Fair Use for the Google Goose; Fair Use for the Book Gander
While many in the library, research, and technology communities have hailed Google’s recent fair use victory, some have expressed skepticism about who is the real beneficiary of the ruling. Some librarians have wondered whether their enthusiasm about the ruling should not “be tempered by the understanding that Google is not the do-no-evil corporation it once represented itself as being?“
Others, like Eric Goldman, observed that even though the ruling opens the door for competitors to replicate Google’s book scanning and presentation of snippets, Google already enjoys a significant first-mover advantage, making it unlikely that others would find it profitable to spend millions of dollars only to catch up with it. I agree that at least for some foreseeable future, Google Books will remain the only game of its kind in town. Even if the ruling stands on appeal, we should not expect any quick entry by any of Google’s existing or future competitors embarking on a similar mass digitization project on a the same scale, to do exactly the same thing. But it does open the door for others to do different, and possibly better, things, and this is important.
Moreover, even though the ruling itself cannot create competition and cannot change the cost structure of mass digitization project, it has removed one barrier–a legal barrier–to competition. This is significant, because any other outcome would only make competition even less likely.Therefore, regardless on whether one stands on perennial questions such as “Google: awesome or evil?”, or “should mass digitization of our cultural assets be done by private enterprise or by public institutions”, or “Google Books: good or bad for privacy”, the fact is that Google has created Google Books, it has digitized millions of books, it will not disappear, and it may not face serious direct competition in the near future. Therefore, anyone who has concerns about a Google-as-monopoly should be happy with the decision. Any other outcome would not have weakened that monopoly, but only entrenched it further.
So yes, Google may be the only game of its kind in town. And yet, Google has to take care about its own self-interest and those of itself shareholders, not of the public. And yes, this may not be ideal, and there is no assurance that it won’t be evil. But even then, the public is still better off as a result of Judge Chin’s ruling, and would have been worse off if the case had been decided differently.
Natural monopoly? Maybe, but weaker
Judge Chin’s ruling means that what’s fair use for the goose is fair use for the gander. If we won’t see a flurry of competitors entering this market it won’t be because they are not allowed to, but because they may not find it profitable. This is a subtle but very important difference. Competitors may not find entry profitable because a project such as Google Books exhibits some features of a natural monopoly: having already incurred many of the very high sunk costs required for such a project, and having already scanned millions of books, Google now has a significant cost advantage over any other competitor contemplating a similar project.
Is such a natural monopoly cause for concern? It could be (see more below). There is no reason to trust a Google monopoly more than we would trust any other monopolist. Google may promise to do no evil, but it cannot credibly promise never to change its mind.
But is this a reason to panic? Should one start drafting amici briefs supporting the Authors’ Guild in the appeal that it promises to file? Not really. If the cost structure in an industry is such that a natural monopoly is likely to emerge, then the important question is not whether there is a monopoly or not, but how easy it is for other firms to enter the market or disrupt the it if the monopoly begins to do evil. In economic lingo, the question is how contestable that monopoly is; the more contestable the monopoly is, the more likely it is to behave like a firm in a competitive industry. Judge Chin’s ruling cannot turn the market into a perfectly contestable one, but if it is upheld on appeal it would ensure that Google’s monopoly becomes much more contestable than it would be under any other outcome of this case.
To see that, let’s consider what would the market look like under three alternative outcomes: an Authors’ Guild victory, another settlement, or legislative action by Congress. Under each of these alternatives, Google would be only slightly worse off, or not at all, while the public would be significantly worse off.
Authors’ Guild victory
The Authors’ Guild goal in this litigation has never been preventing Google from digitizing the books of its members. The Guild’s strategic goal (and that of the publishers who filed their own lawsuit and settled earlier last year) has been to become a real guild, not only by name. They hoped that winning the case would lead to the creation of a new type of collecting society that would collect money not only from Google, but also from anyone else who might embark on any digitization project of books. Like the medieval guilds who controlled their trades, their goal has been to control the digital book trade.
How do we know that? First, because this was one of their main arguments against fair use. The Guild argued that a finding of fair use would impede the development of collective management models for the digital uses of books and excerpts from books. Second, because the Guild’s companion organizations in other countries and the publishers’ trade associations have lobbied for such collective models in other countries, especially in Europe. And third, because that’s exactly what they attempted to do when shortly after filing the lawsuit they began negotiating a settlement with Google, which contemplated exactly that.
Google, of course, was also interested in that settlement, which raises an interesting question: if Google was confident that what it was doing was fair use, then why did it choose to settle rather than litigate. More specifically, why did it prefer a settlement that would force it to share its monopoly rents with un upstream author-publisher monopoly, instead of litigating the fair use question and keep the monopoly to itself? There are three main reasons for that: the first is that as James Grimmelmann told the NY Times “What seemed insanely ambitious and … very dangerous in 2004 now seems ordinary.” Not only the case-law on fair use has become more hospitable to this kind of endeavours, as time went by, it has become easier to demonstrate the public benefits of its project and difficult for the plaintiff to demonstrate any serious harm.
The second reason is that if Google had lost, the amount of statutory damages that it might have been exposed to could have been astronomical. Even if the sabre of statutory damages would only have been used to compel a settlement, Google would have bargained from a less advantageous position.
The third reason is that the settlement would have created a higher barrier to entry, because it would allow Google to continue with its project, while denying this benefit from its competitors. Essentially, Google agreed to install a new monopoly upstream and share its profits with it because that was a price worth paying in order to make entry into its markets more difficult. Fortunately, Judge Chin who rejected the settlements prevented that from happening.
Another settlement
Even though the court rejected the first two proposed settlements, Google and the plaintiffs could still negotiate another settlement instead of litigating the case (and theoretically they might still do that to avoid a ruling on appeal). However, the strong opposition to the earlier settlements, and their empathic rejection by Judge Chin, have reduced the scope of a jointly profit maximizing settlement. After all, much of the surplus that the rejected settlements would have created came from its collusive elements: monetizing the works of all copyright owners, not only those represented by the plaintiffs, from the collusive price-fixing structure that it contemplated, and from the higher barrier to entry that they would have created. When those settlements were rejected, the parties had much less collusive spoils to divide among themselves. At the same time, Google, who never stopped digitizing new books, has been able to entrench its dominant position. At this point, Google’s dominance is secure enough that it probably has more to gain by successfully pursuing its fair use claim than by a settlement. A settlement might make its monopoly less contestable, but still requires it to pay something. A fair use victory means that it doesn’t have to pay anything, and allows it to negotiate licenses for the uses that won’t be fair from a much more convenient position.
Legislative action
In theory, Congress could take action regardless of which way the case were decided, but it’s hard to see how the outcome would be better than a fair use victory. Had the court accepted the Authors’ Guild theory of infringement, the implication would have been that no mass digitization project could be pursued without the consent of all the relevant copyright owners, including those who own the copyright in out-of-print or orphan works. In theory, Congress could have reversed such a ruling, but I wouldn’t bet on that. Rather, my bet is that a victory for the plaintiffs would boost efforts to solve the inevitable market failure by establishing some sort of extended collective licensing, which despite its progressive sound, does not ensure competition at all. Extended collective licensing, as contemplated in some European countries, only means that a non-contestable monopoly of copyright owners can effectively tax all downstream uses (see more here).
Compared to any of these alternatives, Google’s fair use victory is clearly preferable. Yes, it’s good for Google, but Google would be doing fine either way. And yes, Google Books is not likely to face any serious competition anytime soon, but the fair use outcome is the one that makes its monopoly more contestable than any other outcome. Any other outcome would not replace its monopoly with a competitive market, but only make Google’s monopoly more secure.
Google’s natural monopoly: revisited
I suggested above that Google may remain the only player of its kind because of the natural monopoly characteristics of the market. Now it’s time to add some nuance.
Saying that a particular product or service has characteristics of a natural monopoly does not necessarily mean that there will be one and only one producer. It only means that it is not likely that another firm would find it worthwhile to invest in order to do exactly the same thing. However, if entry is permitted, competition may nonetheless exist. For example, if the product or service is integrated with other products or services, existing firms might still find it worthwhile to invest in digitization: if Google Books improves Google Search, Microsoft might find it worthwhile to digitize books to improve Bing; Amazon, which already has an impressive corpus of digital texts, might be interested in offering some of its capabilities for research purposes, etc. There are many examples of competition even when production exhibits characteristics of a natural monopoly: Westlaw and Lexis coexist, JSTOR exists, alongside other journal aggregators. Cable and DSL compete for internet connection, cable and satellite for multichannel TV, ASCAP competes with BMI, and iTunes may be the dominant online retailer of downloadable songs, but it’s not the only one, and it also competes with streaming services such as Pandora. The list goes on.
The point is that as long as entry is permitted–and thanks to Judge Chin some entry will be permitted–we can expect to see entry if someone might figure out some kind of service that it can offer differently or better than Google’s. Google itself is a good example for that. Google was not born a monopolist. Just as Google relied on fair use to outsmart the earlier search engines against which it had competed, someone else might rely on fair use to outsmart Google.
Moreover, Judge Chin’s fair use ruling makes digitization easier not only for potential commercial competitors, but also for non-profit entities, public or private, that rely on different funding structures, and have different goals from commercial ones.
Indeed, even before Google’s latest victory, the field has already been populated various players, who can be expected to take advantage of the ruling. The Internet Archive has been digitizing books–principally public domain books–for years; the HathiTrust has already secured an important victory against the Authors’ Guild, and if the HathiTrust ruling and the Google Books ruling are upheld on appeal, those digitization projects will not only continue with greater force, they might also prompt other projects that so far have waited until the legal landscape has clarified. And above all, the ruling boosts the Digital Public Library of America, who intends to interconnect between all of those different digitization projects and make them accessible through a single platform and portal.
So overall, for Google, Judge Chin’s ruling is but a small victory, but for the public it is a giant one.
Setting aside, for a moment, questions about the applicability of Judge Chin’s ruling in Canada, it certainly appears to facilitate the process of making available offerings in regional history and other niche areas by public and academic libraries. Often such holdings include periodical publications of considerable local (to the niche) interest, but whose presentation on institutional websites has been curtailed by concerns about authorial permissions. Indexing these and supplying contextual snippets while waiting for the term of copyright to expire so that the entire article may be presented will solve major headaches for institutions involved in creating digital archives. With all due respect to Google and the participating libraries, there is more to the world of learning than their large collections combined.
I agree. Your point is another illustration of the need for nuance when analyzing issues of natural monopoly. A natural monopoly exists when there are economies of scale over the entire supply curve. But beyond a certain point there supply curve may begin raising up. E.g., such as when the marginal cost that Google has to incur in order to add another collection of local and niche interest in, say Newfoundland, exceeds Google’s marginal benefit, but not the marginal benefit of the the local institutions in Newfoundland. Such institutions might be direct beneficiaries of the ruling. And now to Canada, this is not legal advice, etc., but Canadian libraries could have started the scanners’ engines a long time ago. What’s SOCAN v Bell Canada if not a case about indexing and supplying contextual snippets?
Ariel, the argument you develop in the last part is reasonable but largely speculative. Books are owned by libraries, and libraries have already digitized their books with Google. A would-be second comer should be able to convince libraries to digitize their books twice. This cannot happen, unless a new technology emerges (perhaps books 3D printing?) The Internet Archive digitizes public domain books, an activity that does not depend on fair use. The same applies for the partners of the DPLA. HathiTrust is a consortium of libraries that are partner of Google.
Google already pays licenses to publishers in order to search and display snippets of news stories in Google News (as explained in AP v Meltwater–incidentally, a decision that is worth comparing with Author’s Guild v Google). Why would a ruling obliging Google to do the same with books’ publishers be a tragedy for the public?
Maurizio, you say “speculative”, I say “theoretical” 😉
But I’m actually fine with “speculative”, because I’m not arguing that there would be a second-comer trying to do exactly what Google has done, and I’m not trying to predict who might be a second-comer and what exactly it would do. If I could predict those things I wouldn’t be teaching law. The point is that the ruling makes entry easier that it would be under any other outcome of the case, and if someone might find it worth its while (economically) to enter the law won’t block it. Will that happen? I don’t know, but we have to take a longer term perspective. If we were discussing that even ten years ago we would probably think that something like Google Books is speculative, and if we were discussing it twenty years ago we would certainly think that something like Google Search is speculative.
The fact that we cannot imagine who might enter does not make diminish the significance of the ruling. In fact, it makes it even more significant, precisely because it helps ensuring that the next Google can enter if it can figure out something different or better to offer.
Regarding the specific points that you’ve raised:
(1) some libraries have digitized their books with Google, but many have not. Those who have partnered with Google might not see any point in digitizing their books twice, but others might. UofT’s library, for example, the third largest research library in North America was offered to participate in Google Books but ultimately decided not to, and I assume that many smaller libraries might be happy to digitize their collections, and as Richard Ellis pointed out in his comment below, there are many collections that await digitization. And by the way, copyright term in Canada is life+50, so we have 20 more years of public domain books.
(2) I believe that the origin of digital files in HathiTrust has been the scans made by Google, but as far as I know HathiTrust is not a partner of Google, and for some functions might actually be a competitor.
(3) The Internet Archive has so far scanned only public domain books, but I’m not sure that necessarily limited to that. In fact, the Archives archives many things that are under copyright (e.g., most of the “internet”).
(4) The Meltwater case is a good demonstration of my point. Google has actually been quite happy to pay license fees when it gets more than what it would be allowed to do under fair use (and that’s a good thing), but also because–at least when it’s a dominant player–whatever it agrees to pay raises the costs of its rivals, who are also expected to pay (case in point, Meltwater). If that strengthens Google’s dominance that’s not a very good thing.