More Reasons to Doubt Rogoff and Reinhart

By Yeva Nersisyan
With unemployment expected to remain high in the U.S. and Europe and the possibility of a double-dip recession growing stronger, some sensible voices are calling for another round of fiscal stimulus. And then there are others who not only argue that we don’t need more stimulus, but make a case for starting to cut spending today, notwithstanding a very fragile “recovery.” Ken Rogoff (see here), who has become the de-facto authority on the issue of sovereign deficits and debt (together with his co-author, Carmen Reinhart), in a recent FT article is trying to make the case for the redundancy of further economic stimulus. Subpar economic performance and unemployment are the usual companions of post-financial crisis recovery, he argues, hence there is no need for a “panicked fiscal response” (even Secretary Geithner has cited their research to demonstrate that the current slow pace of recovery is normal). Rogoff goes on to argue that the long-term effects of government debt accumulation on growth shouldn’t be ignored. The theoretical and empirical bases for his arguments are found in his recent book with Reinhart, This Time is Different, as well as an NBER paper, “Growth in Time of Debt”. This paper, similar to the book, has been very popular, especially among those needing empirical justification for their anti-fiscal policy stance. While the RR book focuses on the short-run, immediate impacts of sovereign debt (i.e. financial and economic crises), the focus of the paper is the impact of sovereign debt on long-term growth. In this blog I want to give a quick, critical evaluation of the paper (a longer version can be found here).

When orthodox economists start their empirical research regarding the long-term impact of deficits and sovereign debt, they do not ask whether deficits contribute to or inhibit long-term economic growth. They do not ask, because they already “know” the answer, as the ECB put it: “Although fiscal consolidation may imply costs in terms of lower economic growth in the short run, the longer-run beneficial effects of fiscal consolidation are undisputed.” (ECB, Monthly Bulletin, June, 2010). What they want to find is some threshold for deficit-to-GDP and debt-to-GDP ratios beyond which debt becomes detrimental to growth. With this goal in mind, Rogoff and Reinhart embark on a “scientific” journey through time and space.
Their method is actually quite simple: they construct some arbitrary ranges for debt-to-GDP ratios (0-30, 30-60, 60-90, >90) and take the average of growth rates for each range. They then take the average of these averages for a large number of countries and conclude that when the government debt-to-GDP ratio crosses the threshold of 90% (again, an arbitrary number), median growth rates fall by one percentage point and the average falls even more. This limit is the same for developed and developing countries, however, when it comes to external debt (which is defined in their book as both public and private debt issued in a foreign jurisdiction, and usually, but not always, denominated in foreign currency), the threshold is much lower, just 60% of GDP. Once a country crosses this lower external debt threshold, annual growth declines by about 2 percentage points and at very high levels, the growth rate is cut almost in half.
Interestingly, however, average growth rates don’t monotonically decline, i.e. the average rate of growth is higher when debt-to-GDP ratio is in the 60-90% range than the lower range of 30-60%. In addition, growth rates don’t slow down for all the countries in their sample. For some countries the average growth rate is higher when debt is over 90% of GDP than for lower levels of debt. Reinhart and Rogoff don’t point out this “anomaly,” nor do they offer any explanations. More importantly, since they take the average of averages of a number of countries, it is possible that countries like the U.S. may drive the results for the whole group. They single out the case of the U.S. in their paper to demonstrate their results. However, a closer look shows that they only have 5 data points for the U.S. when the debt-to-GDP ratio was over 90%. This is only 2.3% of the total of 216 observations. Moreover, 3 out of these 5 observations are for the years 1945, 1946 and 1947, the period after WWII when government debt was high due to war spending. In this period, growth slowed down significantly as the government was withdrawing war spending from the economy. In 1946 alone, GDP contracted at a pace of -10.9%. Rogoff and Reinhart fail to even mention this in their paper. Similar situations might be true for many other countries, where high levels of debt-to-GDP follow extreme economic or political events.
But what is even more important is that what they find in the data is merely a correlation. The causation then is imposed by Reinhart and Rogoff with explanations based on Barro’s Ricardian equivalence theory. “The simplest connection between public debt and growth is suggested by Robert Barro (1979). Assuming taxes ultimately need to be raised to achieve debt sustainability, the distortionary impact imply is likely to lower potential output” [sic].
There is no doubt about the correlation between high debt-to-GDP ratios and low economic growth found in the data. However, there is a more sensible explanation for this correlation. As explained in many past posts on this blog, the government budget balance automatically goes into deficit in a recession leading to an accumulation of public debt. Besides, GDP, the denominator of the ratio shrinks making the ratio even larger. It is sufficient to look at what happened during this most recent crisis to see this. The average rate of growth has been -0.23% for the recession years 2007-2009. At the same time, government debt held by the public has increased from 36% of GDP in 2006 to about 52% in 2009. So if you look at the data, the rate of growth was 2.7% in 2006 corresponding to a debt-to-GDP ratio of 36%. In 2009 growth was -2.6% with a corresponding debt-to-GDP ratio of 52%. Hence there is a correlation between slow growth and high levels of debt which is not surprising. But unless you want to argue that the current recession was caused by high levels of government debt, then it is obvious that causation runs from slow growth to high debt and not the other way around as Reinhart and Rogoff claim.
They also find that growth deteriorates significantly at external debt levels of over 60% and that most default on external debt in emerging economies since 1970s has been at 60% or lower debt-to-GDP ratios (which is the Maastricht criteria). While this might be a surprising finding for them, it should be clear why countries are not tolerant to external debt which is almost always denominated in foreign currency. When a government borrows in foreign currency, even low levels of indebtedness can be unsustainable since the government is not able to issue that foreign currency to meet its debt obligations. As countries need to earn foreign exchange from exports, a sudden reversal in export conditions can render the country unable to meet its foreign debt obligations leading to a crisis and slower growth. Sovereign governments, on the other hand, do not face any financial constraints and cannot run out of their own currency as they are the monopoly issuers of that currency. They don’t need to increase taxes in the future (a la Barro) to pay off the debt as they make interest payments on their “debt” as well as payments of principal by crediting bank accounts, meaning that operationally they are not constrained on how much they can spend. See here for more on this.
While many experts believe that there is an acute possibility of a double-dip recession in the U.S. (see here) and other developed nations, Ken Rogoff is not one of them. And even if we do face the threat, he argues, monetary policy will suffice (if anything, this crisis has demonstrated the ineffectiveness of monetary policy (interest rate management to be more precise) not to be confused with the massive lender-of-last resort operations that the Fed undertook to stabilize the financial system).
Even if there was no threat of a double-dip recession, one could rightly argue that the current high levels of unemployment and underemployment require more government spending. Rogoff’s argument, however, is that sustained high unemployment is the normal consequence of a financial crisis and hence he seems to conclude that fiscal measures to solve the unemployment problem are unnecessary. This is very bad policy advice – we know we have a problem (unemployment), we know how to solve it (public works), but we shouldn’t do so for fear of growth slowing or markets disciplining the government at some indefinite time in the future, a fear based on the wobbly research of Reinhart and Rogoff.
To summarize, the Rogoff and Reinhart research is not a scientific quest but merely a journey with a set destination. It is not based on any sensible theory, and the statistical analysis is of questionable quality as well. Government deficits and debt largely mirror what goes on in the private sector. There are no magic numbers for deficit and debt ratios applicable to all countries and all times. Devising such ratios is a useless exercise.
Even in better times, the U.S. economy is operating with considerably high levels of unemployment and underemployment (see here), underscoring the necessity of government intervention in the economy. In a recession as the private sector cuts back its spending and tries to de-leverage, the role of government, as the only entity in the economy that can run persistent deficits without facing solvency issues, becomes especially important. Regardless of whether there is a threat of a double dip recession, the government should act to solve the unemployment problem through direct job creation TODAY. High levels of unemployment are not compatible with a democratic society.

8 responses to “More Reasons to Doubt Rogoff and Reinhart