COPYRIGHT BACKLASH

DO NOT DELETE 3/22/2012 2:00 PM ARTICLES COPYRIGHT BACKLASH BEN DEPOORTER* ALAIN VAN HIEL‚ SVEN VANNESTEÁ ABSTRACT In the past decade, the entertainment industry has waged a very successful legal campaign against online copyright infringements. In a series of high-profile decisions, content industries have persuaded courts to accept expansive interpretations of contributory enforcement, to create novel doctrines of copyright infringement, and to apply broad interpretations of statutory damage provisions. Many private file sharers, technology companies, university administrators, and Internet service providers have felt the reach of this litigation effort. Yet a significant empirical anomaly exists: even as the copyright industry has ramped up the level of deterrence, online copyright infringements continue unabated. Why has the legal battle against file sharers been so ineffective? The * Professor of Law, University of California, Hastings College of Law; Duke Law School (visiting, Fall 2010) & Lecturer, Ghent University. This Article benefited from comments by Katherine Bartlett, James Boyle, Curtis Bradley, Samuel Buell, Guy-Uriel Charles, Daniel Chen, Robert Cooter, Lisa Griffin, Kimberly Krawiec, Wendy Gordon, Paul Haagen, Gillian Hadfield, Michael Meurer, Gideon Parchomovsky, Pamela Samuelson, Ted Sichelman, Sara Sun-Beale, Molly Van Houweling, Thomas Ulen, Neil Vidmar, and workshop participants at Duke Law School, the University of Pennsylvania Law School, U.C. Berkeley Law School, Stanford Law School, the University of Toronto Law School, Yale Law School, and the 2009 Annual Meeting of the American Association of Law & Economics. Email: [email protected]. ‚ Professor of Social Psychology, Ghent University, Department of Developmental, Personality & Social Psychology. Á BRAI2N/TRI, Department of Neurosurgery, University of Antwerp. 1251 DO NOT DELETE 1252 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 most straightforward explanation is that infringers are not deterred, either because the probability of getting caught remains remote or because the sanctions are not sufficiently salient. If that is the case, the expansive statutory damage award remedies in decisions such as Capitol Records v. Thomas-Rasset and Sony BMG v. Tenenbaum carry renewed promise for the entertainment industry. In this Article we clai m that this deterrence-based approach will prove futile and even counterproductive to the goals of copyright holders. We argue that copyright law faces conditions si milar to Prohibition and other historical episodes of enforcement failure. When infringem ents are widespread, effective deterrence cannot be attained without raising enforcement to levels that undermine the support for the underlying rules. As a result, enforcement has the inadvertent effect of moving behavior in the opposite direction from that intended by the law. In the context of copyright law, enforcement has increased the gap between the social and legal perceptions of the law. Because file sharers, as a group, perceive copyright litigation as excessive, this inadvertently strengthens opposition to the legally protected interests of copyright law. To further our understanding of the interplay between enforcement and public attitudes, we conduct two empirical studies on norms and copyright law. The results confirm that copyright enforcement is a doubleedged sword. While stringent sanctions have a modest deterrent effect on file-sharing behavior, they increase anti-copyright senti ments among frequent offenders. This raises a spectacular challenge for copyright enforcement: the more copyright owners push to step up sanctions for copyright infringements, the more the public resents the protected rights. Consequently, stepping up sanctions tends to increase ²rather than decrease²the rate and frequency of infringing activities. Our key results suggest, therefore, that more stringent copyright enforcement will further erode respect for copyrights and may prove counterproductive to copyright owners. TABLE OF CONTENTS I. INTRODUCTION ............................................................................... 1253 II. BACKGROUND: THE LITIGATION CAMPAIGN BY CONTENT INDUSTRIES .................................................................. 1258 III. LITIGATION AND NORMATIVE OVERDETERRENCE ........... 1264 A. LITIGATION AND NORMATIVE BACKLASH .............................. 1264 B. THE INADVERTENT COSTS OF DETERRENCE ........................... 1267 DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1253 IV. STUDY ............................................................................................. 1272 A. INTRODUCTION & MAIN RESULTS .......................................... 1272 B. METHODOLOGY AND CONSTRUCTION OF VARIABLES............ 1273 1. Study 1 .............................................................................. 1275 2. Study 2 .............................................................................. 1276 3. Analysis ............................................................................ 1277 C. FINDINGS ................................................................................. 1278 V. DISCUSSION .................................................................................... 1284 A. PUBLIC ATTITUDES AND SELF-INTEREST ............................... 1284 B. THE IRONY OF DETERRENCE ................................................... 1287 C. ENFORCEMENT AND EXPECTATIONS ....................................... 1288 VI. CONCLUSION ................................................................................ 1290 APPENDIX ............................................................................................. 1292 I. INTRODUCTION Digital downloading and file sharing present unprecedented challenges to the enforcement of copyright law.1 These new technologies greatly facilitate unauthorized reproduction and distribution of copyrighted material. Evidence suggests that there are roughly ten million simultaneous users of peer-to-peer technology at any given time.2 Sharing unlicensed FRS\ULJKWHG PDWHULDOV LV QRZ D SDUW RI WHHQDJHUV¶ HYHU\GD\ OLYHV WKH amount of time spent downloading unlicensed content significantly outweighs that of visiting record stores or downloading legally from 1. JESSICA LITMAN, DIGITAL COPYRIGHT 111 (2001) (reviewing the various challenges presented by new technology). 2. John Boudreau, Illegal File Sharing Showing No Letup, SEATTLE TIMES, July 3, 2006, available at http://seattletimes.nwsource.com/html/businesstechnology/2003101281_btfilesharing03 .html. See also John Borland, Open-Source P2P Projects Keep Swapping, CNET NEWS (July 15, 2005), http://news.cnet.com/2100-1032_3-5789087.html; Simon Crerar, Illegal File-Sharing As Popular As Ever, TIMES (London), Jan. 19, 2006, available at http://entertainment.timesonline.co.uk/ tol/arts_and_entertainment/music/article715426.ece. After the outcomes in Napster and Grokster, filesharing activities relocated to various BitTorrent sites. See Ernesto, Filesharing Report Shows Explosive Growth for uTorrent, TORRENTFREAK (Apr. 26, 2008), http://torrentfreak.com/p2p-statistics-080426 (reporting that the number of uTorrent users worldwide more than doubled in 2008 compared to 2009). Because of the decentralized nature of the most current generation of file-sharing technologies, it is increasingly more difficult to obtain accurate data on the levels of file sharing; however, several indicators suggest that file sharing remains widespread. See Robert Siciliano, P2P File Sharing on College Ca mpuses, MCAFEE BLOG CENT. (Sept. 25, 2010, 3:46 PM), http://blogs.mcafee.com/ consumer/identity-theft/p2p-file-sharing-on-college-campuses (describing inadvertent instances of file sharing); Kate Holton, U.S. Single Digital Music Sales F lat This Year: Nielsen, REUTERS (Sept. 27, 2010, 4:07 PM), http://www.reuters.com/article/2010/09/27/us-sales-us-idUSTRE68Q2FM20100927 (discussing how digital music sales in the U.S. have stalled). DO NOT DELETE 1254 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 authorized sources, such as iTunes.3 The average teenager has an average of over 800 illegally copied songs on his or her digital music player or personal computer.4 This is a remarkable trend given the intensive efforts by the entertainment industry to halt the unauthorized dissemination of copyrighted material online. Over the past decade, copyright holders have waged a highly successful legal campaign.5 In a series of high-profile court decisions, the entertainment industry has persuaded courts to accept expansive interpretations of intermediary liability,6 to create novel doctrines of copyright infringement,7 and to apply broad interpretations of statutory damage provisions.8 The reach of this litigation strategy has been widely felt. For example, novel, innovative technologies have been banned,9 university administrators10 and content-sharing websites have 3. See David Goldman, 0XVLF¶V/RVW'HFDGH6DOHV&XWLQ+DOI, CNN MONEY (Feb. 3, 2010, 9:52 AM), http://money.cnn.com/2010/02/02/news/companies/napster_music_industry/index.htm? hpt=T2 (reporting a drop in music sales from $14.6 billion in 1999 to $6.3 billion in 2009); Ethan Smith, Sales of Music, Long in Decline, Plunge Sharply, WALL ST. J., Mar. 21, 2007, available at http://online.wsj.com/public/article/SB117444575607043728-oEugjUqEtTo1hWJawejgR3LjRAw_ 20080320.html?mod=rss_free. 4. In a sample of 5600 consumers who were willing to share their iPod listening statistics, the average player held a collection of over 3500 songs. Paul Lamere, :KDW¶V 2Q <RXU L3RG", ORACLE BLOGS (May 22, 2006), http://blogs.sun.com/plamere/entry/what_s_on_your_ipod. See also Dan Sabbagh, Average THHQDJHU¶V L3RG +DV  ,OOHJDO 0XVLF 7UDFNV, TIMES (London), June 16, 2008, available at http://technology.timesonline.co.uk/tol/news/tech_and_web/personal_tech/article4144585. ece. 5. For a description and discussion of this campaign, see infra Part II. 6. A&M Records, Inc. v. Napster, Inc. (Napster I), 239 F.3d 1004, 1021±24 (9th Cir. 2001) (confirming district court finding that plaintiffs would likely be successful in establishing that Napster would be liable as a contributory and vicarious infringer). 7. MGM Studios, Inc. v. Grokster, Ltd. ( Grokster III), 545 U.S. 913, 936±37 (2005) (holding that distributing software with the manifest intent to promote copyright infringement can render the VRIWZDUH¶VGLVWULEXWRUOLDEOHIRUWKHLQIULQJLQJDFWLRQVRIWKLUGSDrties). 8. Sony BMG Music EQWP¶W Y. Tenenbaum, No. 07cv11446-NG, 2009 U.S. Dist. LEXIS 115734 (D. Mass. Dec. 7, 2009). See also Dave Itzkoff, Student F ined $675,000 in Downloading Case , N.Y. TIMES (July 31, 2009, 12:34 PM), http://artsbeat.blogs.nytimes.com/2009/07/31/judge-rulesstudent-is-liable-in-music-download-case/. In subsequent litigation, the award of $675,000 was reduced by the district court to $67,500, but was reinstated by the First Circuit. See 6RQ\%0*0XVLF(QWP¶WY Tenenbaum, 721 F. Supp. 2d 85 (D. Mass. 2010), UHY¶G, Nos. 10-1883, 10-1947, 10-2052, U.S. App. LEXIS 19086 (1st Cir. Sept. 16, 2011). See also Milton J. Valencia, Court Upholds F ine in Music Download Case , BOSTON GLOBE (Sept. 20, 2011), http://www.boston.com/news/local/massachusetts/ articles/2011/09/20/federal_appeals_court_reinstates_675k_fine_against_student_convicted_of_downlo ading_music/. 9. Napster I, 239 F.3d at 1021±24 (confirming that the a peer-to-peer file sharing program could be liable as a contributory and vicarious infringer); UMG Recordings, Inc. v. MP3.com, Inc., 92 F. Supp. 2d 349, 350±52 6'1<  KROGLQJWKDWGHIHQGDQW¶VRQOLQHSRVWLQJRI03ILOHVIRUDFFHVV by individuals who could prove that they owned a CD copy was not protected fair use). 10. Anne Broache, RIAA Threatens 19 Universities with Lawsuits, CNET NEWS (Oct. 18, 2007, DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1255 received cease-and-desist letters,11 and over 20,000 individuals have entered into settlement agreements in order to avoid costly litigation.12 So why has the legal battle against file sharers been so ineffective?13 One plausible explanation is that individuals are not sufficiently deterred because they are convinced that the probability of getting caught remains remote given the vast amount of downloads that occur at any given moment in time. If so, two court decisions bring renewed promise to the HQWHUWDLQPHQWLQGXVWU\¶VOLWLJDWLRQHIIRUW,QRQHKLJK-profile case, a single mother was ordered to pay $1.92 million for sharing twenty-four songs online.14 In another highly publicized file-sharing case, a jury levied $675,000 on a Boston University graduate student for infringing copyrights on thirty songs.15 Moreover, if an increase in the level of deterrence is the solution, the entertainment industry is right to find inspiration in current French legislation that restricts Internet access to copyright violators.16 11:16 AM), http://news.cnet.com/8301-10784_3-9799840-7.html. 11. See, e.g., Seth Sutel, MTV Owner Viacom Demands That YouTube Remove More Than 100,000 Unauthorized Video Clips, ASSOCIATED PRESS, Feb. 3, 2007, available at http://www.usatoday.com/tech/news/2007-02-02-viacom-youtube_x.htm. 12. How To Not Get Sued for File Sharing, ELECTRONIC FRONTIER FOUND. (July 2006), http://www.eff.org/wp/how-not-get-sued-file-sharing. 13. See, e.g., Crerar, supra note 2; Ernesto, supra note 2. 14. Capitol Records, Inc. v. Thomas-Rasset, No. 06-CV-01497 (MJD/LIB), 2009 WL 2030495 (D. Minn. June 18, 2009). See also David Kravets, Jury in RIAA Trial S laps $2 Million F ine on Ja mmie Thomas, WIRED (June 18, 2009, 6:57 PM), http://www.wired.com/threatlevel/2009/06/riaa-jury-slaps2-million-fine-on-jammie-thomas; David Kravets, RIAA Jury F inds Minnesota Woman Liable for Piracy, Awards $222,000, WIRED (Oct. 4, 2007, 2:34 PM), http://www.wired.com/threatlevel/ 2007/10/riaa-jury-finds/. 15. Sony BMG Music EQWP¶W Y. Tenenbaum, No. 07cv11446-NG, 2009 U.S. Dist. LEXIS 115734 (D. Mass. Dec. 7, 2009). See also Itzkoff, supra note 8. 16. On October 22, 2009, the Constitutional Council of France approved the HADOPI law or Creation and Internet Law. Loi 2009-669 du 12 juin 2009 favorisant la diffusion et la protection de la creation sur internet [Law 2009-669 of June 12, 2009 on Promoting the Dissemination and Protection of Creation on the Internet], JOURNAL OFFICIEL DE LA REPUBLIQUE FRANÇAISE [J.O.] [OFFICIAL GAZETTE OF FRANCE], June 13, 2009, No. 2, (English translation available at http://www.conseilconstitutionnel.fr/conseil-constitutionnel/root/bank_mm/anglais/2009_580dc.pdf); Eric Pfanner, F rance Approves Wide Crackdown on Net Piracy, N.Y. TIMES, Oct. 22, 2009, available at http://www.nytimes.com/2009/10/23/technology/23net.html. The law includes a controversial ³threestrike´ procedure, which requires Internet service providers to suspend the Internet service of the connection owner, as opposed to the actual accused infringer, for a period between two months to one year. The accused owner is blacklisted and third-party Internet service providers are prevented from providing the accused owner an Internet connection. See Pfanner, supra . The United Kingdom joined the approach of web disconnection measures with the passage of the Digital Economy Act on June 8, 2010. Digital Economy Act, 2010, 59 Eliz. 2, c. 24, (U.K.). The Digital Economy Act includes provisions that limit the quality of Internet connections or that block access to a location on the Internet ³IURP ZKLFK D VXEVWDQWLDO DPRXQW RI PDWHULDO KDV EHHQ LV EHLQJ RU LV OLNHO\ WR EH PDGH DYDLODEOH LQ LQIULQJHPHQWRIFRS\ULJKW´RUDORFDWLRQZKLFK³IDFLOLWDWH[s]´VXFKEHKavior. Id. § 17. See also Emma DO NOT DELETE 1256 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 In this Article, we claim that this deterrence-based approach will prove futile and even counterproductive to the goals of copyright holders. When noncompliance and infringements are widespread, effective deterrence cannot be attained without raising enforcement to levels that undermine support for the underlying rules. But when enforcement reaches levels that are perceived as normatively excessive, this can have the inadvertent effect of moving behavior in the opposite direction from that intended by the law. If individuals perceive enforcement as excessive, this may reinforce or even strengthen a belief that the legal regime is not legitimate or that a legal rule is unjust.17 Once the public resents the protected rights, the stepping up of sanctions may increase²rather than decrease²the rate and frequency of infringing activities. As historians have argued, countervailing public reactions have played a significant role in the demise of the Prohibition movement and other historical episodes of enforcement failure.18 We argue that the same dynamic is at work today in the enforcement of copyright law by the content industries. It is becoming increasingly clear that the deterrence-based approach to copyright enforcement is severely damaging the public image of copyright industries specifically, and copyright law more generally. Because unauthorized file sharing has a normative component,19 the litigation initiatives of the content industries adversely affect the public attitude toward copyright law. Overall then, the litigation campaign might be counterproductive to the interests of copyright holders. Previous work on counterproductive normative effects has been theoretical in nature.20 We aim to narrow the gap between theory and Barnett, MPs Pass Digital Economy Bill , TELEGRAPH (U.K.), Apr. 8, 2010, available at http://www.telegraph.co.uk/technology/news/7566427/MPs-pass-Digital-Economy-Bill.html. 17. See infra Part III.B. 18. See infra notes 83±89 and accompanying text. 19. Yuval Feldman & Janice Nadler, The Law and Norms of F ile Sharing, 43 SAN DIEGO L. REV. 577, 584±87 (2006) (presenting evidence on the various sources of anti-copyright norms); Daniel J. Gervais, The Price of Social Norms: Towards a Liability Regi me for File-Sharing, 12 J. INTELL. PROP. L. 39, 48±53 (2004); Lior Jacob Strahilevitz, Charismatic Code, Social Norms, and the E mergence of Cooperation on the File-Swapping Networks, 89 VA. L. REV. 505, 534±47 (2003) (describing norms in file-sharing communities). 20. See generally Dan M. Kahan, Gentle Nudges vs. Hard Shoves: Solving the Sticky Norms Problem, 67 U. CHI. L. REV. 607 (2000) (suggesting that gradual approaches are more effective when anti-legal norms are in place); Francesco Parisi & Georg von Wangenheim, Legislation and Countervailing Effects from Social Norms (George Mason Univ. Sch. of Law, Law and Economics Working Paper Series No. 04-31, 2004), available at http://www.gmu.edu/departments/ law/faculty/papers/docs/04-31.pdf (describing a cycle of opinion formation in which public acts of DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1257 empiricism by conducting two experimental studies. Copyright infringement presents a natural case study on this subject, given the emergence of an anti-copyright culture among teenagers and the strong normative component of unauthorized file sharing.21 We examine backlash effects as a possible explanation of the pervasiveness of file sharing in the face of copyright litigation. Specifically, we seek to test the hypothesis that punitive sanctions have the unintended effect of strengthening resistance to copyright enforcement among file sharers. We test this hypothesis in two studies that examine the influence of copyright enforcement on (1) selfreported normative attitudes in the first phase and (2) download behavior when the probability of detection falls to zero in the second phase. In this second phase, we manipulated the severity and certainty of punishment, while holding the expected costs of punishment constant. The results suggest that when normative positions conflict with legal rules, enforcement is a double-edged sword. First, we find that punitive enforcement creates a backlash effect by strengthening prior anti-copyright positions among frequent copyright offenders. Second, punitive enforcement measures are most valuable²both on a normative and behavioral level²among infrequent offenders who have not yet internalized any normative position against copyright law. Third, raising the level of the sanction, rather than increasing the probability of apprehension, is a more effective means to induce deterrence. Our findings present a vexing policy dilemma for copyright holders. Because technological obstacles prevent perfect enforcement on the Internet, the content industries cannot afford to neglect public attitudes. At the same time, in order to maintain effective levels of enforcement, legal sanctions must be elevated to levels that appear counterproductive from a normative perspective. Our findings show a backlash effect for all offenders, but frequent offenders especially, who will download more if previously faced with punitive enforcement measures. The observed backlash effect among frequent infringers casts doubt upon the current enforcement tactics employed by the entertainment industry. By focusing litigation exclusively on frequent offenders, copyright holders bolster anti-copyright positions among this group, while foregoing opportunities to promote pro-copyright norms among occasional infringers. Our key results suggest, therefore, that the imposition of stricter penalties for copyright infringements will further disobedience and protest undermine the legitimacy of legislation, which leads to further opposition). See also William J. Stuntz, Self-Defeating Cri mes, 86 VA. L. REV. 1871, 1872 (2000) (suggesting that misguided enforcement priorities can inadvertently shift public support away from the underlying laws). 21. See supra sources cited in note 19. See also infra Part III. DO NOT DELETE 1258 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 erode respect for copyright law and may prove counterproductive to the copyright industries. This Article unfolds as follows. Part II describes the litigation campaign waged by the content industries in the past decade. Part III discusses the normative effect of litigation as an enforcement strategy. In Part IV, we explain the methodological approach and major findings of our studies. Part V relates our findings to some of the policy issues involved in the enforcement of copyrights. Part VI concludes. II. BACKGROUND: THE LITIGATION CAMPAIGN BY CONTENT INDUSTRIES When online file sharing initially became widespread, content industries disputed the legality of file sharing.22 Once courts established that noncommercial file sharing by private users directly infringed the rights of copyright holders, the entertainment industry pursued the developers of file-sharing platforms on the basis of intermediary liability. And so, the first generation of file-sharing technology, represented by Napster, became the subject of a high-profile litigation battle. At the end of this closely watched litigation, the Ninth Circuit established that copyright law implicates developers of centralized peer-to-peer technology by way of contributory liability: file-sharing technologies materially contributed to copyright infringements if they provided ³WKH VLWH DQG IDFLOLWLHV´ WKDW enabled direct infringements, such as hosting a central list of the files on eacKXVHU¶VFRPSXWHU23 When, however, the final verdict in Napster arrived, a new generation of file-sharing applications was already in use. Though functionally equivalent to their predecessors, these file-sharing platforms evaded the 22. Before file-sharing technologies, content industries had won various copyright claims against commercial applications involved in unauthorized reproductions and novel Internet services that reproduced music. For instance, record companies won a copyright suit against MP3.com, which allowed subscribers to play music that they owned, borrowed, or had previously purchased over the Internet. UMG Recordings, Inc. v. MP3.com, Inc., 92 F. Supp. 2d 349, 350 (S.D.N.Y. 2000). The court GLG QRW XSKROG WKH GHIHQVH¶V DUJXPHQW WKDW WKH VHUYLFH RI 03FRP PHUHO\ DOORZHG VXEVFULEHUV WR ³VSDFH VKLIW´ VRXQG UHFRUGLQJV WKDW WKH\ RZQHG ZLWKRXW FDUU\LQJ DURXQG SK\VLFDO &'V EHFDXVH WKH service was neither transformative nor productive. Id. at 351. According to the court, the use of a different medium did not render the use transformative. Id. See also Infinity Broad. Corp. v. Kirkwood, 150 F.3d 104, 108±09 (2d Cir. 1998) (rejecting the fair use defense by the operator of a service that retransmitted copyrighted radio broadcasts over telephone lines); L.A. News Serv. v. Reuters Television ,QW¶O /WG., 149 F.3d 987, 994±95 (9th Cir. 1998) (rejecting the fair use defense by television news agencies that copied copyrighted news footage and retransmitted it to news organizations). 23. A&M Records, Inc. v. Napster, Inc. (Napster I), 239 F.3d 1004, 1022 (9th Cir. 2001). DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1259 contributory liability standard applied in Napster by decentralizing the technology and removing the role of central servers.24 When content industries challenged this second generation of technologies, they were initially unsuccessful in the district court and appellate court stages. Courts rejected the application of Napster to decentralized file-sharing services because liability for contribuWRU\ LQIULQJHPHQW LPSOLHV ³DFWXDO NQRZOHGJH of infringement at a time when [file-sharing services] can use that knowledge to stop the parWLFXODULQIULQJHPHQW´25 Ultimately, however, the industry was victorious again. The Supreme Court reversed the lower court decisions, holding that software providers are accountable for copyright violations if they invoke copyright-infringing uses and take active steps to that end.26 Adopting the inducement concept from patent law,27 the Supreme Court held that copyright holders can bring VXLWDJDLQVWFRPPHUFLDODJHQWVZKRGLVWULEXWHSURGXFWV³ZLWKWKHREMHFWRI promoting its use to infringe copyright, as shown by clear expression or RWKHU DIILUPDWLYH VWHSV WDNHQ WR IRVWHU LQIULQJHPHQW´28 Lower courts are currently in the process of applying the inducement standard to a third generation of more decentralized, torrent-based file-sharing technologies. The outcome of this litigation, and more generally, the immediate fate of torrent-based technologies, largely depends on the evidence required to esWDEOLVK WKH H[LVWHQFH RI ³DFWLYH VWHSV . . . taken with the purpose of EULQJLQJDERXWLQIULQJLQJDFWV´29 Despite obtaining legal successes against developers of software DSSOLFDWLRQV WKDW ³IDFLOLWDWHG´ RQOLQH FRS\ULJKW LQIULQJHPHQWV30 content industries were unable to prevent the further development and distribution of additional file-sharing technologies. And so, as the number of online 24. See, e.g., Kristina Groennings, &RVWV DQG %HQHILWV RI WKH 5HFRUGLQJ ,QGXVWU\¶V /LWLJDWLRQ Against Individuals, 20 BERKELEY TECH/-   ³7KH>UHFRUGLQJ@LQGXVWU\¶VYLFWRU\LQ Napster was fleeting as publicity over the issue increased awareness of P2P technology and users flocked to decentralized networks like Grokster and KaZaa, making the tracking of P2P use more GLIILFXOW´) (footnote omitted). 25. MGM Studios, Inc. v. Grokster, Ltd. ( Grokster I), 259 F. Supp. 2d 1029, 1037 (C.D. Cal. 2003) (citing Napster I, 239 F.3d at 1020±22), DII¶G, MGM Studios, Inc. v. Grokster, Ltd. ( Grokster II), 380 F.3d 1154 (9th Cir. 2004). 26. MGM Studios, Inc. v. Grokster, Ltd. ( Grokster III), 545 U.S. 913, 940±41 (2005). 27. Id. at 935 (citing Henry v. A. B. Dick Co., 224 U.S. 1, 48±49 (1912); Kalem Co. v. Harper Bros., 222 U.S. 55, 62±63 (1911); Thomson-Houston Elec. Co. v. Kelsey Elec. Ry. Specialty Co., 75 F. 1005, 1007±08 (2d Cir. 1896); Rumford Chem. Works v. Hecker, 20 F. Cas. 1342, 1346 (C.C.D.N.J. 1876) (No. 12,133)). 28. Id. at 936±37. 29. Id. at 938. 30. See supra text accompanying note 23. DO NOT DELETE 1260 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 copyright infringements continued to skyrocket, the entertainment industry redirected its focus to users of file-sharing technologies. In September 2003, the Recording Industry Association of America (³RIAA´) began sending subpoenas to Internet service providers, demanding the names of individuals who were allegedly sharing music on file-sharing networks. The lawsuits targeted individuals who stored large amounts of music files in publicly accessible folders on their computers. These cases settled at an average of $3000.31 A second wave of lawsuits followed in October 2003, when the RIAA initiated eighty additional lawsuits against individual peerto-peer file sharers.32 From September 2003 to December 2004, the recording industry issued over 3400 individual lawsuits against users of peer-to-peer file-sharing technology.33 In November 2004, the Motion Picture Artists Association ³MPAA´ joined the fray34 when it launched a set of lawsuits against individuals who had allegedly shared a substantial amount of movies online.35 By August 2005, the MPAA had initiated six rounds of lawsuits against individual file traders.36 As of November 2010, the RIAA filed at least 18,000 lawsuits against individuals.37 Additionally, content industry representatives have sent pre-litigation letters to universities requesting students to come forward to pay a non-negotiable settlement amount.38 31. This average gradually increased to $3000. Paul Roberts, 5,$$ 6XHV  µ-RKQ 'RHV,¶ PCWORLD (Jan. 21, 2004, 1:00 PM), available at http://www.pcworld.com/article/114387/riaa_sues_ 532_ john_does.html. 32. RIAA Launches Second Wave of File-Swapper Suits, OUT-LAW NEWS (Oct. 31, 2003), http://www.out-law.com/page-4029. 33. F ile Sharing Goes to High Court , WIRED (Dec. 10, 2004), http://www.wired.com/ entertainment/music/news/2004/12/65995. 34. Because the amount of motion pictures being exchanged over peer-to-peer networks had increased dramatically in 2004, mainly due to increased broadband width and improved compression technologies, the movie industry group would no longer sit back. See, e.g., Grant Gross, MPAA to Sue Movie File Swappers, PCWORLD (Nov. 4, 2004, 4:00 PM), http://www.pcworld.com/article/118485/ mpaa_to_sue_movie_file_swappers.html. 35. Hollywood Sues Alleged F ile Swappers, MSNBC.COM (Nov. 16, 2004, 8:34 PM), http://www.msnbc.msn.com/id/6504024/ns/technology_and_science-tech_and_gadgets/t/hollywoodsues-alleged-file-swappers/. 36. See Thomas Mennecke, Movie Studios vs. Internet Movie Thieves, Round Six!, SLYCK NEWS (Aug. 2, 2005), http://www.slyck.com/story877_Movie_Studios_Vs_Internet_Movie_Thieves_Round_ Six. 37. Andrei Toma, U S C G Making the RIAA Look Soft, CYBERBEAR (Nov. 24, 2010), http://cyberbeartracks.com/?p=95 ³7KHFDPSDLJQHQGHGMXVWODVW\HDUZLWKRYHUODZVXLWV´  38. See Press Release, RIAA, 23 New Schools to Receive Latest Round of RIAA Pre-Lawsuit Letters (July 18, 2007), http://www.riaa.com/newsitem.php?id=780E8751-0E03-4258-D651F991B66E1708. See, e.g. , Nick Semenkovich, RIAA Pre-Litigation Letters Sent to MIT , TECH, May 8, 2007, at 1. DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1261 In 2008, a few individuals refused the settlement offers by the RIAA. Although the 5,$$¶V threat letters routinely claimed that court awards could be as much as $150,000, this claim had never been tested in court. In the ensuing litigation, an initial issue concerned the required proof of dissemination of a copyrighted work. Although it can easily be demonstrated that, at one time, a file sharer had a file in a publicly accessible folder, it is very difficult, as a technical matter, to provide evidence that the file has actually been downloaded from the file VKDUHU¶V computer.39 Courts ultimately decided that dissemination could be presumed on the basis of accessibility.40 The next issue concerned the alleged harm caused by an individual accused of file sharing.41 While it is hard to estimate the overall impact of file sharing on the revenue base of content industries,42 it is even more difficult to discern the precise damage imposed by an individual act of file sharing.43 Fortunately for copyright holders, the Federal Copyright Act does not require copyright holders to 39. See Capitol Records, Inc. v. Thomas, 579 F. Supp. 2d 1210, 1218±19 (D. Minn. 2008) (holding that liability for violation of the exclusive right of distribution requires actual dissemination). The Capitol Records court cited Hotaling v. Church of Jesus Christ of Latter-Day Saints, 118 F.3d 199, 203 (4th Cir. 1997) (holding that when a library makes a work available to the public it has completed all the steps necessary for distribution), but the court declined to follow it. 40. See Capitol Records, 579 F. Supp. 2d at 1221. 41. Because most suspected infringers likely prefer not to incur the costs of litigation, some of the more complex legal issues relating to the enforcement of copyright with regard to file sharing have not been subject to rigorous litigation. 42. Some studies, for instance, contest the claim that file sharing has a negative impact on music industry sales. See, e.g., Felix Oberholzer-Gee & Koleman Strumpf, The Effect of File Sharing on Record S ales: An E mpirical Analysis, 115 J. POL. ECON. 1, 3 (2007); Andy McCue, Study: Falling C D 6DOHV&DQ¶W%H%ODPHGRQ32P, CNET NEWS (June 14, 2005, 11:40 AM), http://news.cnet.com/21001027_3-5746291.html (discussing a study finding that file sharing has no negative impact on CD sales). But see Stan J. Liebowitz, F ile Sharing: Creative Destruction or Just Plain Destruction?, 49 J.L. & ECON. 1, 14±17 (2006) [hereinafter Liebowitz, Creative Destruction] (presenting evidence that file sharing diminishes recording industry revenues); Rafael Rob & Joel Waldfogel, Piracy on the High &¶V0XVLF'RZQORDGLQJ6DOHV'LVSODFHPHQWDQd Social Welfare in a Sample of College Students, 49 J.L. & ECON. 29, 30 (2006) (showing data indicating that downloading reduced purchases by individuals in their sample by about 10 percent during 2003). See generally Stan J. Liebowitz, Economists Exa mine File Sharing and Music Sales, in INDUSTRIAL ORGANIZATION AND THE DIGITAL ECONOMY 145 (Gerhard Illing & Martin Peitz eds., 2006) (providing an overview of the empirical work that supports the position that file sharing hurts copyright owners). 43. For instance, in order to estimate the harm imposed by one file in a public accessible folder, one needs to gather information about (1) how many times that the file was downloaded, and (2) how many of these downloads caused individuals not to buy the song through legally available alternatives. Such information is simply not available. Torrent technologies complicate the former inquiry because any act of downloading captures fragmented parts from any number of different users. For a general explanation of the operation of peer-to-peer networks, see Tim Wu, :KHQ&RGH,VQ¶W/DZ, 89 VA. L. REV. 679, 745±50 (2003). DO NOT DELETE 1262 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 provide evidence of actual damages in all circumstances.44 The Copyright Act allows copyright holders to elect to apply statutory damages at any time during the litigation,45 and as it turns out, courts are willing to apply statutory damage provisions quite liberally. In 2009, a defendant was ordered to pay $1.92 million for sharing twenty-four songs online.46 In another case, a jury levied $675,000 against a defendant for sharing thirty songs.47 These high awards are possible because the Copyright Act of 1976 changed the old rule of awards per infringement to awards per infringed work.48 Moreover, courts adopted a broad interpretation of ³willful infringement,´ which in turn enhanced the potential statutory award.49 The generous application of statutory awards certainly bolsters the deterring HIIHFWVRIWKHHQWHUWDLQPHQWLQGXVWU\¶s litigation effort.50 The litigation campaign introduced several legal innovations, each 44. Beginning in 1790, the first Congress enacted the original Federal Copyright Act. Copyright Act of 1790, ch. 15, 1 Stat. 124. The original purpose of statutory damages was to provide a minimum award to copyright owners because of the difficulty of measuring actual damages and profits. See Stephanie Berg, Remedying the Statutory Da mages Remedy for Secondary Copyright Infringement Liability: Balancing Copyright and Innovation in the Digital Age , 56 J. COPYRIGHT SOC¶Y U.S.A. 265, 273±74 (2009). A funGDPHQWDO XQGHUSLQQLQJ RI &RQJUHVV¶V HQDFWPHQW RI WKH &RS\ULJKW $FW RI , Pub. L. No. 94-553, 90 Stat. 2541, was its concern with potentially outrageous statutory awards. Berg, supra , at 297±98. Congress attempted to circumvent such results by using a novel rhetorical explication in the 1976 Act. Pamela Samuelson & Tara Wheatland, Statutory Damages in Copyright Law: A Remedy in Need of Reform, 51 WM. & MARY L. REV. 439, 451±63 (2009). According to Pamela Samuelson and Tara Wheatland, however, Congress did not limit anything and only exacerbated its own unease. Id. at 453±54. In pertinent part, Samuelson and Wheatland argue that the 1976 Act has not had the desired impact Congress anticipated, as evidenced by the latest statutory damage awards in peer-to-peer file sharing cases. Id. 45. 17 U.S.C. § 504(c) (2006). The 1976 Act limits the availability of statutory damages to copyright holders who register their works. If an infringement was committed willfully, the court may increase the award of damages to a sum of $150,000. Id. 46. Capitol Records, Inc. v. Thomas-Rasset, No. 06-CV-01497 (MJD/LIB), 2009 WL 2030495 (D. Minn. June 18, 2009). See also Kravets, supra note 14. 47. Sony BMG Music EQWP¶W Y. Tenenbaum, No. 07cv11446-NG, 2009 U.S. Dist. LEXIS 115734 (D. Mass. Dec. 7, 2009). See also Itzkoff, supra note 8. 48. 17 U.S.C. § 504(c). 49. Kate Cross, Note, David v. Goliath: How the Record Industry is Winning Substantial Judgments Against Individuals for Illegally Downloading Music , 42 TEX. TECH L. REV. 1031, 1041±42 (2010) (discussing whether excessive punitive damages jurisprudence should be applied to statutory damages cases in which defendants have received disproportionately high, and arguably punitive-like, statutory damages judgments against them). It has also been suggested that courts should apply the modern punitive damages test set forth in BMW of N. Am., Inc. v. Gore, 517 U.S. 559 (1996), to copyright cases that involve grossly disproportionate statutory judgments against individual infringers. ³,IRQHVRQJRQL7XQHVFRVWVQLQHW\-nine cents to purchase, then a judgment awarding $80,000 for one song is not only grossly disproportionate but µobviously unreasonable by any measure.¶´Cross, supra, at 1038. 50. Samuelson & Wheatland, supra note 44, at 452±56. DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1263 with a different set of consequences. First, content industries managed to expand the scope of their rights by extending the doctrine of intermediary liability. These doctrinal developments increase the burden on software developers by focusing on (1) the potential for technological enhancements to enable developers to monitor and control direct infringement, and (2) the intentions of indirect infringers as a source of infringement. In this process, courts also removed some of the safe harbor assumptions that were traditionally associated with the dual nature of many technologies.51 Second, despite the legal uncertainty, content industries aggressively enforced these evolving entitlements against private individuals. The resulting mass-settlement campaign introduced a form of reverse class action litigation against dispersed, poorly organized defendants.52 Finally, copyright holders convinced courts to apply expansive remedies to enforce online copyright infringements.53 The legal success of this campaign has not, however, produced a satisfactory reduction of online copyright infringements for content industries.54 This has prompted the industries to forge cooperation with service providers55 and to pursue new legislation that would block Internet access to accused infringers.56 In the next part, we document the impact of the litigation campaign on 51. The courts in Napster and Grokster avoided the issue of noninfringing uses. See A&M Records, Inc. v. Napster, Inc. (Napster I), 239 F.3d 1004, 1021 (9th Cir. 2001); MGM Studios, Inc. v. Grokster, Ltd. ( Grokster I ), 259 F. Supp. 2d 1029, 1037 (C.D. Cal. 2003). 52. See Assaf Hamdani & Alon Klement, The Class Defense, 93 CALIF. L. REV. 685, 688 (2005) (advocating a procedure of claim consolidation among similarly positioned defendants); David W. Opderbeck, Peer-to-Peer Networks, Technological Evolution, and Intellectual Property Reverse Private Attorney General Litigation, 20 BERKELEY TECH. L.J. 1685, 1696   FRLQLQJWKHSKUDVH³UHYHUVH SULYDWH DWWRUQH\ JHQHUDO OLWLJDWLRQ´ WR UHIHU WR DJJUHJDWHG FODLPV LQLWLDWHG E\ ZHOO-resourced plaintiffs against multiple individual defendants). 53. See, e.g., Sony BMG Music EntP¶W v. Tenenbaum, No. 07cv11446-NG, 2009 U.S. Dist. LEXIS 115734 (D. Mass. Dec. 7, 2009); Capitol Records Inc. v. Thomas-Rasset, No. 06-CV-1497, 2009 WL 2030495 (D. Minn. June 18, 2009). 54. See Goldman, supra note 3; Smith, supra note 3 (reporting a decline in music sales over the past decade). See also Crerar, supra note 2; David Kravets, F ile Sharing Lawsuits at a Crossroads, After 5 Years of RIAA Litigation, WIRED (Sept. 4, 2008, 2:55 PM), http://www.wired. com/threatlevel/2008/09/proving-file-sh/ (quoting Electronic Frontier Foundation attorney Fred von /RKPDQQ¶V HYDOXDWLRQ Rf WKH OLWLJDWLRQ FDPSDLJQ ³,I WKH JRDO LV WR UHGXFH ILOH VKDULQJ . . . LW¶V D failure.´ . 55. Larry Dignan, Why RIAA, IS P Cooperation May Deliver Returns for Both S ides, ZDNET (Jan. 29, 2009, 3:33 AM), http://www.zdnet.com/blog/btl/why-riaa-isp-cooperation-may-deliver-returns-forboth-sides/11893; Greg Sandoval, Comcast, Cox Cooperating with RIAA in Antipiracy Ca mpaign, CNET NEWS (Mar. 25, 2009, 9:49 AM), http://news.cnet.com/8301-1023_3-10204047-93.html. 56. Greg Sandoval, 5,$$*LYHV7KXPEV8SWR)UDQFH¶V7KUHH-Strike Law, CNET NEWS, (Apr. 8, 2009, 1:58 PM), http://news.cnet.com/8301-1023_3-10215602-93.html. DO NOT DELETE 1264 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 public attitudes regarding file sharing and the entertainment industry. III. LITIGATION AND NORMATIVE OVERDETERRENCE A. LITIGATION AND NORMATIVE BACKLASH As the brief overview in Part II illustrates, copyright holders have waged a highly successful litigation campaign against file sharing. In a series of high-profile court decisions, the entertainment industry successfully lobbied for the enforcement of expansive interpretations of contributory enforcement, the creation of novel doctrines of copyright infringement, and broad applications of statutory damage provisions. In this process, the entertainment industry was able to close down several new technologies57 and compel many individuals to settle individual infringement cases. As the legal campaign continues to evolve, its negative impact on public attitudes toward content industries is unmistakable. While the enforcement of copyright law against music pirates and commercial bootleggers has traditionally received broad public support, the massive litigation against teenagers and college students has proven to be much more controversial²perhaps, in part, because the general public identifies somewhat with copying PXVLF LQ RQH¶V KRPH especially by individuals living under the same roof. The mass litigation campaign against noncommercial infringers is an exceptional extension of the breadth of copyright enforcement. Historically, copyright enforcement suits have targeted FRPPHUFLDO DJHQWV RU ³SLUDWHV´ ZKR VHOO ERRWOHJV RI OLYH recordings and unlicensed reproductions of records, software, and video games, rather than individuals infringing for personal use.58 57. See supra Part II. 58. LITMAN, supra note 1, at 111 ³2XUFRS\ULJKWODZVKDYHXQWLOQRZIRFXVHGSULPDULO\RQWKH relationships among those who write works of authorship and disseminate those works to thHSXEOLF´  Jane C. Ginsburg, Putting Cars on the ³Information Superhighway´: Authors, Exploiters, and Copyright in Cyberspace , 95 COLUM. L. REV. 1466, 1488 (1995) ³Copyright owners have traditionally avoided targeting end users of copyrighted works´ ; Wu, supra note 43, DW ³2QHLVSUHVVHGWRILQG any example of copyright law being enforced against individuals for home copying (as opposed to comPHUFLDODFWLYLW\ SULRUWR´ See, e.g., Fonovisa, Inc. v. Cherry Auction, Inc., 76 F.3d 259, 264 (9th Cir. 1996) (holding that contributory liability could be imposed on defendant who operated a swap meet where many of the vendors sold counterfeit goods); Arista Records, Inc. v. MP3Board, Inc., No. 00 Civ. 4660(SHS), 2002 U.S. Dist. LEXIS 16165, at *18±19 (S.D.N.Y. Aug. 29, 2002) (record companies brought action against an Internet site operator that provided links to pirated copies of copyrighted musical recordings); 1LQWHQGR RI $P ,QF Y &RPSXWHU  (QWP¶W ,QF No. C96-0187 WD, 1996 U.S. Dist. WL 511619 (W.D. Wash. May 31, 1996) (issuing a preliminary injunction against DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1265 Additionally, the litigation campaign involved a number of awkward public relations incidents that likely contributed to the dwindling public support. For instance, media outlets reported that content industries accused both a twelve-year-old New York girl whose mother lived in low income housing run by the New York City Housing Authority59 and an eightythree-year-old woman who had died over a month earlier.60 To many observHUVWKHLQGXVWU\¶VOLWLJDWH-or-settle practice reflects an unfair power dynamic. Since alleged infringers so frequently choose to settle in order to avoid the expense of litigation, many legal questions remain unanswered. According to some, the almost mechanical administration of the settlement payments is akin to blackmail, extortion, or harassment.61 Moreover, the legitimacy of the litigation effort is undermined by a lack of comprehensiveness. Due to the large number of offenders, any instance of individual enforcement seems random. On the one hand, an indiscriminate ex ante approach to large-scale infringements has some normative appeal. On the other hand, it creates a perception that a few individuals are arbitrarily beLQJVLQJOHGRXWLQD³OLWLJDWLRQORWWHU\´ Another source of discontent is the six-figure statutory damage awards, as applied in Capitol Records v. Thomas-Rasset62 and Sony BMG defendants who sold video game duplication devices for Nintendo game cartridges); Sega Enters. Ltd. v. MAPHIA, 857 F. Supp. 679 (N.D. Cal. 1994) (issuing a preliminary injunction against a defendant who made available unauthorized copies of Sega games on a fee-based Internet bulletin board). See generally Geraldine Szott Moohr, The Cri me of Copyright Infringement: An Inquiry Based on Morality, Harm, and Cri minal Theory, 83 B.U. L. Rev. 731 (2003) (explaining the distinctions between commercial piracy and noncommercial personal infringements). 59. John Borland, RIAA Settles with 12-Year-Old Girl , CNET NEWS (Sept. 9, 2003, 4:05 PM), http://news.com.com/2102-1027_3-5073717.html?tag=st.util.print. 60. Andrew Orlowski, RIAA Sues the Dead, REGISTER (U.K.) (Feb. 5, 2005, 2:30 AM), http://www.theregister.co.uk/2005/02/05/riaa_sues_the_dead/. 61. See Nate Anderson, 7KH ³/HJDO %ODFNPDLO´ %XVLQHVV ,QVLGH D 33-Settlement Factory, WIRED (Oct. 3, 2010, 10:30 AM), http://www.wired.com/epicenter/2010/10/the-legal-blackmailbusiness/; Eric Bangeman, 5,$$³([RUWLRQ´:K\WKH2QO\5,&27KH\)HDU,V6XDYH, ARS TECHNICA (May 7, 2007, 12:41 AM), http://arstechnica.com/tech-policy/news/2007/05/riaa-extortion-why-theonly-rico-they-fear-is-suave.ars; Corynne McSherry, Sensible Data Retention Policies Help Prevent ELECTRONIC FRONTIER FOUND. (Mar. 28, 2007), RIAA Harassment , https://www.eff.org/deeplinks/2007/03/sensible-data-retention-policies-help-prevent-riaa-harassment; The Anti-Anti-Piracy Ca mpaign: CEA Tells RIAA To Stop Harassing People , QUICKJUMP GAMING NETWORK (Apr. 27, 2006, 7:51 PM), http://www.qj.net/qjnet/apple/the-anti-anti-piracy-campaign-ceatells-riaa-to-stop-harassing-people.html. 62. Capitol Records Inc. v. Thomas, 579 F. Supp. 2d 1210, 1213 (D. Minn. 2008). The court HYHQWXDOO\UHGXFHGWKHDZDUGRQUHPDQGVWDWLQJWKDW³[t]he need for deterrence cannot justify a $2 million verdict for stealing and illegally distributing 24 songs for the sole purpose of obtaining free music. Moreover, although Plaintiffs were not required to prove their actual damages, statutory damages must still bear some UHODWLRQWRDFWXDOGDPDJHV´ Capitol Records Inc. v. Thomas-Rasset, 680 F. Supp. 2d. 1045, 1048±49 (D. Minn. 2010). DO NOT DELETE 1266 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 Music Entertainment v. Tenenbaum.63 It has not been lost on the public that these awards extend well beyond the means of the single mother and graduate student that those cases involved. More generally, there is a public sentiment that the awards are disproportionate and excessive.64 Some argue that the punitive nature of the awards is inappropriate because the statutory damage framework, as intended by Congress, merely seeks to substitute for actual damages.65 Arguably, because substantive due process protections prohibit grossly excessive awards, statutory damage awards in copyright actions should be subject to additional scrutiny.66 Others suggest that, given the latest developments, the statutory damage law should be reformed to incorporate a distinction between commercial and noncommercial infringement.67 Generally, such criticism reflects a concern with a lack of balance between the entertainPHQWLQGXVWU\¶VFKRVHQPHDQVDQGits ends.68 The unpopularity of the litigation campaign is illustrated by online initiatives that seek to subsidize the settlement payment of unlucky defendants.69 The normative appeal of file sharing further amplifies the adverse reaction to the strict enforcement of copyright law.70 Teenagers are 63. Sony BMG Music EntP¶W v. Tenenbaum, No. 07cv11446-NG, 2009 U.S. Dist. LEXIS 115734 (D. Mass. Dec. 7, 2009). See also Denise Lavoie, Jury Awards $675K in Boston Music Downloading Case, ASSOCIATED PRESS, July 31, 2009, available at http://www.msnbc. msn.com/id/32236444/ns/technology_and_science-security/t/jury-awards-k-music-downloading-case/. 64. See Pamela Samuelson & Ben Sheffner, Debate, Unconstitutionally Excessive Statutory Da mage Awards in Copyright Cases, 158 U. PA. L. REV. PENNUMBRA 53 (2009), http://www.pennumbra.com/debates/pdfs/CopyrightDamages.pdf (a debate discussing whether large statutory damage awards are constitutional); J. Cam Barker, Note, Grossly Excessive Penalties in the Battle Against Illegal File-Sharing: The Troubling Effects of Aggregating Mini mum Statutory Da mages for Copyright Infringement , 83 TEX. L. REV. 525, 526 (2004). 65. Barker, supra note 64, at 556±59. See also Colin Morrissey, Note, Behind the Music: Determining the Relevant Constitutional Standard for Statutory Da mages in Copyright Infringement Lawsuits, 78 FORDHAM L. REV. 3059 (2010) (arguing in favor of due process evaluation of statutory damages). 66. Barker, supra note 64, at 536. 67. Anna Cronk, Note, 7KH3XQLVKPHQW'RHVQ¶W)LWWKH&ULPH²Why and How Congress Should Revise the Statutory Copyright Da mages Provision for Noncommercial Infringements on Peer-to-Peer F ile-Sharing Networks, 39 SW. L. REV. 181, 195±98 (2009). 68. See Daniel Reynolds, Note, The RIAA Litigation War on F ile Sharing and Alternatives More Compatible with Public Morality, 9 MINN. J.L. SCI. & TECH. 977, 978±87 (2008). 69. See, e.g., Show Your Support, JOEL FIGHTS BACK, http://joelfightsback.com/getinvolved/donate/ (last visited Sept. 6, 2011) (taking donations for Joel Tenenbaum, the defendant in Sony BMG Music Entertainment v. Tenenbaum). 70. See, e.g., Gervais, supra note 19; Strahilevitz, supra note 19 (arguing that computer code may solve collective action probOHPV  7KH GHILQLWLRQ RI ³VRFLDO QRUP´ LV VRPHZKDW LOOXVLYH )RU WKH SXUSRVHDWKDQGZHVLGHZLWKWKHQRWLRQWKDWDVRFLDOQRUPLVD³Vocial regulariW\´²a behavior that is in fact widely adopted in society not only because it is what people do, but also because it corresponds DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1267 convinced that file-sharing technology has many beneficial uses and that copyright law is outdated or biased toward music publishers. In this view, ILOH VKDULQJ LV WR ³EH HPEUDFHG UDWKHU WKDQ IHDUHG´71 EHFDXVH ³WHFKQRORJ\ PDNHV LW SRVVLEOH WR PDNH DQ XQOLPLWHG QXPEHU RI SHUIHFW FRSLHV´72 and GLVWULEXWH WKRVH FRSLHV WR PLOOLRQV RI XVHUV ³DW QR FRVW WR WKH FRQWHQW providHU´73 As early as 2003, surveys have indicated that a substantial number of teenagers believe that sharing digital music is morally acceptable.74 A Gallup poll indicated that 15 percent of teenagers under the age of seventeen EHOLHYHGWKDWLWZDV³morally ZURQJ´WRGRZQORDG ³PXVLF IURPWKH,QWHUQHWIRUIUHH´75 Similarly, survey evidence indicates that less than half of teenagers believe that punishment is appropriate for illegal downloading.76 B. THE INADVERTENT COSTS OF DETERRENCE Clearly, the more salient and publicized features of the litigation campaign have undermined public support for the enforcement of copyright law. But how much should negative public attitudes be of concern to content right holders? Can content industries afford to ignore public backlash? Perhaps content industries are wise to rely on deterrence. From a narrow rational choice perspective, ramping up deterrence certainly holds promise. If the behavior of file sharers is determined simply by the with a normative conception within society, or a subgroup thereof, of what people should do. See Richard H. McAdams, The Origin, Development, and Regulation of Norms, 96 MICH. L. REV. 338 (1997). Where necessary, we will distinguish between personal norms (moral standards attributed to an individual) and social norms (standards attributed to a group or collective). See Michael Wenzel, An Analysis of Norm Processes in Tax Compliance , 25 J. ECON. PSYCHOL. 213, 213±15 (2004). As we will see, however, the distinction between both categories loses some relevance because (1) generally, social norms are often internalized, and (2) in our results, perhaps through a self-serving bias, the social norm evaluation of respondents correlates with personal norms. Differences between personal and social norms, however, are highly relevant to policy recommendations. There is existing literature that provides insightful theoretical and empirical insights into the origins of social norms. 71. Raymond Shih Ray Ku, The Creative Destruction of Copyright: Napster and the New Economics of Digital Technology, 69 U. CHI. L. REV. 263, 268 (2003). 72. Id. at 264. 73. Id. at 268. 74. See, e.g., Steven Hanway & Linda Lyons, Teens O K With Letting Music Downloads Play , GALLUP POLL, Sept. 30, 2003, available at http://www.gallup.com/poll/9373/teens-letting-musicdownloads-play.aspx (reporting that 83 percent of teenagers in the age group of thirteen to seventeen find nothing wrong with file sharing). 75. Id. 76. Memorandum from KRC Research to Interested Parties, Topline Results of Microsoft Survey of Teen Attitudes on Illegal Downloading 2 (Jan. 23, 2008), available at http://www.microsoft.com/ presspass/download/press/2008/ 02-13KRCStudy.pdf. By contrast, 90 percent indicated punishment was appropriate for stealing a bike. Id. at 3. DO NOT DELETE 1268 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 expected costs and benefits of their actions, the statutory remedy precedents in Capitol Records v. Thomas-Rasset and Sony BMG Music Entertainment v. Tenenbaum might be salient enough to discourage some amount of file sharing.77 Similarly, laws that disconnect Internet access could boost deterrence significantly by increasing the expected costs of illegal file sharing.78 Increasingly, however, social science scholarship employs accounts of human behavior that advise against purely coercive approaches to enforcement. In the normative conception of compliance, for instance, individuals observe legal regulations because they believe it is the right thing to do. Social psychologists provide evidence that legal obedience is ³PRUDOity EDVHG´RU³OHJLWLPDF\ EDVHG´79 Human interaction with law thus GHSHQGV ODUJHO\ RQ ZKHWKHU SHRSOH EHOLHYH WKDW WKH ODZ LV ³MXVW´ RU produced by a legitimate regulator.80 Similarly, writings in the fields of sociology, psychology, and economics propose that the effectiveness of law enforcement is influenced by preexisting beliefs and norms as well as by the private costs and benefits of the behavior itself.81 This research suggests that legitimacy and norms influence behavior more than formal sanctions do.82 77. See generally George J. Stigler, The Opti mum Enforcement of Laws, 78 J. POL. ECON. 526 (1970) (constructing a theory of rational enforcement). 78. See generally RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW (5th ed. 1998) (explaining fundamental concepts of the economic approach to law). 79. For instance, in the context of tax compliance, there exists extensive literature on the DVVXPSWLRQWKDW³VRFLDOPRWLYDWLRQVUDWKHUWKDQPHUHVHOILVKQHVV . . . affect taxpaying behaviour, such as HWKLFDOFRQFHUQVDQGVRFLDOQRUPVSHUFHSWLRQVRIIDLUQHVVDQGOHJLWLPDF\´0LFKDHO:HQ]HO, Motivation or Rationalisation? Causal Relations Between Ethics, Norms and Tax Compliance , 26 J. ECON. PSYCHOL. 491, 492 (2005) (citing TOM R. TYLER, WHY PEOPLE OBEY THE LAW (1990); Simon James et al., Developing a Tax Compliance Strategy for Revenue Services , 55 BULL. FOR INT¶L FISCAL DOCUMENTATION 158 (2001)). See also John S. Carroll, Compliance with the Law: A Decision-Making Approach to Taxpaying, 11 LAW & HUM. BEHAV. 319 (1987) (applying decisionmaking models to tax law); James et al., supra at 163. 80. TYLER, supra note 79, at 3±4. 81. Id. at 45±56. See also Tracey L. Meares, S ignaling, Legiti macy, and Compliance: A &RPPHQWRQ3RVQHU¶V/DZDQG6RFLDO1RUPVDQG&ULPLQDO/DZ3ROLF\, 36 U. RICH. L. REV. 407, 410 & nn.27±28 (2002). Similarly, Paul Robinson has argued that some features of criminal law can be explained by the objective to align the law ZLWK³OD\SHrVRQV¶LQWXLWLRQVRIMXVWLFH.´3DXOH. Robinson, Why Does the Cri minal Law Care What the Layperson Thinks is Just? Coercive Versus Normative Cri me Control , 86 VA. L. REV. 1839, 1841 (2000). Robinson also argues that ³WKHH[WHQWRIFULPLQDO ODZ¶Vmoral authority determines the extent of its ability to shape community norms and to influence SHRSOH¶VFRQGXFWWKURXJKQRUPDWLYHforces´ Id. at 1840. 82. A number of empirical studies find that norms and beliefs are a stronger determinant of compliance than deterrence. See, e.g., James Alm, Gary H. McClelland & William D. Schulze, Changing the Social Norm of Tax Compliance by Voting, 52 KYKLOS 141, 141, 163 (1999); Ana de Juan, Miguel A. Lasheras & Rafaela Mayo, Voluntary Tax Complaint Behavior of Spanish Income Tax DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1269 The flipside of this coin is that people might decide not to obey a legal UXOH LI LW LV FRQVLGHUHG ³XQMXVW´ RU LI D ODZPDNHU LV SHUFHLYHG WR EH ³LOOHJiWLPDWH´ $V D UHVXOW HYHQ ZKen it may be irrational from a costbenefit perspective, some unlawful behavior will occur simply because the underO\LQJODZFRQIOLFWVZLWKDQLQGLYLGXDO¶VSUHH[LVWLQJQRWLRQVRIMXVWLFH and legitimacy. Normative intuitions about morality might cause individuals to neglect some of the costs associated with the illegal behavior. But the impact of conflicts between deterrence and morality- or legitimacy-based intuitions might extend even further. When behavior is driven by normative viewpoints, imposing laws that are perceived as ³unjust´ or ³illegitimate´ might reinforce and strengthen the underlying opposition against those laws.83 There is scholarship that highlights the problematic nature of vigorous legal condemnations of norms in some settings, noting that the interaction of law and norms often strengthens a preexisting antisocial norm.84 In the context of criminal law, William 6WXQW]DSSOLHVWKHWHUP³VHOI-GHIHDWLQJFULPHV´WRGHVFULEHVLWXDWLRQVZKHUH ³FULPLQDOL]DWLRQ can work against the very norms on which it rests,´85 FDXVLQJ ³SRSXODU QRUPV . . . to move in the opposite direction from the law.´86 Enforcement may reinforce or strengthen a belief that the legal regime is not legitimate or that a legal rule is unjust, especially if the public perceives the legal sanction to be excessive in relation to the punished behavior. In the context of tax compliance, for example, several studies document that enforcement measures can backfire and cause an overall increase of tax evasion.87 Other notable examples include the failed Payers, 49 PUB. FIN. 90, 90 (1994); Michael Wenzel, The Social S ide of Sanctions: Personal and Social Norms as Moderators of Deterrence, 28 LAW & HUM. BEHAV. 547, 549±50 (2004). For an overview on normative beliefs and tax compliance, see Leandra Lederman, The Interplay Between Norms and Enforcement in Tax Compliance, 64 OHIO ST. L.J. 1453 (2003). 83. See Dan M. Kahan, Social Meaning and the Economic Analysis of Cri me , 27 J. LEGAL STUD. 609 (1998); Timur Kuran & Cass R. Sunstein, Availability Cascades and Risk Regulation, 51 STAN. L. REV. 683, 715±36 (1999); Tracey L. Meares & Dan M. Kahan, Law and (Norms of) Order in the Inner City, 32 LAW & SOC¶Y REV. 805, 816±30 (1998). 84. For a theoretical model, see generally Parisi & Von Wangenheim, supra note 20 (describing a cycle of opinion formation whereby public acts of disobedience and protest undermine the legitimacy of legislation, which leads to further opposition). 85. Stuntz, supra note 20, at 1872. 86. Id. (emphasis added). 87. See John S. Carroll, A Psychological Approach to Deterrence: The Evaluation of Cri me Opportunities, 36 J. PERSONALITY & SOC. PSYCHOL. 1512, 1520 (1978); Harold G. Grasmick & Donald E. Green, Legal Punishment, Social Disapproval and Internalization as Inhibitors of Illegal Behavior , 17 J. CRIM. L. & CRIMINOLOGY 325, 330±35 (1980); Kent W. Smith, Integrating Three Perspectives on Noncompliance: A Sequential Decision Model , 17 CRIM. JUST. & BEHAV. 350, 366±67 (1990). See DO NOT DELETE 1270 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 historical enforcement against alcohol during the Prohibition era,88 and, in the past few decades, the war on drugs as it relates to soft drugs.89 Similarly, countervailing normative effects might explain file shareUV¶ continued persistence despite increased enforcement efforts by content holders. If, by observing the impact of enforcement, individuals conclude that the underlying rules are illegitimate, file sharers may be more likely to disobey the law. The potential undermining effect of normatively excessive deterrence presents a vexing dilemma for content holders. Litigation informs copiers of the danger involved with illegal sharing, but at the same time, the adverse normative impact of enforcement may actually induce additional infringements. The normative costs of elevated penalties can undermine the overall deterrent effect in the following ways. First, normatively excessive enforcement measures might cause potential infringers to disregard the law on nonconsequential grounds. If the deterrent or coercive aspects of enforcement create a belief that the underlying legal rules are unjust, individuals might decide to disobey the law²despite the added costs associated with breaking the law. Second, violating an unjust or immoral law might sometimes increase utility to an individual, perhaps sufficiently so that it outweighs the costs associated with the illegal behavior. Finally, in the long run, normatively excessive deterrence may undermine the political support for the underlying protected rights. Changing social generally Lederman, supra note 82 (providing an overview of various studies on norms in the area of tax compliance). 88. See generally KENNETH ALLSOP, THE BOOTLEGGERS: THE STORY OF CHICAGO¶S PROHIBITION ERA (1968) (describing causes of success and failure in the enforcement campaign on bootlegging); ANDREW SINCLAIR, PROHIBITION: THE ERA OF EXCESS (1962) (tracking the social history of the prohibition movement); Joseph R. Gusfield, Prohibition: The Impact of Political Utopianism, in CHANGE AND CONTINUITY IN TWENTIETH-CENTURY AMERICA: THE 1920¶S 257 (John Braeman, Robert H. Bremner & David Brody eds., 1968) (following the impact of the Prohibition movement in the United States). See also LARRY ENGELMANN, INTEMPERANCE: THE LOST WAR AGAINST LIQUOR 177± 87 (1979); Harry G. Levine & Craig Reinarman, F rom Prohibition to Regulation: Lessons from Alcohol Policy for Drug Policy, 69 MILBANK Q. 461, 466±69 (1991). 89. ³>$@fter the expenditure of billions of dollars on a policy built primarily on the coercion and punishment of drug distributors and users, the War on Drugs has failed to reduce significantly, much less eliminate, drugs as a problem in our society.´ David Rudovsky, The Impact of the War on Drugs on Procedural Fairness and Racial Equality, 1994 U. CHI. LEGAL F. 237, 237 (1994) (discussing the negative impact of drug war policies on constitutional norms and racial equality). For a historical review of the war on drugs, see generally JAMES P. GRAY, WHY OUR DRUG LAWS HAVE FAILED AND WHAT WE CAN DO ABOUT IT: A JUDICIAL INDICTMENT OF THE WAR ON DRUGS (2001) (discussing the 8QLWHG 6WDWHV¶ missteps in addressing the countU\¶V drug problem); STEVEN WISOTSKY, BEYOND THE WAR ON DRUGS: OVERCOMING A FAILED PUBLIC POLICY (1990) (following the United States¶s failure to effectively regulate the cocaine drug trade). DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1271 attitudes might lead political actors to revoke the available rights or remedies. This raises the potential that the normative costs of deterrence may cancel out the benefits of enforcement. At an optimal level, a deterrence strategy balances the marginal deterrent effect of enforcement against marginal normative costs of deterrence.90 Consider for instance, two modalities of enforcement. A first approach imposes higher penalties, while a second approach applies more moderate penalties. Because of the higher penalties, the first modality generates higher levels of deterrence (+5) than the second, more moderate approach (+3). When taking into account the normative effects described above, however, the more stringent enforcement condition may nonetheless be less productive overall. If, for instance, modality one creates a strong adverse normative reaction (-6), while the more moderate sanctions of modality two produce a less adverse normative reaction (-1), the overall value of enforcing modality one is negative (-1) while it is positive for modality two (+2). This proposition stands in stark contrast to the statement, voiced by copyright industry representatives, that strong-armed litigation tactics are sensible because they educate the public on the content of copyright law and, in doing so, may cultivate pro-copyright viewpoints. In this regard, the HQWHUWDLQPHQWLQGXVWU\¶VOLWLJDWLRQFDPSDLJQFDQEHconsidered an effort to promote copyright compliance by inducing a belief that file sharing is unfair to authors.91 Perhaps by focusing attention on the legal rights of copyright holders, the industry hoped to shift preferences in favor of copyright holders. Indeed, the literature on the expressive function of the law states that legal rules, especially if formulated by lawmakers who are perceived to be legitimate, may cause individuals to align preferences with the goals expressed in legal rules.92 90. Additionally, a deterrence policy should consider the non-normative costs of overdeterrence. The conventional concept of overdeterrence in the economic analysis of law describes how the fear of incurring liability may cause individuals to forego socially valuable activities. See A. Mitchell Polinsky & Steven Shavell, Economic Analysis of Law, in THE NEW PALGRAVE DICTIONARY OF ECONOMICS 7 (Steven N. Durlauf & Lawrence E. Blume eds., 2008). Severe malpractice penalties, for instance, might deter malicious or negligent conduct, but they might also cause medical professionals to develop otherwise costly defensive medical practices in order to avoid any possible accusations of wrongdoing. 91. A number of empirical studies report that norms and beliefs can be a stronger determinant of compliance than traditional means of deterrence. See supra note 82. 92. When law creates a focal point by expressing values that might tip norms to a new equilibrium, this process may create a social norm or internalize a normative value. Robert Cooter, Expressive Law and Economics, 27 J. LEGAL STUD. 585, 585 (1998). See also GARY S. BECKER, ACCOUNTING FOR TASTES 1±2 (1996). The idea of law as a focal point that coordinates expectations DO NOT DELETE 1272 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 These competing claims reveal the potential ambiguity with regard to the normative effects of enforcement. In order to examine the normative effects of enforcement, we conducted two studies to test the hypothesis that copyright enforcement may inadvertently move anti-copyright behavior and norms in the opposite direction from that intended by the law.93 Although, as mentioned above, a few studies have explored the interaction between deterrence and norms with regard to tax evasion, the countervailing norm-effect hypothesis awaits further examination.94 Part IV collects and discusses new evidence on the effects of enforcement on social attitudes and public norms in the context of copyright law. IV. STUDY A. INTRODUCTION & MAIN RESULTS In an attempt to advance our understanding of the interaction between norms and deterrence, we explore the dynamics of enforcement on selfreported norm evaluations in the context of two experimental studies. 7KH UHVXOWV VXJJHVW WKDW FRS\ULJKW QRUPV DQG LQGLYLGXDOV¶ HYDOXDWLRQ of the music industry¶V enforcement policies are negatively affected by the enforcement of copyright law. We observed that increases in the severity and certainty of punishment generate more powerful anti-copyright norms. Also, enhanced sanctions appear to bolster adverse positions toward the music industry¶VSROLFLHV.95 We further explored this counterproductive effect by looking at potential differences between various types of infringers. We found that the among citizens is explored further in Richard H. McAdams, A Focal Point Theory of Expressive Law, 86 VA. L. REV. 1649 (2000). 93. $ OHJDO FRQGHPQDWLRQ RIDQRUP PLJKWKDYH WKH XQLQWHQGHG HIIHFW RI PRYLQJ ³HTXLOLEULXP behavior . . . LQWKHRSSRVLWHGLUHFWLRQIURPWKDWLQWHQGHGE\WKHODZ´3DULVL YRQ:DQJHQKHLP supra note 20, at 2. 94. Some scholars have argued that improved enforcement may backfire and produce increased tax evasion, but these viewpoints are contested. For a review of empirical evidence on the proposition that enforcement may backfire, see Lederman, supra note 82. Lederman concludes that In sum, the speculation that sanctions for tax evasion will tend to undermine compliance does not seem to be supported by the evidence. In the experimental context, the availability of sanctions for failure to cooperate increases cooperation. In the tax compliance context, audits increase even compliance of those not threatened with audit. Id. at 1499 (footnotes omitted). 95. Interestingly, as we will describe in more detail, an equivalent increase in the probability of enforcement did not affect anti-industry positions, perhaps because the effectiveness of enforcement is partially determined by technological progress, which is unrelated to the influence of the music industry on the political process. DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1273 degree of backlash positively correlated with the amount of downloading students engaged in before participating in the studies. Increased levels of enforcement produce higher levels of anti-copyright resentment and backlash among students who regularly downloaded music from filesharing networks. Our results suggest that there might be a tipping point involving the amount of downloads and norms: at a certain level of experience with downloading copyright-infringing content, the normative evaluation of enforcement policies becomes more negative.96 Moreover, our results indicate that the backlash effect also extends to illegal downloading. When asked whether they would resume downloading once novel technological developments provided immunity, file sharers who were previously subjected to a more stringent enforcement regime indicated that they planned to download more.97 Further, students faced with more severe sanctions in the first phase reported that they would download more, as compared to participants who were presented with more moderate sanctions.98 B. METHODOLOGY AND CONSTRUCTION OF VARIABLES Our study applied a scenario methodology to examine the sensitivity of respondents to different degrees of copyright enforcement. We paid particular attention to differences between frequent users of file-sharing technology and those individuals who were less experienced with this technology. Participants were invited to participate in the experiments on the basis of their enrollment in an undergraduate, introductory political science class. Students were unaware that the experiment would be conducted in class on the day the study occurred. None of the students had taken classes in intellectual property law or copyright law. To avoid underreporting99 and esteem-based distortions,100 the survey was conducted anonymously in a 96. On the internalization of social norms, see Robert Cooter, Do Good Laws Make Good Citizens? An Economic Analysis of Internalized Norms, 86 VA. L. REV. 1577, 1597±1600 (2000) [hereinafter Cooter, Good Citizens]; Robert D. Cooter, Three Effects of Social Norms on Law: Expression, Deterrence, and Internalization, 79 OR. L. REV. 1, 17±20 (2000) [hereinafter Cooter, Three Effects]. 97. See infra Part IV.C. 98. Id. 99. Liebowitz, Creative Destruction, supra note 42, at 8. 100. Colin F. Camerer, & Ernst Fehr, Measuring Social Norms and Preferences Using Experi mental Ga mes: A Guide for Social Scientists, in FOUNDATIONS OF HUMAN SOCIALITY 55 (Joseph Henrich et al. eds. 2004). DO NOT DELETE 1274 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 large auditorium with a seating capacity of 630. Participants were informed that the survey was part of a study on ³how people respond to copyright enforcement.´ No financial compensation was offered for participating in the study. Students who had never used file-sharing technology completed items that were not relevant for the purposes of the present study and were excluded from further analysis. A total of 404 freshman students participated in two experiments. The participants were predominantly female (252) and ranged in age between eighteen and twenty-four. Only students who indicated that they had illegally downloaded one or more files participated in the scenario study. Students that completed Study 1 did not participate in Study 2. Both studies employed the same materials and procedures. In both experiments we manipulated the severity and certainty of sanctions. In the first study, one group of participants received a scenario with a low severity of sanction ($20) in combination with either a low (0.01%) or high (20%) probability of sanction, while another group received a vignette that had a sanction with a high severity ($20,000) in combination with either a low (0.01%) or high (20%) probability of sanction.101 101. These figures are based on the potential remedies applicable on the basis statutory award provisions in the Copyright Act. We also conducted DSLORWVWXG\WKDWPHDVXUHGVXEMHFWV¶SHUFHSWLRQRI what amounts to low or high probability and low or high severity of punishment in the context of copyright infringement through file sharing. In the pilot study, sixty-three participants were requested to indicate on a 7-point Likert scale whether they agreed with the following statements: (1) if you are caught for illegally downloading music, you will have to pay $2000 for each downloaded song or file²this punishment is severe; (2) if you are caught for illegally downloading music, you will have to pay $20 for each downloaded song or file²this punishment is severe; (3) the chance of getting caught for illegally downloading music is 0.001% (1 in 100,000 file sharers)²the probability of getting punished is high; and (4) the chance of getting caught for illegally downloading music is 20% (1 in 5),²the probability of getting punished is high. A Likert-type scale, named after psychologist Rensis Likert, is a psychometric scale that is widely used in survey research and questionnaires. See Rensis Likert, A Technique for the Measurement of Attitudes, 22 ARCHIVES OF PSYCH. 140 (1932) (explaining the function of the Likert scale). A paired t-test yielded a significant effect (t(62) = 9.25, p < 0.001) for severity of punishment, indicating that participants viewed low severity of punishment (MHDQ ³0´ = -0.22, 6WDQGDUG 'HYLDWLRQ ³SD´ = 1.88) as less severe than high severity of punishment (M = -2.03, SD = 1.57). A similar analysis was conducted for probability of punishment, revealing a significant effect (t(62) = -6.76, p < 0.001). Participants indicated that a low probability of punishment (M = -0.84, SD = 1.89) is less probable than a high probability of punishment (M = 1.71, SD = 1.86). DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1275 1. Study 1 Study 1 consisted of two phases. In the first phase of the scenario, we examined norm evaluations under different modalities of enforcement. In the second phase, participants were asked to indicate how their individual downloading behavior would change if the probability of apprehension was reduced to zero. We presented each group of respondents with different enforcement conditions. We applied the following modalities of enforcement (independent variables): individual conditions varied between (1) low (1/10,000 or 0.01%) or high (1/5 or 20%) rates of probability of punishment, and (2) low ($20) or high ($20,000) severity of punishment for each downloaded song. Each participant was assigned randomly to either the low or high probability of punishment condition and to either the low or high severity of punishment condition. In order to measure the normative effects and behavioral responses to the various enforcement conditions, we presented a list of statements to the participant in the study. The following statements (dependent variables) were used throughout Study 1: (a) Anti-Copyright Viewpoints in Phase 1. Four items measured anticopyright norms: (1) ³these new developments are gradually making me realize that downloading music illegally is not ethical,´  ³because of these changes I will adjust my position regarding illegal downloading,´ (3) ³these developments are causing me to adjust my norms regarding illegal exchanges of music,´ and  ³because of these new developments I ZLOOSXUFKDVHPRUHPXVLFWKURXJKOHJDOEXVLQHVVDYHQXHV´ (b) Evaluation of the Enforcement Policies by the Music Industry in Phase 1. Three items assessed the participantV¶ evaluation of the music industry¶V enforcement policies: (1) ³, DP RI WKH RSLQLRQ WKDW WKH PXVLF industry is conducting an unjust, disproportionate policy,´  ³the policies of the music industry conflict with my sense of justice,´ and (3) ³the policies of the music industry are an attack on my freedom to listen to PXVLF´102 102. These items were combined to form one average score (M = 3.90; SD = 1.37; L.J. &URQEDFK¶V Į ³alpha´ = 0.87). &URQEDFK¶V DOSKD LV D FRHIILFLHQW RI UHOLDELOLW\ FRPPRQO\ XVHG DV D measure of the internal consistency or reliability of a psychometric test score for a sample of examinees. See Lee J. Cronbach, Coefficient Alpha and the Internal Structure of Tests, 16 PSYCHOMETRIKA 297 (1951). DO NOT DELETE 1276 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 (c) Behavioral Adjustment in Phase 2. Four items measured behavioral adjustment: (1) ³, ZLOO UHVXPH GRZQORDGLQJ,´  ³, ZLOO PDNH up for lost time and increase my downloading activities,´  ³,ZLOOUHVXPH downloading because I disagree with the enforcement policy of the record industry,´ and (4) ³,ZLOOFRQWLQXHWRGRZQORDGOHVVEHFDXVH,UHDOL]HWKDW the record industry is losing a lot of money due to iOOHJDOGRZQORDGV´103 After reading a randomly assigned enforcement scenario, respondents provided their responses on the basis of a seven-point scale, assigning scores between two endpoints of (1) ³strong disagreement´ and (7) ³strong agreement´ (an answer RI³´LQGLFDWHGWKDWWKHVWXGHQWVWURQJO\GLVDJUHHG with the statement; an answer of ³´ LQGLFDWHG WKDW WKH VWXGHQW VWURQJO\ agreed).104 2. Study 2 In Study 2 we applied conditions of high probability (and low severity) as well as high severity (and low probability) while holding the expected costs constant across both conditions. By controlling the expected costs in the different conditions, Study 2 enabled us to assess the relative impact of the severity and probability of the sanctions on download behavior in the first phase of the study. We therefore measured download behavior not only in the second phase (as in Study 1), but also in the first phase. Respondents were randomly assigned to one of two groups with the following enforcement conditions (independent variables): one group was subjected to an enforcement regime with a high certainty but low severity of punishment (20% probability of a fine of $50), while a second group faced a converse punishment schedule (0.1% probability of $10,000). At ten dollars each, the expected cost of downloading is identical in both scenarios. Participants in Study 2 completed a survey containing the items from the first study: anti-copyright viewpoints, the evaluation of the music industry¶V enforcement policies, and download behavior in phase 2. Additionally, we included two items to measure download behavior in phase 1: (1) ³, ZLOO GRZQORDG PRUH PXVLF´ DQG  ³, ZLOO GRZQORDG OHVV 103. Statistically, the scale showed sufficient internal consistency (M = 4.54; SD = 1.14; &URQEDFK¶VDOSKD 0.74). 104. Items 2, 3, and 4 were reverse coded and combined with item 1 to form one average anticopyright norm score. DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1277 PXVLF´105 3. Analysis Our analysis of the effect of enforcement modalities on the various groups of individual file sharers and non-file sharers, distinguished between different scenarios that varied the hypothetical probability of enforcement and the severity of sanctions. Again, we randomly assigned the different scenarios to respondents. In analyzing the responses, we looked for differences between respondents on the basis of their prior file-sharing experience. Throughout the study, we used the terms ³occasional,´ ³average´ and ³frequent´ file sharers to distinguish between respondents on the basis of the amount of music that they download on peer-to-peer networks. This distinction is important, as will be discussed, because effective copyright enforcement measures might require a different approach toward habitual file sharers versus occasional file sharers. The scenario methodology, as employed in this study, has certain strengths and weaknesses. Generally, scenario studies combine the benefits of laboratory research and correlation research.106 Our study design enabled us to analyze causality by manipulating the independent variables, while using realistic, everyday situations based on statutory rules and case law.107 As is the case with all laboratory experiments, a study of this nature is, of course, susceptible to the criticism that the results cannot be generalized to the public at large. For the present purpose, however, this criticism is less damaging. By exclusively enlisting college students, our design measures the reactions of the specific group that was targeted in the litigation campaign by the entertainment industry. Finally, a general limitation of any scenario study is that it does not measure actual behavior in the real world. In this context however, a research design involving actual downloading by respondents is not feasible due to the copyright-infringing nature of the activities involved. 105. Item 2 was reverse coded and combined with item 1 to form a separate score indicating the average download behavior among participants. 106. See, e.g., David De Cremer & Daan Van Knippenberg, How Do Leaders Promote Cooperation? The Effects of Charisma and Procedural Fairness, 87 J. APPLIED PSYCHOL. 858 (2002). 107. See Robert L. Dipboye, Laboratory vs. F ield Research in Industrial and Organizational Psychology, 5 INT¶L. REV. INDUS. & ORG¶L PSYCHOL. 1, 25 (1990) (discussing the legitimacy of laboratory research studies). DO NOT DELETE 1278 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 Finally, a few notes on terminology. We refrain from using the term ³DQWLVRFLDO´ FRQGXFW WR GHVFULEH ILOH-sharing practices. This would carry assumptions and arguments that are outside the ambit of this Article. ,QVWHDG ZH HPSOR\ WKH PRUH QHXWUDO WHUPV ³SUR-FRS\ULJKW´ DQG ³DQWLFRS\ULJKW´WRLQGLFDWHWKHDUHDVZKHUHUHVSHFWLYHO\FRQWHQWSURGXFHUVDQG file-sharing consumers are at opposite sides of the spectrum. We realize that even this choice of terminology is open for discussion considering the view, held by some commentators, that peer-to-peer file sharing is not in conflict with the goals of copyright law.108 C. FINDINGS Our results indicate that conditions involving high probabilities of punishment generate higher degrees of anti-copyright norms than lower probabilities of punishment.109 Severity of punishment also had a significant effect, indicating that stronger sanctions produce higher degrees of anti-copyright norms than moderate sanctions. Additional tests revealed that copyright norms are more negative where there is a high probability and high severity of punishment, compared to the other three conditions (that is, the low severity/low probability condition, the low severity/high probability condition, and the high severity/low probability condition). FIGURE 1. Normative Backlash on Anti-Copyright Norms 108. See, e.g., Ku, supra note 71 (suggesting that file sharing necessitates a new business model but benefits artists as a group). 109. We derived this result by conducting separate 2 (probability of punishment) x 2 (severity of punishment) ANCOVAs with total songs downloaded as a covariate for (a) anti-copyright norms, and (b) HYDOXDWLRQRIWKHPXVLFLQGXVWU\¶VHQIRUFHPHQWSROLFLHV$QDQDO\VLVRIFRYDULDQFH ³$1&29$´ LV a general linear model with a continuous outcome variable and two or more predictor variables. An ANCOVA tests whether certain factors have an effect on the outcome variable after removing the variance for which quantitative predictors (covariates) account. See, e.g., Gerard J.P. Van Breukelen & Koene R.A. Van Dijk, Letter to the Editor: Use of Covariates in Randomized Controlled Trials, 13 J. INT¶L NEUROPSYCHOLOGICAL SOC¶Y 903, 903±04 (2007). An overview of the results, including additional figures containing the means, standard deviations, and statistical effects, etc. are provided in table 1 in the appendix below. DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1279 Figure 1 above illustrates these observations. First, note that all students reported relatively high levels of anti-copyright norms (between 4.8±4.9 on a seven-point scale where ³7´ HTXDOV ³, FRPSOHWHO\ GLVDJUHH´ ZLWK VWDWHPHQWV VXFK DV ³GRZQORDGLQJ PXVLF LOOHJDOO\ LV QRW HWKLFDO´ 110 regardless of the particular enforcement regime. Second, while evaluations for other enforcement conditions are relatively uniform, the assessments for high-probability and high-severity conditions are much more negative (as indicated by the light grey line that extends to the top right corner of the graph above). The evaluation of the music industry¶V enforcement policies revealed that strong sanctions caused relatively strong contra-music industry positions, while moderate punishment generated less aversion to the music industry¶VSROLFLHV. In order to interpret the results, we divided the participants into different groups based on the frequency with which they engaged in downloading on file-sharing networks.111 We divided the participants into an ³occasional´ user group (one hundred students with less than seventy downloaded songs), an ³average´ user group (fifty-one students with one hundred downloaded songs), and a ³frequent´ user group (fifty-eight participants with more than 150 downloaded songs).112 We observed that ³frequent´ and ³average´ users had stronger contra-music industry positions than ³occasional´ users (see table 1 in appendix and figure 3). In the second phase of Study 1, we informed students that a new technology now fully shielded them from all enforcement. We asked participants if, and to what extent, they would resume downloading in light of this immunity.113 The results indicated that participants who previously faced severe sanctions in the first phase, engaged in more downloading 110. See infra items reported for (1) Anti-Copyright Viewpoints in Phase 1. 111. We used the covariate for this purpose. 112. More technically, such a division was helpful because the distribution of the total number of downloads reported in our sample departed from the normal distribution. The Kolmogorov-Smirnov Z test indicated a significant deviation from normality, z = 4.81, p < 0.001. Generally, a KolmogorovSmirnov statistic quantifies a distance between the empirical distribution function of the sample and the cumulative distribution function of the reference distribution, or between the empirical distribution functions of two samples. See, e.g., W.T. EADIE ET AL., STATISTICAL METHODS IN EXPERIMENTAL PHYSICS 269±71 (1971). A substantial number of participants (N = 51) indicated that they had downloaded one hundred files, and no other participants reported downloading a number of files in the range of seventy and 150. 113. A separate 2 (probability of punishment) x 2 (severity of punishment) ANCOVA was conducted with the total amount of downloads as a covariate, and individual downloading behavior as the dependent variable. For an overview of the results, see supra Table 1. DO NOT DELETE 1280 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 relative to participants that initially faced more moderate sanctions. Finally, ³frequent´ users downloaded more in the second phase than ³DYHUDJH´ and ³occasional´ users. Overall, these results clearly indicate of a backlash effect on downloading behavior. It is worthwhile to focus on the ³frequent´ users separately. Because an important goal of sanctions is to prevent lawbreakers from continuing their illegal activities, ³frequent´ users are an important target group. Our analysis indicates that ³frequent´ users, who previously faced a low probability of punishment in the first phase of the scenario, would download less than ³frequent´ users who had been previously subject to high probability enforcement conditions. A similar trend was observed for the severity of sanctions. ³Frequent´ users facing moderate punishments in the first phase of the scenario reported lower download activity levels than ³IUHTXHQW´XVHUV previously faced with severe sanctions. To summarize, two main results were obtained in Study 1. During Phase 1, a counterproductive effect on pro-copyright sentiments was revealed in relation to the probability and the severity of sanctions. Raising the level of both severity and certainty of enforcement produced a potentially powerful counterproductive effect, increasing anti-copyright norms. A key result also emerged in Phase 2, where the backlash effect of legal condemnations was observed to extend to underlying download behavior. As we discuss in more detail in the following part, the normative backlash effect has important implications for copyright enforcement policy. In Study 2, our analysiV RI WKH GLIIHUHQW HIIHFWV EHWZHHQ ³KLJKseverity/low-SUREDELOLW\´ DQG ³PRGHUDWH-severity/high-SUREDELOLW\´ enforcement conditions revealed a significant effect of the type of enforcement regime on anti-copyright norms.114 Participants in the highseverity/low-probability condition evaluated copyright norms more negatively than participants in the moderate-severity/low-probability condition. The covariate²total number of downloads²yielded a significant effect, suggesting that high-severity/low-probability conditions had a stronger deterrent effect (that is, when faced with this enforcement condition, students reported that they would reduce their downloading activities more). The elevated, light-shaded grey bars in figure 2 illustrate 114. We conducted three one-way ANCOVAs with total shared files downloaded as a covariate, enforcement regime as the independent variable, and anti-copyright norms, evaluation of the music LQGXVWU\¶VHQIRUFHPHQWSROLFLHVDQGGRZQORDGLQJEHKDYLRUDV the dependent variables. Table 2 presents an overview of the results. DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1281 how high-severity/low-probability enforcement conditions generate higher backlash effects as well as higher levels of deterrence than lowseverity/high-probability enforcement²even though, interestingly, both enforcement regimes impose identical expected costs. DO NOT DELETE 1282 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 FIGURE 2. Effect of Enforcement on Anti-Copyright Norms, Evaluation of Enforcement Policies, and Download Behavior In order to interpret these results, we divided the participants into different user groups115 based on their download background: we differentiated between ³occasional´ users, ³average´ users, and ³frequent´ users. The results indicate that ³frequent´ users hold stronger anticopyright norms than ³occasional´ and ³average´ users. The type of enforcement regime also impacted the evaluation of the music industU\¶V HQIRUFHPHQW SROLFLHV²participants in the highseverity/low-punishment condition endorsed a relatively strong contramusic-industry position, whereas low-severity/high-probability enforcement regimes fostered relatively lower levels of aversion for the 115. As in Study 1, the Kolmogorov-Smirnov Z test indicated a substantial deviation for the total number of downloads, z = 3.12, p < 0.001. DO NOT DELETE 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 2011] 1283 policies of the music industry. The covariate had a significant effect. ³Frequent´ file sharers evaluate the music industry more negatively than ³occasional´ users of file sharing technology. Especially participants in the ³frequent´ user group hold stronger contra-music-industry positions than ³average´ and ³occasional´ users. We also observed a significant effect of the type of enforcement regime on download behavior in phase 1. Participants in the highseverity/low-probability condition reported that they would download less than participants in the low-severity/high-probability condition. The covariate also yielded a significant effect, indicating that ³frequent´ users would reduce their downloading activities more so than ³average´ and ³occasional´ users. As in Study 1, we focused on ³frequent´ users separately because they represent the most interesting group for assessing the effects of copyright enforcement. As figure 3 illustrates, our analysis of download behavior in phase 2 of Study 2 demonstrates that ³frequent´ users are more likely than ³occasional´ and ³average´ users to exhibit backlash effects.116 Our analysis also revealed a significantly stronger backlash effect among ³frequent´ users of file-sharing technologies that were subjected to a highseverity/low-probability condition in the first phase than ³frequent´ users who previously faced low-severity/high-probability enforcement conditions. 116. See infra Table 2. DO NOT DELETE 1284 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 FIGURE 3. Effect of Enforcement on Anti-Copyright Norms, the Enforcement Policies by the Music Industry, Download Behavior, and Backlash Effect for Different User Groups in Study 2 V. DISCUSSION A. PUBLIC ATTITUDES AND SELF-INTEREST The litigation campaign of the content industries has generally focused on the gravest offenders ZKR PDNH DYDLODEOH ³µsubstantial¶ amounts of music online to others over peer-to-peer networks.´117 As a result, university students have been the primary targets.118 This enforcement strategy has several potential rationales. By deterring those who make available a substantial number of music files, the overall access to illegal content online might be reduced. Moreover, by going after the most severe infringers, the industry hopes to draw a normative distinction between small incursions by the average consumers and more extreme forms of 117. Katie Dean, Are You LQ 5,$$¶V &URVV +DLUV", WIRED (June 26, 2003), http://www.wired.com/entertainment/music/news/2003/06/59392. 118. Music Labels Hunt Net µPirates,¶ BBC NEWS (June 26, 2003), http://news.bbc.co.uk/2/hi/entertainment/3021126.stm GHVFULELQJWKH5,$$¶VWDUJHWLQg of peer-to-peer VRIWZDUHXVHUVLQFOXGLQJ³IRXUFROOHJHVWXGHQWV [who] agreed to pay damages after being sued by the 5,$$´ . DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1285 infringing behavior. Some have disputed the wisdom of this strategy, however, because it reduces the deterrent effect on the broader audience of more moderate file sharers.119 Our findings provide some insight into the likely impact of this aspect of the litigation campaign. One main finding of this study is the observed differences between frequent users and more occasional users of filesharing technology. Frequent file sharers exhibit stronger backlash effects: more stringent enforcement conditions result in adverse public attitudes and countervailing behavior. The most straightforward interpretation is that frequent infringers simply have more to lose from stringent enforcement. The public attitude of file sharers might simply reflect a self-serving bias. In this regard, prior research on the bidirectional causality between self-interest and normative evaluations is informative. In a pioneering study on tax aversion, Michael :HQ]HOFRQFOXGHVWKDWWD[SD\HUV³DGMXVWWKHLURZQEHOLHIVVRDVWRMXVWLI\ their behaviour as right and ethical.´120 According to the theory of cognitive dissonance, When we sense something in the world that is inconsistent with the cognitive frame through which we see the world, we initially (unconsciously) ignore or distort our perception. If that becomes impossible, we eventually amend our cognitive frame (i.e., the way we see and understand the world) to incorporate our new perception.121 For instance, the benefits derived from freely downloading music online may induce a normative belief that file sharing should be legal. As a type of self-serving bias, file sharers might adjust their normative positions UHJDUGLQJ GRZQORDGLQJ EHFDXVH WKH\ EHQHILW IURP WKH DFWLYLW\ ³, ZDQW WR download music, so I think prohibiting peer-to-peer actLYLWLHVLVXQMXVW´ 122 Indeed, research on cognitive dissonance suggests that individuals often adjust their attitudes and beliefs123 when they experience a conflict in their 119. Dean, supra note 117 ³2UWRIUHDNHYHU\RQH RXWWKH5,$$FRXOGFKRRVHWRVXHDVPDOOHU offendHU´  120. Wenzel, supra note 79, at 505. 121. Joshua D. Rosenberg, The Psychology of Taxes: Why They Drive us Crazy, and How We Can Make Them S ane, 16 VA. TAX REV. 155, 201 n.213 (1996) (describing how individuals reconcile the conflict between normative positions and legal regulation). 122. See Daniel S. Nagin & Greg Porgarsky, An Experi mental Investigation of Deterrence: Cheating, Self-Serving Bias, and Impulsivity, 41 CRIMINOLOGY 167 (2003) (presenting experimental evidence consistent with a self-serving bias). 123. Rosenberg, supra note 121, at 199±200 (describing how people amend their cognitive frame to reduce conflict by incorporating new perceptions of reality). DO NOT DELETE 1286 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 perceptions of reality.124 Accordingly, file sharers might adjust their beliefs about file-sharing people to remove the uncomfortable feeling of engaging in immoral conduct. Research in cognitive psychology suggests that individuals come to disregard the obvious self-interested origin of such normative adjustments.125 The next step, then, is WR³JHQHUDOLse these views to others, presumably to gain further social support.´126 Indeed, according to the false consensus effect, individuals often believe that others are more like themselves than WKH\UHDOO\DUH$VDUHVXOWSUHGLFWLRQVDERXWRWKHUV¶ beliefs or behaviors, based on casual observation, are very likely to err in the direction of our own personal beliefs or behavior.127 More generally, our findings suggest that there might be a tipping point involving the amount of downloads and normative reactions. At a certain level of experience with unauthorized downloading and file sharing, the evaluation of that behavior is internalized.128 As a result of this process of internalization, the normative position EHFRPHVPRUHUREXVWRU³VWLFN\´ and presents a more pervasive challenge to law enforcement. Another plausible explanation lies with social identity theory.129 As users become more accustomed to using file-sharing technology, they 124. See, e.g., JON ELSTER, SOUR GRAPES: STUDIES IN THE SUBVERSION OF RATIONALITY (1983); LEON FESTINGER, A THEORY OF COGNITIVE DISSONANCE 1±4 (1957). The classic example of this is expressed in the fable where a fox sees some high-hanging grapes and wishes to eat them. When the fox is unable to reach the grapes, he surmises that the grapes are probably not worth eating, as they must not be ripe or that they are probably sour. AESOP, The Fox and the Grapes, in AESOP¶S FABLES 15 (Heidi Holder ill., 1981). 125. See Wenzel, supra note 79, at 505 (describing two-way causality between self-interest and normative beliefs). 126. Id. 127. See generally Brian Mullen et al., The F alse Consensus Effect: A Meta-Analysis of 115 Hypothesis Tests, 21 J. EXPERIMENTAL SOC. PSYCHOL. 262 (1985) (discussing 115 studies that demonstrate the false consensus effect). See also Lee Ross, David Greene & Pamela House, The ³ F alse Consensus (IIHFW´ : An Egocentric Bias In Social Perception and Attribution Processes, 13 J. EXPERIMENTAL SOC. PSYCHOL. 279, 280 (1977). 128. On the process of social norm internalization and the relation to law, see Cooter, Good Citizens, supra note 96; Cooter, Three Effects, supra note 96. 129. HENRI TAJFEL, Social Identity and Social Comparison, in DIFFERENTIATION BETWEEN SOCIAL GROUPS: STUDIES IN THE SOCIAL PSYCHOLOGY OF INTERGROUP RELATIONS 61±76 (1978) (describing the power of collective psychological processes through group identification). A number of articles on law and social norms persuasively argue that effective enforcement policies must take into account how legal sanctions and regulations are perceived by subcultures or social groups. See, e.g., Dan M. Kahan, Social Influence, Social Meaning, and Deterrence , 83 VA. L. REV. 349, 350 (1997); Lawrence Lessig, Social Meaning and Social Norms, 144 U. PA. L. REV. 2181, 2184±85 (1996); McAdams, supra note 70. DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1287 might internalize the norms and ethics of the relevant community or subculture. In the context of tax compliance, for instance, it has been observed that taxpayers often follow the norms and behavior of the group that they identify with.130 In the context of copyright law, as they begin to identify with the anti-copyright subculture, a targeted campaign against file sharers might generate a defensive reaction and have the counterproductive effect of strengthening the community bond and support among file sharers. B. THE IRONY OF DETERRENCE In moving from moderate settlement demands to potential statutory rewards of up to $150,000 per illegally downloaded song, the litigation campaign has ramped up the potential liability exposure involved with file sharing.131 From the data in Study 2, we observed that elevated sanctions have a stronger effect on deterrence than an equivalent increase in the probability of punishment²even when the expected costs of punishment remain identical. This finding violates the axiom of rational choice theory that activity levels remain unaltered when the expected costs and benefits are identical. Behavioral refinements of the rational choice model have updated this belief of course, for instance, by demonstrating the impact of framing RQLQGLYLGXDOV¶ULVNHYDOXDWLRQV132 Our findings contribute to this literature by documenting differences in the sensitivity to varying modalities of enforcement with regard to the severity and probability of enforcement. Because the marginal benefits of increasing severity of punishment are higher than equivalent increases in the probability of punishment, our results suggest that it will often be more cost effective to enhance deterrence by raising the severity of the sanction.133 This result is somewhat ironic given the negative norm backlash effect of punishment observed in Study 1. Throughout our study, we observed that anti-copyright attitudes are more pronounced for participants in the See supra note 125. See supra Part II. PAUL WATZLAWICK, JOHN H. WEAKLAND & RICHARD FISCH, CHANGE: PRINCIPLES OF PROBLEM FORMATION AND PROBLEM RESOLUTION 92±109 (1974) (collecting existing studies); Amos Tversky & Daniel Kahneman, The Fra ming of Decisions and the Psychology of Choice, 211 SCI. 453 (1981) (exploring the framing effect in a number of laboratory experiments). 133. Additionally, the marginal costs of increasing the sanctions will be lower than for raising the rate of apprehension. 130. 131. 132. DO NOT DELETE 1288 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 high-severity/low-probability condition than for their counterparts in the low-severity/high-probability condition. Why do high probabilities induce lower levels of aversion than equivalent increases in the severity of sanctions? Perhaps file sharers react less adversely to elevated probabilities because technological issues, which are not directly controlled by the content industries, also determine the probability of apprehension. Potential sanctions, by contrast, are influenced more directly by the content industries in legal claims (and by prior political lobbying efforts). Additionally, if the rate of apprehension is low, elevated sanctions become more salient because they may create the perception that a few individuals are being singled out.134 The perception of being targeted may, in turn, again bolster a sense of community and kinship among file sharers.135 Finally, low probabilities might reduce the overall support for enforcement because they create an impression that such action will ultimately fail.136 C. ENFORCEMENT AND EXPECTATIONS Throughout our study, we exposed respondents to different modalities of enforcement. In our phased scenarios, some respondents moved from severe conditions to a state of enforcement immunity, while others moved from more moderate enforcement conditions. We observed a stronger backlash effect among frequent downloaders that were previously subjected to more stringent enforcement conditions. In reality, however, the sequence of enforcement and the subjective perception of entitlements proceeded from lower to higher degrees of HQIRUFHPHQW 2YHU WKH FRXUVH RI WKH HQWHUWDLQPHQW LQGXVWU\¶V OLWLJDWLRQ campaign, courts accepted expansive interpretations of intermediary liability,137 created novel doctrines of copyright infringement,138 and 134. See Robert J. Akerlof, A Theory of Social Motivation (MIT Working Paper No. 35, 2008), available at http://mit.academia.edu/RobertAkerlof/Papers/96712/A_Theory_of_Social_Motivation (describing social norm formation processes of groups and subcultures). 135. See Joseph R. Gusfield, Moral Passage: The Symbolic Process in Public Designations of Deviance, 15 SOC. PROBS. 175, 175±76 (1967) (describing the process of reaction and counter-reaction DIWHUDVXEFXOWXUHLVSXEOLFO\GHVLJQDWHGDV³GHYLDQW´  136. LITMAN, supra note 1, at 111. 137. A&M Records, Inc. v. Napster, Inc. (Napster I), 239 F.3d 1004, 1021±24 (9th Cir. 2001) (confirming district court finding that plaintiffs would likely be successful in establishing that Napster would be liable as a contributory and vicarious infringer). 138. MGM Studios, Inc. v. Grokster, Ltd. ( Grokster III), 545 U.S. 913, 936±37 (2005) (holding that distributing software with manifest intent to promote copyright infringement can render the VRIWZDUH¶VGLVWULEXWRUOLDEOHfor the infringing actions of third parties). DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1289 applied broad interpretations of statutory damage provisions.139 One can argue, especially in hindsight, that these decisions constitute sensible adaptations of copyright law in the face of new circumstances. During the initial years of file-sharing technology, however, this was not fully evident. In the early years of a new technology, there is often considerable ambiguity as to the teFKQRORJ\¶V potential social and economic implications. Moreover, new technologies may create enough legal ambiguity so that users of the technology are convinced that no infringements are committed.140 Even if the issues are relatively clean cut to legal experts, the ambiguity might be sufficient to enable self-serving interpretations regarding the legal status of new technology. 141 As research in the field of cognitive psychology demonstrates, individuals are inclined to construct facts in ways that align with their own preconceived beliefs.142 As a result, by the time that the major legal questions were resolved, users of file-sharing technology are no longer neutral bystanders. First, pro-filesharing norms might become internalized at some point.143 When internalized, such pro-sharing norms become more robust144 and present a more ardent challenge to legal enforcement efforts. Second, users of new WHFKQRORJ\ PLJKW H[SHULHQFH ORVV DYHUVLRQ ZKHQ WKH SUHYLRXVO\ ³IUHH´ access to content made available by the new technology is suddenly illegal.145 The perception of having VRPHWKLQJ³WDNHQDZD\´PLJKWDGGWR 139. See, e.g., Sony BMG Music EQWP¶W Y. Tenenbaum, No. 07cv11446-NG, 2009 U.S. Dist. LEXIS 115734 (D. Mass. Dec. 7, 2009). See also Itzkoff, supra note 8. 140. Ben Depoorter, The Several Lives of Mickey Mouse: The Expanding Boundaries of Intellectual Property Law, 9 VA. J.L. & TECH. 4, 67 (2004) (describing cyclical process of technology, norm adaptation, and judicial determination). 141. Ben Depoorter, Technology and Uncertainty: The Shaping Effect on Copyright Law, 157 U. PA. L. REV. 1831 (2009) (exploring the enabling effect of legal uncertainty and delay on norm formation processes). 142. See Linda Babcock & George Loewenstein, Explaining Bargaining Impasse: The Role Of Self-Serving Biases, 11 J. ECON. PERSPECTIVES 109, 111±16 (1997) (authors assigned participants in a study to either the plaintiff or defendant in a hypothetical automotive accident tort case with a maximum potential damages payment of $100,000. The plDLQWLII¶V SUHGLFWLRQ RI WKH OLNHO\ MXGLFLDO award was on average approximately  KLJKHU WKDQ WKH GHIHQGDQW¶V 7KH SODLQWLII¶V DYHUDJH nomination of a ³fair´ figure was approximately  KLJKHU WKDQ WKH GHIHQGDQW¶V  See generally Dan M. Kahan & Donald Braman, More Statistics, Less Persuasion: A Cultural Theory of Gun-Risk Perceptions, 151 U. PA. L. REV. 1291 (2003) (describing culturally biased interpretations of data). 143. On the internalization of social norms, see Cooter, Good Citizens?, supra note 96; Cooter, Three Effects, supra note 96, at 17±20. 144. See Geoffery Neri, Note, Sticky Fingers or Sticky Norms? Unauthorized Music Downloading and Unsettled Social Norms, 93 GEO. L.J. 733, 757±58 (2003). 145. 7KLVEULQJVWRPLQGWKHIDPRXVVWDWHPHQWRI$WWRUQH\*HQHUDO5REHUW.HQQHG\WKDW³WKHSRRU man looks upon the law as an enemy, not as a friend. For him the law is always taking something DZD\´Attorney General Robert Kennedy, Address at University of Chicago Law Day (May 1, 1964) DO NOT DELETE 1290 3/22/2012 2:00 PM S O UTH ERN CALI F ORNIA LAW REVIEW [Vol. 84:1251 the normative resistance to the enforcement of newly established rights in copyright law.146 VI. CONCLUSION In situations where a social group perceives legal sanctions to be excessive with respect to the behavior that is being punished, enforcement may inadvertently strengthen preexisting adverse normative positions. This Article has examined the hypothesis that counterproductive effects can help explain the pervasiveness of copyright infringements in the face of lawsuits against file sharers. We investigated empirically the interaction between copyright enforcement and preexisting norms in two studies that examined the reactions of copyright infringers to varying probabilistic penalties. Our results suggest that copyright enforcement is a double-edged sword. Sanctions have a deterrent effect on file-sharing behavior. Sanctions, however, can backfire and increase anti-copyright sentiments among file sharers. This raises particular difficulties for copyright enforcement because negative norm effects might prevent sanctions from being set at levels that attain deterrence. Second, our findings suggest that enforcement efforts would likely be more effective if targeted specifically to different types of copyright offenders. The results show a backlash effect for all users, but frequent users especially will download more if previously faced with punitive enforcement measures. The observed backlash effect among frequent users casts doubt upon the enforcement tactics employed by the entertainment industry. By focusing litigation on frequent offenders, copyright holders bolster anti-copyright norms among this group, while foregoing opportunities to promote pro-copyright norms among occasional infringers. For content holders, the limitations of deterrence-based approaches increase the importance of fostering social attitudes that support the rights of copyright holders.147 Indeed, the attempts by the music and film (cited in Werner Z. Hirsch, 5HGXFLQJ/DZ¶V8QFHUWDLQW\DQG&RPSOH[LW\, 21 UCLA L. REV. 1233, 1247 n.28 (1974)). 146. Experimental research demonstrates that individuals value certain resources or legal entitlements that they possess more than they would value the exact same thing had they never possessed it at all. See Daniel Kahneman, Jack L. Knetsch & Richard H. Thaler, Experi mental Tests of the Endowment Effect and the Coase Theorem, 98 J. POL. ECON. 1325, 1326 (1990); George Loewenstein & Samuel Issacharoff, Source Dependence in the Valuation of Objects, 7 J. BEHAV. DECISION MAKING 157, 165 (1994). 147. See Kahan, supra note 20 (suggesting that gradual approaches are more effective when antilegal norms are in place). DO NOT DELETE 2011] 3/22/2012 2:00 PM C OPYRIG HT BAC KLAS H 1291 industries to associate file sharing with theft and stealing in public relations campaigns seek to manipulate the underlying social meaning of copyright infringements. Social norm theorists recognize, however, that norms are notoriously hard to control.148 Moreover, in the copyright context, the entertainment industry faces opposition from normative frameworks that advocate for more communal models of cultural exchange and relaxed copyright standards.149 Additionally, the potential role of the government in fostering pro-copyright norms is limited in this context. Legislative developments have created a public perception that copyright law is driven by lobbying and political interest group pressure on behalf of the entertainment industry.150 If litigation against teenagers and universities was part of a grand strategy to influence the social meaning of file sharing, LWLV FOHDUWKDW WKH LQGXVWU\¶V FRS\ULJKW FDPSDLJQ KDV IDLOHG 2QH SRVVLEOH explanation for this failure lies with the counterproductive norm effects documented in this Article. 148. See, e.g., Jeffrey J. Rachlinski, The Li mits of Social Norms, 74 CHI.-KENT L. REV. 1537 (2000); Jeffrey J. Rachlinski, The ³New´ Law and Psychology: A Reply to Critics, Skeptics, and Cautious Supporters, 85 CORNELL L. REV. 739 (2000). 149. See, e.g., JAMES BOYLE, THE PUBLIC DOMAIN: ENCLOSING THE COMMONS OF THE MIND 179±204 (2008); LAWRENCE LESSIG, FREE CULTURE: HOW BIG MEDIA USES TECHNOLOGY AND THE LAW TO LOCK DOWN CULTURE AND CONTROL CREATIVITY 282±86 (2004). 150. See, e.g., Ian McPherson, Copyright Becomes a Tool of the Cartels, NETNACS, Oct. 2002 (on file with author); John Naughton, Mickey Mouse Threatens to Block All Ideas in Future , OBSERVER, Feb. 24, 2002, available at http://www.guardian.co.uk/technology/2002/feb/24/ business.columnists. For an analysis of the political process of copyright reform, see generally Jessica D. Litman, Copyright Legislation and Technological Change , 68 OR. L. Rev. 275 (1989) (describing the entrenched practice of drafting copyright statutes through negotiations among industry representatives). Effects of punish ment Probability of Total High Low Total High Low Total High Low Severity of punishment 4.88(1.2 4.90(1.26 4.87(1.17) 3.45(1.20) 3.43(1.27) 3.46(1.17) 4.30(1.71) 4.49(1.37) 4.11(1.74) High 5.37(1.3 5.71(1.32 5.03(1.49) 4.32(1.38) 4.27(1.32) 4.41(1.49) 4.79(1.34) 4.99(1.22) 4.59(1.54) Total 5.29(1.37 4.96(1.37) 3.85(1.36) 3.93(1.38) 4.74(1.40) 4.35(1.66) 4.62* Severity of punishment 9.13** Interaction effect 0.63 23.41** 5.50* 9.69** 4.41* 0.95 0.28 Occasional 4.88(1.08) 3.44( 0.98) 4.21(1.51) Average 5.04(1.23) 4.28(1.76) 4.66(1.80) Frequent 5.45(1.17) 4.16(1.38) 4.94(1.39) 0.27 3.10 ‚ Effects of user type User type S tatistical effect, 4.01* APPENDIX S tatistical effects, F (1,204) Probability of S O UTH ERN CALI F ORNIA LAW REVIEW Low DO NOT DELETE Backlash on Behavior 1292 Evaluation of the enforcement policies by the music industry 3/22/2012 2:00 PM [Vol. 84:1251 TABLE 1. Study 1, means, standard deviations, and statistical effects of sanction and user type Anti-copyright norms 4.41(1.36) 4.05(1.64) 4.49(1.61) 4.70(1.13) High severity/Low probability 4.88(1.30) 4.69(1.00) 5.46(1.31) 4.74(1.14) S tatistical effect, F (1,108) 4.60* 4.31* 6.25* 0.19 4.51(1.46) 4.11(1.05) 4.66 (1.62) 4.55( .95) 4.12(1.00) 4.12(1.92) 5.05 (1.50) 4.48(1.36) 4.93(1.16) 4.32(1.35) 5.70 (1.10) 5.13(1.15) 8.46** 4.49* 8.81** 12.36*** Effects of user type User type Occasional Average Frequent S tatistical effect, F (1,108) 1293 3/22/2012 2:00 PM Low severity/High probability Effects of punish ment DO NOT DELETE Backlash on Behavior C OPYRIG HT BAC KLAS H Download behavior 2011] Evaluation of the enforcement policies by the music industry TABLE 2. Study 2, means, standard deviations, and statistical effects of sanction and user type Anti-copyright norms