joegratz.net

Ed Felten, a professor at Princeton active in the “freedom-to-tinker” movement, has posited a Grand Unified Theory of Peer-to-Peer File Sharing. It brings together the results of all of the recent studies of P2P usage and its effect on CD sales.

The studies show that (1) when asked, people who admit to file-sharing buy fewer CDs, and (2) based on time-series correlation of downloads and CD sales, downloads have no effect on CD sales. These results seem contradictory, but Felten’s theory explains both.

Under Felten’s GUT, there are two kinds of P2P users: Free-riders and Samplers. Free-riders are mostly younger and less risk-averse; they admit to file sharing and they buy less CDs because of it, downloading most of the music they listen to. Samplers are mostly older and more risk-averse; they don’t admit to file sharing, and they buy more CDs because of file sharing, buying CDs of music they listen to (which they find out about via downloads). Because P2P causes more CD sales to samplers and less CD sales to free-riders, the net effect is zero, as Strumpf and Oberholzer-Lee found.

The key data we’re missing that would tell us whether Felten’s theory is correct is a simple correlation between number of downloads and number of CD sales across genres. It seems that Samplers would be more likely to listen to jazz than Free-riders, for instance, and Free-riders would be more likely to listen to music that everyone agrees is teenybopper crap — say, Hilary Duff. I suspect that his findings would hold, at least in large part.