Reinhart and Rogoff provide a detailed response in WSJ, acknowledging the Excel mistake but downplaying the other criticisms. Justin Fox identifies a deeper problem with macroeconomics in general:
This is watching the sausage of macroeconomics being made. It’s not appetizing. Seemingly small choices in how to handle the data deliver dramatically different results. And it’s not hard to see why: The Reinhart-Rogoff data set, according to Herndon-Ash-Pollin’s analysis, contained just 110 “country-years” of debt/GDP over 90%, and 63 of those come from just three countries: Belgium, Greece, and the UK.
This is a problem inherent to macroeconomics. It’s not like an experiment that one can run multiple times, or observations that can be compared across millions of individuals or even hundreds of corporations. In the words attributed to economist Paul Samuelson, “We have but one sample of history.” And it’s just not a very big sample.
Free Exchange weighs in:
The latest dust-up does nothing to answer the question of causation. Slower GDP growth could be the cause of a rising debt load rather than the result. Ms Reinhart and Mr Rogoff acknowledge in their academic work that this conundrum “has not been fully resolved”, but have sometimes been less careful in media articles. This is perhaps their biggest mistake. The relationship between debt and growth is a politically charged issue. It is in these areas that economists must keep the most rigorous standards.
On that point, Yglesias unpacks Arindrajit Dube’s analysis, which argues that low growth causes debt more than debt causes low growth. Previous Dish on the R&R paper here and here.
The Dish 