The 9th PE-DPS Jan H. Höffler on Data Use in Economics

Greetings!

In our last session of the year on November 25th, 4pm CET, we're inviting

Jan H. Höffler on Data Use in Economics: What can we do better

Data science permeates all scientific realms and, slowly but surely, our surrounding environment. How does this impact economics? This presentation aims to tackle data use and transparency by addressing the "replication crisis" phenomenon while actively presenting a step towards a solution: The ReplicationWiki. "Replication crisis" refers to the phenomenon that research results can often not be independently replicated. Though this takes place in several sciences it is rarely revealed. One reason for this is a lack of data and code transparency. There has been some improvement mostly at top journals, however, journals still tend to block corrections, and there are hardly any retractions in the social sciences, even when it is pointed out that there were problems with data collection/data use and transparency. In economics, already the first issue of Econometrica mentioned that empirical work published in that journal should be replicable.

In 1986, most of the articles of an issue of the Journal of Money, Credit, and Banking, whose editor had archived their data, were found not to be replicable in an article published in the American Economic Review. The Journal of Applied Econometrics already in 1995 introduced an online data archive after a special issue on the topic. Only after a series of scandals, in particular about fake data in psychology, did most leading journals start to run mandatory data archives, and more and more code is shared. But the bulk of empirical work in economics still lacks transparency. Despite that, it keeps getting cited in academia, by practitioners, politicians and, in the media. With the originally INET-funded ReplicationWiki and its accompanying replication working paper series and workshops we have been contributing to change.

CLAIM: Economists should study epistemology and I would say that up until now, we have failed miserably.

QUESTION: What else can young scholars do to change the status quo of economic research of limited reliability?

Short bio:
Jan H. Höffler founded the ReplicationWiki during his PhD with an INET grant. On replication he has taught at universities in Germany, Toronto, Nanjing, Geneva, and Maastricht, presented at international conferences and as invited speaker at various departments, (co-)organized workshops and conference sessions and published articles.