Can Public Policy Influence Private Innovation?
Media
Part of Can Public Policy Influence Private Innovation?
- Title
- Can Public Policy Influence Private Innovation?
- extracted text
-
Can Public Policy Influence Private
Innovation?
JAMES BESSEN
Abstract
Private innovation appears to have played a major if not dominant role in the
growth of output per capita since the Industrial Revolution. Yet economic theory
indicates that the level of investment in private innovation will generally be less
than is socially optimal unless public policies such as patents encourage additional
investment. Therefore, public innovation policy would seem to be critical to
economic growth. Surprisingly, a large body of research fails to find unambiguously
positive effects of patents on innovation and economic growth, even though patents
have been used for hundreds of years. While some industries in some nations clearly
benefit from patents, many other industries do not and patents might actually
discourage innovation in some industries. Economic theory has provided valuable
insights, yet real-world policy apparently needs to reflect a richer set of behavior
and a more complex legal environment. New research is developing a more nuanced
understanding including research on alternative means of providing rewards to
innovators, research on the costs of litigation and disputes arising from the failure
of patents to provide clear boundaries, research on cumulative innovation and
strategic uses of large blocks of patents, and research on the extension of patent
coverage to new technologies and to developing nations. In addition, major new
sources of data permit much more extensive empirical research.
INTRODUCTION
In 1529, one Anton Müller of Danzig (Gdansk) invented a loom that wove
multiple ribbons simultaneously. The Danzig city council, however, worried
about the effect this invention would have on the employment of weavers,
secretly had Müller drowned. This policy succeeded in suppressing the
invention for nearly a century until it reemerged in Flanders nearly a century
later. This invention eventually developed into a more automated form
known as the Dutch Loom, the first automated loom and a predecessor of one
of the key technologies of the Industrial Revolution.
Clearly, policy can have an effect on innovation, but not always the effect
that is best for society. The question is whether policies can and do have net
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
positive social effects on innovation. Indeed, there appear to be large differences in the abilities of nations to develop and employ new technologies.
Is this because of differences in innovation policies? Alternatively, is it due
to other differences in general economic institutions, culture, or even geography? In particular, can social science tell us how well past policies have
worked and provide helpful guidance to policymakers for the future?
Perhaps surprisingly, the positive evidence for major past innovation policies, such as patents, is mixed at best. Patents seem to work well in some
circumstances but not in others. It appears that past economic theory has
been a bit too simple, not recognizing some of the costs that patents can
impose on innovators as well as the contingent nature of the benefits. Perhaps
not surprisingly, economists and other social scientists have only had limited
success providing guidance to policymakers that is implemented. However,
change is visible on both fronts. Richer economic theory, empirical analysis
of newly developed datasets and detailed policy analysis have begun to provide a better understanding of what works, where the problems lie and how
better policy can be crafted.
FOUNDATION
Almost all of the social science research relevant to public policy for innovation has been in economics, so I will begin with a brief overview of the
relevant economic theory. Modern economic analysis of innovation begins
with Kenneth Arrow’s 1962 essay where he argued that inventions would
be underpriced in competitive markets without patents and where he considered the incentives for innovation in both monopolies and competitive
markets. Arrow formalized a long held intuition that free-riding externalities would undermine the returns to innovation without patents. In Arrow’s
analysis, inventions are a form of information, an “idea” that can be replicated at a negligible marginal cost. Without a patent, a competitive market
would price the idea at this marginal cost, so inventors would have difficulty
recouping their initial investments without patents. Patents, by creating a
temporary monopoly, allow the patent holder to charge a higher price and
thereby provide a greater return on the inventor’s investment in innovation.
In 1969, William Nordhaus developed a model of the tradeoff between the
social losses associated with the high price under this temporary monopoly
and invention incentives. In his model, the tradeoff was determined by a
policy variable, namely the patent term. Subsequent models added additional policy variables such as patent “breadth” and “scope.” These were
followed by models of patent races, where inventors would race to get a
patent, depending on the strength of the incentives, and models of sequential
Can Public Policy Influence Private Innovation?
3
innovation, where the tradeoffs between pioneer inventors and improvers
were affected by policy variables.
Yet these foundations had significant limitations in regard to guiding policy. First, many of the putative policy variables such as patent “breadth” did
not correspond to any real-world policy levers. Although patent lawyers talk
about the breadth of a patent, this is based on specific analysis of particular patents, it is often highly uncertain and is not something that could be
measured by economists or easily regulated by the Patent Office or courts.
Second, the basic economic models presented a highly idealized version of
patent rights, not reflecting, for example, costs that the patent system imposes
on innovators. Indeed, all property rights rely on a system of “public notice”:
a way to determine the boundaries and owners of the rights. Much policy
is focused on providing efficient public notice. When a property rights system works well, economic agents can conduct clearance searches at little cost
and can obtain the necessary licenses to conduct their business. However,
when public notice functions do not work well, clearance search can become
too costly to perform and the boundaries of property rights can be unpredictable. In this case, economic agents become prone to inadvertent infringement, giving rise to unwanted disputes and litigation. Bessen (the author)
and Meurer (2008) have explored the role of public notice for patents. The
expected costs of possible disputes and litigation over unclear boundaries of
patents impose disincentives on prospective innovators, tending to counter
the positive incentives provided by patent rents.
Third, despite the seeming consensus on the centrality of patents to
innovation, many economists remained deeply skeptical about the benefits
provided by real-world patent systems. The empirical support for patents
was not strong. Survey research showed that patents did not provide
strong incentives in most industries. There were other ways of protecting
innovations and other incentives, including prizes and government contracting. Moreover, in some cases innovators freely shared their knowledge,
seemingly contradicting the canonical models. In addition, some researchers
had difficulty finding the relevance of patent race models to real-world innovative competition. Some economists, such as Boldrin and Levine (2008),
questioned whether patent monopolies fundamentally slowed innovation
when each new innovation cumulatively builds on previous innovations.
While the simplified models do provide valuable insights, they can prove
of limited value for guiding policy where the law is complex and where innovators act under a richer set of conditions than envisioned in these models.
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
THE STATE OF RESEARCH: HAVE INNOVATION POLICIES
PROMOTED INNOVATION?
By far, most of the empirical research in this area has focused on patents,
which is not surprising because patents have been used for centuries. Douglass North argued that secure property rights facilitated the British Industrial Revolution, including patent rights on inventions. However, economic
historians studying patents have been more doubtful. The reason for their
skepticism is that few of the major inventions of the Industrial Revolution
had patents that brought substantial rewards to their inventors. Many inventions were not patented. Indeed, it appears that only 11% of British inventions
shown at the 1851 World’s Fair were patented. Other major inventions were
patented but did not make money, sometimes only becoming commercially
feasible after the patent expired. Some inventors obtained patents and then
were ruined by costly litigation. Some, such as Arkwright, had their patents
invalidated; Arkwright made a fortune nevertheless. In addition, a few, such
as James Watt, had profitable patents. If most inventors did not benefit from
patents, it is hard to argue that patents provided them a strong incentive to
innovate.
Other sorts of evidence also show what is at best—a mixed picture of the
ability of patents to spur innovation. A number of cross-country studies have
run panel regressions on national rates of gross domestic product (GDP)
growth using a variety of right-hand-side controls, including the strength
of the nation’s patent rights. While the strength of general property rights
is generally correlated with economic growth, the strength of patent rights
or intellectual property rights is not. Patent rights do seem to be correlated
with national rates of R&D spending, but some evidence suggests that the
causality might run in the reverse direction. That is, in one study changes
in the law to strengthen patent rights follow national R&D spending. This
might happen if firms invest in R&D and then lobby for legal changes once
they have achieved some success.
A number of other studies look at natural economic experiments where
patent law was changed, often in response to obligations associated with
trade treaties. In general, pro-patentee changes in a nation’s patent laws tend
to increase patenting by foreign firms. This makes sense because their products now can be sold with a greater degree of exclusivity if protected with a
patent. However, such changes in patent law do not, in general, lead to more
domestic patenting or more domestic R&D spending. A possible exception
exists for highly developed nations.
Another sort of evidence comes from studies of patent value. The rents that
a patent generates are a partial measure of the incentive that patents provide.
A discounted stream of rents provides a measure of patent value. Two sorts
Can Public Policy Influence Private Innovation?
5
of econometric methods have been used to measure the value of patents to
their owners. One method uses data on the payment or nonpayment of the
renewal fees that must be paid to keep a patent in force. Economists infer
that if a patent holder is unwilling to pay a fee to keep a patent in force for,
say, an additional year, then the rents that the patent generates must be less
than the fee. Under some simple assumptions, the mean value of rents can be
estimated for a group of patents and these can be compared to R&D expenditures. This literature finds that patent rents tend to run about 10–15% of
R&D—suggesting that patents provide a rather modest but positive incentive
for R&D spending and that most R&D spending might take place without
patents. However, renewal estimates typically do not observe the rents on
the most valuable patents because all of these patents are renewed to the full
term of the patent. In effect, the renewal method estimates the values of these
patents by extrapolation and therefore it is possible that these estimates are
understated.
A second method produces estimates that are overstated. This method estimates the value of patents held by public firms by regressing a measure of
the market value of the firm against measures of its assets, including a stock
of patents. The coefficient on the patent stock provides a measure of patent
value. The corresponding rents turn out to be about 18% of R&D spending,
slightly above the estimates found using renewal data. In addition, this estimate will be overstated to the extent that firms obtain additional patents in
response to success in the market place (a reverse causality).
Despite the small size of patent rents relative to R&D, the rents are
positive and significant. Nevertheless this does not necessarily imply
that patents spur innovation. The problem is that patents also impose
costs—disincentives—on innovators, as noted above. The costs of patent
litigation, including both the costs of lawyers and the effect of the lawsuit on
business, are one of those costs when patent boundaries are unpredictable
or too costly to conduct clearance search. Researchers have estimated these
costs for public firms using stock market event studies around the filing
of patent lawsuits. These estimates suggest that: (i) the costs have grown
rapidly since the mid-1990s, (ii) in technologies such as pharmaceuticals
and chemicals—where patent boundaries are well-defined—the costs
imposed by litigation are small compared to patent rents, but, (iii) in other
technologies, especially information technologies, litigation costs are now
far greater than patent rents. In these industries, patents appear to provide a
net disincentive to innovation.
Much recent research has focused on specific aspects of patents and
innovation, thus helping to fill out a more detailed picture of the innovation
landscape. In addition to estimating the burden of litigation on innovators,
research on patent litigation has also identified factors that affect a firm’s
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
hazard of being sued or of having to go to court to enforce its patents. Other
research, both theoretical and empirical, has sought to better understand the
nature and timing of dispute settlement.
One important area in litigation research has been on patent “nonpracticing
entities” (NPEs), known popularly as patent trolls. These are organizations
that do not develop a technology to be implemented, but, instead, acquire
patents to assert against others in order to collect royalties. The number of
lawsuits involving NPEs has soared in recent years; over 5000 defenses were
mounted in 2011 in lawsuits filed by NPEs according to one source. This
new development has been controversial. The patent system has long had
a variety of nonpracticing agents who have facilitated markets for technology. Indeed, one of the recent advances in research has been recognition that
patents can serve to facilitate markets for technology and these have been
historically important. They allow small inventors who might not have the
resources and skills to bring a product to market to sell or license the technology to a party who can produce and distribute it. However, the recent spate
of NPEs tend to be much more interested in litigating against firms that have
already developed technologies than in licensing a technology to firms that
need it. That is, much of the NPE activity is about licensing freedom from the
threat of a lawsuit rather than licensing an actual technology, with training,
know-how, and so on. Some research shows that NPE litigation appears to do
far more damage to the defendants, who tend to be firms that invest heavily
in technology, than any benefit it provides to the small inventors who create
the patents used in NPE lawsuits.
Another area of study has been on those subject matter areas where the
courts have extended patent coverage during recent years. Before 1994, for
example, software algorithms were seen as unpatentable by themselves.
Court decisions since the mid-1990s have extended patent coverage not only
to software, but also to methods of doing business and mental processes.
Over a quarter of a million patents have now been issued in these new
areas, making for something of a natural economic experiment. Research
has shown that patents in these new areas are much more likely to be
involved in litigation. Indeed, NPE lawsuits are heavily concentrated in
software and business methods. It appears that the high rate of litigation is
not so much because these fields are new to the patent system, but rather
that these subjects of major problems with patent notice. In particular, the
interpretation by the courts of patent boundaries in these fields has been
highly unpredictable, giving rise to opportunistic litigation. Research on
software patents has been influential in legislative proceedings in Europe
and in court decisions in the United States.
Software patents have also featured in another area of research, namely,
patent thickets. Because firms acquired tens of thousands of software patents
Can Public Policy Influence Private Innovation?
7
(being conceptual, they often require very little R&D), economists became
aware that they were using them strategically as a block. A number of theoretical papers have studied the nature of this strategic interaction and a
number of empirical papers have sought to identify the effects of patent thickets on innovation. The findings appear to be double-edged: patent thickets
do deter entry of innovators into new markets, but prospective entrants are
more likely to enter if they acquire patents themselves.
Economists have also studied the operation of the patent office and the
patent examination process. Research on one administrative procedure—a
“post-grant review” proceeding that allows third parties to challenge patents
immediately after they are issued—influenced legislation in the United States
such that the proceeding was included in the patent act passed into law in
2011.
This review has focused on research on patents. Research has also been
done on other sorts of policies intended to foster innovation. For example,
economists have attempted to estimate the effect of R&D tax credits on R&D
spending.
NEW DIRECTIONS
Three sorts of factors influence those areas where future research will be
fertile:
Better, more realistic understanding of how innovation works. For example,
“spillovers” have long been identified as important, but this concept
includes a variety of different sorts of actual behavior. Some new
research has focused on the effect of laws that restrict employee
mobility, one source of spillover. Other research is looking into the free
exchange of knowledge among innovators, something long recognized,
but little explored.
Better data. Patent offices and courts are making more data available to
researchers, opening up new areas of inquiry, for example, better information on patent prosecution and on payment of renewal fees. In addition, a number of surveys have been initiated in recent years, covering
all aspects of the innovation process.
New developments. Nonprofit organizations have proposed using prizes
and similar measures in order to promote the development of pharmaceuticals for neglected diseases. This is promoting research on prizes.
In addition, international trade treaties have resulted in dramatically
changed patent systems in developing countries such as China, also
opening up new avenues for research.
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
REFERENCES
Arrow, K. (1962). Economic welfare and the allocation of resources for invention.
The rate and direction of inventive activity economic and social factors (pp. 609–626).
Cambridge, MA: National Bureau of Economic Research.
Bessen, J., & Meurer, M. J. (2008). Patent failure: How judges, bureaucrats, and lawyers
put innovators at risk. Princeton, NJ: Princeton University Press.
Boldrin, M., & Levine, D. K. (2008). Against intellectual monopoly. Cambridge, England: Cambridge University Press.
Nordhaus, W. D. (1969). Invention, growth, and welfare: A theoretical treatment of technological change. Cambridge, MA: MIT Press.
JAMES BESSEN SHORT BIOGRAPHY
James Bessen is a scholar on the economics of innovation and patents
who has also been a successful innovator and CEO of a software
company. In 1983, Bessen developed the first commercially successful
“what-you-see-is-what-you-get” desktop publishing program, founding
a company that delivered PC-based publishing systems to high end commercial publishers. Since leaving the software business, Mr. Bessen has
done economics research on innovation, including theoretical and empirical
studies of the economics of patents as well as studies of innovation in the
nineteenth century and its effects on skills, wages, and productivity. His
work on patents has influenced policymakers and courts in the United
States, Europe, and Australia. He is the author of Learning by Doing: The Real
Connection between Innovation, Wages, and Wealth (Yale 2015). With Michael
J. Meurer, he wrote Patent Failure: How Judges, Bureaucrats, and Lawyers Put
Innovators at Risk (Princeton 2008). Currently, Mr. Bessen is a Lecturer in Law
at the Boston University School of Law and Fellow at the Berkman Center
on Internet and Society at Harvard.
See also http://www.researchoninnovation.org/bio.php
RELATED ESSAYS
The Public Nature of Private Property (Sociology), Debbie Becher
Technology Diffusion (Economics), Adam B. Jaffe
Knowledge Transfer (Psychology), Timothy J. Nokes-Malach and J. Elizabeth
Richey
Translational Sociology (Sociology), Elaine Wethington
-
Can Public Policy Influence Private
Innovation?
JAMES BESSEN
Abstract
Private innovation appears to have played a major if not dominant role in the
growth of output per capita since the Industrial Revolution. Yet economic theory
indicates that the level of investment in private innovation will generally be less
than is socially optimal unless public policies such as patents encourage additional
investment. Therefore, public innovation policy would seem to be critical to
economic growth. Surprisingly, a large body of research fails to find unambiguously
positive effects of patents on innovation and economic growth, even though patents
have been used for hundreds of years. While some industries in some nations clearly
benefit from patents, many other industries do not and patents might actually
discourage innovation in some industries. Economic theory has provided valuable
insights, yet real-world policy apparently needs to reflect a richer set of behavior
and a more complex legal environment. New research is developing a more nuanced
understanding including research on alternative means of providing rewards to
innovators, research on the costs of litigation and disputes arising from the failure
of patents to provide clear boundaries, research on cumulative innovation and
strategic uses of large blocks of patents, and research on the extension of patent
coverage to new technologies and to developing nations. In addition, major new
sources of data permit much more extensive empirical research.
INTRODUCTION
In 1529, one Anton Müller of Danzig (Gdansk) invented a loom that wove
multiple ribbons simultaneously. The Danzig city council, however, worried
about the effect this invention would have on the employment of weavers,
secretly had Müller drowned. This policy succeeded in suppressing the
invention for nearly a century until it reemerged in Flanders nearly a century
later. This invention eventually developed into a more automated form
known as the Dutch Loom, the first automated loom and a predecessor of one
of the key technologies of the Industrial Revolution.
Clearly, policy can have an effect on innovation, but not always the effect
that is best for society. The question is whether policies can and do have net
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
positive social effects on innovation. Indeed, there appear to be large differences in the abilities of nations to develop and employ new technologies.
Is this because of differences in innovation policies? Alternatively, is it due
to other differences in general economic institutions, culture, or even geography? In particular, can social science tell us how well past policies have
worked and provide helpful guidance to policymakers for the future?
Perhaps surprisingly, the positive evidence for major past innovation policies, such as patents, is mixed at best. Patents seem to work well in some
circumstances but not in others. It appears that past economic theory has
been a bit too simple, not recognizing some of the costs that patents can
impose on innovators as well as the contingent nature of the benefits. Perhaps
not surprisingly, economists and other social scientists have only had limited
success providing guidance to policymakers that is implemented. However,
change is visible on both fronts. Richer economic theory, empirical analysis
of newly developed datasets and detailed policy analysis have begun to provide a better understanding of what works, where the problems lie and how
better policy can be crafted.
FOUNDATION
Almost all of the social science research relevant to public policy for innovation has been in economics, so I will begin with a brief overview of the
relevant economic theory. Modern economic analysis of innovation begins
with Kenneth Arrow’s 1962 essay where he argued that inventions would
be underpriced in competitive markets without patents and where he considered the incentives for innovation in both monopolies and competitive
markets. Arrow formalized a long held intuition that free-riding externalities would undermine the returns to innovation without patents. In Arrow’s
analysis, inventions are a form of information, an “idea” that can be replicated at a negligible marginal cost. Without a patent, a competitive market
would price the idea at this marginal cost, so inventors would have difficulty
recouping their initial investments without patents. Patents, by creating a
temporary monopoly, allow the patent holder to charge a higher price and
thereby provide a greater return on the inventor’s investment in innovation.
In 1969, William Nordhaus developed a model of the tradeoff between the
social losses associated with the high price under this temporary monopoly
and invention incentives. In his model, the tradeoff was determined by a
policy variable, namely the patent term. Subsequent models added additional policy variables such as patent “breadth” and “scope.” These were
followed by models of patent races, where inventors would race to get a
patent, depending on the strength of the incentives, and models of sequential
Can Public Policy Influence Private Innovation?
3
innovation, where the tradeoffs between pioneer inventors and improvers
were affected by policy variables.
Yet these foundations had significant limitations in regard to guiding policy. First, many of the putative policy variables such as patent “breadth” did
not correspond to any real-world policy levers. Although patent lawyers talk
about the breadth of a patent, this is based on specific analysis of particular patents, it is often highly uncertain and is not something that could be
measured by economists or easily regulated by the Patent Office or courts.
Second, the basic economic models presented a highly idealized version of
patent rights, not reflecting, for example, costs that the patent system imposes
on innovators. Indeed, all property rights rely on a system of “public notice”:
a way to determine the boundaries and owners of the rights. Much policy
is focused on providing efficient public notice. When a property rights system works well, economic agents can conduct clearance searches at little cost
and can obtain the necessary licenses to conduct their business. However,
when public notice functions do not work well, clearance search can become
too costly to perform and the boundaries of property rights can be unpredictable. In this case, economic agents become prone to inadvertent infringement, giving rise to unwanted disputes and litigation. Bessen (the author)
and Meurer (2008) have explored the role of public notice for patents. The
expected costs of possible disputes and litigation over unclear boundaries of
patents impose disincentives on prospective innovators, tending to counter
the positive incentives provided by patent rents.
Third, despite the seeming consensus on the centrality of patents to
innovation, many economists remained deeply skeptical about the benefits
provided by real-world patent systems. The empirical support for patents
was not strong. Survey research showed that patents did not provide
strong incentives in most industries. There were other ways of protecting
innovations and other incentives, including prizes and government contracting. Moreover, in some cases innovators freely shared their knowledge,
seemingly contradicting the canonical models. In addition, some researchers
had difficulty finding the relevance of patent race models to real-world innovative competition. Some economists, such as Boldrin and Levine (2008),
questioned whether patent monopolies fundamentally slowed innovation
when each new innovation cumulatively builds on previous innovations.
While the simplified models do provide valuable insights, they can prove
of limited value for guiding policy where the law is complex and where innovators act under a richer set of conditions than envisioned in these models.
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
THE STATE OF RESEARCH: HAVE INNOVATION POLICIES
PROMOTED INNOVATION?
By far, most of the empirical research in this area has focused on patents,
which is not surprising because patents have been used for centuries. Douglass North argued that secure property rights facilitated the British Industrial Revolution, including patent rights on inventions. However, economic
historians studying patents have been more doubtful. The reason for their
skepticism is that few of the major inventions of the Industrial Revolution
had patents that brought substantial rewards to their inventors. Many inventions were not patented. Indeed, it appears that only 11% of British inventions
shown at the 1851 World’s Fair were patented. Other major inventions were
patented but did not make money, sometimes only becoming commercially
feasible after the patent expired. Some inventors obtained patents and then
were ruined by costly litigation. Some, such as Arkwright, had their patents
invalidated; Arkwright made a fortune nevertheless. In addition, a few, such
as James Watt, had profitable patents. If most inventors did not benefit from
patents, it is hard to argue that patents provided them a strong incentive to
innovate.
Other sorts of evidence also show what is at best—a mixed picture of the
ability of patents to spur innovation. A number of cross-country studies have
run panel regressions on national rates of gross domestic product (GDP)
growth using a variety of right-hand-side controls, including the strength
of the nation’s patent rights. While the strength of general property rights
is generally correlated with economic growth, the strength of patent rights
or intellectual property rights is not. Patent rights do seem to be correlated
with national rates of R&D spending, but some evidence suggests that the
causality might run in the reverse direction. That is, in one study changes
in the law to strengthen patent rights follow national R&D spending. This
might happen if firms invest in R&D and then lobby for legal changes once
they have achieved some success.
A number of other studies look at natural economic experiments where
patent law was changed, often in response to obligations associated with
trade treaties. In general, pro-patentee changes in a nation’s patent laws tend
to increase patenting by foreign firms. This makes sense because their products now can be sold with a greater degree of exclusivity if protected with a
patent. However, such changes in patent law do not, in general, lead to more
domestic patenting or more domestic R&D spending. A possible exception
exists for highly developed nations.
Another sort of evidence comes from studies of patent value. The rents that
a patent generates are a partial measure of the incentive that patents provide.
A discounted stream of rents provides a measure of patent value. Two sorts
Can Public Policy Influence Private Innovation?
5
of econometric methods have been used to measure the value of patents to
their owners. One method uses data on the payment or nonpayment of the
renewal fees that must be paid to keep a patent in force. Economists infer
that if a patent holder is unwilling to pay a fee to keep a patent in force for,
say, an additional year, then the rents that the patent generates must be less
than the fee. Under some simple assumptions, the mean value of rents can be
estimated for a group of patents and these can be compared to R&D expenditures. This literature finds that patent rents tend to run about 10–15% of
R&D—suggesting that patents provide a rather modest but positive incentive
for R&D spending and that most R&D spending might take place without
patents. However, renewal estimates typically do not observe the rents on
the most valuable patents because all of these patents are renewed to the full
term of the patent. In effect, the renewal method estimates the values of these
patents by extrapolation and therefore it is possible that these estimates are
understated.
A second method produces estimates that are overstated. This method estimates the value of patents held by public firms by regressing a measure of
the market value of the firm against measures of its assets, including a stock
of patents. The coefficient on the patent stock provides a measure of patent
value. The corresponding rents turn out to be about 18% of R&D spending,
slightly above the estimates found using renewal data. In addition, this estimate will be overstated to the extent that firms obtain additional patents in
response to success in the market place (a reverse causality).
Despite the small size of patent rents relative to R&D, the rents are
positive and significant. Nevertheless this does not necessarily imply
that patents spur innovation. The problem is that patents also impose
costs—disincentives—on innovators, as noted above. The costs of patent
litigation, including both the costs of lawyers and the effect of the lawsuit on
business, are one of those costs when patent boundaries are unpredictable
or too costly to conduct clearance search. Researchers have estimated these
costs for public firms using stock market event studies around the filing
of patent lawsuits. These estimates suggest that: (i) the costs have grown
rapidly since the mid-1990s, (ii) in technologies such as pharmaceuticals
and chemicals—where patent boundaries are well-defined—the costs
imposed by litigation are small compared to patent rents, but, (iii) in other
technologies, especially information technologies, litigation costs are now
far greater than patent rents. In these industries, patents appear to provide a
net disincentive to innovation.
Much recent research has focused on specific aspects of patents and
innovation, thus helping to fill out a more detailed picture of the innovation
landscape. In addition to estimating the burden of litigation on innovators,
research on patent litigation has also identified factors that affect a firm’s
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
hazard of being sued or of having to go to court to enforce its patents. Other
research, both theoretical and empirical, has sought to better understand the
nature and timing of dispute settlement.
One important area in litigation research has been on patent “nonpracticing
entities” (NPEs), known popularly as patent trolls. These are organizations
that do not develop a technology to be implemented, but, instead, acquire
patents to assert against others in order to collect royalties. The number of
lawsuits involving NPEs has soared in recent years; over 5000 defenses were
mounted in 2011 in lawsuits filed by NPEs according to one source. This
new development has been controversial. The patent system has long had
a variety of nonpracticing agents who have facilitated markets for technology. Indeed, one of the recent advances in research has been recognition that
patents can serve to facilitate markets for technology and these have been
historically important. They allow small inventors who might not have the
resources and skills to bring a product to market to sell or license the technology to a party who can produce and distribute it. However, the recent spate
of NPEs tend to be much more interested in litigating against firms that have
already developed technologies than in licensing a technology to firms that
need it. That is, much of the NPE activity is about licensing freedom from the
threat of a lawsuit rather than licensing an actual technology, with training,
know-how, and so on. Some research shows that NPE litigation appears to do
far more damage to the defendants, who tend to be firms that invest heavily
in technology, than any benefit it provides to the small inventors who create
the patents used in NPE lawsuits.
Another area of study has been on those subject matter areas where the
courts have extended patent coverage during recent years. Before 1994, for
example, software algorithms were seen as unpatentable by themselves.
Court decisions since the mid-1990s have extended patent coverage not only
to software, but also to methods of doing business and mental processes.
Over a quarter of a million patents have now been issued in these new
areas, making for something of a natural economic experiment. Research
has shown that patents in these new areas are much more likely to be
involved in litigation. Indeed, NPE lawsuits are heavily concentrated in
software and business methods. It appears that the high rate of litigation is
not so much because these fields are new to the patent system, but rather
that these subjects of major problems with patent notice. In particular, the
interpretation by the courts of patent boundaries in these fields has been
highly unpredictable, giving rise to opportunistic litigation. Research on
software patents has been influential in legislative proceedings in Europe
and in court decisions in the United States.
Software patents have also featured in another area of research, namely,
patent thickets. Because firms acquired tens of thousands of software patents
Can Public Policy Influence Private Innovation?
7
(being conceptual, they often require very little R&D), economists became
aware that they were using them strategically as a block. A number of theoretical papers have studied the nature of this strategic interaction and a
number of empirical papers have sought to identify the effects of patent thickets on innovation. The findings appear to be double-edged: patent thickets
do deter entry of innovators into new markets, but prospective entrants are
more likely to enter if they acquire patents themselves.
Economists have also studied the operation of the patent office and the
patent examination process. Research on one administrative procedure—a
“post-grant review” proceeding that allows third parties to challenge patents
immediately after they are issued—influenced legislation in the United States
such that the proceeding was included in the patent act passed into law in
2011.
This review has focused on research on patents. Research has also been
done on other sorts of policies intended to foster innovation. For example,
economists have attempted to estimate the effect of R&D tax credits on R&D
spending.
NEW DIRECTIONS
Three sorts of factors influence those areas where future research will be
fertile:
Better, more realistic understanding of how innovation works. For example,
“spillovers” have long been identified as important, but this concept
includes a variety of different sorts of actual behavior. Some new
research has focused on the effect of laws that restrict employee
mobility, one source of spillover. Other research is looking into the free
exchange of knowledge among innovators, something long recognized,
but little explored.
Better data. Patent offices and courts are making more data available to
researchers, opening up new areas of inquiry, for example, better information on patent prosecution and on payment of renewal fees. In addition, a number of surveys have been initiated in recent years, covering
all aspects of the innovation process.
New developments. Nonprofit organizations have proposed using prizes
and similar measures in order to promote the development of pharmaceuticals for neglected diseases. This is promoting research on prizes.
In addition, international trade treaties have resulted in dramatically
changed patent systems in developing countries such as China, also
opening up new avenues for research.
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
REFERENCES
Arrow, K. (1962). Economic welfare and the allocation of resources for invention.
The rate and direction of inventive activity economic and social factors (pp. 609–626).
Cambridge, MA: National Bureau of Economic Research.
Bessen, J., & Meurer, M. J. (2008). Patent failure: How judges, bureaucrats, and lawyers
put innovators at risk. Princeton, NJ: Princeton University Press.
Boldrin, M., & Levine, D. K. (2008). Against intellectual monopoly. Cambridge, England: Cambridge University Press.
Nordhaus, W. D. (1969). Invention, growth, and welfare: A theoretical treatment of technological change. Cambridge, MA: MIT Press.
JAMES BESSEN SHORT BIOGRAPHY
James Bessen is a scholar on the economics of innovation and patents
who has also been a successful innovator and CEO of a software
company. In 1983, Bessen developed the first commercially successful
“what-you-see-is-what-you-get” desktop publishing program, founding
a company that delivered PC-based publishing systems to high end commercial publishers. Since leaving the software business, Mr. Bessen has
done economics research on innovation, including theoretical and empirical
studies of the economics of patents as well as studies of innovation in the
nineteenth century and its effects on skills, wages, and productivity. His
work on patents has influenced policymakers and courts in the United
States, Europe, and Australia. He is the author of Learning by Doing: The Real
Connection between Innovation, Wages, and Wealth (Yale 2015). With Michael
J. Meurer, he wrote Patent Failure: How Judges, Bureaucrats, and Lawyers Put
Innovators at Risk (Princeton 2008). Currently, Mr. Bessen is a Lecturer in Law
at the Boston University School of Law and Fellow at the Berkman Center
on Internet and Society at Harvard.
See also http://www.researchoninnovation.org/bio.php
RELATED ESSAYS
The Public Nature of Private Property (Sociology), Debbie Becher
Technology Diffusion (Economics), Adam B. Jaffe
Knowledge Transfer (Psychology), Timothy J. Nokes-Malach and J. Elizabeth
Richey
Translational Sociology (Sociology), Elaine Wethington