ArticlePDF Available

Abstract and Figures

For many decades, the hyperinflation of subscription prices for scholarly journals have concerned scholarly institutions. After years of fruitless efforts to solve this “serials crisis”, open access has been proposed as the latest potential solution. However, also the prices for open access publishing are high and are rising well beyond inflation. What has been missing from the public discussion so far is a quantitative approach to determine the actual costs of efficiently publishing a scholarly article using state-of-the-art technologies, such that informed decisions can be made as to appropriate price levels. Here we provide a granular, step-by-step calculation of the costs associated with publishing primary research articles, from submission, through peer-review, to publication, indexing and archiving. We find that these costs range from less than US$200 per article in modern, large scale publishing platforms using post-publication peer-review, to about US$1,000 per article in prestigious journals with rejection rates exceeding 90%. The publication costs for a representative scholarly article today come to lie at around US$400. We discuss the additional non-publication items that make up the difference between publication costs and final price.
Content may be subject to copyright.
University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln
Copyright, Fair Use, Scholarly Communication, etc. Libraries at University of Nebraska-Lincoln
Assessing the size of the aordability problem in
scholarly publishing
Alexander Grossman
HTWK Leipzig, Fakultät Informatik und Medien,,
Björn Brembs
University of Regensburg,
Follow this and additional works at: h7ps://
Part of the Intellectual Property Law Commons,Scholarly Communication Commons, and the
Scholarly Publishing Commons
6is Article is brought to you for free and open access by the Libraries at University of Nebraska-Lincoln at DigitalCommons@University of Nebraska
- Lincoln. It has been accepted for inclusion in Copyright, Fair Use, Scholarly Communication, etc. by an authorized administrator of
DigitalCommons@University of Nebraska - Lincoln.
Grossman, Alexander and Brembs, Björn, "Assessing the size of the a5ordability problem in scholarly publishing" (2019). Copyright,
Fair Use, Scholarly Communication, etc.. 112.
Assessing the size of the
affordability problem in
scholarly publishing
Alexander Grossmann1; Björn Brembs2
1 HTWK Leipzig, Fakultät Informatik und Medien, Karl-Liebknecht-Straße 145, 04277 Leipzig, Ger-
2 University of Regensburg, Institute of Zoology Neurogenetics, Universitätsstraße 31, 93040 Re-
gensburg, Germany,
For many decades, the hyperinflation of subscription prices for scholarly journals
have concerned scholarly institutions. After years of fruitless efforts to solve this
“serials crisis”, open access has been proposed as the latest potential solution. How-
ever, also the prices for open access publishing are high and are rising well beyond
inflation. What has been missing from the public discussion so far is a quantitative
approach to determine the actual costs of efficiently publishing a scholarly article
using state-of-the-art technologies, such that informed decisions can be made as to
appropriate price levels. Here we provide a granular, step-by-step calculation of the
costs associated with publishing primary research articles, from submission,
through peer-review, to publication, indexing and archiving. We find that these
costs range from less than US$200 per article in modern, large scale publishing
platforms using post-publication peer-review, to about US$1,000 per article in pres-
tigious journals with rejection rates exceeding 90%. The publication costs for a rep-
resentative scholarly article today come to lie at around US$400. We discuss the
additional non-publication items that make up the difference between publication
costs and final price.
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Introduction 1
The affordability problem of scholarly publish-2
ing, i.e., the hyperinflationary price increases with 3
stagnating library budgets, has been discussed for 4
decades (see, e.g., Chan 2004; Harnad et al. 2004; 5
Douglas 1990; Fisher 2008; Houghton 2001; 6
Tananbaum 2003; Rose-Wiles 2011). In recent years, 7
perhaps precipitated by some so-called ‘gold’ open ac-8
cess (OA) journals charging article-processing charges 9
(APCs; fees usually charged to authors or their institu-10
tions upon acceptance for publishing an article and 11
making it openly available), the average cost of an arti-12
cle has emerged as a useful measure with which to 13
compare different business models. However, most 14
authors refer to the prices charged by the publisher, 15
not the actual cost to the publisher (e.g., Van Noorden 16
2013; Schimmer et al. 2015; Odlyzko 2013; Johnson et 17
al. 2018). One consequence of this mis-attribution is a 18
potential overestimation of the actual costs of schol-19
arly publishing due to the inclusion of the business 20
models and pricing strategies of publishers into the 21
calculation. To close this gap, here we provide a bot-22
tom-up calculation of the cost of efforts and services 23
which are required to achieve a certain service level in 24
order to publish an academic journal article. We com-25
pare our cost estimate with the current pricing 26
schemes of publishers. 27
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Traditionally, access to scholarly publications 28
has been provided through a subscription model. 29
Non-disclosure agreements, commonly used by sub-30
scription publishers today (with the explicit intent to 31
increase prices (Tempest 2013)), make it difficult to cal-32
culate per-article prices at the level of journals, pub-33
lishers or countries. However, it is known how many 34
scholarly articles are being published every year on a 35
world-wide basis (2.4 million in 2017, (White 2019)) 36
and there are converging estimates on the subscrip-37
tion revenue spent world-wide each year (approx. 38
US$10 billion; Van Noorden 2013; Schimmer et al. 39
2015; Odlyzko 2013; Johnson et al. 2018). Dividing 40
these two figures leads to a widely agreed per-article 41
price of approx. US$5,000 paid largely by libraries for 42
the subscription system (Johnson et al. 2018). Both fig-43
ures are reportedly slightly higher today, but the final 44
per article price is relatively unchanged and still re-45
markably close to a long-standing US$4,000/article es-46
timate (Odlyzko 1995; Johnson et al. 2018). Taken to-47
gether, with both the revenue and the publication vol-48
ume increasing over the last decades, the per-article 49
price of the subscription system has remained rela-50
tively constant between US$4,000-5,000, further vali-51
dating the value of this measure. 52
While most OA journals do not charge APCs (or 53
other author-facing fees, such as submission fees) and 54
instead finance their services via alternative routes 55
(71% of journals listed in the Directory of Open Access 56
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Journals, DOAJ), most OA articles are being published 57
in the minority of journals which do charge APCs (58%, 58
Crawford 2019). So far, in contrast to subscription 59
prices, APCs are commonly not covered by non-disclo-60
sure agreements. On the contrary, most journals pub-61
licly list their APCs. Moreover, in those cases where 62
APCs are paid by research organizations, universities 63
or academic libraries on behalf of their authors, there 64
are data available on a more granular basis compared 65
to the subscription-based business model. For in-66
stance, Jahn and Tullney calculated from APCs for 67
7,417 journals which have been paid by 30 German ac-68
ademic libraries between 2005 and 2015 an average 69
APC of 1,298€ (~US$1,470)(Jahn & Tullney 2016). In 70
contrast, Schimmer et al. (2015) project an average 71
APC of 2,000€ (~US$2,260) for their scenario of transi-72
tioning to a full OA system. In a sample covering the 73
USA and Canada, APCs averaged US$1,775 (Solomon 74
& Björk 2016). Confirming these numbers, Morrison 75
(2018a) finds that the most common APC in her sam-76
ple is US$1,780. In the UK, JISC reports average APCs 77
around 1,700£ (~US$2,240)(Shamash 2017). Covering 78
all DOAJ-listed journals, Crawford finds an average 79
APC paid of US$1,569 (Crawford 2019). Interestingly, 80
this year, the German DEAL consortium agreed to pay 81
2,750€ (~US$3,110) per article in their “publish & read” 82
contract with the publisher Wiley (Haufe 2019). Thus, 83
the prices incurred vary from zero to several thou-84
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
sands of $/£/€, an additional reason why these num-85
bers - while accurate - are not useful for a reliable cal-86
culation of what the scholarly publishing of public re-87
search could or should cost. 88
From the figures available, it is straightforward 89
to hypothesize that publishers, by and large, deter-90
mine their price structure according to what they esti-91
mate the market to be able to carry, i.e., with a value-92
based (or prestige) pricing strategy in a market with 93
status consumption (Goldsmith et al. 2010; Kumcu & 94
McClure 2003). Both the subscription approach and 95
the APC approach share the same basic property, 96
which uncouples the price charged from the costs in-97
curred: non-substitutability. In the subscription sys-98
tem, due to rules such as the Ingelfinger rule (Marshall 99
1998; Angell & Kassirer 1991) that prevent duplicate 100
publications, each article can be found at only one 101
journal of one publisher exclusively. Hence, due to this 102
lack of competition, subscription pricing need not be 103
coupled to publication costs, but purely to reader de-104
mand. Analogously, the more than 34,000 scholarly 105
journals are not only differentiated by the areas of 106
scholarship they serve, they are also stratified in a 107
ranking system where no two journals share the same 108
position, conveying prestige and status to authors. 109
Thus, as duplicate publications are still prevented in 110
OA as in subscription journals, the number of journals 111
in a particular field and prestige stratum effectively 112
equals one. The APC-OA ‘market’ hence suffers from 113
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
analogous non-substitutability problems as the sub-114
scription market, leading to market failure and hyper-115
inflation also there (Crawford 2019; Morrison 2018a; 116
Shamash 2017; Khoo 2019). Corroborating these ob-117
servations are data that also APCs fluctuate with au-118
thor demand rather than with costs and that authors 119
appear to be price-insensitive (Schönfelder 2018; An-120
drew 2012; Khoo 2019). In fact, at least two publishers 121
have publicly stated that their pricing was driven not 122
by costs, but by market and competitor analysis 123
(Poynder 2015; Morrison 2018b). Thus, in both sys-124
tems, monopolistic situations have arisen that let de-125
mand, prestige and purchasing power, rather than 126
cost drive the prices. The non-substitutability in these 127
markets appears to be a major contributing factor 128
leading to value-based pricing. This argument entails 129
that in order to arrive at a truly competitive market 130
where the main driver for price is cost (i.e., promoting 131
a cost-plus pricing strategy), the goods in this market 132
need to be substitutable. As scholarly articles are writ-133
ten and reviewed by the scholars themselves, the 134
goods in this market are publishing services. 135
The editorial, reviewing, processing, production 136
and publication workflows do not differ with regard to 137
the way they are paid, i.e., via subscriptions, APCs or 138
other modes of payment. For example, so-called hy-139
brid journals derive their revenue simultaneously 140
from APCs and from subscription fees. Whereas this 141
business practice, to charge both parties, libraries and 142
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
authors of one and the same journal, has been criti-143
cized as “double-dipping” (Mittermaier 2015), it simul-144
taneously proves that editorial workflows and produc-145
tion service levels must be identical for both business 146
models. Such internal workflows and service levels are 147
usually set by industry standards and the policy of the 148
publisher. Consequently, when calculating the cost of 149
publishing a scholarly article, to arrive at a cost-plus 150
pricing scheme, besides fixed costs, we only have to 151
consider the workflow and associated services, ac-152
cording to current practice. 153
In this article, we list the various steps and pro-154
cedures for a representative publishing workflow ac-155
cording to current industry standards. Each step in-156
curs a cost which can be determined by analyzing the 157
market rates for each service or procedure. These 158
costs comprise the direct costs. We also add several 159
indirect (or fixed) cost items which do not accrue on a 160
per article basis. The final per-article costs are then 161
specified as a range depending on the number of arti-162
cles published and the service level desired. 163
Methodology 164
To arrive at a meaningful figure denoting how 165
much the publication of an article does costs on aver-166
age, it is necessary to arrive at the exact cost for each 167
step in the processing workflow of a manuscript being 168
submitted for publication. These direct or variable 169
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
costs then have to be combined with the indirect or 170
fixed costs of running a publishing enterprise, such as 171
staff costs, real estate and energy costs, etc. The for-172
mer requires granular insight and expertise about the 173
different service levels for the entire publishing work-174
flow. The latter is commonly calculated as staff over-175
head. In this work, we have therefore calculated the 176
cost for each step in the standard publication work-177
flow under consideration of both fixed and variable 178
costs. Both external and internal expenses have been 179
taken into account as well as overhead costs to cover 180
fixed non-direct company costs of the publishing ven-181
ture. 182
Direct or variable costs 183
Expenses and fees for each individual service 184
have been arrived at from two main sources. Some 185
standard services have been taken from openly avail-186
able price lists (Table 1). 187
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Service Provider
Permalink to fee page
Long-term preservation
Peer-review, publishing, type-
Akron Aps
Peer-review management
Table 1: Publishing services and their fees. 190
Second, we requested quotes from vendors 192
without publicly available fees, or turned to other 193
sources (ECAT 2009).. For services such as manuscript 194
submission and peer review management systems we 195
considered vendors such as Manuscript Central (Clari-196
vate) and Editorial Manager (ARIES). 197
Other costs such as internal staff costs (includ-198
ing overhead, EU/US standard) were estimated taking 199
into account not only current market costs we have re-200
quested ourselves, but also numbers from major pub-201
lishing houses (MDPI, Wiley, Springer, DeGruyter, 202
Frontiers, Ubiquity, SciELO, Open LIbrary of the Hu-203
manities). While some of these publishers have made 204
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
their costs public (Table 2), others have either pro-205
vided their numbers under the condition of confiden-206
tiality or the numbers were gained from internal 207
sources. 208
Open Library of the Humanities
Table 2: Published itemized cost structures from publishers/service providers. 210
For certain tasks, for example copyediting or 212
typesetting, there are hundreds of individual compa-213
nies worldwide providing those services on a industry-214
standard level. In our quote requests, we have consid-215
ered only those with which we have collaborated in 216
real business life so far or from which we know the 217
performance and service level in detail from co-oper-218
ations over two decades. Having compared the pricing 219
of those service providers with others, we found only 220
a very small variation of cost for such tasks, which jus-221
tifies our practical approach. It was never our ambition 222
to perform an exhaustive but always incomplete mar-223
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
ket study of service providers worldwide, but an at-224
tempt to provide an authoritative documentation of 225
approximate current publishing costs as a valuable in-226
formation tool for decision-makers and other stake-227
holders in policy drafting, contract negotiations or 228
public discourse. 229
There are three main areas in which production 230
steps have to be considered: content acquisition, con-231
tent preparation (production) and content dissemina-232
tion/archiving. Importantly, ‘content acquisition’ does 233
not imply active acquisition of authors and/or manu-234
scripts. 235
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
1. Content acquisition 237
a. Searching and assigning reviewers 238
b. Communication with reviewers 239
c. Communication with authors 240
d. Handling of re-submission process 241
e. Plagiarism check 242
f. Online submission system 243
g. CrossRef Similarity Check 244
h. CrossRef DOI for article 245
i. CrossRef DOI for 2 or more reviews 246
j. APC collection 247
2. Content preparation (production) 248
a. Manuscript tracking system 249
b. Production system check-in 250
c. Technical checking of manuscript 251
d. Copyediting 252
e. Language editing 253
f. Typesetting 254
g. Formatting figures/graphs/tables 255
h. Altmetric badge 256
i. XML and metadata preparation 257
j. Handling author corrections 258
3. Content dissemination/archiving 259
a. Web OA platform and hosting 260
b. CLOCKSS/Portico 261
c. OAPEN 262
d. Upload to Scopus, PMC, etc. 263
Pricing figures have been deducted by openly 265
available price lists of vendors, as for example for 266
Scholastica, Akron Aps, CrossRef, CLOCKSS (see Tables 267
1, 2). In all other cases where pricing list or fees were 268
not openly available on the web, prices were indicated 269
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
after a direct request for proposal or communicated 270
privately. For the latter we have checked with other 271
partners to validate that information. Some service 272
vendors have not split their services in a granular 273
manner but offer a full service for more steps of the 274
publishing workflow. In those cases we have tried to 275
split those costs or consider the full cost as part of one 276
of the scenarios (see below) which cover the complete 277
manuscript acquisition and article production pro-278
cess. 279
Indirect or fixed costs 280
The calculation of per-article figures from costs 281
that do not accrue on a per-article basis (e.g., salaries, 282
annual fees, etc.) was based on the following assump-283
tions: (i) The average STM article contains 12 printed 284
pages (Johnson et al. 2018). (ii) We estimated an aver-285
age STM article to contain 10 non-text items such as 286
figures or tables. (iii) We also assumed an average re-287
jection rate of 50% after conventional (pre-publica-288
tion) peer-review with at least two reports and ten 289
contact requests to secure one reviewer. (iv) We as-290
sume a desk-rejection rate of 10% after editorial re-291
view. (v) We also base our staff costs on the granular 292
work load per article and not on full-time equivalents 293
(FTE). These assumptions entail that all editorial duties 294
(on average 7.5 person-hours per submitted manu-295
script) are handled by in-house staff and none by aca-296
demic editors, while peer-review is still performed by 297
volunteer academics. In this way, staff costs, including 298
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
overhead expenses, are calculated on a per-article ba-299
sis. Salary costs are based on industry standards in 300
more economically developed countries for the differ-301
ent editorial tasks. Overhead expenses can vary signif-302
icantly depending on the profit and loss structure of 303
the publisher and include rent, repairs, depreciation, 304
interest, insurance, travel expenditures, labor burden, 305
telephone bills, supplies, taxes, accounting fees, etc. 306
We have estimated an average 33% overhead on top 307
of salary costs. The following publication tasks are 308
commonly covered by annual (membership) fees plus 309
an initial, one-time set-up or installment fee: Web OA 310
platform and hosting, CLOCKSS/Portico, OAPEN, Alt-311
metric Badge and Crossref. Because these costs ac-312
crue regardless of how many articles are published 313
(i.e., fixed costs), we have calculated per-article costs 314
for journals with different numbers of articles pub-315
lished per year. 316
While some general fixed costs are covered by 317
salary overheads (see above), we deliberately chose to 318
not include certain fixed costs: Cost of sales have not 319
been considered because for open access journals no 320
longer sales representatives are required which have 321
to negotiate renewals of subscriptions with libraries 322
on an annual basis. We also excluded management 323
costs as these are highly variable and in large publish-324
ers with many journals (and hence articles), per article 325
costs of management are often negligible. We realize 326
that this may be different for publishers which publish 327
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
low-volume journals but with nevertheless highly paid 328
executives (see Discussion). Because making an article 329
public (i.e., ‘publishing’) is distinct from locking it be-330
hind a paywall, we have also not calculated the often 331
very significant paywall costs. While innovation (or ac-332
quisition of innovative technologies) as well as brand-333
ing and advertising/marketing are crucial for a com-334
pany to succeed and thrive in a market in the long 335
term, we have also not included these costs as they 336
are not directly related to publishing scholarly articles. 337
Such costs would include conference attendance, ad-338
vertisement in print, online, social media and search 339
platforms, as well as search engine optimization (SEO). 340
Similarly, government relations (lobbying) may be con-341
sidered a necessary expense for any business, but as 342
it does not directly relate to the process of publishing 343
academic papers, we did not include these costs in our 344
calculations either. However, we do discuss the prob-345
able extent to which these non-publication costs may 346
affect pricing. 347
Scenarios 348
The motivation for the above assumptions was 349
to combine a robust cost estimate (i.e., sourced from 350
measurable time efforts and industry salaries) with an 351
upper bound cost estimate which would come to lie 352
above most academic-run journals. We also calculated 353
a cost estimate for articles handled exclusively by vol-354
unteer academics. Prices for journals where volunteer 355
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
and compensated editors cooperate, will hence fall 356
between these two extremes. 357
With a modern, decentralized/federated plat-358
form providing publishing functionalities without jour-359
nals, some of these steps become obsolete, while oth-360
ers remain relevant. Steps that may become obsolete 361
include DOIs, long-term archiving such as CLOCKSS or 362
Portico, indices such as Scopus. Relevant steps re-363
maining are typesetting/copyediting, XML prepara-364
tion, format conversion, plagiarism checks. 365
Scenario A
Scholastica including ms submission, standard peer-review, track-
ing system, OA webpage, hosting
Scenario A2
Scenario A, but PPPR
Scenario B
Generic service providers, ms submission, standard peer-review
tracking system; OA webpage, hosting
Scenario B2
Scenario B, but PPPR
Scenario C
Generic service providers for content preparation with online plat-
form; without external submission, reviewing, and tracking system;
with DOI; no external hosting/archiving; volunteer editors
Scenario C2
Scenario C, but Scholastica
Table 3: Publishing scenarios for which detailed cost calculations have been performed. 367
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
We have grouped the various combinations of 369
tasks and publication options into six broad scenarios, 370
for which we have calculated all associated publication 371
costs (Table 3). These scenarios correspond either to 372
existing publishing options or to options that have 373
been discussed in the literature. For each of the six 374
scenarios, we have also calculated the same costs, but 375
assuming a 90% rejection rate (see raw data file). 376
All the data we have based our calculations on 378
are available at Figshare (DOI: 379
10.6084/m9.figshare.8118197). 380
Results 381
One of the first findings of our calculations is 382
that in order to employ at least one 50% FTE of an in-383
house editor, a journal has to publish approx. 100 ar-384
ticles per year or more. Hence, in the following, we will 385
base our estimates on journals publishing at least 100 386
articles per year (corresponding to 50% FTE) or 1,000 387
articles (corresponding to 5 FTEs), to show the spread 388
of fixed and indirect costs over the number of articles 389
published. 390
Our estimate of per-article publishing costs in a 391
conventional pre-publication peer-review (50% rejec-392
tion rate) scenario where all editorial duties are per-393
formed by in-house staff (Scenario B) ranges from 394
US$643.61 for a journal that publishes 100 articles per 395
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
year down to US$565.15 for such a journal that pub-396
lishes 1,000 articles (or more, as the indirect costs be-397
come increasingly negligible around this value). These 398
values consist of US$266.53 direct publishing costs 399
(i.e., CrossRef Similarity Check, CrossRef DOI for an ar-400
ticle, CrossRef DOI for two or more reviews, copyedit-401
ing, typesetting, formatting figures/graphs/tables, alt-402
metric badge, upload to Scopus and XML and 403
metadata preparation), US$ 289.91 for editorial staff 404
and US$8.72 to US$87.18 for 1,000 to 100 articles, re-405
spectively, in indirect costs (i.e., Web OA platform and 406
hosting, CLOCKSS, OAPEN, Altmetric Badge and Cross-407
ref). 408
These numbers were calculated using generic, 409
full-service providers based in India, where applicable. 410
There are open access service providers that provide 411
packaged deals for the same services as these generic 412
service providers. We have calculated the same steps 413
using a well-known provider in this area, Scholastica 414
(Scenario A). Interestingly, these figures are slightly 415
higher: US$ 374.08 for direct publishing costs and 416
US$5.92 to US$59.18 for 1,000 to 100 articles, respec-417
tively, for indirect costs (editorial staff costs remain the 418
same). 419
While these costs have been calculated for a ge-420
neric journal with 50% rejection rate, per-article costs 421
will increase with increased rejection rates and de-422
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
crease with less rejections as in, e.g., a post-publica-423
tion peer-review (PPPR) model. In a journal that uses 424
generic service providers and publishes all submitted 425
manuscripts as PDF preprints with a DOI before per-426
forming otherwise identical peer-review as described 427
above (i.e., PPPR with in-house editors and volunteer 428
reviewers), per article editorial services drop from 429
US$289.91 to US$140.69 (Scenario A2/B2), with all 430
other costs remaining nearly identical. Conversely, 431
prestigious journals with rejection rates of around 432
90% see their costs rise to US$1053.87 for 100 articles 433
per year or US$770.53 for the larger journals with 434
about 1,000 articles per year (generic service provid-435
ers). 436
These numbers also show that for a conven-437
tional journal today, where academics perform their 438
editorial duties on a volunteer basis (i.e., Scenario B, 439
but no editorial costs as editor salaries are paid for by 440
their academic institutions), direct publication costs 441
come to lie at US$266.53 with generic service provid-442
ers and total costs depend on the scale at which the 443
journal operates. Small journals with 100 articles 444
would face average per article total publication costs 445
of US$353.71, while journals with 1,000 or more arti-446
cles would only face costs of US$275.25 or less per 447
published article. Even at the highest convenience for 448
a small, volunteer-run journal, costs come to lie at 449
US$454.63 where a full-service provider (Scholastica) 450
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
handles all of the technical aspects of the work (Sce-451
nario C2). 452
The above calculations (summarized in Table 4) 453
demonstrate economies of scale. The more articles 454
are being published, the lower the costs for each arti-455
cle, approaching the fixed costs for each article. 456
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Conventional peer review, Scholastica, 100 articles (A)
Conventional peer review, Scholastica, 1,000 articles (A)
Conventional peer review, generic providers, 100 articles
PPPR, Scholastica, 100 articles (A2)
Conventional peer review, generic providers, 1,000 arti-
cles (B)
PPPR, Scholastica, 1,000 articles (A2)
PPPR, generic providers, 100 articles (B2)
Volunteer editors, Scholastica, 100 articles (C2)
Volunteer editors, Scholastica, 1,000 articles (C2)
PPPR, generic providers, 1,000 articles (B2)
Volunteer editors, generic providers, 100 articles (C)
Volunteer editors, generic providers, 1,000 articles (C)
Table 4: Different scenarios of journal organization, ordered by total per article costs (in 458
US$). The scenarios are labeled with A, A2, B, B2, C, C2 (see table 3). 459
Because of the economies of scale and recent 461
calls for the replacement of journals with a modern 462
publishing platform (Brembs 2019; Stern & O’Shea 463
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
2019; Grossmann 2015; Nosek & Bar-Anan 2012; Hart-464
gerink 2019), we have also calculated the cost of pub-465
lishing the annual output of the STM community, ap-466
prox. 3 million articles, on such a platform that facili-467
tates PPPR organized by academic editors on a single, 468
decentralized, federated platform running modern 469
software solutions. Such a platform would dispose of 470
several production steps which are necessitated by 471
the current balkanization of the literature in different 472
journals published by different publishers, but keep 473
others (see Methodology). In this scenario, the indirect 474
and fixed costs per article approach zero due to the 475
high number of published articles (but see Discus-476
sion), such that the only remaining costs would be the 477
direct publishing costs of US$190.17 per published ar-478
ticle. 479
Finally, taking a ballpark cost figure of US$600 480
for a scholarly article with full editorial services (i.e., 481
scenario A/B) and comparing it to the low end of the 482
average price estimate for a subscription article of 483
about US$4,000, it becomes clear that publication 484
costs only cover 15% of the subscription price (Fig. 1). 485
Assuming a conservative profit margin of 30% (i.e., 486
US$1,200 per article) for one of the large publishers 487
(McGuigan & Russel 2008; Larivière et al. 2015; 488
Beverungen et al. 2012; Harvie et al. 2012), there re-489
mains a sizeable gap of about US$2,200 in non-publi-490
cation costs, or 55% of the price of a scholarly sub-491
scription article (Fig. 1). 492
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Fig. 1: Subscription price and cost items. Assuming the commonly accepted US$4,000 price 495
tag for a subscription article, published profit margins of 30% and our calculation of 496
US$600 in publication costs for a full-service subscription article (scenario A/B, see Table 497
4), there remain US$2,200 in non-publication costs per article. 498
Discussion 500
Since the 1990s, it has been recognized that the 501
prices of scholarly journals were escalating at unsus-502
tainable rates (Douglas 1990). In the last 30 years, this 503
“serials crisis” has never been coherently addressed, 504
let alone solved. With this work, we aim to provide 505
more financial evidence for future evidence-based 506
policies addressing the affordability problem of schol-507
arly communication (Chan 2004; Harnad et al. 2004). 508
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Subscription prices and publication costs 509
Not only current discussions are addressing the 510
affordability problem in the unit of cost per article (Van 511
Noorden 2013; Schimmer et al. 2015; Odlyzko 2013; 512
Johnson et al. 2018; Odlyzko 1995; Jahn & Tullney 513
2016; Solomon & Björk 2016; Morrison 2018a) and we 514
follow this precedent. Drawing from publicly available 515
price lists and industry-standard service costs, we find 516
that publishing costs per article vary from US$194.89 517
to US$723.16, depending on the level of service and 518
publishing volume (Table 4). It is important to note 519
that these are conservative estimates, likely to consti-520
tute upper bounds, where innovation and changes in 521
practice can be expected to decrease costs. 522
Perhaps not surprisingly, the convenience of 523
outsourcing the main publishing services to a special-524
ized full-service provider comes with a small increase 525
in cost (scenario A vs. scenario B), when compared to 526
an itemized sourcing of publishing services. In our cost 527
estimate, we have not factored in the management 528
cost of sourcing the itemized services, as we have not 529
included company management in our calculations. 530
Any decision between these two options will thus have 531
to be made after factoring in such costs as well. 532
Even in the rare, most expensive case, these 533
costs compare very favorably to the current subscrip-534
tion pricing of around US$4,000-5,000. Our highest 535
value encompasses conventional, journal-based pre-536
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
publication peer-review with a generic 50% rejection 537
rate at a small journal (~100 articles per year) where 538
all management of peer-review is performed by in-539
house editorial staff with no volunteer academic edi-540
tors. Our data suggest that increasing only the rejec-541
tion rate, for example from 50% to 90%, leads to an 542
increase in publication costs of around 30-40% (e.g., in 543
scenario B from US$565.15 to US$770.53 for 1,000 ar-544
ticle journals or from US$643.61 to US$1,053.87 for 545
100 article journals). Apparently, this is a consequence 546
of the respective increase of direct personnel ex-547
penses for managing the peer review process and 548
communicating with both reviewers and authors for 549
classical pre-publication peer review. As currently 550
most highly selective journals publish on the order of 551
800-900 research articles per year about US$1,000 per 552
article can be seen as an upper bound of total publica-553
tion costs at such journals. 554
Article processing charges and publication costs 555
The reported average APCs charged by the mi-556
nority of journals with such fees vary between 557
US$1,400-2,200 depending on the sample (see above 558
and, e.g., Table 2). The large difference between these 559
values and even our most expensive cost estimate is 560
at least partly consistent with our hypothesis that the 561
quasi-monopolistic situation of the publishers, due to 562
the non-substitutability of their goods and services, al-563
lows them to adopt a value-based pricing strategy also 564
in the APC-OA case, similar to subscription pricing. It is 565
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
therefore straightforward to hypothesize that any pol-566
icy that fails to address the non-substitutability prob-567
lem in scholarly communication will also fail to solve 568
the affordability problem and lead to a similar market-569
failure as in the subscription model. An analogous ar-570
gument has previously also been endorsed by the Eu-571
ropean Commission Directorate-General for competi-572
tion (Tennant & Brembs 2018), even before our calcu-573
lations were available. Further reducing the odds of 574
APC-OA solving the affordability problem is the fact 575
that authors are not only price-insensitive (Khoo 576
2019), but seem to prefer publishing in journals that 577
charge APCs as opposed to those that do not, as 578
evinced by the fact that most OA articles are published 579
in the minority of journals that charge APCs (Crawford 580
2019). Above and beyond authors’ preference for jour-581
nals with APCs over those without, among those APC 582
journals, authors are incentivized to publish in high-583
APC, rather than low-APC journals, because APCs in-584
crease with the prestige of the journal (Tennant & Lo-585
max 2019; Andrew 2012). Consequently, a recent 586
study observed APC increases of 2.5-6 times inflation 587
over six years in their sample (Khoo 2019). This con-588
verging evidence all points towards both APC-OA and 589
subscriptions to suffer from analogous flaws which 590
lead to hyperinflation and market-failure in both 591
cases. Our data now add further evidence in support 592
of this hypothesis. 593
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Aiming for a cost-plus market 594
Starting from current subscription pricing of 595
around US$4,000-5,000 per article (Van Noorden 596
2013; Schimmer et al. 2015; Odlyzko 2013; Johnson et 597
al. 2018; Odlyzko 1995), we confirm previous esti-598
mates that current subscription moneys are sufficient 599
to pay for a complete transition to OA, even at current 600
inflated APCs on the order of about US$2,000 per arti-601
cle (Schimmer et al. 2015; Odlyzko 2013; Johnson et al. 602
2018; Jahn & Tullney 2016; Solomon & Björk 2016; 603
Morrison 2018a). Calculated globally, this hypothetical 604
transition to APC-OA would cut the US$10 billion 605
world-wide annual subscription budget roughly in 606
half, at least in the short term. At the same time, if 607
there were a way to enforce cost-plus pricing strate-608
gies in publishers, even the current prices would at 609
least be 100% above actual publishing costs at the 610
highest level of service and even more for a lower level 611
of service and higher article volume, which is the norm 612
at many journals. Thus, the mere transition to a mar-613
ket where the current value-based pricing strategies 614
are not deployed any more, all else being equal, 615
stands to save the global taxpayer at least 75% of the 616
current subscription budget, or the equivalent of 617
about US$7.5 billion annually. However, the current 618
journal system does not provide for such a solution as 619
journals are non-substitutable (see above). 620
Replacing journals with a modern, server-621
based, decentralized solution (Brembs 2019; Stern & 622
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
O’Shea 2019; Grossmann 2015; Hartgerink 2019) im-623
plements substitutability of services and, hence, com-624
petition, providing for the largest savings: even when 625
the volume of articles amounts to 3 million per year 626
(Johnson et al. 2018), the global taxpayer stands to 627
save about 95% of the current subscription budget, or 628
the equivalent of approx. US$9.5 billion annually, on 629
publishing prices. 630
Cost-plus pricing technically feasible today 631
There are more conclusions to be drawn from 632
the evidence we provide here. For one, while the cur-633
rent APC-OA prices would, if applied universally, ad-634
dress the affordability problem and substantially 635
lower the cost to the taxpayer in the short term, the 636
available evidence suggests that the current value-637
based pricing strategy of publishers (together with the 638
price-insensitivity of authors (Khoo 2019)) is likely to 639
quickly eat into these gains and again lead to unsus-640
tainable inflation, as in the subscription case. 641
Second, because the workflow we model con-642
sists of verifiable, modular components, we demon-643
strate that a cost-plus pricing scheme is possible to-644
day. Phrased differently, customers of commercial 645
publishers can use these numbers as tools in contract 646
negotiations to demand more cost-oriented contracts. 647
However, at the same time, as long as the ultimate lev-648
erage in such negotiations, namely to walk away and 649
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
opt for the goods and services of a competitor, re-650
mains inaccessible due to the non-substitutability 651
problem, the effectiveness of this tool will remain 652
comparatively limited. 653
Third, our calculations show that with publish-654
ing volumes exceeding 1,000 articles per year, fixed 655
costs shrink below 1% of the direct article costs and 656
become negligible. This was expected and already 657
concluded in a previous analysis (Bogich & Ballesteros 658
2016). These insights are important for designing a 659
transition towards a scholarly publishing platform in-660
stead of journals. 661
Fourth, due to the limited possibility in dividing 662
labor contracts into arbitrarily small portions, we find 663
that journals with volumes below approx. 100 articles 664
per year would be best served financially if they oper-665
ated on the concept of volunteer academic editors 666
handling the peer-review, instead of in-house staff. 667
Targeting the non-substitutability problem 668
Synthesizing all of these conclusions, it be-669
comes clear that any solution to the affordability prob-670
lem must aim at eliminating non-substitutability and 671
strive towards large volume strategies. Historically, 672
non-substitutability has been solved with, e.g., indus-673
try standards that allow substitution of products and 674
services. For instance, multimedia standards allow for 675
media from any producer to be played on any player. 676
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
In the case of scholarly communication, the non-sub-677
stitutability is granted via prevention of duplicate pub-678
lications of discoveries in different journals together 679
with prestige stratification of the journals. Both of 680
these factors are conveyed by the journals where the 681
individual articles are published. Consequently, one 682
straightforward approach to mitigate this non-substi-683
tutability is to eliminate journals as venues and imple-684
ment technical standards to allow publication services 685
to become substitutable. 686
One technical implementation of this principle 687
is to collect all articles in a single, decentralized, feder-688
ated venue that is governed by the scholarly commu-689
nity and designed using common, evolvable standards 690
to allow for the substitution (and, consequently, com-691
petition) of service providers (Brembs 2019; Stern & 692
O’Shea 2019; Grossmann 2015; Hartgerink 2019). This 693
concept mimics other infrastructure arrangements 694
such as water, electricity, HVAC, email, etc. This ap-695
proach would, at the same time, solve the problem of 696
large publication volume: the STM field is on course to 697
publish about 3 million articles every year (Johnson et 698
al. 2018), allowing fixed costs to effectively converging 699
towards zero in the per-article currency (Bogich & Bal-700
lesteros 2016). However, even if the per-article costs 701
of such infrastructure are negligible, they remain a 702
substantial item in absolute terms that scholarly insti-703
tutions need to pay. In a recent tender, the European 704
Commission provided an indicative estimate for the 705
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
cost of “development of the platform, its services and 706
business processes, communication and sustainabil-707
ity” (European Commission 2017), of around 250,000€ 708
per year. Perhaps an order of magnitude higher costs 709
may be estimated to implement and run a system that 710
is scaled for the world-wide scholarly output, arriving 711
at approx. US$3 million per year. Given that there are 712
about 10,000 universities world-wide (Förster 2019) 713
(plus a large number of non-university research insti-714
tutions) which would stand to participate, these costs 715
to establish and maintain such an infrastructure 716
would likely amount to approx. US$300 per institution 717
per year. Even if only the 3,300 European Union uni-718
versities (European Commission 2003) were to imple-719
ment and run the platform by themselves with other 720
institutions only contributing article costs, these indi-721
rect costs would amount to less than US$1,000 per 722
year and institution. These numbers demonstrate that 723
even under conservative estimates, the fixed costs of 724
a publishing platform remain within feasible bounds. 725
While these numbers demonstrate not only the imme-726
diate feasibility of the transition towards such a plat-727
form, but, indeed, the fiscal imperative for it, it is far 728
from clear how the transition should be accomplished 729
practically. Because it is beyond the scope of this arti-730
cle to provide such policy recommendations, we refer 731
to those already provided elsewhere (see, e.g., 732
Brembs 2019; Stern & O’Shea 2019). 733
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Such a solution would preserve the rules aimed 734
at preventing duplicate publication, but eliminate jour-735
nal hierarchy as a signal for prestige. Given that, at 736
least in the experimental sciences, journal prestige is 737
associated with lower reliability (Brembs et al. 2013; 738
Brembs 2018), it may be argued that eliminating jour-739
nal prestige ought to be a goal in and of itself, in order 740
to tackle any decline in reproducibility (e.g., Karp 2018; 741
Baker 2016; Schooler 2014; Berg 2018; Sayre & Riegel-742
man 2018; Saltelli & Funtowicz 2017; Lilienfeld 2017; 743
Everett & Earp 2015; Brembs 2019). 744
Non-publication costs 745
If the lowest publication costs for journals with 746
volunteer editors constituted merely 5-10% of current 747
subscription prices and publicly reported publisher 748
profits only amount to an additional 30-40%, which 749
non-publication costs are publishers currently facing 750
and taxpayers paying for? While these costs are 751
opaque and variable between publishers and, indeed, 752
between journals, some estimates can be made from 753
publicly available data. If one assumes revenue of 754
about US$4,000 per subscription article (i.e., on the 755
low end of the converging estimates), a conservative 756
30% profit margin (i.e., US$1,200 per article) for one of 757
the large publishers (McGuigan & Russel 2008; Lari-758
vière et al. 2015; Beverungen et al. 2012; Harvie et al. 759
2012) and generous publication costs of US$600 per 760
article (scenario A/B; table 4), then there remains a 761
sizeable gap of about US$2,200 in non-publication 762
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
costs per article - more than the sum of publication 763
costs and profits combined, or 55% of the subscription 764
cost of a scholarly article (Fig. 1). While some of these 765
costs may be considered necessary for any business, 766
none of them are associated with publishing primary 767
research articles (see Methods). 768
Running a business: Management 769
While our cost calculations include generic run-770
ning costs such as rent, repairs, depreciation, interest, 771
insurance, travel expenditures, labor burden, tele-772
phone bills, supplies, taxes, accounting fees, etc., we 773
have explicitly omitted some indirect costs such as 774
management cost and paywalls. For instance, accord-775
ing to their 2016 tax statement, the New England Jour-776
nal of Medicine spends 4% of its publication revenue 777
on their top ten management staff alone (which would 778
translate to about US$160 per article if applied to our 779
example above; Fig. 1). 780
Preventing access: Paywalls 781
Subscription journals also face costs associated 782
with paywalls. It’s difficult to estimate the cost of such 783
technology for publishers, but the cost of a new pay-784
wall for the New York Times was reported to lie be-785
tween US$25-50 million (Pulley 2011; Kramer 2011). 786
Alternatively, as the functional distinction between 787
subscription articles and OA articles is precisely the 788
missing paywall in OA articles, one could also assume 789
that publishers arrive at their current APC pricing of 790
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
around US$2,000 by subtracting paywall costs from 791
their subscription price. This assumption would entail 792
paywall costs of approx. US$2,000 per article (i.e., the 793
difference between APC and subscription pricing). 794
On top of the technical cost of a paywall, one 795
may also consider the legal fees for defending pay-796
walls for this cost item. Publishers have a track record 797
of litigation with regard to articles outside of their pay-798
walls and regularly seek damages in court for actual or 799
perceived threats to their subscription business 800
model (Hansen 2019; Chawla 2017; Van Noorden 801
2017; Association Of American Publishers 2015; Cox 802
2018; Flaherty 2013; Schiermeier 2017). These costs 803
accrue by seeking to enclose the scholarly literature 804
within the paywalls of publisher via alternative routes 805
in addition to the digital paywalls. 806
News, advertising, sales, marketing, public relations: branding 807
Another cost item is publishing non-research 808
content. For instance, for 2016, PubMed lists a total of 809
1,632 articles published by the New England Journal of 810
Medicine, while Clarivate Analytics only counts 328 ar-811
ticles for their Impact Factor. Assuming that only the 812
latter articles amount to primary research publica-813
tions, this journal’s revenue also pays for 1,304 non-814
research articles. Similar numbers also hold for other 815
prestigious journals (e.g.: Nature: 880/2765, Science: 816
805/1938; research/total), often with their own jour-817
nalist and editorial staff commissioning articles and/or 818
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
reporting themselves on research and policy news. 819
However, the number of journals where this can con-820
stitute a significant fraction of their total costs is pre-821
sumably small, likely restricted to the most prestigious 822
journals. 823
Prestigious journals also often practice active 824
author or materials acquisition, by traveling to confer-825
ences and laboratories, building networks in a strat-826
egy to entice the next exciting research finding to be 827
published in their journals. Active author acquisition 828
accrues costs both in terms of travel and time spent 829
networking and communicating with authors that is 830
not covered in our cost estimates (see Methods). 831
Sometimes, new journals also need to engage 832
in such author acquisition practices, which, perhaps, 833
can be best subsumed under general marketing or 834
public relations costs required for building and main-835
taining a brand. These marketing costs also include, 836
e.g., advertising in various venues targeting both au-837
thors and subscribers. For many publishers it is also 838
common to promote their brand at conferences and 839
institutions with, e.g., hosted speakers, travel grants or 840
sponsored awards. 841
Because of the complex, time-consuming nego-842
tiations with libraries on ever tighter budgets due to 843
the hyperinflationary subscription price increases, 844
publishers also need to employ expert sales teams. 845
The task of these sales teams is not only to find the 846
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
most irresistible way to package and bundle subscrip-847
tion journals and/or databases, but also to device the 848
most inexorable psychological strategy for their nego-849
tiations with librarians. These sales teams need to op-850
erate in close connections with the various advertis-851
ing, marketing and public relations teams of the pub-852
lisher to accomplish a coherent brand image. One may 853
argue that in times of OA, these sales costs are not 854
necessary expenses any more and more associated 855
with paywall costs than with publication costs. On the 856
other hand, in an OA world, one may argue that brand-857
ing was never more important for author acquisition. 858
New technologies: innovation and acquisitions 859
Publishers also need to invest in innovation, in 860
order to stay current with their technologies and func-861
tionalities. While scholarly publishers have been quick 862
to transition from print to web-based technologies in 863
the past, the digital functionalities of most of the schol-864
arly literature today lag at least a decade behind cur-865
rent functionalities of other digital objects outside of 866
the scholarly literature. The level of investment in in-867
novation thus remains unclear and its effects ques-868
tionable. Instead of investments into their own tech-869
nological innovation, publishers today appear to ac-870
quire companies that have invented desired function-871
alities around the scholarly workflow, with the goal to 872
provide services beyond publications (Bosman & Kra-873
mer 2018; Crunchbase 2019; Posada & Chen 2018; 874
Campfens 2019). 875
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Government relations: Lobbying 876
Most international publishers, as any other cor-877
poration, also spend significant amounts of money on 878
government relations (i.e., lobbying). Some of these 879
corporations employ staff at the vice president level 880
not only in the most important research nations, but 881
also at the level of supra-national bodies such as the 882
European Commission (Jonathan Tennant 2018). 883
These staff, in turn, employ assistants and other mem-884
bers of their teams. Obviously, the task of these em-885
ployees is to protect current revenue streams, e.g., 886
subscription or APC income. For instance, one pub-887
lisher, Elsevier, spends more than 400,000€ per year 888
on lobbying at the level of the European Commission 889
alone (Anon 2018). The consequences of such efforts 890
have been observable, e.g., in the so-called Finch Re-891
port” in the UK (Finch 2012), which surprised many 892
commentators with its publisher-friendly recommen-893
dations (see, e.g., Prior 2013; Jonathan Tennant 2018). 894
Lack of competition: Inefficiencies 895
Finally, with profit margins exceeding 30% in 896
many cases, there may be less pressure to optimize 897
the workflow to cut down further on already marginal 898
publication costs (on the order of 15% of total costs in 899
the example above, Fig. 1). It is thus conceivable that 900
large publishers, where the economies of scale al-901
ready have decreased costs, are operating at such low 902
efficiencies that their publication costs may come to 903
lie higher than we calculated. 904
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Which non-publication costs should remain bundled up with pub-905
lishing? 906
Regardless of all of these estimates necessarily 907
remaining vague and imprecise, the fact remains that 908
the scholarly community must eventually make a 909
number of decisions, if it is to tackle the affordability 910
problem. Which of the above non-publication costs 911
such as lobbying, start-up acquisition, executive sala-912
ries in the millions of US$, non-research article pub-913
lishing, marketing/advertising, sales/negotiations, in-914
efficiencies etc., should remain bundled up with the 915
process of publishing scholarly research articles? 916
Which of these costs are avoidable, which necessary 917
and which even desirable? Are profit margins of 30-918
40% on taxpayer funds tolerable? 919
In fact, one may even ask whether many of the 920
services we list as part of the scholarly publishing 921
standard are actually necessary for scholarly publish-922
ing. After all, journals such as the Journal of Machine 923
Learning Research, Discrete Analysis or the Journal of 924
Open Source Software publish their articles with inter-925
nal costs below US$10 (Jon Tennant 2018). Likewise, 926
the preprint archive arXiv publishes their articles at 927
similar costs (Cornell University Library 2010). 928
Acknowledgments 929
We are indebted to Michael Dowling, Jon Tennant, Isabell Welpe and Bernhard MItter-930
meier for critically reading earlier versions of this document. We are also grateful to Abel 931
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Packer from SciELO and Brian Cody from Scholastica for privately sharing cost data from 932
their organizations with us. 933
References: 934
Andrew, T., 2012. Gold Open Access: Counting the Costs. Ariadne, (70). Available at: http://www.ari-935 [Accessed April 2, 2019]. 936
Angell, M. & Kassirer, J.P., 1991. The Ingelfinger Rule revisited. The New England journal of medicine, 325(19), 937
pp.13711373. 938
Anon, 2018. LobbyFacts: RELX. Lobbyfacts. Available at:
tive/714c6d1fe1764c059631d861816468a8/relx-group [Accessed June 5, 2019]. 940
Association Of American Publishers, 2015. Statement on Sci-Hub Litigation. Statement on Sci-Hub Litigation. 941
Available at: [Accessed May 9, 942
2019]. 943
Baker, M., 2016. 1,500 scientists lift the lid on reproducibility. Nature, 533(7604), pp.452454. 944
Berg, J., 2018. Progress on reproducibility. Science, 359(6371), p.9. 945
Beverungen, A., Böhm, S. & Land, C., 2012. The poverty of journal publishing. Organization, 19(6), pp.929946
938. 947
Bogich, T.L. & Ballesteros, S.P., 2016. On the Marginal Cost of Scholarly Communication. Research/ 948
a journal of science and its applications. Available at: 949
Bosman, J. & Kramer, B., 2018. Workflows. Innovations in Scholarly Communication. Available at: 950 [Accessed May 7, 2019]. 951
Brembs, B., 2018. Prestigious Science Journals Struggle to Reach Even Average Reliability. Frontiers in hu-952
man neuroscience, 12, p.37. 953
Brembs, B., 2019. Reliable novelty: New should not trump true. PLoS biology, 17(2), p.e3000117. 954
Brembs, B., Button, K. & Munafò, M., 2013. Deep impact: unintended consequences of journal 955
rank. Frontiers in human neuroscience, 7, p.291. 956
Campfens, Y., 2019. Market research report: What has become of new entrants in research work-957
flows and scholarly communication? Open Science Framework. 958
Chan, L., 2004. Supporting and Enhancing Scholarship in the Digital Age: The Role of Open Access 959
Institutional Repository. Canadian Journal of Communication, 29(3). Available at: 960 961
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Chawla, D., 2017. Publishers take ResearchGate to court, alleging massive copyright infringement. Science. 962
Available at:
ing-massive-copyright-infringement. 964
Cornell University Library, 2010. arXiv Business Model White Paper | arXiv e-print repository. arXiv. Availa-965
ble at: [Accessed June 7, 2019]. 966
Cox, K., 2018. Eleventh Circuit Reverses and Remands Georgia State E-Reserves Case (Again) | 967
ARL Policy Notes. Available at: [Accessed May 9, 2019]. 968
Crawford, W., 2019. Gold Open Access 2013-2018: Articles in Journals (GOA4), Lulu. 969
Crunchbase, 2019. Website. Available at:
tions/acquisitions_list [Accessed May 7, 2019]. 971
Douglas, K., 1990. The Serials Crisis: Adjusting to Change. The Serials librarian, 18(1-2), pp.111121. 972
ECAT, 2009. Managing Peer Review Online. Available at:
review-online [Accessed April 14, 2019]. 974
European Commission, 2017. Open Research Europe - The European Commission Open Research Publish-975
ing Platform. Available at: [Ac-976
cessed April 14, 2019]. 977
European Commission, 2003. The role of universities in the Europe of knowledge. EUR-Lex. Available at: 978 [Accessed May 2, 2019]. 979
Everett, J.A.C. & Earp, B.D., 2015. A tragedy of the (academic) commons: interpreting the replica-980
tion crisis in psychology as a social dilemma for early-career researchers. Frontiers in psy-981
chology, 6, p.1152. 982
Finch, J., 2012. Accessibility, sustainability, excellence: How to expand access to research publica-983
tions: Report of the Working Group on Expanding Access to Published Research Findings 984
[“The Finch Report”]. Available at:
work/finch-report-final [Accessed May 7, 2019]. 986
Fisher, J.H., 2008. Scholarly Publishing Re-invented: Real Costs and Real Freedoms. The journal of electronic 987
publishing: JEP, 11(2). Available at: [Accessed April 6, 988
2019]. 989
Flaherty, C., 2013. An academic press sues a librarian, raising issues of academic freedom. Available at: 990
academic-freedom [Accessed May 9, 2019]. 992
Förster, K., 2019. Universities Worldwide: Homepage. Available at: [Accessed April 10, 2019]. 993
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Goldsmith, R.E., Flynn, L.R. & Kim, D., 2010. Status Consumption and Price Sensitivity. Journal of 994
Marketing Theory and Practice, 18(4), pp.323338. 995
Grossmann, A., 2015. Publishing in transition do we still need scientific journals? ScienceOpen Research. 996
Available at:
e00d84a72cd1/ScienceOpen/3077_XE696973259861784096.pdf [Accessed April 2, 2019]. 998
Hansen, D., 2019. Giving the Authors a Voice in Litigation? An ACS v. ResearchGate Update - 999
Scholarly Communications @ Duke. Scholarly Communications @ Duke. Available at: 1000
gation-an-acs-v-researchgate-update/ [Accessed May 9, 2019]. 1002
Harnad, S. et al., 2004. The green and the gold roads to Open Access. Nature. Available at: [Accessed April 3, 2019]. 1004
Hartgerink, C., 2019. Verified, Shared, Modular, and Provenance Based Research Communication 1005
with the Dat Protocol. Publications, 7(2), p.40. 1006
Harvie, D. et al., 2012. What are we to do with feral publishers? Organization, 19(6), pp.905914. 1007
Haufe, G., 2019. FAQ Wiley Contract. Projekt DEAL. Available at:
tract [Accessed June 7, 2019]. 1009
Houghton, J.W., 2001. Crisis and transition: the economics of scholarly communication. Learned 1010
publishing: journal of the Association of Learned and Professional Society Publishers, 14(3), 1011
pp.167176. 1012
Jahn, N. & Tullney, M., 2016. Data and code used from: A study of institutional spending on open 1013
access publication fees in Germany. Available at:
load/2905588/2905590/paper_openapc-final.tar.gz [Accessed April 2, 2019]. 1015
Johnson, R., Watkinson, A. & Mabe, M., 2018. 2018 STM Report, STM Assic. Available at: https://www.stm-1016 [Accessed December 21, 2018]. 1017
Karp, N.A., 2018. Reproducible preclinical research-Is embracing variability the answer? PLoS biology, 16(3), 1018
p.e2005413. 1019
Khoo, S.Y.-S., 2019. Article Processing Charge Hyperinflation and Price Insensitivity: An Open Ac-1020
cess Sequel to the Serials Crisis. LIBER Quarterly, 29(1), p.1. 1021
Kramer, S.D., 2011. New York Times Paywall Cost: More Like $25 Million. Available at: https://gi-1022 [Accessed June 7, 1023
2019]. 1024
Kumcu, E. & McClure, J.E., 2003. Explaining Prestige Pricing: An Alternative to Back-Bending De-1025
mand. Marketing Education Review, 13(1), pp.4957. 1026
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Larivière, V., Haustein, S. & Mongeon, P., 2015. The Oligopoly of Academic Publishers in the Digital Era. PloS 1027
one, 10(6), p.e0127502. 1028
Lilienfeld, S.O., 2017. Psychology’s Replication Crisis and the Grant Culture: Righting the Ship. Per-1029
spectives on psychological science: a journal of the Association for Psychological Science, 1030
12(4), pp.660664. 1031
Marshall, E., 1998. EMBARGOES: Franz Ingelfinger’s Legacy Shaped Biology Publishing. Science, 282(5390), 1032
pp.861861. Available at: 1033
McGuigan, G.S. & Russel, R.D., 2008. The business of academic publishing: A strategic analysis of 1034
the academic journal publishing industry and its impact on the future of scholarly pub-1035
lishing. Electronic Journal of Academic and Special Librarianship, 9(3). Available at: 1036
ing-a-strategic-analysis-of-the-a [Accessed May 14, 2019]. 1038
Mittermaier, B., 2015. Double Dipping in Hybrid Open Access Chimera or Reality? ScienceOpen Research. 1039
Available at:
682b1e6fe3b8/ScienceOpen/2729_XE6708374319518627865.pdf [Accessed April 2, 2019]. 1041
Morrison, H.G., 2018a. Global OA APCs (APC) 2010-2017: Major Trends. In 22nd International Con-1042
ference on Electronic Publishing. 22nd International Conference on Electronic Publishing. 1043
OpenEdition Press. Available at: 1044
Morrison, H.G., 2018b. MDPI pricing (thanks to MDPI CEO Franck Vazquez, PhD). Sustaining the 1045
Knowledge Commons / Soutenir les savoirs communs. Available at: https://sustaining-1046
phd/ [Accessed May 10, 2019]. 1048
Nosek, B.A. & Bar-Anan, Y., 2012. Scientific Utopia: I. Opening Scientific Communication. Psychological in-1049
quiry, 23(3), pp.217243. 1050
Odlyzko, A., 2013. Open Access, library and publisher competition, and the evolution of general 1051
commerce. Available at: [Accessed December 14, 2018]. 1052
Odlyzko, A.M., 1995. Tragic loss or good riddance? The impending demise of traditional scholarly 1053
journals. International journal of human-computer studies, 42(1), pp.71122. 1054
Posada, A. & Chen, G., 2018. Inequality in Knowledge Production: The Integration of Academic In-1055
frastructure by Big Publishers. In 22nd International Conference on Electronic Publishing. 1056
22nd International Conference on Electronic Publishing. OpenEdition Press. Available at: 1057 1058
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Poynder, R., 2015. Emerald Group Publishing tests ZEN, increases prices: what does it mean? Available at: 1059 [Accessed April 1060
2, 2019]. 1061
Prior, A., 2013. Key Issue - The “Finch Report”: the future is gold, but many challenges lie ahead. 1062
Insights: the UKSG journal, 26(1), pp.7781. 1063
Pulley, B., 2011. New York Times Fixes Paywall to Balance Free and Paid. Bloomberg. Available at: 1064
balance-free-vs-paid-on-the-web [Accessed June 7, 2019]. 1066
Rose-Wiles, L.M., 2011. The High Cost of Science Journals: A Case Study and Discussion. Journal of 1067
Electronic Resources Librarianship, 23(3), pp.219241. Available at: 1068 1069
Saltelli, A. & Funtowicz, S., 2017. What is science’s crisis really about? Futures, 91, pp.511. 1070
Sayre, F. & Riegelman, A., 2018. The Reproducibility Crisis and Academic Libraries. College & Research Librar-1071
ies, 79(1). Available at: 1072
Schiermeier, Q., 2017. US court grants Elsevier millions in damages from Sci-Hub. Nature, 6, p.541. 1073
Schimmer, R., Geschuhn, K.K. & Vogler, A., 2015. Disrupting the subscription journals’ business 1074
model for the necessary large-scale transformation to open access. Available at: 1075
Transition_White_Paper.pdf [Accessed December 14, 2018]. 1077
Schönfelder, N., 2018. APCsMirroring the impact factor or legacy of the subscription-based model? Avail-1078
able at:
felder%202018%20APCs.pdf [Accessed April 2, 2019]. 1080
Schooler, J.W., 2014. Metascience could rescue the “replication crisis.” Nature, 515(7525), p.9. 1081
Shamash, K., 2017. Article processing charges in 2016 | Jisc scholarly communications. Available at: 1082
2016 [Accessed April 2, 2019]. 1084
Solomon, D. & Björk, B.-C., 2016. Article processing charges for open access publicationthe situ-1085
ation for research intensive universities in the USA and Canada. PeerJ, 4, p.e2264. 1086
Stern, B.M. & O’Shea, E.K., 2019. A proposal for the future of scientific publishing in the life sciences. PLoS 1087
biology, 17(2), p.e3000116. 1088
Tananbaum, G., 2003. Of wolves and and boys: the scholarly communication crisis. Learned pub-1089
lishing: journal of the Association of Learned and Professional Society Publishers, 16(4), 1090
pp.285289. 1091
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
Tempest, D., 2013. Elsevier’s David Tempest explains subscription-contract confidentiality clauses, UK: Osford 1092
University. Available at: [Accessed June 5, 2019]. 1093
Tennant, J., 2018. Democratising Knowledge: a report on the scholarly publisher, Elsevier, Education Interna-1094
tional. 1095
Tennant, J., 2018. Why the term “Article Processing Charge” (APC) is misleading - Green Tea and 1096
Velociraptors. Green Tea and Velociraptors. Available at:
term-article-processing-charge-is-misleading/ [Accessed June 7, 2019]. 1098
Tennant, J. & Brembs, B., 2018. RELX referral to EU competition authority. Zenodo. Available at: https://ze-1099 [Accessed April 4, 2019]. 1100
Tennant, J.P. & Lomax, D.R., 2019. An overview of Open Access publishing in palaeontology. Palaeontologia 1101
electronica , 22(2), pp.110. 1102
Van Noorden, R., 2013. Open access: The true cost of science publishing. Nature News, 495(7442), p.426. 1103
Van Noorden, R., 2017. Publishers threaten to remove millions of papers from ResearchGate. Nature, 112, 1104
p.241. 1105
White, K.E., 2019. Publication output in peer-reviewed science and engineering journals. NSF. Available at: 1106 [Accessed June 6, 2019]. 1107
PeerJ Preprints | | CC BY 4.0 Open Access | rec: 18 Jun 2019, publ: 18 Jun 2019
... Publishers provide services of selection, screening, and dissemination of articles, including ensuring (proper) meta-data indexing in databases. Although several careful estimates are available regarding the cost of academic publishing e.g., [1], one aspect these estimates often neglect is the cost of peer reviews [2]. Our aim was to provide a timely estimation of reviewers' contribution to the publication system in terms of time and financial value and discuss the implications. ...
... They offer their time and highly specialized knowledge to provide a detailed evaluation and suggestions for improvement of manuscripts. On average, a reviewer completes 4.73 reviews per year, 1 yet, according to Publons, 2 1 Based on Personal communication with the Publons team. 2 months=1&order_by=num_reviews certain reviewers complete over a thousand reviews a year. This contribution takes considerable time from other academic work. ...
... The high price of scientific publishing receives a lot of attention, but the focus is usually on journal subscription fees, article processing charges, and associated publisher costs such as typesetting, indexing, and manuscript tracking systems e.g., [1]. The cost of peer review is typically no included. ...
Full-text available
Background The amount and value of researchers’ peer review work is critical for academia and journal publishing. However, this labor is under-recognized, its magnitude is unknown, and alternative ways of organizing peer review labor are rarely considered. Methods Using publicly available data, we provide an estimate of researchers’ time and the salary-based contribution to the journal peer review system. Results We found that the total time reviewers globally worked on peer reviews was over 100 million hours in 2020, equivalent to over 15 thousand years. The estimated monetary value of the time US-based reviewers spent on reviews was over 1.5 billion USD in 2020. For China-based reviewers, the estimate is over 600 million USD, and for UK-based, close to 400 million USD. Conclusions By design, our results are very likely to be under-estimates as they reflect only a portion of the total number of journals worldwide. The numbers highlight the enormous amount of work and time that researchers provide to the publication system, and the importance of considering alternative ways of structuring, and paying for, peer review. We foster this process by discussing some alternative models that aim to boost the benefits of peer review, thus improving its cost-benefit ratio.
... Indeed, publisher-imposed paywalls prevent not only the public from accessing the latest research, the vast majority of which is in fact funded by public money, but also the scientists themselves, who must pay both to publish and to read their own articles. While the recent launch of the 'open access' publishing initiative, Plan S, is beginning to challenge this system, the cost of the resultingly mandatory open access is born not by the publishers, but by the scientists and academic institutions who must often pay increasingly larger sums in order to publish their work free of paywalls [1,2]. Hence, one of the main advantages of preprints is in enabling the free dissemination of scientific findings. ...
... Physics platforms were typically driven by single individuals, and often peaked within two years from launch before disappearing in the following years [11]. While having multiple experimental preprint review platforms will no doubt also benefit the field of biology -because many substitutable options leads to a healthier publishing ecosystem [2] -we should still aim for sustainability: to establish self-sustaining systems of self-organizing peer-review, dependent on and run by the wider community rather than a few individuals [12]. The key in our view consists of approval by the field, an operating system that is intuitive and easy to use, low workload, adequate mitigation of perceived risks to reputation and journal submission and reliable stewardship far into the future under a managing board that is refreshed periodically. ...
Full-text available
The dissemination of scientific results and new technologies in biomedical science is rapidly evolving from an exclusive and fee-oriented publishing system towards more open, free and independent strategies for sharing knowledge. In this context, preprint servers such as bioRxiv answer a very real scientific need by enabling the rapid, free and easy dissemination of findings, regardless of whether these are novel, replicated, or even showcasing negative results. Currently, thousands of manuscripts are being shared via bioRxiv each month, and neuroscience is the largest and fastest growing subject category. However, commenting on bioRxiv is declining and no structured scientific validation such as peer-review is currently available. The Peer Community In (PCI) platform addresses this unmet need by facilitating the rigorous evaluation and validation of preprints, and PCI Circuit Neuroscience (PCI C Neuro) aims to develop and extend this tool for the neuroscience community. Here we discuss PCI C Neuro’s mission, how it works, and why it is an essential initiative in this new era of open science.
... Regardless of any potential waiver or discount systems in place (including through negotiations between universities and publishers), the total APC paid by the geochemistry community is Geochemistry has proposed an APC of US$ 900. Such an APC is still high when compared with US$ 400 estimated by Grossmann and Brembs [13] for the true costs of article production, and there are cases where the production costs are 2 orders of magnitude even lower than this [14]. At the moment, it remains eminently unclear why such vast differences appear, and what such an excess of money is being spent on, which represents billions of dollars of taxpayer money being spent each year. ...
In this short communication, we look at Open Access (OA) practices in geochemistry based on articles published in 2018 and 2019 in a list of 56 journals and whether Article Processing Charge (APC) and Journal Impact Factor (JIF) drive publication or not. More than 40% of articles in 2018-2019 were published OA, and about 70% of them in fully OA journals with a mean APC of US$ 900 whereas the remaining were published in historical hybrid journals with higher APC of more than $US 1,800. A good correlation is observed between the number of OA articles published in hybrids journals and JIF whereas there is a positive relationship between the number of OA articles published in fully OA journals and APC. For OA articles published in hybrid journals (i.e. the majority of historical journals in geochemistry), it seems that the proportion of OA articles tends to increase in journals with higher JIF.
This paper investigates whether free access to scientific literature increases the participation of under‐represented groups in scientific discourse. To this end, we aggregate and match data tracing access to Sci‐Hub, a widely used black open access (OA) repository or shadow library, and publication data from the Web of Science (WoS). We treat the emergence of Sci‐Hub as an exogenous event granting relatively unrestricted access to publications, which are otherwise hidden behind a paywall. We analyze changes in the publication count of researchers from developing countries in a given journal as a proxy for general participation in scientific discourse. Our results indicate that in the exemplary field of mathematics, free access to academic knowledge is likely to improve the representation of authors from developing countries in international journals. Assuming the desirability of greater international diversity in science (e.g., to generate more original work, reproduce empirical findings in different settings, or shift the research focus toward topics that are overlooked by researchers from more developed countries), our findings lend evidence to the claim of the OA movement that scientific knowledge should be free and widely distributed.
Nesta obra reúnem-se cerca de meia centena de textos que retratam, a partir de diferente experiências, perspetivas e olhares disciplinares, a forma como a pandemia provocada pelo vírus SARS-CoV-2 foi sentida na Universidade do Minho, a primeira universidade portuguesa a ser atingida pelos seus efeitos, a partir de 7 de março de 2020. Os contributos aqui reunidos exprimem o modo como a comunidade académica se organizou para garantir o funcionamento da instituição e o cumprimento da sua missão, quer no âmbito do ensino, quer da investigação e da inovação, não dispensando uma necessária reflexão crítica sobre a nova realidade que se afirmou com a pandemia, o seu significado e os impactos da mesma na sociedade. Dado o elevado número e a diversidade dos contributos, a obra organiza-se em três volumes, que tentam dar resposta a três possíveis interrogações. Assim, o 1º volume, com o subtítulo de ‘Reflexões’, procura equacionar diferentes perspetivas em torno da questão: Mas o que é isto? Já o 2º volume, que recebe o subtítulo de ‘(Re)Ações’, corporiza um conjunto de textos que reflete as atuações em diferentes áreas da dimensão académica, elucidando quanto ao modo: Como reagimos? Finalmente, o 3º volume, que acusa o subtítulo de ‘Projeções’, dá expressão a uma inevitável pergunta: E agora? Pois, afinal, todos sabemos que isto não vai, nem pode, ficar tudo bem.
Full-text available
Background Latin America has historically led a vital and open access movement and leads the worldwide region with the adoption of wider open access practices. Argentina has expressed its commitment to join Plan S, an initiative from a European consortium of research funders with a mandate to promote the open access publishing of scientificoutputs. Analysis This opinion article suggests that the potential adhesion of Argentina orother Latin American nations to Plan S ignores the reality and tradition of LatinAmerican open access publishing. Conclusion and implications Plan S must demonstrate that it will encourage at a regional and global level the advancement of non-commercial open access initiatives. Latin America should invest and promote open nonprofit scholarly-led infrastructures, allowing the academy to regain control of scholarly communications. Keywords Plan S; Latin America; Open access publishingContexte Historiquement, l’Amérique latine a été un chef de file dans le mouvement pour fournir un accès aux communications savantes qui soit vital et ouvert. Aujourd’hui, ce continent demeure un leader mondial en adoptant des pratiques de libre accès qui sont plus englobantes. L’Argentine, notamment, a manifesté sonengagement pour se joindre au Plan S, lequel est l’oeuvre d’un collectif de bailleurs de fonds européens dont le mandat est de promouvoir l’édition de recherches scientifiques à libre accès. Analyse Cet article d’opinion suggère que l’adhésion potentielle au Plan S par l’Argentine ou d’autres pays latino-américains ignore la réalité et l’histoire de l’édition à libre accès latino-américaine. Conclusion et implications Le Plan S doit démontrer qu’il encouragera des initiatives non-commerciales envers le libre accès à un niveau régional et mondial. L’Amérique latine doit investir dans des infrastructures sans but lucratif qui soient ouvertes et menées par des universitaires, de manière à permettre à ceux-ci de reprendre le contrôle sur les communications savantes.Mots clés Plan S; Amérique latine; Édition à libre accès
La science ouverte est devenue une priorité de premier rang de la politique de recherche de l’État français. Parmi les axes de cette politique, se trouve l’ouverture des données de recherche et des publications scientifiques. A partir de ses fonctions traditionnelles (qualité, propriété intellectuelle, diffusion, conservation), l'article interroge l'impact de la politique de la science ouverte sur les revues scientifiques, en particulier dans trois domaines : leur rôle dans le fonctionnement des dispositifs d’évaluation et de suivi (fonction politique), leur rôle dans le fonctionnement des dispositifs de données, en particulier pour l’acquisition et l’exploitation des données massives par l’industrie de l’information (fonction Big Data), et leur rôle dans le fonctionnement économique des éditeurs, agrégateurs, agences etc. (fonction économique). Dans la mesure où l’accès aux revues devient gratuit, la question se pose si l’usage des revues ne devient pas le véritable produit des plateformes, par le biais des informations générées.
Full-text available
A scholarly communication system needs to register, distribute, certify, archive, and incentivize knowledge production. The current article-based system technically fulfills these functions, but suboptimally. I propose a module-based communication infrastructure that attempts to take a wider view of these functions and optimize the fulfillment of the five functions of scholarly communication. Scholarly modules are conceptualized as the constituent parts of a research process as determined by a researcher. These can be text, but also code, data, and any other relevant pieces of information that are produced in the research process. The chronology of these modules is registered by iteratively linking to each other, creating a provenance record of parent and child modules (and a network of modules). These scholarly modules are linked to scholarly profiles, creating a network of profiles, and a network of how profiles relate to their constituent modules. All these scholarly modules would be communicated on the new peer-to-peer Web protocol Dat, which provides a decentralized register that is immutable, facilitates greater content integrity than the current system through verification, and is open-by-design. Open-by-design would also allow diversity in the way content is consumed, discovered, and evaluated to arise. This initial proposal needs to be refined and developed further based on the technical developments of the Dat protocol, its implementations, and discussions within the scholarly community to evaluate the qualities claimed here. Nonetheless, a minimal prototype is available today, and this is technically feasible.
Full-text available
Open access publishing has frequently been proposed as a solution to the serials crisis, which involved unsustainable budgetary pressures on libraries due to hyperinflation of subscription costs. The majority of open access articles are published in a minority of journals that levy article processing charges (APCs) paid by authors or their institutions upon acceptance. Increases in APCs is proceeding at a rate three times that which would be expected if APCs were indexed according to inflation. As increasingly ambitious funder mandates are proposed, such as Plan S, it is important to evaluate whether authors show signs of price sensitivity in journal selection by avoiding journals that introduce or increase their APCs. Examining journals that introduced an APC 4-5 years after launch or when flipping from a subscription model to immediate open access model showed no evidence that APC introduction reduced article volumes. Multilevel modelling of APC sensitivity across 319 journals published by the four largest APC-funded dedicated commercial open access publishers (BMC, Frontiers, MDPI, and Hindawi) revealed that from 2012 to 2018 higher APCs were actually associated with increased article volumes. These findings indicate that APC hyperinflation is not suppressed through market competition and author choice. Instead, demand for scholarly journal publications may be more similar to demand for necessities, or even prestige goods, which will support APC hyperinflation to the detriment of researchers, institutions, and funders.
Full-text available
Science advances through rich, scholarly discussion. More than ever before, digital tools allow us to take that dialogue online. To chart a new future for open publishing, we must consider alternatives to the core features of the legacy print publishing system, such as an access paywall and editorial selection before publication. Although journals have their strengths, the traditional approach of selecting articles before publication ("curate first, publish second") forces a focus on "getting into the right journals," which can delay dissemination of scientific work, create opportunity costs for pushing science forward, and promote undesirable behaviors among scientists and the institutions that evaluate them. We believe that a "publish first, curate second" approach with the following features would be a strong alternative: authors decide when and what to publish; peer review reports are published, either anonymously or with attribution; and curation occurs after publication, incorporating community feedback and expert judgment to select articles for target audiences and to evaluate whether scientific work has stood the test of time. These proposed changes could optimize publishing practices for the digital age, emphasizing transparency, peer-mediated improvement, and post-publication appraisal of scientific articles.
Full-text available
Although a case can be made for rewarding scientists for risky, novel science rather than for incremental, reliable science, novelty without reliability ceases to be science. The currently available evidence suggests that the most prestigious journals are no better at detecting unreliable science than other journals. In fact, some of the most convincing studies show a negative correlation, with the most prestigious journals publishing the least reliable science. With the credibility of science increasingly under siege, how much longer can we afford to reward novelty at the expense of reliability? Here, I argue for replacing the legacy journals with a modern information infrastructure that is governed by scholars. This infrastructure would allow renewed focus on scientific reliability, with improved sort, filter, and discovery functionalities, at massive cost savings. If these savings were invested in additional infrastructure for research data and scientific code and/or software, scientific reliability would receive additional support, and funding woes—for, e.g., biological databases—would be a concern of the past.
Full-text available
We believe that Elsevier and other major publishers are continuing to engage in anti-competitive practices, which are continuously worsening, and that information gained in the last 15 years urges immediate investigation and intervention into this unregulated market space. This could be, for example, through an empirical analysis of the scholarly publishing market; by having an independent regulatory body monitoring and overseeing the digital services provided by Elsevier and others within the industry; banning the use of non-disclosure clauses in licensing contracts; requiring transparency into the production costs of research articles and publishing operations; banning the use of inappropriate journal-level metrics in hiring, granting, and promotion decisions; abolishing copyright on journal articles; and encouraging the wider establishment of library consortia to increase buyer power (and therefore simultaneously potentially reduce the monopoly power of publishers). It is our view that action of this sort is clearly required in order to transform the present scholarly publishing sector into one characterised by sustainable service-based competitive market conditions, and this cannot be achieved without taking explicit measures against the business practices of Elsevier and other large publishers. We finish by noting that Plan S is a small and welcome step in this direction, but requires substantially more conviction if it is to lead to the functionality of a truly competitive scholarly publishing market.
p>The pros and cons of hybrid open access are heavily disputed. A main point of discussion is whether ‘double dipping’ takes place, i.e. paying twice to publish and read the same article. To examine publishers’ assertions that they do not double dip, a survey was conducted of 24 publishers with detailed questions about their pricing policy using hypothetical examples of hybrid open access publication in their journals. The outcome is quite sobering: while a small number of publishers appear to be fully offsetting their hybrid open access income, or making no additional charge for hybrid open access at all, for the rest of those surveyed no clear evidence could be gathered that double dipping does not take place, and many appeared to be double dipping to some extent.</p