In a recent working paper by Robert G. Hammond of North Carolina State University the impact of album pre-releases in file-sharing networks on physical and digital album sales is analyzes. The paper comes to the conclusion that album sales benefit from album leaks. “[A]n album that became available in file-sharing networks one month earlier would sell 60 additional units”. In addition the results also suggest that popular artists benefit more from file-sharing than newcomers and less establised artists. In the following the analytical and methodolocigal background and the results of this paper are highlighted.
The paper focuses on album sales instead of single tracks and relies on data from private – invitation only – BitTorrent trackers. Hammond included all albums that were released between May 2011 and January 2012 in the U.S. Thus, he identified 1,095 album from 1,075 artists covering a variety of genres. Descriptive statistics show that 37.1% of the albums were released by the then four major companies – Universal Music Group (15.0%), Warner Music Group (9.5%), Sony Music Entertainment (7.8%) and EMI (4.8%). Another 22.4% of the albums were released by indie companies distributed by the majors. The rest of 40.6% of albums were released and distributed by indie labels. The analysis also highlights that 90.5% of the albums leaked before or at least after the official release date. 59.8% of the albums already leaked days or even weeks before the release date. The median album leaks 3.7 days and the mean album 7.7 days prior to the release date. In addition it can be shown that most of the albums leak within two weeks prior or soon after the official release date.
To assess the impact of file-sharing on album sales, the author purchased Nielsen SoundScan data for each album for the first six weeks after the release, since most of the the albums’ sales occur in that period – 55.8% of an album’s sale occur within the first two weeks.
In order to overcome the endogenous problem of file-sharing and record sales, Hammond introduced availability of albums on file-sharing networks as an instrumentel variable. Availability is measured as a ratio of the number of seeders of an album and its leechers. According to the author, this ratio does not vary with an unobserved album quality and is thus exogenous.
Considering all model specifications, file sharing has a positive effect on physical and digital album sales. “[A]n album that leaked one month earlier will receive 59.6 additional sales” (p. 15). However, more established artists – with two previous albums, both of which sold at least 100,000 units – benefit more from file sharing than less established ones. The author speculates “(…) that artists with established fan bases are positively predisposed toward the [new] album” than younger and less established artists (p.19).
In respect of music genres the file sharing effect on more popular genres such as pop, country and hiphop/rap is larger on less popular or niche genres such as folk, metal, jazz. In addition, major labels benefit more from pre-releases on file sharing networks than major-distributed indie labels, which outperform pure independent labels. Among the major companies Sony Music Entertainment benefits most from file sharing, followed by Universal Music Group, EMI and Warner Music Group.
With a similar methodological approach than Oberholzer-Gee and Strumpf (2007), Hammond comes to the conclusion that file sharing has a weak but positive effect on album sales. Hammond adds in this respect, that he does not “find any evidence of a negative effect in any specification, using any instrument” (p. 21). Thus, an artist should not expect his/her sales to decline if his/her album is pre-released on a file-sharing network. However, the results, which are in line with Will Page’s and Eric Garland’s (2009) observations that “the same artists are popular with both legal and illegal downloaders” (p. 23), conflicts with David Blackburn (2004) findings that album sales of well-established artist are hurt by file-sharing whereas younger, less established artist benefit from sharing activities. Hammond addresses this discrepancy by explaining that Blackburn’s data do not contain information on the number of downloads, but only on the number of available files, whereas his data provides both information on availability and popularity. A second reason why Blackburn comes to a different conclusion is the instrument variable in use. His instrument is the RIAA’s announcement to sue file-sharers in 2003. Instead of Blackburn’s simple dual instrument, Hammond provides with pre-release availability a continuous instrument variable, which offers in his own words “econometric and theoretical improvements” (p. 22).
To sum up, Hammond’s paper confirms the results not only of Oberholzer-Gee and Strumpf (2007) and in some respect also of Blackburn (2004), but of more recent studies by Bhattacharjee et al. (2007) and McKenzie (2009), who use the same methodological approach of directly measuring file sharing activity. Although different instrument variables are used in each of the papers, the message is with one consent: There is no reason to believe that file sharing hurts regular music sales.
Bhattacharjee, Sudip, Ram D. Gopal, Kaveepan Lertwachara, James R. Marsden and Rahul Telang, 2007, “The Effect of Digital Sharing Technologies on Music Markets: A Survival Analysis of Albums on Ranking Charts.” Management Science, Vol. 53, No. 9 (September 2007), pp. 1359-1374.