The glitchy data comes from a hallmark study on debt, Growth in a Time of Debt, from 2010. Since it was published, it's become a favorite of people like Paul Ryan, who mentioned it frequently during the Romney campaign. But a new meta-analysis of the study's original excel spreadsheet, by three UMASS economists, uncovered an incredible coding error that excluded as much as a quarter of the data concerning economic growth. To put that in perspective, the original paper found that countries with high debt grew -0.1% every year. Without the Excel bug, that number changes to 2.2%. "We will redouble our efforts to avoid such errors in the future," responded the scientists in a Wall Street Journal post today.
Of course, it's not like politicians were basing their ideas on the paper alone—in reality, these kinds of studies are used to post-rationalize prior beliefs about the economy. More importantly, Paul Krugman points out that there are plenty of instances where countries boomed during high debt periods, as well as plenty of examples of countries slogging through depressions with tons of debt. In other words, it's a more complex issue than simple causation.
You can't blame Excel alone for the error—after all, the data should've been reviewed back when the paper was published. Still, there are plenty of other examples of program glitches leading to major inaccuracies. The moral of the story? Double check your sources before making decisions that affect the global economy.
[Ars Technica]
Thank a Simple Excel Glitch For Austerity Economics