The Great Reinhart/Rogoff Error
With talk of immigration, guns, and bombs, an important wonk-story has slipped under the radar: the Reinhart/Rogoff debacle. (Warning: this gets wonky).
Paul Ryan’s austerity budget proclaims that there is “conclusive empirical evidence” that high levels of federal debt cuts off growth. He is citing the work of Reinhart and Rogoff, who published a paper in 2010 called, “Growth in a Time of Debt,” which purported to show that economic growth decreased drastically once a country’s debt to gdp ratio crossed over 90%. Economists like Paul Krugman initially called foul, arguing that the caution should be reversed: that slow growth caused debt to balloon. But Rogoff/Reinhart pushed forward, conceding that while causation may go in both ways, countries should still pursue austerity (recent evidence suggests their causation is bunk).
But a crucial blow has been struck to that assumption by Thomas Herndon, Michael Ash and Robert Pollin in their new paper. Synopsis: Herndon et al find that the GDP growth rate for countries carrying a public debt to GDP ratio of over 90 percent is 2.2 percent, not the -0.1 percent found by R/R (which is a huge difference). The reason: “We replicate RR and find that coding errors, selective exclusion of available data, and unconventional weighting of summary.”
Essentially, Reinhart and Rogoff first, failed to include some countries in their analysis through a routine Excel error. Incredulous readers can see a screenshot here. Second, they used an incredibly strange weighing mechanism that essentially discarded years of solid growth in the UK because of one year of poor growth in New Zealand. Third, they just flat-out exclude three time periods (Australia (1946-1950), New Zealand (1946-1949), and Canada (1946-1950)) where countries had solid growth and high debt to gdp ratios. Plug in the (distorted) numbers, and voila! empirical “proof” that the E.U. and U.S. should pursue harsh austerity.
Business Insider is now reporting that the paper failed to go through proper channels, and was not peer reviewed. This leads to questions that have been raised already about empirical economics (especially given how many universities have abandoned economic history and other relevant disciplines). Krugman has been harping (most recent post here) on the fact that most people just accepted the results without question because it confirmed the conventional wisdom (remember Galbraith’s warning about the “conventional wisdom”?).
The real refutation of Reinhart/Rogoff isn’t Herndon et al. It’s what Europe is going through right now – and no amount of Excel errors will change the high unemployment rates. Well, unless you believe Jack Welch…
All told, this doesn’t change much. Reinhart/Rogoff will continue to argue that high debt correlates to low growth, which is true, but the correlation is the reverse of what the asusterity Cassandra’s declare. We need to cut deficits, but only once the crisis has passed.