended by:
              Roadbed Guy

              I've included the mention of those errors in the paragraph discussing the methodological issues, so as not to gloss over them.

              Thanks.

              Deficits don't matter, jobs do.

              by aguadito on Wed Apr 17, 2013 at 08:02:36 AM PDT

              [ Parent ]

            •  er (0+ / 0-)

              what i mean is I've ADDED them now per your suggestion.

              not that I had already included them :P

              thanks again, I didn't realize I had totally avoided going into the details of that.

              good catch and thanks for your input :)

              Deficits don't matter, jobs do.

              by aguadito on Wed Apr 17, 2013 at 08:04:00 AM PDT

              [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site