To see more on this project, click here.

This is big news, as it casts doubt on one of the most influential pieces of economic work published over the past several years.  Reinhart and Rogoff’s work has been used to create (or justify) alarm over America’s mounting public debt, and in turn press austerity to the top of the nation’s economic agenda.

Many sociologists will probably agree with Posen’s view that it seems unlikely that there exists a strict cutoff in debt-to-GDP ratios beyond which an economy stagnates.  Economies are highly complex organisms, and there are a range of factors that could shape public debt’s ultimate impact on prosperity.  Economies are embedded in society — the mechanics by which economic forces operate are shaped by the broader economic, political and social institutions under which they operate.

That being said, credit should be given to Reinhart and Rogoff for making their data available publicly.  This type of reanalysis would have been very difficult to accomplish without this type of transparency.  Comparatively few researchers are willing to subject themselves to type of possible scrutiny.

Moreover, my experience working with this type of long-run data is that issues related to sample representation are not as clear cut as analyses of Census data or political-economy data from the 1990s or 2000s.  For example, researchers have to struggle with questions about balancing longitudinal and cross-sectional representation, and it is hard to deal with these problems through methods like data imputation because they are not missing at random.  Moreover, researchers have to make decisions about whether or not to include unique historical periods in their analyses (like the post-WWII demobilization of 1946 – 1950).  Granted, the report states that these data inclusion rules were not applied consistently, which is a clear problem.  Yet I would resist the conclusion that the problems uncovered by Herndon et al. suggest something nefairous on Reinhart and Rogoff’s part.

A final note from Posen’s article, with which I agree a great deal:

Certainly, academics who oversell their conclusions to get their lines into a speech or to come up with a rule that can be named after them can become disproportionately influential. And the academic establishment does not help matters by encouraging the production of new and counter-intuitive results irrespective of their validity in practice. It also disincentivises the replication and empirical testing of previously published work. Yet, eventually, the truth will out – as it did this week.

The research community may strive too hard for novel, headline-grabbing breakthroughs, and overlook the need for – and value of – normal science.