Thorsten Hennig-Thurau, Victor Henning, & Henrik Sattler Consumer File Sharing of Motion Pictures Illegal consumer file sharing of motion pictures is considered a major threat to the movie industry. Whereas industry advocates and some scholars postulate a cannibalistic effect on commercial forms of movie consumption, other researchers deny this effect, though sound evidence is lacking on both sides. Drawing on extant research and utility theory, the authors present hypotheses on the consequences and determinants of consumer file sharing and test them with data from a controlled longitudinal panel study of German consumers. The data contain information on the consumers��� intentions toward and actual behavior in relation to the consumption of 25 new motion pictures, allowing the authors to study more than 10,000 individual file-sharing opportunities. The authors test the effect of file sharing on commercial movie consumption using a series of ReLogit regression analyses and apply partial least squares structural equation modeling to identify the determinants of consumer file sharing. They find evidence of substantial cannibalization of theater visits, DVD rentals, and DVD purchases responsible for annual revenue losses of $300 million in Germany. Five categories of file-sharing behavior drive file sharing and have a significant impact on how consumers obtain and watch illegal movie copies. Thorsten Hennig-Thurau is Professor of Marketing and Media, Depart- ment of Marketing and Media Research, Bauhaus-University of Weimar, and Research Professor of Marketing, Cass Business School, London (e-mail: tht@medien.uni-weimar.de). Victor Henning is a doctoral student of Marketing, Department of Marketing and Media Research, Bauhaus- University of Weimar (e-mail: victor.henning@medien.uni-weimar.de). Henrik Sattler is Professor of Marketing, Institute of Marketing and Media, University of Hamburg (e-mail: uni-hamburg@henriksattler.de). The authors thank Gary King, Marc Fischer, and the four anonymous JM reviewers for their helpful and constructive comments on previous ver- sions of this article. To read and contribute to reader and author dialogue on JM, visit http://www.marketingpower.com/jmblog. �� 2007, American Marketing Association ISSN: 0022-2429 (print), 1547-7185 (electronic) Journal of Marketing Vol. 71 (October 2007), 1���18 1 Eworldwide ver since the ascent of Internet file-sharing services and the parallel sharp decline of the music industry���s sales, movie executives have feared that their industry would be similarly affected by illegal file sharing (The Economist 2002). Recent figures show that approximately 130,000 movies are downloaded each day through file-sharing networks in the United States alone (Motion Picture Association of America [MPAA] 2004a), and movie theater admissions in 2005 fell by 9% in the United States and even more in other major markets. Against this backdrop, the MPAA (2004b) claims that ���ille- gal movie trafficking represents the greatest threat to the economic basis of moviemaking in its 110-year history,��� and it has declared ���war on piracy��� (Fritz 2005). However, sound evidence for the proclaimed effect of file sharing on movie consumption is lacking. A multitude of industry reports postulates a cannibalization effect of file sharing on movie industry revenues, but the results of acad- emic studies are inconclusive. No peer-reviewed article has yet investigated the effects of movie file sharing on com- mercial distribution channels, and the limited work that reports a negative effect of music file sharing on legal music consumption uses highly abstract proxies, such as ���Internet penetration,��� to measure consumer file sharing (e.g., Liebowitz 2006). At the same time, some researchers argue that file sharing does not damage the (music) industry and provide empirical (Oberholzer-Gee and Strumpf 2005) and theoretical (Gopal, Bhattacharjee, and Sanders 2005) argu- ments for the absence of a cannibalization effect���or even the presence of a positive effect of file sharing on legal consumption. We shed light on this controversial issue by employing controlled longitudinal panel data from 770 to 813 con- sumers that encompass information on more than 10,000 movie file-sharing opportunities. We use this data to investi- gate whether illegal movie file sharing influences revenues generated through theatrical visits, DVD rentals, and DVD purchases and, if so, how strong the effects are. In addition, we present���to our knowledge, for the first time���a com- prehensive, theory-based model of the factors that drive consumers��� movie file-sharing activity. This model offers the movie industry a more thorough understanding of why consumers engage in file sharing, suggesting more effective antipiracy strategies. We structure the rest of the article as follows: After reviewing the relevant literature, we derive a set of hypothe- ses regarding the consequences and determinants of movie file sharing from extant research and utility theory. We then report our data set and use ReLogit regression analysis and partial least squares (PLS) structural equation modeling to test the hypotheses. We conclude by discussing the results and implications. Motion Picture File-Sharing Literature File-Sharing Consequences Industry representatives unanimously argue that illegal motion picture file sharing has a negative impact on other

2 / Journal of Marketing, October 2007 kinds of movie consumption, and industry-commissioned studies, such as those of the German Federal Film Board (hereinafter FFA) (2006) and MPAA (2004c), support their claims. For example, in a study of movie piracy by the FFA (2006), respondents indicated how movie downloading or copying movies with a CD/DVD burner influenced their consumption of motion pictures through other channels. Of the respondents, 42% reduced their number of movie theater visits (though 8% stated they went to the movies more often), 45% said they rented fewer DVDs, and 44% replied that they bought DVDs less often. Similarly, the findings of an eight-country study commissioned by the MPAA (2004c) indicate that ���about one in four Internet users (24%) have downloaded a movie��� (MPAA 2004c, p. 1) and that, on a global level, 26% of downloaders purchase movies ���much less��� or ���a little less��� often than in the past (excluding Korea, the outlier, lowers the unweighted mean from 26% to 14%). The insights generated by these and other industry studies are limited by their methodological approaches and lack of transparency. In all cases, the results rely on an ex post ���what-if��� approach that asks consumers who have already seen movies as illegal copies (and there- fore know the cinematic quality) to speculate whether they would have paid for the movies if they had not been avail- able as illegal copies. To the best of our knowledge, no scholarly research has addressed the effects of sharing illegal movie copies on commercial distribution channels. In the related context of music file-sharing studies, researchers are split into two opposing groups. The first group reports a negative impact of music file sharing on industry sales (Liebowitz 2006 Michel 2006 Montoro-Pons and Cuadrado-Garc��a 2006 Peitz and Waelbroeck 2004 Zentner 2006), but these stud- ies all rely on aggregate household Internet penetration in a given city as a proxy for file sharing and do not monitor file sharing on an individual basis. Therefore, this approach raises serious questions regarding spurious correlations and paves the way for alternative explanations. The second group of researchers questions these find- ings and argues that file sharing has either no impact or a positive impact on industry revenues. Specifically, Gopal, Bhattacharjee, and Sanders (2005) propose a model of online music-sharing economics and derive implications for consumer surplus and producer profits. Following the train of thought that consumer file sharing represents a form of ���sampling��� for experience goods, they conclude that file- sharing networks lower the total costs of evaluating and acquiring experience goods, which increases purchases and industry profits. In other words, file sharing reduces con- sumers��� risk in evaluating new music (an argument that eas- ily extends to movies), a major obstacle in consumer deci- sion making. Using a different argument, Boldrin and Levine (2002) and Grgeta (2004) model competition with sunk costs and argue that, with certain assumptions, the decreasing costs of reproduction that result from file sharing make it easier, not more difficult, for the producer to recoup his or her invest- ment and that as the rate of reproduction increases, competitive rents increase. Their conclusion is based on the concept of indirect appropriability, which assumes that an original product attains greater consumer utility when it can be copied and that this utility increase can be captured by the producer through a price increase. However, similar to Gopal, Bhattacharjee, and Sanders (2005), these researchers do not provide empirical findings to substantiate their conclusions. Oberholzer-Gee and Strumpf (2005) present empirical results that show no negative impact of file sharing on tradi- tional music distribution channels. Over the course of four months, they monitor 1.75 million file downloads on file- sharing networks and then match the downloads to U.S. album sales data. Their empirical analysis shows that music file sharing has no significant impact on album sales. Again, however, the generalizability of their findings is somewhat limited because the authors use the ���number of German school kids on vacation��� as an instrumental variable for file-sharing activity to bypass endogeneity problems caused by the simultaneity of downloading and purchasing activity in their aggregate level data. To summarize, movie industry representatives argue that file sharing serves as a substitute for commercial movie consumption, but no peer-reviewed research has studied this relationship for movies, and the results from music file- sharing research are inconclusive and limited by method- ological constraints. Moreover, no existing study has sur- veyed actual consumer decision making on an individual level, and no study has used longitudinal data. File-Sharing Determinants: Rochelandet and Le Guel���s Model Related to the consequences of movie file sharing for com- mercial channels are the factors that drive consumer file sharing. Research into these factors is also rare we are not aware of a single academic study that directly addresses this question. Again, some scholars have researched file-sharing determinants in the related context of music. Most authors focus on the role of individual constructs for file sharing (e.g., ethical predispositions [Gopal et al. 2004] consumer expertise, social networking, and moral judgments [Huang 2005]), whereas Rochelandet and Le Guel (2005) attempt to integrate different drivers of sharing illegal music copies in a comprehensive model. Building on the Beckerian consumer utility framework, Rochelandet and Le Guel (2005) propose that consumers prefer illegal copies of music over the original product (i.e., a CD) when consuming the illegal copy offers greater util- ity. More specifically, they argue that three groups of factors influence consumers��� utility perceptions of the original and the illegal copy: (1) the utility derived from buying an origi- nal (including both gross utility and costs), (2) the costs of the illegal copy (mainly transaction costs), and (3) the degree of substitution between an original and its illegal copies. Rochelandet and Le Guel find partial support for their model from a convenience sample of 2500 French consumers. With an ordered logit approach, the factors in their model explain 10% of the music file-sharing intensity.

Consumer File Sharing of Motion Pictures / 3 Consequences and Determinants of Motion Picture File Sharing Motion Picture File Sharing as the Focal Construct We define the file sharing of motion pictures as consumers��� consumption of illegal copies of full-length motion pictures. This definition considers not only watching but also the mere act of obtaining illegal movie copies as forms of con- sumption. Although these two behaviors are closely related, they are conceptually distinct because consumers do not necessarily watch every illegal copy they obtain. Our use of the phrase ���illegal copies��� excludes original movies that consumers have the legal right to watch, such as those made available by their copyright owners to file-sharing networks or Internet video forums, such as YouTube, and commercial video-on-demand services, such as Movielink. Finally, our conceptualization of file sharing involves not only access- ing illegal movie copies from file-sharing networks (���Inter- net piracy���) but also the personal exchange of illegal movie copies among consumers (e.g., on CD-Rs and DVD-Rs ���hard goods piracy���) this is consistent with the conceptual- ization of movie file sharing used by the movie industry (MPAA 2006). The Effects of Motion Picture File Sharing on Commercial Channels Consistent with a consumer utility perspective of file shar- ing (Rochelandet and Le Guel 2005), we propose the exis- tence of negative (i.e., cannibalistic) effects of movie file sharing on movie consumption in the three key commercial channels: theater visits, DVD rentals, and DVD sales (e.g., Liebowitz 2006 MPAA 2004c). In all three channels, we distinguish among three related but distinct potential canni- balization effects. The first hypothesized effect refers to consumers��� inten- tions to watch an illegal copy of a movie. We propose that when a consumer has such intentions, he or she is less sus- ceptible to offers from theaters, DVD rental outlets, and DVD retailers because his or her intention to watch an ille- gal copy usually entails the expectation to obtain a copy of the movie for free instead of paying for it through legal channels. As a consequence, the consumer will refrain from using those commercial channels. This should be the case regardless of whether the consumer actually obtains an ille- gal copy of the movie. H1: A consumer���s intentions to watch an illegal movie copy reduce the probability that the consumer will (a) watch the movie in a movie theater, (b) rent the movie on DVD, or (c) purchase the movie on DVD. The second hypothesized effect refers to a consumer���s actual obtainment of illegal movie copies. Here, we argue that consumers who have gained access to an illegal copy of a movie have a lesser probability of seeing the movie in a theater or on DVD, regardless of (1) their original intentions toward watching an illegal copy of the movie and (2) whether they actually watch the illegal copy. Distinguishing between consumers��� intentions and their actual behaviors is important from a managerial perspective because if inten- tions influence commercial channel usage, the movie indus- try should focus its antipiracy activities on consumers who intend to watch a copy. In contrast, if actually obtaining illegal copies harms movie theaters and other channels, it is the copies that should be the focus of the industry���s antipiracy actions because any obtained copy would canni- balize commercial channels regardless of the consumers��� intentions. H2: For a given level of file-sharing intentions, a consumer���s obtainment of an illegal movie copy reduces the probabil- ity that the consumer will (a) watch the movie in a movie theater, (b) rent the movie on DVD, or (c) purchase the movie on DVD. The third hypothesized effect is related to the con- sumer���s watching of illegal copies. We postulate that con- sumers who watch an illegal movie copy have a lesser probability of seeing that movie in a theater or on DVD, regardless of their original intentions toward watching an illegal copy of the movie. Whereas our second hypothesis factors out what happens after the consumer obtains a copy, our third hypothesis posits that the specific act of watching the copy cannibalizes revenues. The relevance of this hypothesis stems from its associated managerial implica- tions specifically, it suggests that antipiracy actions should be directed toward preventing consumers from watching illegally obtained movie copies. H3: For a given level of file-sharing intentions, a consumer���s watching of an illegal movie copy reduces the probability that the consumer will (a) watch the movie in a movie theater, (b) rent the movie on DVD, or (c) purchase the movie on DVD. Determinants of Motion Picture File Sharing When modeling the determinants of movie file sharing, we build on Rochelandet and Le Guel���s (2005) utility theory approach but substantially refine and extend this approach in several ways. In general, we distinguish among five cate- gories of factors that we expect to drive consumers��� movie file-sharing behavior: perceived degree of substitution between an original movie and its illegal copies, utility of the original, (transaction) costs of the illegal copy, specific utility of the copy, and a consumer���s file-sharing knowl- edge. The first three categories come from Rochelandet and Le Guel, and the last two are unique to this study. We dis- cuss the categories and the individual drivers they encom- pass next and summarize them in Figure 1. Degree of substitution. A direct implication of the utility theory approach is that the degree to which a consumer per- ceives illegal movie copies as providing the same utility as watching the original movie in a theater or on DVD deter- mines the intensity of consumer file sharing. This perceived degree of substitution influences the utility of the illegal copy (Rochelandet and Le Guel 2005) and therefore should have a positive effect on the intensity with which consumers obtain and watch illegal movie copies. H4: The degree to which a consumer judges illegal movie copies as substitutes for movies in commercial channels