Review of Patent Failure (James Bessen & Michael J. Meurer)
Certainly every person with an interest in patents shud read this book. It is rather clearly written, it is not overly long (260 pages), has a good use of illustrations. The authors take an admireable clear-headed, disinterested, empirical look at the patent system. I definitely recommend this book.
James Bessen and Michael J. Meurer - Patent Failure How Judges, Bureaucrats, and Lawyers Put Innovators at Risk
Below are some quotes and comments to the book.
Claims to veins of minerals create the third, hybrid case, where surface
claims can not entirely avoid costly disputes and the tragedy of the com-
mons might occur, even when miners hold fairly broad rights. A remark-
able example is the so-called War of the Copper Kings in Butte, Montana
(Glasscock 1935). The mountain standing outside of Butte was once
known as the Richest Hill on Earth. It was mined for gold, silver, and,
most notably, copper. The early miners at Butte exhausted the relatively
small supplies of gold and silver in the 1860s and 1870s. At that point
four large mining interests began to buy old claims in a search for copper
ore. By the mid-1880s it was becoming clear that the mountain was laced
with a rich tangle of copper veins that penetrated deep into the mountain.
It was very difﬁcult to trace the copper veins to the surface of the moun-
tain. As a result, it did not become clear until about twenty years later who
owned what copper.
Glasscock explains the source of uncertainty:
The federal mining laws . . . protect[ed] the prospector who ﬁrst lo-
cated an outcropping mineral vein. Such surface indication of valuable
ore was known as the apex of the vein. The owner was guaranteed the
right to follow that vein downward, even when it led under the hold-
ings of claims located behind it. That would have been ﬁne if veins
were always continuous from the surface down, but too frequently
they are not. They are broken or faulted, cut off here and elsewhere by
worthless rock. If a vein leading down from the surface is lost near the
vertical side wall of a claim, and a similar vein of identical ore is found
below it or to one side in the adjoining claim, who is to decide
whether the second discovery is a geological continuation of the ﬁrst?
Who but the courts, basing decision on the expert testimony of geol-
ogists and engineers?
The interlaced veins meant that different mining companies often dug
tunnels beneath or beside the tunnels of their rivals. Occasionally, miners
would break through into a neighboring tunnel. Sales (1964) reports that
gun ﬁghts and chemical warfare occurred in the mines. Sales and Glasscock
both suggest malicious blasting by one mining company injured miners in
other mines. Glasscock reports that one company would develop its claims
so that the water in its mines would drain into rivals’ mines. And both writ-
ers relate that the mining companies would use inefﬁcient extraction meth-
ods in their race to mine a contested vein before their rival was able to.
Legal control over these socially harmful tactics was difﬁcult to achieve be-
cause ownership was unclear and litigation was protracted and costly.17
In some cases, when tangible property is taken from nature, the scope
of the property rights is not so clear. In these cases, simple physical char-
acteristics are not so useful for establishing legal boundaries because the
relevant characteristics change over time or are not fully known initially
(that is, they are revealed over time). The mining disputes discussed in the
previous chapter make this point. Another example comes from water law.
In certain jurisdictions, the right to use water from a stream running
through a property depends on the consumption of others elsewhere on
the streamcourse. Hence, a newcomer will need to investigate her neigh-
bors’ water use to determine whether and to what extent property rights
already exist for the stream flow.
In the case of migratory wild animals, property law follows the “rule of
capture”: you can own what you capture, but not the stock from which it
came. Thus, when someone shoots a wild duck, she does not gain rights to
the flock. It is easy to see how the rule of capture promotes clear notice.
Suppose the first hunter to shoot a duck in a flock actually gained owner-
ship over the flock. It would be virtually impossible for hunters in the next
county to recognize the flock was owned. Furthermore, the counterfactual
property rule would invite endless disputes about who was the true owner
of the flock, and which ducks belong to which flock.31
Similarly, the possession rule in patent law is designed to mitigate notice
problems. Paragraph 1, Section 112 of the patent statute, United States
Code Title 35, requires that the patent describe how to make and use the
invention in sufficient detail so that others can do so. This “enablement” re-
quirement makes the patentee demonstrate the practical knowledge needed
to usefully own the claimed invention.32
This possession requirement allows courts to invalidate patent claims
that are “too broad” insofar as the inventor did not really possess all the
claimed technology. A famous example concerns patents on the light bulb.
Thomas Edison was not the first inventor of the incandescent light bulb.
He had many competitors, and his light bulb built on many earlier contri-
butions.33 William Sawyer and Albon Man together obtained a light bulb
patent before Edison achieved his famous invention and they sued Edison.
Their patent claimed a light bulb with a “conductor of carbon, made of fi-
brous or textile material.” Edison made a light bulb with a bamboo filament
that fell within the language of the broad Sawyer and Man claim. The court
ruled in favor of Edison because Sawyer and Man had actually only made a
light bulb using carbonized paper as a filament. They did not make light
bulbs with other filaments drawn from the wide range of fibrous and textile
carbon-based filaments—in fact, most of those filaments would not work.
Edison labored mightily to find a bamboo filament, which worked very
well—he tried over six thousand different substances before settling on
bamboo. But the Sawyer and Man patent did not describe this important
detail. They possessed the specific invention of a light bulb using car-
bonized paper, but they did not possess the claimed knowledge to make
and use all “fibrous or textile” forms of carbon, including the bamboo later
discovered by Edison. Therefore, the court invalidated Sawyer and Man’s
claim because it claimed more than they actually possessed—they claimed
technology that had not yet been invented.34
Ideally, enablement restricts patent scope so that inventors’ property
rights do not stray far from the invention they actually possess. In the past,
inventors had to demonstrate a working prototype or scale model of the
invention in order to demonstrate possession. Inventors no longer need to
provide a working prototype in order to obtain a patent; the general pos-
session requirement, however, remains central to patent law.
Thus, we are troubled by the many recent examples of patent claims
that have been read broadly to cover infringing technologies that are dis-
tant from the invention actually possessed by the patent owner. Many of
these infringers have arrived at significant inventions independent of any
information contained in the patent at issue. Consider, for example, the
following two cases.
Perhaps one of the clearest lessons of the Cold War was that private-property
and market economies can be powerful engines of economic growth and
innovation. While centralized economies have mustered impressive eco-
nomic efforts, especially during times of war, they have generally failed to
provide a high and rapidly growing standard of living. Moreover, what they
have achieved has sometimes come at a horrible human cost.
The experience of the Cold War seems to lend force to arguments that
intellectual property, too, promotes economic growth and innovation.
Indeed, it is now often argued that the institutions responsible for the suc-
cess of Western economies are “the rule of law and private property rights,
including intellectual property.”1 Similarly the Intellectual Property Own-
ers Association suggests that property-based incentives explain U.S. tech-
nological leadership: “The possibility of patent rights gives incentives to
inventors and their employers to create new technology and to invest in
commercializing technology. Policy makers have generally agreed that the
American tradition of strong patent laws has contributed to making this
country the world’s technological leader, a position it has held for more
than a century.”2 This is a seductive argument. There is solid empirical ev-
idence that secure property rights are conducive to economic growth. So it
might seem to follow that “strong” patent laws should also promote inno-
vation and economic growth. But what is the actual empirical evidence
that patents and other forms of intellectual property are responsible for
the technological leadership of the United States in particular and the
Casual observation suggests that the United States and other Western
nations share both technologically advanced economies and well-developed
patent systems. But this is a correlation, not evidence of causation. That is,
well-developed patent systems might cause economic growth in these na-
tions. Or it might be, instead, that successful technology companies or
other groups, such as the patent bar, have lobbied for patent protection. In
this latter case, economic success promotes the expansion of the patent
system, not the other way around. Indeed, the patent systems in advanced
nations today consist of highly sophisticated institutions supported by sub-
stantial funding. These institutions were not simply legislated, but rather
developed, along with a wide variety of other legal and social institutions.
Their evolution required both extensive experience and a large allocation
of resources and they would seem as out of place in nineteenth-century
America as they would in many of today’s less-developed nations. Thus the
correlation between the sophistication of a nation’s technology and the
sophistication of its patent system does not provide evidence of a causal
link in and of itself; a more advanced analysis is required.
It might well be true, as the Intellectual Property Owners maintain, that
most policymakers see a link between “strong” patent laws and U.S. tech-
nology leadership.3 But as James Boyle acerbically notes, policymakers have
too often ignored empirical evidence, basing policy, instead, on “faith-
based” reasoning about property rights with regard to such matters as soft-
ware patents, broadcast rights, copyright term, and database rights.4
Of course, the economic effectiveness of all forms of property depends
on details of the supporting institutions—this is evident from the disparate
growth-paths of Soviet Bloc economies. But the economic effectiveness of
patents might be much more sensitive to the details of the relevant institu-
tions than are general property rights. Perhaps this is because patent law
might be much more specialized, complex, and sophisticated than, say, real
property law, and thus effective institutions might be more difficult to
develop and maintain.
In any case, the empirical economic evidence strongly rejects simplistic
arguments that patents universally spur innovation and economic growth.
“Property” is not a ritual incantation that blesses the anointed with the
fruits of innovation; legislation of “stronger” patent rights does not auto-
matically mean greater innovation. Instead, the effectiveness of patents as
a form of property depends critically on the institutions that implement
patent law. And there appear to be important differences in the effective-
ness of the implementation across different technologies and industries.
On the other hand, we can also reject the view that patents uniformly
stifle innovation. In the pharmaceutical industry and in the nineteenth-
century United States, we see definite evidence that patents do and did
sometimes provide positive private incentives for innovation.
Of course, we have asked and answered an intentionally narrow ques-
tion here. We have not asked whether the patent system is the best way
to encourage innovation. Nor have we even asked whether the total net
effect of the patent system is positive. Some argue, for instance, that
mechanisms such as rewards or purchase contracts would be more so-
cially efficient ways of encouraging pharmaceutical research. Others,
such as Boldrin and Levine (2005), argue that even though patents pro-
vide some individuals with rewards, they are not necessary to encourage
innovation and that they are socially wasteful because they make subse-
quent innovations more difficult. These are interesting and important
questions, but we doubt that they can be answered very well at this time
based strictly on the empirical evidence. That is, the evidence is incon-
clusive with regard to these questions.
Our approach in the following chapters is to focus on the narrower
questions of whether and where today patents do function effectively as a
property system, what factors affect this performance, and what institu-
tional changes might improve the effectiveness of the patent system. We
limit our inquiry to the extent that we seek to obtain definitive answers. We
do, however, think that the effectiveness of patents as a property system is
central in any case to some of the other considerations noted above. If the
patent system can be made more effective, then this necessarily affects any
comparison to alternative policies. It also affects any assessment regarding
the balance between private incentives for initial innovation and those for
follow-on innovations. If patents can be made to work like property, then
this constitutes a powerful argument in favor of the patent system.
Moreover, even after controlling for a wide range of variables, the more a
ﬁrm spends on R&D, all else being equal, the more likely it is to be sued for
infringement. This is inconsistent with the notion that infringers cheat to
avoid R&D. We would expect cheaters to spend less on R&D, all things
again being equal. And to the extent that R&D expenditures can be used to
hide infringing technology, we would also expect greater R&D spending
to be associated with a lower risk of detection. Instead, this pattern is en-
tirely consistent with the inadvertent-infringement explanation—the more
a ﬁrm invests in technology, the more it inadvertently exposes itself to
patents of which it is not aware.3
The idea that patent infringers are large R&D spenders also seems to be
at odds with the picture of pirates we hold from other areas of law. Copy-
right and trademark pirates are often small-time operators such as street
vendors. They hope to “ﬂy under the radar” of the property owners’ mon-
itoring efforts. Large retailers, on the other hand, take great pains to make
sure that they are not selling counterfeit goods because any infractions
would likely come to the notice of the property owners and their cus-
tomers. We would expect large technology companies to take great pains
to avoid infringement (as Kodak did) precisely because they are so visible.
This raises yet another point: if RIM consciously stole NTP’s property,
then one would expect RIM to at least make some effort to hide its crime.
Instead, RIM publicized its allegedly infringing technology. RIM came to
NTP’s attention because of a press release that RIM put out—the func-
tional description of RIM’s product in the press release was sufﬁcient for
NTP to determine that an infringement lawsuit could be ﬁled.
It would appear that actual evidence of hiding seems rather limited. Al-
leged infringers often act like RIM. For example, in lawsuits involving soft-
ware, the alleged infringer typically has a publicly available product or service.
Quite frequently patent holders claim that certain publicly observable prod-
uct features are infringing. Moreover, the powerful reverse-engineering tools
available for software mean that publicly available products can easily be
checked for infringement. If most alleged infringers were cheaters, then we
would expect relatively few lawsuits over publicly observable products—
cheaters would avoid technologies where they could not hide their theft. But,
in fact, most patent lawsuits involving software appear to involve publicly
observable features and litigation rates on software patents are relatively high
(Allison et al. ; see also chapter 9 in the present volume). And in gen-
eral, ﬁrms report that they can detect infringement in most products, but not
in most processes.4 This does not seem to inhibit patent lawsuits over prod-
ucts relative to processes.
Even simple delay can impose large business costs. Consider, for example,
litigation against Cyrix, a start-up ﬁrm that introduced Intel-compatible
microprocessors. Intel, the dominant maker of microprocessors, sued Cyrix
and the litigation lasted four years (there were multiple suits). During much
of that time Cyrix had difﬁculty selling microprocessors to computer manu-
facturers because most of them were also customers of Intel and they were
reluctant to buy a product that might infringe. Cyrix also had difﬁculty ﬁnd-
ing fabricators willing to manufacture their chips—again, for fear of being
sued themselves. In the meantime, Intel responded by accelerating its devel-
opment of chips that would compete against Cyrix’s offerings. In the end,
Cyrix won the lawsuit, but lost the war, having lost much of its competitive
advantage. In effect, Cyrix lost the window of opportunity to establish itself
in the marketplace. Litigation exacted a heavy toll, indeed.
Never heard of that. Fuck you Intel!
More notable still is that some of the most successful individual inven-
tors succeeded not because of their inventive contribution but because of
their patents. Jerome Lemelson, a proliﬁc inventor with close to 600
patents, is renowned among patent lawyers as the master of “submarine”
patents—patents kept hidden for many years. Lemelson slowed the prose-
cution of his patents, sometimes for over twenty years.3 He waited until
his technologies were independently invented and commercialized, and
then he brought his patent to the surface and negotiated royalties after the
potential licensees were locked into the patented technology.4 Although
his patents covered breakthrough technologies such as bar-code scanning,
he did not contribute these breakthroughs to society.
In sum, patents on software are not just like other patents. The evidence
shows that software patents are particularly prone to litigation and to dis-
putes over patent boundaries, a concern that has been raised about them
since the 1960s. We attribute these problems to the abstract nature of soft-
ware technology; too many software patents claim all technologies with
similar form or all means of achieving a result, when the actual invention
is much more limited and often trivial.
Patent law has developed a number of doctrines to circumscribe ab-
stract patent claims. Unfortunately, the Federal Circuit has set software-
speciﬁc precedents that essentially remove most restrictions on abstract
claims in software. Perhaps the court acted out of a desire to promote
patents in this ﬁeld of technology that has historically not used patents.
The result has been both a proliferation of software patents and lawsuits.
Software patents are not the only patents to suffer problems of abstract
claims. Any technology can be claimed abstractly and, to make matters
worse, the Federal Circuit has recently eroded limits on abstract patents
for nonsoftware business processes and even basic scientiﬁc ideas (for ex-
ample, Laboratory Corp. of America v. Metabolite Laboratories). But overall,
software patents likely have a far greater inﬂuence on the performance of
the patent system than do nonsoftware business processes.
Software patents are, in fact, responsible for a major share of patent
lawsuits. They thus play a central role in the failure of the patent system as
a whole. Any serious effort at patent reform must address these problems
and the failure to deal with the problems of software patents—either with
software-speciﬁc measures or general reforms—will likely doom any reform
effort. We turn to possible changes in patent policy in the next chapter.
Real property rights, as opposed to abstract conceptions of property, have
limits, however. The messy, practical details of deﬁning boundaries, provid-
ing public notice, facilitating clearance, and so forth, place real constraints
on where property can be effective. A reasonable property system recognizes
such limits. A landowner gets no rights to untapped oil ﬂowing beneath her
land nor to migratory ducks who put down on it nor to the airplanes that
ﬂy over it. Property rights should be granted only when property owners
can manage them efﬁciently, and only if third parties can effectively cope
The same is true with property rights in inventions. Economics re-
search conﬁrms that the effectiveness of patents varies by type of inven-
tion. For example, patents have worked best where boundaries can be
staked in veriﬁable physical characteristics, like small molecules. With
many chemical patents, third parties can test alternative substances and
unambiguously determine whether they fall within the patent claims or
not. In this case, the boundaries are clear, disputes and litigation are rela-
tively infrequent, and the economic beneﬁts of patents are high.2
On theother hand, patents work poorly when they are highly abstract, claiming
technologies that are not known to the patentee or not even developed at
the time of application. As was seen in chapter 9 with respect to software,
it is sometimes difﬁcult, or even impossible, to distinguish which tech-
nologies are covered by abstract patent claims; not surprisingly, software
patents have high litigation rates and high costs, as do patents on ﬁnancial
and other business inventions.
And so we return to our theme of abstraction in another guise. As with
limitations on other property, the law has long recognized that there are
substantive limits on which inventions can be patented, including limita-
tions on abstract patents. Yet implementing this limitation is one of the
most intractable problems facing any property rights system for inven-
tions. Since the eighteenth century, patent law has attempted to proscribe
abstract patents, but the doctrines used and their application have not al-
ways been successful or uncontroversial. It bears repeating that we do not
claim to know how to craft the best policy regarding abstract patents. Yet
the empirical evidence convinces us that allowing patents on “everything
under the sun” while simultaneously encouraging that patenting by relax-
ing non-obviousness and enablement standards for key technologies con-
stitute a major departure from the policy of the past. And although this
departure might sound good in the abstract, its record, like the record
regarding claim construction, has been one of failure.
The problem with mistaking abstract conceptions of property for the
real thing is that this substitutes rhetoric for reasoned policy, where per-
formance can be measured, evaluated, and adjusted. The result is policy
that loses touch with reality. In the worst case, abstract rhetoric about
property rights or about the sanctity of the patent statute simply provides
cover for special interests.3 The antidote is empirical evidence, and the ev-
idence we have assembled unequivocally shows an all-too-real patent sys-
tem far removed from the ideal found in so much of the rhetoric. But the
picture we paint is also far removed from what the patent system could be.
We think the historical record is clear—the patent system can perform
well, and it can perform badly. The legal and institutional details are criti-
cal. So is the economic and technological environment. Like other times
in American history, we face a challenge today to improve the perform-
ance of the patent system. Yet the data in ﬁgure 12.1 give us pause. The
challenge facing the patent system today might be more difﬁcult and the
stakes might be higher than they have been in the past. A unitary patent
system simply cannot survive if it works well in some industries, but fails
critically in others. If patent institutions prove inﬂexible, then perhaps we
will be left with a patent system for chemicals and pharmaceuticals and lit-
tle else. In any case, the future of the patent system will depend on getting
beyond rhetoric and abstract thinking to build institutions that improve
patent notice, even if this comes with realistic limits on what can be
patented and how it can be claimed. Then, perhaps, the patent system can
deliver on its promise as a property system for inventions.
These are the closing words.