An Empirical Study of U.S. Copyright Fair Use Opinions Updated, 1978-2019

Barton Beebe*

Download a PDF version of this article here.

Watch our second episode of The Author Series featuring Professor Beebe here.

Read JIPEL’s letter regarding data validation for this article here.

A copy of the raw data analyzed for this article can be found here.

This article presents a brief update through 2019 of the author’s previous quantitative study of all reported federal court opinions that applied the Copyright Act’s four-factor test for copyright fair use. This updated study systematically analyzes 579 copyright fair use opinions from 435 cases over the 42-year period from 1978 through 2019. The updated data show that, for better and worse, much has remained the same in our fair use case law since 2005. Most notably, the fourth factor, going to the effect of the defendant’s use on the market for the plaintiff’s work, continues ultimately to dominate the test. However, the data record a significant shift toward summary adjudication of the fair use defense, a decline in the influence of the courts of the Second Circuit, and a substantial recovery in recent years in courts’ attention to the transformativeness test for fair use.

 

Introduction

Section 107 of the Copyright Act of 1976 sets forth four factors that courts “shall” consider in determining whether an unauthorized use of a copyrighted work qualifies as a non-infringing “fair use.”1 The four-factor test is central to the operation of American copyright law and the system of commercial and artistic speech that it regulates. Since the January 1, 1978 effective date of the Copyright Act, the test has been the subject of four Supreme Court cases2 and perhaps a fifth one this term in Google LLC v. Oracle America, Inc.3 Over the decades, the four-factor test has also featured in hundreds of lower court cases, incited an enormous amount of scholarly commentary,4 and received widespread international attention from nations considering, and then usually declining to adopt, the American four-factor approach to copyright fair use.5

In light of the continuing importance of and attention to the four-factor test, this article presents an update through 2019 of my previous quantitative study of all reported federal court opinions that made substantial use of the section 107 four-factor test from 1978 through 2005.6 That previous study analyzed 306 opinions from 215 cases over that 28-year period.7 This updated study analyzes 579 opinions from 435 cases over what is now a 42-year period from 1978 through 2019. It applies the same method of “systematic content analysis” of the fair use opinions that I pursued in the previous study.8 In doing so, it seeks to assess the daily life of our fair use case law outside of the headline-making, blockbuster cases that tend to populate our casebooks and serve as the usual suspects in our fair use scholarship. As the original study showed and this update will affirm, the “leading cases” do not necessarily reflect or influence everyday fair use adjudication, particularly in the district courts.9

The updated data show that much has remained the same in our fair use case law since 2005. The same two factors that drove the test through 2005 have continued to do so: the first factor, going to the “purpose and character” of the defendant’s use,10 including whether it qualifies as “transformative,”11 and the fourth factor, going to “the effect of the [defendant’s] use upon the potential market for or value of the copyrighted work.”12 As between these two factors, the updated data indicate that at least in the general population of our fair use case law, factor four has remained the single dominant factor in courts’ adjudication of the fair use defense—economic analysis continues ultimately to define fair use in the American copyright system.13 Meanwhile, appeal, reversal, and dissent rates remain unexceptional as compared to other areas of private law.14 Courts continue to apply the four-factor test mechanically and rarely consider additional factors beyond the four that are prescribed by the statute.15 The lower courts continue mistakenly to cite old, obsolete dicta from the Supreme Court case law, dicta that the Court itself has repeatedly tried to overwrite.16 Admirably, however, judges continue to resist stampeding all the factor outcomes to conform with the overall test outcome; instead, they generally continue to admit in their written opinions when a factor does not support their overall fair use finding.17

But the updated data also show some notable developments in the case law since 2005. A substantially higher proportion of our fair use case law now takes the form of motion to dismiss and summary judgment opinions.18 There has also been a significant rise in the proportion of opinions addressing unauthorized uses of photographs.19 Among the lower courts, the courts of the Second Circuit still exert the most influence on the fair use case law, but that influence is declining.20 More substantively, while the transformativeness test appeared to be waning in influence by 2005, it has since recovered its previous level of influence, even in the lower-profile, workaday fair use opinions that make up the majority of the data.21 However, the data indicate that in these same opinions, while courts now routinely consider transformativeness, a finding that the defendant’s use is transformative is still far from necessary to trigger a finding of fair use.

This update assumes some familiarity with and roughly follows the structure of the original study. Part I provides background on section 107 and the general characteristics of the opinions, including the proportion that found fair use, and of the judges who wrote them. Part II reports how the factor outcomes correlated with each other and the overall test outcome, to which factors judges devoted the bulk of their analysis, and the degree to which judges stampeded the factor outcomes. Part III focuses more specifically on the factual findings judges made under each of the four factors.

I. Background

A. Section 107 of the Copyright Act of 1976

My original study briefly addressed the history of the fair use doctrine and section 107.22 Here, for reference purposes, I quote section 107 in full:

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.23

Though section 107 consists of three parts (the preamble, the four factors, and the final statement added in 1992 regarding unpublished works), nearly all courts and commentators continue to focus only on the four factors. Over the 42-year period from 1978 through 2019, only 21.6% of the opinions cited the preamble to justify their fair use determination, and that proportion has remained essentially unchanged throughout the period.24 Meanwhile, since 1992, only eight opinions have explicitly alluded to the added statement regarding unpublished works as informing their fair use analysis.25 More significantly, despite courts’ frequent recitation of the principle that the factors “are not meant to be exclusive,”26 courts appear to have grown even less likely to consider factors beyond the four statutory factors. The original study reported that 17.0% of the 306 opinions through 2005 considered other factors;27 since 2006, that proportion has dropped to 4.7%.28 The original drafters of section 107 were concerned that the section would freeze what had up to then been an evolving judge-made doctrine.29 The data show this concern to have been valid.

As for the manner in which courts have used the factors, courts continue to apply them “mechanistically,”30 perhaps even more so than they did in the past. The original study reported that 59.5% of the 306 opinions through 2005 adopted the practice of explicitly stating which party, if either, each factor favored;31 since 2006, 77.0% of fair use opinions have done so. Overall, of the 579 opinions in the updated dataset, 87.5% considered all four factors, and 85.9% did so in order. (As in the past, this all greatly facilitated reliable coding of the opinions.)

B. Distributions of the Opinions

1. Distribution of Opinions by Year and Posture

Figure 1 shows the distribution by year of the 433 district court opinions, 139 circuit court opinions, and seven Supreme Court opinions in the updated dataset.32 The annual number of circuit court fair use opinions has remained relatively steady since the 1980s. By contrast, since 2010, there has been an uptick in the annual number of district court opinions employing the four-factor test.33 This uptick may partly reflect a significant increase in fair use opinions involving defendants’ unauthorized appropriation of photographs, especially in the internet context. Of the 25 district court opinions in 2018, 18 (or 72.0%) involved the copying of photographs; of the 27 district court opinions in 2019, 14 (or 51.9%) did so.34

Figure 1
Number of Reported Supreme Court, Appellate Court, and

District Court Fair Use Opinions by Year, 1978–2019

A second explanation for the increase in district court opinions since 2010 may emerge out of the data represented in Table 1 and Figure 2. Table 1 reports the distribution of opinions by posture for the full 42-year period from 1978 through 2019.35 Figure 2 shows the 5-year moving average of the annual number of fair use opinions by posture (for the leading postures) over the same period. No doubt due at least in part to the influence of Bell Atlantic Corp. v. Twombly36 and Ashcroft v. Iqbal,37 the past decade has seen a significant rise in the number of opinions addressing motions to dismiss based on the fair use defense.38 Of the 53 motion to dismiss opinions appearing in the dataset, only three predate Twombly.39

There has also been a significant rise in opinions addressing motions for summary judgment on the fair use defense, suggesting that litigants and courts have become more comfortable in the past decade with addressing the defense in the summary judgment posture.40 In some instances, this has led to individual cases producing multiple district court opinions on the fair use issue in relatively quick succession.41 For the period 1978 through 2019, 58.7% of our fair use case law consisted of opinions addressing motions for summary judgment on the fair use issue; more specifically, 52.5% of fair use circuit court opinions did so.

Table 1
Distribution and Fair Use Wine Rate by Posture of All Fair Use
Opinions, 1978–2019

Figure 2
5-Year Moving Average of Annual Number of Motion to Dismiss, Preliminary
Injunction, Summary Judgment, and Bench Trial Fair Use Opinions, 1978–2019

2. Reversal, Dissent, and Appeal Rates

The original study observed that the conventional wisdom at the time viewed fair use as an extraordinarily unpredictable and unstable area of copyright law.42 Yet the data up to 2005 showed that appeal, reversal, and dissent rates in the fair use case law were entirely unexceptional in comparison to other areas of private law.43 Since that time, a body of important scholarship has emerged arguing that copyright fair use is in fact a reasonably predictable and stable area of American law.44

The updated data further support this revised and reassuring view of the fair use case law. If anything, the statistics have trended toward more stability. From 1978 through 2005, the percentage of circuit court majority fair use opinions that reversed the district court was 33.8%;45 from 2006 through 2019, the percentage that did so declined to 27.3%.46 The overall reversal rate for the full 42-year period was 31.3%. As for dissent rates, the original study reported that from 1978 through 2005 the percentage of circuit court fair use majority opinions producing a dissent was 14.1%;47 for the period 2006 through 2019, it declined to 6.8%.48 The overall dissent rate from 1978 through 2019 was 11.3%. None of these statistics are out of the ordinary for federal civil cases.49

3. Distribution of Opinions by Venue

Table 2 details the distribution of the 572 non-Supreme Court opinions in the dataset by venue. The courts of the Second Circuit continue to dominate our fair use case law, but not as much as they once did. Figure 3 shows, for all district and circuit court fair use opinions from 1978 through 2019, the 5-year moving average of the annual proportion of such opinions that originated in the district courts or the circuit court of either the Second Circuit or the Ninth Circuit. Cases involving high-technology and the Internet do not alone explain the rise in the proportion of opinions originating in the Ninth Circuit; the proportion of such opinions coming out of the Ninth Circuit courts is not substantially different from that of the Second Circuit courts. Meanwhile, the Southern District of New York (S.D.N.Y.) still dominates the district court case law, but again, not as much as it once did. On its own, the S.D.N.Y. contributed more opinions to the overall dataset than all the district courts of the Ninth Circuit combined. But the original study reported that up to 2005, the S.D.N.Y. accounted for 31.3% of fair use district court opinions;50 since 2006, that proportion has declined to 26.6%.

Perhaps more interesting—and telling—are the new data on the influence that Second and Ninth Circuit appellate court opinions exert on opinions outside of their respective circuits. As in the original study, both appellate courts enjoy significant trade surpluses in the export and import of opinion citations among the district and appellate courts of the various circuits in the fair use case law. Overall, for the period 1978 through 2019, individual district and circuit court opinions outside of the Second Circuit cited to an average of 1.51 circuit court opinions from the Second Circuit, while opinions outside of the Ninth Circuit cited to an average of 0.92 circuit court opinions from the Ninth Circuit. No other appellate courts come close to these levels of extracircuit influence. As in the original study, by this measure, the S.D.N.Y. on its own exerts more extracircuit influence than every other circuit other than the Second and Ninth, with an average of 0.56 S.D.N.Y. opinions cited by district and appellate courts outside the Second Circuit for the period from 1978 through 2019. But trends over time for these statistics show that the Ninth Circuit has been gaining influence.51

Thus, as with the original data, our fair use case law continues to consist essentially of what are now the 207 opinions of the Supreme Court, the Second Circuit, the Ninth Circuit, and the S.D.N.Y., along with the remaining 372 opinions that are largely their progeny in the other federal courts.

Table 2
Distribution by Venue of Circuit & District Court Fair Use Opinions,

1978–2019*

Figure 3
5-Year Moving Average of Annual Proportion of District and Circuit Court
Fair Use Opinions Coming from the District and Circuit Courts of the Second Circuit or the Ninth Circuit, 1978–2019

4. Fair Use Win Rates

As reported above in Table 1, for the full period 1978 through 2019, 40.2% of the opinions sampled (regardless of posture and disposition on appeal) found fair use. Because 99 of the 435 cases included in the dataset were responsible for more than one opinion, a better statistic may be the fair use win rate for the last-in-time, non-concurrence, non-dissent opinion in each particular case, which I will refer to as a case’s final opinion. Of these 435 final opinions, 38.2% found fair use. Table 3 below details the distribution and fair use win rates of these opinions by posture.

Table 3
Distribution and Fair Use Win Rate by Posture of Final, Non-

Concurrence, Non-Dissent Fair Use Opinions in 435 Fair Use
Cases, 1978–2019

As they did in the original study, the fair use win rates for preliminary injunction and bench trial opinions in Table 3 fall significantly below 50%, with, for example, only one in four final bench trial opinions in the dataset finding fair use. I had suggested in the original study that the low fair use win rates in these postures may be the result of cases in which the defendant pled a relatively weak or even frivolous fair use defense because it is inexpensive to do so.52 Good judges will dutifully work through the four-factor analysis even for the most ridiculous claims of fair use, and so such opinions are for better and worse included in the dataset.

The original study introduced one viable (though hardly flawless) method of filtering out opinions addressing a frivolous fair use defense, which is to filter out those opinions that devoted a small proportion of their overall word count to the fair use analysis.53 For the 435 final opinions in the updated dataset, there is a moderate positive correlation between the proportion of the opinion devoted to the fair use analysis and the likelihood that the opinion will find fair use (r=0.274, N=435, p<.001). This should not be surprising; all else equal, accepting a fair use defense tends to require more analysis than rejecting it.54 But filtering for final opinions in the updated dataset that devoted at least 10% of their word count to the fair use analysis, which is comparable to what the original study did as part of its examination of fair use win rates, still yields fair use win rates significantly below 50% for preliminary injunction and bench trial opinions.55 To reach a roughly 50% fair use win rate requires filtering only for opinions that devoted at least 40% of their word count to the fair use analysis.56 This is all to say that even more so than in the original dataset, the updated dataset shows that a significant majority of our fair use case law—and precedent—consists of analyses that found no fair use.

Overall, the updated data show no statistically significant variations among the circuits in fair use win rates. To be sure, taking all 572 non-Supreme Court opinions in the data set regardless of posture or fate on appeal, the 178 opinions of the courts of the Second Circuit yielded a statistically-significantly higher fair use win rate (47.8%) than did the 394 opinions of courts of the other circuits (36.6%). (X2(1, N=572)=6.412, p=0.011). The courts of no other circuit were significantly different when analyzed in this way. But when we focus only on final opinions, any significant differences among the circuits disappear. For example, of the 431 non-Supreme Court final opinions in the dataset, the 124 opinions coming from the courts of the Second Circuit yielded a fair use win rate of 41.9%, while the 307 opinions from outside of the Second Circuit yielded a fair use win rate of 36.5%. (X2(1, N=431)=1.114, p=0.291). Further focusing only on final opinions since 2006 (i.e., the new data) does not produce significant differences. There was some evidence in the data from 1978 through 2005 that the case law of the courts of the Second Circuit was relatively fair use friendly.57 Whatever intercircuit differences there may have been, however, appear to have faded.

Finally, as for fair use win rates over time, a variety of different approaches to parsing the data produce no clear trends over time. Results greatly depend on how the opinions are filtered. For example, Figure 4 reports five-year moving averages of the proportion of opinions by year that found fair use for all 579 opinions in the dataset and for the subset of the 314 final opinions in the dataset that dispositively found fair use or no fair use (rather than outstanding fact issues). The former trend is basically flat. The latter suggests some uptick in fair use outcomes since 2000 (starting notably before the Supreme Court case Campbell v. Acuff-Rose in 1994). But overall, the data are inconclusive.

Figure 4
5-Year Moving Average of Annual Proportion of Opinions

Finding Fair Use

C. Judge-Specific Characteristics and Fair Use Outcomes

In a brief 2008 follow up to the original study, and largely inspired by the work of Matthew Sag, Tonja Jacobi, and Maxim Stych,58 I tested whether there was any relation between individual judges’ ideological or partisan leanings and their propensity to find fair use or no fair use.59 To do so, I used a variety of widely-used proxies for and measures of judicial ideology, including most notably the judge’s Judicial Common Space (JCS) score.60 Regardless of which measure I used, I found no relation between judges’ partisan ideology and their adjudication of the fair use defense in the original dataset of fair use opinions from 1978 through 2005.61

For the updated data set through 2019, I tested a number of judge-specific variables, including the judge’s JCS score, the political party of the judge’s appointing president, the judge’s age at the time of the opinion, the judge’s race, and the judge’s gender.62 I again found no significant relation between a judge’s partisan ideology as measured by the JCS or the judge’s appointing president and a judge’s adjudication of the fair use defense. Age and race also yielded no significant difference. But interestingly, a simple two-by-three, gender-by-outcome comparison of the updated data suggests that male judges were significantly more likely to find fair use than female judges. Of the 572 non-per curiam opinions in the updated dataset, male judges accounted for 448 opinions, of which 43.1% found fair use and 43.3% found no fair use (the remaining 13.6% found fact issues or produced mixed rulings). By contrast, female judges accounted for 124 opinions, of which only 28.2% found fair use and 51.6% found no fair use. (X2(2, N=572)=9.435, p=0.008). It is not at all clear why the data would produce this difference. I consider gender further below.

II. Interfactor Analysis

A. Correlation Analysis

Among the goals of the original study was to establish quantitatively the degree to which the outcomes of each of the factors in the four-factor fair use test correlated with each other and with the overall outcome of the test.63 One means of doing so involves pairwise correlation analysis among the factor outcomes and the overall test outcome. Table 4 sets forth these pairwise correlation results.

Table 4
Pairwise Correlations Among Overall Fair Use Findings and Factor
Outcomes in 579 Fair Use Opinions, 1978–2019***
(using ternary variables: 1=favors FU, -1=disfavors FU, 0=other)

The results in Table 4 are very much in line with those reported in the original study. As expected, the outcome of factor four (“the effect of the use upon the potential market for or value of the copyrighted work”) correlates most strongly with the overall test outcome and the other factor outcomes, followed by the outcomes of factors one and three (respectively, “the purpose and character of the use” and “the amount and substantiality of the portion used in relation to the copyrighted work as a whole”). Though the correlation between the outcome of factor two (“the nature of the copyrighted work”) and the overall test outcome is statistically significant, it remains, as it always has been, relatively weak. These basic correlation coefficients have not substantially changed over time.

The outcomes of factors one and four are also strongly correlated with each other. Indeed, in 72.5% of the 579 opinions, the court arrived at exactly the same outcome under both factors (out of six possible coded outcomes per factor).64 As the original study explained, this result should not be surprising.65 A finding under factor one that the defendant’s use is transformative or non-commercial will often support a finding under factor four that the use has no appreciable effect on the potential market for the plaintiff’s work. Of greater interest may be those relatively rare instances in which factors one and four pointed in opposite directions. They did so in 44 opinions. The court’s overall fair use determination aligned with factor four in 50.0% of these opinions, aligned with factor one in 36.0% of the opinions, and otherwise found that the fair use determination was a fact issue. As in the original study, correlation analysis and simple percentages suggest, though admittedly weakly, that factor four continues to drive the overall test outcome. However, as I will discuss further below, other analyses of the data present a more complicated picture.

B. Word Count Analysis

The analysis of trends over time in the proportion in word count of courts’ fair use analysis that they devoted to each of the four factors may yield further insight into the degree to which courts have relied on certain of the factors in making (or at least defending in their opinions) their fair use determination.66 In assessing the word count data from 1978 through 2005, the original study suggested that courts exhibited two waves of attention to factor one.67 The first came after the 1984 Supreme Court case of Sony Corp. of America v. Universal City Studios, Inc., in which, as I discuss further below, the Court established the presumption under factors one and four that commercial uses are presumptively unfair.68 The second came after the 1994 Supreme Court case of Campbell v. Acuff-Rose Music, Inc., in which the Court fully embraced the concept of transformativeness under factor one.69 The original study further suggested that both of these waves were followed by troughs in which courts’ attention to the factor lessened, so that by 2005, attention to factor one appeared to have been steadily declining. However, in a subsequent article, Neil Netanel quantitatively studied courts’ use of the transformativeness doctrine from the 1994 Campbell case through 2010 and documented a recovery of interest by courts in the first factor—and specifically in the transformativeness inquiry—for the period 2006 through 2010.70

Figure 5 presents a five-year moving average of the proportion in word count of each opinion’s fair use analysis devoted to discussion of each of the four factors. It largely confirms Netanel’s argument that despite any apparent decline in attention to factor one from 2000 through 2005, courts renewed their interest in the factor in subsequent years. What is striking is that in the meantime, at least as measured by word count, courts’ attention to factor four has noticeably diminished. Whether this is evidence of a decline in the influence of factor four is a question I will return to below.

Figure 5
5-Year Moving Average of the Proportion in Word Count of Fair Use Analysis Devoted to Each of the Four Factors in 579 Fair Use Opinions, 1978–2019

C. Stampeding

The original study sought to dispel the conventional wisdom at the time that judges first decide whether the defendant’s conduct constituted fair use and then align the factor outcomes to support that result.71 To this end, the study established a stampede score for each opinion, which was simply the sum of the factors that the opinion found to favor fair use minus the sum of the factors that the opinion found to disfavor fair use.72 The study showed that courts do not in fact stampede the factors to conform with the overall test outcome.73 Whether finding fair use or no fair use, judges tended to call the factors as they saw them and openly admitted when a factor did not support their overall determination.74

The data continue to support this view of the case law. Figure 6 shows the mean stampede score by posture and fair use outcome in the 579 opinions in the updated dataset. None of the means for the period 2006 through 2019 are significantly different from the corresponding means for the period 1978 through 2005.

Figure 6
Mean Stampede Score by Posture in 579 Fair Use Opinions, 1978–2019
(Stampede score = (number of factors determined to favor fair use) – (number of factors determined to disfavor fair use))

III. Intrafactor Analysis

I now turn to the various subfactor considerations that courts have traditionally relied on to determine the outcomes of each of the four factors.

A. An Overview of the Significance of the Subfactor Findings

To help in understanding what role each of the various subfactor considerations play in the overall fair use analysis, Table 5 presents the results of various logistic regression models estimated for all opinions that were dispositive (i.e., either found fair use or no fair use), were unreversed, and in the case of appellate opinions were neither dissents nor concurrences. I will refer to such opinions, for lack of a better term, as core opinions.75 The models regress a binary dependent variable representing whether the opinion found fair use on various binary independent variables representing underlying factual findings, whether the opinion came out of the courts of the Second or the Ninth Circuits, and, following Sag,76 whether the plaintiff or defendant was a natural person. Table 5 reports odds ratios, which are not particularly easy to understand. For each independent variable, the odds ratio reports the ratio of the odds of a finding of fair use to the odds of a finding of no fair use when the condition the independent variable represents is satisfied and all other variables are held constant77 Though the regression models are admittedly rough, the odds ratios can at the very least provide a sense of the relative impact of various factual findings and other variables on the overall fair use determination.

I will discuss the results reported in Table 5 in more detail below. I briefly note here with respect to the various objective variables that the regression results suggest that litigating a case in the courts of the Second Circuit or the Ninth Circuit as opposed to other venues does not significantly affect the likelihood of a fair use finding. Additionally, once we control for the factual characteristics of the cases, the impact of gender is only marginally statistically significant. Finally, consistent with Sag’s work, the fact that the owner of the copyrighted work is a natural person (rather than, for example, a corporation) appears to be associated with a significant increase in the likelihood that a court will find fair use.78

B. Factor One: Purpose and Character of the Use

We saw above that the outcome of factor one continues to correlate very strongly with the overall outcome of the fair use test. Expressed in percentage terms, of the 354 core opinions in the updated dataset, 153 found that factor one favored fair use and 141 of these (or 92.2%) found fair use; of the 174 core opinions that found that factor one disfavored fair use, 168 (or 96.6%) found no fair use. Indeed, the correlation between the outcome of factor one and the overall test outcome is so strong as to suggest that factor one is nearly dispositive of the overall test outcome. Alternatively, and perhaps more cynically, courts may feel the need to align the outcome of factor one with their overall fair use determination. Regardless, the importance of factor one prompts the question of which subfactor considerations drive the outcome of factor one itself. As in the original study, I focus here on the most important of these subfactors: whether the defendant’s use was transformative, whether the defendant’s use was commercial or noncommercial, and whether the defendant acted in good or bad faith.

Table 5
Logistic Regression of Fair Use Outcome on Subfactor Factual Findings
and Objective Variables in Core Fair Use Opinions, 1978–2019

1. Transformativeness

Since the 1994 Campbell case, the consideration of whether a defendant’s use qualifies as “transformative” has emerged as among the most important to a court’s overall fair use determination;79 indeed, Netanel argues that it “overwhelmingly drives” that determination.80 The original study suggested that by the end of 2005, courts’ attention to the transformativeness inquiry, as compared to factor one more generally, was waning.81 However, as the gray line in Figure 7 shows, the proportion of fair use opinions considering transformativeness has returned to very high levels.82

Other data further illustrate how big of a role a finding that the defendant’s use is transformative can play in the court’s overall fair use determination. The regression models suggest that when a court finds that the defendant’s use is transformative, the ratio of the odds a defendant will prevail in its fair use defense to the odds it will fail is anywhere from 86 to 91 times greater. By this measure, a finding of transformativeness exerts by far the greatest impact of any finding on a court’s likelihood of making an overall determination of fair use. More simply, as Table 6 indicates, of the 78 core opinions since Campbell in which a court found transformativeness, in all but three the court went on to find fair use. Tellingly, in each of the three outlying opinions, the court took pains to minimize the significance of its transformativeness finding.83

Figure 7
5-Year Moving Average of Proportion of Opinions Considering Various Factor One Subfactors in 579 Fair Use Opinions, 1978–2019
84

Table 6
Transformativeness Finding by Overall Fair Use Outcome in 246 Core Opinions, from Campbell v. Acuff-Rose Through 2019

Yet it remains important to emphasize that while a finding of transformativeness is nearly sufficient to trigger an overall finding of fair use, it is far from necessary to trigger such a finding. The data indicate that the transformativeness inquiry has not in fact replaced the overall fair use analysis.85 Consistent with Rebecca Tushnet’s defense of non-transformative copying, even non-transformative copying of entire works will sometimes qualify as fair use.86 As Table 6 shows, since Campbell 118 core opinions have found fair use. Of these, 32 (27.1%) did not consider transformativeness and six (5.1%) explicitly found that the defendant’s use was not transformative.87 Overall, only 75 (63.6%) of the 118 core opinions that found fair use explicitly found the defendant’s fair use to be transformative in nature. Of the remaining 43 opinions that did not find transformativeness but did find fair use, 19 involved facts in which the defendant copied the entirety of the plaintiff’s work.88

The data further show that a finding of transformativeness continues to have a mixed effect on a court’s treatment of other factors in the fair use analysis. Consistent with the original study, a finding of transformativeness still does not appear to stampede the factors. The mean stampede score for the 78 core opinions since Campbell that found transformativeness was 1.99, with the mode stampede score of 2 reported by 38.5% of the opinions. As the original study explained, this makes sense because transformative uses most commonly target works classified as creative under factor two and often involve quite substantial uses of those works under factor three.89 At the same time, however, a finding of transformativeness correlates very strongly with a finding that the defendant’s use will have no substantial adverse effect on the potential market for the plaintiff’s work under factor four. In the 78 core opinions since Campbell that found that the defendant’s use was transformative, only six (7.7%) found that the defendant’s effect on the plaintiff’s mark disfavored fair use under factor four.90

A subset of transformativeness cases consists of cases in which the defendant made a parody of the plaintiff’s work. In nearly all of these, the defendant’s conduct was deemed to be transformative and a fair use. Overall, the updated dataset consists of 35 opinions from 26 cases in which the court explicitly found that the defendant’s work constituted a parody. In all but three of these cases, the defendant’s use was ultimately found to be a fair use,91 and the last time a court found a parody not to be a fair use was 1988, six years before Campbell.92 Of the 23 cases in which the courts ultimately found the parody to be a fair use, 20 involved parodies deemed to be commercial in nature. Even more so than generally transformative works, the species of such works that qualify as parodic are especially privileged under factor one and the overall four-factor fair use analysis.

2. Commerciality

Though transformativeness continues to attract the bulk of scholarly attention, the commerciality of the defendant’s use remains the subfactor that courts most consistently invoke in their factor one analysis, as the black line in Figure 7 shows. Of the 545 fair use opinions following the 1984 Supreme Court majority opinion in Sony, which put substantial weight on the commerciality or noncommerciality of the defendant’s use,93 85.3% have explicitly addressed the commerciality subfactor. The regression results further show the substantial effect that the commercial or noncommercial status of the defendant’s use has on the overall outcome of the test, though the effect is not nearly on the order of that exerted by a finding of transformativeness. While transformative uses nearly always qualify as fair uses, noncommercial uses may fail to so qualify. Of the 47 core opinions since Sony that found that the defendant’s use was noncommercial in nature, a respectable minority of 11 opinions (23.4%) found overall that it was nevertheless not a fair use. Meanwhile, commercial uses may occasionally qualify as fair uses. Of the 233 core opinions that found that the defendant’s use was commercial in nature, 36.9% found the use to be a fair use.

Though there is thus some flexibility in the relation between the commerciality or noncommerciality of the use and a court’s determination of whether it qualifies as a fair use, there remains a continuing problem in the case law under the factor one commerciality analysis, which is shown by the dashed line in Figure 7. As the original study detailed, the Sony majority opinion in 1984 quite explicitly established a presumption under factor one that a use “to make copies for a commercial or profit-making purpose . . . would presumptively be unfair,” while the “contrary presumption” would apply to noncommercial uses.94 Leading authorities have long condemned this formulation, with David Nimmer flatly stating in his treatise: “Such a categorical rule is unwarranted.”95 Clearly sensing that it had gone too far, the Supreme Court sought in subsequent cases to undo the Sony presumption, but it could never apparently bring itself simply to say that the Sony presumption was a mistake. Instead, even in Campbell, where the Court went to great lengths to criticize the “Court of Appeals’s elevation of one sentence from Sony to a per se rule,”96 the Court implied that the Sony treatment of commerciality was still essentially valid and that the Sixth Circuit had just misinterpreted it.97 The result is that, as the original study showed, lower courts continued to cite to the Sony commerciality presumption under factor one without recognizing Campbell’s intervention, though, by 2005, the presumption appeared to finally be on its last legs.98 Yet, Figure 7 shows that courts have since revived its force, with 37 opinions since Campbell (and 26 opinions since 2005) ignoring Campbell’s intervention and citing the original, unqualified Sony factor one commerciality presumption.99 This all provides a good example of how the leading cases do not always reflect, let alone influence the everyday state of our fair use case law, especially when those leading cases decline to make clear when they are abrogating previous precedent.100

3. Other subfactors

The original study concluded that despite courts’ routine invocation of the equitable nature of the fair use defense, findings of good and bad faith did not have an appreciable effect on the overall outcome of courts’ fair use analysis.101 The regression results reported in Table 5 suggest that this remains true for the updated dataset. To be sure, as in the original study, a finding that the defendant acted in bad faith still correlates strongly with a determination of fair use; of the 20 opinions in the updated dataset that found bad faith, only four found fair use overall.102 But 84.5% of the 579 opinions never bothered to address whether the defendant had acted in bad faith.

C. Factor Two: Nature of the Copyrighted Work

Fair use opinions continue routinely to denigrate factor two as unimportant to the overall fair use analysis,103 and the updated data support the view that the factor typically has a relatively minimal impact on the outcome of the overall four-factor test. In general, courts rarely find that the factor supports a finding of fair use.104 Nevertheless, the data suggest that certain findings under both of factor two’s subfactors—whether the work is creative or factual and whether the work is published or unpublished—continue to have an at least statistically significant effect on a court’s overall determination. The regression results reported in Table 5 indicate, as expected, that the odds a court will find fair use increase when the plaintiff’s work is factual in nature and decrease when it is creative in nature. More interestingly, and consistent with the original study and the irony it noted,105 while the unpublished nature of the plaintiff’s work has no apparent impact on a court’s overall fair use determination, the fact that a work is published appears to increase the odds that a court will find fair use. Specifically, the ratio of the odds of a finding of fair use to a finding of no fair use more than doubles when the court finds that the plaintiff’s work was published.106

D. Factor Three: Amount and Substantiality of the Use

Overall, the status of factor three has not changed appreciably since the original study.107 Its outcome continues to correlate very strongly with the overall test outcome, particularly when factor three is found to favor the defendant. In the 101 core opinions in which the court found that factor three favored a finding of fair use, all but two found in favor of fair use overall.108 However, in contrast to the findings of the original study, Table 5 suggests that the fact that the defendant copied the entirety of the plaintiff’s work no longer appears to significantly impact a court’s overall fair use determination. In the updated data, of the 148 core opinions in which the court found that the defendant’s use copied the entirety of the plaintiff’s work, a reasonably balanced 58.8% found no fair use and 41.2% found fair use.109

One trend that has become clearer with the updated data is the increasing degree to which courts explicitly assess under factor three whether the defendant’s use of the plaintiff’s work was necessary to the defendant’s purpose, and even if it was, whether the defendant used more than was necessary to accomplish that purpose.110 Figure 8 estimates a five-year moving average of the proportion of opinions by filing year approaching factor three in this manner. Though this gradual shift toward a necessity requirement could be understood to benefit plaintiffs, the data show no strong correlation between a court’s invocation of necessity and either the outcome of factor three or the overall outcome of the fair use test.

Figure 8
5-Year Moving Average of Proportion of Opinions Assessing Necessity
Under Factor Three, 1978–2019

E. Factor Four: Effect on the Market

The scholarly literature continues to debate whether the leading factor in courts’ fair use analysis is currently factor one, with its focus on the “purpose and character” of the defendant’s use, or factor four, with its focus on the effect of the defendant’s use on the “potential market for or value of” the plaintiff’s work. Looming in the background of this debate is the larger question of whether courts are generally pursuing a more defendant-friendly, purpose-centered approach to fair use or a more plaintiff-friendly, economic approach to fair use. Most scholars endorse the former approach, believing it to be more supportive of the public domain. This debate has largely relied on the latest headline-making cases. Prominent cases such as Blanch v. Koons,111 Cariou v. Prince,112 and Authors Guild v. Google, Inc.113 may serve as evidence that factor one and transformativeness have emerged as dominant. Other cases such as Kienitz v. Sconnie Nation LLC114 and Fox News Network, LLC v. TV Eyes, Inc.115 may indicate that factor four is dominant.

The updated data suggest that with respect to our fair use case law as a whole, beyond the headline cases, the status of factor four has not appreciably changed over the past three or four decades.116 It continues to be the factor whose outcome correlates most strongly with the overall test outcome, as Table 4 indicates. Of the 169 core opinions that found that factor four disfavored fair use from 1978 through 2019, all but three ultimately found no fair use, and none of the three outlying opinions that found fair use offers particularly compelling analysis to explain its divergence between factor four and the overall outcome.117 Meanwhile, of the 154 opinions that found that factor four favored fair use, all but nine found fair use. A majority of these nine outlying opinions ruled that factor four favored the defendant because there was no market for the plaintiff’s work.118

As with factor one, the very strong correlation between the outcome of factor four and the overall test outcome prompts the question of which subfactor factual findings drive the outcome of factor four itself. The problem with factor four, however, is that historically courts have not developed any subfactor factual findings under it. Instead, as I argued in the original study, courts typically treat factor four as essentially a “metafactor” in which they integrate their analyses of the preceding three factors.119 In doing so, they balance the justification for the defendant’s use of a work against its effect on the plaintiff’s economic incentives to create and further exploit that work. Crucially, however, when courts engage in this balancing test in the analytical space provided by factor four, courts do so in economic terms, within the wheelhouse of law and economic analysis. By contrast, had the four-factor test been designed so that the analysis of the justification for the defendant’s use came fourth, in the cleanup position, one imagines that outcomes might be different in close cases.

Recent scholarship has brought to light one important, more specific function that factor four plays—or should play.120 In instructing courts to assess “the effect of the use upon the potential market for or value of the copyrighted work,” factor four requires courts to define the limits of the “potential market” and “value” that the copyright owner should have the exclusive right to exploit. Thus, for example, courts have established under factor four that the owner should not have the exclusive right to exploit the market for harsh reviews of its work or for parodies that ridicule the work.121 But here too factor four plays a largely synthetic role. Under it, the first three factors aid the court in determining whether the defendant’s use falls within the category of uses that should be reserved exclusively to the copyright owner as a matter of copyright policy or simply of industry custom.122

What is remarkable is how little guidance the doctrine of factor four itself provides to this market-definition inquiry.123 At best, a few doctrinal memes guide the courts. The first is that, as Harper & Row declared, factor four is “undoubtedly the single most important element of fair use.”124 Though Campbell sought to override this dictum by emphasizing that courts should consider all the factors,125 the proposition itself remains oft-cited, as Figure 9 shows. Courts’ citation to the principle does not correlate one way or the other with the outcome of factor four or the overall outcome of the fair use test. But its slight resurgence in recent case law may indicate that in the overall benefit-cost, access-versus-incentives fair use analysis, many courts have internalized the instruction that the costs to the plaintiff’s incentives are ultimately to be weighted more heavily than the benefits to the defendant in terms of access.

Figure 9
5-Year Moving Average of Proportion of Opinions Referencing Various Factor Four Subfactors in 579 Fair Use Opinions, 1978–2019

The second doctrinal meme is the slippery slope principle first established by Sony and reinforced by Campbell that courts should “consider not only the extent of market harm caused by the particular actions of the alleged infringer, but also ‘whether unrestricted and widespread conduct of the sort engaged in by the defendant . . . would result in a substantially adverse impact on the potential market’ for the original.”126 This proposition has gradually gained ground over time. By its terms, it is plaintiff-friendly, though as with the “most important element” proposition, there is no correlation between its citation and the outcome of factor four or the overall test.

Third and finally, as shown in Figure 9, courts continue to cite the presumption established by Sony under factor four that commercial uses presumptively harm the plaintiff’s market. The Sony court had stated:

What is necessary [under factor four] is a showing by a preponderance of the evidence that some meaningful likelihood of future harm [to the market for the plaintiff’s work] exists. If the intended use is for commercial gain, that likelihood may be presumed. But if it is for a noncommercial purpose, the likelihood must be demonstrated.127

As it did with Sony’s factor one commerciality presumption, the Campbell court sought to defuse Sony’s factor four commerciality presumption, this time by stating that the presumption did not apply to any “case involving something beyond mere duplication for commercial purposes.”128 But just as the original study noted for opinions through 2005,129 so now, lower courts continue to ignore Campbell’s limitation. Since Campbell, 54 opinions have cited to Sony’s factor four commerciality presumption, 30 of them without recognizing Campbell’s limitation. All but three of these opinions ruled against the defendant and found no fair use.130

Conclusion

In the spirit of the original study, this brief update has focused on the mass of everyday copyright fair use cases. On their own, most individual fair use cases may not seem to be particularly important except to the parties involved, if even to them. However, taken together, they form a body of case law whose patterns and characteristics reveal the daily life of our fair use case law, a daily life that is in some ways surprisingly different from the life of the “leading cases.” The lower courts sometimes ignore higher court cases or are simply unaware of them. Old, ill-advised dicta can keep cropping up like perennial weeds even decades after efforts to eradicate them. But perhaps most surprisingly, when viewed as a whole, the fair use case law presents itself as far more stable and predictable—and unchanging—than the headline-making cases might suggest.


Footnotes

*John M. Desmarais Professor of Intellectual Property Law, NYU School of Law. This long-running project has benefitted from presentations to the Los Angeles Copyright Society in 2008, the Fordham Intellectual Property Law & Policy Conference in 2012, the University of Washington School of Law Campbell at 21 Conference in 2015, New York City Bar Association Copyright Law Committee in 2015, and the Intellectual Property Scholars Conference in 2018. Thanks to Emiliano Catan, Jeanne Fromer, and Matthew Sag for comments. Thanks to Lauren Bobersky and Matthew Sumner for excellent research assistance.

  1. Copyright Act of 1976, Pub. L. No. 94-553, sec. 101, § 107, 90 Stat. 2541, 2546 (codified as amended at 17 U.S.C. § 107 (2018)).
  2. Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994); Stewart v. Abend, 495 U.S. 207 (1990); Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539 (1985); Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417 (1984).
  3. Oracle Am., Inc. v. Google LLC, 886 F.3d 1179 (Fed. Cir. 2018), cert. granted, 140 S. Ct. 520 (2019).
  4. See 4 Melville B. Nimmer & David Nimmer, Nimmer on Copyright § 13.05[A][6] (Matthew Bender rev. ed. 2019) (“A wealth of scholarly theories have been advanced to bring order to the tangled brier that is fair use.”). Indeed, there are far more law review articles on copyright fair use than there are fair use court opinions. For the period 1978 through 2019, the overall ratio of law review articles on the issue to actual court opinions on the issue was 1.61—or, stated differently, over that period there were only 0.62 fair use court opinions for every law review article published on copyright fair use. Furthermore, these data very likely underestimate the number of law review articles on fair use during the period. They are based on the number of entries appearing each year in the Hein Online Law Journal Library that used the term “fair use” in their titles and addressed copyright fair use (rather than, for example, trademark fair use).
  5. See generally Jonathan Band & Jonathan Gerafi, The Fair Use/Fair Dealing Handbook (2013), infojustice.org/wp-content/uploads/2013/03/band-and-gerafi-
    2013.pdf
    [https://perma.cc/5XCL-BWJF] (providing a typology of fair use laws around the world); Niva Elkin-Koren & Neil Weinstock Netanel, Transplanting Fair Use Across the Globe: A Case Study Testing the Credibility of U.S. Opposition, 72 Hastings L.J. (forthcoming 2020) (reporting the results of an empirical study of Israeli fair use case law).
  6. Barton Beebe, An Empirical Study of U.S. Copyright Fair Use Opinions, 1978–2005, 156 U. Pa. L. Rev. 549 (2008).
  7. Id. at 564–65.
  8. See id. at 623–24 (describing the method of collecting and coding the opinions studied).
  9. It cannot be stressed strongly enough, however, that the statistics set forth in this update—and indeed in the original article—cannot present a complete picture of our fair use case law, let alone of the larger operation of copyright fair use outside of the case law. The data are a necessary supplement to, but they cannot substitute for, traditional systematic doctrinal analysis. At best, each statistic is an additional form of circumstantial evidence, a description of a different part of the elephant, that may be adduced to develop a better understanding of our overall fair use case law. See Neil Weinstock Netanel, Making Sense of Fair Use, 15 Lewis & Clark L. Rev. 715, 732–34 (2011) (discussing the problem of selection bias and defending the utility of systematic content analysis of reported federal opinions); Matthew Sag, Predicting Fair Use, 73 Ohio St. L.J. 47, 83 (2012) (“The potential for selection bias notwithstanding, the fact remains that litigated cases are important and they are constantly subject to ad hoc empirical assessments. . . . The unrepresentative nature of written opinions does not imply that scholars should abandon the field of empirical legal studies, but it does call for some caution in interpreting results.”); see also Beebe, supra note 6, at 565–66 (discussing selection bias).
  10. 17 U.S.C. § 107(1) (2018).
  11. See Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 578–85 (1994) (discussing the fair use concept of transformativeness).
  12. 17 U.S.C. § 107(4).
  13. See infra Part III.E.
  14. See infra Part I.B.2.
  15. See infra Part I.A.
  16. See infra Parts III.B.2, III.E.
  17. See infra Part II.C.
  18. See infra Part I.B.1.
  19. See infra Part I.B.1.
  20. See infra Part I.B.3.
  21. See infra Part III.B.1.
  22. See Beebe, supra note 6, at 557–61.
  23. 17 U.S.C. § 107.
  24. Interestingly, however, 3.3% of the opinions (or 19 of the 579 opinions) made an adverse inference that to the extent the defendant’s use did not take the form of “criticism, comment, news reporting, teaching . . . , scholarship, or research,” this supported a determination that the use did not qualify as a fair use. See, e.g., Peteski Prods. v. Rothman, 264 F. Supp. 3d 731, 739–40 (E.D. Tex. 2017) (“Here, [defendant] made no alteration to the work nor did she use it as part of a commentary or criticism.”).
  25. See, e.g., Swatch Grp. Mgmt. Servs. v. Bloomberg L.P., 742 F.3d 17, 31 (2d Cir. 2014), amended and superseded by 756 F.3d 73 (2d Cir.). Three of these eight opinions nevertheless found no fair use. See, e.g., Monge v. Maya Magazines, Inc., 688 F.3d 1164, 1168 (9th Cir. 2012).
  26. Harper & Row Publishers, Inc. v. Nation Enters., 471 U.S. 539, 560 (1985).
  27. Beebe, supra note 6, at 564.
  28. See also Pamela Samuelson, Unbundling Fair Uses, 77 Fordham L. Rev. 2537, 2620 (2009) (“It is curious . . . how reluctant courts have been to consider factors beyond those set forth in § 107 in the fair use caselaw.”).
  29. See Beebe, supra note 6, at 559.
  30. Harper & Row, 471 U.S. at 588 (Brennan, J., dissenting).
  31. Beebe, supra note 6, at 562.
  32. Of the 435 cases that produced these opinions, 336 produced only one opinion. By contrast, five cases produced five opinions each. For the final majority opinion in each of these cases, see Harper & Row, 471 U.S. 539; New Era Publications International, APS v. Henry Holt, Co., 884 F.2d 659 (2d Cir. 1989); Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994); Princeton University Press v. Michigan Document Services, Inc., 99 F.3d 1381 (6th Cir. 1996); Bouchat v. Baltimore Ravens Ltd. Partnership, 737 F.3d 932 (4th Cir. 2013). This study coded the United States Court of Federal Claims as a district court.
  33. The rapid increase in district court opinions has meant that for the first time since 1992, there were more federal court opinions on copyright fair use in both 2018 and 2019 than there were law review articles focused on the issue in each of those years. The ratio of court opinions to law review articles in 2018 was 1.11 and in 2019, 1.03.
  34. Viewed differently, of the 83 district court opinions in the overall dataset involving the unauthorized reproduction of photographs, 61 (or 73.5%) date from 2010. The Liebowitz Law Firm represented the plaintiff in three district court opinions involving photographs in 2018 and four involving photographs in 2019; the firm also represented a videographer in an additional 2019 opinion. The plaintiff prevailed on the fair use issue in five of these eight opinions. On the Liebowitz Law Firm, see Mike Masnick, Copyright Troll Richard Liebowitz Benchslapped and Sanctioned AGAIN in a Massive Filing Detailing Pages upon Pages of Him Lying Under Oath, Techdirt (June 30, 2020, 3:33 PM), https://www.techdirt.com/articles/20200626/18131744799/copyright-troll-richard-liebowitz-benchslapped-sanctioned-again-massive-filing-detailing-pages-upon-pages-him-lying-under-oath.shtml.
  35. For purposes of simplifying the data, this study coded as preliminary injunction opinions two temporary restraining order opinions that included complete four-factor fair use analyses. See United Tel. Co. v. Johnson Publ’g Co., 855 F.2d 604 (8th Cir. 1988); James Castle Collection & Archive, LP v. Scholastic, Inc., No. 1:17-CV-437-BLW, 2017 U.S. Dist. LEXIS 181801 (D. Idaho Oct. 30, 2017).
  36. 550 U.S. 544 (2007).
  37. 556 U.S. 662 (2009).
  38. William F. Patry, Patry on Fair Use § 7:5, Westlaw (database updated May 2020) (“Increasingly, courts have considered fair use on a rule 12(b)(6) motion to dismiss for failure to state a claim . . . .”).
  39. See Gregerson v. Vilana Fin., Inc., 446 F. Supp. 2d 1053 (D. Minn. 2006) (denying motion to dismiss); Chi. Sch. Reform Bd. of Trs. v. Substance, Inc., 79 F. Supp. 2d 919 (N.D. Ill. 2000) (granting motion to dismiss); Int-Elect Eng’g, Inc. v. Clinton Harley Corp., No. C-92-20718 JW, 1993 U.S. Dist. LEXIS 11510 (S.D. Cal. June 24, 1993) (same). With respect to post-Twombly motion to dismiss opinions in the data, less than half were in cases involving photographs.
  40. See Sony BMG Music Ent. v. Tenenbaum, 672 F. Supp. 2d 217, 221 (D. Mass. 2009) (“The doctrine of fair use is not infinitely malleable, requiring a jury determination every time it is intoned, no matter what the facts.”). See also Pierre N. Leval, Campbell as Fair Use Blueprint, 90 Wash. L. Rev. 597, 613 (2015) (“Fair use disputes should generally be amenable to disposition on the pleadings or on summary judgment.”).
  41. E.g., BWP Media USA, Inc. v. Gossip Cop Media, LLC, 87 F. Supp. 3d 499 (S.D.N.Y. 2015) (motion to dismiss granted based on the fair use defense with regard to one of plaintiff’s photographs and denied with regard to other plaintiff’s photographs); BWP Media USA, Inc. v. Gossip Cop Media, Inc., 196 F. Supp. 3d 395 (S.D.N.Y. 2016) (finding after bench trial that defendant’s use of certain plaintiff’s photographs did not qualify as fair use); Dr. Seuss Enters., L.P. v. ComicMix LLC, 256 F. Supp. 3d 1099 (S.D. Cal. 2017) (denying motion to dismiss on fair use defense); Dr. Seuss Enters., L.P. v. ComicMix LLC, 372 F. Supp. 3d 1101 (S.D. Cal. 2019) (granting summary judgment to defendant on fair use defense).
  42. See Beebe, supra note 6, at 574.
  43. See id. at 574–75.
  44. See, e.g., Netanel, supra note 9; Sag, supra note 9; Samuelson, supra note 28.
  45. See Beebe, supra note 6, at 574.
  46. This decline did not constitute a statistically significant difference. (X2(1, N=115)=0.539, p=0.463).
  47. See Beebe, supra note 6, at 574.
  48. Though apparently a substantial difference, the updated data indicate that the difference in dissent rate for the period from 1978 to 2005 and the dissent rate for the period from 2006 to 2019 did not constitute a statistically significant difference (X2(1, N=115)=1.912, p=0.167).
  49. See Beebe, supra note 6, at 574.
  50. See id. at 567.
  51. The original study reported that for the period from 1978 through 2005, district and appellate courts outside of the Second Circuit cited to an average of 1.55 appellate court opinions from the Second Circuit, while the equivalent statistic for citations outside of the Ninth Circuit to Ninth Circuit appellate court opinions was 0.68. See Beebe, supra note 6, at 568. For the period from 2006 through 2019, the statistic for the Second Circuit declined from 1.55 to 1.45 and the statistic for the Ninth Circuit increased from 0.68 to 1.19.
  52. See Beebe, supra note 6, at 575–80.
  53. See id. at 567–81.
  54. See id. at 580–81.
  55. Specifically, the fair use win rate was .368 for the 144 preliminary injunction and bench trial opinions whose fair use analysis consisted in word count of at least 10% of the opinion’s overall word count.
  56. Specifically, the fair use win rate was .450 for the 60 opinions whose fair use analysis consisted in word count of at least 40% of the opinion’s overall word count.
  57. See Beebe, supra note 6, at 594 n.142; Sag, supra note 9, at 81–82 (discussing intercircuit differences in fair use win rates).
  58. See Matthew Sag, Tonja Jacobi & Maxim Stych, Ideology and Exceptionalism in Intellectual Property: An Empirical Study, 97 Cal. L. Rev. 801 (2009).
  59. See Barton Beebe, Does Judicial Ideology Affect Copyright Fair Use Outcomes?: Evidence From the Fair Use Case Law, 31 Colum. J.L. & Arts 517 (2008). See also Lee Epstein et al., The Judicial Common Space, 23 J.L. Econ. & Org. 303 (2007).
  60. See id. at 520.
  61. See id. at 521.
  62. For the JCS scores of district court judges, I used Christina L. Boyd, 113th Congress Data Downloads (1937-2014 appointments), Federal District Court Judge Ideology Data (2015), http://cLboyd.net/ideology.html. For the JCS scores of appellate court judges and Supreme Court justices, I used Lee Epstein et al., The Judicial Common Space, Wash. Univ. St. Louis (last updated Dec. 15, 2019), http://epstein.wustl.edu/research/JCS115.02.zip (data updated through 2017). For other judge and justice characteristics, I used Gary Zuk et al., Attributes of U.S. Federal Judges Database, Songer Project, http://www.songerproject.org/attributes.html (last visited Oct. 23, 2020) (data updated through 2000). For recently appointed judges not covered in these datasets, I coded their characteristics (other than their JCS scores) by hand. The data were combined and analyzed in Stata, a common statistical analysis software program.
  63. See Beebe, supra note 6, at 582–85.
  64. The outcome of each factor was coded as one of the following: favored fair use, disfavored fair use, presented fact issues, was neutral, was not addressed, or was unclear.
  65. See Beebe, supra note 6, at 583.
  66. Because judges have continued to apply the four-factor test in a highly mechanical manner (and often use subheadings to label their analysis under each factor), determining the proportion of each opinion’s word count that the judge devoted to each of the four factors has remained a straightforward undertaking.
  67. See Beebe, supra note 6, at 587–88.
  68. See infra notes 94 and 127 and accompanying text.
  69. See Campbell v. Acuff-Rose Music, Inc, 510 U.S. 569, 578–85 (1994).
  70. See Netanel, supra note 9, at 736–38.
  71. See Beebe, supra note 6, at 589–90.
  72. See id. at 590.
  73. See id. at 590–91.
  74. See id.
  75. The subset of final opinions (N=435) and the subset of core opinions (N=354) intersect (they share 288 opinions in common) but are not identical. Both exclude concurrences, dissents, and reversed opinions. Core opinions further exclude opinions that found that the fair use determination raised issues of fact. By contrast, final opinions include opinions finding a fact issue, but are limited only to the final-in-time (non-concurrence, non-dissent, unreversed) opinion in the case.
  76. See Sag, supra note 9, at 74–78.
  77. For example, according to model 1 in Table 5 (which omits judge gender, circuit, and natural person status of each party), when a court finds that the defendant’s use is transformative, the odds of a finding of fair use are 91.3 times as large as the odds of a finding of no fair use. Expressed as predicted probabilities under model 1, the predicted probability that a court will find fair use when the use does not qualify as transformative is .42 (with all other variables set to 0) as against a predicted probability of .99 when the use qualifies as transformative.
  78. See Sag, supra note 9, at 74–78.
  79. See generally Jiarui Liu, An Empirical Study of Transformative Use in Copyright Law, 22 Stan. Tech. L. Rev. 163 (2019) (reviewing all federal court fair use opinions addressing transformativeness through January 1, 2017); R. Anthony Reese, Transformativeness and the Derivative Work Right, 31 Colum. J.L. & Arts 467, 485 (2008) (reviewing all federal appellate court opinions addressing fair use from Campbell through 2007).
  80. See Netanel, supra note 9, at 734 (“[C]ontrary to Beebe’s finding that the transformative use doctrine has had quite limited influence on fair use case law, the transformative use paradigm, as adopted in Campbell v. Acuff-Rose overwhelmingly drives fair use analysis in the courts today.”).
  81. See Beebe, supra note 6, at 603.
  82. This is consistent with Netanel’s findings. See Netanel, supra note 9, at 736–40.
  83. See Fox News Network, LLC v. TVEyes, Inc., 883 F.3d 169, 177–78 (2d Cir. 2018) (characterizing the use as “at least somewhat transformative”); Warner Bros. Ent. v. RDR Books, 575 F. Supp. 2d 513, 544 (S.D.N.Y. 2008) (characterizing the use as “not consistently transformative”); Castle Rock Ent. v. Carol Publ’g Grp., 955 F. Supp. 260, 272 (S.D.N.Y. 1997) (characterizing the use as transformative “only by a generous understanding of what it means for a work to be ‘transformative’”).
  84. The figure includes reversed opinions because we are analyzing whether courts consider these subfactors relevant to a fair use analysis.
  85. See Kim J. Landsman, Does Cariou v. Prince Represent the Apogee or Burn-Out of Transformativeness in Fair Use Jurisprudence? A Plea for A Neo-Traditional Approach, 24 Fordham Intell. Prop. Media & Ent. L.J. 321, 324 (2014) (expressing concern that the transformativeness test “has in practice often dominated or replaced” the four-factor test). Cf. Kienitz v. Sconnie Nation LLC, 766 F.3d 756, 758 (7th Cir. 2014) (“We’re skeptical of Cariou’s approach, because asking exclusively whether something is ‘transformative’ not only replaces the list in § 107 but also could override 17 U.S.C. § 106(2), which protects derivative works.”).
  86. See Rebecca Tushnet, Copy This Essay: How Fair Use Doctrine Harms Free Speech and How Copying Serves It, 114 Yale L.J. 535, 587 (2004) (“Courts should recognize that various kinds of copying, not united by some overall theory about creating new works, promote freedom of speech . . . . The point is not to denigrate fair use, but to recognize that many kinds of uses of copyrighted material may be justified, not just uses that put a critical spin on a prior work.”).
  87. See Swatch Grp. Mgmt. Servs. v. Bloomberg L.P., 742 F.3d 17 (2d Cir. 2014), amended and superseded by 756 F.3d 73 (2d Cir.); L.A. News Serv. v. CBS Broad., Inc., 305 F.3d 924 (9th Cir. 2002); Super Future Equities, Inc. v. Wells Fargo Bank Minn., 553 F. Supp. 2d 680 (N.D. Tex. 2008); S&L Vitamins, Inc. v. Austl. Gold, Inc., 521 F. Supp. 2d 188 (E.D.N.Y. 2007); Gulfstream Aerospace Corp. v Camp Sys. Int’l, 428 F. Supp. 2d 1369 (S.D. Ga. 2006); Religious Tech. Ctr. v. Netcom On-Line Commc’n Servs., No. C-95-20091 RMW, 1997 U.S. Dist. LEXIS 23572 (N.D. Cal. Jan. 3, 1997).
  88. See, e.g., Swatch Grp. Mgmt. Servs., 742 F.3d 17; Bill Graham Archives v. Dorling Kindersley Ltd., 448 F.3d 605 (2d Cir. 2006); Amsinck v. Columbia Pictures Indus., 862 F. Supp. 1044 (S.D.N.Y. 1994). See also Samuelson, supra note 28, at 2557 (discussing “iterative copying for orthogonal purposes”).
  89. See Beebe, supra note 6, at 606.
  90. In three of these opinions, the court eventually overrode its transformativeness finding to find no fair use overall. These three opinions are referenced supra in note 81. The four opinions that found fair use (and in which transformativeness thus trumped factor four) were Bouchat v. NFL Properties, LLC, 910 F. Supp. 2d 798 (D. Md. 2012); Fuentes v. Mega Media Holdings, Inc., No. 09-22979-CIV-MORENO/TORRES, 2011 U.S. Dist. LEXIS 70996 (S.D. Fla. June 9, 2011); Warren Publishing Co. v. Spurlock, 645 F. Supp. 2d 402 (E.D. Pa. 2009); Williamson v. Pearson Education, Inc., No. 00 Civ. 8240 (AGS), 2001 U.S. Dist. LEXIS 17062 (S.D.N.Y. Oct. 19, 2001).
  91. The three outliers are New Line Cinema Corp. v. Bertlesman Music Group, Inc., 693 F. Supp. 1517 (S.D.N.Y. 1988); DC Comics, Inc. v. Unlimited Monkey Business, Inc., 598 F. Supp. 110 (N.D. Ga. 1984); and Dr. Pepper Co. v. Sambo’s Restaurants, Inc., 517 F. Supp. 1202 (N.D. Tex. 1981).
  92. See New Line Cinema Corp., 693 F. Supp. 1517. See also Samuelson, supra note 28, at 2550 (noting that after Campbell, “every subsequent parody case has been adjudged a fair use”).
  93. See Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 447–53 (1984).
  94. Id. at 449.
  95. See Nimmer & Nimmer, supra note 4, at § 13.05[A][1][c].
  96. Campbell v. Acuff-Rose Music, Inc, 510 U.S. 569, 585 (1994).
  97. See id.
  98. See Beebe, supra note 6, at 601–03.
  99. See, e.g., BMG Rights Mgmt. U.S., LLC v. Global Eagle Ent., Inc., No. 2:18-CV-03723-VAP-JEMx, 2019 U.S. Dist. LEXIS 206738, at *9 (C.D. Cal. Sept. 9, 2019); Disney Enters. v. Vidangel, Inc., 371 F. Supp. 3d 708 (C.D. Cal. 2019); Lucasfilm Ltd. v. Ren Ventures, No. 17-cv-07249-RS, 2018 U.S. Dist. LEXIS 144116, at *5 (N.D. Cal. June 29, 2018).
  100. See Beebe, supra note 6, at 596–97 (noting that “the [Supreme] Court has repeatedly sought to reconstrue what it should have explicitly rescinded and replaced”).
  101. See id. at 607–09. See also Leval, supra note 40, at 612–14 (strongly criticizing the consideration of good faith as part of the fair use analysis).
  102. See Swatch Group Mgmt. Servs. v. Bloomberg L.P., 742 F.3d 17 (2d Cir. 2014), amended and superseded by 756 F.3d 73 (2d Cir.); NXIVM Corp. v. Ross Inst., 364 F.3d 471 (2d Cir. 2004); Narell v. Freeman, 872 F.2d 907 (9th Cir. 1989); Yang v. Mic Network, Inc., 405 F. Supp. 3d 537 (S.D.N.Y. 2019).
  103. See, e.g., Estate of Smith v. Cash Money Recs., Inc., 253 F. Supp. 3d 737, 751 (S.D.N.Y. 2017) (“As the Second Circuit has noted, this factor ‘is rarely found to be determinative.’” (citing Davis v. Gap, Inc., 246 F.3d 152, 175 (2d Cir. 2001)).
  104. Of the 579 opinions in the overall dataset, only 22.6% found that factor two favored a finding of fair use as against 47.7% that found that the factor disfavored a finding of fair use.
  105. See Beebe, supra note 6, at 614 (“[T]he Supreme Court sought to establish that a certain finding (here, that the work is unpublished) disfavors fair use. Lower courts appear not to have acted on that dictum, however, other than to invert it to conclude that the opposite of that finding (here, that the work is published) favors fair use.”).
  106. Stated in terms of predicted probabilities, the regression results under model 1 suggest that the predicted probability that a court will find fair use when the work does not qualify as published is .42 (with all other variables set to 0) as against a predicted probability of .65 when the work qualifies as published.
  107. For a thorough quantitative analysis of the role of the third factor in the overall fair use analysis, see Anthony Reese, How Much Is Too Much: Campbell and the Third Fair Use Factor, 90 Wash. L. Rev. 755 (2015).
  108. The two outliers were Video Pipeline, Inc. v. Buena Vista Home Ent., Inc., 342 F.3d 191 (3d Cir. 2003); Video-Cinema Films, Inc. v. Lloyd E. Rigler-Lawrence E. Deutsch Found., 2005 U.S. Dist. LEXIS 26302 (S.D.N.Y. Nov. 1, 2005).
  109. By contrast, in the original data set of all opinions from 1978 through 2005, of the 99 opinions that found that the defendant had taken the entirety of the plaintiff’s work, only 27.3% found fair use. See Beebe, supra note 6, at 616.
  110. See, e.g., Salinger v. Colting, 641 F. Supp. 2d 250, 267 (S.D.N.Y. 2009) (“Because Defendants have taken much more from Salinger’s copyrighted works than is necessary to serve their alleged critical purpose, the third factor weighs heavily against a finding of fair use.”), vacated, 607 F.3d 68 (2d Cir. 2010).
  111. 467 F.3d 244 (2d Cir. 2006).
  112. 714 F.3d 694 (2d Cir. 2013).
  113. 804 F.3d 202 (2d Cir. 2015).
  114. 766 F.3d 756 (7th Cir. 2014).
  115. 883 F.3d 169 (2d Cir. 2018).
  116. But see Jane C. Ginsburg, Fair Use in the United States: Transformed, Deformed, Reformed?, Sing. J. Legal Stud. (forthcoming 2020) (discussing the “reinvigoration of the fourth factor” at least in prominent cases).
  117. See Bouchat v. NFL Props. LLC, 910 F. Supp. 2d 798, 812 (D. Md. 2012); Warren Publ’g Co. v. Spurlock, 645 F. Supp. 2d 402, 427–28 (E.D. Pa. 2009); Williamson v. Pearson Educ., Inc., No. 00 Civ. 8240, 2001 U.S. Dist. LEXIS 17062, at *19–20 (S.D.N.Y. Oct. 19, 2001).
  118. See, e.g., Designer Skin, LLC v. S & L Vitamins, Inc., 560 F. Supp. 2d 811, 824–25 (D. Ariz. 2008).
  119. See Beebe, supra note 6, at 617–18. Indeed, regressing the outcome of factor four (rather than the overall test outcome) on the various subfactor findings and other objective factors listed in Table 5 results in largely the same odds ratios as those reported in Table 5—which should not be surprising given the tight correlation between factor four and the overall test outcome.
  120. See David Fagundes, Market Harm, Market Help, and Fair Use, 17 Stan. Tech. L. Rev. 359 (2014); Jeanne C. Fromer, Market Effects Bearing on Fair Use, 90 Wash. L. Rev. 615 (2015); Xiyin Tang, Can Copyright Holders Do Harm to Their Own Works? A Reverse Theory of Fair Use Market Harm, 54 U.C. Davis. L. Rev. (2021) (forthcoming). See also Shyamkrishna Balganesh, Copyright as Market Prospect, 166 U. Pa. L. Rev. 443 (2018); Shyamkrishna Balganesh, Foreseeability and Copyright Incentives, 122 Harv. L. Rev. 1569 (2009).
  121. See Fromer, supra note 120, at 646–49.
  122. On the role of industry custom in the fair use analysis, see Jennifer E. Rothman, The Questionable Use of Custom in Intellectual Property, 93 Va. L. Rev. 1899, 1930–45 (2007).
  123. This may help to explain why factor four reported such a proportionally low word count in Figure 5.
  124. Harper & Row Publishers, Inc. v. Nation Enters., 471 U.S. 539, 566 (1985).
  125. Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 578 (1994) (“All are to be explored, and the results weighed together, in light of the purposes of copyright.”).
  126. Id. at 590 (citations omitted) (omission in original).
  127. Sony v. Universal City Studios, Inc., 464 U.S. 417, 451 (1984).
  128. Campbell, 510 U.S. at 591.
  129. See Beebe, supra note 6, at 618–21.
  130. For an example of an opinion that relied on the Sony fourth factor commerciality presumption to find no fair use, see TD Bank v. Hill, Civil No. 12-7188 (RBK/JS), 2015 U.S. Dist. LEXIS 97409, at *52 (D.N.J. July 27, 2015) (where defendant’s book copied elements of plaintiff’s manuscript, finding that “[i]f the intended use is for commercial gain, [the] likelihood [of market harm] may be presumed.” (quoting Video Pipeline, Inc. v. Buena Vista Home Entertainment, Inc., 275 F. Supp. 2d 543, 565 (D.N.J. 2003) (quoting A&M Records, Inc. v. Napster, Inc., 239 F.3d 1004, 1016 (9th Cir. 2001), amended (Apr. 3, 2001), aff’d sub nom. A&M Records, Inc. v. Napster, Inc., 284 F.3d 1091 (9th Cir. 2002)))).