Libertarians intuitively understand the case for patents: just as other property rights internalize the social benefits of improvements to land, automobile maintenance, or business investment, patents incentivize the creation of new inventions, which might otherwise be undersupplied.

So far, so good. But it is important to recognize that the laws that govern property, intellectual or otherwise, do not arise out of thin air. Rather, our political institutions, with all their virtues and foibles, determine the contours of property—the exact bundle of rights that property holders possess, their extent, and their limitations.

Outlining efficient property laws is not a trivial problem. The optimal contours of property are neither immutable nor knowable a priori. For example, in 1946, the U.S. Supreme Court reversed the age-old common law doctrine that extended real property rights to the heavens without limit. The advent of air travel made such extensive property rights no longer practicable—airlines would have had to cobble together a patchwork of easements, acre by acre, for every corridor through which they flew, and they would have opened themselves up to lawsuits every time their planes deviated from the expected path. The Court rightly abridged property rights in light of these empirical realities.

In defining the limits of patent rights, our political institutions have gotten an analogous question badly wrong. A single, politically captured circuit court with exclusive jurisdiction over patent appeals has consistently expanded the scope of patentable subject matter. This expansion has resulted in an explosion of both patents and patent litigation, with destructive consequences.

Patent Politics without Romance

James Buchanan coined the phrase “politics without romance” to define public choice, a research agenda that studies political outcomes in light of the incentives and constraints that politicians, bureaucrats, judges, and voters face. Contra elementary-school civics lessons (to say nothing of the academic literature), many public policies are enacted not in the general interest, but to benefit groups that successfully manipulate the political system. By concentrating benefits and dispersing costs, the few are able to take from the many.

In patent politics, the romance has been gone for at least three decades. Here’s why. In most areas of law, the loser in a federal trial appeals to the circuit court corresponding to the federal judicial district in which the trial was held. But in 1982, at the urging of the patent bar, Congress consolidated appellate review of all patent cases in a newly created Court of Appeals for the Federal Circuit.

The patent bar had sought this consolidation for some time. As early as 1951, Simon Rifkind, a former federal judge in New York City, warned in a prescient essay that then-current proposals to create a specialized patent court would lead to “decadence and decay.”

The patent Bar is already specialized. At present, however, patent lawyers practice before nonspecialized judges and accommodate themselves to the necessity of conveying the purposes of their calling to laymen. Once you complete the circle of specialization by having a specialized court as well as a specialized Bar, then you have set aside a body of wisdom that is the exclusive possession of a very small group of men who take their purposes for granted.

The consolidation of patent appeals into a single court also concentrated the benefits that pro-patent interests could get by lobbying to affect the judicial nomination and confirmation process. They now could focus their scrutiny on just 12 seats, rather than monitoring appointments across the entire federal appellate judiciary.

The creation of the court has significantly altered the law. Using a dataset of district and appellate patent decisions for the years 1953–2002, economists Matthew Henry and John Turner find that the Federal Circuit has been significantly more permissive with respect to affirming the validity of patents. They estimate that patentees are three times more likely to win on appeal after a district court ruling of invalidity in the post-1982 era. In addition, following the precedents set by the Federal Circuit, district courts have been 50 percent less likely to find a patent invalid in the first place, and patentees have become 25 percent more likely to appeal a decision of invalidity.

With patents more likely to be upheld in the Federal Circuit era, the incentive to patent has increased. Bronwyn Hall finds a highly significant structural break in patent applications occurring between 1983 and 1984. The number of patents granted by the U.S. Patent and Trademark Office also increased, from 63,005 in 1982 to 275,966 in 2012—a quadrupling of the rate in only 30 years.

Figure 1. Patents Issued by Year

 

Source: U.S. Patent and Trademark Office

The change in the law wrought by the Federal Circuit can also be viewed substantively through the controversy over software patents. Throughout the 1960s, the USPTO refused to award patents for software innovations. However, several of the USPTO’s decisions were overruled by the patent-friendly U.S. Court of Customs and Patent Appeals, which ordered that software patents be granted. In Gottschalk v. Benson (1972) and Parker v. Flook (1978), the U.S. Supreme Court reversed the Court of Customs and Patent Appeals, holding that mathematical algorithms (and therefore software) were not patentable subject matter. In 1981, in Diamond v. Diehr, the Supreme Court upheld a software patent on the grounds that the patent in question involved a physical process—the patent was issued for software used in the molding of rubber. While affirming their prior ruling that mathematical formulas are not patentable in the abstract, the Court held that an otherwise patentable invention did not become unpatentable simply because it utilized a computer.

In the hands of the newly established Federal Circuit, however, this small scope for software patents in precedent was sufficient to open the floodgates. In a series of decisions culminating in State Street Bank v. Signature Financial Group (1998), the Federal Circuit broadened the criteria for patentability of software and business methods substantially, allowing protection as long as the innovation “produces a useful, concrete and tangible result.” That broadened criteria led to an explosion of low-quality software patents, from Amazon’s 1-Click checkout system to Twitter’s pull-to-refresh feature on smartphones. The GAO estimates that more than half of all patents granted in recent years are software-related. Meanwhile, the Supreme Court continues to hold, as in Parker v. Flook, that computer software algorithms are not patentable, and has begun to push back against the Federal Circuit. In Bilski v. Kappos (2010), the Supreme Court once again held that abstract ideas are not patentable, and in Alice v. CLS (2014), it ruled that simply applying an abstract idea on a computer does not suffice to make the idea patent-eligible. It still is not clear what portion of existing software patents Alice invalidates, but it could be a significant one.

Observers on all sides widely recognize that the Federal Circuit routinely undermines Supreme Court precedent. Patent attorney Gene Quinn reacted with dismay to the Supreme Court’s ruling in Mayo v. Prometheus (2012), but consoled himself with the thought that the ruling would not stand forever.

How long will it take the Federal Circuit to overrule this inexplicable nonsense? The novice reader may find that question to be ignorant, since the Supreme Court is the highest court of the United States. Those well acquainted with the industry know that the Supreme Court is not the final word on patentability, and while the claims at issue in this particular case are unfortunately lost, the Federal Circuit will work to moderate (and eventually overturn) this embarrassing display by the Supreme Court.

Supreme Court justices also recognize the Federal Circuit’s insubordination. In oral arguments in Carlsbad Technology v. HIF Bio (2009), Chief Justice John Roberts joked openly about it:

Mr. Rhodes: I can’t suggest what the Court might finally decide other than to say that—that, again, the circuit courts of appeal have uniformly applied this. They seem to be—

Chief Justice Roberts: Well, they don’t have a choice, right? They can’t say, I don’t like the Supreme Court rule so I’m not going to apply it, other than the Federal Circuit.

[Laughter]

The economic consequences of the Federal Circuit’s misbehavior, however, are not so funny.

The Opportunity of the Commons

As a result of the Federal Circuit’s pro-patent jurisprudence, our economy has been flooded with patents that would otherwise not have been granted. If more patents meant more innovation, then we would now be witnessing a spectacular economic boom. Instead, we have been living through what Tyler Cowen has called a Great Stagnation. The fact that patents have increased while growth has not is known in the literature as the “patent puzzle.” As Michele Boldrin and David Levine put it, “there is no empirical evidence that [patents] serve to increase innovation and productivity, unless productivity is identified with the number of patents awarded—which, as evidence shows, has no correlation with measured productivity.”

While more patents have not resulted in faster economic growth, they have resulted in more patent lawsuits. Software patents have characteristics that make them particularly susceptible to litigation. Unlike, say, chemical patents, software patents are plagued by a problem of description. How does one describe a software innovation in such a way that anyone searching for it will easily find it? As Christina Mulligan and Tim Lee demonstrate, chemical formulas are indexable, meaning that as the number of chemical patents grow, it will still be easy to determine if a molecule has been patented. Since software innovations are not indexable, they estimate that “patent clearance by all firms would require many times more hours of legal research than all patent lawyers in the United States can bill in a year. The result has been an explosion of patent litigation.” Software and business method patents, estimate James Bessen and Michael Meurer, are 2 and 7 times more likely to be litigated than other patents, respectively (4 and 13 times more likely than chemical patents).

Software patents make excellent material for predatory litigation brought by what are often called “patent trolls.” Trolls use asymmetries in the rules of litigation to legally extort millions of dollars from innocent parties. For example, one patent troll, Innovatio IP Ventures, LLP, acquired patents that implicated Wi-Fi. In 2011, it started sending demand letters to coffee shops and hotels that offered wireless Internet access, offering to settle for $2,500 per location. This amount was far in excess of the 9.56 cents per device that Innovatio was entitled to under the “Fair, Reasonable, and Non-Discriminatory” licensing promises attached to their portfolio, but it was also much less than the cost of trial, and therefore it was rational for firms to pay. Cisco stepped in and spent $13 million in legal fees on the case, and settled on behalf of their customers for 3.2 cents per device. Other manufacturers had already licensed Innovatio’s portfolio, but that didn’t stop their customers from being targeted by demand letters.

Litigation cost asymmetries are magnified by the fact that most patent trolls are nonpracticing entities. This means that when patent infringement trials get to the discovery phase, they will cost the troll very little—a firm that does not operate a business has very few records to produce. But discovery can cost a medium or large company millions of dollars. Using an event study methodology, James Bessen and coauthors find that infringement lawsuits by nonpracticing entities cost publicly traded companies $83 billion per year in stock market capitalization, while plaintiffs gain less than 10 percent of that amount.

Software patents also reduce innovation in virtue of their cumulative nature and the fact that many of them are frequently inputs into a single product. Law professor Michael Heller coined the phrase “tragedy of the anticommons” to refer to a situation that mirrors the well-understood “tragedy of the commons.” Whereas in a commons, multiple parties have the right to use a resource but not to exclude others, in an anticommons, multiple parties have the right to exclude others, and no one is therefore able to make effective use of the resource. The tragedy of the commons results in overuse of the resource; the tragedy of the anticommons results in underuse.

Consider a new product that must license ten patents in order to come to market. If each patent holder demands a third of the product’s revenue, the product won’t be viable. The holdout problem and other transaction costs associated with bargaining to an efficient solution can be prohibitive. Now consider that products like smartphones build on many thousands of licensed patents. Although Apple and Samsung have been able to navigate these licensing waters well enough to bring profitable products to market, the patent licensing and litigation thicket creates economies of scale in the legal department that make smaller firms less competitive.

In order to cope with the tragedy of the anticommons, we should carefully investigate the opportunity of  the commons. The late Nobelist Elinor Ostrom made a career of studying how communities manage shared resources without property rights. With appropriate self-governance institutions, Ostrom found again and again that a commons does not inevitably lead to tragedy—indeed, open access to shared resources can provide collective benefits that are not available under other forms of property management.

There is evidence that open access leads to greater innovation in science. Fiona Murray and coauthors exploited a natural experiment in research involving patented mice, in consequence of an unanticipated shock in which some varieties of engineered mice, but not others, suddenly became available on an open access basis. They found that greater openness encouraged the exploration of more diverse research paths, including those followed by entrants new to the mouse research arena. Importantly, they found no decline in the creation of new varieties of mice. This suggests that—litigation costs aside—patent law could be reducing the stock of ideas rather than expanding it at current margins.

Advocates of extensive patent protection frequently treat the commons as a kind of wasteland. But considering the problems in our patent system, it is worth looking again at the role of well-tailored limits to property rights in some contexts. Just as we all benefit from real property rights that no longer extend to the highest heavens, we would also benefit if the scope of patent protection were more narrowly drawn.

Reforming the Patent System

This analysis raises some obvious possibilities for reforming the patent system. Diane Wood, Chief Judge of the 7th Circuit, has proposed ending the Federal Circuit’s exclusive jurisdiction over patent appeals—instead, the Federal Circuit could share jurisdiction with the other circuit courts. While this is a constructive suggestion, it still leaves the door open to the Federal Circuit playing “a leading role in shaping patent law,” which is the reason for its capture by patent interests. It would be better instead simply to abolish the Federal Circuit and return to the pre-1982 system, in which patents received no special treatment in appeals. This leaves open the possibility of circuit splits, which the creation of the Federal Circuit was designed to mitigate, but there are worse problems than circuit splits, and we now have them.

Another helpful reform would be for Congress to limit the scope of patentable subject matter via statute. New Zealand has done just that, declaring that software is “not an invention” to get around WTO obligations to respect intellectual property. Congress should do the same with respect to both software and business methods.

Finally, even if the above reforms were adopted, there would still be a need to address the asymmetries in patent litigation that result in predatory “troll” lawsuits. While the holding in Alice v. CLS arguably makes a wide swath of patents invalid, those patents could still be used in troll lawsuits because a ruling of invalidity for each individual patent might not occur until late in a trial. Current legislation in Congress addresses this class of problem by mandating disclosures, shifting fees in the case of spurious lawsuits, and enabling a review of the patent’s validity before a trial commences.

What matters for prosperity is not just property rights in the abstract, but good property-defining institutions. Without reform, our patent system will continue to favor special interests and forestall economic growth.

 

 

Some parts of this essay are drawn from my Mercatus working paper (in press at Public Choice), which is coauthored with Alex Tabarrok, whom I wish to thank, credit, and saddle with blame for any errors I have introduced.

Patent grants have always been at the center of controversy, and the U.S. patent system is no exception. Today many critics of the patent system employ rhetoric reminiscent of Yeats’ poem about the first World War, implying that “things fall apart; the centre cannot hold;” and some even call for “mere anarchy” or the abolition of patents.[1] Eli Dourado’s lead essay similarly proposes significant reforms based on the specific hypothesis that the Court of Appeals for the Federal Circuit (CAFC) contributes to pervasive problems. It is impossible to directly prove or disprove that, while we have not attained Bethlehem, alternative routes would get us there faster. But economic historians are not deterred by the merely impossible, and I will draw on long-term patterns to suggest that the current experience is not anomalous. Patent and legal institutions are not perfect, but the news of a crisis and the need for a Second Coming have been much exaggerated.

 

The True Story

The lead essay supposes that the current environment is overly favorable to, and legal institutions have been captured by, patent interests. Both theory and evidence would lead us to exercise caution before accepting such a claim. We cannot objectively gauge the attitudes of courts by parsing their legal decisions, for many reasons.[2] Further, “pro-patent” decisions that uphold patent rights are distinctly different from improper bias toward “patent interests.” Joseph Story, the most influential Supreme Court Justice in the realm of intellectual property law, stated that we need to remind ourselves that it is the U.S. Constitution itself that bestows secure property rights on patents.[3] The legal default is thus a presumption of strong protections for patented inventions, and the burden of proof is on anyone seeking to circumscribe or overturn those rights.

Dourado points an accusatory finger at the Federal Circuit, whose “misbehavior” in many dimensions has created significant deleterious outcomes. We are encouraged to infer from Figure 1 in his essay (which does not adjust for population or other legal and economic changes) that the introduction of a specialized patent court has caused the observed significant increase in patents granted in the past few decades. We can gain valuable insights from the historical experience and from disaggregation of these basic data, as in Figure 1 below. From a longer-term perspective, the patterns of patenting per capita for domestic residents today is not substantively different from the cyclical grants of the past.[4] The most dramatic change relates to the filing for patents in the United States by foreigners, which now exceeds grants to domestic residents. As such, clearly a substantive part of the explanation for post-1980 patterns lies well beyond the borders of the United States.[5]

 

Figure 1: Domestic and Total Patents Per Capita Filed in the United States, 1840-2013

 

Source: U.S. Patent and Trademark Office; and U.S. Bureau of the Census

 

Tesla the Magician, Trolls, and Other Fables

The Federal Circuit, it is argued, has also misbehaved by expanding the subject matter of patents and allowing protection for software. I am not certain why the United States, one of the most creative countries in the world in the realm of information technology, should choose to follow the leadership of New Zealand (?!) in revoking property rights for software inventions. The legal boundaries of patent grants in disruptive technologies have always been difficult to delineate, and in the nineteenth century Joseph Story characterized them as “the metaphysics of the law.”[6] In that period, just as today, the subject-matter of patents filed necessarily had to change to incorporate economically important new discoveries, such as the application of the principles of electricity to innovations that dramatically increased productivity and social welfare – and litigation. To nineteenth-century observers, electrical innovations appeared magical, and patents on them were even more abstract and problematic than software might seem to us today. The opinion in Edison Electric Co. v. Westinghouse noted that “it is exceedingly hazardous for one not an expert to express an opinion on a question so wholly within the domain of scientific exposition… . It is difficult—impossible, perhaps—to describe what invention is.”[7] Congress, the courts, and the Patent Office nevertheless resisted the temptation to prohibit patents for electrical inventions.

Figure 2: Patent Litigation Rate (lawsuits relative to patents granted), 1990-2012

Source: U.S. Patent Office and Annual Report of the Director: Judicial Business of the United States Courts.

The lead essay suggests that the Federal Circuit is to be held responsible for an “explosion” in patent litigation.[8] Figure 2 illustrates the actual rate of litigation (relative to patent grants), which hardly seems to warrant the concern it has attracted, especially given the recent changes in legal rules that have contributed to an increase in the number of separate lawsuits filed. Moreover, the rate of litigation for copyright has risen faster than that of patents. This indicates a general increase in intellectual property concerns, rather than a patent-specific rationale. In the first century of the patent system rates of litigation were at times significantly higher than today, and growth was also more rapid in the transformative era that preceded the Second Industrial Revolution. The point here is that, if litigation at the present time is similar to the long-term norm, the primary causes are likely to lie beyond recent issues and institutions.

Information technology has transformed the world around us, and such markets are expanding, so it is not surprising if software and related innovations now account for a larger fraction of patents filed and litigated. Patent litigation is but one facet of the changes created by all disruptive technologies.  In the past, every major new discovery has led to a rapid increase in litigation, not just in the realm of patent law, but in contracts, torts, crime, and indeed all areas of the legal system. As Justice Cardozo noted, “the great inventions that embodied the power of steam and electricity, the railroad and the steamship, the telegraph and the telephone, have built up new customs and new law.”[9] Disputes about smartphones and software will similarly in the future give way to litigation about the Next New Thing.

The author calls for reforms to mitigate the negative consequences of “predatory litigation” by “patent trolls,” who “legally extort” large sums from “innocent parties.”  In less colorful terms, patents comprise property rights in invention that offer a legal right of exclusion and appropriation enforceable by the courts. Economic efficiency implies that property rights should be based on the nature of the patented invention, and not on the identity of the patentee. Proposals that policymakers should distinguish between patent trolls (however defined) and other patentholders essentially advocate that patent policy should be driven by moral judgments about who is deemed to be more deserving of protection. Even if “non-trolls” (hobbits?) were favored, other patent holders could obviously circumvent the restrictions, but at greater costs to the economic system. In short, limitations on the rights of any patent holder because of such arbitrary factors are incompatible with both market efficiency and the fundamental principles of the American patent system.  

The patent controversy in the United States today is not significantly different from previous eras. It is worth noting that, in the past, populist clamor for congressional action to deal with “patent wars” and perceived excessive patent litigation was repeatedly ignored by legislators who wisely realized that the costs of such reforms would likely outweigh their alleged benefits. A longer-term perspective leads one to be much more cautious about lobbying for drastic reforms in a system that has functioned effectively for over two centuries.

 

            From its founding in 1847 and throughout the remainder of nineteenth century, the American Medical Association was convinced of one thing about patents: patented medicines were not only bad, they were unethical. The AMA’s initial code of ethics—which again did not change on this point throughout the nineteenth century—held that it was “derogatory to professional character” for any “physician to hold a patent for any surgical instrument, or medicine” and that it was “reprehensible for physicians to give certificates attesting the efficacy of patent or secret medicines, or in any way to promote the use of them.”[1] Such was the AMA’s hostility to patented medicines that in 1849 Dr. Thomas Edwards, an AMA member and Congressman, headed up a congressional committee that recommended as “an important measure of reform” legislation banning all patents on medicines.[2]

            Today, of course, the AMA has changed its position, and indeed it would be hard to imagine a modern medical practice in which a doctor abstained from promoting “in any way” the use of patented pharmaceuticals. My point here is quite straightforward: Not all proposed “reforms” of the patent system are good ones. In particular, readers should be skeptical—I think extremely skeptical—of calls to reform patent law by banning patents altogether or in a particular field. Such brute-force “reforms” have been proposed in the past, and history has shown that approach to be profoundly shortsighted.

Eli Dourado’s essay reviews some of the recent criticisms of the patent system and concludes with a call for “good property-defining institutions.” I’m with him on that. Who, after all, wants bad property-defining institutions? Indeed, with respect to the Federal Circuit, my co-author Craig Nard and I were two of the earliest critics of complete centralization of appeals in the Federal Circuit.[3] Last year Chief Judge Diane Wood joined us in that view (with a gracious acknowledgement of our work), and now in this essay Eli Dourado endorses the view too. I welcome the company.  

If, however, we want good institutions for the patent system (rather than merely embracing the simplistic approach of abolishing patents altogether), then details of reform are going to matter. And it’s in the details that Dourado and I disagree.

The most dramatic “reform” proposed in Dourado’s essay is that Congress should abolish by statute all patents on “both software and business methods.” That purported reform is precisely the sort of targeted patent abolition championed by the AMA and many doctors in the nineteenth century. The similarity does not end there.

Dourado’s essay spends quite a bit of time arguing, correctly, that open access to scientific and technical information can lead to significant innovation. That’s an old insight, and one that the patent system fully accepts. After a relatively short interval, all patented inventions fall into the public domain, where they are freely available to everyone as the building blocks of new innovations. Moreover, even during the duration of the patent life, the information contained in patents is freely available to all, and that information can, and indeed sometimes does, spur insights and innovation in completely different fields.

The crucial question is not whether the public domain can support innovation (it certainly can), but whether a system with no property rights and only the public domain produces the optimal level of innovation. On that point, the evidence supporting patent abolition—either generally or in a particular sector—remains utterly wanting.

Dourado’s essay cites a “natural experiment” supposedly providing evidence that “open access leads to greater innovation in science,” but that natural experiment emphatically did not involve eliminating patent incentives. Rather, the natural experiment involved providing researchers with more access to so-called “upstream” research products—in that case genetically engineered mice—and thus allowing those researchers to engage in more “downstream” research (even better genetically engineered mice). But, and this is the crucial point, those downstream researchers were not in any way prohibited from patenting the results of their research and, as the paper cited by Dourado makes perfectly clear, those downstream researchers were motivated by the potential for licensing revenue.

Here I will go into some of the all-important details. The paper cited by Dourado notes that one aspect of the “open access” agreement was that DuPont, the upstream or pioneering innovator and patent holder, eliminated its demand for what are called “reach through” rights, which would have given DuPont a share in the “licensing revenue from any commercial applications developed” by the downstream or follow-on researchers. Such reach-through rights reduce “the incentives to pursue more applied research stages” because the downstream researchers would have had to share their licensing revenue. Eliminating reach-through rights allows the downstream innovators to keep all the licensing revenue for themselves and, not surprisingly, the increased incentives yielded more downstream research. While this natural experiment may give insights into the optimal balance between pioneering and follow-on patent rights, it has (to put it mildly) limited value for showing how researchers would react to a hypothetical patent-free zone where there is no licensing revenue for anyone because neither upstream nor downstream research is patentable.

Again, I do not want to overstate my points. Invention without any promise of licensing revenue can certainly occur, but this is not news. Even in the nineteenth century, the opponents of patented medicines had examples in which doctors had struggled to invent without seeking patents. Thomas Edwards’s 1849 congressional report cited the “thirty years” of laborious work by Edward Jenner, who ultimately developed but did not patent a smallpox vaccine. (Of course, better historical work might have cast doubt on trumpeting the smallpox vaccine as the model of innovation without property rights, for Jenner merely rediscovered a vaccination technique that had been in the public domain for decades but that was neither commercially exploited nor appreciated.)

It is also not news that the patent system is imperfect. The critics of nineteenth century medicinal patents had some valid complaints. The patent system of the era was then (and is now) imperfect; it needed (and still needs) improvement. But the right reform then was not (and now is not) giving up and embracing a property-rights free zone in whole fields of endeavor.

If the real goal is “good property-defining” institutions—a goal I think worth pursuing—then I think policymakers should eschew blunderbuss solutions that call for the elimination of whole swaths of patents. Instead, I have long supported better enforcement of patent law’s key standard for patentability, which is supposed to prevent the patenting of “obvious” inventions. That doctrine has evolved over a century and half of steady and incremental progress in improved understanding about the appropriate limits of patentability.[4] (Indeed, this crucial doctrine didn’t even exist in 1849, which is one of the reasons that the AMA of the era had some justified complaints against the then-existing patent system.) In the twentieth century, the Supreme Court correctly identified the goal of the obviousness doctrine as being to restrict the award of patents only to “those inventions which would not be disclosed or devised but for the inducement of a patent,” but, as Michael Abramowicz and I have shown, current legal doctrine being applied at the Patent Office and the courts could (and should) do a much better job of reaching that goal.[5] 

Worrying about matters such as patent law’s obviousness doctrine has none of the flamboyance of a call to eliminate all software patents, but the targets of reform can be the same. Eli Dourado rightly criticizes Amazon’s 1-click patent as an example of a “low-quality software patent.” So have I—repeatedly.[6] But the problem with the 1-click patent is not that it’s on software but that it’s “low-quality”—it’s trivial; it’s obvious. The right solution to such “low-quality software patents” is not to get rid of all software patents (low-quality and high-quality), but to do a better job of getting rid of all low-quality patents (software or not). That’s not a flashy “let’s-kill-all-the-software-patents” approach. It’s boring; it’s incremental. But if the goal really is to get “good property-defining institutions” rather than simply to abolish property rights across fields of innovation, such incremental, boring improvements are what’s needed.

 

Eli Dourado argues that software patents tend to have a negative effect on innovation (as James Bessen and Michael Meurer have argued previously in Cato’s Regulation)[1] and should be eliminated. John F. Duffy agrees that many software patents should not have been granted, citing Amazon’s 1-click patent an example. But he emphasizes that the reason 1-click is a low quality patent is “not that it’s on software but [because] . . . it’s trivial; it’s obvious.” Rather than eliminating types of patents, like software and business methods, Duffy argues that courts and the PTO should focus on correctly applying the requirement that patented inventions be nonobvious.

One way to understand Dourado’s and Duffy’s differing solutions is to see their discussion as revisiting the longstanding debate over whether “rules” or “standards” make for better law.

Generally, this debate involves asking whether in particular circumstances it is better to create rules (“No driving over fifty miles per hour”) or standards (“Drive at a safe speed”) to govern some action.

Rules are more challenging to articulate well in advance, and usually they are over- or under-inclusive of what behavior the rule writer wants to reach. For example, the factors that affect safe driving speed are too numerous to take into account in a finite rule — weather, road conditions, temperature, and traffic patterns are just a few variables that interact with each other in complicated ways. So the strict application of any rule concerning driving speed will either exclude some safe driving conditions, include some unsafe ones, or both. Rules are nonetheless attractive because they are, compared to standards, easier to apply ex post and to anticipate ex ante.

In contrast, standards hold the promise, at least in theory, of always reaching the right result. If drivers must only pay fines for driving at an “unsafe” speed, then no one driving quickly down an empty road on a clear day will get penalized, and no one would get away with driving forty-five miles per hour, weaving in and out of traffic, in pouring, icy rain. But the downside of standards is that they are comparatively costly to apply, and those costs multiply when several parties must apply them. A police officer must assess what a safe speed is; if the driver disagrees he must go to a judge to argue and provide evidence about the weather, traffic conditions, road conditions, and so forth, in order for the judge to have the information necessary to apply the standard correctly.

Which are better: rules are standards? The answer changes based on the situation — how significant are the harms of an over- or underinclusive rule? Which way does one want to err in a given case? How high are the costs of applying the standard compared to the benefits of getting the ideal result?

Dourado’s solution involves the creation and application of a rule: no patents on software. While a rigorous definition of “software patents” would have to be chosen and articulated, in theory the application of the rule could be fairly straightforward. Duffy’s preference for eliminating obvious patents instead is a preference for a standard — many elements and documents have to be considered when evaluating each patent application to decide whether or not it is an obvious invention.

Framing this discussion as a choice between rules and standards helps explain why Dourado’s solution may not just be viable, but preferable. Some software inventions are nonobvious, easy to copy, and costly to develop — in other words, they are good candidates for deserving a patent. But perhaps software patents, as a whole, hurt innovation and innovative companies more than they help, by fostering trolling and inviting the patenting of many obvious programs. A clear, but admittedly imperfect, rule — no software patents — could be preferable over trying to apply, case by case, the standards for patentability, particularly because of the incentives and institutional structure of the Federal Circuit and Patent & Trademark Office.

In other words, the benefits of “good” software and business method patents may be outweighed by the costs associated with trying to weed out the “bad” ones: costs such as challenging low-quality patents’ validity in an adjudicative setting and choosing how to respond to demand letters sent by non-practicing entities with a portfolio of low-quality patents.[2] The law sometimes makes rough cuts when the cost of precision is too high, and perhaps software patents are an area where such a rough rule is optimal — not because all software patents are necessarily bad, but because software patents are so strongly correlated with other activities or qualities that we want to eliminate from the patent system.

Whatever the ideal solution is, the perfect need not be the enemy of the good. A variety of solutions would help improve the patent system, because the system has many problems which interact with and compound each other. Many software patents are merely mathematical formulas or abstract ideas and should not be considered patentable subject matter because they remove too much “raw” material from the public domain.[3] Other patents are granted because the standards for “nonobviousness” are, in practice, too forgiving. The strict liability of our patent system allows parties who contribute nothing to the “progress of science and the useful arts” to extract payments from productive companies and inventors who happen to have independently created something already patented.[4] And finally, the patent bar has a vested interest in a strong patent system, and its influence puts pressure on that system to evolve in a way that may not be in the public’s good. Each of these phenomena interacts with the others.

As a result, the patent troll who has acquired an obvious software patent that was never practiced or licensed, and who exacts unjustified settlements from small companies who may not have even infringed, is a phenomenon that exists and proliferates because of all of these imperfections in the system working in tandem. And so, it is not surprising that all of these problems would be mitigated by a proposal designed to reach just one.

Which brings us back to Dourado’s initial observation: the optimal contours of property are neither immutable nor knowable a priori, and so we necessarily need good property-defining institutions to choose what they are. The notion that patents are or aren’t property doesn’t help us decide the contours of patent rights — how long patents should last, what exactly can be patented, how and when patents can be enforced, and what institutions are best situated to decide the answers to these questions in practice. And these questions are difficult. Even Ayn Rand sidestepped suggesting a length for intellectual property terms, stating that if intellectual property “were held in perpetuity . . . it would lead, not to the earned reward of achievement, but to the unearned support of parasitism.”[5]

Perhaps even less intuitively, the question of whether patents are property also does not tell us whether the contours of property must be roughly or precisely defined. In fact, the contours of real property rights usually are rough — as Henry Smith observes, “The architecture of property emerges from the process of solving the problem of how to serve use interests in a roughly cost-effective way.”[6]

The patent system would be greatly improved if the PTO and courts could draw the line between obvious and non-obvious inventions correctly and at low enough cost to applicants, litigants, and society as a whole. But this is challenging precisely because of public choice theory. Applying standards — where the correctness of individual assessments of particular patent applications and suits is what constitutes the law working well — is harder for the public to see and criticize, and provides more of an opportunity for smaller groups to influence the patent system at the expense of the public interest. Eliminating the institutional structures that favor these groups is the next step to improving both the patent system and the innovation economy.

 

What a delight it has been to receive comments on my lead essay from such excellent scholars! I am grateful to all three respondents for their thoughtful remarks.

The debate seems to have split into two streams—Zorina Khan disputes the premises of my argument for reform, while John Duffy and Christina Mulligan have discussed instead whether one of my proposed reforms is appropriate. I will therefore address these points separately and in turn.

Is the Explosion of Patents a Myth?

 Zorina Khan begins her essay with a claim to which I object: She says that the U.S. Constitution “bestows secure property rights on patents.” I don’t believe this is a fair reading of the plain text of the Constitution, which states that “The Congress shall have Power To… promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” Congress has the power, therefore, but not the obligation, to set up a patent system. It is well within Congress’s constitutional authority to abolish the patent system altogether, or (as I have proposed instead) to limit the patent system’s scope to non-software inventions. From a legal perspective, there is not, as Khan claims, a special burden of proof that falls on those who seek a less expansive patent system.

What is the best way to evaluate whether there has been an explosion in patenting activity after the creation of the Federal Circuit? In my essay, I cited econometric studies that found structural breaks in court rulings and patent applications consistent with such an effect. In addition, I showed visually the raw increase in the number of patents granted by the USPTO over time, which displays a startling increase shortly after the Federal Circuit was created. Khan takes raw patent filing data, makes two adjustments—controlling for population growth, and separating out foreign sources of patents—and shows that per-capita American-originated patent filings have increased much less dramatically (although they are now at an all-time high).

Do these data adjustments make sense? If what matters is the number of restrictions imposed on individuals and businesses bringing products to market, they do not. If there are now four times as many patents being granted annually, that’s four times as many ways businesses can get dragged into a lawsuit. It certainly doesn’t matter whether the holder of a patent is American or not; American businesses can be sued by foreign plaintiffs in American courts. And even if we did want to adjust for population, since 1982, U.S. population has increased by about 37 percent, whereas the number of patents granted has more than quadrupled—adjusting for population doesn’t seriously affect my argument.

Neither population growth nor increases in foreign patenting can explain why Bronwyn Hall found a statistically significant structural break in patent applications in the early 1980s. Population grows pretty smoothly, and the increase in foreign patent applications was underway before the Federal Circuit was created. In addition, Khan’s data adjustments say nothing about the statistically significant effects on patent rulings that were discovered by Matthew Henry and John Turner. Wouldn’t it be odd, after the law became more friendly to patent interests, for patent applications not to increase?

Khan also casts doubt on whether there has been an explosion in patent litigation. The number of lawsuits per patent granted, she argues, has been roughly constant (although, again, they are at a local maximum). This data underscores my point, however. If patents are now being issued at over four times the rate that they were before the Federal Circuit was created, and the rate at which they are being litigated is roughly constant, then there are now approximately four times as many patent lawsuits as there were before the Federal Circuit came into existence. That is a large increase. Furthermore, this data does not capture what many firms have been complaining about—threatened lawsuits, often with little legal basis, that nevertheless result in settlements.

Finally, Khan is mischaracterizing my comments on patent “trolls” by implying that I believe that patent rights should vary depending on “the identity of the patentee.” While it is true that some patent reformers have suggested we need special laws that apply to non-practicing entities, that is not my view. There is nothing illegitimate about the secondary market in patents per se. Rather, as I said in my essay, the problem arises when some parties (including non-practicing entities) exploit asymmetries in the legal system. The troll problem should be alleviated by removing those asymmetries, for example by adopting a loser-pays rule and by ruling on the validity of the patent at issue before costly trial preparations begin.

Sometimes You Need a Blunderbuss

John Duffy and I agree on much about the patent system’s problems, so I am grateful to him for homing in on a remaining area of disagreement: whether it is desirable to eliminate a class of particularly problematic patents by subject matter, or whether we should instead rely on the enforcement of existing legal doctrines to eliminate the worst patents individually. I have always thought that Duffy’s proposed solution—stricter application of the non-obviousness standard to software patents—has a certain elegance. While I have some sympathy for his view, I remain convinced that eliminating the class would be preferable.

Christina Mulligan has already articulated an important response to Duffy; I would have expressed roughly the same concerns, only not as well. One of the key insights of public choice is that elegant academic ideas do not always work well when they meet the reality of interest group politics. If we are designing policy for this messy world, sometimes policy will have to be somewhat less tidy than we might otherwise prefer. Mulligan’s essay on rules versus standards provides an excellent way to think about these issues; please consider her arguments incorporated by reference.

Duffy calls the statutory elimination of software patents a “blunderbuss solution.” To an extent, I accept this characterization, but it is nevertheless worth asking exactly how imprecise this particular blunderbuss would be compared to Duffy’s more surgical response. The non-obviousness standard in § 103 says that one cannot patent an innovation that would have been obvious to “a person having ordinary skill in the art.” For software, the relevant art is programming a general-purpose computer. Someone with ordinary skill in this art is capable of quite a lot—the whole point of using a general-purpose computer is that one can compute anything at all that is computable. Would Duffy agree that rigorous application of the non-obviousness standard would eliminate, if not all, then almost all software patents? Insofar as our solutions yield similar invalidations in practice, my blunderbuss starts to look somewhat less haphazard.

Evaluation of policy options must also contain a cost element. If we agree that there are many low-quality software patents that should be invalidated, what is the cheapest way to invalidate them? Duffy’s solution requires socially costly litigation to determine whether particular software patents represent a non-obvious invention. Even granting that this would not require literally every software patent to be litigated for obviousness, it seems likely that it would result in thousands of trials. Whatever the drawbacks of simply abolishing software patents, it has the virtue of economizing on litigation costs. These cost savings must of course be weighed against the supposed gains of Duffy’s less aggressive standard, but they are significant.

It is worth speculating about exactly which software innovations would survive a more rigorous non-obviousness standard. I can think of a number of important, world-changing software innovations that are arguably non-obvious: for example, the invention of TCP, which created a virtual circuit over a packet-switched network, or the creation of the World Wide Web, which married (pre-existing) hypertext pages with the (pre-existing) Internet. For the sake of argument, assume these were non-obvious inventions. What is interesting about these innovations is not that they were not patented (patents on software were not generally allowed at the time these inventions were made), but that copyrights on the software implementations were not even enforced by the creators. Copyrights and patents are not the same thing, but we might consider the stance of an inventor toward the copyright on her innovative software as a proxy for the extent to which that inventor is motivated by directly monetizing her invention. In the case of at least these non-obvious software innovations, direct monetization does not appear to have been the goal.

Similarly, the most non-obvious consequential software invention of the post-State Street era is arguably Bitcoin. Skilled programmers had been trying for decades to create decentralized digital cash. Yet when Satoshi Nakamoto finally invented cryptocurrency, it was released under a non-restrictive license and not patented. Silicon Valley is now investing many millions of dollars in Bitcoin-related businesses, despite the currency’s tender age.

Admittedly, this is anecdotal evidence, but it is striking: to my eyes, the most consequential innovations that might remain patentable under Duffy’s proposal but would be ineligible for patent protection under mine do not appear to have been motivated by direct monetization. It’s possible that this pattern might not hold up in the future, but there seems to be something fundamentally generative about software that makes the most important innovations something that inventors want to share. Combined with the arguments above, this makes me skeptical that we would actually lose anything significant by abolishing software patents altogether. I am eager to understand why Duffy disagrees with this analysis.

To take this conversation down a somewhat tangential path, it might be interesting to explore the question of how we should choose to define a software patent. A few years ago, in an amicus brief[1] for the Cato Institute, Reason Foundation, and Competitive Enterprise Institute, I and others proposed the following as a definition of software and processes not eligible for patent protection:

The term “process” in section 101 is properly limited to processes which aim to have an effect on matter. Although software and business method patents have proliferated over the past two decades, many software and business-method patents . . . should not qualify as patentable subject matter because the purpose of performing the processes is not to have an effect on the physical world.

Prior Supreme Court precedent is consistent with this historic understanding of patentable process. The aim of the [Diamond v.] Diehr patent was to cure synthetic rubber — to effect a physical change in the world — and that patent was upheld. See [450 U.S. 175, 177 (1981)]. In contrast the purpose of the [Gottschalk v.] Benson patent was to represent a number in binary, [409 U.S. 63, 64 (1972)], and the purpose of the [Parker v.] Flook patent was to calculate updated alarm limits, [437 U.S. 584, 585 (1978)]. Although the algorithms in Benson and Flook would happen to effect a change on a computer, these changes did not transform the claimed algorithms into patentable processes because the purpose of performing the claimed processes was not to alter the physical world. The aim of the Benson and Flook patents was to represent or calculate a number, respectively. Those patents were correctly invalidated.[2]

The reason for limiting patent-eligible processes to those “which aim to have an effect on matter” — as opposed to those which incidentally do (as any math requiring a pencil and paper would have an incidental effect on the paper and pencil tip) — comes from the historical understanding of “art” (the previous, and supposedly equivalent term, for “process” in 35 U.S.C. § 101).[3]

An early authority on patent law, George Ticknor Curtis, described the term “art” as “a new process or method of working or of producing an effect or result in matter.” George Ticknor Curtis, A Treatise on the Law of Patents for Useful Inventions § 9 (Boston, Little, Brown & Co. 3d ed. 1867) []. He illustrated the definition with examples drawn from case law, each of which referred to improved manufacturing techniques involving physical objects. See id. §§ 9-19. A later treatise by William Robinson defined “art” as “an act or a series of acts performed by some physical agent upon some physical object, and producing in such object some change either of character or of condition.” 1 William C. Robinson, The Law of Patents for Useful Inventions § 159 (Boston, Little, Brown & Co. 1890).[4]

There are, of course, other good ways to identify and define problematic patents. As my previous essay noted, a variety of measures would help improve the patent system, because the system has many problems that interact with and compound each other. To provide one additional solution, Mark Lemley has gotten enormous traction arguing for the elimination of “functional claiming” in software patents[5]— i.e. the practice of effectively claiming any invention that solves a problem, rather than the particular solution that has actually been invented.

The challenge behind a call to end software patents, or any type of problematic patent, is to clearly articulate the kind of patents that ought to be eliminated. Currently, we are seeing some software patents being invalidated after the Alice Corp. v. CLS Bank decision.[6] Is Alice enough? Do we need a more clearly articulated rule? What should it be? And, to return to Eli’s initial discussion, I am curious to what degree the development of better law would allay concerns about the Federal Circuit and patent bar being “captured”?