Open Access policies for universities and research institutions

There are two basic types of policy – voluntary and mandatory. The former requests or encourages researchers to make their work Open Access by self-archiving it in the institutional repository: the latter requires this action. Whilst research managers, in the spirit of the academy and out of a reluctance to add more to the administrative burden of researchers, may shy away from requiring certain behaviours from their staff, in the case of Open Access it has been shown that voluntary policies have little effect.
The spontaneous rate of self-archiving by researchers to make their work Open Access is around 15-20% and this is not increased in institutions with voluntary policies on Open Access. Only mandatory policies bring the high level of self-archiving that provides a university with the increased visibility and impact that Open Access promises (see a summary of Arthur Sale's findings that evidence this). Moreover, researchers themselves do not object to being required to make their work Open Access (see data to support this statement). The number of research funders and universities that have already introduced mandatory policies bears this out. Policies that rely on voluntary action by researchers fail. The clearest example of this was the original Open Access policy from the National Institutes of Health (NIH) in the US.It was a voluntary policy, requesting its grantholders to make their work Open Access. Only 4% of them did so. it has since been upgraded to a requirement, as a result of which the compliance rate has jumped to 56% at the last count (the NIH estimates that 80000 journal articles result each year from work it funds).
The fact is that if research managers wish to establish a repository that successfully gathers together the whole output of the institution, then a mandatory Open Access policy is needed. This is a summary of existing Open Access policies:
Institutional mandatory policies              131
Multi-institutional mandatory policies          1
Departmental mandatory policies             31
Funder mandatory policies                       52
Total mandatory policies                    215
The chart below shows the cumulative pattern of growth of these policies, with a clear acceleration in the rate of adoption in recent times.
ROARMAP Growth Chart last update Aug 2 2011
The first university-wide mandatory policy was implemented by Professor Tom Cochrane, Deputy Vice-Chancellor at Queensland University of Technology in Australia, in 2004. Previously there had only been one mandate – at School level, in the School of Electronics & Computer Science at the University of Southampton – so QUT’s mandate was a world first.

Since then, growing numbers of universities have followed this route, along with research funders such as the European Research Council.

Mandatory policies do not need to appear over-officious. Coupled with a clear case explaining why the university wishes to collect its research outputs in one place – for internal record-keeping, for national research assessment exercises, as a central locus for access to the outputs of any individual, group or department, and so forth – a mandate becomes a non-controversial part of the institutional modus operandi.

There are brief case studies of university mandatory policies and how they were introduced here.

The Registry of Open Access Policies (ROARMAP)

A list of policies developed by universities, research institutes and research funding agencies is maintained at the University of Southampton. As this is a self-registering service, supplemented by the list owners adding policies that they have discovered serendipitously, this list under-represents the actual number of policies in existence.
Stevan Harnad's presentation on how to design an institutional policy