The Economics of Movie Downloads in the Film Industryby Jie Zhou
Despite a growing amount of economic literature on music downloads, the topic of movie downloads has largely remained unexplored. This paper uses survey data from a sample of college students to create demand curves which are used to simulate models for various movie products. Analysis results show that movie downloaders value legal products such as cinema tickets and DVDs as much as non-downloaders, although downloaders are much less receptive to new pay per view and online movie products. Maximum societal surplus is captured when industry can price discriminate between downloaders and non-downloaders. The industry also seems to have little incentive to create competitive new online legal alternatives due to the strong consumer preference for traditional movie products.
Piracy of information goods has become a vicissitude of the Internet. Despite much publicity relating to music piracy, only recently have people come to realize the feasibility of movie piracy. In 2003 the MPAA estimated that between 400,000 and 600,000 films were being illegally downloaded on the Internet everyday1. If that figure was correct, then the true figure is most certainly higher now.
Unlike the music industry, which has attributed online piracy to the slight fall in music sales during 2000-2002, the movie industry saw record revenue in both box office and home video sales in 20042. However, determined not to fall victim to "napsterization" like the music industry, the movie industry has acted swiftly in issuing lawsuits against and providing legal alternatives to piracy. Despite the comparison between the two industries, inherent differences still cloud the future of illegal downloads and its impact on the movie industry.
So far there has been very little academi erature on the economics of movie downloading. This study will use survey data to simulate markets with varying product combinations to evaluate the effect of illegal movie downloads on traditional movie markets as well as the market potential of new pay per view and streaming products.
The data set used in this paper is derived from a fifteen-question multiple-choice online survey conducted over a nine-day period. Respondents were asked to select their maximum willingness to pay for a range of movie products and services as well as their patience in searching and waiting for movie downloads. The sample used for analysis consists of 97 unique data sets, most of which were derived from responses from Yale undergraduates. This is by no means a representative sample.
Inherent Differences Between Music and Movie Piracy
Typically a movie file is 200 times larger than an audio file. The larger size of movie files results in a two-fold impact on download time. First, it lengthens the minimum time it takes to download the movie (200 times longer than the time needed to download the average mp3 song). Second, a movie download requires the download connection to be stable for a longer period of time, increasing both the likelihood of a failed download due to a peer user going offline before the download finishes and the probability of data corruption during the download process.
Large files also have supply side effects. Movies are more costly to store and share because of the extremely high opportunity cost of keeping large files in limited hard disk space. As a result, fewer movies are shared, and searching for movies (especially older movies) is especially difficult.
These are the key obstacles that the online movie distribution community is still trying to overcome.
The survey tackled the search-time issue directly by asking the maximum time each respondent would spend searching for and downloading a movie. Results varied greatly. The correlation between the number of downloads in the past 12 months and download patience is 0.517493. However, it is ambiguous how the two variables are interrelated. One can imagine that innately patient people are willing to wait longer for downloads and ultimately download a greater amount of movies. This also suggests that download speed is a significant inhibitor to downloading - one which will become less significant as internet speeds increase. On the other hand, it is also possible that some Yale students download more than others due to other exogenous factors (high demand for fresh movies, megalomania) and that those who do download more become more tolerant of slow download speeds. This possibility suggests that patience is a dynamic variable that will only increase over time with greater internet use thus also lowering the perceived cost of downloading. Overall, this result does not reveal anything we don't already suspect - that movie downloading will only increase in the future.
However, respondents' patience exceeded my expectations. The means for search time and download time were 32 minutes and 38.12 hours respectively. More than a third of the respondents were willing to endure modem speeds to download a typical 700 megabyte movie. More than half the users were willing to download movies at a mere 8.3 kilobytes/sec and only a sixth of respondents demanded no less than a 1Mbit connection. These results challenge the perception that broadband is necessary for practical movie distribution. There is some sample bias here. Almost all respondents have access to permanent broadband connection allowing users to download without fear of sporadic disconnections. Most respondents relied on Yale College Internet, which used bandwidth shaping tools to limit most peer to peer traffic to approximately 3 kilobytes per second. These extenuating circumstances may have made Yale students more tolerant of slow speeds than most other college students. For further study, a larger survey with a more diversified sample is recommended.
Movies as an Experience Good
Like music, a movie is an experience good whose attributes and quality are hard to discern prior to purchase3. However, unlike the music case in which people enjoy and demand hearing songs multiple times, the utility derived from a movie is usuall exhausted in the first viewing. There are some special cases in which certain movies demand multiple viewings and certain types of people (e.g. small children) enjoy the same movie multiple times. However, in most cases, it makes more sense to delete or backup the downloaded movie after watching it once. This further decreases the incentive for users to leave movies on the hard disk space where they are available for sharing.
Effects of Downloads on Traditional Movie Products
One of the arguments that attempts to discount piracy's negative effects is that pirates value the product very little and thus would not have purchased the product had the option of piracy not been available. In such a case, not only is revenue unaffected, but deadweight loss shrinks as well4. In the survey, respondents were asked the maximum amount of money they were willing to pay to see a cinema released movie, rent a movie, and buy a movie on VHS or DVD. The survey was specifically worded to extract the ex ante valuation of a non-specific movie which the respondent would be willing to see. This information was used as the dependent variable in a regression involving the number of movies the respondent had downloaded in the past 12 months. The regression coefficients for cinema, rental and videos were -0.02489, -0.01616, -0.03288 respectively, none of which are statistically significant. In fact, if we split the sample into only those who did not download anything and the rest, the difference in average willingness to pay for cinema, rental and video between the two groups is only -2%, -8% and -3% respectively.
The natural interpretation of this result is that downloaders value movie products just as highly as non-downloaders. However, this does not mean that movie downloads have a negligible impact on willingness to pay. For the question "what is your maximum willingness to pay for X", I believe that downloaders either assume that they do not already have an illegal copy of X but could attempt to obtain it or assume that piracy is not an option at all. If they assume the former, then the lack of correlation between willingness to pay and movies downloaded is significant since it shows that the option of movie downloads is not an inhibiting factor when deciding to purchase legitimate products.
Another relevant survey result is the revaluation of willingness to pay for movie products after obtaining an illegal copy. Again, there is no significant correlation between the number of movies the respondent downloaded and their revaluation of willingness to pay.
Movie rentals are not devalued by a statistically significant amount while cinema and video are devalued by an average of $5.44, and $12.28 (or 46% and 56%) respectively among all respondents. Thus theoretically if all respondents had an illegally downloaded copy of every single movie they would consider purchasing, then they should only be willing to pay a maximum of $6.42 for cinema tickets and $9.53 for DVDs. One possible theory to explain this result is that there is a large cost involved in actually obtaining illegal movies (such as search effort and download time) that cannot be justified by the gains in consumer surplus.
All market simulations in this paper were conducted in Microsoft Excel using logical functions, trial and error, and brute force optimization with Excel Solver. Consumer demand is directly based on discrete survey data with no attempts at smoothing or creating a probabilistic distribution. We assume a monopolistic market with all prices controlled by a single industry entity. The customers consist of the 97 respondents of the survey. They will buy the movie product if the price is at or below their maximum willingness to pay. In the first simulation, a new movie is about to be released on either cinema, rental or video and the industry is trying to decide on a revenue maximizing price. We assume a world without piracy. The revenue maximizing expression is given by:
Where Wi is the highest willingness to pay for the movie product for the ith respondent, and P is the price of the movie product.
The expression inside the summation gives the value $P if the maximum willingness to pay (W) of customer i is greater or equal to the Price (P) or $0 otherwise.
Results of the optimal solution are shown here:
|Optimal Pricing||Total Surplus||Producer Surplus||Consumer Surplus||Total Surplus Achieved|
The number in the brackets is the percentage of total surplus captured.
In the second model, we add piracy and assume that all downloaders (those who have downloaded at least one movie) have the option of downloading the movie being released at no cost. Maximum willingness to pay for non-downloaders remains the same. The revenue maximizing expression is given by:
Where Wi is the highest willingness to pay for the non-downloader i.
W*i is the highest willingness to pay for the downloader i who has the costless option of obtaining an illegal copy of the movie.
Downloaders who do not purchase the movie product legally still gain benefit from having the illegal copy (calculated by Wi - W*i).
After solving the optimization problem, the following results were found:
|Optimal Price||Producer Surplus||Consumer Surplus||Total Surplus|
|Cinema||10||$600 (-25%)||$322 (56%)||$922 (-8%)|
|Rental||5||$300 (0%)||$28 (-59%)||$328 (-11%)|
|DVD||15||$1005 (-23%)||$923.5 (28%)||$1928.5 (-5%)|
The numbers in brackets show percentage change from the first model. Industry revenue for cinema and home video fell by approximately a quarter while rental was unharmed. Prices remained more or less unchanged. As expected, producer surplus decreased while consumer surplus increased. Interestingly total market surplus decreased slightly. This happens because the marginal benefit downloaders gain from the legal option of going to the cinema or having an authentic DVD cannot be justified by the price the industry is charging. Thus the extra value associated with the legal experience is not captured and becomes a dead weight loss.
In the third simulation, we assume that the movie industry can price discriminate between downloaders and non-downloaders. The revenue maximizing expression is:
Where P1 and P2 are the respective prices for non-downloaders and downloaders.
Results of the optimal solution are as follows:
|Non-Downloader's Price||Downloader's Price||Producer Surplus||Consumer Surplus||Surplus Achieved|
|Cinema||10||5||730 (-9%)||370 (79%)||1100 (96%)|
|Rental||4||5||325 (8%)||63 (-9%)||388 (90%)|
|DVD||15||15||1005 (-23%)||923.5 (28%)||1928.5 (91%)|
Percentage figure in surplus achieved column relates to proportion of total market surplus captured. All other percentage figures relate to percentage change in comparison with the results in simulation 1.
Price discrimination worked very well in capturing both consumer and producer surplus for cinema tickets but not so well for rental and video. Results for rentals have been rather inconsistent due to a large number of irrational survey answers. Price discrimination was not even worth attempting in the case of video sales since a large number were still willing to pay a premium price for DVDs despite having illegal copies. Overall, this model was able to capture the greatest proportion of total market surplus.
This analysis model has many faults. First, it assumes that the downloading of illegal movies for downloaders is costless. This has already been proven false in Part 1 of this essay. Finding and downloading requires much time and effort not to mention the high probability that the movie could not be found at all. Second, in Models 2 and 3, I have limited total benefit to the downloader to Wi (achievable if P = 0). The rationale is that benefit from the illegal download is represented by Wi-W*i and any extra benefit from the legal product is W*i. This is not necessarily true because the illegal download is not always an inferior substitute compared to the legal product and can sometimes be complementary. Compared to rentals and cinema screenings, downloads can be re-watched for an indefinite number of times and be shared among friends. Also, since 90% of movies on the Internet are released between its theatrical release date and its DVD release date5, the 'freshness' of the illegal download can generate additional benefit before legal alternative is available. Thus I believe the consumer surplus for downloaders is somewhat underestimated.
Conclusion and PredictionsThe movie industry faces a tough battle against online movie piracy. Download speed is a major inhibitor for those who are not downloading though for many people current speeds on the BitTorrent and eMule networks are more than tolerable. Capitalizing on consumer preference for freshness by releasing movies early, illegal downloads weaken sales of future rather than current movie products. Both downloaders and non-downloaders value movie products equally while almost all survey participants admitted that possession of a movie download will lower willingness to pay for legal products. At the same time, new alternative movie products are hampered by lack of selection, restrictive user rights, and uncompetitive prices. With that being said, the movie industry is still making record sales and should have little reason to endanger traditional movie product revenue with new products to combat piracy.
The industry should impose stronger internal security to prevent high quality movies from appearing on the Internet before its home video release date. It should also differentiate legitimate products to render online downloads poor substitutes. This includes adding more extra features on DVDs and accelerating the development of newer technologies such as BlueRay and HD-DVD. Better legal products will not eliminate but rather coexist with piracy and minimize revenue loss to industry.
Notes1: Valenti, Jack. "The Peril of Piracy and the value of movies and Intellectual Property to this Nation." MPAA.org. 30 Sep. 2003. MPAA. 16 Mar. 2005 .
2: "Yearly Box Office." Box Office Mojo. 2005. 15 Apr. 2005 .
3: Nelson, P. 1970. "Information and Consumer Behaviour". Journal of Political Economy, March/April, 311-29.
4: Rob, Rafael, and Joel Waldfogel. "Piracy on the High C's:." Prelim Draft (2004). 31 Jan. 2005 .
5: S. Byers, L. Cranor, E. Cronin, D. Kormann, and P. McDaniel, Analysis of Security Vulnerabilities in the Movie Production and Distribution Process. Telecommunications Policy, 28(8):619-644, August 2004.