House of Commons - Open Access - Business, Innovation and Skills Committee

Open Access - Business, Innovation and Skills Committee Contents

4 The transition to open access: costs and hidden costs

Pure Gold and Hybrid Gold

36. Neither the Finch Report nor the Government nor RCUK make any detailed distinction between hybrid journals and fully open access or "pure" Gold journals.[51] Consequently, publishers operating either type of model are therefore equally eligible to receive APCs paid by UK research funders and research organisations.[52] We received many submissions arguing that this would have two unintended consequences: allowing "double dipping"[53] by hybrid publishers, and promoting artificially high APCs.

37. Due to the Government's clear preference for Gold funded by APCs, and the fact that RCUK has said it will make available £100m to fund the implementation of this policy[54], it seems likely that publishers will be incentivised to add a Gold option to their subscription model, and become hybrid as a direct result of the Government's open access policy. Evidence from Reed Elsevier supported this assumption:

Since 1 April, when the new UK policies were implemented, there has been a differential pattern of open-access publishing. There are many more hybrid open-access journals, in particular, in the landscape now. The Finch group was looking forward to that new landscape. [...]With the Government's policy that we are all implementing, we will see an increase in the amount of hybrid open-access publishing done at scale.[55]

38. It seems likely that for-profit publishers of hybrid journals with a diversified income comprised of both subscriptions and APCs will seek to maintain that joint revenue stream for as long as possible, especially if in doing so they would be actively complying with Government and funder policies. Clearly, one of the factors that will determine the cost of any move to open access in the UK will be the extent to which subscriptions need to be maintained, while payments of APCs increase. This is referred to in the Finch Report as "stickiness".[56] As we have seen, even if the UK reaches total open access by paying for APCs for 100% of research produced in this country, it must continue to pay subscription charges in order to access the vast majority of the rest.

39. The Government and RCUK have reiterated the Finch Report's conclusion[57] that HEIs and others must "negotiate hard"[58] with publishers to drive down subscription charges to reflect the additional funding they are receiving through APCs. RCUK has said it expects publishers to introduce differential pricing to reflect its additional income.[59] We heard evidence from Reed Elsevier that it operates a strict no double dipping policy, which it operates by "modify[ing] [its] list prices two years in arrears to reflect the number of hybrid open-access articles that are published".[60] The evidence we received suggested that this cannot operate as a genuine rebate since the reduction in subscriptions is applied to all subscribers worldwide, regardless of whether they have paid APCs or not.[61] In this way, the Government's open access policy means that the UK subsidises access to research for the rest of the world.[62]

40. Our evidence suggested that average APCs charged by pure Gold publishers are considerably lower than those of publishers operating hybrid journals. The major Gold publisher PLOS published 11% of all peer reviewed research articles funded by the Wellcome Trust in 2012, for which the average APC was £1060. This was around two thirds of the average paid by the Wellcome Trust.[63] The published accounts[64] of PLOS reveal it is making a modest operating profit, and its publicly available price list[65] shows that it has not raised its prices since 2009, in contrast to subscription based publishers. Some submissions to this inquiry argued that lower APCs led to low quality publishing content, but this was contradicted by the evidence from PLOS whose largest journal PLOS ONE published in 2012 more peer-reviewed research papers funded by the Wellcome Trust, MRC, BBSRC and Cancer Research UK than any other journal.[66]

41. RCUK has undertaken to publish data on "how the open access block grants are being used, specifically the numbers of research papers which are being made open access through payment of an APC and the actual APCs being paid to publishers"[67]. We recommend that RCUK also requires data on subscription expenditure from UK HEIs to establish the impact of its policy on subscription purchasing and pricing.

Embargo periods

42. David Willetts, the Minister with responsibility for Open Access policy, explained his assessment of the advantages of Gold funded by APCs over Green as follows:

First of all, it is honest, it is explicit. There is a cost and a value to publishing and we should recognise it [...]It is explicitly recognised and I like that about it. I like the fact of course that you do get the work openly accessible straight away. The hidden cost in green is waiting six months, 12 months, 24 months, whatever it is, until the layperson outside the academic community can access a piece of work that he or she as a taxpayer has already paid for once. That is the second advantage.[68]

43. Evidence to our inquiry, however, does not back up this point of view and runs counter to the Minister's assessment. A 2011 study showed that 60% of journals allow immediate unembargoed self-archiving of the peer-reviewed version of the article, that is, immediate open access.[69] Where publishers impose an embargo, would-be readers can send an automatic email to the author requesting an e-copy.[70]

44. The Finch Report observed of embargoes that "the restrictions imposed by publishers seem to have succeeded so far in limiting any potential impact on take-up of subscriptions to their journals".[71] We heard conflicting views on whether short embargo periods pose a risk to publishers,[72] but there is no available evidence base to indicate that short or even zero embargoes cause cancellation of subscriptions. Evidence from the field of high energy physics shows that despite nearly 100% immediate, unembargoed deposit (Green), subscriptions have not been damaged.[73] The 4 million EU funded PEER (Publishing and the Ecology of European Research) project (2012) showed that traffic to journal websites increased when articles were made available through a publicly accessible repository,[74] possibly because interest grew as articles were disseminated more widely.

45. Several submissions argued that short embargo periods were more harmful to HASS (humanities, arts and social sciences) than STEM (science, technology, engineering and medicine) disciplines. The most frequently deployed argument in HASS subjects is that since these articles have longer citation half-lives (i.e. are referred to over a longer period) a longer embargo is necessary. The Finch Report was persuaded by this distinction, recommending embargoes for HASS subjects of 24 months and for STEM subjects of 12 months. We did not receive any evidence to support this recommendation, which was adopted by the Government and RCUK though their endorsement of the Publishers Association's decision tree.[75]

Figure 2: The Publishers Association's decision tree for policy on access to UK publicly funded research outputs[76]

46. Some of the evidence we received predicted that the open access policies of the Government and RCUK regarding embargoes would incentivise publishers to introduce embargoes or lengthen existing embargo requirements, in order to force UK authors to pay for Gold.[77] Longer embargoes result in reduced access to published research findings for the UK, and reduced access to the rest of the world. We heard reports of a UK publisher in social sciences switching from a zero embargo policy to require a 24 month embargo, citing its accordance with the recent change to UK open access policy.[78]

47. The Government's response to this was that the publisher's interpretation of UK policy was inaccurate, since Government and RCUK policy includes embargoes of up to 12 and 24 months respectively, as a "maximum allowance" rather than "a requirement".[79] However, the Government acknowledged that the publisher "was acting within the limits of UK policy" and acknowledged that "any move to extend embargo periods where shorter ones are otherwise sustainable for publishers would be counter to the overall aims of the Government OA policy".[80] The Government told us that the Publishers Association would "welcome" being informed of instances where "inappropriate" embargoes were implemented.[81]

48. Since then, the world's second largest journal publisher, Springer, has imposed a 12 month embargo on papers deposited in institutional repositories, where previously unembargoed deposit was permitted.[82]

49. We note the absence of evidence that short embargo periods harm subscription publishers. We have seen evidence that current UK open access policy risks incentivising publishers to introduce or increase embargo periods. This has serious implications for open access in the UK and the rest of the world. We agree with the Government that lengthened embargoes are counter to its aim to increase access.

50. The stated policy objective of the Government and RCUK is to increase access to publicly funded research. Long embargoes are a barrier to access. We recommend that the Government and RCUK revise their policies to place an upper limit of 6 month embargoes on STEM subject research and up to 12 month embargoes for HASS subject research, in line with RCUK's original policy published in July 2012.[83]

51. Given the importance of ensuring that UK open access policy does not result in reduced access in the UK or worldwide, the Government and RCUK must monitor and evaluate the impact of their open access policy on embargo lengths imposed by UK publishers. The impact on different subject areas must also be carefully monitored. That information must inform future meetings of the Finch Group and RCUK's reviews of open access policy.

Levels of Article Processing Charges

52. The Finch Report's economic modelling of the costs to the UK of moving towards open access used a range of APCs, beginning with a starting point of £1,450, a middle ground scenario of £1,750, and a higher scenario of £2,175. The Finch group arrived at the figure of £1,450, the lowest figure in its modelling, because it was the average APC paid by the Wellcome Trust at the time.[84]

53. We received a considerable volume of evidence to suggest that the prices of APCs used in the Finch Report's modelling were very high. Indeed, BIS's own 'Economic Analysis of the cost of alternative options for open access in the UK' stated that "the Finch Group reported £12m spent nationally in Open Access in 2010 for around 16,000 open access articles",[85] which works out at an average price per article of only £750. A 2012 study found the global average APC of open access journals to be £571.[86] The evidence of major open access publisher PLOS gave more detail about Wellcome's average payments for APCs:

One of our concerns, and a concern of many pure open-access publishers, is to ensure that we have an effective, functioning market in article-processing charges. If we do not have a market and we do not have some price sensitivity for authors, there is a concern that prices can run out of control. One example of that is again in our evidence, where the charges we made to the Wellcome Trust in 2012 were actually around about two thirds of the average of what the Wellcome Trust pays. Of course, the Wellcome Trust pays those bills, so it is very easy to spend more money when you are spending someone else's money.[87]

54. While the Finch Report did not attempt to set market levels for APCs, the figures it used in its modelling undoubtedly created a risk that publishers would see these as a benchmark. RCUK explained in its evidence to us that in calculating the total amount of funding that it would make available to support Gold funded by APCs, "the average cost of an APC has been taken from the Finch Report (estimated as £1727 plus VAT)[88]." The Government confirmed that RCUK's expected spend of £100m on Gold OA over the next five years was based on the same figure.[89]

55. We have not seen any evidence to suggest that either the Government or RCUK have attempted to establish whether the figures used in the Finch Report's modelling accurately represented current APC averages. The minutes of the meeting of the Finch working group on 22 February 2012 record that two BIS economists who "provided assistance in economic modelling for the Group", also "worked on the advice to Ministers once the Group's report [was] finalised".[90] When BIS's own economic analysis was released in March 2013, the Rt Hon David Willetts MP wrote to the House of Lords Science and Technology Committee:

This analysis was done in mid 2012 prior to the publication of the recommendations of the Finch Group and prior to the announcement of RCUK policy. It was not intended as an evaluation of the policy that has now been adopted but it supports the Government preference for Gold access independently of Finch and RCUK.[91]

56. The same economist was involved with the Government's initial pre-Finch Report analysis of open access, the Finch Report modelling, and the Government's subsequent acceptance of those recommendations. This draws the independence of the Finch Report and its economic analysis into question. We are of the view that had the Finch working group commissioned entirely independent economic analysis this would have been avoided.

57. We conclude that the Finch Report, the Government and RCUK have failed to assess adequately the existing levels of APCs that are charged by a range of open access journals, both within the UK and worldwide, and instead formed a plan of expenditure based on payments to publishers that, compared to a range of benchmarks including APCs of the largest "pure" Gold publisher, are rather less than competitive.

58. We recommend that the Finch working group commissions an independent report on APC pricing, which should include average APC prices of pure Gold journals and hybrid journals, domestically and internationally.

59. We strongly support the recommendation of the Science and Technology Committee of the House of Lords that the Government undertake a full cost-benefit analysis of open access policy, including the impact on different subject areas. This analysis must include data to reflect actual rather than projected costs during the transition period.

Affordability of APCs for authors and UK research organisations

60. A large number of contributors to our inquiry argued that the block grants provided by RCUK to fund APCs would be insufficient to facilitate the transition to APC funded publishing at scale, and would not compensate research organisations for the increased costs of both maintaining subscription purchasing and funding APCs. The Head of Politics at the University of Exeter estimated that block grant funds for a department of 35 full—time academics and many more post-doctoral students would cover only a single APC for the entire department.[92] We asked the Minister whether this raised concerns around affordability of APCs. He responded:

The University of Exeter runs on a budget of several hundred million pounds. Part of what Exeter does is attract academics who wish to publish research findings. It is reasonable to say that part of the academic and research mission of Exeter is to ensure that good quality research is funded. This argument that I get a bit presents a picture of the relationship between academics and university management that I do not recognise.[93]

61. Though not entirely clear, this would seem to suggest that the Government's view is that institutions must look elsewhere in their budgets to find the funds to cover the costs of maintaining subscriptions and paying APCs, which raises serious issues for HEIs. Several representatives of the HE community said that in contrast to Gold, Green was achievable and sustainable for their institutions.[94] This was supported by recent research which suggested that with APCs at £1500 (i.e. at the lower end of the range used in the Finch Report's modelling), adopting Gold would cost UK universities 12 times the cost of adopting Green, and for the more research intensive universities, Gold could cost 25 times as much as going Green.[95] Many submissions referred to the administrative and management costs for universities of processing, distributing and monitoring the block grant for open access. These will need to be monitored.

62. We also received differing views on the total opportunity cost represented by the Government's open access policy. The Government submitted that this "will only be about one per cent of the science budget of £4.6 billion per annum".[96] The Russell Group described this in different terms:

The funding for open access is being taken directly from Research Council budgets, which could otherwise be used to support research, doctoral training and knowledge transfer activities that have a more direct impact on economic growth, jobs and future quality of life. RCUK has earmarked £37 million over two years for Gold open access: £17 million in the first year of policy implementation and £20 million in year two. This is equivalent to the stipends and fees for 964 doctoral students in year one, 1,135 in year two and even more in subsequent years as open access requirements increase. In research terms this is three times more than EPSRC is investing in the UK Catalysis Hub at Harwell, which will have major applications in energy, manufacturing and healthcare. Over a five year period, RCUK has estimated it will need to put over £100 million into Gold open access, similar to the whole of the Government's recent announcements on capital spending for synthetic biology research.[97]

63. At a time when the budgets of research organisations and HEIs are under great pressure, it is unacceptable that the Government has issued, without public consultation, an open access policy that will require considerable subsidy from research budgets in order to maintain journal subscriptions and cover APCs. Significant public investment has already been made in institutional repositories, of which there are 120 in the UK, and they could represent a more cost-effective and sustainable route to full open access.

64. We are concerned that the expectation appears to be that universities and research organisations will fund the balance of APCs and open access costs from their own reserves. We look to the Government and RCUK to mitigate against the impact on university budgets. The Government must not underestimate the significance of this issue.

The shared ultimate goal of full Gold open access

65. Within the longstanding debate on open access policy, proponents for Green and Gold are often described as being diametrically opposed in their views. In fact, the evidence we have seen demonstrates considerable common ground between the two positions. Professor Stevan Harnad, an advocate of Green, argued that a well managed transition process using strong deposit mandates would:

Induce the transition to Gold open access [...] to allow publishing to evolve toward the obvious, optimal and inevitable outcome: Gold open access at a fair, affordable, sustainable price, with no inflation and no double-dipping.[98]

66. The most comprehensive economic analysis available supports this. A 2012 report commissioned by the UK Open Access Implementation Group assessed the costs and benefits of Gold for UK research institutions and found that in a fully open access world, the net benefits of Gold would be greater than those of Green. In their article 'Planting the green seeds for a golden harvest', Dr Alma Swan and Professor John Houghton said:

If open access were adopted worldwide, the net benefits of Gold open access would exceed those of Green open access. However, we are not in an open access world [...] At the institutional level, during a transitional period when subscriptions are maintained, the cost of unilaterally adopting Green open access is much lower than the cost of Gold open access—with Green open access self-archiving costing average institutions sampled around one-fifth the amount that Gold open access might cost, and as little as one-tenth as much for the most research intensive university. Hence, we conclude that the most affordable and cost-effective means of moving towards open access is through Green open access, which can be adopted unilaterally at the funder, institutional, sectoral and national levels at relatively little cost.[99]

67. The key to the success or failure of the UK's open access policy therefore is the management of the transitional stage to full open access. The almost immediate acceptance of the Finch Report's recommendations by the Government and RCUK in the formation of their policies, together with the radical change in policy direction from predominantly Green policies and mandates towards Gold funded by APCs, has resulted in insufficient consideration being given to this transitional period. As discussed earlier, the understanding of how that transition might be most effectively managed for all stakeholders has also been hampered by the Finch Report's failure to evaluate the success of existing UK open access policies and the vital role the Green route will make to that transition.

68. Slowly, the messages from RCUK are changing. There have been more than half a dozen substantive revisions to RCUK policy and guidance since its official issue in July 2012. In January 2013, RCUK said that "at the current time the Gold option provides the best way of delivering immediate, non-restricted access to research papers".[100] By April, its position had changed to "we have a preference [for Gold]... but it is not an absolute preference," and "we should not place all our bets on one particular horse".[101] RCUK's current position is that even if APCs are available, either the Green or Gold route is acceptable, with authors or their institutions free to decide.[102] Both routes now count towards RCUK's expected rates of compliance, which before recognised only Gold open access.[103] The decision tree, originally produced by the Publishers Association and publicly endorsed by both the Government and RCUK, therefore does not represent RCUK's position.[104]

69. The pro-active stance the Government has taken in the formation of open access policy is to be welcomed. However, we are of the view that the Government has failed to communicate effectively that Gold open access is the ultimate goal at the end of a transition phase. Because insufficient attention has been given to the transitional route, the Government has neglected the opportunity to ensure that costs are constrained, and that institutions and research authors are convinced of the merits of open access policy.

70. The Government and RCUK should clarify that Gold open access is the ultimate goal of, rather than the primary route to, their open access policies. We recommend that the Government and RCUK reconsider their preference for Gold open access during the five year transition period, and give due regard to the evidence of the vital role that Green open access and repositories have to play as the UK moves towards full open access.

71. RCUK's current guidance provides that the choice of Green or Gold open access lies with the author and the author's institution, even if the Gold option is available from the publisher. This is incompatible with the Publishers Association's decision tree, and RCUK should therefore withdraw its endorsement of the decision tree as soon as possible, to avoid further confusion within the academic and publishing communities.

51 See Glossary Back

52 The Finch Report, recommendation ii Back

53 See Glossary Back

54 Ev 39 Back

55 Q2 Back

56 The Finch Report, p 38 Back

57 The Finch Report, p 11 Back

58 Ev 50 Back

59 Ibid Back

60 Q17 Back

61 Ev 75 Back

62 Ibid Back