Music sharing doesn't kill CD sales, study says

For the study, released Monday, researchers at Harvard University and the University of North Carolina tracked music downloads over 17 weeks in 2002, matching data on file transfers with actual market performance of the songs and albums being downloaded. Even high levels of file-swapping seemed to translate into an effect on album sales that was "statistically indistinguishable from zero," they wrote.

"We find that file sharing has only had a limited effect on record sales," the study's authors wrote. "While downloads occur on a vast scale, most users are likely individuals who would not have bought the album even in the absence of file sharing."

The study, the most detailed economic modeling survey to use data obtained directly from file-sharing networks, is sure to rekindle debates over the effects of widely used software such as Kazaa or Morpheus on an ailing record business.

Big record labels have seen their sales slide precipitously in the past several years, and have blamed the falling revenue in large part on rampant free music downloads online. Others have pointed to additional factors, such as lower household spending during the recession, and increased competition from other entertainment forms such as DVDs and video games, each of which have grown over the same time period.

Executives at file-sharing companies welcomed the survey, saying it should help persuade reluctant record company executives to use peer-to-peer networks as distribution channels for music "We welcome sound research into the developing peer-to-peer industry, and this study appears to have covered some interesting ground," said Nikki Hemming, chief executive officer of Kazaa parent Sharman Networks. "Consider the possibilities if the record industry actually cooperated with companies like us instead of fighting."

The study, performed by Harvard Business School associate professor Felix Oberholzer and University of North Carolina, Chapel Hill associate professor Koleman Strumpf, used logs from two OpenNap servers in late 2002 to observe about 1.75 million downloads over their 17 week sample period.

That sample revealed interesting behavioral, as well as economic, data. Researchers found that the average user logged in only twice during that period, downloading about 17 songs. Some people vastly overshot that average, however--one user apparently logged in 71 times, downloading more than 5,000 songs.

The two professors narrowed their sample base by choosing a random sample of 500 albums from the sales charts of various music genres, and then compared the sales of these albums to the number of associated downloads.

Even in the most pessimistic version of their model, they found that it would take about 5,000 downloads to displace sales of just one physical CD, the authors wrote. Despite the huge scale of downloading worldwide, that would be only a tiny contribution to the overall slide in album sales over the past several years, they said.

Moreover, their data seemed to show that downloads could even have a slight positive effect on the sales of the top albums, the researchers said.

The study is unlikely to be the last word on the issue. Previous studies have been released showing that file sharing had both positive and negative effects on music sales.

The Recording Industry Association of America was quick to dismiss the results as inconsistent with earlier findings.

"Countless well-respected groups and analysts, including Edison Research, Forrester, and the University of Texas, among others, have all determined that illegal file sharing has adversely impacted the sales of CDs," RIAA spokeswoman Amy Weiss said in a statement. "Our own surveys show that those who are downloading more are buying less."