The portability or non-portability of research

Our REF consultation had 44 questions. The one provoking the most polarised and contentious responses concerned the Stern proposal that outputs should not be portable. Stern proposed that when staff move from one institution to another the credit for outputs they have authored should not move with them. Previously outputs were characterised as ‘portable’. The ‘originating institution’ lost all credit for the outputs, even if the author left the day before census date after working there for many years.

The case for non-portability

Stern and others set out two particular reasons for making this change, both compelling arguments.  First, the originating institution may have invested considerably in the recruitment, start up and future career of a faculty member, only to see the reward lost if the staff member moved near to the end of the cycle. This was described by Stern as ‘a distortion to investment incentives in the direction of short-termism’. Second, many vice-chancellors have complained about salary inflation, often a consequence of the originating institution paying a premium to retain credit for outputs which they had supported when the author has a competing offer, described by Stern as ‘rent-seeking by individuals’.

The Stern proposal of non-portability received only minority support in our consultation, albeit a significant minority.  Respondents in favour cited the two reasons above, often with narrative comments which indicated the strength of their support.

Concerns raised

But a majority of respondents opposed the proposals, often with equally persuasive arguments. Considerable concern was raised about the implications and implementation of the proposal. Many suggested its effects were likely to be worse than the ‘gaming’ it is intended to address. The reasons included the difficulty of determining the institution where the output was generated, stagnation in staff mobility, deliberate delays in publication of outputs, unfairness to early career researchers, and penalties for new units. We heard particular criticism about changing the rules mid-cycle. We are now slightly more than halfway through the REF submission cycle (December 2013 to December 2020), and many significant staffing and financial decisions have already been taken. It was felt that, in principle, the REF should not impose retrospective conditions.

Implementing Stern

So what do the funding bodies propose to do? We note that the respondents opposing the proposals made strong points, but they did not undermine the strength of the central argument about ‘disincentive to investment’ and ‘rent-seeking’. These points are compelling, and so we accept, and will implement, the Stern proposal that credit for outputs should remain with the originating institution.

But we also accept that retrospective decisions are unfair. This means we need transitional arrangements for this cycle.  We also accept that some cogent points were made by those in favour of the status quo (notably about early career researchers) and that the rules must include some exceptions or refinements to address those points.

We have recently been testing out proposals for transitional arrangements in several sounding-board meetings. Overwhelmingly, transitional arrangements have been welcomed in principle. But we have also heard criticism of the complexity and consequent burden of the transitional proposals we have suggested. As with staff selection, this reflects the challenge of developing a simple set of transitional arrangements that address the full detail of the concerns.

We now propose to discuss this more widely, though over a short time period. There is, however, a simple proposal which adds no new burden over the previous arrangements – although it will involve less precision than full-blown transitional arrangements. We propose it on the grounds that the sector is concerned about burden, and saving on burden may more than compensate for any loss of precision.

We simply propose to implement the Stern proposal: that the institution where the research output was demonstrably generated and at which the member of staff was employed should retain full credit. Following our open access guidelines we would expect the originating institution to hold details of the publication in their repository. However in this cycle, credit will also go to the receiving institution. This will result in some outputs being credited to more than one institution. This already happens on a large scale with co-authored publications, where the credit is not split between authors. Instead, all authors in different institutions receive full credit.

An alternative hybrid approach would be to introduce limited non-portability from a set point in time (in itself contentious).  Eligibility to return outputs to the REF would depend on when the staff involved were recruited relative to this date.  This would mean the operation of two rules around portability for this exercise, with the outputs of staff employed before the specified date falling under the 2014 rules of full portability. Outputs from staff employed after this date would fall under the new rules. The new rules would allow a limited number of outputs to transfer with staff (for example, two outputs). Eligibility to submit outputs would otherwise be linked to where the staff member was employed when the output was first made publicly available. Our sounding-boards have welcomed the prospect of transitional arrangements, but blanched at the complexity when we have explained this hybrid model.

Whichever of these special arrangements is used for this cycle, a review will follow, to inform how the policy on portability is best implemented in subsequent exercises. We will start with the expectation that the Stern proposals will be implemented, but with scope to acknowledge the points made by those who favoured no change in the portability rules.

We now need to open this choice of transition arrangements out to the community. It is a question of burden and complexity vs simplicity, low cost and minor loss of precision. Which is it to be?

Please comment below, email any views to researchpolicy@hefce.ac.uk or take part in the general discussion we are holding from now until September.