In January 2010, two prominent Harvard University economists, Carmen Reinhart and Kenneth Rogoff, published a highly influential paper in which they argued that high levels of government debt are associated with low levels of economic growth.
They concluded that above the threshold where government debt exceeds 90 percent of national income, “median growth rates fall by one percent, and average growth falls considerably more.”
Following on the heels of the 2008 global financial crisis and the associated spike in government borrowing in Europe and the United States, the Reinhart-Rogoff paper quickly became a touchstone for the small-government crowd. Austerity is the order of the day. Reinhart and Rogoff are its prophets.
Now three economists at the decidedly less upscale University of Massachusetts – Thomas Herndon, Michael Ash and Robert Pollin – have uncovered a series of errors and outright blunders in the Reinhart-Rogoff results.
Not only did the trio show that Reinhart and Rogoff misinterpreted and misanalyzed their data, they also found a simple spreadsheet error that dramatically changed the statistical results. Austerity, it turns out, only works if you don’t know how to use Excel.
Reinhart and Rogoff have acknowledged their errors, though they are at pains to stress that the errors are largely immaterial to their overall conclusions that government debt levels of more than 90 percent of national income are associated with low levels of economic growth.
They also somewhat disingenuously point out that “We are very careful in all our papers to speak of ‘association’ and not ‘causality.’ ” Disingenuously, since their pro-austerity stance shines through all their work. After all, the title of their 2010 paper was “Growth in a Time of Debt,” not “Debt in a Time of Recession.”
Especially misleading is a chart in their paper that shows US economic growth rates for four different levels of US government debt, with the bars becoming alarmingly redder as the debt levels increased.
Reinhart and Rogoff analyzed 220 years of US economic history to conclude that, on average, the US economy has consistently grown at rates over 3 percent per year at all levels of government debt from 0 percent to 90 percent of national income. But when US government debt has risen above 90 percent, the US economy has contracted, they found.
Nowhere in their paper do they mention just when it was that US government debt rose above 90 percent of national income. Was it the Great Depression? No. The Bush or Obama years? No. Perhaps back in the 19th century? No.
In fact, in its 220-year recorded economic history, the United States has only ever experienced four years in which federal government debt exceeded 90 percent of US national income: 1944, 1945, 1946 and 1947.
In those four years, real economic growth was 8.1 percent, -1.1 percent, -10.9 percent and -0.9 percent, respectively. Which tells us absolutely nothing, except that after a huge world war it takes some time for an economy to readjust to peacetime production. Anyone who says that America’s sudden recession in 1946 was due to government debt, not the end of the war, is either crazy, deceitful or stupid.
Carmen Reinhart is the Minos A. Zombanakis Professor of the International Financial System at Harvard’s Kennedy School of Government. Kenneth Rogoff is the Thomas D. Cabot Professor of Public Policy and Professor of Economics at Harvard University. You be the judge.
Actually, Reinhart and Rogoff do recognize the warping effects of World War II – on Australia and New Zealand. In those two countries Reinhart and Rogoff found that high debt was actually associated with stronger than average economic growth. But they (correctly) wrote this off as a distortion caused by the war.
In fact, of the 20 rich countries studied by Reinhart and Rogoff, only one example shows negative growth resulting from high debt: the United States after World War II. But despite the fact that they are American, live in America and mainly study the US economy, they fail to note that the only period of high debt coupled with recession in US history was the 1946 demobilization after World War II.
Apparently war distorts the data when it makes debt look good, but war isn’t worth mentioning when it makes debt look bad.
It gets worse. University of Southern California professor Richard Green raises an even bigger issue. In a column for Forbes magazine, he suggests that it may be the case that debt doesn’t cause low growth. It may be that low growth causes governments to go into debt.
Green presents very preliminary statistical results in his column based on a standard econometric technique called the Granger causality test. His results suggest that the impact of high debt on economic growth is either positive or neutral, while the impact of economic growth on high debt is either negative or neutral.
This is strong first-look evidence that recessions cause debt, not the other way around. But no one – Green included – expects to solve this complex statistical issue in a 600-word column. The travesty is that Reinhart and Rogoff didn’t even raise the issue in a 25-page academic paper.
Lies, damned lies, and statistics. It is easy to massage data. For example, why should one expect high government debt to have an immediate impact on economic growth? Reinhart and Rogoff could just as well have studied the impact of government debt on growth rates several years later.
If they had, they might have found that in the United States, high government debt was associated with rapid economic growth. US government debt peaked in 1945 at 112.7 percent of national income. Five years later, in 1950, the US economy was racing ahead at an 8.7 percent growth rate.
The potential lesson for today? If we borrow heavily in 2013, we can enjoy a huge growth dividend in 2018.
Of course, that lesson is no more valid than Reinhart and Rogoff’s austerity lesson. But it’s no less valid.
If we borrow now to invest in education, job training and infrastructure, it’s likely we will have robust growth in 2018. But we don’t know that from Reinhart and Rogoff’s historical data. We know that from common sense.
Even if the expected economic growth doesn’t materialize, we will still have the education, the job training and the infrastructure to show for our spending. That’s something.
At a time when the US government can borrow for five years for less than 1 percent annual interest and for 30 years for less then 3 percent annual interest, it’s crazy to be cutting government spending instead of investing in our future. Well, it’s either crazy, deceitful or stupid. You be the judge.