The study, entitled "The Effect of File Sharing on Record Sales: An Empirical Analysis," looks at .01% of the world's downloads from the last third of 2002, and in doing so examines only observed behaviour - there are no surveys integrated into the study, only empirical evidence. After some analysis, the study concludes that the effect that download have on album sales is "statistically indistinguishable from zero."
We find that file sharing has only had a limited effect on record sales. OLS estimates indicate a positive effect on downloads on sales, though this estimate has a positive bias since popular albums have higher sales and downloads. After instrumenting for downloads, most of the impact disappears. This estimated effect is statistically indistinguishable from zero despite a narrow standard error. The economic effect is also small. Even in the most pessimistic specification, five thousand downloads are needed to displace a single album sale.This study puts in the limelight accusations made by the major labels over the course of the last ten or so years. If downloads are having such a powerfully negative effect on album sales, why is there no hard evidence of this causation? Do these accusations hold any merit?
According to some quick calculations performed in about six seconds by me, in order to offset the decline in album sales last year (about 30.3 million less than 2005), downloaders would have had to download 151.55 billion more albums (at 100mb per album, that would require a 15.155 more terabytes of music) in 2006 than in 2005. As a contrast, small independents have seen increases in popularity due to the advent of digital distribution and the publicity afforded by torrent sites such as indietorrents.com.
The Effect of File Sharing on Record Sales: An Empirical Analysis
Top 50 Albums 2005
Top 50 Albums 2006