INNOVATION AND INCARCERATION:

AN ECONOMIC ANALYSIS OF
CRIMINAL INTELLECTUAL PROPERTY
LAW

CHRISTOPHER BUCCAFUSCO*

JONATHAN S. MASURt

TABLE OF CONTENTS
I. INTRODUCTION .............................. ...... 276
II. THE ECONOMICS OF INTELLECTUAL PROPERTY AND
CRIMINAL LAW ................................... 280
A. THE ECONOMICS OF INTELLECTUAL PROPERTY ..... ...... 281
B. THE ECONOMICS OF CRIMINAL LAW: INCARCERATION,
DAMAGES, AND DETERRENCE ................ ...... 284
C. PROPERTY (AND INTELLECTUAL PROPERTY) CRIMES .............. 289
III. THE LIMITED CASE FOR CRIMINAL COPYRIGHT
LIABILITY ....................................... 293
A. THE HARM OF COPYRIGHT INFRINGEMENT ........ ..... 294
1. Copyright Infringement and Incentives ........ ..... 294
2. The Efficient Level of Copying ................... 295

* Associate Professor of Law and Co-Director of the Center for Empirical Studies of
Intellectual Property, Chicago-Kent College of Law.
t Deputy Dean and Professor of Law, University of Chicago Law School. The authors wish to
thank Kathy Baker, Stefan Bechtold, Rochelle Dreyfuss, Harry First, Rebecca Giblin, Eric Goldman,
Sarah Harding, Lital Helman, Margot Kaminski, Mark Lemley, Fred von Lohmann, Irina Manta, Lisa
Larrimore Ouellette, Matthew Sag, Pam Samuelson, David Schwartz, Chris Sprigman, Jennifer Urban,
Polk Wagner, and the attendees of the NYU Colloquium on Innovation Policy, the Max Planck-ETH
Zurich Conference on the Economics of IP and Antitrust, and the Chicago-Kent Faculty Workshop for
comments on a previous draft of this Article. The authors are also grateful for excellent research
assistance by Leah Eubanks, Matthew Schock, and Dayron Silverio. Masur would like to thank the
David and Celia Hillard Fund for research support.

275

Electronic copy available at: https://ssrn.com/abstract=2297488

276 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

B. THE ECONOMICS OF DETERRING COPYRIGHT

INFRINGEMENT .................................... 298
1. Self-help and Digital Rights Management..........................298
2. Civil Sanctions and Deterrence ............ ....... 302
a. Comparing Benefits and Harms ................ 302
b. Detection, Enforcement, Punitive Damages, and
Judgment-Proof Infringers .............. ..... 304
3. Secondary Liability............................307
C. THE THEORETICAL CASE FOR CRIMINAL COPYRIGHT
SANCTIONS ................................... 309
1. The Benefits of Criminal Sanctions ........... ..... 309
2. The Costs of Criminal Sanctions ......... ......... 312
D. WHAT KIND OF CRIMINAL LIABILITY COULD BE
DESIRABLE? . .. .. .. .. .. .. .. .. .. .. .. .. .. . .. . . .. 315
E. ASSESSING CURRENT U.S. CRIMINAL COPYRIGHT LAW .......... 317
1. Current Criminal Copyright Law............. ............... 317
2. Criminal IP Enforcement .................. ..... 320
IV. CRIMINAL PATENT LAW? ................... ........ 322
A. THE ECONOMIC HARM FROM PATENT INFRINGEMENT ............ 322
B. LIKELIHOOD OF DETECTION ............................ 324
C. JUDGMENT-PROOF PATENT DEFENDANTS? ......... . .. . .. . . 328
D. OPTIMAL ACTIVITY LEVELS AND INCAPACITATION.................329
V. CONCLUSION.............................. ....... 334

I. INTRODUCTION
The scope and enforcement of intellectual property ("IP") laws are
becoming salient, for the first time, to a wide cohort of U.S. and
international communities. National and international legislation, including
the Stop Online Piracy Act ("SOPA"),' the PROTECT IP Act ("PIPA"), 2
and the Anti-Counterfeiting Trade Agreement ("ACTA"), 3 have generated
protests online and in the streets by people who are concerned about the
expansion of IP rights.4 Common to each of these proposals was an

1. Stop Online Piracy Act, H.R. 3261, 112th Cong. § 103(a)(1)(B) (2011).
2. Preventing Real Online Threats to Economic Creativity and Theft of Intellectual Property
("PROTECT IP") Act of 2011, S. 968, 112th Cong. § 3(a)(1)(B) (2011).
3. Anti-Counterfeiting Trade Agreement, 50 I.L.M. 243, 243-57 (2011).
4. See Michael A. Carrier, The Proposed New Copyright Crime of "Aiding and Abetting,"
OUPBLOG (Oct. 28, 2010, 2:30 PM), http://blog.oup.com/2010/10/copyright-crime/ (describing the
academic protests to ACTA); Mike Masnick, An Updated Analysis: Why SOPA & PIPA Are a Bad
Idea, Dangerous & Unnecessary, TECHDIRT (Jan. 18, 2012, 7:32 AM),

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 277

expansion of the use of criminal sanctions to deter IP violations.5 Many

copyright owners and the associations that represent them support criminal
enforcement of IP rights, including the use of imprisonment, to combat the
threat of increased IP piracy on the internet and throughout a globalized
economy.6 Others, including a heterogeneous coalition of scholars,
activists, and internet-based companies like Google and Wikipedia, fear
that using criminal sanctions to protect IP will expand already overgrown
rights and chill valuable expressive and inventive behavior. 7
There is likely some truth in both of these positions. The internet and
global markets do make infringing IP easier. And with so many more
potential infringers, it may be getting harder to enforce copyright laws.
Accordingly, the deterrent effect of civil sanctions for IP violations may be
weakening. Enhanced criminal sanctions could, however, tilt the balance
between owners and users of P too far toward owners, thereby limiting the
opportunities for creative and innovative developments in the future. They
might also impose unnecessary costs upon the convicted infringers
themselves, channeling resources toward incarceration rather than research
and development. In situations where proposed rules present both costs and
benefits, economic analysis of law can play an important role in
policymaking.8
This Article offers an economic analysis of the use of criminal liability
in two areas of IP: copyright law and patent law.9 Our goal is to analyze the
relative costs and benefits of criminal sanctions for IP violations. Economic

http://www.techdirt.com/articles/20120117/23002717445/updatedanalysis-why-sopa-pipa-are-bad-idea-
dangerous-unnecessary.shtml (explaining past and present criticisms to SOPA and PIPA).
5. Stop Online Privacy Act, H.R. 3261, §§ 102(a)(2), 201(a); PROTECT IP Act, S. 968, § 6(a);
Anti-Counterfeiting Trade Agreement, 50 I.L.M. at 250-53. See also Christophe Geiger, The Anti-
Counterfeiting Trade Agreement and Criminal Enforcement of Intellectual Property: What
Consequencesfor the European Union?, in INTELLECTUAL PROPERTY AT THE CROSSROADS OF TRADE
167, 169-74 (Jan Ros6n ed., 2012) (asserting that ACTA's use of criminal enforcement of intellectual
property law was one of the most controversial aspects of the agreement); Michael A. Carrier, SOPA,
PIPA, ACTA, TPP: An Alphabet Soup ofInnovation-StiflingCopyright Legislation and Agreements, 11
Nw. J. TECH. & INTELL. PROP. 21, 21-24 (2013) (summarizing the legislative changes and sanctions
implemented under SOPA, PIPA, and ACTA).
6. See infra note 105 and accompanying text.
7. See supranote 4 and accompanying text.
8. John Bronsteen, Christopher Buccafusco & Jonathan S. Masur, Well-Being Analysis vs. Cost-
Benefit Analysis, 62 DUKE L.J. 1603, 1605 (2013).
9. This Article does not discuss trademark law, although criminal liability is used to combat
counterfeit goods. The principal reason we have excluded trademark law from this discussion is that the
underlying economic rationale for trademark protection-avoiding consumer confusion-is different
from the underlying economic rationale for copyright and patent law-solving a public goods problem
for informational goods.

Electronic copy available at: https://ssrn.com/abstract=2297488

278 SOUTHERN CALIFORNIA LAW RE VIEW [Vol. 87:275

analysis is particularly appropriate to this discussion for a number of

reasons. Copyright and patent law are widely recognized as resting on
utilitarian foundations of promoting social welfare by incentivizing
investment in informational goods. 10 They do this by providing exclusive
rights to creators of IP that allow them to charge prices for use that are
above marginal cost." Yet IP's commitment to incentives to create must be
balanced by the costs that those rights impose on others who want to use or
further develop the works and inventions that have been created already.
Economic analysis's explicit focus on utilitarian welfare calculus, which
compares the costs and benefits of legal rules, can aid in striking the correct
balance. 12
An explicitly economic focus on criminal sanctions in IP-to the
exclusion of other normative methodologies-is important for an additional
reason. Many IP stakeholders and scholars have operated under the
assumption that there is an economic case for criminal IP sanctions.13
Criminal sanctions for [P infringement are thought to be justified by the
possibility of deterring or incapacitating would-be infringers.14 When
criminal [P sanctions have been criticized, the criticism has usually come
from a normative position outside of economics-moral rights, for
instance." The economic case for criminal sanctions is usually treated as

10. See wILLIAM M. LANDES & RICHARD A. POSNER, THE ECONOMIC STRUCTURE OF
INTELLECTUAL PROPERTY LAW 13 (2003) (stating that a "dynamic benefit of a property right is the
incentive that possession of such a right imparts to invest in the creation or improvement of a
resource"). But see ROBERT P. MERGES, JUSTIFYING INTELLECTUAL PROPERTY 13-15, 31-67 (2011)
(proposing a Lockean foundation for IP protection).
11. LANDES & POSNER, supra note 10, at 20 ("Unless there is power to exclude, the incentive to
create intellectual property in the first place may be impaired. Socially desirable investments
(investments that yield social benefits in excess of their social costs) may be deterred if the creators of
intellectual property cannot recoup their sunk costs.").
12. Bronsteen, Buccafusco & Masur, supranote 8, at 1612.
13. See, e.g., I. Trotter Hardy, Criminal Copyright Infringement, 11 WM. & MARY BILL RTS. J.
305, 305, 314-23 (2002) (suggesting that the basic economic underpinnings of IP law explain the
reasoning behind Congress's increasing criminalization of IP sanctions).
14. Id. at 312.
15. See Miriam Bitton, Rethinking the Anti-Counterfeiting Trade Agreement's Criminal
Copyright Enforcement Measures, 102 J. CRIM. L. & CRIMINOLOGY 67, 78-81 (2012) (explaining that
criminal sanctions might be ineffective at deterring copyright infringement because social norms may
not categorize the behavior as being morally wrong and may view it as being inconsequentially
harmful); Stuart P. Green, Plagiarism,Norms, and the Limits of Theft Law: Some Observations on the
Use of Criminal Sanctions in Enforcing Intellectual Property Rights, 54 HASTINGS L.J. 167, 240-41
(2002) ("From the perspective of criminal law, moreover, words like 'theft' and 'stealing' have
particular expressive and moral resonances that are unlikely to find easy equivalence in the law of
intellectual property."); Irina D. Manta, The Puzzle of Criminal Sanctions for Intellectual Property
Infringement, 24 HARV. J.L. & TECH. 469, 492-512 (2011) (analyzing why "moral, utilitarian, and
public choice rationales" are important factors to consider when determining the effectiveness of

Electronic copy available at: https://ssrn.com/abstract=2297488

20 14] INNOVATION AND INCARCERATION 279

unassailable. 16
In this Article, we mean to assail it. We do not believe that the case for
criminal sanctions in IP is nearly as strong as many have treated it. Indeed,
we will argue that criminal liability may be justified only in one small
corner of IP law, in response to one discrete type of infringement.17 We
believe that our analysis is particularly important because it undermines the
economic case for criminal sanctions at its very foundation. There is no
need to engage in complicated and frequently unresolvable debates about
economic versus moral concerns if both modes of thought militate equally
against criminal IP penalties. Even without venturing into the deep waters
of normative debate, we believe the range of economically justifiable
criminal IP sanctions is quite narrow.
Additionally, as we will explain below, our analysis in this Article
generally accepts the traditional assumptions of law and economics
scholarship regarding IP law. That is, we assume for purposes of this
Article that IP is necessary to solve the public goods problem in
information, that it incentivizes people to create, and that the government
will do an appropriate job of balancing the rights of creators with the rights
of users.18 These assumptions have all been subjected to withering critiques
by scholars, including the authors of this Article.' 9 We believe, however,
that it is important to keep these assumptions intact as an initial matter. If
the case for criminal IP sanctions is weak when the assumptions are
accepted, it will only be weaker once those assumptions have been relaxed.
According to our analysis, there is a limited and tentative case for the
use of criminal liability, including imprisonment and alternative sanctions,
for only some types of copyright infringement-and none at all for patent
infringement. Because of the large scale of some copyright violations and
the high possibility of judgment-proof defendants, civil remedies with

criminal sanctions in IP law); Geraldine Szott Moohr, The Crime of Copyright Infringement: An Inquiry
Based on Morality, Harm, and Criminal Theory, 83 B.U. L. REV. 731, 733 (2003) (arguing that the
moral consensus that would condemn infringement of IP rights "is far from robust and the harm
rationale provides only an equivocal basis for criminalization" of this behavior).
16. See Hardy, supra note 13, at 305, 312 (stating that a basic economic analysis of punishment
as deterrence illustrates why criminal sanctions can be effective when applied to IP law).
17. See infra Part III.
18. See LANDES & POSNER, supra note 10, at 11-36 (outlining the economic theory of property
law and the government's role in the allocation of property rights).
19. See, e.g., Bronsteen, Buccafusco & Masur, supra note 8, at 1603 (arguing that traditional,
cost-benefit economic analysis "creates significant systemic errors"); Christopher Buccafusco & Paul J.
Heald, Do Bad Things Happen When Works Enter the Public Domain?: EmpiricalTests of Copyright
Term Extension, 28 BERKELEY TECH. L.J. 1, 1 (2013) (stating that current research "indicates that the
significant costs of additional copyright protection for already-existing works may not be justified").

Electronic copy available at: https://ssrn.com/abstract=2297488

280 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

punitive damages will create insufficient deterrence.20 In response,

copyright owners will resort to self-help, using digital rights management
("DRM") technologies to prevent unauthorized copying. 21 But DRM tends
to be overbroad and can eliminate much socially valuable copying. Thus,
criminal sanctions could provide a way of deterring harmful copying while
keeping valuable copying free from the restrictions of DRM. Imprisoning a
few people whose behavior is inarguably harmful may create opportunities
for others who are engaging in welfare-enhancing expression. As we
explain, however, it is difficult to predict whether criminal sanctions will
have these intended effects. The behaviors of various parties are difficult to
anticipate and may not coincide with the assumptions of economic
analysis.22
The situation in patent law is different. Patent violations tend to be
easier to detect than copyright violations.23 More importantly, unlike many
copyright infringers, typical patent infringers will have sufficient resources
to satisfy judgments against them. This means that civil remedies should
produce adequate deterrence, and there is little need for the liberty
deprivations associated with criminal sanctions. 24
This Article is organized in three main parts. In Part II, we introduce
the standard economic analysis of IP and criminal law. Part III applies this
analysis to copyright law. We first consider whether any case can be made
for incorporating criminal sanctions into copyright law. After determining
that such a case exists, we then discuss the appropriate scope of criminal
copyright law. Part IV applies economic analysis to the possibility of
criminal patent law. Our analysis suggests that criminal liability is
unnecessary to deter patent violations.

II. THE ECONOMICS OF INTELLECTUAL PROPERTY AND

CRIMINAL LAW
In this part, we present an overview of the economic theories of IP and
criminal punishment. Although the economic theory of IP may be familiar
to many readers, we offer a brief synopsis here in the interest of

20. See infra Part ll.B.2.

21. See infra Part III.B.1.
22. See infra Part II.B.1; Stefan Bechtold, The Present and Future of Digital Rights
Management-Musings on Emerging Legal Problems, in DIGITAL RIGHTS MANAGEMENT:
TECHNOLOGICAL, ECONOMIC, LEGAL AND POLITICAL ASPECTS 597, 597-98 (Eberhard Becker et al.
eds., 2003) (listing numerous criticisms of DRM).
23. See infra Part V.B.
24. See Manta, supra note 15, at 507 n.241 (stating that copyright infringement is "often-
dispersed [and] can be more difficult to detect than" patent infringement).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 28 1

completeness. We then explain why criminal laws, and sanctions such as

incarceration, are often necessary despite the availability of civil damages.
We also expound the economic rationale for the criminalization of property
offenses despite the lack of any obvious economic loss. These explanations
are offered with an eye toward IP crimes, which will be the focus of the
succeeding parts.

A. THE ECONOMICS OF INTELLECTUAL PROPERTY

We offer here a simple snapshot of the economics of intellectual

property. Much important work has been done on this subject, and we will
not attempt to summarize all of it.2 5 For our purposes, a baseline
understanding of the economic theory surrounding IP law is sufficient.
Inventive ideas and creative works are public goods: once they have
been created, they can be shared among many individuals to the benefit of
all of them simultaneously. 26 In this respect, they are also nonrival goods:
multiple people can make use of an idea or a creative work without
depriving one another of the enjoyment and use of that work.2 7 Two firms
can sell DVD players that use the same technology simultaneously, and
multiple people can read War and Peace simultaneously without interfering
with one another. Because intellectual and creative ideas are public goods,
they are liable to be underproduced.2 8 The first individual to think of an
invention or a creative work bears all the costs of creating, while others can
copy the idea or the work and dissipate the creator's advantage. 29 The
threat that the creator will bear all of the costs and only reap a fraction of
the benefits raises the possibility that creators might not think it worth their
time and resources to create in the first place.30

25. See, e.g., LANDES & POSNER, supra note 10, at 11-36 (using the laws that govern physical
property to illustrate the general economic underpinnings of IP law); Mark F. Grady & Jay 1.
Alexander, Patent Law and Rent Dissipation, 78 VA. L. REV. 305, 308-10 (1992) (proposing a rent
dissipation theory of patent law); Edmund W. Kitch, The Nature andFunction of the Patent System, 20
J.L. & ECON. 265, 267-71 (1977) (proposing a patent prospect theory); Robert P. Merges & Richard R.
Nelson, On the Complex Economics of PatentScope, 90 COLUM. L. REV. 839, 868-84 (1990) (offering
a New Institutional Economics approach to patent law).
26. See Mark A. Lemley, Ex Ante Versus Ex Post Justificationsfor Intellectual Property, 71 U.
CHI. L. REV. 129, 129 (2004) ("Ideas are public goods: they can be copied freely and used by anyone
who is aware of them without depriving others of their use.").
27. See Peter Lee, The Accession Insight and PatentInfringement Remedies, 110 MICH. L. REV.
175, 239-40 (2011) (using patents to illustrate the nonrival nature of intellectual property rights).
28. See Lemley, supra note 26, at 129 ("[A]bsent intellectual property protection, most would
prefer to copy rather than create ideas, and inefficiently few new ideas would be created.").
29. See id. at 129-35 (explaining the standard justification for intellectual property).
30. Id. at 129,131.

Electronic copy available at: https://ssrn.com/abstract=2297488

282 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

IP rights exist in order to solve this public goods problem and to

encourage the production of innovative and creative goods.' Property
rights in inventions and creative works allow their creators to capture all of
the rents from the creation of those goods. Without having to fear that their
profits will be dissipated, creators will have greater incentives to innovate
in the first place.3 2 However, these property rights simultaneously allow
their owners to charge monopoly prices.3 3 This leads to deadweight
losses, 34 since some consumers who would consume the goods if they were
priced competitively cannot afford the monopoly prices. 35 IP law is thus
thought to trade off dynamic efficiency-greater innovation and
creativity-against static inefficiency-namely, deadweight losses. 36
Efficiently solving the public goods problem does not, however, mean
that creators and innovators should be given maximal rights in their
contributions. In addition to creating static inefficiencies, strong IP rights
also prevent others from making welfare-enhancing uses of protected
works and inventions.37 Others may wish to create derivative works of a
copyrighted play or new improvements of a patented machine. These
contributions are more expensive when the underlying contribution is
protected by an IP right that must be licensed. In addition, there is a
concern that some socially valuable behavior will not be generated at all
because the IP owner may refuse to license the right.3 9 For example,
copyright owners may be unwilling to license parodies and critical reviews
of their works even though these uses seem to enhance social welfare.

31. Id. at 131.

32. See id. at 131-33 (explaining how the "incentive theory of intellectual property" solves the
public goods problem in ideas).
33. Id. at 131.
34. Ian Ayres & Gideon Parchomovsky, Tradable Patent Rights, 60 STAN. L. REv. 863, 867
(2007).
35. See Lemley, supra note 26, at 131 (stating that due to "supracompetitive" pricing, "some
people who would be willing to pay more than the marginal cost of a copy of the idea will not be able to
access it").
36. Ayres & Parchomovsky, supra note 34, at 867 (explaining that patents "involve[] a
fundamental tradeoff between dynamic and static efficiency: patents spur innovation but only at the
cost" of higher prices for current consumers).
37. See, e.g., Lee, supra note 27, at 180 (using patents to demonstrate this concept in stating that
"a patentee enjoys exclusive rights over any technology that infringes her claims, even if that
technology substantially improves on her original creation").
38. See LANDES & POSNER, supra note 10, at 15 (stating that judicial decisions are not
copyrighted for this reason, because "the transaction costs of obtaining licenses by the myriad of
lawyers, litigants, judges, and law professors . . .would be immense").
39. See Lee, supra note 27, at 180 (stating that "transaction costs and strategic behavior," such as
the refusal to license an IP right, may prevent parties from realizing the true potential of an improved
work).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 283

For these reasons, the duration and scope of IP rights should be

limited. IP law's goal is to balance the incentives provided to creators with
the interests of the public and of subsequent creators. 40 The current
generation of creators and inventors should only be given enough rights in
their contributions to ensure that they can recapture sufficient profits to
make creating them worthwhile. Giving additional rights to current creators
produces inefficiencies both through deadweight losses and by increasing
the costs of creating and inventing for subsequent generations.
One final point bears emphasis. Many scholars believe that intellectual
property rights do not encourage substantially greater invention or
creativity, at least in some contexts.4 1 In addition, many believe that current
IP laws are not efficiently balanced, because they provide excessive
incentives to the current generation of creators at the expense of the public
and subsequent creators. 42 We count ourselves among this group to some
extent. We do not adopt that view here, however, because it would render
the inquiry we seek to undertake far too easy. If intellectual property rights
are not necessary to induce innovation or creativity, then there is no
economic justification for such rights and they should simply be
eliminated.4 3 Or, in lieu of eliminating IP rights, there should be no
remedy--civil or criminal-for infringement.4 4 If creators and inventors
are already getting inappropriately strong rights in their works, enhancing
their rights through greater enforcement will only further tip the balance in
their favor.4 5
We do not intend to minimize these concerns-as noted, we agree
with them to some extent. Nonetheless, we think that it is important to take
the basic- economic assumptions seriously as an initial matter. First, if

40. See LANDES & POSNER, supra note 10, at 11 (referring to this underlying goal of IP law as a
basic "tradeoff between 'incentive' and 'access").
41. See DAN L. BURK & MARK A. LEMLEY, THE PATENT CRISIS AND HOW THE COURTS CAN
SOLVE IT 137-65 (2009) (detailing the many industries in which patents may do more harm than good
and describing the few instances in which patent rights may be socially valuable).
42. See, e.g., Grady & Alexander, supra note 25, at 306, 311-13 (stating that some advocates
believe that "[w]hen the patent system induces a race for invention rights, competing inventors work
too quickly and dissipate much of the social benefit that would otherwise come from their efforts").
43. See Michael Abramowicz & John F. Duffy, The Inducement StandardofPatentability, 120
YALE L.J. 1590, 1594-98 (2011) (explaining the economic value of patents and proposing when they
should and should not be granted).
44. See LANDES & POSNER, supra note 10, at 13 (stating that incentive for creative innovation
and improvement is a "dynamic benefit of a property right," suggesting that if this benefit does not in
fact exist, then there would be less reason to impose sanctions for property right violations).
45. See Manta, supra note 15, at 498-99 (suggesting that this type of imbalance might result in
"the prosecution and conviction of relatively harmless infringers and perhaps even innocent parties,"
and that it could also lead to "overdeterrence of innovation").

Electronic copy available at: https://ssrn.com/abstract=2297488

284 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

economic analysis does not support criminal sanctions, then we need not
even bother analyzing those sanctions in light of even more hostile
normative frameworks. Second, it may be wrong to doubt the importance
of IP rights in inducing research, development, and creativity. If that is the
case, it will be necessary to consider what remedies should be applied in
the event of infringement. Accordingly, we will proceed as if patent and
copyright law play some role in incentivizing innovation (as they are meant
to), and as if Congress has properly balanced the interests of creators and
the public. On this account, IP infringement will harm both the IP owner
and society at large (resulting in the loss of some benefits from future
creativity), though it is difficult to know to what degree. 46

B. THE ECONOMICS OF CRIMINAL LAW: INCARCERATION, DAMAGES, AND

DETERRENCE

The economic goal of any set of laws is to produce the greatest

possible benefits at the lowest possible cost-in other words, to generate
the maximum achievable net benefits. 47 Criminal and tort laws are of
course directed at deterring costly behavior, and so the "benefits" produced
by these laws come in the form of crimes and accidents that have been
avoided. 48 The law's objective is to deter as much harm as possible while
imposing the fewest costs. 49 From an economic perspective, criminal law
and punitive incarceration at first seem to present puzzles. Incarceration
generates very substantial economic costs, costs that are imposed upon the
prisoner, hisso friends and family, and the government that is charged with
imprisoning him.51 The principal economic objective of punishment is
deterrence, and tort law should be able to efficiently deter misconduct

46. See LANDES & POSNER, supra note 10, at 21-24 (describing the cost-benefit tradeoff of
property rights and noting that "the high social costs of intellectual property rights create uncertainty as
to whether on balance such rights are, from an overall social standpoint, cost-justified at all").
47. See A. MITCHELL POLINSKY, AN INTRODUCTION TO LAW AND ECONOMICS 7-10 (1983)
(outlining a utilitarian theory of law centered on the maximization of net benefits).
48. See RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW 167-71 (7th ed. 2007) (describing
the Learned Hand theory of optimal tort liability).
49. See Stephen P. Garvey, Can Shaming Punishments Educate?, 65 U. CHI. L. REv. 733, 775
(1998) ("Utilitarianism, for example, encourages architects of punishment to get the most deterrence at
the lowest possible cost.").
50. Because the vast majority of incarcerated prisoners are male, we will employ the male
pronoun here although it is not standard convention.
51. See John Bronsteen, Christopher Buccafusco & Jonathan Masur, Happiness andPunishment,
76 U. CHI. L. REv. 1037, 1039-55 (2009) (describing the hedonic costs of incarceration); Garvey, supra
note 49, at 743-44 (stating that prison is "an expensive public problem" in that it is "both costly and
overused").

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 285

through monetary sanctions alone.52 Moreover, while tort lawsuits might

involve significant transaction costs, the civil fines themselves are not
economic costs, but rather transfer payments from one party to another that
involve little or no efficiency loss.s On this account, prison sentences
would seem unnecessarily wasteful.54 However, the economic justification
for criminal law lies with the possibility that defendants will be insolvent or
otherwise unable to satisfy a civil judgment.ss
Consider a straightforward case of assault. An assault victim might
suffer damages in the tens or even hundreds of thousands of dollars in
medical bills, lost wages, and pain and suffering. The perpetrator of an
assault may not have assets worth nearly that much, and thus it might be
impossible to collect the necessary measure of damages. (This is the reason
that many states require automobile drivers to have insurance-in order to
guard against otherwise judgment-proof defendants. 56) If the amount of
damages is set at a figure that exceeds the perpetrator's ability to pay, then
it will be impossible to fully deter the perpetrator. Thus, the judgment-
proof defendant cannot be effectively deterred."7 This problem is much
more severe for more serious harms, such as murder, and is accentuated by

52. See POLINKSY, supra note 47, at 73-84 (assuming that if individuals are risk neutral, an
economically "efficient system of law enforcement is one in which the fine is as large as possible-
equal to the wealth of the individuals whose behavior is being controlled").
53. See Charles Silver, "We're Scared to Death": Class Certfication and Blackmail, 78 N.Y.U.
L. REv. 1357, 1418 (2003) (stating that in both conventional lawsuits and class actions, "the payment is
a transfer made to satisfy a demand, and thus is not an economic loss").
54. The contrast we draw in this Article is between monetary penalties and incarceration. We
realize that this contrast does not map perfectly onto the line between criminal and civil penalties:
criminal fines are quite common, and nonpunitive civil incarceration exists as well. POSNER, supra note
48, at 215. In some cases there will be important differences between civil tort damages and criminal
fines. However, the distinction between monetary penalties and incarceration is the starkest and most
important contrast between criminal and civil law, and it is the one that matters most from an economic
perspective because the two types of penalties are experienced so differently by offenders. See id.
(describing the differences between civil and criminal money damages, as well as acknowledging
criminal law's use of "nonpecuniary sanction[s]"). Accordingly, we focus our attention upon this
critical distinction while acknowledging the nuance involved in both civil and criminal sanctions.
Henceforth, when we refer to "criminal" penalties we mean incarceration, and when we refer to "civil"
penalties we mean monetary sanctions. In this part we will also refer to some alternative criminal
sanctions that might deter infringement, as we do in Part M with respect to copyright.
55. See Richard A. Posner, An Economic Theory of the Criminal Law, 85 COLUM. L. REv. 1193,
1195 (1985) ("The optimal damages that would be required for deterrence would so frequently exceed
the offender's ability to pay that public enforcement and nonmonetary sanctions such as imprisonment
are required.").
56. See POLINSKY, supra note 47, at 65-71 (outlining the economic reasoning as to why
insurance reduces this risk of insolvency).
57. Posner, supra note 55, at 1204-05.

Electronic copy available at: https://ssrn.com/abstract=2297488

286 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

the fact that not every harm is detected and punished.ss

In order to achieve optimal deterrence, standard economic theory
suggests that the expected penalty imposed upon the defendant must be
equivalent to the defendant's likely gain from the crime.59 The expected
penalty is the likelihood that a defendant will be detected, apprehended,
and fined, multiplied by the magnitude of the sanction that will be
imposed.60 Imagine that a hypothetical perpetrator expects to obtain the
equivalent of $10,000 in benefits from an assault, and the perpetrator has
only a 40 percent chance of being caught. If the only potential liability for
the assault was $10,000, the cost to the offender is only $4000 ($10,000 x
0.40). This is insufficient to deter the behavior. To deter the assault, the law
would have to set the sanction at $25,000 ($10,000 + 0.40), an amount that
many defendants charged with assault may not possess. 61
As the foregoing discussion indicates, criminal damages should be set
not by the harm done to the victim-as are tort damages-but by the
benefit from the crime to the perpetrator.6 2 Because the majority of torts
involve unintentional harms in the course of otherwise valuable activities,
tort damages must be equal to the expected harm to the victim in order to
induce potential tortfeasors to take efficient precautions. 63 On the other
hand, many crimes involve intentional conduct that bears no resemblance
to valuable activity, and the issue is not the level of precaution taken but
the criminal's conscious decision to commit the crime in the first
instance. 64 Accordingly, the penalty need only be high enough to place the

58. See POLINSKY, supra note 47, at 83 (clarifying that in discussing an economically efficient
law enforcement system, the author made an important and "implicit" assumption that there were no
mistakes in the system).
59. Gary S. Becker, Crime and Punishment:An Economic Approach, 76 J. POL. EcON. 169, 181-
84(1968).
60. See Emmett H. Miller IH, FederalSentencing Guidelinesfor OrganizationalDefendants, 46
VAND. L. REv. 197, 204 (1993) ("Thus, under both approaches, the expected gain or harm represents
the actual gain or harm resulting from the offense multiplied by the probability of detection and
conviction.").
61. See Posner, supra note 55, at 1204-05 (suggesting that for this reason, "criminal law is
designed primarily for the nonaffluent").
62. See Jeremy Bentham, Principles of Penal Law, in I THE WORKS OF JEREMY BENTHAM 367,
396 (John Bowring ed., 1843) ("If the apparent magnitude, or rather value of that pain be greater than
the apparent magnitude or value of the pleasure or good he expects to be the consequence of the act, he
will be absolutely prevented from performing it.").
63. See POSNER, supra note 48 at 167-71 (explaining the Learned Hand formula of liability for
negligence).
64. See Posner, supra note 55, at 1194 (stating that criminal law has a "pervasive
emphasis ... on the mental state of the accused, and .. . on the moral character rather than the
consequence of behavior").

Electronic copy available at: https://ssrn.com/abstract=2297488

2014]1 INNOVATION AND INCARCERATION 287

criminal in a better situation if he refrains from committing the crime than

if he commits it.65 In many cases, this means that surprisingly low criminal
penalties may be appropriate. 66 For instance, the appropriate penalty for
some murders might be only a few years in prison, despite the terrible harm
involved in murder. The reason is that many murderers might gain very
little benefit from their crimes, and murderers are apprehended and
prosecuted at relatively high rates." Accordingly, in economic terms, it is
not necessary that the punishment match the crime. 68
This analysis is muddied somewhat when one considers optimal
activity and precaution levels. One goal of tort law is to avoid
overdeterrence: it is inefficient for potential tortfeasors to take precautions
that are more expensive than the damages that they might cause. 69 For
instance, imagine that an automobile driver has a 10 percent chance of
causing an accident that will result in $10,000 in damages. Suppose,
however, that tort damages are set at $50,000, rather than $10,000, in the
event of an accident. The driver would now have an incentive to take a
precaution that costs $5000 ($50,000 x 0.10) despite the fact that the
expected cost of the damage that this driver will cause is only $1000
($10,000 x 0.10). Excessive tort damages lead to inefficient behavior. 70
This analysis applies straightforwardly to accidents and cases of
negligence, but its application to intentional harms is not nearly so clear.
Society is willing to accept that some accidents may occur in the course of
lawful driving because it would be too expensive to take every precaution
necessary to prevent them entirely. 7 1 Accordingly, the optimal level of
accidents is nonzero. 72 Yet it may be that there is no "precaution" against
intentional murder that is not worth taking. The harm caused by murder is
enormous, and the offender can very cheaply avoid causing that harm

65. See id. at 1205 (stating that criminal sanctions must be designed "so that the criminal is made
worse off by his act").
66. See Garvey, supra note 49, at 738 (proposing that "shaming penalties" might be an effective
new punishment model).
67. See Posner, supra note 55, at 1197 ("[I]t is unlikely that every disutile experienced by the
wretched victim confers an equal and opposite utile on the offender.").
68. See id. at 1207 ("[M]ore serious crimes might not always be punishable by more severe
penalties than less serious ones.").
69. See Kyle D. Logue, CoordinatingSanctions in Tort, 31 CARDOZO L. REv. 2313, 2324 (2010)
(explaining that tort law seeks to avoid overdeterrence).
70. See POSNER, supra note 48, at 167-71 (describing a general theory of tort law and efficient
behavior).
71. Logue, supranote 69, at 2324.
72. Jennifer H. Arlen & William J. Carney, Vicarious Liabilityfor Fraudon Securities Markets:
Theory and Evidence, 1992 U. ILL. L. REv. 691, 718 ("In standard torts cases, the optimal level of
accidents is not zero.").

Electronic copy available at: https://ssrn.com/abstract=2297488

288 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

simply by deciding not to intentionally kill the victim. Thus, the magnitude
of the harm will always exceed the costs of avoiding it.73 Accordingly, the
optimal level of intentional murder might be exactly zero. If this is the case,
damages for murder and other serious intentional crimes should be
infinite. 74 Indeed, the strongest economic argument typically made against
very high penalties for serious intentional crimes is the need for marginal
deterrence, not the desire to avoid overdeterrence. 75
Marginal deterrence is the idea that more serious crimes should be
punished more harshly than less serious crimes in order to deter criminals
who might cause a less serious harm from instead causing a more serious
harm. 76 For instance, if assault and murder were both punished by life in
prison, a criminal who has committed assault would have no incentive not
to commit murder (in an attempt to flee, for instance). He would face no
greater penalty for doing so, and he could increase his chances of escape. 77
The necessity of marginal deterrence means that the penalty for very
serious crimes, such as aggravated assault (the optimal level of which is
zero), should not, in fact, be as serious as the penalty for murder.
Marginal deterrence also provides some of the reason why
policymakers have turned to expensive criminal sanctions like
imprisonment, rather than relying solely on monetary penalties.7 1 If a
criminal is contemplating causing $100,000 worth of harm but only has the
means to pay a $50,000 judgment, then there is no incentive for him not to
cause $200,000 worth of harm-he will reap the extra benefit without
being subject to any additional risk. The threat of a prison sentence will
take from him a resource of which he has a significant stock-his liberty-
thus reintroducing a deterrent effect that will prevent the greater harm. 79
Because prison sentences can be scaled in terms of their length and

73. See Eyal Zamir & Barak Medina, Law, Morality, and Economics: Integrating Moral
Constraintswith Economic Analysis ofLaw, 96 CALF. L. REv. 323, 354 (2008) (stating that, according
to "commonsense morality," the value of human life may be "lexically superior to pecuniary losses").
74. See Posner,supra note 55, at 1202 ("If the average person ... were asked how much money
he would demand to surrender his life on the spot, his answer would be that no finite offer would be
high enough, since he would get no utility out of the money.").
75. Id. at 1207.
76. Zamir & Medina, supra note 73, at 379 n.21 1 (noting that economic theory requires marginal
deterrence if a criminal is to have an incentive not to commit a more serious crime).
77. See Posner, supra note 55, at 1207 (making a similar analogy in stating that "[i]f robbery is
punished as severely as murder, the robber might as well kill his victim to eliminate a witness").
78. See id. at 1205-08 (describing optimal criminal penalties and the role of marginal
deterrence).
79. See id at 1204 (stating that imprisonment is one way of addressing the "infeasibility of
primary reliance on fines to deter" crime).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 289

severity, the law can create marginal deterrence along a much wider
spectrum of harm-causing behavior.
Finally, even if every potential defendant were properly capitalized
and able to pay the full measure of damages required after causing harm,
there might nonetheless be some individuals who could not be deterred.
This could be due to mental instability, a tendency to succumb to fits of
rage, overoptimism (regarding the chances of being caught), or any number
of other factors that cause the individual to fall short of perfect rational
calculation.so Incapacitation is the proper remedy with respect to such
individuals.81 And of course, incapacitation can only be achieved via
incarceration, not monetary damages. 82
Accordingly, carceral criminal sanctions are justifiable from an
economic perspective because of difficulties in properly deterring potential
offenders with civil sanctions, most frequently because offenders are
insolvent or unable to pay the full judgments against them.8 3 This central
point will inform much of the discussion that follows.

C. PROPERTY (AND INTELLECTUAL PROPERTY) CRIMES

The foregoing discussion focused on harms involving injury to a

person. This is appropriate, as these are the most serious types of crime and
present the clearest case for criminal incarceration. 84 Yet our true focus is
on property crimes. This is an Article about intellectual property, and
property crimes such as theft require a somewhat different economic
justification. Unlike harms involving personal injuries, there is no obvious
economic harm when someone commits a property crime." Destroying a
home or wrecking a car would certainly create economic harm, but if a
thief merely steals a piece of property, nothing of value is destroyed-the

80. See Jonathan S. Masur, Richard H. McAdams & Thomas J. Miles, For GeneralDeterrence,
in CRIMINAL LAW CONVERSATIONS 122, 122-24 (Paul H. Robinson et al. eds., 2009) ("[N]ot all people
are fully rational, informed, and capable of self-control.").
81. See SANFORD H. KADISH ET AL., CRIMINAL LAW AND ITS PROCESSES: CASES AND
MATERIALS 120-24 (9th ed. 2012) (explaining a general theory of incapacitation).
82. See Masur, McAdams & Miles, supra note 80, at 123 ("[P]recisely when individuals prove
themselves undeterrable is when we may think it worthwhile to incarcerate them in order to incapacitate
them.").
83. Posner, supra note 55, at 1195.
84. See id. at 1202 ("This nonlinearity suggests why tort law may be adequate for many small
risks of death (for example, the risk of being killed in an automobile accident caused by negligence),
but not for the large risks that are created by crimes of violence.").
85. See LANDES & POSNER, supra note 10, at 47 (implying that if a thief is not a potential
purchaser of the property, then the "owner is not deprived of any revenue," and thus does not lose
anything from the unauthorized taking of the property).

Electronic copy available at: https://ssrn.com/abstract=2297488

290 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

property still exists. It is possible that the theft is just a transfer from one
user to another, not an efficiency loss.86 If there is no economic harm from
theft, then there is no economic reason to punish it with criminal sanctions.
There are, however, powerful arguments that theft creates economic
harm.8 7 The first is that theft may result in an item being transferred from a
higher-value user to a lower-value user, which decreases overall welfare.88
If an automobile is worth $10,000 to its owner and only $5000 to the thief
who steals it, then the theft destroys $5000 in value just by virtue of the
fact that the automobile is transferred from the owner to the thief. By
blocking these types of involuntary transactions, the law forces individuals
into the world of voluntary transactions through which property finds its
way to the highest-value user.89 Because inventions and creative works are
nonrival, however, this argument does not apply in the context of
intellectual property. "Theft" of intellectual property does not deprive the
original owner of its use, so there can be no fear that a higher-value owner
has been deprived.90 We must, accordingly, look elsewhere for economic
justifications for (intellectual) property crimes.
Theft may also create economic harm by reducing incentives to
efficiently maintain or improve property. 9 ' If property owners have to fear
that their property might imminently be stolen, they will have no reason to
invest in improving that property, even when it is efficient to do so. 92
Individuals might even shy away from acquiring property in the first place,
even if they would be higher-value users than the current owners. 93 This
argument has particular resonance for IP law. If someone "steals"
intellectual property-that is, infringes it without paying royalties to the
owner-the owner will be able to earn only competitive returns, rather than
monopoly profits. 94 As noted above, economic theory justifies the ability to

86. See Lee, supra note 27, at 226 ("Of course, the Coase theorem holds that in the absence of
transaction costs, initial allocations do not matter, and costless transactions will allow resources to flow
to their highest-valued use.").
87. See, e.g., Posner, supra note 55, at 1196 ("[lIt is inefficient to allow pure coercive transfers
of wealth-'pure' implying that the transfer is not an incident of a productive act.").
88. See id. ("[Property is] less valuable in an economic sense in [the] hands [of a thief.]").
89. Id at 1195-96.
90. See Lee, supra note 27, at 192-93 (arguing that the nonrival nature of intellectual property
"cuts directly against exclusive rights").
91. See Manta, supra note 15, at 476-77 ("The most obvious harm[] caused by violations such as
theft . .. [is] the potential disincentive for... owners to engage in future productive and socially
beneficial endeavors because their resulting profits are at risk of being stolen.").
92. See id. (describing the incentives of property owners in a world where theft is rampant).
93. See LANDES & POSNER, supra note 10, at 18 (stating that this is a possibility when the total
cost of protection "exceed[s] the benefits of propertization").
94. Lemley, supra note 26, at 131.

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 291

earn monopoly profits on copyrighted works and patented inventions

because competitive returns will be inadequate to incentivize IP owners to
create and commercialize their ideas. 95 Accordingly, IP "theft" could
reduce incentives to invent or create IP, just as theft of real or chattel
property might reduce incentives to acquire or improve that property.96
This would frustrate the economic rationales underlying intellectual
property.
The third argument that theft causes economic harm relates to the
steps that owners will take to protect their property from theft. Imagine a
country that decriminalized theft (or that had unreliable police, prosecutors,
and courts). Without the state to protect their property, citizens of that
country would understand that their property was at risk of being taken if
they did not protect it themselves.97 Consequently, they might build high
walls around their homes, install security systems, hire armed guards, train
vicious guard dogs, and so forth." These are socially wasteful activities.
They produce little social value, and they exist only to prevent the transfer
of property from one user to another. 99 Economists thus view the criminal
prohibition on theft as a less costly means of protecting property. 0 Rather
than giving private property owners incentives to invest in wasteful self-
help, economists favor centralizing the task of protecting property in the
government, a body that can take advantage of economies of scale. 101

95. See id (explaining how the "incentive theory of intellectual property" solves the public
goods problem in ideas).
96. Id. at 131-33.
97. See George P. Fletcher, The Metamorphosis of Larceny, 89 HARv. L. REv. 469, 474 (1976)
("(T]he thief upset[s] the social order not only by threatening property, but by violating the general
sense of security and well-being of the community; .. .theft [is] feared as a socially unnerving event.").
98. Posner, supra note 55, at 1196 ("[I]f [a thief is] allowed to take the car [of his neighbor,
his] ... neighbor will have an incentive to expend resources on preventing it from being taken, and
these expenditures considered as a whole, yield no social product."). See also ROBERT P. MERGES,
PETER S. MENELL & MARK A. LEMLEY, INTELLECTUAL PROPERTY IN THE NEW TECHNOLOGICAL AGE
34 (6th ed. 2012) (making this point about trade secrets).
99. See LANDES & POSNER, supra note 10, at 119 (defining this as "inefficient self-protection").
100. See William E. Nelson, Emerging Notions of Modern Criminal Law in the Revolutionary
Era: An Historical Perspective, 42 N.Y.U. L. REV. 450, 452-66 (1967) (examining criminal law's
historical development from an instrument enforcing the community's moral values to one focused on
"punish[ing] and deter[ring] ... attacks on property"); Posner, supra note 55, at 1196 (arguing that the
role of the criminal law is discouraging market bypassing).
101. See, e.g., A. Mitchell Polinsky, Private Versus Public Enforcement of Fines, 9 J. LEGAL
STUD. 105, 107 (1980) ("[T]he benefits from coordinating enforcement-for example, avoiding
duplication of investigative effort and exploiting economies of scale in information processing-are
obtained under public enforcement and monopolistic enforcement, but not under competitive
enforcement."). Cf Amitai Aviram, Allocating Regulatory Resources, 37 J. CORP. L. 739, 765 (2012)
("[P]rivate actors are often poor enforcers because they lack the economies of scale (such as
investigative expertise and litigation experience) that come with repeated enforcement, and because

Electronic copy available at: https://ssrn.com/abstract=2297488

292 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

This rationale also has relevance for IP law. Copyright owners have a
number of available avenues of self-help, including (perhaps most
importantly) digital rights management ("DRM").' 02 For patent owners, the
mechanism for engaging in self-help is to shift from patents to trade
secrets-literally locking innovation secrets away behind closed doors
instead of disclosing them to the public.o 3 We will discuss these
mechanisms in much greater detail below. For now, the general point is
that intellectual property owners, like other property owners, may use
expensive self-help as an alternative if legal prohibitions on theft appear
inadequate.
There is one final distinctive feature of property crimes worth
addressing, and it is the likelihood that a thief (as opposed to, say, a
murderer) would be able to pay off a civil judgment. At first glance it might
appear that property criminals are less likely to be judgment-proof than
criminals who have injured a person. One reason is that property crimes
often cause less harm than personal injury crimes. Another reason is that a
thief will often have one asset that can be used to pay the judgment: the
stolen property itself. This is not a panacea; as we explained above, the
proper measure of damages for a crime might well be many times greater
than the actual harm caused if the probability of detection and punishment
is low. But it nevertheless places property criminals on a somewhat
different footing than criminals who injure other people.
We have identified a number of factors relevant to the question of
whether criminal penalties should be applied to prevent some type of
undesirable activity. In the interest of clarity, we summarize these factors
here:
(1) What is the magnitude of the economic harm from the activity, and
what are the benefits to the infringer?
(2) What is the probability that violations will be detected and
prosecuted?
(3) Will the defendant likely have sufficient funds to pay a judgment?
(4) Is the economically efficient activity level zero or greater than
zero?
(5) Will it be desirable to incapacitate offenders?

most people have fewer financial means to pursue enforcement than public actors.").
102. See infra Part III.
103. J. Jonas Anderson, Secret Inventions, 26 BERKELEY TECH. L.J. 917, 923 (2011) (describing
the tradeoffs between trade secret and patent protection, and the decisions faced by innovators).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 293

In the parts that follow, we consider the issue of criminal sanctions for
copyright and patent infringement through the economic lens of these
factors.

III. THE LIMITED CASE FOR CRIMINAL COPYRIGHT LIABILITY

Over the course of the last century, the availability of criminal
sanctions for copyright infringement has expanded dramatically. 104 Driven
in large part by lobbying from the Recording Industry Association of
America ("RIAA") and the Motion Picture Association of America
("MPAA"), criminal copyright law has grown to cover a wider variety of
content and behavior, and the penalties for violations have multiplied.'os
According to these advocacy groups, criminal sanctions, including
substantial imprisonment terms, are needed to deter the rising tide of
copyright "theft."l 0 6
This part addresses those arguments from the perspective of an
economic analysis of law. We suggest that, according to the standard
economic account of criminal law, there might be narrow situations in
which criminal sanctions could be theoretically justified for deterring
copyright violations and improving overall social welfare. Below, we
describe when criminal copyright liability might be valuable and when it
might not, and we assess current U.S. copyright law in light of our findings.

104. See Hardy, supra note 13, at 315-16. Congress enacted the first criminal copyright statute in
1897. It applied only to the unauthorized performance of plays and music. See Act of Jan. 6, 1897, ch.
4, 29 Stat. 481. In 1909, Congress expanded liability to cover all copyrightable works. See Copyright
Act of 1909, ch. 320, 35 Stat. 1075. Over the last quarter of the twentieth century, Congress further
expanded criminal liability and enhanced the magnitude of criminal penalties. Copyright infringement
could result in felony convictions, defendants could be found guilty even in the absence of a purpose for
financial gain, and maximum sentence lengths grew. Piracy and Counterfeiting Amendments Act of
1982, Pub. L. No. 97-180, 96 Stat. 91 (creating felony liability for copyright violations); No Electronic
Theft ("NET') Act of 1997, Pub. L. No. 105-147, 111 Stat. 2678 (amending the criminal provisions to
allow criminal conviction in the absence of a showing that the defendant was motivated by commercial
purpose or financial gain); Sentencing Reform Act of 1984, Pub. L. No. 98-473, 98 Stat. 1987
(extending maximum sentence length for criminal copyright violation from two to five years of
imprisonment).
105. See Brian P. Heneghan, The NET Act, Fair Use, and Willfulness-Is Congress Making a
Scarecrowof the Law?, 1 J. HIGH TECH. L. 27, 28 (2002) ("The Net Act enjoys strong support from the
Software Publishers Association, U.S. Copyright Office, the Department of Justice, Adobe Software,
Microsoft Corporation, the Recording Industry of America, and the Motion Picture Association of
America.").
106. See generally WILLIAM PATRY, MORAL PANICS AND THE COPYRIGHT WARS (2009)
(discussing property metaphors in copyright).

Electronic copy available at: https://ssrn.com/abstract=2297488

294 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

A. THE HARM OF COPYRIGHT INFRINGEMENT

Economic analysis suggests that criminal law can be valuable in
situations where people must be deterred from conduct that reduces overall
social welfare. 107 When people steal, rape, and kill, the reduction in social
welfare and the need for deterrence is obvious. In Part II, we explained how
and when criminal sanctions can be used to reduce occurrences of these
behaviors. But in what way is illegally downloading a song or streaming a
bootlegged signal of a soccer match equivalent to these heinous acts? How
is watching a prerelease version of the latest blockbuster movie a crime?
From an economic perspective, the answer depends on the extent to which
copyright infringement causes harm that cannot be deterred by other
means. Copyright infringement can enter the domain of criminal law when
other mechanisms for deterring socially harmful conduct, including civil
sanctions and self-help remedies, are either too costly or unsuccessful.108
1. Copyright Infringement and Incentives
In order to determine the extent to which deterrence, whether criminal
or otherwise, is appropriate, we must first understand the harms caused by
the unauthorized copying of creative works. Although copyright is a form
of intellectual property, and although many people refer to copyright
infringement as "theft,"l 09 infringement is importantly different from theft
of real or personal property.1 10 With copyright infringement, the original
owner is not deprived of the use of the infringed content."I When someone
illegally photocopies a novel, the copyright owner is not prevented from
reading the novel herself. Yet the issue is not with the copyright owner's

107. See supra Part I.

108. Geraldine Szott Moohr, Defining Overcriminalization Through Cost-Benefit Analysis: The
Example of Criminal Copyright Laws, 54 AM. U. L. REV. 783, 788-89 (2004) ("Whether criminal
penalties are an appropriate way to deal with copyright infringement depends on whether the strategy
confers a net social benefit, which is found in intellectual property policy.").
109. During Congressional debates on the NET Act, Senator Patrick Leahy declared, "Just as we
will not tolerate the theft of software, CD's, books, or movie cassettes from a store, so will we not
tolerate the stealing of intellectual property over the Internet." 143 CONG. REC. S12689, S12691 (daily
ed. Nov. 13, 1997) (statement of Sen. Leahy). See also Green, supranote 15, at 240 (noting that many
recently enacted criminal IF laws have in common "at least a literal commitment to the idea that
intellectual property of various sorts might be subject to 'theft' or to being 'stolen').
110. The Supreme Court has recognized the difference between copyright infringement and theft
of real or personal property. In Dowling v. United States, 473 U.S. 207 (1985), the Court rejected the
government's interpretation of the National Stolen Property Act that "would make theft, conversion, or
fraud equivalent to wrongful appropriation of statutorily protected rights in copyright." Id at 216.
111. Dowling, 473 U.S. at 216 ("The copyright owner ... holds no ordinary chattel. A copyright,
like other intellectual property, comprises a series of carefully defined and carefully delimited interests
to which the law affords correspondingly exact protections.").

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNO VA TION AND INCARCERATION 295

use of the novel but with her incentives to write it in the first place. The
harm of unauthorized copying arises from the public goods problem in IP
and the law's solution to that problem.11 2 When someone downloads a song
for free off the internet rather than paying for it on iTunes, the creator's ex
ante incentives to create diminish. For every person who opts out of an
available market transaction for the song, the creator will tend to invest that
much less in the quality and quantity of new songs she produces. And if
enough people download songs for free rather than paying for them, the
creator may cease making music altogether. To ensure that the appropriate
level of incentives remains, people must be deterred from making
unauthorized copies of works.

2. The Efficient Level of Copying

Like theft and murder, unauthorized copying of expressive works can
create harm and reduce overall social welfare. But unlike theft and murder,
where the optimal amount of those behaviors is at or very near zero, the
optimal amount of unauthorized copying may be significantly greater than
zero. Certain kinds of copying may be beneficial, and deterring them would
be harmful."'
First, as we noted in Part II.A, the ability to charge monopoly prices
above marginal cost creates deadweight losses that can be inefficient. For
example, imagine that the marginal cost of producing a copy of a movie is
$2. Because the copyright owner has an exclusive right to copy the movie,
she can charge $10 for a copy of the movie. The owner's ability to sell the
movie at $10 to the many people who value having a copy of the movie at
$10 or greater provides the owner with the opportunity to recoup her
investment. But what if some person values the movie at $5? He is
unwilling to pay the price for the movie, and without the ability for the
owner to price discriminate,114 the difference between what he is willing to
pay and the marginal cost of the copy represents inefficient, deadweight

112. See supra Part II.A.

113. William M. Landes & Richard A. Posner, An Economic Analysis of Copyright Law, 18 J.
LEGAL STUD. 325, 335 (1989) ("Some copyright protection is necessary to generate the incentives to
incur the costs of creating easily copied works, but too much protection can raise the costs of creation
for subsequent authors to the point where those authors cannot cover them even though they have
complete copyright protection for their own originality.").
114. Price discrimination can occur in some situations involving copyrightable works. For
example, copyright owners charge higher prices for hardback editions and first-run movies because
some people value being able to consume the work rapidly. See LANDES & POSNER, supra note 10, at
39 (explaining that because of the ability to price discriminate, prices charged for first-run movies are
generally higher than prices in the aftermarket). Price discrimination is also possible through the use of
DRM technologies, which we discuss in Part III.B.1.

Electronic copy available at: https://ssrn.com/abstract=2297488

296 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

loss.115
The situation here is different from that of theft of real property. As
William Landes and Richard Posner explain,
If a thief steals a Rolls-Royce from the dealer's lot, it is no consolation to
the dealer that the thief was not a potential customer because he could
not afford to pay the dealer's price; the theft deprived the dealer of the
opportunity to sell the Rolls to someone else. But when the purchaser of
a software program makes a copy for someone else, he does not reduce
the number of copies in the software producer's inventory. If the
someone else was not a potential purchaser from the producer, the
producer loses nothing from the unauthorized copying." 11 6
Thus, if the owner is able to sufficiently recoup her investment in the
movie, deterring people who value the work at above marginal cost but
below the copyright owner's price may be unnecessary or even
counterproductive.
Second, and more importantly, some unauthorized uses of a work
might be efficient and socially valuable but nonetheless fail to arise due to
market failures. Although economic analysis generally trusts markets to
efficiently allocate goods to valuable uses,117 some users of copyrighted
works engage in a variety of behaviors that are unlikely to result in market
transactions. For example, authors of copyrighted works are unlikely to
license use of their works to others for criticism and parody, yet these are
generally believed to be socially beneficial activities. 1 8 In other situations,
transaction costs may prevent otherwise efficient bargains from arising. 119

115. See Eric Goldman, A Road to No Warez: The No Electronic Theft Act and Criminal
Copyright Infringement, 82 OR. L. REv. 369, 426-31 (2003) (suggesting that Congress should redefine
the scope of the NET Act to properly criminalize copyright infringement); Stan J. Liebowitz & Stephen
Margolis, Seventeen Famous Economists Weigh in on Copyright: The Role of Theory, Empirics, and
Network Effects, 18 HARv. J.L. & TECH. 435, 440-41 (2005) ("Copyright engenders a deadweight loss
as a by-product of the incentives to create that it provides. A system of private ownership providing the
incentive for creation cannot give a reward to the creator without also having an apparent deadweight
loss in the consumption market.").
116. LANDES & POSNER, supra note 10, at 47 (giving the weak demand for AIDS drugs in Africa
as an example of how piracy does not necessarily reduce an intellectual property owner's revenue).
117. See R. H. Coase, The Problem ofSocial Cost, 3 J.L. & EcoN. 1, 15 (1960) (explaining how,
in a world without transaction costs, voluntary market transactions will result in entitlements being
owned by their highest valuing user).
118. See Alfred C. Yen, When Authors Won't Sell: Parody, FairUse, and Efficiency in Copyright
Law, 62 U. COLO. L. REv. 79, 85-90 (1991) (explaining that parody is consistent with copyright's goal
of promoting social welfare).
119. See Christopher Buccafusco & Christopher Jon Sprigman, The Creativity Effect, 78 U. CHI.
L. REv. 31, 33-34 (2011) (describing inefficiencies in IP markets that may arise from biased
assessments of a work's value); Christopher Buccafusco & Christopher Sprigman, Valuing Intellectual
Property:An Experiment, 96 CORNELL L. REv. 1, 32 (2010) (same).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 297

Someone wishing to transform a work or use it in a way that creates

significant value that she is unable to capture may be unwilling to pay the
necessary licensing fee. In these situations, the "fair use" provisions of the
U.S. Copyright Act allow others the right to freely use copyrighted works,
because those uses are considered to be welfare enhancing and because
market transactions are unlikely to provide them in sufficient quantity.120
Deterring the behaviors discussed above is not valuable. 2 1 They tend
to create significant benefits with relatively little harm to creators'
incentives. Uses of a copyrighted work for criticism, parody, or education
make many people's lives better off, and the inability to charge for them is
unlikely to deter creation. 122 Although they may marginally decrease
incentives to create, the optimal amount of these behaviors is decidedly
above zero.
Other uses of a work, however, will tend to generate little benefit but
have significant social costs. When people download copies of movies or
songs for free when they would have been willing to pay the copyright
owner's price, they are creating inefficiencies. Assuming that the law has
properly balanced the costs and benefits of copyright protection,123 this
diminution in the copyright owner's ability to recoup her investment costs
is harmful. 124

120. See Wendy J. Gordon, Fair Use as Market Failure: A Structuraland Economic Analysis of
the Betamax Case and Its Predecessors, 82 COLUM. L. REV. 1600, 1604-05 (1982) (suggesting that
economic principles of market failure should be used to disseminate authored works and avoid the
erosion of incentives for authors).
121. Landes & Posner, supra note 113, at 340 ("An increase in copyright protection is likely to
reduce the welfare benefits (consumer plus producer surplus) generated by a given work-assuming it
will be created.").
122. See Amy Kapczynski, The Cost of Price: Why and How to Get Beyond Intellectual Property
Internalism, 59 UCLA L. Rev. 970, 997-98 (2012) (explaining that the field of IP should be analyzed
using values of efficiency, as well as justice and privacy).
123. We believe that Congress has done a poor job of balancing the incentives of authors with the
rights of the public. Copyright terms are far too long, and the scope of copyright holders' rights is far
too broad. See Christopher Buccafisco & Paul J. Heald, Do Bad Things Happen When Works Enter the
Public Domain?: Empirical Tests of Copyright Term Extension, 28 BERKELEY TECH. L.J. 1, 19-29
(2013) (presenting empirical data which suggest that the works suffer little economic harm when they
enter the public domain). Cf Liebowitz & Margolis, supra note 115, at 437, 457 (providing a
"counterweight" to the economists' amicus brief against copyright term extensions in Eldred v.
Ashcroft).
124. See Stan J. Liebowitz, File Sharing: CreativeDestruction or Just PlainDestruction?, 49 J.L.
& ECON. 1, 24 (2006) (showing that recent empirical research on the effects of music and movie
downloading on sales has suggested significant decreases in sales as a result of downloading). We wish
to emphasize again, however, that a mere decrease in sales does not alone show that overall social
welfare has been diminished. See Felix Oberholzer-Gee & Koleman Strumpf, The Effect of File Sharing
on Record Sales: An Empirical Analysis, 115 J. POL. ECoN. 1, 40 (2007) (discussing why file sharing
may have increased social welfare in the music industry); Manta, supra note 15, at 475 (comparing the

Electronic copy available at: https://ssrn.com/abstract=2297488

298 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

B. THE ECONOMICS OF DETERRING COPYRIGHT INFRINGEMENT

The unauthorized copying of expressive works, at least in some
instances, can harm social welfare by undermining the incentives of authors
to invest in their creations. Deterring these cases of unauthorized copying
would be valuable. Economic analysis describes the various tools that
might be brought to bear to do so, including private self-help and public
law solutions.125 This section explores the effectiveness and desirability of
self-help and civil law sanctions in the copyright context.

1. Self-help and Digital Rights Management

Like owners of real property who attempt to prevent others from
trespassing on their land by building fences, copyright owners who seek to
deter unauthorized copying can attempt to limit access or use of their
works. Traditionally, this was a difficult task for copyright owners. Once a
book was put into circulation, there was little the copyright owner could do
to prevent others from copying it, lending it, or performing it publicly.
Copyright law and the civil sanctions it creates are a response to this
problem. With the rise of digital copies of expressive works that are
embodied in computer software and hardware, however, copyright owners
can create technological limits on the things that users can do with copies.
Known as digital rights management ("DRM"), these technological
measures can limit the ability of users to make copies of a work, to transfer
the work between different devices, or to modify the work.' 26
DRM can be very attractive to copyright owners because it can
provide considerable control over the behavior of users. 127 For example, if
the copyright owner of a movie makes copies of the movie available in
video cassettes, a purchaser of a video cassette could easily copy the movie
for a number of her friends with only limited loss in quality. Some of these
friends might have been potential purchasers of the movie, and their ability

harm of copyright infringement to vandalism).

125. Owners of informational goods do not necessarily require copyright protection. They have a
variety of techniques available to them. Hardy describes four different types of protection against
unauthorized copying: "l) entitlement-like protection; 2) contract-like protection; 3) state-of-the-art
limitations; and 4) special purpose technical limitations." Trotter Hardy, Property (and Copyright) in
Cyberspace, 1996 U. CHI. LEG. F. 217, 223. Our analysis will primarily focus on the first and fourth
techniques that Hardy mentions.
126. Bechtold, supra note 22, at 597.
127. Dan L. Burk, Anticircumvention Misuse, 50 UCLA L. REV. 1095, 1100 (2003) ("Copyright
holders might prefer a world in which the rights granted under statute or asserted via license became
self-enforcing. Something close to this can be achieved through the employment of technological
devices accompanying copies of a work as they are distributed." (footnote omitted)).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 299

to get the free copy could harm the copyright owner. Now, however, the
copyright owner can release the movie on a DVD that is coded with DRM
software that prevents the purchaser from making copies. 128 The same is
potentially true for purchasers of e-books, CDs, software, and other digital
content that is downloaded online.
Although DRM seems like an efficient self-help measure to prevent
unauthorized copying, its use is highly controversial,' 29 and its overall
effect on social welfare is probably negative. DRM's principal economic
benefit is that it can be relatively cheaply created and implemented. Coding
DRM software is not difficult and, once created, including the DRM code
in copies of a work is effectively costless for the producer.13 0 Accordingly,
it is unlike the relatively costly self-help measures that property owners
must create and maintain to prevent trespass or theft.'"' To prevent
someone from stealing a book from a store, the owner must pay security
personnel or install other physical detection devices in each copy. To
prevent someone from making unauthorized copies of an e-book, the
producer simply needs to include a string of code in the e-book digital
file.1 32
Despite its ease and low cost, DRM might be an inefficient solution to
the problem of unauthorized copying of expressive works.133 First, users of
many kinds of products object to the limitations that DRM establishes.
Consumers of music, movies, and software have developed norms about
the kinds of uses that should be allowed, and they balk when content
providers attempt to restrict those uses. Recently, in response to consumer
complaints, Apple's iTunes service and other online music distributors
have dramatically scaled back their use of DRM. Similarly, consumers

128. Id. at 1100-01.

129. See Bechtold, supra note 22, at 598 (highlighting several problems that contribute to the
controversial nature of DRM).
130. Of course, consumers may prefer to own copies of works that are not encrypted with DRM,
so the amount they are willing the pay for DRM-encrypted copies will be less. Consumer resistance to
DRM likely had some effect on Apple's decision to release music on iTunes in "DRM-free" versions.
Yet even these versions are not truly DRM free as they only allow purchasers to play songs on a limited
number of devices.
131. See supra Part II.B.
132. See Bechtold, supra note 22, at 598 (listing technologies that may be used to establish a
secure distribution channel for digital content, including encryption and tamper-resistant hardware).
133. Julie E. Cohen, Lochner in Cyberspace: The New Economic Orthodoxy of "Rights
Management, " 97 MICH. L. REV. 462, 471 (1998) ("Digital technologies allow more effective fencing
of intellectual property, and thus cure some of the market failure problems associated with creative and
informational works-although ... they have the potential to create market failures of a different
sort."); Burk, supra note 127, at 1097 ("[T]he anticircumvention right is being, and will continue to be,
abused.").

Electronic copy available at: https://ssrn.com/abstract=2297488

300 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

have expressed dismay at Microsoft's announcement that it will include

highly restrictive DRM provisions in the release of its new Xbox game
console.134 Although DRM still plays a major role in attempting to protect
against unauthorized uses of movies, e-books, and computer software, there
is reason to think that negative consumer responses will increasingly limit
its effectiveness,
In addition, DRM has proven less valuable than was hoped, because
the code can be circumvented. Programmers who desire to use a work in a
way that is forbidden by DRM can "hack" the code and nullify its
effectiveness."' As with real or personal property, this leads to an
inefficient "arms race" between content producers and hackers as they each
expend costly resources attempting to alternately protect and break DRM
code. Although initially creating and installing the DRM is inexpensive,
creating multiple iterations of the code to outpace hackers can get very
expensive. In the United States, the law has attempted to avert this arms
race by prohibiting the circumvention of technological barriers to use. The
Digital Millennium Copyright Act ("DMCA") prohibits the use or
distribution of circumvention measures and subjects violators to civil and
criminal penalties.1 36
While the DMCA minimizes the risk of an inefficient arms race
between content providers and hackers, DRM-based self-help creates other
significant inefficiencies. As noted above, not all unauthorized copying of
expressive works is harmful to social welfare. Many unauthorized uses of
copyrighted works, especially those classified as fair uses under the
copyright act, significantly enhance social welfare. As yet, and perhaps
inevitably, DRM technology is unable to distinguish between infringing
and noninfringing uses of protected works.' 3 ' DRM affects the professor
who wants to excerpt a portion of a book for educational purposes in class
just as much as it does the e-book purchaser who wants to make copies of
the file for all of his friends. While the latter use is welfare diminishing, the

134. Anthony Tosie, Nintendo Disagrees with Microsoft's DRM Policies, NEOWIN (June 13,
2013), http://www.neowin.net/news/nintendo-disagrees-with-microsofts-drm-policies.
135. Burk, supra note 127, at 1102 ("The drawback to reliance primarily upon technological
controls is that technically sophisticated users may find ways to circumvent or disable the control
system, and may even assist unsophisticated users in doing so.... A skilled user may be able to 'hack
around' the controls built into technological content systems.").
136. 17 U.S.C. § 1201(a) (2012). It is important to note that violation of the DMCA
anticircumvention provisions is not considered violation of copyright. DMCA violations create
independent civil and criminal liability. Hardy, supranote 13, at 322.
137. Bechtold, supra note 22, at 604 (discussing the possibility of creating DRM technologies that
can distinguish between appropriate and inappropriate uses of a protected work).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] ]
INNOVATION AND INCARCERATION 301

former is welfare enhancing, but both are prohibited by DRM.

Users who wish to make socially valuable fair uses of a DRM-
protected work can attempt to circumvent the technological protection, but
they do so at considerable risk. First, although the Librarian of Congress
exempts certain uses of copyrighted works from the anticircumvention
provisions of the DMCA,"' U.S. courts have not agreed about whether
every fair use of a copyrighted work is exempt.139 Accordingly, unless
users fall within one of the express exemptions (which appear to change
annually), they risk being subject to heavy civil and criminal fines and
possibly imprisonment. Second, even if circumvention of DRM for fair use
of a work were unambiguously acceptable, the contours of the fair use
doctrine are notoriously fuzzy. Whether a given behavior is copyright
infringement or fair use depends on a difficult-to-predict balancing of
multiple factors. Risk averse parties whose conduct lies on the boundaries
of fair use may elect to avoid engaging in potentially valuable behavior to
avoid liability under the DMCA.1 40 It is unlikely that DRM software itself
will develop to the point at which software can validly distinguish between
fair and unfair uses of a work given the complex, fact-intensive nature of
the distinction.
Importantly, variations in individuals' willingness and ability to
circumvent DRM creates a perverse effect on the kinds of copying that
does and does not get prevented by DRM. Law-abiding individuals who
wish to make fair or personal uses of copyrighted content will generally
also be those who are unable or unwilling to circumvent DRM technology,
while those individuals who engage in large scale direct copying of works
for economic or reputational reasons will tend to be able and willing to
hack DRM code. This means that DRM tends to have its greatest deterrent
effect on socially valuable conduct and only limited effect on harmful
conduct, thereby shifting the costs generated by harmful copiers to
beneficial copiers.

138. Burk, supra note 127, at 1104.

139. Court opinions have differed on the relationship between DMCA and fair use. Compare
Storage Tech. Corp. v. Custom Hardware Eng'g & Consulting, Inc., 421 F.3d 1307, 1318-19 (Fed. Cir.
2005) (holding that a copyright holder must show a connection to copyright infringement in order to
succeed in a claim under the DMCA), and Chamberlain Grp., Inc. v. Skylink Techs., Inc., 381 F.3d
1178, 1202-03 (Fed. Cir. 2004) (holding that distribution of a circumvention device-in that case a
garage door opener-did not violate the anticircumvention provisions because its use did not lead to
any copyright violation), with Universal City Studios Inc. v. Reimerdes, 111 F. Supp. 2d 294, 322
(S.D.N.Y. 2000) ("If Congress had meant the fair use defense to apply to such actions, it would have
said so.").
140. James Gibson, Risk Aversion and Rights Accretion in Intellectual PropertyLaw, 116 YALE
L.J. 882, 887-906 (2007).

Electronic copy available at: https://ssrn.com/abstract=2297488

302 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

A final, and significant, concern about the use of DRM self-help is

that it can be used to protect content that does not enjoy copyright
protection.' 4 1 For example, DRM can be applied to works whose copyright
terms have expired and should be freely available in the public domain.
DRM might also be inappropriately applied to content that fails to meet the
requirements of copyrightability, including nonoriginal data.142 The
balance copyright law creates between creator incentives and public uses is
threatened if content providers respond to the risk of unauthorized copying
by protecting these works with DRM.
It is difficult to judge the net welfare effects of DRM self-help. In its
favor are low cost and ease of use, but set against these are the significant
chilling effects that DRM has on valuable behaviors. DRM prevents many
important uses of expressive works, and the penalties attached to DRM
circumvention likely diminish behaviors that are at the core of human well-
being. Given the risks of DRM, it is appropriate to consider other methods
for deterring harmful unauthorized copying.

2. Civil Sanctions and Deterrence

The standard economic response to costly self-help is the provision of
civil sanctions with penalties sufficient to deter harmful behavior.14 3
According to economic theory, people will be deterred from engaging in
harmful activities if the law establishes penalties that equal or exceed the
magnitude of the benefits violators can hope to achieve from their conduct.
Copyright law does precisely this. Infringement of a copyright owner's
exclusive rights subjects the violator to civil liability that is intended to
deter such violations. For a number of reasons, however, the effectiveness
of civil remedies in certain situations is doubtful.
a. Comparing Benefits and Harms
Economic theory predicts that the law can deter undesirable conduct,
including copyright infringement, by creating penalties that exceed the
benefits that can be obtained from engaging in the conduct. For example, if
someone is contemplating either purchasing or illegally downloading for
free $100 worth of music, the existence of a fine of $100 or more for the
illegal behavior will result in her choosing to purchase the music rather

141. See Burk, supra note 127, at 1102 ("Indeed, content owners may prefer to rely on
anticircumvention laws to prohibit tampering with the technological controls, leaving the technology to
prohibit whichever uses the content owner unilaterally chooses, rather than relying on copyright law to
prohibit certain statutorily determined uses of the work.").
142. See id.
143. See supra Part II.B.

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 303

than violate the law. The benefits of violating the law to the would-be
infringer are smaller than the costs of doing so, and a rational person would
be deterred from downloading the music.
Typically, economic analysis assumes that the magnitude of the
penalty sufficient to achieve deterrence is marginally higher than the value
of the harm done to the victim. Thus, if the penalty for stealing a $10,000
car is $10,001, rational people will prefer to purchase the car rather than
steal it. The benefits to the violator are generally considered to be
equivalent to the harms of the victim (plus any additional social harm). In
the context of copyright infringement, however, this may not be the case.
First, as mentioned above, much copyright infringement is the result of
people who value the work at a price between the marginal cost of
producing it and the monopoly price charged by the copyright owner. 14 If,
for example, someone illegally downloads ten albums that cost $10 each,
but if she only values those albums at $5 each, setting the penalty for
infringement at $100 will lead to overdeterrence of behavior that is not
costly.
More difficult still is the situation involving the large category of
copyright infringement that is mostly divorced from economic
considerations. Much large-scale copyright infringement on the internet
does not appear to be motivated by a desire to obtain works for free. 145
Many hackers and members of "warez" communities derive benefits from
copyright infringement from the reputational effects of successfully
circumventing DRM and providing copies of works for free on the
internet. 146 Comparing the magnitude of the benefits these hackers
experience with the harms to copyright owners is extremely difficult. On
one hand, the market-based values that hackers experience are likely much
smaller than the net value of the infringements that result from the hackers'
conduct. On the other hand, the emotional and reputational benefits that

144. KANTER MEDIA, ONLINE COPYRIGHT INFRINGEMENT TRACKER WAVE 3, 26 (2013) (finding
that approximately 16 percent of survey respondents list "I can't afford to pay" as a reason for copyright
infringement), available at http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/online-
copyright/w3/wave3.pdf.
145. See Goldman, supra note 115, at 405-10 (discussing how the NET Act has been ineffective
because infringers are not motivated by a desire to obtain works for free).
146. Id. at 370 (defining a "warez trader" as "an individual who copies and distributes computer
software simply for self-aggrandizement-the reputation, the thrill, the 'fun' of having the latest
programs or the biggest 'library' of 'warez' titles" (quoting Copyright Piracy, and HR. 2265, the No
Electronic Theft (Net) Act: HearingBefore the Subcomm. on Courts & Intellectual Prop., 105th Cong.
99 (1997) (statement of Sandra Sellers, Vice President, Software Publishers Association) (internal
quotation marks omitted)).

Electronic copy available at: https://ssrn.com/abstract=2297488

304 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

they experience may be very great and thus difficult to deter.147

Accordingly, setting the appropriate magnitude of civil sanctions necessary
for deterrence of copyright violation will be challenging.
It is worth noting, in addition, that to the extent that some copyright
infringers are motivated by political rather than economic concerns, they
may be even more difficult to deter. According to one study, 9 percent of
people who obtain some content through illegal online channels and 16
percent of people who obtain all content through such channels report that
nothing would make them stop infringing.' 4 8 If this is true, efforts at
deterrence are useless, and such people should be incapacitated from
further infringement (at least if doing so is less expensive than the value of
the harm they are causing).
b. Detection, Enforcement, Punitive Damages, and Judgment-Proof
Infringers
By far the most problematic issue with the use of civil remedies to
deter copyright infringement involves the difficulty of detecting and
litigating infringing behavior. As we described in Part II, setting the
magnitude of the penalty equal to a potential lawbreaker's benefits will
only achieve deterrence when the probability of a violation being detected
is 100 percent. If the penalty equals the lawbreaker's benefits and if
detection and enforcement are imperfect, the expected benefits of violating
the law will exceed the expected harms. In order to solve this problem,
economic theory suggests that the law rely on punitive (or augmented)149
damages to create a damages multiplier that will account for imperfect
detection.' Copyright law, with the availability of statutory damages,' 5'

147. Lydia Pallas Loren, Digitization, Commodification, Criminalization: The Evolution of

Criminal Copyright Infringement and the Importance of the Willfulness Requirement, 77 WASH. U. L.
Q. 835, 855 (1999) ("[T]he damage done by non-commercially motivated infringers can equal and
sometimes surpass the damage done by infringers engaged in reproduction and distribution for financial
gain."); Andrea Wechsler, CriminalEnforcement oflntellectualProperty Law: An Economic Approach,
in CRIMINAL ENFORCEMENT OF INTELLECTUAL PROPERTY: A HANDBOOK OF CONTEMPORARY
RESEARCH 128, 143 (Cristophe Geiger ed., 2012) ("[S]elected cases of criminal enforcement have
shown that infringers-such as the world's most resilient bittorrent site, Pirate Bay-with a policy
agenda beyond economic costs and benefits might still not be deterred . . .
148. KANTAR MEDIA, supranote 144, at 27.
149. Thomas C. Galligan, Jr., Augmented Awards: The Efficient Evolution ofPunitive Damages,
51 LA. L. REV. 3, 68 (1990).
150. A. Mitchell Polinsky & Steven Shavell, Punitive Damages: An Economic Analysis, Ill
HARV. L. REv. 869, 874 (1998) ("Punitive damages ordinarily should be awarded if, and only if, an
injurer has a chance of escaping liability for the harm he causes." (emphasis omitted)).
151. 17 U.S.C. § 504(c) (2012). A successful plaintiff in a civil copyright suit can recover actual
damages or statutory damages up to $30,000 for any one work. Id. § 504(c)(1). In addition, if the
violation was willful, the plaintiff can receive an additional $750 to $150,000 per work. Id. § 504(c)(2).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 305

provides something similar to this, but for a number of reasons it is

unlikely to create sufficient deterrence.1 52
Deterring copyright infringement is different in the digital age.1 53
When making an unauthorized copy of a work required expensive
professional machinery to print books or copy film strips, detecting large
scale infringement was relatively easy. 154 Now, millions of people own
devices that can instantly create and distribute millions of infringing copies
around the world. Successfully policing all of these people is impossible,
and the likelihood that any individual infringement will be detected is
infinitesimal.15 5 According to one study of online content consumers, only
about 10 percent of respondents indicate that one of the reasons they use
paid content services is the fear of getting caught.1 56 Even large scale
infringements of multiple works are unlikely to be detected at levels
significantly above zero. In addition, even if copyright owners are able to
detect the existence of an infringement, connecting it to a specific person
can be incredibly difficult. Potential infringers are spread out around the
world and hidden behind internet protocol addresses that mask their
identities. Thus, although in some ways detecting the existence of
infringement might be easier on the internet than it was in the days of tape
recorders, the sheer magnitude of infringement makes litigation to enforce
copyrights cost prohibitive. '

152. H.R. Rep. No. 106-216, § 2 (1999) (describing congressional findings that copyright
damages "are too low to deter individuals from trademark counterfeiting and copyright piracy").
153. Wechsler, supra note 147, at 138 ("It is, in particular, IP crimes in a digital environment
where infringers are intentionally evasive that are most difficult to detect-a problem that is
corroborated by the increasing sophistication of organized crime.").
154. We should note that there are some copyrightable works the infringement of which may not
be very difficult to detect. For example, infringement of architectural works will likely be relatively
easy to detect. And while infringement of sculptural works may currently be easy to detect, the rise of
3D printing may make it easier to copy and share sculptural works online. Michael Weinberg, It Will Be
Awesome If They Don't Screw It up: 3D Printing,Intellectual Property, and the Fight over the Next
Great Disruptive Technology, PUB. KNOWLEDGE (Nov. 2010), http://www.publicknowledge.org/it-will-
be-awesome-if-they-dont-screw-it-up.
155. Hardy, supra note 13, at 313 ("In many cases of copyright infringement today, the chance
that an infringer will be caught is substantially less than fifty percent. For many individuals using the
Internet, for example, the chance of being caught for occasionally downloading a copyrighted song, or
uploading a copyrighted piece of software, is almost zero.").
156. KANTAR MEDIA, supra note 144, at 25.
157. Mark A. Lemley & R. Anthony Reese, Reducing Digital Copyright Infringement Without
Restricting Innovation, 56 STAN. L. REv. 1345, 1350 (2004) ("It is not currently cost-effective for
copyright owners to sue individual infringers, because there are tens of millions of them, because
lawsuits are expensive, and because many infringers would only be liable for (or able to pay) minimal
damages.") As Lemley and Reese note, when it comes to the economics of deterring copyright
infringement, "In applying the deterrence model to the digital environment, likelihood of enforcement
substitutes for likelihood of detection." Id. at 1392.

Electronic copy available at: https://ssrn.com/abstract=2297488

306 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

From the perspective of economic analysis, the damages multiplier

that would be necessary to raise expected penalties from infringement
above expected benefits will be very large. If only one-tenth of one percent
of copyright infringements are detected (probably a generous figure),"' the
damages multiplier is 1000. This means that if an infringer causes $200
worth of harm, she will be responsible for $200,000 in damages.
Copyright infringement damages of this magnitude create two
significant problems. First, statutory damages awards of this magnitude
may run afoul of due process. In recent cases before the Firstl 59 and
Eighth 60 Circuits, copyright infringement defendants have challenged
large awards of statutory damages as violations of due process. These types
of challenges have not yet succeeded, but it is not hard to imagine massive
statutory damages awards running afoul of due process if they are "so
severe and oppressive as to be wholly disproportioned to the offense and
obviously unreasonable."' 61 Yet with the low levels of detection of
copyright infringement, the damages multiplier sufficient to create
deterrence might often result in awards that raise constitutional questions.
Second, damage awards of this magnitude will tend not to have the
appropriate effect on marginal deterrence because many defendants will be
judgment proof for these amounts. 162 In the past, significant resources were
necessary to engage in large-scale copyright infringement.1 63 The
commercial enterprises that were capable of producing many copies of
bootlegged recordings or unauthorized books generally would have been
able to pay significant infringement damages. Now, when an eighteen-year-
old college student can infringe thousands, or even hundreds of thousands
of dollars' worth of copyrights, the likelihood that defendants will have the
resources necessary to compensate plaintiffs is low.164 Since the magnitude

158. Tom Tyler lists the objective probabilities of being caught, convicted, and imprisoned for
burglary and larceny at one percent. Tom R. Tyler, Compliance with Intellectual Property Laws: A
PsychologicalPerspective, 29 N.Y.U. J. INT'L L. & POL. 219, 222 (1997).
159. Sony BMG Music Entm't v. Tenenbaum, 660 F.3d 487, 489 (1st Cir. 2011).
160. Capitol Records, Inc. v. Thomas-Rasset, 692 F.3d 899, 907 (8th Cir. 2012).
161. Id. (quoting St. Louis, I. M. & S. Ry. Co. v. Williams, 251 U.S. 63, 67 (1919)) (internal
quotation marks omitted).
162. According to a survey of online content users in the United Kingdom, people who engage in
any illegal copyright infringement online are more likely to be of a lower socioeconomic status than are
people who engage in no online copyright infringement. KANTAR MEDIA, supranote 144, at 16.
163. See ADRIAN JOHNS, PIRACY: THE INTELLECTUAL PROPERTY WARS FROM GUTENBERG TO
GATES 145-78 (2009) (discussing how resource-sharing partnerships were often required to carry out
large-scale infringement projects in eighteenth-century Europe).
164. Manta, supra note 15, at 503 ("Copyright infringers also often do not have 'deep pockets,' so
it is difficult for owners to recover much in civil litigation compared to the amounts spent on attorneys'
fees and other litigation costs.").

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 307

of copyright liability will often exceed infringers' ability to pay,' 65

impecunious infringers will not be subject to significant deterrence. If a
potential infringer is unable to pay $2,000,000 worth of damages, there is
nothing deterring him from causing $3,000,000 worth of harm.
Thus, for certain classes of copyright infringers-private actors with
limited resources engaged in large-scale infringement-the threat of civil
sanctions may be insufficient to establish deterrence. These actors will tend
not to have the capital available to meet the kinds of high magnitude
damage awards that low levels of detection necessitate.
3. Secondary Liability
Although individual infringers may not have the resources necessary
to satisfy large civil damages awards, many of them use intermediaries to
distribute unauthorized copies of protected works. By suing these
intermediaries, including file-sharing websites and internet service
providers ("ISPs"), can copyright holders efficiently deter infringement?' 6 6
In theory, subjecting intermediaries to secondary liability can lower the
costs of detecting and suing individual infringers. It also attaches liability
to parties that are less likely to be judgment proof and creates incentives for
intermediaries to monitor the behavior of their users. 167 Copyright law's
secondary liability doctrines of contributory infringement and vicarious
liability enable such lawsuits, but as with all such strategies, we must
determine whether they generate more benefits than costs.'6 8 We suspect
that enhanced use of secondary liability may not deter enough harmful
behavior to offset its potential effects on legitimate behavior.

165. Consider the sanctions in the Tenenbaum and Thomas-Rasset cases mentioned. See supra
notes 159-60 and accompanying text. Damages awards in the hundreds of thousands of dollars might
well have exceeded the defendants' resources. Tenenbaum, 660 F.3d at 490-91, 515; Thomas-Rasset,
692 F.3d at 910.
166. See Lital Heiman, Pull Too Hard and the Rope May Break: On the Secondary Liability of
Technology Providers for Copyright Infringement, 19 TEX. INTELL. PROP. L.J. 111, 113 (2010)
("Secondary liability of technology providers is designed to promote effective and efficient copyright
enforcement through three main objectives. First, it provides a cost-effective litigation mechanism
compared to the alternative path of suing countless direct infringers. Second, it attempts to generate
adequate compensation for plaintiffs via the deep pockets of technology companies. Third, it positions
technology providers as gatekeepers that can hinder infringement by a copyright-friendly design and
utilize their service fees to both distribute revenues to copyright owners and discourage infringement."
(footnote omitted)).
167. Id. at 145.
168. See Douglas Lichtman & William Landes, Indirect Liabilityfor Copyright Infringement: An
Economic Perspective, 16 HARV. J.L. & TECH. 395, 396-99 (2003) (noting that tracking down third
parties who may have been indirectly involved in copyright infringement and filing lawsuits against
them may be too costly for copyright holders to pursue).

Electronic copy available at: https://ssrn.com/abstract=2297488

308 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

First, although file-sharing websites, like Napster and Grokster, and

content hosting platforms, like YouTube, have been largely responsible for
the expansion in unauthorized copying in the digital age,169 much copyright
infringement still occurs without intermediaries. Parties who produce
bootlegged copies of CDs and DVDs to sell on street corners and over the
internet typically do not rely on intermediaries to distribute content. When
they do, the intermediary's relationship to the infringing behavior will often
be too insignificant to trigger secondary liability.' 70 In these circumstances,
secondary liability simply does not reach the infringing content.
Second, efforts to deter many intermediaries would fall prey to the
same deterrence problems that arise for individual infringers; that is, they
are undercapitalized relative to the magnitudes of judgments and are often
run by groups with political rather than economic motivations that make
them hard to deter.17 ' The cost of the infrastructure necessary to run a file-
sharing website consisting of servers and other hardware, while not
inconsiderable, is still relatively small compared to the size of a judgment
that could be levied against it. The value of the infringing content shared on
platforms like BitTorrent, even ignoring the availability of statutory
damages, likely dwarfs that of the resources used to make it available.
Further, even after websites like the Pirate Bay are closed through legal
action, infringing behavior is simply displaced to other sites, usually
instantaneously.1 72
Finally, secondary liability doctrines, which make intermediaries
responsible for the behavior of their users, could chill valuable speech.173
Although much of the behavior taking place on BitTorrent websites
involves socially harmful unauthorized copying, there is considerable
activity on websites like YouTube that is socially beneficial. To the extent
that intermediaries are subject to liability for copyright violations of their

169. See Helman, supra note 166, at 125-31 (enumerating examples of lawsuits brought against
file-sharing websites despite these websites never having engaged in any infringement directly).
170. Contributory infringement requires that the defendant "knowingly" induces, causes, or
otherwise materially contributes to the infringing conduct of another. Gershwin Publ'g Corp. v.
Columbia Artists Mgmt., Inc., 443 F.2d 1159, 1162 (2d Cir. 1971). Vicarious liability requires that the
defendant have a direct financial stake in the infringing behavior. Shapiro, Bernstein & Co. v. H. L.
Green Co., 316 F.2d 304, 307 (2d Cir. 1963).
171. See Goldman, supra note 115, at 405-10 (noting that deterring infringers is often difficult
because many are judgment proof and some hold strong ideological beliefs that copyrighted works
should be freely reproduced).
172. The Pirate Bay website was operational again within a week of a raid on its premises for
copyright infringement. Quinn Norton, The Pirate Bay Bloodied but Unbowed, WIRED (June 6, 2006),
http://www.wired.com/science/discoveries/news/2006/06/71089.
173. Lemley & Reese, supra note 157, at 1350.

Electronic copy available at: https://ssrn.com/abstract=2297488

20 14] INNO VA TION AND INCARCERATION 309

users, they are likely to be risk averse in the face of threatened lawsuits and
thus readily limit access for users who are engaging in valuable speech at
the boundary of infringement. 7 4 For example, although the DMCA
establishes "safe harbor" provisions that shield intermediaries who respond
to copyright takedown notices sent by content owners, 7 s those
intermediaries have very little interest in carefully screening contested
claims of unauthorized use.'7 6 It is easier and less risky for them to simply
remove the contested content. Accordingly, we fear that expanded use of
secondary liability might be more costly in terms of overall social welfare.

C. THE THEORETICAL CASE FOR CRIMINAL COPYRIGHT SANCTIONS

Deterring socially harmful copyright infringement with DRM and
civil sanctions may be difficult and costly in some settings. Creating
deterrence for judgment-proof infringers using civil sanctions alone may be
impossible. In response, copyright owners can turn to technological
measures to prevent infringement in the first place. But as we have shown,
reliance on self-help can itself create social costs. 177 DRM will tend to
protect uncopyrightable content, and it will prevent fair and valuable uses
of copyrightable material. Accordingly, it is worth considering the possible
value of criminal sanctions, including imprisonment, as a deterrent to
harmful copyright infringement. In this section, we propose a limited scope
for the efficient use of criminal sanctions in copyright law. In addition, we
discuss some possible caveats to our proposal, and we address some
arguments in favor of criminal copyright liability that we believe are
incorrect.

1. The Benefits of Criminal Sanctions

As we explained in Part II, economic analysis recognizes certain cases
when criminal penalties can be more efficient than other mechanisms for
deterring harmful conduct. Criminal penalties are valuable when self-help
is costly and when civil remedies are insufficient to deter behavior. These
two factors are likely true for some kinds of copyright infringement. Yet
the existence of some deterrence benefits does not necessarily mean that
criminal sanctions are efficient. They should only be adopted if the

174. See Neal Kumar Katyal, CriminalLaw in Cyberspace, 149 U. PA. L. REv. 1003, 1007-08,
1100 (2001) ("Because an ISP derives little utility from providing access to a risky subscriber, a legal
regime that places liability on an ISP for the acts of its subscribers will quickly lead the ISP to purge
risky ones from its system.").
175. 17 U.S.C. § 512(c) (2012).
176. See Katyal, supra note 174, at 1007-08, 1100.
177. See supra Part II.B. 1.

Electronic copy available at: https://ssrn.com/abstract=2297488

310 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

deterrence benefits exceed the costs of their use.

When a rock star infringes the copyright of another singer's song,
criminal sanctions will not be necessary. The likelihood that the
infringement will be detected is high, and the rock star will probably be
able to pay the resulting damages award. This is not the case, however, for
a private person who uploads hundreds of copyrighted files onto the
internet. Detection of these behaviors is low, so the magnitude of the
damages award necessary to create deterrence will almost certainly swamp
the person's assets. A potential infringer may not be dissuaded at all by the
threat of a judgment so large that it can never be paid.17 8 And high-
magnitude damages awards eliminate the possibility of marginal
deterrence-once the damages award exceeds the infringer's present and
future resources, there is nothing stopping her from infringing further.
Because monetary damages are unlikely to deter individual large scale
infringers, criminal sanctions may be appropriate. The threat of
imprisonment or some type of alternative sanction may create the necessary
deterrence against socially harmful copyright violations.' 7 9 The alternative
sanctions that we have in mind are those that have often been used in
cybercrime prosecutions-prohibitions on the ownership or use of
technologies that are capable of violating copyrights.' 80 This means that for
some period of time, people convicted of criminal copyright infringement
would be subject to prohibitions or limitations on their use of computers,
cell phones, tablets, and other electronic devices.' 8 ' For deterrence
purposes, alternative sanctions will be cheaper to enforce than

178. Some copyright infringement judgments may be dischargeable in bankruptcy if the

infringements were not "willful and malicious." 11 U.S.C. §523(a)(6) (2012). The Ninth Circuit has
recently indicated that the "willful and malicious" inquiry in bankruptcy law is not equivalent to the
"willfulness" inquiry in copyright law, and that the former must be established separately by a
bankruptcy court. In re Barboza, 545 F.3d 702, 707 (9th Cir. 2008) (requiring separate analysis of
whether copyright infringement was willful and malicious as those terms are understood in bankruptcy
law for determining whether discharge is suitable).
179. Hardy, supranote 13, at 314.
180. See COMPUTER CRIME & INTELLECTUAL PROP. SECTION CRIMINAL Div., U.S. DEP'T OF
JUSTICE, PROSECUTING COMPUTER CRIMES 146 (2009), available at
http://www.justice.gov/criminal/cybercrime/docs/ccmanual.pdf ("Where a networked computer has
been used to perpetrate online fraud, to receive contraband such as child pornography or stolen credit
card numbers, or as the instrument of intrusions into or attacks on other computers, these considerations
may militate in favor of imposing a restriction on computer use as a condition of supervised release.").
181. See, e.g., United States v. Holm, 326 F.3d 872, 878-79 (7th Cir. 2003) (overturning a one-
year ban on the use of the internet by a convicted child pornographer but allowing more narrowly
tailored restrictions on internet use, including random searches and filtering software). See generally
Robin Miller, Validity of Condition of Probation,Supervised Release, or ParoleRestricting Computer
Use or Internet Access, 4 A.L.R. 6th 1 (2005).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNO VA TION AND INCARCERATION 311

incarceration, and they may have an important deterrent bite between

potentially underdeterring civil sanctions and potentially overdeterring
incarceration. Perhaps more important than their deterrent effect, however,
is the incapacitative effect of alternative sanctions. If enforced, these
sanctions should prohibit offenders from engaging in future copyright
infringements. Incapacitation will be particularly valuable from an
economic perspective for those defendants, like warez traders and the
ringleaders of file-sharing sites, who are difficult to deter because of their
political beliefs. If they are causing harm and other penalties do not deter
them, incapacitation through imprisonment or alternative sanctions can
prevent the harmful behavior.
The recently implemented Copyright Alert System ("CAS") may go a
significant way toward creating the kinds of deterrence and incapacitation
that would be established by alternative criminal sanctions.' 8 2 CAS is a
private system created through negotiations between the RIAA and MPAA,
on one hand, and leading ISPs, on the other, to create a mechanism for
"graduated response" to online copyright infringement.183 CAS, which was
implemented in February 2013 by the Center for Copyright Information,
tracks the online behavior of users of BitTorrent websites. Users who are
detected engaging in illegal file sharing will receive multiple notifications
of their detection and warnings against continued file sharing. After the
fifth or sixth such alert, users may be subject to severe temporary
limitations on their internet bandwidth that would prevent more file
sharing. 184 Although CAS has received criticism from some academics and
the Electronic Frontier Foundation, 8 its use may be less controversial and
more efficient than the alternative of highly punitive civil or criminal
sanctions.
Evidence on the effectiveness of graduated response systems is limited
by available data. Similar systems that have been implemented in Europe
and Asia seem to issue very few final warnings to users and impose very
few penalties on them.186 One empirical study of France's HADOPI three-

182. Annemarie Bridy, Graduated Response American Style: "Six Strikes" Measured Against
Five Norms, 23 FORDHAM INTELL. PROP. MEDIA & ENT. L.J. 1, 60-62 (2012).
183. Rebecca Giblin, Evaluating GraduatedResponse, 37 COLuM. J.L. & ARTS 147, 176 (2014).
184. Id. at 177-78.
185. Bridy, supra note 182, at 63; Mary LaFrance, Graduated Response by Industry Compact:
Piercingthe Black Box, 30 CARDOZO ARTS & ENT. L.J. 165, 171 (2012); Abigail Phillips, The Content
Industriesand ISPs Announce a "Common Frameworkfor Copyright Alerts ": What Does It Mean for
Users?, ELECTRONIC FRONTIER FouND. (July 7, 2011), https://www.eff.org/deeplinks/2011/07/content-
industry-and-isps-announce-common.
186. See Giblin, supra note 183, at 157-59, 162-66 (examining the systems of France, New

Electronic copy available at: https://ssrn.com/abstract=2297488

312 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

strikes law, however, finds a significant effect of the law on consumption

of legally purchased online music. 187 According to this study, increased
consumer awareness of the law was associated with a 22.5 percent increase
in iTunes song sales, or about $6.3 million in annual iTunes track
revenues.188 Although these data cannot indicate whether the systems like
HADOPI and CAS are more cost-effective means of decreasing copyright
infringement than other methods such as civil and criminal lawsuits, they
suggest that various forms of alternative sanctions could generate
considerable deterrent effect at relatively low cost.
Before discussing the costs of criminal copyright liability, we should
address one argument in favor of criminal liability that we do not find
persuasive in this context. As noted in Part II, public detection and
prosecution of harmful conduct can be more efficient than private
enforcement because economies of scale favor a single police force and
because private individuals may struggle to enforce laws on their own. If
two people are both robbed, it is inefficient for both of them to investigate
the crime when the police could combine the investigations. Owners of
copyrights, however, are well organized and represented by major industry
associations that can devote the necessary resources to detecting and
litigating infringements. So, although a single infringer may be violating
copyrights of many different owners, the RIAA and MPAA appear willing
and able to police this conduct. In fact, much of the criminal investigation
performed by the Justice Department occurs after notification of infringing
behavior has been made by these associations. 189

2. The Costs of Criminal Sanctions

Although imposing criminal penalties for copyright infringement may
have some deterrent benefits, it is not without costs. The most obvious
costs of criminal copyright infringement are the costs of detection,
enforcement, prosecution, and sanction. The costs of detecting and
prosecuting copyright infringement are typically shared by the public (via
the Department of Justice) and organizations that represent victims of
infringement (the RIAA, MPAA, and software organizations and
companies). These costs are often steep. In addition, imprisoning offenders

Zealand, Taiwan, and South Korea).

187. Brett Danaher et al., The Effect of GraduatedResponse Anti-Piracy Laws on Music Sales:
Evidence from an Event Study in France (2012), available at
http://papers.ssm.com/sol3/papers.cfm?abstract-id=1989240.
188. Id. at 18.
189. Kim F. Natividad, Stepping It Up and Taking It to the Streets: Changing Civil and Criminal
Copyright Enforcement Tactics, 23 BERKELEY TECH. L.J. 469, 480-82 (2008).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 3 13

creates significant costs: prisons are expensive to operate, inmates are no

longer productive members of society, and imprisonment creates a variety
of additional harms to offenders and society.1 90 The social harms caused by
unauthorized copying would have to be extremely large in order to justify
the significant costs of detecting, prosecuting, and imprisoning copyright
infringers. Accordingly, if the technological prohibitions mentioned above
have significant deterrent effect, they will be much less costly to
administer.
Infringement liability is intended to deter people from engaging in
socially harmful unauthorized copying, but because copying can also serve
socially valuable purposes, enhanced penalties risk chilling some beneficial
conduct. 19 ' Some people may refrain from valuable but legally
questionable conduct if the penalties for infringement increase.19 2 The
likelihood of this happening will depend on the similarity between
acceptable and unacceptable conduct and the difficulty of predicting ex
ante whether certain behaviors are legal or not.193
As we describe above, one of the principal benefits of civil or criminal
sanctions to deter copyright infringement is the reduction in socially
wasteful self-help. If criminal sanctions lead to improved deterrence of
copyright infringement, copyright holders would no longer need to rely on
DRM to prevent infringement. The reduction in overbroad DRM usage that
eliminates (or at least makes more costly) much socially valuable conduct
is the chief goal of enhanced deterrence. Whether content owners would
refrain from using DRM is, however, unclear. The creation and use of
DRM code is relatively cheap for content owners, and this is especially true
because the DMCA minimizes the costs of an arms race with hackers.' 94
Moreover, DRM generates significant benefits for copyright holders in
addition to preventing illegal copying. DRM enables copyright holders to
protect works beyond their copyright term and to protect content that is not
an appropriate subject for copyright law. DRM also can significantly
increase the costs of fair use. As we described, circumventing DRM for fair
uses may not be exempted from the DMCA, and even if it is, the

190. See Bronsteen, Buccafusco & Masur, supra note 51, at 1040-41 ("People who have spent
any time in prison are significantly more likely to experience chronic, stress-related health impairments,
unemployment, and the breakdown of psychologically vital social ties.").
191. See Moohr, supranote 108, at 802 ("[C]riminal enforcement, to the extent it deters infringing
activity, may also deter socially valuable conduct that is not unlawful.").
192. Mania, supra note 15, at 499.
193. See Moohr, supra note 108, at 802-04 (discussing the complexities users face when
determining whether a copyrighted work is free to use).
194. Lee Kovarsky, A Technological Theory of the Arms Race, 81 IND. L.J. 917, 950-53 (2006).

Electronic copy available at: https://ssrn.com/abstract=2297488

314 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

uncertainty of whether conduct is in fact fair use will deter many risk-
averse users. This may be entirely consistent with the preferences of
copyright holders; much fair use, especially criticism and parody, will be
objectionable to copyright owners. Since the reduction in DRM usage is the
main benefit we anticipate from criminal liability, it is important to know
whether it would actually occur.
One solution to this problem is for Congress to disallow DRM
entirely. Since DRM can be socially harmful, and since the justification for
the costs of criminal sanctions is the reduction of DRM usage, Congress
could simply prevent the distribution of copyright works in media that
include technological protection measures. In that sense, our claim about
the value of criminal IP sanctions is largely a political one. The more
effective that law is in deterring copyright infringement, the stronger the
political case for prohibiting or restricting various types of socially costly
self-help. If law were able to deter much large-scale copying, this might
provide the political impetus necessary to catalyze the enactment of laws
restricting or prohibiting DRM.
A final category of costs worth mentioning includes the possible
effects of imposing criminal liability on conduct that many people believe
does not warrant it. There is considerable variation in people's beliefs
about the harms of copyright infringement.195 Many people believe that
downloading music and movies from the internet is not harmful at all and
certainly should not subject offenders to imprisonment. 196 Stark divergence
between social norms about behavior and legal rules governing it can be
very costly. 197 When norms and law diverge, people's respect for the law
decreases, and their compliance rate declines." Thus, if people believe
that criminal sanctions are inappropriate for copyright infringement, the
benefits of deterrence will be weaker.199 As with DRM, however, the

195. Hardy, supra note 13, at 326 ("We are witnessing, then, a growing split between the
government and various copyright-dependent industries, like music, computer software, and motion
pictures, on the one hand, and highly critical copyright consumers and Internet users on the other.").
196. Id. at 314 ("[T]he trend toward greater deterrence through greater punishment of
infringement, especially criminal punishment, seems increasingly at odds with the public's perception
of what is fair and what is consistent with the American tradition of individual liberty.").
197. See Tyler, supranote 158, at 227 (explaining that any symbolic power formal law has is lost
when the law diverges from public morality).
198. Moohr, supra note 108, at 805 ("Respect and legitimacy are threatened when a community
norm that condemns prohibited conduct is not yet in place.").
199. See TOM R. TYLER, WHY PEOPLE OBEY THE LAW 64 (1990) (noting that compliance with the
law is stronger when people believe the law "ought" to be followed); Moohr, supra note 108, at 795
("Under any theory of deterrence, it is more difficult to induce law-abiding behavior when underlying
social norms do not support the law.").

Electronic copy available at: https://ssrn.com/abstract=2297488

2 014] INNO VA TION AND INCARCERATION 3 15

precise effect of criminal copyright liability on norms is difficult to predict.

It is also possible that the imposition and effective enforcement of criminal
sanctions for violating copyright law could help people understand the
harm of copyright infringement, resulting in greater convergence between
social attitudes and the law.
Subject to these important qualifications, there is a sound, if limited,
economic case for the use of criminal penalties to deter some copyright
infringement. Civil sanctions alone appear ineffective at deterring some
socially harmful copying. In response, copyright owners will increasingly
turn to DRM to combat this behavior. But while DRM can be used to
prevent socially harmful copying, it will also tend to limit socially
beneficial uses of copyrighted works. If criminal sanctions can deter
harmful conduct that would not have been deterred by civil sanctions alone,
copyright owners should be willing (or compelled) to forego DRM, thereby
decreasing the costs of valuable behavior. The chief economic rationale for
criminal copyright law is not that it will allow copyright holders to capture
more rents, but rather the expectation that deterring some harmful copying
will generate more beneficial behavior.

D. WHAT KIND OF CRIMINAL LIABILITY COULD BE DESIRABLE?

Part III.C made the theoretical case for using criminal sanctions to
deter copyright infringement. We believe that some kind of criminal
copyright liability may be economically justified. In this section, we
describe the nature of the criminal sanctions that are most likely to be
beneficial. The use of criminal sanctions to deter copyright infringement
offers not only significant benefits, but also considerable costs. As with
most such projects, accurately estimating the costs and benefits is difficult.
This is especially true given the empirical uncertainty surrounding some of
the most important issues. In the face of this uncertainty, the use of
criminal sanctions for copyright violation should be narrowly
circumscribed and limited to situations that are likely to produce the
highest benefits at the lowest costs.
Criminal liability is especially important with respect to individual or
poorly capitalized infringers of large quantities of copyrighted material
who are difficult to detect and identify. Accordingly, criminal liability
should focus on these situations. It would certainly be objectionable to
disproportionately prosecute and imprison poor offenders. Yet focusing
prosecutorial effort on those acts of infringement which are the most costly
and most difficult to detect makes economic sense. Additionally,
infringement prosecutions should address large-scale counterfeiting

Electronic copy available at: https://ssrn.com/abstract=2297488

316 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

operations that market bootlegged music, movies, and software. These

behaviors likely cause the most social harm, and these are the violators for
whom criminal liability will pack the most deterrent punch.
Because imposing criminal liability risks chilling valuable conduct,
criminal sanctions should be reserved for behavior that is most clearly
harmful. This is the case for exact duplication of copyrighted works that
will directly substitute for legitimately available copies. Although
copyright law gives authors the exclusive right to make derivative works
and substantially similar copies of their works, 200 a great deal of behavior
that falls into these categories also could constitute fair use. 20 1 By contrast,
there will be an extremely limited number of situations when exact
duplication of copyrighted works can constitute fair use. 202 By confining
criminal liability to the latter cases, the risks of chilling socially valuable
behavior can be minimized.
These risks can be further minimized by including a strict mens rea
requirement for criminal liability. Mens rea standards can be helpful in
distinguishing harmful from valuable conduct. Accordingly, criminal
copyright infringement should condition liability on proof that the
defendant intentionally violated a known copyright. That is, the prosecution
should have to prove that the defendant knew that the infringed works were
protected by copyright and that he knew that his conduct was unlawful.
Good-faith belief that his behavior was not copyright infringement or
constituted fair use should preclude criminal liability.
Finally, in order to further guard against deterring socially valuable
behavior, the DMCA should be amended to clearly exempt defendants
from liability when their circumvention of technological protective
measures is done for purposes of fair use or to obtain access to content that
is not protected by copyright.2 03 In theory, perfect enforcement of copyright

200. 17 U.S.C. § 106 (2012).

201. Moohr, supra note 108, at 27 ("Most copyrighted works are based, directly or indirectly, on
the work of others. Potential creators may be reluctant to use copyrighted work, which is a sort of 'raw
material' for new work, when the potential cost of doing so is very high." (footnote omitted)).
202. See 17 U.S.C. § 107 ("In determining whether the use made of a work in any particular case
is a fair use the factors to be considered shall include-(I) the purpose and character of the use,
including whether such use is of a commercial nature or is for nonprofit educational purposes; (2) the
nature of the copyrighted work; (3) the amount and substantiality of the portion used in relation to the
copyrighted work as a whole; and (4) the effect of the use upon the potential market for or value of the
copyrighted work.").
203. See Jacqueline Lipton, Information Property: Rights and Responsibilities, 56 FLA. L. REV.
135, 159 (2004) ("The legal status of fair use and the scope of the fair use defense might be clarified for
the digital age if Congress amends the DMCA so that copyright holders are not permitted to deny
access to fair users.").

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] ]INNOVATION AND INCARCERATION 3 17

law would eliminate the need for technological protections, but perfect
enforcement is not possible in practice and, given the cost of achieving it, is
not warranted. Thus, the use of DRM can help prevent harmful copyright
violations, but users who circumvent it for valuable reasons should not be
subject to liability.

E. ASSESSING CURRENT U.S. CRIMINAL COPYRIGHT LAW

Having made the limited and tentative economic case for criminal
sanctions for some copyright infringement, it remains to assess the extent
to which current U.S. law conforms to our judgments. For the most part, it
matches our prescriptions fairly well. The provisions of criminal copyright
law 'and the enforcement efforts of the Department of Justice generally
comport with our analysis of when such liability is appropriate. We do have
some concerns about the monetary thresholds for criminal liability,
however.

1. Current Criminal Copyright Law

U.S. copyright law provides for the use of criminal sanctions to deter
copyright infringement in 17 U.S.C. § 506.204 The law provides for
punishment of "[a]ny person who willfully infringes a copyright ... if the
infringement was committed-(A) for purposes of commercial advantage
or private financial gain; [or] (B) by the reproduction or distribution,
including by electronic means, during any 180-day period, of 1 or more
copies or phonorecords of 1 or more copyrighted works, which have a total
retail value of more than $1,000.",205 Thus, to be found liable, a defendant
must be shown to have acted willfully and to have been motivated by
commercial advantage or to have met the damages threshold established in
the second provision of the statute.
The mens rea requirement of willfulness is helpful in limiting the
potential for chilling valuable copying behavior. Courts have generally
interpreted this requirement to require proof of a voluntary, intentional
violation of a known legal duty. 206 This typically means that the defendant
subjectively knew that the infringed works were subject to copyright and
that his conduct was unlawful. Merely intending to copy is insufficient.

204. 17 U.S.C. § 506 (2012).

205. Id. § 506(a)(l)(A)-(B).
206. United States v. Liu, 731 F.3d 982, 985 (9th Cir. 2013) (requiring the government to prove
that the defendant knew he was acting illegally in order to satisfy the "willfully" mens rea requirement);
Rebecca E. Hatch, CriminalInfringement of Copyright Under 17 U.S.C.A. § 506, 120 AM. JUR. PROOF
OF FACTS 3D 181, § 3 (2011).

Electronic copy available at: https://ssrn.com/abstract=2297488

318 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

Given this fairly strict requirement, it is unlikely that the existence of

criminal sanctions will deter fair use.
We are less confident that the thresholds for criminal liability in
subsections (A) and (B) are sufficient to limit criminal liability to those
cases when it is most effective. The requirement in (A) that the defendant
be motivated by commercial gain or profit covers much conduct where
civil liability would be sufficient. This is especially true since the No
Electronic Theft ("NET") Act of 1997 modified the requirement of
commercial gain to mean "the receipt, or expectation of receipt, of anything
of value, including the receipt of other copyrighted works." 207 Selling a
copy of a CD to a friend for $10 would meet the requirement, but there is
no reason to think that the apparatus of the criminal law is necessary to
deter this behavior. Even with the low likelihood of detection, the harm it
causes is slight.
Subsection (B)'s requirement, which was also added by the NET Act,
removes the necessity of proving the defendant's motive, but it does add a
monetary threshold. It is difficult to predict whether $1000 is the
appropriate place to set this limit. The answer will turn on the probability
of detection and the availability of a damages multiplier in civil suits. If a
defendant willfully infringes $1000 worth of copyrighted works, he may be
liable for statutory damages of up to $30,000 per infringed work and up to
$150,000 in enhanced damages for willfulness. 208 If the total civil damages
award for the $1000 infringement is $200,000, then a rational person will
only be deterred from infringing if the probability of detection is at least
0.005 (one-half of one percent). If the damages multipliers cannot be
constitutionally raised to reflect likelihoods of detection that are lower than
this, the $1000 threshold may be set close to the optimal value. In addition,
defendants who cannot pay such large judgments also might not be deterred
by civil fines, in which case criminal penalties could be necessary.
The penalties for criminal copyright infringement are specified in 18
U.S.C. § 23 19.209 Defendants found guilty under 17 U.S.C. § 506(a)(1)(A)
(the commercial gain or profit subsection) are subject to imprisonment for
not more than five years if they reproduce or distribute at least ten copies of
one or more works with a total retail value of more than $2500 in any 180-
day period. 21 0 The maximum prison term doubles upon subsequent

207. No Electronic Theft ("NET") Act of 1997, Pub. L. No. 105-147, 111 Stat. 2678 (codified at
17 U.S.C. § 101).
208. 17 U.S.C. §504.
209. 18 U.S.C. §2319 (2012).
210. Id. §2319(b)(1).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014}] INNOVATION AND INCARCERATION 3 19

convictions. 211 For defendants found guilty under subsection (B) (the
monetary threshold section), the maximum prison term is three years for a
first offense and six years for subsequent offenses. 2 12
In addition, some courts have been willing to employ computer use
restrictions and other alternative sanctions as conditions of probation and
supervised release. One defendant was subjected to three years of probation
during which he "shall maintain a daily log of all addresses he accesses via
any personal computer or other computer used by him, other than for
authorized employment, and he shall make this log available to the
Probation Officer." 213 Additionally,
The defendant shall refrain from accessing, via a computer, any
"material" that relates to the activity in which he was engaged while
committing the instant offense, namely, (Internet Relay Chat] channels;
e-mail and instant messaging that relate to warez activities; and any
intemet warez related web sites, channels, and private sites. 2 14
Although monitoring these sanctions may be difficult, that is always the
case with probation and supervised released. If these sanctions are backed
with the threat of a return to prison, they may nonetheless be successful
even if they are not perfectly enforced.
As always, it is difficult to say whether the penalties imposed for
copyright infringement are well-balanced to maximize their deterrent
benefit and minimize social costs. Sending people to prison is very
expensive. Moreover, as we have argued elsewhere, longer terms of
imprisonment may have diminished deterrent capacity relative to their
length and cost.215 Given the high social cost of imprisonment, terms
exceeding a few years are probably not efficiently deterring copyright
infringement. The use of alternative sanctions such as technological use
limitations is worth exploring. 216 The threat of having limited internet and
cell phone access for five years may be sufficient to deter many twenty-

211. Id. § 2319(b)(2).

212. Id. §2319(c)(I)-(2).
213. United States v. Fong, 2006 WL 5388953 (N.D. Cal. Nov. 2, 2006).
214. Id. See also United States v. Bailey, 286 F. App'x 678, 679 (11th Cir. 2008) (discussing
where the defendant's supervised release was revoked for failure to adhere to computer restrictions).
215. See generally Bronsteen, Buccafusco & Masur, supra note 51 ("The negative post-prison
repercussions of having served a prison term accrue to essentially any inmate: the differences in post-
prison outcomes between offenders who were incarcerated for shorter terms and those incarcerated for
longer terms are minimal.").
216. See Sam Cowin, Note, You Don't Have Mail: The Permissiblity of Internet-Use Bans in
Child PornographyCases and the Need for Unformity Across the Circuits, 80 GEO. WASH. L. REV.
885 (2012) (arguing how internet-use bans can be an effective punishment in child pornography cases).

Electronic copy available at: https://ssrn.com/abstract=2297488

320 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

year-olds from engaging in large scale copyright infringement. Further, the

costs of imposing such sanctions are relatively low: only monitoring costs
and the productivity losses from limiting offenders' use of communications
technologies. In addition, many copyright violators who are not motivated
by economic gain, especially those whose behavior is politically motivated,
may be difficult or impossible to deter with any size sanction. 2 17
Accordingly, alternative sanctions that prevent them from engaging in
violations in the first place may have valuable incapacitative effects.
Without access to the technology that enables infringement, these offenders
will be less likely to cause harm. Although alternative sanctions impinge
upon offenders' free speech rights, they are almost certainly less extreme
than the liberty losses associated with imprisonment.2 18

2. Criminal IP Enforcement
As we have discussed, there are reasons to be concerned that the scope
and magnitude of criminal penalties for copyright infringement may be
overbroad. In addition, the recent high-profile prosecution of Aaron Swartz
for computer crimes suggests that the U.S. government could be
overzealously enforcing copyright laws. For the most part, however, recent
criminal copyright prosecutions seem to conform to the situations in which
such liability is most valuable.
The Department of Justice tends to prosecute between fifty and one
hundred cases each year, 219 and the defendants are typically those engaged
in the kinds of conduct most appropriate for criminal liability-large-scale
counterfeiting that is difficult to detect. Most recently, for example, the
leaders of the internet piracy group "IMAGiNE" were convicted of running
"an organized online piracy ring that sought to become the premier group
to first release Internet copies of movies only showing in theaters." 220 The
defendants received sentences ranging from twenty-three to sixty
months. 22 1 In another recent case, the defendant was convicted of
producing and selling thousands of counterfeit music CDs and movie

217. See Goldman, supranote 115, at 405-10.

218. United States v. Holm, 326 F.3d 872, 878 (7th Cir. 2003) (noting when limiting a complete
internet ban that "such a ban renders modem life-in which, for example, the government strongly
encourages taxpayers to file their returns electronically, where more and more commerce is conducted
on-line, and where vast amounts of government information are communicated via website-
exceptionally difficult").
219. Hatch, supra note 206, § 1.
220. Leader of Internet Piracy Group "IMAGiNE" Sentenced in Virginia to 60 Months in Prison
for Criminal Copyright Conspiracy, U.S. DEP'T JUST. (Jan. 3, 2013),
http://www.justice.gov/opalpr/2013/January/1 3-crm-010.html.
221. Id.

Electronic copy available at: https://ssrn.com/abstract=2297488

2014]1 INNOVATION AND INCARCERATION 321

DVDs at flea markets in California. The estimated retail value of the

counterfeit goods was $2.6 million. The lead defendant in the case received
a four-year sentence. 222
These are the kinds of defendants against whom criminal liability will
be most valuable. 223 Their behavior is hard to detect and occurs on a large
scale. If they were subject to civil liability only, the magnitude of the
damages awards would almost certainly exceed their ability to pay.
Moreover, their behavior-counterfeiting music, movies, and software-is
the kind that creates substantial social harm and little, if any, social benefit.
The fact that someone was convicted and sentenced for selling thousands of
copies of bootlegged music is unlikely to deter others who are engaged in
conduct that is socially valuable.
This part has analyzed the theoretical economic case for imposing
criminal sanctions on copyright infringers and found that criminal penalties
may be warranted to deter some behaviors that might otherwise go
unchecked by civil sanctions or DRM. The existence of criminal penalties
for copyright infringement could minimize copyright owners' reliance on
DRM, thereby reducing the costs of socially beneficial uses of copyrighted
works. Deterring socially harmful copying while promoting socially

222. San Jose Man Sentenced to Prisonfor Extensive Conspiracy to Sell Counterfeit Media, U.S.
DEP'T JUST. (Nov. 19, 2012), http://www.fbi.gov/sacramento/press-releases/2012/san-jose-man-
sentenced-to-prison-for-extensive-conspiracy-to-sell-counterfeit-media.
223. See also Annandale Man Sentenced to 36 Months for $2.5 Million in Software Piracy,U.S.
DEP'T JUST. (Nov. 9, 2012), http://www.justice.gov/usao/vae/news/2012/11/20121108nguyennr.html
(discussing a defendant who pleaded guilty to copyright infringement for selling $2.5 million in
copyright-infringing computer software and defrauding more than 2000 customers. He was sentenced to
thirty-six months in prison, followed by three years of supervised release. Additionally, he was "ordered
to pay restitution of $2.5 million and a forfeiture money judgment of $1.4 million."); Internet
Businessmanfrom New York Sentenced for Infringing Copyrights, U.S. DEP'T JUST. (Dec. 3, 2012),
http://www.justice.gov/usao/vae/news/2012/12/20121203newsomenr.html (discussing a defendant who
pleaded guilty to four counts of criminal copyright infringement and was sentenced to eleven months in
prison followed by three years of supervised release, which includes nine months of home confinement.
As an additional part of his sentence, the defendant "forfeited assorted computers, hard drives, and
other electronic equipment." The defendant owned and operated an online business selling copyrighted
computer software, education, and training materials worth approximately $345,021.68. A civil
copyright infringement suit was previously filed against the defendant in 2006.); Leader of
NinjaVideo.Net Website Sentenced to 22 Months in Prisonfor Criminal Copyright Conspiracy, U.S.
DEP'T JUST. (Jan. 6, 2012), http://www.justice.gov/criminal/cybercrime/press-
releases/2012/besharaSent.pdf (announcing that the leader of NinjaVideo.net Website was sentenced to
twenty-two months in prison followed by two years of supervised release following her prison stay, and
she must complete 500 hours of community service. In addition, the defendant must "repay
$209,826.95 that she personally obtained from her work at NinjaVideo.net and forfeit to the United
States several financial accounts and computer equipment involved in the crimes." The website, which
operated from February 2008 until June 2010, enabled visitors to view movies still in the theaters,
movies not yet released, and newly aired television shows.).

Electronic copy available at: https://ssrn.com/abstract=2297488

322 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

beneficial copying promotes the overall welfare. As noted, however,

important questions about the precise effects of criminal liability remain
unanswered and unanswerable with current data.

IV. CRIMINAL PATENT LAW?

Although we believe that there is a limited economic case for criminal
copyright sanctions, this does not automatically mean that the same is true
for patent law. The economics of patents and inventions are very different
from those of copyrights and creative works. Accordingly, this part applies
the economic theory discussed in Part II to the unique issues that arise for
patent law.

A. THE ECONOMIC HARM FROM PATENT INFRINGEMENT

In order to determine whether criminal sanctions would be
economically beneficial in patent law, we must first determine what (if
any) economic harm is caused when a firm or individual infringes a patent.
There are two possible sources of harm. The first is the reduction in the
patent holder's incentives to commercialize, improve, or create the
invention in the first place-that is, the reduced incentives to innovate in
light of the reduced returns to owning a patent.224 These costs are
extremely difficult to calculate because they require answering a
hypothetical counterfactual: How much innovative work would the patent-
holding firm have done had it known that its patent rights would be
infringed by a competitor? This in turn implicates the question of how
much other firms will reduce their innovative efforts in the future if they
have reason to believe that their future patent rights will not be secure. As
we did with copyright law, we will treat some forms of patent infringement
as causing social welfare losses. 2 25
The second potential source of harm from patent infringement lies
with the possibility of expensive self-help by the patent holder. What form
might that self-help take in the patent context? The kinds of technological
measures that are available for copyrighted works will generally not be
available to prevent unauthorized use of new inventions. Innovators could,

224. See supra Part II.A (describing an economic theory of patent rights as an inducement to
innovation).
225. As with copyright law, there will be some authorized uses of patents that will not be socially
costly. Again, when infringers fall in the category of deadweight losses, their infringement (at least
without effective price discrimination) does not cause social welfare harms. Additionally, as discussed
infra, some forms of patent infringement, such as "inventing around" a patent, may create net social
benefits (analogously to fair use of copyrighted works).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 323

however, switch from patents to trade secrets, thereby preventing

unauthorized use of the invention by refusing to share it publicly. 226
Inventors could try to limit others' ability to know essential aspects of an
invention, giving them a competitive advantage. A trade secret is the IP
equivalent of building high walls around a piece of property. (Some trade
secrets are in fact protected by high walls.) Trade secrets are privately
costly because they require private expenditures to secure them: the owner
of a trade secret must expend resources protecting that secret via
technological measures (such as walls, both physical and electronic) and
nondisclosure agreements. 227 They may also be socially costly (compared
with patents) precisely because information surrounding the technological
advance is kept secret rather than shared with the public. 228 One of the
principal benefits of the patent system is the anticipated quid pro quo
between inventors and the public-in return for getting exclusive rights in
their inventions, inventors must disclose those inventions to the public who
benefit from the opportunities for further innovation. 229
We suspect, however, that the risk of firms electing trade secrets over
patents is relatively small, even in an environment in which patent
infringement goes underdeterred. Many inventions will simply be
impossible to protect with trade secrets. 23 0 When an invention is
commercialized and sold, competitors will be able to reverse-engineer the
product and reveal the innovation underlying it. 23 1 Reverse engineering is
not prohibited by trade secret law, so competitors will quickly gain access
to the underlying invention. 23 2 This will be true for nearly all
pharmaceuticals and biotechnology (whose chemical formulae are
discernible from the drugs themselves), medical devices, and machinery, as
well as many electronics and semiconductor inventions. Inventions of
processes that can be practiced behind closed doors, the fruits of which can

226. Anderson, supra note 103, at 923-24 (describing the options available to innovators choosing
between different forms of intellectual property protection).
227. Mark A. Lemley, The Surprising Virtues of Treating Trade Secrets as IP Rights, 61 STAN. L.
REV. 311, 331-32 (2008) (describing trade secret law and the requirements it places on owners of trade
secrets).
228. Id. at 340-41 (analyzing the social losses that accompany decisions to opt for trade secrets
over patents).
229. Jeanne C. Fromer, PatentDisclosure,94 IOWA L. REv. 539, 542, 548-50 (2009) (describing
the benefits of patent disclosure and the tradeoffs at the heart of intellectual property).
230. See Lemley, supra note 227, at 315-19 (making this point about the difficulty of protecting
trade secrets for commercial products).
231. Dunlop Holdings, Ltd. v. Ram Golf Corp., 524 F.2d 33, 37 (7th Cir. 1975) (discussing the
possibility of reverse-engineering commercial products in an opinion by Justice Stevens).
232. Lemley, supra note 227, at 319, 340-41 (discussing reverse engineering).

Electronic copy available at: https://ssrn.com/abstract=2297488

324 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

then be sold without revealing the secrets, may still be protectable as trade
secrets. (Ironically, business methods and software-the most commonly
cited examples of bad patents-may often fall into this category. 233) And
certainly there will be some inventions that are too complex or opaque to
be successfully reverse-engineered. But by and large, the types of
innovations that are currently protected by patent law will not be
protectable by trade secrets.
Of course, the harm that patent infringement will do is only part of the
equation. Although the existence of meaningful harm is necessary in order
answer the threshold question of whether sanctions for infringement are
worthwhile, determining the appropriate level of those sanctions will
require understanding the value of the infringement to the infringer. Here,
the story is somewhat clearer. The benefit to an infringer from patent
infringement is just the profit that the infringer received from the
infringement, over and above whatever profit the infringer might have
made had it sold a noninfringing product instead. 23 4 This is similar to the
inquiry that a court must conduct when determining the "reasonable
royalty" that an infringer must pay to a patent holder that has prevailed in
court.2 35 For the most part, it depends upon the number of sales that the
infringer has made and the profits per sale. This inquiry can be quite
technically difficult, as it often involves answering a complicated
counterfactual.23 6 But it is at least conceptually straightforward, and a court
will often be able to compute the measure to within a reasonable
approximation.

B. LIKELIHOOD OF DETECTION
As we noted above, the deterrent power of sanctions and their
appropriate magnitude will depend on how often harmful behavior will be
detected. How likely is detection of patent infringement? Here, it is useful
to separate two types of infringers: those who infringe accidentally and
would prefer to negotiate licenses, and those who infringe deliberately (and
could more accurately be characterized as engaging in theft). Patent owners
will always have incentives to search for possible infringers. Unintentional
infringers-the first group-would have similar incentives to search for

233. See BuRK & LEMLEY, supra note 41, at 72-73 & n.32, 156-58 (describing the arguments
typically leveled against software and business method patents).
234. LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51, 67 (Fed. Cir. 2012) (describing
the proper measure of damages in patent cases).
235. Id. at 66-67 (discussing the reasonable royalty measure of damages).
236. See id. at 67-70 (illustrating the difficulties inherent to properly calculating a running
royalty).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 325

patents that their products might infringe, in order to secure licenses. 237
They also would have no incentive to attempt to hide or disguise their
products or to obscure the fact that those products are infringing. These
unintentional infringers are thus quite likely to be detected (at least
eventually) by patent owners.
For intentional infringers who do not wish to be caught, the picture is
not quite so clear. These parties will attempt to hide their infringement, and
the onus will be entirely on patent holders to locate them. 238 This increases
the search burden placed on patent holders, but does it mean that these
infringers are unlikely to be discovered? We suspect that the answer is
generally no. Most infringing products that are placed on the shelves and
sold openly are quite likely to be detected due to the strong incentives that
patent holders have to police their property rights. However, it is always
possible that a domestic black market for innovative goods could develop,
much like the black market for pirated CDs and DVDs. (We will discuss
international markets below.) But no significant such market appears to
exist, and we think that is unsurprising for several reasons. First, the cost of
producing patented products is typically much higher than the cost of
copying copyrighted works.23 9 Any individual with a computer and the
right software can produce burned DVD copies, but it would take a factory
to make the computer used to burn the DVD. The same holds true for
semiconductors, biotechnology, and the like. Second, quality is much more
at issue in patented products than it is for copyrighted works. Partly
because it is so easy to produce pirated copyright works, and partly because
they are so inexpensive, purchasers do not typically worry about receiving
a copy of unusable quality. But this is not the case for drugs or electronics,
for example, where obtaining a low-quality copy would be expensive and
even potentially harmful. Accordingly, consumers will tend to opt for
purchasing patentable goods from trusted sources, not on the black
market.2 40
It is in part for this reason that criminal investigative tools are less
important and less valuable in the patent context than elsewhere. Police
have mechanisms for investigating crimes and gathering evidence, such as
search warrants, that are unavailable to private litigants. In many cases,

237. See Jonathan S. Masur, Patent Liability Rules as Search Rules, 78 U. CHI. L. REv. 187, 191-
93 (2011) (analyzing patentee and infringer search incentives).
238. Id at 194-96 (discussing this case).
239. See Ted Sichelman, Commercializing Patents, 62 STAN. L. REv. 341, 367 (2010) (explaining
the economics of bringing a patented invention from the laboratory to the marketplace).
240. See id. at 378 (discussing marketplaces for inventive goods).

Electronic copy available at: https://ssrn.com/abstract=2297488

326 SOUTHERN CALIFORNIA LAWREVIEW [Vol. 87:275

police are thus able to investigate wrongdoing with greater accuracy and at
lower cost than similarly situated private plaintiffs might be. This could
provide an independent reason for public enforcement of intellectual
property rights, though it would not necessarily justify coupling public
enforcement with carceral sanctions. With regard to patents, however, the
advantage from involving the police in infringement investigations is much
smaller. The reason, as we explained above, is that the bulk of critical
information is already publicly available via the product being used or sold.
There would be little advantage to having the police investigate
infringement, because private parties are already able to detect it at equally
low cost. Moreover, private parties are best situated to decide which cases
of patent infringement are most serious and most worth investing resources
to stop. Patent owners understand the value of their own IP better than any
public official could. Accordingly, when it comes to patents, a system of
disaggregated private search and enforcement is likely superior (or at least
no worse) than one of public enforcement.
The greatest likelihood of infringement going undetected involves
patents on processes, including software patents, that can be performed
behind closed doors. Just as these processes are the easiest to conceal, and
thus lend themselves most directly to the use of trade secrets, so too are
they the easiest to infringe in secret without detection. 241' A patent holder
could always study the products or services being performed by a
competitor and deduce that the competitor must be violating its patent; or it
could take a guess, file an action alleging infringement, and then seek
discovery that would reveal the competitor firm's secret activities.
Accordingly, even secret infringing activity is by no means impossible to
detect. But there is at least a chance that it could go undetected for a
meaningful length of time. The result is symmetry between infringers' and
patent holders' actions: it is easier to infringe a process patent than a
product patent without being detected, and it is easier to use trade secrets to
protect a process than to protect a product, making innovators significantly
more likely to opt out of the patent system and into trade secret protection
when their innovations take the form of processes. This trend will be
heightened when patent holders view infringement remedies as inadequate.
Thus, the case for criminal penalties for patent violations is strongest with
respect to process patents, particularly patents on inventions that could
conceivably be practiced in secret.

241. See W.L. Gore & Assocs., Inc. v. Garlock, Inc., 721 F.2d 1540 (Fed. Cir. 1983) (describing
the use of a patented process in secret, in a fashion that did not inform the public).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNOVATION AND INCARCERATION 327

While we believe that infringing activities are likely to be detected in

general, findings of liability are another matter entirely. Federal district
court judges find patent law notoriously difficult, 242 and the Federal Circuit
itself has been widely criticized for issuing confusing decisions,
promulgating ambiguous doctrines, and simply deciding cases
incorrectly. 243 One of us has separately suggested that courts should award
heightened damages to patent owners that prevail at trial (and assess
penalties against patent owners that fail) in order to compensate parties for
the risk that their cases will be decided incorrectly. 244 The risk that an
infringer will be discovered but not sanctioned is thus very real.
Yet the risk may not be all that great. Even if patent courts are no
better than a coin flip, there is still a 50 percent chance that an infringer
will be sanctioned. Those are much better odds than exist for most crimes,
where criminals are caught and sentenced to prison at a far lower rate. A 50
percent rate would require only that damages be doubled, which is not
insignificant, but is again far lower than the types of multiples typically
contemplated for other crimes, including copyright infringement. But we
suspect that patent courts, for all of their flaws, are actually better than a
coin flip. This will decrease the necessary multiples accordingly. If, for
instance, patent courts are 75 percent accurate, then the necessary multiple
would be 4/3, a relatively modest figure.
Of course, patent law awards heightened damages, but not to
compensate for the possibility that infringement will go undetected or
unpunished. Rather, judges in patent trials can award up to treble damages
in cases of "willful" infringement: those instances in which a party
knowingly and deliberately infringed a patent.2 45 It should be clear from the
preceding paragraphs that this practice makes little sense from the
perspective of optimal deterrence, at least on its face. However, as we will
explain below, there is one respect in which enhanced damages for willful
infringement tracks a rational economic theory of sanctions for intellectual
property violations.

242. See David L. Schwartz, PracticeMakes Perfect? An EmpiricalStudy of Claim Construction

Reversal Rates in Patent Cases, 107 MICH. L. REV. 223, 255 (2008) (demonstrating that even
experienced patent judges have over 30 percent of their claim construction opinions reversed).
243. Jonathan S. Masur, Patent Inflation, 121 YALE L.J. 470, 478-79 (2011) (listing the various
criticisms that have been leveled at the Federal Circuit).
244. Anup Malani & Jonathan S. Masur, Raising the Stakes in Patent Cases, 101 GEo. L.J. 637,
661-62 (2013) (suggesting a mechanism for compensating patent holders for the risk of error in the
courts).
245. 35 U.S.C. § 284 (2006).

Electronic copy available at: https://ssrn.com/abstract=2297488

328 SOUTHERN CALIFORNIA LAWREVIEW [Vol. 87:275

C. JUDGMENT-PROOF PATENT DEFENDANTS?

The next issue is whether patent defendants would typically be able to

pay judgments levied against them if those judgments took the form of
monetary sanctions, or whether they would be insolvent or judgment proof
in a significant number of cases. In Part IV.A, we noted that the benefit to
an infringer from infringing a patent is equal to the infringer's profits from
the sale or use of the patented product. These profits could in theory be
quite substantial: if a firm sold a patented drug on the black market for a
price just below the monopoly price, it could earn tens or even hundreds of
millions of dollars. Profits from the infringement of less valuable
inventions would likely be smaller but could still be significant.
However, in Part IV.B we observed that patent infringement will
generally be easy to detect, with some notable but limited exceptions.
Because most infringement will be detected, it will not be necessary to
substantially increase the sanctions to account for the likelihood of
nondetection in order to achieve appropriate deterrence. There may have to
be some multiplier to account for the fact that "conviction" is never certain
in a patent court. Nonetheless, the proper measure of sanctions for
infringement will typically hover near the measure of infringer profits.
With rare exceptions, then, the overall sanction should be at or near the
level of profits earned by the infringer.
We believe that patent infringers will tend not to be judgment proof
and will be able to afford this measure of damages. We reach this
conclusion for two reasons. First, as we noted above, manufacturing a
patented product for sale is often quite expensive.246 A firm that wishes to
sell computers, smartphones, or any other type of consumer electronics will
require a factory of some sort to assemble those computers, and even a
pharmaceutical infringer will need a significant operation to produce
prescription drugs in any quantity. To be sure, some inventions, such as
software, can be duplicated with a minimum of capital investment.24 7 But
the median patented product will require a far greater investment of
resources to produce in quantity than will the median copyrighted DVD or
CD. Accordingly, patent infringers will generally be larger, better-
capitalized operations with substantial resources that can be seized in order
to pay judgments. Against such firms, monetary damages will create
adequate deterrence.

246. See supraPart IV.B.

247. See supraPart III.

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNO VA TION AND INCARCERATION 329

Of course, there is always the possibility that an infringer will be a

foreign firm whose assets cannot be seized by American courts. To the
extent that black markets for patented products have developed, they have
typically been foreign based, for precisely this reason. 248 But if American
courts cannot reach a firm's assets, they will not be able to reach the
persons running those firms for purposes of criminal penalties either. Thus,
while deterrence may be an issue with regard to infringers operating
abroad, imposing criminal penalties in cases of patent infringement will not
ameliorate the problem.
The second and even more general reason that patent infringers will
not be judgment proof is that they will have the profits they have earned
from the infringement available to them. Simple disgorgement of these
profits should be nearly adequate to deter firms from infringing a patent.
Recall that this is not the case in many instances of standard property
crimes. In a typical property crime, the property may be damaged or
spoiled before it is recovered, or it may never be recovered at all.249 In
addition, because property criminals are caught and prosecuted at low rates,
it is necessary to multiply the punishment imposed upon them many times
in order to achieve adequate deterrence. 250 This means that forcing the
wrongdoer to disgorge the stolen property, by itself, cannot serve as an
adequate sanction. It is of course possible that patent infringers might
"spoil" the profits they have stolen by absconding with them or investing
them in an asset that has lost value. But this is much less of an issue with
respect to monetary profits that will often be recoverable as cash than it is
with respect to stolen property that might have been fenced, destroyed, or
otherwise reduced in value. By consequence, we believe it is quite likely
that patent infringers will be able to pay the monetary judgments against
them, even when those judgments are large enough to induce deterrence.

D. OPTIMAL ACTIVITY LEVELS AND INCAPACITATION

Before we can reach any final conclusions regarding the propriety of

criminal sanctions for patent infringement, there are two other questions we
must consider. The first is the optimal level of patent infringement: Should
there be zero infringement, or more than zero? Of course, in a utopian

248. See Black Market Pharmacies and the Spam Empire Behind Them, NPR PLANET MONEY
(Jan. 15, 2013), http://www.npr.org/blogs/money/2013/01/15/169424047/episode-430-black-market-
pharmacies-and-the-spam-empire-behind-them (describing the internet black market for patented drugs
and how it operates abroad).
249. See supraPart II.C (describing the effects of property crimes on property).
250. See supraPart I.C.

Electronic copy available at: https://ssrn.com/abstract=2297488

330 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

world there would in fact be zero patent infringement, but the question is
whether the level of precaution necessary to bring that about would be
more or less socially costly than the harm of infringement itself. Put
another way, the question is: How close is patent infringement to socially
valuable conduct? Returning to the explanation we offered in Part II, it
would be extremely costly if every person were to drive his or her car so
carefully that no automobile accidents ever occurred. Because driving is
(typically) a socially valuable activity that should be allowed to take place,
society tolerates some number of accidents in order to enable individuals to
drive without having to bear the costs of excessive precautions. On the
other hand, intentional murder bears no resemblance to any socially useful
activity, and thus there is no precaution too costly. The optimal rate of
murder is zero, even accounting for the costs of precautions, and that is
why the penalty for murder is infinite or close to infinite-the death
penalty or life in prison-in many jurisdictions.
It should be clear from this brief exposition that the socially optimal
level of patent infringement is well above zero. Engaging in innovative
research and development is immensely valuable, and it would be
tremendously costly for inventors to take the precautions necessary to
ensure that they are never violating an existing patent.25 ' Merely searching
through the existing stores of patents is an immensely costly endeavor;
determining what technologies those patents actually cover is even
costlier. 252 Demanding that a firm never infringe a patent would almost
certainly mean asking that firm to cease operations entirely. 253
The case for tolerating some instances of patent infringement may be
even stronger than it is for automobile accidents. Conduct that presses
against the line between permissible and impermissible may be socially
productive in the patent context, while it is generally not in others.
Consider efforts to "engineer around" a patent. When a firm understands
the boundaries of a patent and tries to evade that patent by developing a
process or product that does not infringe it, sometimes the firm is merely

251. See ADAM B. JAFFE & JOSH LERNER, INNOVATION AND ITS DISCONTENTS: How OUR
BROKEN PATENT SYSTEM Is ENDANGERING INNOVATION AND PROGRESS, AND WHAT TO Do ABOUT IT
59-68 (2004) (describing cases of simultaneous invention and the difficulties that even honest firms
have in avoiding accidental patent infringement).
252. See Jonathan S. Masur, Costly Screens and Patent Examination, 2 J. LEGAL ANALYSIS 687,
697 (2010) (describing the costs created by the "patent thicket" as firms search for patents they might
be infringing).
253. See Manta, supra note 15, at 498-99 (explaining that because of possible overdeterrence by
criminal sanctions, "entire classes of technologies and medicines, even life-saving ones, might never be
developed and reach the market due to inventors' fears of criminal sanctions").

Electronic copy available at: https://ssrn.com/abstract=2297488

20 14] SINNOVA TION AND INCARCERATION 33 1

trying to deny another patent holder its due rewards. But much of the time
the research and development that go into engineering around a patent are
highly socially valuable and result in a superior alternative to the patented
invention. 254 In the course of engineering around, a firm might discover a
cheaper or more efficient way to produce a good or process, or it might hit
upon an improvement to the invention that carries additional benefits. This
is activity that society should encourage, not ban. It would be inappropriate
to increase the penalties for infringement to a level at which they would
discourage these types of innovative activities. Unlike crimes like murder,
there is generally no case for assessing penalties for infringement beyond
the minimum needed to ensure deterrence.
This is not the same as arguing that patent infringement should not be
punished criminally because it is inadvertent. It is extremely common for
states to criminalize inadvertent crimes, even crimes that occur due to risks
the criminal did not even know she was creating. 255 The rationale is the
same as it is for any crime: the need to deter risky behavior and the belief
that civil sanctions will be insufficient to create this deterrence. The key
questions with respect to patent infringement or any other crime are how
risky the behavior is and how close it comes to being socially valuable.
Within criminal law, the criminalization of inadvertent acts is reserved for
situations in which the behavior creates a great deal of risk and does not
resemble socially productive conduct. The law is structured with the
objective of eliminating risky conduct without chilling valuable behavior.
Here, these same considerations militate against criminalizing patent
infringement. As we explained above, the key reason is that inadvertent
infringement, and particularly conduct close to the line, is frequently
innovative and highly socially productive. The optimal level of such
activity is well above zero. To criminalize infringement, under
circumstances when most infringers can pay civil judgments, would be to
excessively and unnecessarily curb this valuable activity.
The relative unimportance of most patent infringement being
inadvertent helps explain why criminal penalties would be improper even
in cases of willful infringement. It might appear to some observers that
willful infringement describes a category of cases for which the optimal
activity level is conceivably zero. Perhaps, they might argue, there is no
reason infringers who know that they are using another party's intellectual

254. See Malani & Masur, supra note 244, at 653-54 (describing the virtues of engineering
around as a socially beneficial activity).
255. See, e.g., CAL. PENAL CODE § 192 (West 2008) (California's manslaughter rule); N.Y.
PENAL LAW § 125.10 (McKinney 2009) (New York's criminally negligent homicide rule).

Electronic copy available at: https://ssrn.com/abstract=2297488

332 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

property should be allowed to do so without licensing it. 256 The patent

owners are being deprived of the just rewards for their inventions, and the
infringers are not engaging in any useful research and development of their
own. Damages for such an activity should be set at an extremely high level,
or perhaps criminal penalties should even be imposed. According to this
line of argument, patent law is right to impose heightened damages in
precisely these cases.
Yet we do not believe the story is so simple. First, many cases of
willful infringement start out as nonwillful infringement and only transform
into willful infringement when the patent owner notifies the infringer of the
existence of the patent.2 57 In these "mixed" cases only the part of the
infringement that occurred after the infringer gained notice of the
infringement should count as willful, and often this is the minority of the
activity. Second, in many cases an infringer will have attempted to obtain a
license from the patent owner but failed. Bargaining breakdowns in
bilateral monopoly situations are common and well understood. The patent
owner may also be making unreasonable demands or be in a position to
hold up the infringer for significant rents. Under these circumstances, one
cannot say that the optimal level of infringement is zero: the concessions
by an infringer necessary to negotiate a license in every case may well be
more socially costly than the infringement itself, particularly when the
patent owner has the right to seek an injunction to halt the infringing
conduct entirely. 258 That is, it may be more socially costly for the infringer
to either cease activity entirely or pay the patent holder's ransom than it is
for the infringement to have taken place.
That is not to say that patent holders should not be compensated for
willful (or any other type of) infringement. A thief is not entitled to steal a
car just because the owner of the car refuses to sell it for the thief's price.
But it is to say that even willful infringement can occur under
circumstances in which it could be very costly to prevent. Accordingly, the
case for even greater damages-beyond what is necessary for standard
deterrence-is weak even with respect to willful infringement.
The final question is whether individuals who infringe patents should

256. See In re Seagate Tech., LLC, 497 F.3d 1360, 1368-72 (Fed. Cir. 2007) (en banc)
(describing willfully infringing conduct and explaining the arguments against tolerating it).
257. Underwater Devices, Inc. v. Morrison-Knudsen Co., 717 F.2d 1380, 1384-85, 1389-90 (Fed.
Cir. 1983) (detailing the practice of patentees using letters to inform potential infringers that they might
be infringing and thereby create the knowledge necessary to trigger liability for willful infringement),
overruled by In re Seagate Tech., 497 F.3d at 1371.
258. 35 U.S.C. § 283 (2006).

Electronic copy available at: https://ssrn.com/abstract=2297488

2014] INNO VA TION AND INCARCERATION 333

be incapacitated-that is, whether they should be actively prevented from

committing further acts of infringement, rather than deterred.259 This
question relates to the preceding discussion because it raises the issue of
whether infringement can ever resemble a socially productive activity, or
whether it is purely harmful. If we believe that infringers merely create
social costs, and we believe that an individual who has infringed once is
likely to do so again, then there is an argument for using incapacitation.
Better directly to block the person from engaging in future acts of
infringement than to rely upon deterrence-which failed once-to
accomplish the same task.
As the foregoing paragraphs should make clear, we do not believe that
this is an accurate depiction of infringing activities. Individuals who
infringe once are not necessarily likely to do so again. The infringement
might have been based upon a mistake or a lack of information.260 Or it
might have been the result of innovative activity, activity that generally
should not be quashed. The individuals involved in the infringement might
also be especially talented, innovative, or productive members of society,
the types of individuals whom it would be most costly to incarcerate. At
some level this is an empirical question about which we can only offer an
educated guess-it might be that there are serial patent infringers who seem
to have no regard for others' intellectual property rights. But we suspect
that infringement is typically a byproduct of normal innovative and
commercial activities, and if this is the case then incarceration would be too
heavy a cudgel to employ.
Moreover, even if we are entirely wrong about this analysis, there are
less costly ways to prevent an individual from infringing in the future than
incarcerating that person. Depending on the individual's prior activities,
she could be barred from serving as an officer or director of a corporation,
or even prevented from working in a particular technological field. This is
analogous to the sanction of taking away an individual's computer that we
discussed in Part II. Such a measure would no doubt be costly, but not

259. This raises interesting questions regarding how the government would go about
incapacitating a corporation, if that is the entity that is found to infringe. We thank Mark Lemley for
raising this point. The corporation itself could be liquidated, though that seems highly overbroad. More
likely, we imagine that the corporate employees involved in the infringement-the scientists who
developed the innovation, and the executives or general counsel who authorized the infringement-
could be imprisoned, or even prevented from working in an innovation-related field after being
released. This penalty would mirror the criminal sanctions in securities fraud cases. There, guilty parties
are often prohibited from working in the financial industry. This penalty may strike the reader as
excessive. It certainly strikes us that way, and in the discussion that follows we will reject it.
260. See supra text accompanying notes 251-53.

Electronic copy available at: https://ssrn.com/abstract=2297488

334 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 87:275

nearly so costly as actually locking the individual behind bars. It should

serve only as an absolute last step-but that means that prison need not be
any type of step at all.
Because the benefit to a patent infringer is capped at the measure of
profits, because patent infringement is likely to be detected, and because
the vast majority of infringers will be capitalized firms that have access to
the profits they have just reaped from the infringement, the case for
criminal sanctions for patent infringement is especially weak. Arguments
regarding activity levels or the need for incapacitation have no greater
purchase. Accordingly, we believe it is appropriate from an economic
perspective that American law imposes no criminal penalties for any acts
related to patent infringement.2 6 ' The only area in which we feel that
criminal penalties might conceivably be desirable involves secret
infringement of patented processes. Even here, where detection is much
less certain, criminal penalties should be used extremely sparingly, if at all.
The economic structure of patents, and the economic realities of innovation
and commercialization, provide superior alternatives.

V. CONCLUSION
Scholars and stakeholders have been wrong to assume that criminal
sanctions for IP infringement are justified on economic grounds. It is true
that criminal sanctions could play an important role in preventing harmful
behavior that cannot be deterred through other means. But imposing
criminal sanctions is costly. For patent infringement, the costs of imposing
criminal sanctions are very unlikely to exceed its benefits. Civil sanctions
will probably be sufficient for creating the optimal amount of infringing
behavior. For copyright infringement, civil sanctions alone should be
sufficient to deter nearly all types of harmful conduct. There is an
economic case for imposing criminal sanctions for copyright infringement
only with respect to a discrete set of activities: massive reproduction and
sales of commercially valuable works. We have arrived at the conclusion
that the case for criminal IP sanctions is weak or nonexistent while putting
to one side both noneconomic considerations, such as moral or
deontological objections to criminal IP sanctions, and economic concerns
about the efficacy of lIP in promoting innovation. Once those considerations
are added to the calculus, we suspect that the argument for criminal
sanctions for IP infringement will disintegrate almost entirely.

261. See 35 U.S.C. §§ 101-329 (Patent Act as amended).

Electronic copy available at: https://ssrn.com/abstract=2297488