I recently interviewed Thomas Krichel, founder of RePEc, the content repository for economics. RePEc, which he founded as long ago as 1997, is interesting for many reasons. RePEc is one of the largest subject repositories, with over 55,000 registered authors, and 14,000 institutions providing content. It is a pioneering domain-based resource created to meet the working practice of economists, and their use of working papers. RePEc was built from the start as a domain-specific tool, while arXiv, initially based around physics papers, is now widely used for papers in many subject areas (including some economics papers, incidentally). You could say that arXiv provides the model that subsequent preprint servers followed – well over a hundred, at the last count, but few collections of this type have reached the widespread adoption of arXiv, SSRN and RePEc. Krichel emphasises that RePEc is more than a preprint server. It contains preprints, but he describes it as ‘a digital library for economics, with at its heart a metadata database about papers’. Could RePEc provide inspiration for new thinking towards open science?
RePEc is indeed distinctive from other subject repositories in several ways:
- It is distributed – there is no single central RePEc server
- It contains both working papers, and articles provided by publishers
- It provided tools for author registration and disambiguation in 1999, well before ORCID existed
- It provides current awareness tools, focusing on recent changes in thinking
- It provides a ranking system for economists, using no fewer than 36 separate rankings, some of them based on citation.
Since it has no central server, RePEc has low costs compared with arXiv. In fact, it would be difficult to establish the real costs, since so many institutions host a portion of RePEc. The distributed model is of course in favour today, since it provides a more sustainable service than relying on a single large funder, who may at any point stop providing (see Microsoft Academic Graph for an example). In contrast, RePEc would survive the closure of a few institutional servers. As a further demonstration of its distributed model, RePEc is enhanced by several additional digital services that make use of the underlying content, such as the Munich Personal RePEc archive, for authors in institutions that do not have a participating RePEc archive.
Working papers
What is a working paper? It’s somewhat similar to a preprint, in that it is an early version of a paper that has not been through peer review, although working papers tend to be published in thematic series. Submitting papers to each series is typically by invitation only; this provides a kind of quality control. For years before the Internet, Krichel, a former economics lecturer, revealed, working papers were circulated by economics departments in institutions, sending photocopies to other institutions. Quite apart from the labour involved in managing such a system, the process must have created a two-tier society. If you were an institution that took the care to send out your papers, and to circulate the papers you received, all well and good, but if your institution didn’t have the resources to photocopy and to manage the process, you simply lost out. So in this sense, working papers in digital form are a huge advance: circulation is instant, and labour-free. “All I did”, stated Krichel, rather laconically, was to create a service to circulate those papers digitally. It may have been a small step back in 1993, but today all economists make use of the service. You could say that the creation of a digital sharing service for working papers was one of the biggest steps towards democratization of economics, from a small number of prestigious (and highly-funded) institutions to a tool that provided a level playing field for all academic economists.
An interesting 2020 scholarly article found that around 67% of the papers were subsequently published in a journal, and 8% as a book chapter, not dissimilar to the proportion of bioRxiv preprints that reach publication (although not typically in books). The article suggests that the number of working papers is decreasing, but I’m not sure if that represents a move away from the idea of sharing content before publication.
A 2015 article compared what they call “subject repositories”: arXiv, RePEc, SSRN and PubMed Central, by counting citations to these repositories in Scopus, although I don’t think that metric is a very effective way to judge the success of RePEc, given the distinctive characteristics of economics working papers described above. Nonetheless, the article suggests that subject repositories have a higher impact than institutional repositories, and suggest that the success of arXiv and RePEc may be due to practice of circulation of early drafts of papers within physics and economics respectively.
Clearly, the size and continuing success of RePEc (by my count one of the four largest subject repositories) suggest it fulfils a real need in economics. Perhaps this digital transfer of an effective working practice in economics provides a model for other domains, and one that conforms well with open science principles.
Thanks to Thomas Krichel for his comments on an earlier version of this post.
Leave a Reply