An Empirical Study of U.S. Copyright Fair Use Cases, 1978 …fordhamipconference.com/wp-content/uploads/2010/08/3A.Beebe_.pdf · • LEXIS Federal Court Cases, Combined database:

1

1

An Empirical Study of U.S. Copyright Fair Use Cases, 1978-2011

adapted from 156 PA. L. REV. 549 (2008)

Prof. Barton BeebeNYU School of Law

.

2

Outline

I. Background

II. The Data Set

III. Findings

A. Summary Statistics

B. Interfactor Analysis

C. Factor-Specific AnalysisIV. Conclusion

2

3

Notwithstanding the provisions of sections 106 and 106A, the fair use of a copyrighted work, including such use by reproduction in copies or phonorecords or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright. . . .

17 U.S.C. 107

I. BackgroundA. The Preamble and the Four-Factor Test

4

Notwithstanding the provisions of sections 106 and 106A, the fair use of a copyrighted work, including such use by reproduction in copies or phonorecords or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright. . . .

17 U.S.C. 107


3

5

. . . In determining whether the use made of a work in any particular case is a fair use the factors to be considered shall include—

(1) the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;

(2) the nature of the copyrighted work;

(3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and

(4) the effect of the use upon the potential market for or value of the copyrighted work.

The fact that a work is unpublished shall not itself bar a finding of fair use if such finding is made upon consideration of all the above factors.

17 U.S.C. 107


6

• Test-Wide Questions

– Which factor outcomes drive the outcome of the test?

– How do the factor outcomes interact?

– To what extent do the factor outcomes stampede?

– How often does the FU defense succeed and is there significant intercircuit variation?

• Factor-Specific Questions

– How does the subfactor doctrine operate in practice (e.g., transformativeness, commercial use, “heart” of the work)?

• Judge-specific questions:

– Is there any relation between the ideology, age, gender, or tenure of the judge and the outcome of the test or of specific factors?

• What does the Section 107 test as applied tell us about legal multifactor reasoning in general?

I. BackgroundB. Research Questions

4

7

• More generally, what can a quantitative method tell us that a qualitative method cannot?

– “Systematic content analysis” of court opinions

– See generally Mark A. Hall & Ronald F. Wright, Systematic Content Analysis of Judicial Opinions, 96 CAL. L. REV. 63 (2008)

I. BackgroundB. Research Questions

8

• David Nimmer, Fairest of Them All and Other Fairy Tales of Fair Use, 66 LAW & CONTEMP. PROBS. 263 (2003)

I. BackgroundC. Literature Review

5

9

• David Nimmer, Fairest of Them All and Other Fairy Tales of Fair Use, 66 LAW & CONTEMP. PROBS. 263, 280 (2003)


10

• David Nimmer, Fairest of Them All and Other Fairy Tales of Fair Use, 66 LAW & CONTEMP. PROBS. 263, 280 (2003)


6

11

• Barton Beebe, An Empirical Study of U.S. Copyright Fair Use Opinions, 1978-2005, 156 PA. L. REV. 549 (2008)

• Anthony Reese, Transformativeness and the Derivative Work Right, 31 COLUMBIA J. L. & ARTS 101 (2008)

• Pamela Samuelson, Unbundling Fair Uses, 77 FORDHAM L. REV. 2537 (2009)

• Neil Netanel, Making Sense of Fair Use, 15 LEWIS & CLARK L. REV. 715 (2011)

• Matthew Sag, Predicting Fair Use (working paper available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1769130)


12

• Dan Simon, A Third View of the Black Box: Cognitive Coherence in Legal Decision Making, 71 U. Chi. L. Rev. 511 (2004)

– Defines coherence-based reasoning as reasoning in which the “decision-making process progresses bidirectionally: premises and facts both determine conclusions and are affected by them in return. A natural result of this cognitive process is a skewing of the premises and facts toward inflated support for the chosen decision.”

– An Empirical Study of the Multifactor Tests for Trademark Infringement, 95 Cal. L. Rev. 1581 (2006)

I. BackgroundD. Theoretical Approaches to Multifactor Decisionmaking

7

13

• LEXIS Federal Court Cases, Combined database: copyright and "fair use" and 107 and date(geq(1-1-1978) and leq (12/31/2011))

• Yielded 785 opinions (on 1/1/2012)

• Filtered down to 395 opinions that made “substantial use” of the four-factor test or that were concurrences or dissents to appellate opinions that did so

– “Substantial use” defined as analyzing at least two factors from the test

– Examples of excluded opinions:

• Ty, Inc. v. Publications International, Ltd., 292 F.3d 512 (7th Cir. 2002) (declining to apply factors on ground that the “statutory definition” of fair use, “though extensive[,] is not illuminating”) (Posner, J.)

II. The Data SetA. Collecting/Coding the Opinions

14

– Examples of excluded opinions (cont.):

• Pacific and Southern Company, Inc. v. Duncan, 572 F. Supp. 1186, 1195 (N.D. Ga. 1983) (“[I]t becomes readily apparent that Ms. Duncan's use is not for a purpose such as ‘criticism, comment, news reporting, teaching, scholarship, or research.’ That being the case, analysis of the four factors listed under § 107 is unnecessary.”), aff’d in part and rev’d in part on other grounds, 744 F.2d 1490 (11th Cir. 1984)

• Madison River Mgmt. Co. v. Business Mgmt. Software Corp., 387 F. Supp. 2d 521 (M.D.N.C. 2005) (declining to apply factors and analogizing instead to previous 7th Circuit case)


8

15

• 74 variables

• Coding of factor outcomes

0 = Not addressed

1 = Found to favor plaintiff

2 = Found to favor defendant

3 = Found to be neutral

4 = Found to be a fact issue

5 = Found to be not relevant

6 = Unclear

• Software: Atlas.ti, Excel, Stata


16

Atlas.ti

9

17

• 63% of opinions explicitly stated valence of factors; 36% explicitly balanced the factors at the conclusion of their analysis

• 20-Opinion Moving Average of the Proportion of Opinions Stating and Reviewing the Valences of Each Factor

II. The Data SetB. Amenability to Coding

18

10

19

III. FindingsA. Summary Statistics

• 281 district court opinions; 107 circuit court opinions (87 majority opinions, 6 concurrences, 14 dissents); 7 Supreme Court opinions (4 majority opinions; 1 concurrence; 2 dissents)

• Distribution by year

20

11

21


• 281 district court opinions; 107 circuit court opinions (87 majority opinions, 6 concurrences, 14 dissents); 7 Supreme Court opinions (4 majority opinions; 1 concurrence; 2 dissents)

• Distribution by court

– Of the 281 district court opinions, 76 (27.1%) came from the SDNY, 22 (7.8%) from the C.D. Cal., and 18 (6.4%) from the N.D. Cal.

– Of the 107 circuit court opinions, 40 (37.4%) came from the 2d Circuit, 28 (26.2%) from the 9th Circuit, and 10 (9.4%) from the 6th Circuit.

– The Supreme Court, 2d Circuit, 9th Circuit, and SDNY dominate our fair use case law.

22


• Intercircuit influence

– Circuit and district court citation practices to other circuits

• Opinions outside of the 2d Circuit cited on average to 1.47 2d Circuit opinions in their fair use analyses

• Comparable numbers for opinions outside of:

– Outside 9th Circuit: .71 citations to 9th Circuit opinions

– Outside 5th Circuit: .11 citations to 5th Circuit opinions

– Other circuits: negligible

– Outside SDNY: .580 citations to SDNY opinions

• Reversal, Dissent, and Appeal Rates

– Of the 87 circuit court majority opinions, 30 reversed the district court (34.5% reversal rate) and 14 met with dissents (13.1% dissent rate)

– Of the 281 district court opinions, 55 were appealed (24.8% appeal rate), with 36 (16.2%) affirmed and 19 (8.6%) reversed

– NB: This includes opinions from 2011, appeals of which may not yet be recorded

12

23

Distribution and Disposition of Unreversed District Court Opinions by Circuit and Posture (for leading postures)

III. FindingsA. Summary Statistics – FU Win Rates

24

13

25

Distribution and Disposition of Unreversed Circuit Court Majority Opinions (for leading postures)


26

14

27


• Intercircuit and interdistrict variation in win rates

– Circuit courts

• National FU win rate in dispositive opinions: 43.8% of 80 majority opinions found FU

– 2d Circuit: 53.9% of 26 2d Circuit dispositive majority opinions found FU; 38.9% of non-2d Circuit dispositive majority opinions found FU

– 9th Circuit: 44.0% of 25 9th Circuit dispositive majority opinions found FU; 43.6% of non -9th Circuit dispositive majority opinions found FU

– District courts

• National FU win rate in dispositive opinions: 43.3% of 252 opinions found FU

– SDNY: 52.9% of 68 dispositive SDNY opinions found FU; 39.7% of non-SDNY dispositive district court opinion found FU

– C.D. Cal.: 52.4% of 21 dispositive C.D. Cal. opinions found FU; 42.4% of non-C.D. Cal. dispositive district court opinions found FU

28


15

29

III. FindingsA. Summary Statistics – Win Rates Over Time

Proportion of Dispositive Opinions Per Year Finding Fair Use (All Courts)

30


Proportion of Dispositive Opinions Per Year Finding Fair Use (All Courts)

16

31


Proportion of Dispositive Opinions Per Year Finding Fair Use (2d Circuit Appellate and District Courts)

32


Proportion of Dispositive Opinions Per Year Finding Fair Use (9thCircuit Appellate and District Courts)

17

33


Sag, at 34

34


Sag, at 42

18

35


Netanel, at 755

36

III. FindingsB. Interfactor Analysis – Pairwise Correlations

19

37

III. FindingsB. Interfactor Analysis – Logistic Regression

38

III. FindingsB. Interfactor Analysis

Twenty-Opinion Moving Average of the Proportion of the Opinion’s Fair Use Analysis Devoted to Each Factor

Factor 1

Factor 4

Factor 3

Factor 2

20

39

III. FindingsB. Interfactor Analysis

39

Stampede Score = number of factors found to favor fair use minus the number of factors

found to disfavor fair use

Stampeding?

(1978-2005)

40

21

41

III. FindingsC. Factor-Specific Analysis: Factor One

Twenty-Opinion Moving Average of the Proportion of Opinions Analyzing Transformativeness, Commerciality, and Applying the Sony Presumption

Transformativeness

Sony Presumption

Commerciality

42


• Transformativeness:

– A finding of transformativeness is generally sufficient but not necessary to trigger a finding of fair use.

– Sufficient: Of the 395 opinions from 1978 through 2011, 72 opinions found a transformative use and 63 (87.5%) of these found fair use.

– Four of the 72 found a transformative use and no fair use:• Castle Rock Entertainment v. Carol Publ. Group, 955 F. Supp.

260 (S.D.N.Y. 1997)

• SunTrust Bank v. Houghton Mifflin Co., 136 F. Supp. 2d 1357 (N.D. Ga. 2001)

• Warner Bros. Entertainment Inc. v RDR Books, 575 F. Supp. 2d 513 (S.D.N.Y. 2008)

• Bridgeport Music, Inc. v. UMG Recordings, Inc., 585 F.3d 267 (6th Cir. 2009)

22

43


• Transformativeness (cont.):

– Not Necessary: 161 of the 395 opinions from 1978 through 2011 found fair use. Of these, only 63 found a transformative use. 9 explicitly found no transformative use.

• See, e.g.:

– Religious Tech. Ctr. v. Netcom On-Line Commun. Servs. | 1997 U.S. Dist. LEXIS 23572 (N.D. Cal. 1997)

– Gulfstream Aerospace Corp v. Camp System Intl., 428 F.Supp.2d 1369 (S.D. Ga. 2006)

• What is a transformative use?

– Netanel: a use for a “different expressive purpose from that for which the work was created”

– Reese: transforming content vs. transforming purpose

– Samuelson: “orthogonal uses”

– Sag: “creativity shift”

44


• Commercial use

– A finding that the defendant made a commercial use does not appear to affect the overall outcome of the fair use analysis.

• But see Sag: “direct commercial use” makes fair use finding less likely

– The commercial use issue is interesting for what it shows of the indiscipline of the lower courts made possible by the S Court’s effort to maintain appearances.

23

45


• Supreme Court statements about commerciality in factor one:– Sony at 451: “If the Betamax were used to make copies for a

commercial or profit-making purpose, such use would presumptively be unfair. The contrary presumption is appropriate here, however…”

– Harper & Row at 562: “The fact that a publication was commercial as opposed to nonprofit is a separate factor that tends to weigh against a finding of fair use. ‘Every commercial use of copyrighted material is presumptively an unfair exploitation of the monopoly privilege that belongs to the owner of the copyright.’ Sony Corp. of America v. Universal City Studios, Inc., 464 U.S., at 451.”

– Stewart v. Abend at 237: “’Every [unauthorized] commercial use of copyrighted material is presumptively an unfair exploitation of the monopoly privilege that belongs to the owner of the copyright.“” Sony Corp. of America v. Universal Studios, Inc., supra, 464 U.S. at 451

– Acuff-Rose at 585: “[A]s we explained in Harper & Row, Sony stands for the proposition that the ‘fact that a publication was commercial as op-posed to nonprofit is a separate factor that tends to weigh against a finding of fair use.’”

46

Twenty-Opinion Moving Average of the Proportion of Opinions Analyzing Transformativeness, Commerciality, and Applying the Sony Presumption

Transformativeness

Sony Presumption

Commerciality

24

47


• Bad Faith

– 53 opinions explicitly found bad faith or no bad faith. Of these, 46 ruled on fair use in the direction of the bad faith finding.

• Preambular Uses

48

III. FindingsC. Factor-Specific Analysis: Factor Two

• Creative vs. factual works:

– 161 opinions addressed facts where plaintiff’s work was held by the court to be creative. Of these, 36.7% found fair use.

– 60 opinions addressed facts where plaintiff’s work was held by the court to be factual. Of these, 63.3% found fair use.

– 107 opinions did not address the creative/factual issue.

• Published vs. non-published works:

– Opinions claiming unpublished supports no FU on privacy or economic grounds: 33

– Opinions claiming unpublished supports FU because the work is not available otherwise: 4

– Opinions claiming published supports no FU because work can be purchased: 6

– Opinions claiming published supports FU because work is already available to the public: 51

– Opinions not addressing publication status: 279

25

49

III. FindingsC. Factor-Specific Analysis: Factor Three

• Amount of the plaintiff’s work that defendant copied

– 139 opinions found that the defendant copied the entirety of the plaintiff’s work. Of these 33.8% found fair use.

• Defendant took the “heart” of the plaintiff’s work

– 44 opinions found that the defendant did so. 84.0% found no fair use.

– 311 opinions did not address this subfactor

50

III. FindingsC. Factor-Specific Analysis: Factor Four

26

51


• Correlation between outcome of factor four and outcome of overall test:

– Of 337 dispositive opinions, 158 found that factor four disfavored fair use. 155 (98.1%) of these found no fair use.

• Williamson v. Pearson Educ., 2001 U.S. Dist. LEXIS 17062 (S.D.N.Y. 2001)

• Warren Publishing Co. v. J. David Spurlock, 645 F. Supp. 2d 402 (E.D. Pa. 2009)

• Fuentes v. Mega Media Holdings, Inc., 2011 U.S. Dist. Lexis 70996 (S.D. Fla. 2011)

– Of 337 dispositive opinions, 145 found that factor four favored fair use. 136 (93.8%) of these found fair use.

52


Campbell Presumption

Slippery Slope

Most Important Factor

Sony Presumption

27

53

III. FindingsD. Does Judicial Ideology Affect Fair Use Outcomes?

See also Matthew Sag, Tonja Jacobi & Maxim Sytch, Ideology and Exceptionalism in Intellectual Property – An Empirical Study, 97 CAL. L. REV. 801 (2009)

From Barton Beebe, Does Judicial Ideology Affect Copyright Fair Use Outcomes?: Evidence From the Fair Use Case Law, 31 COLUMBIA J. L. & ARTS

101 (2008)

54

IV. Conclusions

• Basic descriptive conclusions

– Unexceptional reversal and appeal rates

– No persuasive evidence of stampeding

– Zombie precedent lives

– Some difference in win rates among the courts

• The four-factor test

– Judges take the factors seriously

– Commerciality appears to be unimportant

– Transformativeness appears to be nearly dispositive and increasingly so (Netanel)

• Transformativeness as repurposing

– Factor four largely makes conceptual space for a balancing of factor one and factor three

An Empirical Study of U.S. Copyright Fair Use Cases, 1978 …fordhamipconference.com/wp-content/uploads/2010/08/3A.Beebe_.pdf · • LEXIS Federal Court Cases, Combined database:

Documents