Three economists from the University of Massachusetts, Amherst recently published a paper, claiming they have found an uncommon spreadsheet coding error in a famous AER paper, 'Growth in a Time of Debt', coauthored by Harvard economists Carmen Reinhart and Kenneth Rogoff. This immediately went viral since the key finding in Reinhart and Rogoff's paper had already become some politicians' cannonball in order to fire pro-austerity plans across the US and the Europe. Specifically, R-R have found that, on average, a greater than 90% debt to GDP ratio would lead to a growth rate of -0.1%. But according to the critics, the authors haven't taken the averages right in their Excel spreadsheet, and the correct number should be 2.2% instead of -0.1%, which is far from stagnation. Later, R-R offered their responses, twice - first rebutting all the critiques but then admitting the coding error while rebutting the rest (here is a nice summary of the entire episode).
I bought most of their second response, yet I still think their Excel error was dumb. Besides, if it were 2.2% rather than the eye-dazzling -0.1%, I would doubt they could ever publish it in the first place. Although Reinhart and Rogoff referred to their JEP piece where they've shown similar results of negative correlations, I still believe what really caught the eyes of economists as well as politicians was the magnitude of that correlation, particularly when the debt-to-GDP ratio passes the 90% threshold.
Interestingly, adding new flavor to this debate is neither economists (see here and here) nor politicians but comedian host Stephen Colbert. He took on this debate yesterday and turned it into something hilarious, something epic. He brought onto the stage one of the critics from UMass, an econ grad student Thomas Herndon, and shouted at these "left-leaning academi-aholes", "nerds! I bet you found them [the errors] on a Friday night with your mom, while the rest of us were going up to The Point and drinking PBR!"