Barton Beebe, a copyrights and trademarks expert and professor at Cardozo School of Law, recently published an article titled “An Empirical Study of U.S. Copyright Fair Use Opinions, 1978-2005“.[1] The Study is a thorough statistical examination of every federal court opinion that significantly involved fair use since 1978. The study was so insightful that it is important to summarize and share some of the key empirical data to the copyright law professional community. In short, much of the data went “against the grain” of conventional thinking with respect to fair use litigation and the Section 107 four-factor test. Specifically, many personal assumptions and understandings about fair use litigation and § 107 were tested and negated.

Hopefully, some of the empirical data will help others better understand fair use litigation, and possibly assist in future copyright infringement and/or fair use cases. In particular, the data exposes trends and judicial tendencies that tell lot about what should be the focal point of a fair use argument or an attack on a fair use argument. This data could either be used in a motion for summary judgment, or it could be used to decide whether or not to even file a motion for summary judgment (cite to Beebe, not me!). Still, William Patry, in his blog in May, 2007, warned that the data could be skewed because of subjective application:

Professor Barton Beebe has undertaken to understand how fair use has been decided on the ground . . . [e]ntitled “An Empirical Study of U.S. Copyright Fair Use Opinions, 1978-2005,” Professor Beebe canvasses how courts have treated each factor, the phenomenon of stampeding, and reversal rates. It is an important work that should be carefully studied. Whether it will help advocates in actual cases is a different issue, and is complicated by the strong divergence between the rhetoric employed by courts in reaching results and how they reach those results. Even when courts say, for example, that they are engaging in an analysis of whether defendant’s use is “transformative,” what any particular judge means by transformative may vary greatly. In the end, like beauty, fair use tends to be in the eye of the beholder.[2]

Nevertheless, the data is very telling.

Section 107 of the Copyright Act of 1976

“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. 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.”[3]


The Empirical Data

The following facts and empirical data were taken from a close study of 306 federal court opinions focusing on fair use and § 107. All data and accompanying quotes are taken from Professor Beebe’s article. Behold the truth about Fair Use and § 107 litigation:

Macro Issues

  • Exclusivity of the four factors? It is frequently stated by practitioners and courts that the four factors of consideration under § 107 are not exclusive, but in practice it seems that this is not true. “Commentators have long called upon courts to look to additional factors, such as ‘fairness’, and in the opinions studied, the courts themselves frequently acknowledged that the section 107 test is illustrative rather than limitative. Yet only 17.0% of the opinions explicitly considered one or more additional factors, and only 8.8% stated that the additional factor was relevant to the fair use determination.”[4]
  • Potential fair uses “chilled” by the threat of litigation? Despite the fact that “Federal Judicial Center data . . . suggests that a steady average of approximately 2000 copyright infringement complaints were filed per year in federal district courts” between 1978 and 2005, only 10.9 fair use opinions were rendered annually, with only 4.6 of those actually finding fair use.[5] Thus, “[e]ven while fair use concerns may pervade the public sphere (and legal-academic commentary on it), many fair use controversies may never take the form of a filed complaint, as potential fair uses are “chilled” by the mere threat of litigation. Empirical work supports this explanation.”[6]
  • Fair use case law comes from four courts. Of the 306 opinions examined from every federal court in our legal system, 122 of them were from only four courts - the Supreme Court, the Second and Ninth Circuits, and the Southern District of New York. Thus, “when we speak of modern U.S. fair use case law, we are speaking primarily of [these courts] . . . and the progeny of these opinions in the other federal courts.”[7] The Sixth Circuit and the federal courts within its authority are third on the list of Circuit Courts, after the Second and Ninth Circuits, but have only rendered 9% of all Circuit Court opinions and 5.7% of all District Court opinions (half (6/12) of which are from the Eastern District of Michigan).
  • Fair use parties frequently resort to summary judgment motions; SMJ motions are almost always granted. Despite the fact that “fair use is a mixed question of law and fact,”[8] the reality from the empirical data is that fair use is often resolved at the summary judgment stage, where there are no genuine issues of material fact. “More than half of the opinions addressed a motion or cross-motion for summary judgment, and of these 121 opinions, 86% granted the motion or one of the cross-motions.”[9]
  • First Amendment is still a large component of the fair use argument, but is not necessary to it. Although it seems quite logical that a First Amendment argument, or at least First Amendment concerns, would accompany a fair use opinion, not as many courts incorporate the First Amendment as first thought. Only 25% of the 306 opinions invoked the First Amendment in its analysis or found it relevant, and fewer than that actually analyzed First Amendment concerns within the fair use analysis itself.[10]
  • Although it is conventional wisdom that fair use case law is littered with reversals and divided courts, this is not the case. Of the 71 Circuit Court majority opinions examined, only “24 reversed the district court’s fair use holding (for a reversal rate of 33.8%) . . .”[11] This percentage is on par with estimates of overall circuit court reversal rates, and well below estimates of Federal Circuit reversal rates in patent cases (47%). Of the 211 district court opinions, only 53 were appealed, and only 18, or 8.5%, were reversed.[12] It is important to note that most of the unreversed opinions were findings of no fair use. 10/10 Seventh Circuit affirmations were of findings of no fair use at the district court level, while the district courts of the Second Circuit did not find fair use in any of their 9 unreversed bench trial opinions, and 9/11 affirmations in the Ninth Circuit were findings of no fair use.[13]
  • Factors one and four of the fair use test dictate the outcome of the test. So you know which basket in which to lay most of your eggs: “The outcome of factor four coincided with the outcome of the overall test in 83.8% of the 297 dispositive opinions while the outcome of factor one coincided with the outcome of the overall test in 81.5% of these same opinions. By comparison, the outcome of factor two coincided with the outcome of the overall test in 50.2% of these opinions. As for the combined influence of factors one and four, in 214 (or 72.1%) of the opinions, factors one and four either both favored or both disfavored fair use. In all but one of these opinions, the outcome of the fair use test followed the outcome of these two factors.”[14]

Micro Issues

  • Factor one, “purpose and character of the use”, is a good bet if you can get it on your side. ” . . . 95.3% of the 148 opinions that found that factor one disfavored fair use eventually found no fair use, while 90.2% of the opinions that found that the factor favored fair use eventually found fair use. Only factor four can boast of higher percentages.”[15]
  • The “commerciality” inquiry is much more popular than the “transformativeness” inquiry under factor one, but less influential. “Of the 306 opinions, 84.0% explicitly considered whether the use was commercial or noncommercial in nature under factor one, while only 38.2% explicitly considered the transformativeness of the defendant’s use under the factor.”[16] The attention that litigants and courts lend to the commerciality inquiry is a direct result of the Sony Betamax case, in which the Supreme Court stated that, “if the intended use is for commercial gain, [the likelihood of future harm] may be presumed.”[17] Despite this doctrine, “a finding that the defendant’s use was for a commercial purpose (which was made in 64.4% of the opinions) did not significantly influence the outcome of the fair use test in favor of an overall finding of no fair use. Rather, it was a finding that the defendant’s use was for a noncommercial purpose (which was made in 15.4% of the opinions) that strongly influenced the outcome of the test in favor of an overall finding of fair use.”[18] Of the 29 opinions examined between Sony and Campbell v. Acuff-Rose Music, Inc. (which altered the presumption),[19] that found a commercial use, only 15 found no fair use.[20] However, although “transformativeness” is deployed and explored much less frequently, it is much more controlling of the outcome. “[T]he data suggest that while a finding of transformativeness is not necessary to trigger an overall finding of fair use, it is sufficient to do so. We can infer that it is not necessary in light of the fact that 25 (or 36.8%) of the 68 post-Campbell opinions that found fair use made no reference to transformativeness. We can infer that it is sufficient . . . in light of the fact that each of the 13 circuit court opinions and 27 of the 29 district court opinions that found the defendant’s use to be transformative also found it to be a fair use . . .”[21] Furthermore, transformativeness seems to supersede a finding of commercial use. “In 28 opinions, the court found the defendant’s use to be both commercial and transformative under factor one, and in 26 of these opinions, the court found both that factor one and the overall test favored fair use . . .”[22] Put in different terms, “a defendant has a 35.5% chance of winning the fair use defense where it has made a nontransformative, commercial use of a creative, published work . . . [but] if that same use were found to be transformative, the defendant’s chance of winning the fair use defense would increase to 94.9%.”[23]
  • Bad faith is rarely argued, but when it is, it usually means no fair use. Only 14 opinions of the 306 found improper conduct, and in 12 of those the court found no fair use. Conversely, in 28 opinions where the court expressly stated that the defendant’s conduct was not improper, the court found fair use in 24 of them.[24]
  • The preambular purposes don’t mean much. Despite § 107’s inclusion of examples of fair purposes for use of copyrighted material, empirical data tell us that even if a use falls squarely under one of the categories, it doesn’t mean it’s a fair use. Only 40.9% of 22 cases where the use was found to be an enumerated research use was found to be fair use. Even more surprising, only 48.2% of uses found to be educational were found to be a fair use.[25]
  • Factor Two, the “nature of the copyrighted work”, is not only not influential, but often not even discussed. Despite § 107’s command that “the factors to be considered shall include” factor two, “17.7% of the 306 opinions failed even to refer to the factor, while an additional 6.5% did so only to call it irrelevant.”[26]
  • If you can prove that the nature of the copyrighted work is factual, your chances of winning a fair use defense increase dramatically. ” . . . a defendant has a 35.5% chance of prevailing in its fair use defense when it has made commercial use of a published work that is found to be creative in nature . . . and an 80.3% chance of prevailing in its fair use defense when it has made commercial use of a published work that is found instead to be factual in nature.”[27]
  • Despite the Supreme Court’s attempt at establishing that a finding that the copyrighted work is unpublished disfavors fair use, the facts show the opposite. Based on empirical data, “the defendant’s chances of succeeding in its fair use defense increase from 27.4% to 57.8% when the status of the plaintiff’s work shifts from not-published to published - and not at all when the status of the plaintiff’s work shifts from unpublished to [published].”[28]
  • Factor three, “amount and substantiality of use”, is only important if the “heart” of the work is taken, and should not be measured quantifiably. It makes sense to think that the more of a copyrighted work one takes, the less chance that a fair use will be found. It would follow, then, that if the defendant takes the entire copyrighted work and reproduces it that this should undoubtedly favor a finding of no fair use. The empirical data tell us differently. “Of the 99 opinions that addressed facts in which the defendant took the entirety of the plaintiff’s work, 27.3% [still] found fair use.” The better test is whether the “essence” or the “heart” is taken. “Courts explicitly found that the defendant took the heart of the plaintiff’s work in 37 opinions, and found no fair use in 35 of these. Courts made the opposite finding - that the defendant did not take the heart of the plaintiff’s work - in 25 opinions, and found fair use in 23 of these.”[29] Specifically, “if a defendant makes a commercial use of a creative, published work, its chances of succeeding in its fair use defense decline from 35.5% to 12.0% if it is found to take the entirety of that work, and to 1.0% if it is found to take the heart of that work.”[30]
  • Factor Four is a MUST for plaintiffs attempting to prove no fair use, and for defendants attempting to prove fair use. Consider the following: “of the 141 opinions that found that factor four disfavored fair use, 140 found no fair use.”[31] Enough said. Consider the following: “of the 116 opinions that found that factor four favored fair use, all but 6 found fair use.”[32] Enough said.

[1] 156 Penn. L. Rev. 549 (2008).

[2] William Patry, The Patry Copyright Blog, “Barton Beebe’s Fair Use Study”, found at (last visited 12/2/08).

[3] 17 U.S.C. § 107.

[4] Barton Beebe, An Empirical Study of U.S. Copyright Fair Use Opinions, 1978-2005, 156 Penn. L. Rev. 549, 564 (2008).

[5] Beebe, An Empirical Study, at 565.

[6] Id., at 565, n. 66.

[7] Id., at 568.

[8] Castle Rock Entm’t, Inc., v. Carol Publ’g Group, Inc., 150 F.3d 132, 137 (2d. Cir. 1998).

[9] Beebe, An Empirical Study, at 570.

[10] Id., at 573.

[11] Id., at 574.

[12] Id., at 575.

[13] Id.

[14] Id., at 584.

[15] Id., at 597.

[16] Id., at 598.

[17] Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 451 (1984).

[18] Beebe, An Empirical Study, at 602.

[19] 510 U.S. 569 (1994).

[20] Beebe, An Empirical Study, at 601.

[21] Id., at 605.

[22] Id., at 606.

[23] Id.

[24] Id., at 606.

[25] Id., at 610.

[26] Id.

[27] Id., at 611-612.

[28] Id., at 614.

[29] Id., at 616.

[30] Id.

[31] Id., at 617.

[32] Id.

This entry was posted on Tuesday, January 27th, 2009 at 3:32 am.
Categories: Copyright Caucus ~ by Ian McClure.

No Comments, Comment or Ping

Reply to “Copyright Fair Use Empirical Study by Barton Beebe”