One of the most frequently used arguments in favor of depositing papers in a repository is to preserve the rights of the hosted institutions that pay the research done by the authors of those papers. Most of these rights usually refer to exploitation rights but many times the even more important moral rights are forgotten. Moral rights intends to preserve the authorship and integrity of the document (nobody can overwrite the text or usurp the author) that can be also extended to the institutional authorship. That means that the name of the university or research institute should be clearly identified in a meaningful way. This apply of course to the mention of the affiliation both in the full paper and in the metadata records but also equally to the web locator and name of the file, that is the full URL of the webpage of the file and the webpage of the metadata record. As it is known the host part of those URLs is the web address of the repository itself.
The prestige of an institution is clearly linked to both intangibles (for example quality of teaching) but also tangibles like research results. i.e. the collections of scientific papers published by its faculty members (plus dissertations/thesis from students, technical reports, books, etc...) that usually are the core of its institutional repository. This is clearly the treasure of an academic institution, so then...
There is no problem in building consortia for sharing resources and expertise but there are several very important ones (HAL, BePress, Univ. California, DiVa and others) that consists of repositories that do not exist individually or do not use their own individual institutional domain. There are no addresses like repository.ucla.edu, repositoire.ensmp.fr or publikationsdatabas.kth.se, and most important, the presidents and rectors of these institutions are unaware of these transfer of rights.
If we checked the moral rights at the level of author we realize that the web addresses also excluded the recognition owed to the authors, as:
Today mandates are not been very useful, probably because most of the authors are reluctant to deposit their papers. This is an issue that should be explored in depth.