Study Finds P2P Has No Effect on Legal Music Sales

MBrichacek writes "The Journal of Political Economy is running the results of a study into P2P file-sharing, reports Ars Technica. The study has found that, contrary to the claims of the recording industry, there is almost no effect on sales from file-sharing. Using data from several months in 2002, the researchers came to the conclusion that P2P 'affected no more than 0.7% of sales in that timeframe.' 803 million CDs were sold in 2002, according to the study, which was a decrease of about 80 million from the previous year. While the RIAA has been blaming that drop (and the drop in subsequent years) on piracy, given the volume of file-sharing that year the impact from file sharing could not have been more than 6 million albums total. Thus, 74 million unsold CDs from that year are 'without an excuse for sitting on shelves.'"

The Original Report

[eldavojohn]

The paper that The Journal of Political Economy [uchicago.edu] is citing is The Effect of File Sharing on Record Sales: An Empirical Analysis [unc.edu][PDF Warning!] which I found hosted on Koleman Strumpf of UNC Chapel Hill's school homepage although it is also available via one of my favorite (though not very comprehensive) research sites, Citeseer [psu.edu].

Something interesting to note is that this paper is dated March of 2004 (not too new as Ars Technica reported) and it causes me great wonder why I've never come upon this before (or why it's never been cited in the news). I recall reading tons of reports from one of the Associations where piracy is proven to hurt record sales but several years after this one is published, I finally see it.

For those of you interested in the data, pages 34 on contain some very interesting data whereby downloads are broken down by song, album, country & genre (in case everyone was trying to pin illegal downloads on those damned teeny boppers).

For those of you who wish to question the sample size, see Section B. "File Sharing Data and Album Sample" of the paper. You will also be interested in reading Appendix A in which they call into question their own sample sizes and weigh in on how accurate they might or might not be. To quote the paper for some more detail on the downloads samples,

Over the sample period we observe 1.75 million file downloads or roughly ten per minute.10 This is about 0.01% of all the downloads in the world. A significant majority of the downloads were music files. U.S. users accounted for about one third of the downloads (and the data contain about 0.01% of all music downloads by U.S. users).

To quote the paper on album sales samples,

The mean of sales for these albums during our observation period is 151,786 copies, ranging from 71 copies to 3.5 million copies.

Don't kid yourself, this is a difficult study to do. Both the downloads and album sales must be sampled and modeled correctly to draw correct conclusions. In the end, it would be hard to verify/discredit any studies done on this topic since A) consumers are human and therefore erradic & B) macro economics still isn't well understood.

Now, for those of you who just want the bottom line at the end of the paper,

We find that file sharing has no statistically significant effect on purchases of the average album in our sample.

And, from the very end of the paper,

If we are correct in arguing that downloading has little effect on the production of music, then file sharing probably increases aggregate welfare. Shifts from sales to downloads are simply transfers between firms and consumers. And while we have argued that file sharing imposes little dynamic cost in terms of future production, it has considerably increased the consumption of recorded music. File sharing lowers the price and allows an apparently large pool of individuals to enjoy music. The sheer magnitude of this activity, the billions of tracks which are downloaded each year, suggests the added social welfare from file sharing is likely to be quite high.

Yeah, that's right, the research concluded that "file sharing probably increases aggregate welfare." I'll bet if we all got drills & augers, we could get that into the brains of the people running the RIAA & MPAA.

Re:Where is the study on how much CDs suck?

[penguin_dance]

Now that people can buy individual songs, there's no excuse for a record company or artist to put 2 or 3 good songs and then fill the rest up with crap. And they're slitting their own throats doing so.

But if the RIAA would get their heads out of their ass and realize that the majority of the population doesn't want to hear the crap they put out they might turn things around. First off they're marketing to the wrong bunch. They're marketing to the decling population of teens to twenties. This worked in the 60s and 70s when the baby boomers were that age. Then they dropped them for the younger market. I would argue that most boomers are unlikely to download warez, would gladly purchase some music they'd like to hear and have the disposable income to do so. Just look at the sales of a release of Beatles No. Ones a few years back. And (get ready for the "old foggie" comment) it doesn't have to be old groups, but groups that can sing, not lipsync for for MTV. (Have you ever thought about how many former #1 selling artists would never have made it on American Idol today because they weren't good looking enough!?) Everything today is either rap or Beyonce. I would really like to find some new artists that I can like consistently, but they're few and far between.

Re:Exactly

[randomblast]

As someone who works in the live music industry, I can tell you that the majority of tours only just break even; the rest of them just don't. Gigs serve to promote the album. Why do you think the Grateful Dead haven't retired yet? Can't put much of a pension aside when you're working to pay for your job.