Threat of job loss as motivation for Research Works Act: real or fear-mongering?

I find it difficult to silently read the AAP endorsement of the Research Works Act.  Multiple readings later I still find myself yelling aloud at its twisted positions and statements.

Most of my rebuttals take the form of expletive + “no, that isn’t true” or “give me a break” + a silent prayer that policymakers won’t take the positions seriously.

But what about the AAP threat that the industry and 30,000 jobs are in danger if the government supports, requests, or requires more open scholarship?  I suspect many politicians in the US government will take that very seriously.  Threatening job losses in the US right now is a Great Card, and publishers keep bringing it to the table.

The Jobs Card makes me furious because *if* scholarship can be done better it *should* be done better, full stop.  But with the government really considering about these issues and soliciting feedback, it is time to get beyond that and see if, you know, it happens to be true that 30,000 US jobs are in jeopardy.  If so, we need to figure out how to address that in our discussions, because we can be darn sure the government will balance that against the benefits of openness.

First, what is at risk?  Scholarly publishing as an industry isn’t going away any time soon: physicists have arXiv and they still publish in journals.  What is at risk is the traditional subscription-based business model.  If the scholarly article-of-record is made immediately available with no restrictions, there is no reason for anyone to pay subscription fees.  Luckily, BMC and PLoS have already demonstrated that an author/author’s funder-pays model can work (at least for fields funded by federal money, with access to publication budget line items… the same fields that would be subject to federal mandates on openness).  Traditional publishers could move to this model if they wanted to.  So what is at risk is the business model, not the industry.

Second, what about the jobs?  Well, the publishing industry would still need to employ people, even if it changes to a PLoS business model. Maybe fewer people, though, since PLoS has been designed from the ground up to be a streamlined operation?

Queue the Saturday-morning kitchen-table back of envelope analysis (ok, post-its).
*** btw if others have more refined or robust estimates of the job situation, please share!

In 2010, the first year PLoS reported covering its costs with revenue, PLoS published 8662 papers.  PLoS has about 120 employees.  That’s 72 papers/employee.

Let’s compare that to traditional publishing models.  Elsevier publishes 250,000 papers per year.  Elsevier has 7000 employees worldwide.  That’s 35 papers/employee.  However, Elsevier also works on non-article products and services, like book publishing, SciVerse software, Gray’s Anatomy, MD Consult, and conference organizing.  I have no idea what proportion of their employees work on these things.  If 20%, the number of employees working on articles is 5600.  So 250,000/5600 = 45 papers/employee.

What does this mean?

At 45 papers/employee, publishing 250,000 papers at a PLoS-level of efficiency would take 250,000/72=~3500 employees.  We estimated that Elsevier has 5600 worldwide employees working on research articles now, so 2100 more than would be necessary at PLoS levels of efficiency.  Would these people have to be let go?  Heck no!  Publishers often argue that they offer a better product than the bare-bones PLoS offering — those 2100 people are doing something useful — so they could charge more money for their product than PLoS!  Compete!  Tout your high impactness and careful copyediting!  $500 more per article would do it (250000*500/2100=$60k).  Totally reasonable, and much less than many existing OA fees.

Let’s think about how many jobs it really is in the USA.  Björk et al estimates there were about 1.35 million articles are published per year in 2008: let’s call it 1.5 for 2010.  1.5 million papers/72 employees per paper=20,100 scholarly communication jobs would be required to publish that many papers, at a PLoS levels of efficiency.  1.5 million papers/45=33,333 jobs would be required at an Elsevier level of efficiency.  The difference is 13k jobs worldwide.  Worldwide.  AAP claims that 45% of worldwide scholarly publishing is by North American science publishers.  Let’s say US science+non-science publications is 50%, maybe?  50% of 13k jobs =~ 7,000 jobs.  I reiterate that these jobs do NOT have to be lost.  Presumably these employees are indeed creating value (through more thorough copyediting or peer-review routing or paper-article-publishing or whathaveyou) then the publishers could (and should) CHARGE EXTRA and pay for these jobs that create this added value.

But let’s give a bit of perspective to 7,000 jobs.  Google “jobs lost USA.”  This was the top hit when I did it: nearly 1000 people out of work because a hospital is closing.

Enough said.

Publishers are fear mongering with talk of loss of jobs.   As far as I can tell, there is no big risk, if publishers are willing to move with the times and embrace new business models.

The US government has no business sacrificing the progress of science and taxpayer access to maintain a particular business model.

The opportunity cost, however, of continuing with the current pathetic dissemination of research results?  Don’t get me started.

Go set your government official straight about the Research Works Act and Public Access in general (due Jan 12)!   Feel free to use analysis in this post with or without attribution.  

If you know of other sources of evidence and analysis around this issue, on either side, please let me know.