May 23, 2017

GSU e-Reserves Decision | Peer to Peer Review

Kyle_CourtneyOr, How I Learned How to Stop Worrying, Reject Licensing, and Embrace Fair Use and e-Reserves

The infamous Georgia State University (GSU) e-reserves case (Cambridge University Press v. Patton) emerged last month from its long winter slumber to give us yet another 200+ page decision which librarians, lawyers, and publishers have begun to parse and analyze. And, like me, they are probably asking themselves: What does this decision actually mean?

For a quick backdrop, here are the salient facts: In April 2008, three academic publishers were invited by the Copyright Clearance Center (CCC) to sue GSU for “pervasive, flagrant and ongoing unauthorized distribution of copyrighted materials” on their library’s e-reserve system—specifically for the nonfiction textbook chapters. CCC agreed to pay half the legal fees if these presses brought a lawsuit.

In 2012, Judge Orinda Evans found copyright violations in only five of 99 excerpts. GSU’s e-reserves policy, she held, had been a good faith interpretation of the fair use provision under copyright law. Evans’s 323-page decision spelled out very specific guidelines to be used in evaluating the fair use of textbook chapters in an e-reserve system.

The ruling was appealed by the presses to the Eleventh Circuit Court of Appeals.

In October 2014 the 11th Circuit Court of Appeals vacated the earlier decision and sent it back to Judge Evans for another fair use review, this time using the different, specific directions the Court laid out for a “proper” fair use analysis. It was thought, at the time, the 11th Circuit did not like Judge Evans’s mechanical analysis—she had used numbers to calculate portions of her analysis, e.g. ten percent or less; one chapter or more, etc.

On March 31 of this year Judge Evans issued a new opinion after reanalyzing the allegedly infringing works according to the directions of the 11th Circuit Court of Appeals. This time, Judge Evans found only four cases of infringement among the 48 works, and designated GSU the prevailing party. (More on “prevailing party” later.)

Note that only 48 were appealed. Of 26 of the texts discussed in the original case, Evans had ruled there was no case for copyright infringement because 1) the publishers could not show they held the copyright, or 2) there was no evidence that any students had used the excerpt.


Two appeals and three decisions later, the publishers have lost one count of infringement—from five infringements in 2012 to four in 2016. That’s a heck of a lot of legal fees for one less finding of infringement.

The path of this particular victory for GSU and the e-reserves system is stranger, than I would have ever predicted.

The 11th Circuit in its appeal specifically asked that Judge Evans review the fair use analysis without any of her earlier mechanical analysis. Yet in the opening of the fair use portion of the case, she states:

“This Court estimates the initial, approximate respective weights of the four factors as follows: 25% for factor one, 5% for factor two, 30% for factor three, and 40% for factor four.”

That is certainly one of the more bizarre fair use statements I have seen. Is Judge Evans thumbing her nose at the 11th Circuit’s rejection of her mechanical fair use analysis, which worked, in my opinion, rather well?

Either way, this is how she weighs the four factors of the fair use test in the opinion. Of course, this is not a transformative fair use case, in which, arguably, the percentages would be far different.

Let us go quickly through Judge Evan’s new fair use analysis:

Factor One: Thankfully, “[f]actor one will favor fair use in all cases. It will not ‘strongly favor’ fair use.” Here Evans reaffirmed that the first fair use factor—the purpose and character of the use—will still favor fair use when that purpose is an educational, nonprofit purpose.

Factor Two: While Judge Evans had previously decided that the use of nonfiction works caused factor two to weigh in favor of fair use, the 11th Circuit rejected this presumption and it was sent back for review. Relying on the 11th Circuit’s instructions, Judge Evans now considered whether the nonfiction chapters were intertwined with scholarly analysis.  Despite this new formula, she mitigated this factor as relatively unimportant by saying it only makes up five percent of the total fair use calculation. Therefore, factor two— nature of the work—is generally neutral, and does not weigh against fair use.

Factor Three: This factor under the 11th Circuit’s instructions has a new specific analysis, which, in my opinion, mingles factors one, three, and four in a way that could never be used in a real-world scenario. Factor three now

“will take into account the effect of the favored nonprofit educational purpose of the use under factor one, plus the impact of market substitution as recognized under factor four, in determining whether the quantity and substantiality (value) of Defendants’ unlicensed copying was excessive. All relevant record evidence will be considered; the factor three outcomes will vary.”

So Evans’ initial useful and easy “ten percent or one chapter” rule of thumb is gone. Instead, Evans looks at each excerpt, notes the number of pages, and calculates 1) if it’s appropriate to the purpose (from factor one) 2) its potential to substitute for a purchase of the book (market substitution from factor four). Where the excerpts are a very small number of pages and a small percentage of the entire work, and suitable to accomplish a specific pedagogical objective, then fair use will be found.

Factor Four: This factor also receives a new analysis. If I were to summarize it simply, the new analysis seems to be that the fourth factor—market harm—must be so extensive that it would completely undercut the incentive to create or publish the work in the first place. That is a very high burden to overcome. The market has to be harmed to such a degree that the publisher feels it’s not even worth publishing the book.


How does the court analyze this? The decision reprints many charts revealing a lot of financial information about each book, including sales over specific time periods and the amount of money resulting from the permissions market. This seems to be the only method to apply this new fourth factor calculation: finding and analyzing the financial data.


Sales permissions from GSU case decision

How exactly are libraries supposed to perform this analysis when we are doing our best fair use calculation for e-reserves? Many of these publishers, I assume, would not want to give away this data at all, but had to for the court case.

In my opinion, we are not supposed to. As I stated earlier, even with the data, the standard for this fourth factor is high—the market harm must be so extensive that it would completely undercut the incentive to create or publish the work. I am doubtful that a well-crafted fair use calculus for nonprofit, educational use of a textbook chapter would ever reach this standard. And even if the well-meaning e-reserves librarian has misjudged fair use, §504(c)3 of the Copyright Act protects employees of nonprofit educational institutions or libraries from serious statutory damages.

Final calculus: of the 48 remaining excerpts, 44 were found to be fair use.

How should the library community read this decision? Should we alter our e-reserve policies? Do we have to start soliciting sales data?

I say we stay the fair use course. If anything, this decision has led me to conclude that licensing is for the birds. How can I say this? Well, in the analysis, the greater the chapter licensing revenue was for the publisher, the lesser the scope for fair use. So, by analogy, the less we license, the more room for fair use. Therefore, let’s continue to make the best fair use decisions possible instead of licensing: let’s make our e-reserves more fair use–friendly, and focus on fair use and other alternatives to licensing fees. The moment we start paying licenses, we will actually be hurting our own cause. (Not to mention that paying licensing fees directly to licensing companies helps fund lawsuits against libraries, as the CCC does in the present case—taking our money and using it against us).


So that’s it, right?

Far from it.

Remember the part about the “prevailing party”? Under the law, the prevailing party has the right to recover attorney fees and costs. And when you have litigation that has stretched out over eight years and three court decisions, those fees can be hefty. GSU attorneys recently asked for $3,236,235 for their attorneys’ fees, and an additional $85,746 in costs. And GSU attorneys have stipulated, those costs could go up yet again—there might be more legal action ahead.

The publishers have filed, yet again, to be able to collect current evidence about GSU’s practices. The case was decided based on prior conduct of GSU, but the publishers want a list of online e-reserve readings from the most recent academic terms to “refresh” the record and look for more “pervasive, flagrant and ongoing unauthorized distribution of copyrighted materials.” That same argument has been rejected several times in the past by Judge Evans.

Why are they trying again? The publishers do have those four infringements to deal with, and if they are to ask for an injunction, the publishers feel the court should give them the latest information—as they recently demanded in their brief, the injunctions must be “tailored to current conduct at GSU.”

Briefs from both parties will be filed during the second week of May, and Judge Evans will find herself ruling on the record yet again, which could mean another appeal.

Maybe we will see the close of this case by 2018?

Register for the LJ Day of Dialog Livestream!
  For the first time in Day of Dialog history, we’re offering livestream access for those unable to attend in person! Tune in on May 31 to watch a full-day of author presentations and panels. You’ll also receive access to the event archive for two months.