Home > corruption, economics profession > Reinhart and Rogoff: One year later

Reinhart and Rogoff: One year later

from Dean Baker

It has been a bit more than a year since the Excel Spreadsheet error that shook the world. For those who may have missed it, in April of 2013, Thomas Herndon, a University of Massachusetts graduate student in economics, found an error in the calculations of Harvard Professors Carmen Reinhart and Ken Rogoff on the relationship between government debt and economic growth.

Reinhart and Rogoff had done analysis showing that countries experienced sharply slower growth once their debt to GDP ratio exceeded 90 percent. With the United States and many European countries reaching debt to GDP ratios in this 90 percent range, Reinhart-Rogoff’s work was seen as a warning alarm. It was taken as providing evidence that they would have to reduce spending and/or raise taxes to get or stay below the 90 percent cutoff.

Political leaders and central bankers around the world were happy to trumpet the Reinhart-Rogoff findings. The story was that cutting deficits may slow growth in the short-term, and seriously hurt those directly affected by the cuts such as laid off government workers, but it was essential medicine for sustaining a healthy economy.

The spreadsheet error uncovered by Herndon, and analyzed in a paper co-authored with two University of Massachusetts professors, Michael Ash and Robert Pollin, showed that the Reinhart and Rogoff story was not true. Working off the spreadsheet that Reinhart and Rogoff had created, they showed there was no 90 percent cliff. Reinhart and Rogoff’s cliff depended both on the spreadsheet error and also a peculiar way of aggregating growth rates across countries.

If the numbers were entered correctly and added up across countries with more typical methods, growth did not fall off steeply for debt levels above 90 percent. The data still showed a negative relationship with higher debt levels associated with lower growth rates, but the sharpest reduction in growth rates occurred with debt levels of less than 30 percent. That was a very different story than what Reinhart and Rogoff were telling publicly and presumably also in private meetings with central bankers, finance ministers, and members of Congress.

Perhaps even more importantly, a number of analyses looked at the direction of causation between growth and debt. While Reinhart and Rogoff never directly tested for causation, they certainly implied that the causation went from high debt to low growth rates.

Following the discovery of the spreadsheet error several papers analyzed the Reinhart-Rogoff data and found that the causation went almost entirely from slow growth to high debt. In other words the story was not that countries ran up big debts and then their economies stopped growing. The story was that countries that had serious growth problems tended to run larger deficits to boost growth. Also, if an economy is growing rapidly, its debt to GDP ratio would decline (other things equal) as its GDP rose. When a country’s GDP is not rising much, it’s much harder to bring down the debt to GDP ratio.

With the academic basis for deficit reduction undermined by this new research, it might have been reasonable to expect there would be a renewed push for measures to stimulate the economy and reduce the high unemployment rates that plague most wealthy countries. However nothing like this happened. The push for deficit reduction in the United States and Europe went on just as it had before.

The one exception was Japan, where the government of Shinzo Abe embarked on an aggressive stimulus program. Abe took this path in spite of the fact that Japan has by far the highest debt to GDP ratio of any wealthy country. And Abe’s policies appear to have worked to date, as growth jumped and employment surged.

But Abe embarked on this path even before the spreadsheet error had come to light. The economics profession can’t claim that this new evidence was responsible for the change of policies in Japan.

There isn’t much that the economics profession can claim in this debate that makes it look very good. While there is now a large and growing body of evidence that larger budget deficits would boost growth and employment in the current economic environment, those in the political establishment in both Europe and the United States seem impervious to evidence at this point. They got all the evidence they needed when they had the Reinhart-Rogoff study they could cite. Now that it turns out that Reinhart and Rogoff were mistaken, they see no reason to re-examine their policies.

It is also instructive that Reinhart and Rogoff don’t seem to have suffered much in their professional standing. While both of them have produced a large body of research over their careers, so that their reputations did not rely on the 90 percent debt-to-GDP cliff, this was a rather egregious error. They justified their mistake by pointing out that it only appeared in a working paper that they had rushed to finish. A revised version of the paper included the correct numbers.

However they surely knew that the dramatic 90 percent cliff story from the original working paper was being used in policy debates around the world. Knowing that they had been rushed when they wrote that version of the paper, surely they had time in the subsequent 3 years until the error was discovered to go back and examine their work, or more likely have a research assistant re-examine their work. Obviously they never chose to do so. If either of them has suffered any professional consequence from this failure, it is difficult to see what it is.

Economics is a profession that fixates on the idea of getting incentives right. When two prominent economists can make a major error on work that had a huge impact on economic policy across the world, and face no real consequences, it says a great deal about the incentives in the economic profession.

  1. Garrett Connelly
    May 6, 2014 at 1:15 pm

    It is good to see more interesting details to this story but sad to see another economist unaware that growth leads to human extinction.

    • Herb Wiseman
      May 6, 2014 at 3:42 pm

      Depends on the type of growth does it not? Meaning it depends on relying on the GDP instead of a measure of life quality. Surely you mean that the GDP is a bad measure to sue for economics.

  2. Bruce E. Woych
    May 6, 2014 at 1:31 pm

    A political policy requires crisis and legitimation to give it steam and justifications. Rarely do we get a direct look at the people that create those supports, and the institutions that finance them in serial fashion.

  3. Bruce E. Woych
    May 6, 2014 at 1:46 pm

    Two of these co-authors of the study you sited also showed discrepancies in Stock Market research that received much less attention:
    Stock Market Liquidity and Economic Growth: A Critical Appraisal of the Levine/Zervos Model
    Andong Zhu | Michael Ash | Robert Pollin | 9/1/2002
    “Levine and Zervos (1998) presented cross-country econometric evidence showing that, in a sample of 47 countries, stock market liquidity contributed a significant positive influence on GDP growth between 1976-93.
    We show that the Levine-Zervos results are not robust to alternative specifications because of the incomplete manner in which they control for outliers in their data. We show that when one properly controls for outliers, stock market liquidity no longer exerts any statistically observable influence on GDP growth.”

    One can only speculate how Washington pandered to this misguided trust in fueling the fires of Wall Street policy incentives…., but it is pretty clear the correcting data was never used to dampen Wall Street’s favorite sons!

  4. Herb Wiseman
    May 6, 2014 at 3:46 pm

    I wonder about who owns the debt? That factor has not been studied from what I can tell. If the private sector is owed the government debt, the interest it receives from the tax base through the government is taxed lightly if at all and often placed in tax havens.

    But if the debt is held by government owned banks the interest paid by the tax base is returned as a revenue stream for government which means it remains available in the economy for services and infrastructure. That then can be used to repair, rebuild and expand needed infrastructure such as water and sewer. That increases the GDP but in a positive way.

  5. May 7, 2014 at 2:31 am

    If the proposition P = “Debt/GDP > 90% implies slower growth”, then Reinhart and Rogoff (RR) has failed to prove P. This fact has led to logical fallacies committed by many economists.

    1. No one has proved that P is false. It may still be true; only the proof has failed.
    2. If Q = “Fiscal stimulus with public debt is bad”, P does not necessarily imply Q.
    3. Even if P is false, ~P does not imply ~Q, i.e. it does not imply that stimulus is good.

    From falsifying the RR proof, Herndon has not proved (his words) “the implication for policy is that, under particular circumstances, public debt can play a key role in overcoming a
    recession.” The conclusion does not follow from his research.

    The research has not “settled the ‘stimulus vs. austerity’ argument once and for all”, as claimed by Blodget or that “Paul Krugman has won” the case for stimulus. Inconclusive
    statistical research proves nothing.

    The research so far has settled nothing and proved nothing, perhaps only shows the lack of clear thinking by the economics profession.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: