Share |

The Hysteric Response to the Reinhart-Rogoff Row

e21 | 04/24/2013 |

Last week, careful research by UMass-Amherst professors found that the seminal paper by Ken Rogoff and Carmen Reinhart contained a data error that biased their conclusion that debt-to-GDP ratios in excess of 90% lead to dramatically slower economic growth.  The 90% of GDP figure was arbitrary with no theoretical basis.  As a result, it’s not terribly surprising to find that its significance was oversold.  Yet, the basic contention that high levels of public debt weigh on growth was and remains perfectly reasonable.  It is unfortunate that Reinhart and Rogoff’s mistake has been used by proponents of government activism to invalidate the basic premise that increases in public indebtedness slow growth.

First, it is important to point out that the 90% threshold never had any real economic meaning.  Empirical results without theoretic substantiation are always suspect because they could be the product of data mining.  If the original 90% threshold was selected because it delivered the most salient difference in average or median economic growth rates, then it was an artifact of the data rather than a meaningful threshold.   Moreover, reliance on the perils of some specific debt threshold with no theoretical basis to explain why that specific debt threshold was meaningful was always subject to claims of reverse causality.  As Paul Krugman argued:

“As soon as the paper was released, many economists pointed out that a negative correlation between debt and economic performance need not mean that high debt causes low growth. It could just as easily be the other way around, with poor economic performance leading to high debt." 

These were always legitimate questions that deserved serious answers.  But when the UMass paper was released, it felt as though that discussion had been short-circuited.  Suddenly, the entire “austerity” movement has been invalidated:  if the Reinhart-Rogoff study’s 90% threshold was mistaken, then it must mean that increases in public debt had no ill-effects on GDP growth, or at least were small relative to the boost government spending could provide.  One could only imagine the sense of validation that gripped the pro-government commentariat upon learning of the error. 

The irony is that now that a data error has been found, progressives wish to inflate the significance of Reinhart-Rogoff and pretend as though it was the sole basis for concern about elevated public debt levels.  This is nonsense: numerous studies found that increased levels of public debt were associated with slower growth and they generally identified a lower threshold.  The Bank for International Settlements (2011) found that debt overhang problems become aggravated once debt hits 85% of GDP.   A 2013 paper from a former Fed Governor and famous econometrician found that economies are especially vulnerable to a debt crisis when debt exceeds 80% of GDP.     

Higher debt levels increase the required level of future taxation, increase the probability of a financial crisis, and reduce governments’ flexibility to deal with future crises that may emerge.  Higher expected future tax rates reduce incentives to invest today.  The probability of a fiscal crisis makes investors seek liquidity and safety rather than productive investment.  Both factors increase savings, reduce current spending, and slow growth rates.  On the last point, it is important to recognize that TARP did not succeed because it was a genius plan with perfect execution.  TARP succeeded because federal debt was just 36.5% of GDP at the start of 2008.  Had the U.S. entered 2008 with today’s debt level, it is not clear the federal government would have had the capacity to backstop the banking sector.  The result would have been a situation similar to the one in Europe, where the weakness of the sovereign and the banking system create a negative feedback loop that reduces credit availability and investment demand.  Moreover, the higher the debt level, the more dependent the fiscal authority becomes on the central bank suppression of interest rates.  At 1980s interest rates, it would cost more than 10% of GDP to service current federal debt.

Accumulating more debt for additional costly fiscal experimentation with stimulus makes no sense.  Yet, the basic argument against it, however strong, does not seem to resonate.  Instead, there is a strange fascination with “tipping points,” or knowing precisely when the debt level is going to precipitate a crisis.  As Fed governor Jerome Powell explained, while we don’t know precisely where that tipping point is, we should all recognize that we’re much closer to it today than we were previously.  It would be a shame if policymakers were to discover precisely where the tipping point is by allowing the country to go over it. 


e21 Projects & Partnerships