Accessibility, sustainability, excellence: how to expand access to research publications/Annex E: Analysis of the Possible Financial Impact of Shifts to Open Access Publishing in the UK

Annex E

TRANSITION TO THE OPEN ROAD
GOLD SCENARIO MODELLING UPDATE

Prepared by:

Cambridge Economic Policy Associates Ltd.

INTRODUCTION

Research Information Network (RIN) has commissioned CEPA to provide analysis of the costs to the UK of transitioning to a situation where there is a greater proportion of author-side payments or “Gold Open Access”. This analysis is a follow-up to previous work by RIN, CEPA and Mark Ware (i.e. the Heading for the Open Road report, hereafter referred to as ‘OR Report’).

In response to similar work which has been undertaken by other stakeholders, this note carries out additional sensitivity analysis on the Gold Scenario in the OR Report.

There are two key conceptual differences between the results in this note and those in the OR Report:

• First, the sensitivities here have been based on UK article output and funding in 2010, as opposed to a forecast 2015 Baseline (in the OR report), which itself was dependent on a series of assumptions.

• Second, we have set the starting point article processing charge (APC) at the ‘breakeven’ point for academic institutions (as opposed to all institutions in the OR report). This means that the sensitivities are concerned with the point at which it is possible that UK academic institutions would become indifferent (at the margin) between subscriptions and author-side payments.

Notwithstanding these changes, all of the caveats relating to the Gold Scenario modelling set out in the OR Report apply here. (For example, the breakeven points relate to ‘average’ institutions; and the analysis of costs assumes that publishers reduce subscription prices in proportion to the increase in Gold OA articles).

The sensitivities covered in this note are:

Sensitivity 1	Sensitivity 2	Sensitivity 3	Sensitivity 4
The level of article processing charge (APC)	The level of Gold uptake in the UK and globally	The level of Gold uptake in rest of world versus UK	The % of corresponding authors from the UK

ASSUMPTIONS

Assumptions that we have used to define the Gold OA Scenario in this note, which differ from those in the OR Report, include:

a) Number of articles produced. As requested by RIN, we have updated the model inputs to reflect two estimates, from a recent BIS commissioned report, for article production:

• The UK produced 123,594 articles in 2010.
• The world produced 1,935,954 articles in 2010.^[1]

These figures are roughly 20% higher than those previously assumed in the OR Report 2010 baseline (98,280 for UK and 1,576,218 for the world).

b) Library costs. In the OR Report we used top-down estimates of total library costs to calculate the per article library costs which were used in the model. In light of the assumed increase in articles produced (see (a) above) and the desire to keep total library costs consistent with the OR Report, we made the per article library costs have been reduced in order to leave total library costs unchanged.

c) Subscription costs from 'other' organisations. The model splits costs between academic institutions (RLUK, HEC, etc.) and other organisations (government and independent/corporate libraries). In the OR Report we assumed that these 'other' libraries had access to a specific number of articles/journals. Having increased the number of articles produced in the 2010 baseline, we adjusted the model to ensure that the number of subscriptions for ‘other’ organisations remained constant (i.e. didn't increase).

Given these changes to the modelling assumptions (and the conceptual differences referred to above), the estimates in this note differ from those presented in the OR Report. The 2010 estimates against which the sensitivities are compared are shown in Table 0.1 below.

Table 0.1: UK 2010 publishing and distribution funding

	Cost Category	2010
Academic Institutions	Author-side payments	11.1
	Subscriptions	112.0
	Access Provision	52.1
	Transition costs	n/a
	Total	175.2
Others	Author-side payments	2.8
	Subscriptions and other revenues[1]	67.5
	Access Provision	51.4
	Transition costs	n/a
	Total	121.7
Annual Total		296.6

For ease of reference, Table 0.2 provides details of the key assumptions in the OR Report cost modelling that relate to the Gold Scenario. The same assumptions are used in the Gold Scenario modelled for the purpose of this note.

Table 0.2: Key cost-related assumptions that affect the Gold Scenario

Input	Discipline (location)	2015 Baseline[2]	Gold Scenario
Article allocation — Share of articles published via APCs	S/T/S (UK and Global)	5.4%	15%
	M (UK and Global)	12.8%	40%
	A/H (UK and Global)	1%	5%

RESULTS - SENSITIVITIES

For our sensitivity analysis, our starting point is the updated Gold Scenario (hereafter referred to as the ‘Central Case’). As stated in Section 1, we run sensitivities based on:

• The level of article processing charge (APC). The variations are the percentage change from the APC, set at the level where academic institutions breakeven, in the Central Case, and we consider sensitivities of ±10%, ±20% and +50% of this figure.

• The level of Gold uptake in the UK and rest of the world. This sensitivity considers the impact from changing the level of Gold uptake globally (i.e. by the same proportion in both the UK and the rest of the world). In the Central Case the level of Gold uptake is 23%, which is a weighted average of Science and Technology articles (15% open access), Mathematics (40%) and Arts and Humanities (5%). We consider average levels of Gold open access ranging from 10% (i.e. lower than the Central Case) up to 50%.

• Gold uptake in rest of world versus the UK. This sensitivity explores the impact on the UK of the rest of the world moving to a lower proportion of author-side payments than the UK.

• Percentage of UK-produced articles with a UK corresponding author. As stated above, UK article production (i.e. published article having at least one UK author) is 123,594 in 2010, but some of these articles may not have a UK corresponding author. If we assume that author-side payments will be made by the corresponding author only, then a reduction in the percentage of UK corresponding authors would reduce the costs to the UK. An estimate arising from the BIS study, but unpublished, is that 65% of UK article production has a UK corresponding author listed.

The Central Case that we have used in the report is one in which the APC is set a level at which there is a zero impact on academic institutions (on average).

Table 0.3 below provides an explanation of each of the cost categories shown in our results. These categories are consistent with the OR report.

Table 0.3: Explanation of Cost Categories

Cost Category	Description
Total average annual net cash costs	Costs that will occur both now (as of 2015) and in perpetuity. These include: Transition Costs. The additional ongoing cash costs incurred under the transition to the Gold Scenario, which are split between academic institutions and other organisations. This includes the costs of advocacy, marketing, communications and information provision regarding Gold open access at each institution. Steady state costs. Ongoing cash costs of the scholarly communication system, which are split between academic institutions and other organisations. Costs include: Publication & Distribution costs. Peer review management and editing of articles received, composition of approved articles into journals, marketing, and transportation of journals to libraries and other buyers. Access provision costs. Activities carried out mainly by libraries to make journals accessible to end users: Procurement, receiving and indexing, archiving, management and administration, IT systems, and library storage.
20-year NPV Costs and Savings	This calculation takes the future stream of net costs over a 20 year period (starting in 2015), and determines the net present value of those future costs/savings by discounting their value at 3.5% per year.[3]

Some additional notes to consider in relation to the cost categories:

• Costs are split between academic institutions (e.g. RLUK, etc.) and other organisations (government, corporate and independent libraries).
• NPV stands for 'Net Present Value'.
• Figures exclude 'transaction costs' ('one-off' costs of moving to the Gold Scenario as they are not ongoing).
• Figures exclude 'non-cash costs' (peer review 'time' costs, and user search/print costs).

Sensitivity 1 - Level of APC

Table 0.4: Results of sensitivity 1 (figures in £m)^[4]

	Cost Category	Central Case	Sensitivity 1: APC
			-20%	-10%	+10%	+20%	+50%
Academic Institutions	Author-side payments	22.0	17.6 (-4.4)	19.8 (-2.2)	24.2 (+2.2)	26.4 (+4.4)	33.0 (+11.0)
	Subscriptions	-18.7	-18.7 (0)	-18.7 (0)	-18.7 (0)	-18.7 (0)	-18.7 (0)
	Access Provision	-3.5	-3.5 (0)	-3.5 (0)	-3.5 (0)	-3.5 (0)	-3.5 (0)
	Transition costs	0.2	Constant at £200k for academic institutions
	Total	0.0	-4.4 (-4.4)	-2.2 (-2.2)	2.2 (+2.2)	4.4 (+4.4)	11.0 (+11.0)
Others	Author-side payments	5.5	4.4 (-1.1)	5.0 (-0.6)	6.1 (+0.6)	6.6 (+1.1)	8.3 (+2.8)
	Subscriptions and other revenues[5]	-9.5	-9.8 (-0.3)	-9.6 (-0.2)	-9.3 (+0.2)	-9.1 (+0.3)	-8.6 (+1.2)
	Access Provision	-2.1	-2.1 (0)	-2.1 (0)	-2.1 (0)	-2.1 (0)	-2.1 (0)
	Transition costs	0.8	Constant at £800k for Others
	Total	-5.2	-6.7 (-1.4)	-6.0 (-0.7)	-4.5 (+0.7)	-3.8 (+1.4)	-1.7 (+3.5)
Annual Total		-5.2	-11.1 (-5.8)	-8.1 (-2.9)	-2.3 (+2.9)	0.6 (+5.8)	9.4 (+14.6)
20-year NPV Net Total		-76.9	-162.7 (-85.8)	-119.8 (-42.9)	-34.0 (+42.9)	9.0 (+85.8)	137.7 (+214.6)

Points to note are as follows:

• Even in the Central Case other sectors of the UK economy (‘Others’) make significant savings (c£5.2m). This reflects the fact that the non-academic institutions (i.e. the Others’ produce a small number of articles that they would have to pay for in the Gold Scenario—and this cost is small relative to the savings that they make from not having to pay subscription fees. The APC would have to increase significantly, well over 50%, for this group to incur cash costs from a global shift to author-side payments.
• Variation in APC can have a fairly significant impact on UK academic institutions annual cash costs. If the APC rises by 20%, UK academic institutions annual net costs will rise by £4.4m. However, for the UK as a whole some of these cash costs are offset by savings in the other sectors of the economy. The ‘Others’ save £3.8m, meaning that the additional costs to the UK are only £0.6m.

• If the APC was 50% greater than the academic institution breakeven point then costs to the institutions would be £11.0m more than the Central Case, with Others’ savings only offsetting this by £1.7m.

Sensitivity 2 - level of Gold uptake in the UK and rest of the world

Table 0.5: Results of sensitivity 2 (figures in £m)

	Cost Category	Central Case	Sensitivity 2: level of Gold uptake in the UK and rest of the world
			10%	30%	40%	50%
Academic Institutions	Author-side payments	22.0	3.2 (-18.9)	31.6 (+9.6)	45.9 (+23.8)	60.1 (+38.1)
	Subscriptions	-18.7	-2.7 (+16)	-26.9 (-8.2)	-39.0 (-20.3)	-51.1 (-32.4)
	Access Provision	-3.5	-0.5 (+3)	-5.0 (-1.5)	-7.2 (-3.8)	-9.5 (-6)
	Transition costs	0.2	Constant at £200k for academic institutions
	Total	0.0	0.2 (+0.2)	-0.0 (-0.1)	-0.2 (-0.2)	-0.3 (-0.3)
Others	Author-side payments	5.5	0.8 (-4.7)	7.9 (+2.4)	11.5 (+6)	15.0 (+9.5)
	Subscriptions and other revenues	-9.5	-1.4 (+8.1)	-13.6 (-4.1)	-19.7 (-10.2)	-25.8 (-16.3)
	Access Provision	-2.1	-0.3 (+1.8)	-3.0 (-0.9)	-4.3 (-2.3)	-5.7 (-3.6)
	Transition Costs	0.8	Constant at £800k for Others
	Total	-5.2	-0.0 (+5.2)	-7.9 (-2.6)	-11.8 (-6.5)	-15.7 (-10.4)
Annual Total		-5.2	-0.0 (+5.3)	-7.9 (-2.7)	-11.9 (-6.7)	-16.0 (10.7)
20-year NPV Net Total		-76.9	1.3 (+78.2)	-116.7 (-39.8)	-175.7 (-98.8)	-234.7 (-157.8)

Key points to note are as follows:

• Varying the level of Gold uptake does not have a significant impact on academic institutions’ net costs, because the APC has been set at the academic institutions breakeven point. For example, an increase from 23% (in the Central Case) to 50% causes a small change in annual academic institutions’ net costs (a saving of £0.3m). However, the savings for Others is much greater, £15.7m when the level of Gold uptake is set at 50%.

• Substantial savings are made in access provision as the level of Gold increases. This results from libraries undertaking fewer activities in relation to procurement of journals and administration.

Sensitivity 3 - Level of uptake in rest of world versus UK

Table 0.6: Results of sensitivity 3 (figures in £m)

	Cost Category	Central Case	Sensitivity 3: level of uptake in the rest of world versus UK
			80%	60%	40%	20%
Academic Institutions	Author-side payments	22.0	22.0 (0)	22.0 (0)	22.0 (0)	22.0 (0)
	Subscriptions	-18.7	-15.2 (+3.5)	-11.7 (+7.1)	-8.1 (+10.6)	-4.6 (+14.1)
	Access Provision	-3.5	-2.8 (+0.7)	-2.2 (+1.3)	-1.5 (+2)	-0.9 (+2.6)
	Transition costs	0.2	Constant at £200k for academic institutions
	Total	0.0	4.2 (+4.2)	8.4 (+8.4)	12.6 (+12.6)	16.8 (+16.8)
Others	Author-side payments	5.5	5.5 (0)	5.5 (0)	5.5 (0)	5.5 (0)
	Subscriptions and other revenues	-9.5	-7.7 (+1.8)	-5.9 (+3.6)	-4.1 (+5.3)	-2.3 (+7.1)
	Access Provision	-2.1	-1.7 (+0.4)	-1.3 (+0.8)	-0.9 (+1.2)	-0.5 (+1.6)
	Transition Costs	0.8	Constant at £800k for Others
	Total	-5.2	-3.1 (+2.2)	-0.9 (+4.4)	1.3 (+6.5)	3.5 (+8.7)
Annual Total		-5.2	1.1 (+6.4)	7.5 (+12.7)	13.9 (+19.1)	20.2 (+25.5)
20-year NPV Net Total		-76.9	16.7 (+93.6)	110.4 (+187.2)	204.0 (+280.9)	297.6 (+374.5)

Key points to note are as follows:

• If, as this sensitivity shows, the rest of the world has a lower uptake of author-side payments compared to the UK (i.e. a lack of global participation) the increase in costs to the UK is potentially very large. For example, if the rest of the world only has 20% of the UK's level of Gold uptake, UK annual costs increase by £25.5m relative to the Central Case.

• Looking at the annual costs by group, around two thirds of these increased costs are incurred by academic institutions. Author-side payments do not change (as expected) from the Central Case, but subscription costs increase significantly as the rest of the world's level of author-side payment reduces.

Sensitivity 4 - UK corresponding authors

Table 0.7: Results of sensitivity 4 (figures in £m)

	Cost Category	Central Case	Sensitivity 4: UK corresponding authors as % of articles with UK authors
			85%	75%	65%	55%
Academic Institutions	Author-side payments	22.0	18.2 (-3.9)	16.0 (-6.1)	13.8 (-8.3)	11.6 (-10.5)
	Subscriptions	-18.7	-18.7 (0)	-18.7 (0)	-18.7 (0)	-18.7 (0)
	Access Provision	-3.5	-3.5 (0)	-3.5 (0)	-3.5 (0)	-3.5 (0)
	Transition costs	0.2	Constant at £200k for academic institutions
	Total	0.0	-3.8 (-3.8)	-6.0 (-6.0)	-8.2 (-8.2)	-10.5 (-10.6)
Others	Author-side payments	5.5	4.5 (-1)	4.0 (-1.5)	3.4 (-2.1)	2.9 (-2.6)
	Subscriptions and other revenues	-9.5	-9.5 (0)	-9.5 (0)	-9.5 (0)	-9.5 (0)
	Access Provision	-2.1	-2.1 (0)	-2.1 (0)	-2.1 (0)	-2.1 (0)
	Transition Costs	0.8	Constant at £800k for Others
	Total	-5.2	-6.2 (-1)	-6.8 (-1.5)	-7.3 (-2.1)	-7.9 (-2.6)
Annual Total		-5.2	-10.0 (-4.8)	-12.8 (-7.6)	-15.6 (-10.3)	-18.3 (-13.1)
20-year NPV Net Total		-76.9	-147.8 (-70.9)	-188.3 (-111.4)	-228.8 (-151.9)	-269.3 (-192.4)

Key points to note are as follows:

• In 2010 it is assumed that UK authors contributed to approximately 123,600 articles. Based on the BIS study’s estimate of 65% for corresponding authors the UK would make a large net savings, of £15.6m, in comparison to the 2010 baseline (an increase in savings over the Central case of £10.3m). These savings are split relatively evenly between academic institutions (£8.2m) and others (£7.3m).

• If there were a greater number of corresponding authors, or a greater number of UK authors funding APCs, the savings would fall. For example, if UK authors funded 85% of articles they contributed to then the annual savings to the UK would be £10m (£4.8m greater than in the Central Case). Over 60% of these savings (£6.2m) would be to Others.

• The proportion of savings to academic institutions increases as the number of corresponding authors falls. This reflects the fact that the majority (over 80%) of articles produced in the UK are by authors in academic institutions.

• COMBINED SENSITIVITY

Assumptions:

• APC 20% higher than central case;

• UK Gold uptake is 50% (weighted average across different subjects);

• Rest of the world Gold uptake is 25% (i.e. half of the UK);

• UK pays for 75% of articles containing UK authors (i.e. 75% corresponding authors).

Table 0.8: Results of additional sensitivity (figures in £m)

	Cost Category	Central Case	Sensitivity
Academic Institutions	Author-side payments	22.0	53.7 (+31.7)
	Subscriptions	-18.7	-22.6 (-3.9)
	Access Provision	-3.5	-4.2 (-0.7)
	Transition costs	0.2	Constant at £200k for ac. institutions
	Total	0.0	27.1 (+27.1)
Others	Author-side payments	5.5	13.4 (+7.9)
	Subscriptions and other revenues[6]	-9.5	-11.0 (-1.5)
	Access Provision	-2.1	-2.5 (-0.4)
	Transition costs	0.8	Constant at £800k for Others
	Total	-5.2	0.7 (+5.9)
Annual Total		-5.2	27.8

	Cost Category	Central Case	Sensitivity
			(+33.0)
20-year NPV Net Total		-76.9	409.1 (+486.0)

Notes:

• In central case, Academic Institutions are cash neutral.
• May be rounding errors

---

1. Both of these figures are taken from ‘International Comparative Performance of the UK Research Base’, A report prepared for the Department of Business, Innovation and Skills (BIS), Appendix F: Supplementary Data, page 26.
2. Estimates from Björk 2010.
3. The discount rate of 3.5% (real) is taken from the HM Treasury Green Book. This differs from the 6.8% used in the OR Report which was a real rate of 6.8% and chosen to remain consistent with Houghton (2009).
4. Note, some values may not add due to rounding.
5. Other revenue includes: Pay per view, reprints, memberships fees & individual subscriptions, and advertising.
6. Other revenue includes: Pay per view, reprints, memberships fees & individual subscriptions, and advertising.