WRRO received a request from The York Management School to add a series of Working Papers to the Institutional Repository - and the School was particularly keen that the papers should also appear in RePEc.

RePEc (Research Papers in Economics) is a decentralized database of research papers in economics.

RePEc requires a RePEc archive to be created locally using a metadata format called ReDIF (Research Documents Information Format). What does this mean if works are already housed in an institutional repository?

RePEc will point to the full text of files in an institutional repository so there is no need for duplication. However, in order to be included in RePEc, a local archive needs to be created holding basic information about the collection of papers being added to RePEc - some examples are given below. Full instructions for creating an archive are available from the RePEc site and RePEc provides templates for the various files you need to create on your server:

You will need an archive code - this is assigned by RePEc


e.g. White Rose has the code RePEc:wrc

RePEc looks for a particular directory structure on your server:


Basically, the RePEc folder contains:

(a) a description of your archive. Here is an example:

(b) a description of the collections within your archive. Here is an example:

(NB there is only one collection at present but this file could contain several series e.g. Economics Department, Management School etc.)

(c) a subdirectory for each collection containing a ReDIF metadata description of the individual items in the collection. Here is an example:

This file includes the URLs for the individual papers - which can point to the copies in the institutional repository.

RePEc checks the archives for changes - usually daily.

The resulting display of working papers in RePEc looks like this:


1. This work was done manually as so is clunky and non-scalable. The creation of the subdirectories required by RePEC and the creation of the files describing the RePEc archive and its subcollections are not particularly onerous tasks.

2. The metadata files describing archives could easily sit on the institutional repository server, as in this example, or on a departmental server. Overall, a distributed architecture may be more feasible so that departments control - and have responsibility for updating - their information. For instance, individuals can create their own identifier in RePEc and may wish to include this in the metadata for specific items so that their items which reside in different collections can be drawn together through this shared identifier (useful if an academic moves institutions, for example). However, we'll offer a hosting service for departments without the time or resources to create their own RePEc repository, so long as we can automate metadata creation (see 3.)

3. The most time consuming element is the creation of metadata for the individual items. A DC -> ReDIF export option from ePrints would be useful.

One example of a script which has been used to convert OAI data to ReDIF is available:

Some kind of dynamic link between the IR and the RePEC metadata files would be very helpful.

Deposit scenario