Software Patents Bad, Pharmaceutical Patents Good?: confused “libertarian” “opponents” of state monopoly privileges

by Stephan Kinsella on April 17, 2023

In Alex Tabarrok’s recent post End Software Patents! (see video below), he argues against software patents. He states:

In an industry like pharmaceuticals, patents make sense. It costs about a billion dollars to develop the average new drug. But a generic imitation might cost just 50 cents a pill. If innovators are not able to recoup their costs of research and development, there’ll be no one left to innovate. But does every innovative idea need a 20 year monopoly? A concept like one-click shopping does not have the same sunk costs of research and development as does a new drug. In industries like software, where innovation costs and imitation costs are more balanced, patent protection isn’t needed to incentivize innovation.

This is all well and good, but unfortunately he evinces no principled opposition to patents per se. This greatly weakens the case against software patents. True, this is a short video, and one might at first assume Tabarrok is just granting the validity of some patents, like pharmaceutical patents, for the sake of argument, so that he can focus on the “egregious” and obviously bad types of patents. And we should be in favor of incremental improvement, so long as it is unambiguously positive. A tax rate of 33% is better than 34%. A patent system covering a wide swath of technology but not software is better than one that covers software as well.

But Tabarrok has elsewhere made it clear he means what he says: some patents protection is really good … just not “too much.”¹ In this, he is like other moderate, utilitarian libertarians, who are never for patent and copyright abolition, only for patent “reform.” And the “reform” they favor is rarely significant or radical; always ginger and almost trivial. Even some Republicans seem to be for real IP reform,² but then they get fired.³ Real reform would significantly reduce patent terms, scope, and penalties. Not just arbitrarily carve out “software patents” from patentable subject matter.⁴ Back in the 1990s, the Patent Act was amended to protect doctors from being sued under patent law for performing patented “medical procedures.”⁵ What a way to “make” “law”!⁶ The proposals to carve out software from patent law are reminiscent of this ad hoc, unprincipled “patch” of patent law for the medical procedure field. Such ad hoc patches cannot address the real problems of the schizophrenic arguments behind the patent system.

The real reason to oppose software patents is that, like all patents, they infringe on property rights.⁷ That is why all patents, and indeed all copyrights, should be abolished. Not reformed. Utterly and completely abolished. Immediately. Patents and copyrights are a complete perversion of justice, property rights, and liberty. Software patents are not the problem, nor are patent trolls, low-quality patents, etc. Focusing on incidental details of the system in a futile quest for “reform” loses sight of the issues at stake. As I noted previously:

…patent trolls are not the real problem. … the problem is not software patents, or patent trolls (“non-practicing entities”), or even “bad” or junk or low-quality patents. The problem is not incompetent patent examiners. The problem is not ambiguous or arbitrary nonobviousness standards, or inadequate prior art databases or searching capabilities. The solution is not patent reform, or reducing the term of or even getting rid of “software patents.” The problem is the good patents—”legitimate,” hiqh-quality, unassailable patents granted not to patent trolls but to practicing entitieswho use these patents and the force of law to squelch competition. Even if you get rid of all “bad” patents, all software patents, all trolls, all business method patents–the fundamental problem remains: companies can acquire patents that they can use to squelch competition.⁸

But neither Tabarrok nor other free market “moderate” types, such as Cato’s Tim Lee and George Mason’s Jerry Brito,⁹ take this justice- and property-rights-based approach. Rather, Tabarrok apparently believes that interventions in the free market are justified if they are “needed to incentivize innovation.” But this is similar to the approach shared by those “free-market” economists who bristle at “unbridled competition”; sure, sure, we want some competition—but not too much!¹⁰

These people are not genuine free-marketeers. The libertarian and free market approach to determining how to allocate property rights is to use a Lockean (first-use first-own, plus contractual title transfer) property-allocation framework, in an attempt to do justice. It is not about tweaking “incentives” to “optimize” the “right” amount of some predetermined outcome (usually predetermined by the central planner or court intellectuals, like economists).

As Rothbard explained:

[T]he basic axiom of libertarian political theory holds that every man is a self-owner, having absolute jurisdiction over his own body. In effect, this means that no one else may justly invade, or aggress against, another’s person. It follows then that each person justly owns whatever previously unowned resources he appropriates or ‘mixes his labor with’. From these twin axioms—self-ownership and “homesteading“—stem the justification for the entire system of property rights titles in a free market society.¹¹

Rothbard is right: free markets and libertarian ideas rest on two basic principles: self-ownership and homesteading, plus related notions of contract (Rothbard has made significant contributions to the crucially important field of contract theory).¹²

By contrast, the idea of adjusting property rules to achieve varying goals of optimizing-judges or legislators—such as increasing or maximizing “social wealth”—is utterly contrary to justice and free markets. As Hoppe writes:

… an ethic must not only have permanency and stability with changing circumstances; an ethic must allow one to make a decision about “just or unjust” prior to one’s actions, and it must concern something under an actor’s control. Such is the case for the classic private property ethic with its first-use-first-own principle. According to this ethic, to act justly means that a person employs only justly acquired means—means originally appropriated, produced, or contractually acquired from a previous owner—and that he employs them so that no physical damage to others’ property results. Every person can determine ex ante whether or not this condition is met, and he has control over whether or not his actions physically damage the property of others. In distinct contrast, the wealth maximization ethic fails in both regards. No one can determine ex ante whether or not his actions will lead to social wealth maximization. If this can be determined at all, it can only be determined ex post. Nor does anyone have control over whether or not his actions maximize social wealth. Whether or not they do depends on others’ actions and evaluations. Again, who in his right mind would subject himself to the judgment of a court that did not let him know in advance how to act justly and how to avoid acting unjustly but that would judge ex post, after the facts?¹³

(In my view, consequentialism is roughly compatible with deontological and principled, “natural rights” views of libertarian rights,¹⁴ but utilitarianism not, if only because it is utterly incoherent, methodologically and ethically.)

But let’s say Tabarrok is right: that one reasonable way to determine what legal rules ought to be in place is to look at whether these rules produce “incentives” that are “necessary.” If there is “a determination” that the law produces such net benefits, then it should be passed by the legislature. If not, the law should not be passed.¹⁵

But what reason is there to think that a legislature would ever abide by the advice of tweakers like Tabarrok? The patent system has come into being and become “broken” and “corrupted” by special interests for a reason. Why would anyone ever think that such a system would ever achieve anything like ideal results? Even if an ideal patent system is desired, better to have none at all than to have the real one that real states will inevitably produce—ones that are beholden to special interests, that are used to squelch competition and form oligopolies, to impose hundreds of billions of dollars or more of costs on the economy every year.¹⁶

But let’s ignore even these problems. If one is to argue for the legitimacy of patents based on empirical effects—e.g. the claim that a certain type of patent is “needed to incentivize innovation” in a given field—then surely the burden of proof is on him making that claim. After all, this is Tabarrok’s implicit claim here in criticizing software patents: that there is no showing that they are needed. But to argue either way, based on these standards, one would need to have some facts at one’s disposal.¹⁷ One needs to produce the facts that buttress one’s case for patents. For example, to argue that “In an industry like pharmaceuticals, patents make sense” because patent protection is here “needed to incentivize innovation”—and that this is not the case for software patents—one would need to know (a) the value of the marginal innovation “incentivized” by the patent system, (b) the value of the innovation lost because of patent distortions, and (c) the cost the patent system imposes on the economy. And there is no guarantee that (a) is even positive! (I believe it is certainly negative.) As Tom Palmer writes:

[U]tilitarian arguments of a certain class can cut for or against intellectual property rights claims. As dealt with in much of the economics literature, for example, the utility gains from increased incentives for innovation must be weighed against the utility losses incurred from monopolization of innovations and their diminished diffusion. Some have argued that the first part of the comparison may be either negative or positive; patents or copyrights may actually decrease innovation, rather than increase it.¹⁸

Even if (a) minus (b) is positive, that is, we suppress some innovation but stimulate a lot more, this is still not a social benefit unless this net innovation gain is greater than the cost of the system. For example if the patent system induces an extra $100B worth of innovation in a given industry, and loses us $30B, for a net innovation gain of $70B, but the patent system imposes $200B cost on the economy in implementation, then the $30B net innovation gain is not worth the $200B hard cost. (Indeed, as I estimate, the patent system imposes an overall net cost on the US economy of at least $100B a year.)¹⁹

Now, which of these numbers do empiricist IP proponents really know? None of them. Yet they pretend to know their sums without knowing the components. This is illustrated in Tabarrok’s previous post Patent Policy on the Back of a Napkin, which I discuss in Tabarrok: Patent Policy on the Back of a Napkin. As I noted there, Tabarrok makes a Laffer-curve style “argument” that patent rights are currently “too strong.” Of course, he is correct that patent rights are too strong. But there are so many problems with his position. Here Tabarrok is implicitly adopting a wealth-maximization/utilitarian approach to policy. There are several problems with this, as I detailed in Against Intellectual Property. For one, it assumes that values are cardinal and interpersonally comparable, even though they are not. And it assumes that if a given policy helps A more than it hurts B, this is justified. But this is false; it does not follow that it is justified to violate B’s rights just because A feels some more intense joy greater than B’s pain. Also: there is no evidence anyway that patent policy produces any kind of net societal gains at all. This should be evident from Tabarrok’s ridiculous Laffer-curve of innovation vs. patent strength, which has no numbers associated with it; it’s pure academic whim.²⁰ Further, such an unprincipled, utilitarian approach leads to other unlibertarian ideas, such as Tabarrok’s proposal that the state take tens of billions of dollars from taxpayers and award it to people it deems “innovative”, to encourage innovation.²¹ As for numbers: the Founders adopted a quasi-empirical case for patents in 1789,²² but never backed it up with any empirical studies. It was just their “hunch” that allowing the new centralist federal government to grant monopoly privileges (copyright and patent) might “generate” more gain than the system costs.²³ But they had no proof of this. And guess what? In the 200+ years since, no one has come up with any. For example, Austrian economist Fritz Machlup, in a Congressional study in 1958, concluded:

No economist, on the basis of present knowledge, could possibly state with certainty that the patent system, as it now operates, confers a net benefit or a net loss upon society. The best he can do is to state assumptions and make guesses about the extent to which reality corresponds to these assumptions. … If we did not have a patent system, it would be irresponsible, on the basis of our present knowledge of its economic consequences, to recommend instituting one.”²⁴

And in a recent study, economists Michele Boldrin and David Levine, authors of Against Intellectual Monopoly, conclude:

The case against patents can be summarized briefly: there is no empirical evidence that they serve to increase innovation and productivity, unless the latter is identified with the number of patents awarded—which, as evidence shows, has no correlation with measured productivity.²⁵

Yet we have people like Tabarrok confidently concluding that there is a “Laffer curve” in the “balance” between patent “strength” and “innovation.” Technocrats like Tabarrok are talking utter nonsense, and they must know it. There is no “evidence” for their empirical case for some patent protection whatsoever, even if we magnanimously forget the huge and insurmountable ethical and methodological difficulties with their whole empirical, positivist, utilitarian, wealth-maximizationalist unprincipled approach. Tabarrok’s curve is flawed. The correct diagram is obvious: it should slope monotonically downward to the right: the more IP “strength,” the less economic freedom, competition, property rights, and innovation. Tabarrok’s tentative sally against maximalist patent rights (i.e.: state-granted anti-competitive monopoly privileges) is attacked by an Objectivist influenced economist, Geoffrey Manne, under the influence of the waning pro-IP approach of Objectivists like Adam Mossoff.²⁶ Bottom line: it is good that some people think we should reduce patent rights. But they are wrong to be half-hearted about it—to say that we should “reform” the law, but that some patent grants by the central state are necessary. Proponents of the free market who favor monopoly privilege grants by a bureaucratic agency of the criminal central state—and in the name of “property rights,” no less—are deeply confused about the nature of property rights and free markets.

***

Now, to return to the most recent post, Tabarrok says “In an industry like pharmaceuticals, patents make sense” because patent protection is here “needed to incentivize innovation”. How does he know this? He doesn’t know any of the numbers, and doesn’t even pretend to. The closest he gets is to his dashed-off figure “It costs about a billion dollars to develop the average new drug.” Even if this were true—so what? It doesn’t prove that we “need” patents to incentive innovation. It certainly doesn’t justify his outrageous claim that absent pharma patents “there’ll be no one left to innovate”. Really? Without patents there will literally be no pharma innovation? Tabarrok does not cite the work of Boldrin and Levine, who in ch. 9 of Against Intellectual Monopoly show that countries without patents in pharmaceuticals, like Switzerland, have historically been very innovative, and who show that the estimates of costs of developing a drug attributed to innovation are vastly overblown; see aslo Marcia Angell‘s new book The Truth About the Drug Companies, making similar arguments that many of the costs of bringing a new drug to market have to do with marketing, bribery, and so forth (discussed with Russ Roberts on Econ Talk in Angell on Big Pharma). As Sheldon Richman writes, quoting Boldrin and Levine:

Historically, intellectual monopoly in pharmaceuticals has varied enormously over time and space. The summary story: the modern pharmaceutical industry developed faster in those countries where patents were fewer and weaker… . [I]f patents were a necessary requirement for pharmaceutical innovation, as claimed by their supporters, the large historical and cross-country variations in the patent protection of medical products should have had a dramatic impact on national pharmaceutical industries. In particular, at least between 1850 and 1980, most drugs and medical products should have been invented and produced in the United States and the United Kingdom, and very little if anything produced in continental Europe. Further, countries such as Italy, Switzerland, and, to a lesser extent, Germany, should have been the poor, sick laggards of the pharmaceutical industry until recently. Instead, the opposite was true for longer than a century.

And any libertarian would suspect that whatever the costs of drug innovation are, are exacerbated by the federal government’s FDA. Surely the response to increases costs imposed by one governmental regulatory agency is not to create another one, but to abolish the first.

So here we have Tabarrok overestimating the cost of pharmaceutical development (at least, the part related to innovation and patents), ignoring the fact that much of these costs are imposed by the FDA, and claiming that without drug patents “there’ll be no one left to innovate” despite clear historical evidence to the contrary.

Tabarrok wants to adopt an empirical approach to policy, without having data to back up his claims. He seems to think that you can just make some verbal argument that reveals that it’s “obvious” that we “need” patents on drugs, but that it’s “not needed” on software. Without any numbers. It just seems obvious to him. It’s “too easy” to compete in the field of pharmaceuticals; while in software, there is more of a “balance” between “innovation costs” and “imitation costs.” (By the way, I’ve written a lot of software patents; why laymen like Tabarrok think it’s possible to legislatively permit patents on apparatuses, processes, compositions of matter, and to nonetheless carve out some arbitrary except for software, is beyond me; patent lawyers are very creative and can find ways to draft patent claims to cover processes that are implemented on computers, and so on, which some people call “software.”) And yet, in the field of pharma, the “scale” leans the other way—here, “innovation costs” are “very high” while “imitation costs” are “very low,” making the case an easy one. So: yes on pharma patents, no on software. Well, great—what about in between? Who gets to decide this? A great big panel of experts, no doubt.

Such are the perils of unprincipled, utilitarian thinking.

Or he could point out that the software industry would need 6 million patent attorneys (there are only about 40,000 now) at a cost of $2.7 trillion per year (US GDP is only $15T) to avoid infringing software patents.²⁷ But who knows—maybe it’s worth it. I mean, the software industry is about $200B a year, but maybe it’s worth it to spend $2.7T on it.

And where is the concern about copyright? After all, copyright enforcement causes censorship, chilling of free speech, jail time, and is being used as an excuse to ratchet up the police state and choke off Internet freedom; and it is costly too: Cost to Google to Pre-Screen YouTube Videos to Prevent Copyright: $37 Billion Per Year.

how would you get rid of software patents, anyway? Richard Stallman says . But what is a “general purpose computer”? Does that count an ipad? A smartphone (the state seems to distinguish smart phones from tablets and ipods for copyright-jailbreaking purposes, and has difficulty deciding whether an iPad is a computer or not, for import tax purposes). And as Mike Masnick notes:

People would certainly question why general purpose computing gets a pass. Also, the “generally used computing hardware” standard could be kind of hard to define as well. It still seems like there are more elegant solutions that focus on the real root of the problem, rather than trying to “carve out” certain impacts that we don’t like.²⁸

Arbitrarily removing software patents from patentable subject matter, even if someone could find a coherent way to amend the patent statute to do this, is not the way to improve the patent system. The only way to significantly “improve” the patent system, short of abolition, is to reduce the costs, the harms, that it imposes on society; and the only way to do this is to reduce patent strength, term, and scope: reduce the patent term; get rid of injunctions; provide an independent inventor defense; eliminate enhanced damages; add a working/reduction to practice requirement; provide that the losing patent plaintiff pays the defendant’s costs; exclude IP from trade negotiations.²⁹

***

Appendix:

some further notes not yet integrated into this post:

numerous empirical studies showing that software patents actually reduce innovation. But if we are going to rely on empirical studies to bolster the case against software patents, why stop there? After all, as

And why stop at patents? After all, empirical studies also demonstrate the harm copyrights impose as well.³⁰