CEPR - Center for Economic and Policy Research

Multimedia

En Español

Em Português

Other Languages

Home Publications Blogs CEPR Blog The Reinhart-Rogoff Debt-to-GDP Error: Why it Matters

The Reinhart-Rogoff Debt-to-GDP Error: Why it Matters

Print
Written by Dean Baker   
Thursday, 18 April 2013 12:02

The Carmen Reinhart and Ken Rogoff (R&R) paper purported to show that countries with debt-to-GDP ratios above 90 percent see sharply slower growth rates, and has been widely cited in policy discussions in the United States and Europe and used as a rationale for a near-term focus on deficit reduction. Politicians and policy analysts relied on the results of this paper to insist on spending cuts and tax increases even in economies that are operating at levels of output far below full employment. Based on R&R’s findings, they argued that it was important to keep debt levels from crossing the 90 percent threshold.

This debate is important because the threat to growth from high debt levels was one of the main arguments against the aggressive use of fiscal policy to boost growth. The work of HAP and UMass economist Arindrajit Dube has essentially undermined the basis for this argument. No one can still maintain that we have good evidence that debt levels of the size we could conceivably face in the near future would impair growth.

The new paper from HAP works off the original spreadsheet used by R&R and uncovers several important calculation errors.

  • The most important of these errors was excluding four years of relevant data from New Zealand. Correcting this mistake raised New Zealand’s average growth rate in high debt years from the -7.9 percent R&R reported to a positive 2.6 percent.
  • Because only 7 countries have crossed the 90 percent debt-to-GDP barrier highlighted by R&R, this change alone raises the growth rate among the high debt countries by 1.5 percentage points.

When this and other adjustments are made to R&R’s data, the sharp falloff in growth rates for countries with debt to GDP ratios above 90 percent disappears.

  • While the corrected growth rate is still lower for high debt countries, the difference is much smaller and nowhere close to being statistically significant.
  • Furthermore, the sharpest falloff in growth rates occurs at very low debt levels (less than 30 percent of GDP).
  • If the corrected results from R&R could be taken as a basis for policy, then the implication would be that countries should strive to have extremely low debt to GDP ratios, certainly well below the levels that the United States and other wealthy countries have generally sustained.

Of course many economists have long raised conceptual objections to R&R’s basic approach. The most obvious of these is the problem of causation.

  • Countries with bad economic performances are likely to have high debt to GDP ratios.
  • This is both because they are borrowing a lot to support their economy and also because if GDP grows quickly then the debt to GDP ratio is likely to fall. (Since GDP is the denominator in the ratio, a faster growing GDP directly lowers the ratio.)
  • University of Massachusetts economist Arindrajit Dube did an analysis of the R&R data, “Growth in a Time Before Debt,” that compellingly demonstrated this pattern of reverse causation. He found that high levels of debt to GDP were not a good predictor of future GDP growth.
  • Specifically, the current year’s debt level tells us almost nothing about growth rates over the next three years.
  • However Dube found that the current year debt to GDP ratio is a very good predictor ofpast growth. There was a very strong relationship between high debt to GDP ratios in the current year and slow growth in the prior three years.
  • This finding strongly suggests that the weak relationship between growth and debt-to-GDP ratios in the corrected R&R paper are simply an artifact of reverse causality.Weak or negative growth leads to more debt.

The policy prescription from the corrected R&R paper is that people troubled by high debt-to-GDP ratios should focus their energies on maintaining strong economic growth.

  • Japan can be poster child for this story of reverse causation. Its economy soared in the 80s driven by bubbles in both its stock and housing markets.
  • During this period it had very low debt to GDP ratios.
  • Then the bubbles burst in 1990. It subsequently had slow growth and also ran large deficits to support the economy.

There are other important conceptual reasons for questioning the R&R claim that high debt leads to slow growth.

  • Debt is to some extent arbitrary. A country like the United States holds literally tens of trillions of dollars of assets (e.g. land, mineral rights, fishing rights, the airwaves etc.) that it could sell at any time.
  • No one could really believe that if we sold of $1.5 trillion in assets and thereby shaved 10 percentage points off the debt-to-GDP ratio that the economy would suddenly boom.
  • Nor is the value of debt constant through time. The bonds issued at very low interest rates today will sell at steep discounts when interest rates return to more normal levels.
  • Again, no one could really believe that if we bought back these bonds at sharp discounts, and thereby reduced our debt-to-GDP ratio, that the economy would then grow more rapidly.

We know that spending on infrastructure, education and other areas could boost growth and increase employment. At this point there is no reason other than personal prejudices against debt not to pursue policies to increase growth and create jobs.

 

Tags: austerity | debt-to-GDP ratio | economic crisis | Reinhart/Rogoff

Comments (0)Add Comment

Write comment

(Only one link allowed per comment)

This content has been locked. You can no longer post any comments.

busy
 

CEPR.net
Support this blog, donate
Combined Federal Campaign #79613
budget economy education employment Haiti health care housing inequality jobs labor labor market minimum wage OccupyWallSt poverty recession retirement Social Security taxes unemployment unions wages Wall Street women workers working class

+ All tags