Google v. Oracle, a years-long copyright infringement case in which Oracle accuses Google of reimplementing its application programming interfaces (APIs) in its free-to-use operating system Android, now heads to the Supreme Court.
Many on the political right, most recently Rachel Bovard writing in National Review, David Hogberg in The Federalist, and Seton Motley in RedState, righteously rage against Google “stealing” 11,500 lines of code from Oracle’s Java APIs. In reality, Oracle’s claims of infringement threaten innovation and competition by claiming a monopoly on something that cannot be stolen because it cannot be owned.
Niskanen Center has joined Public Knowledge and R Street Institute on an amicus brief in favor of Google, arguing that the use of the APIs shouldn’t be protected by copyright in the first place, but I would like to set aside the legal arguments and instead to take aim at the unearned moral high ground claimed by Bovard, Hogberg, and many other IP hawks that view copying as theft—something that just isn’t the case, despite the ease with which it is asserted.
Supporters of free markets and property rights must stop thinking of the unauthorized use of ideas and ideal objects (in this case a series of commands) as “theft.” Indeed, they must wholesale reject the notion that patents and copyrights are property and instead recognize them as the government grants of monopoly they are, as my colleague Brink Lindsey and I have argued. Google’s use of Oracle’s APIs may be found to constitute infringement, but Oracle has not been deprived of anything; their APIs work just as well as before. This mismatch between the nature of “stealing” an idea and stealing actual property requires us to take a broader view of why we have intellectual property in the first place.
There are two main arguments for the treatment of ideal objects, as embodied in patents and copyrights, as property. One extends the Lockean desert theory to the realm of ideal objects. The other is consequentialist, law and economics view of the issue, and sees intellectual property as necessary to protect the incentive to innovate in the face of potential free-riding.
I will begin with the latter view. In his seminal article, “Toward a Theory of Property Rights,” Harold Demsetz uses the (historically inaccurate but nonetheless useful) narrative of Native American tribes living in present-day Canada and their strategy to prevent over-hunting of the beaver population for their pelts. To prevent a depletion of the commons through overhunting, different lands were allocated to different groups, each with an exclusive claim to the beavers in that area. By dividing up a scarce resource, the different groups had an incentive to maintain the beaver population lest it be hunted to extinction.
Demsetz closes the paper by carrying over his logic to the case of intellectual property. “If a new idea is freely appropriable by all,” he writes, “incentives for developing such ideas will be lacking [and] benefits derivable from these ideas will not be concentrated on their originators.” He only briefly touches on the issue, but the extension of his argument for the utility of physical property rights has become the elevator pitch for why intellectual property should exist.
A basic comparison of the nature of real versus ideal objects will show the mismatch between patents and copyrights and physical property and the weakness of Demsetz’s consequentialist case for intellectual property. Ideas are nonrivalrous. Only one of us can use a given piece of physical property at a given piece of time; any number can use the same idea. Property rights in the physical world are necessary as a means of resolving claims to scarce resources, and in the process encouraging productive use through their husbandry or efficient consumption. Intellectual property rights artificially create scarcity and grant the owners the benefits of monopoly rights while passing monopoly costs onto the rest of society.
Patents and copyrights make sense when the costs of innovation are high and the costs of imitation are low. But such costs, and thus the strength of the Demsetzian argument, vary based on the industry and technologies in question.
Based on (somewhat dated) surveys of practitioners in various industries, pharmaceutical R&D would decline dramatically without the patent system, but other tools, such as lead time, secrecy, and offerings of complementary services or products are the weapon of choice for other industries looking to recoup the upfront costs of innovation and keep ahead of the competition.
Software is on the opposite end of the spectrum. Not only is the reimplementation of computer code common practice, but it is essential to innovation. Look at the popularity of open source programs, like Android, Linux, Firefox, and WordPress. The permissionless reimplementation of software benefits all parties involved, as Wired’s Klint Finley explains:
Why would Google give away [the Android operating system] something so central to its business? Because it hoped outside developers would make the software better as they adapted it to their own needs. And they have: Google says more than 1,300 outsiders have worked on TensorFlow. By making it open source, Google helped TensorFlow become one of the standard frameworks for developing AI applications, which could bolster its cloud-hosted AI services…
So Google stands to benefit, but why would an outsider contribute improvements to TensorFlow? Let’s say a company makes its own version of TensorFlow with unique elements, but keeps those elements private. Over time, as Google made its own changes to TensorFlow, it might become harder for that other company to integrate its changes with the official version; also, the second company would miss out on improvements contributed by others.
In short, open source provides a way for companies to collaborate on technology that’s mutually beneficial. [Emphasis added].
Interestingly, Oracle itself may be engaging in the practice at which critics of Google’s “theft” are wagging their fingers. In ArsTechnica, R Street Institute’s Charles Duan argues that Oracle’s use of code copied from APIs used in Amazon’s data storage platform, S3. To ensure compatibility, Oracle’s “Amazon S3 Compatibility API” reimplements many lines of code from the original API. Et tu, Oracle?
Like a Baptist in a liquor store, there’s an element of hypocrisy to Oracle engaging in the same practice that it’s squaring off with Google over in the highest court in the land. But Oracle, like Google and the Baptist with his Budweiser, is engaging in a common and beneficial practice in software development.
The beneficial effects of this ubiquitous free flow of code and ideas are lost on Google’s critics from the right, alas. Bovard writes that “no one wants to pour resources into the development of new code when it will ultimately just be copied by competitors.” This reimplementation (pun intended) of Demsetzian logic simply doesn’t apply in this case.
Despite being precisely wrong on nature of copying and innovation in the software industry, however, the IP maximalist right’s arguments could survive if, despite the economic consequences, they demonstrate that the use of an ideal object created by another is theft. The clear economic benefits of the private property system aside, I am sympathetic to arguments that it is acceptable (at least on the margins) to prioritize the wishes of property holders even in the face of sub-optimal economic outcomes.
Enter the Lockean argument for the treatment of ideas as property: why shouldn’t people be able to enjoy the fruits of their intellectual labor the same way they are able to enjoy the fruits of their physical labor?
Locke’s argument for property deals with how, if ever, one can take something from the commons, which is granted to everyone by God, and make it exclusively their own. Locke’s theory allows the drawing of clear lines around what can and can’t be appropriated via the mixing of labor, while extension of Locke’s logic to ideas has no natural stopping point.
As a matter of law, you can only copyright an expression, not an idea. But why not? If I came up with a good but not copyrightable idea, why shouldn’t I be entitled to a return from those who use it? Who’s to say that the creator of the buddy comedy, cargo shorts, or buffalo wings shouldn’t be able to claim exclusive rights to those creations? Why couldn’t Joseph Campbell demand George Lucas get his permission for the latter’s telling of the Hero’s Journey in space?
For that matter, why not have patent and copyright terms that last forever? Jack Valenti of the lobbying group formerly known as the Motion Picture Association of America wanted copyright terms to last “forever less one day.” Thankfully, the Constitution requires grants of intellectual monopoly be temporary, but if ideas were property, why not have them last forever? Land, widgets, and buildings can remain in private hands indefinitely, but claims to indefinite copyright or patent terms, though internally consistent, are laughable.
That protections for the producers of creative works should exist to prevent blatant copying has strong intuitive and practical appeal. The problem with treating such copying as theft is that the Lockean case for intellectual property proves too much.
Beyond this reductio ad absurdum, grants of intellectual monopoly also infringe on the property rights of others. I may justly acquire all of the components necessary to produce a mousetrap, but if I use those parts to produce a patented mousetrap, all of the sudden I have violated the patent owner’s property rights. The sum of my parts became someone else’s whole, despite no productive labor on their part. By design, monopoly rights to intellectual property allow the collection of above-normal market returns (called “rents” in economic jargon) that is not only disconnected from the upfront cost of innovation, but requires those who wish to use those ideas in novel or productive ways to hand over a cut of the loot.
Oracle’s claim to the exclusive rights over the reimplementation of their APIs represents a restriction on the rights of Google to use such code in smartphones. A drug patent represents a restriction on the rights of a generic manufacturer to produce the same pharmaceutical. The anti-circumvention provisions of the Digital Millennium Copyright Act prevent the owners of tractors, phones, and printers from fixing their own property. Copyright and patent holders suffer no loss, other than that of monopoly profits, when others are able to use their own property as they see fit. Such restrictions on property rights are far more direct and produce material harms, yet their loss is not bemoaned with the same fervor.
Lest there be any confusion, I am by no means anti-patent or anti-copyright. Rather, my position is that they do not deserve the same moral or rhetorical treatment as the right to physical property. These policies have their place in the innovation ecosystem and like any other regulations or subsidies, their restriction on other property rights may be justified by the gains in innovation they yield. My position is essentially the same as Milton Friedman’s when he says in Capitalism and Freedom, “[t]he specific conditions attached to patents and copyrights…are not a matter of principle. They are matters of expediency to be determined by practical considerations.”
In the case of Google v. Oracle, and the tech industry more broadly, good policy would be to deny copyright protection for the APIs Google reimplemented. Though a favorable ruling by the Supreme Court would benefit Google most directly, the free reimplementation of APIs would benefit newcomers to the industry and innovation in general, as the American Antitrust Institute’s amicus brief for Google argues.
Matters of expediency to be determined by practical considerations, as essential as they are in discussions of IP policy, do not excite the passions. The treatment of Google as a “thief” gives the IP hawks of the right a rhetorical cudgel that they have no business wielding. Ironically, the extensive use of “theft” and similar language was partly responsible for a reduction in Oracle’s damages against SAP SE in an unrelated copyright infringement case to the tune of $1 billion.
To flatly assert, as Bovard, Hogberg, and others who view copyrights as property do, that Google or any other infringers are thieves is wrong. Even those holdouts unconvinced by my case against the treatment of copyrights and patents as property must concede that there are strong arguments in favor of my position which must be seriously grappled with before terms like “theft” and “stealing” can be thrown around.
The use of the term property to describe what are, in fact, monopoly rights and claims on the property rights of everyone is a branding coup. It is also a serious limitation on the ability of supporters of the free market system and property rights to positively contribute to discussions on how much, if at all, government should intervene in free exchange and curtail property rights. The right’s talent for looking critically at regulations and special privileges granted to private actors has been hamstrung by this framing. Rather than unquestioningly embracing it, it should be closely scrutinized, and abandoned entirely.
In the interest of full disclosure, the Niskanen Center receives support from Google for its Captured Economy project. The project is an outgrowth of the book The Captured Economy, which was written by Brink Lindsey and Steven Teles prior to this support and contains a theoretical discussion on the nature of ideal objects similar to those presented here.