A controversial new study by economists at Harvard and the University of North Carolina has found that file-sharing is not the cause of declining CD sales. Researchers spent a year and a half analyzing downloads and sales figures for 680 albums -- and what they found contradicts the record industry's claim that online piracy has led to a fifteen percent decline in sales since 2000.
"No matter how we use our statistical models, we cannot find a connection between decreased sales and downloads," says Felix Oberholzer-Gee, co-author of the report and a professor at Harvard Business School. "If you want to understand why sales have changed as dramatically as they have, do not look to file-sharing."
The fifty-one-page study -- arriving six days after the record industry sued another 532 file sharers -- is the most rigorous economic analysis available. It tracks downloading spikes and declines that are caused by factors unrelated to a song's popularity and uses the ebbs and flows to analyze file-sharing's impact on CD sales. "If it were true that increases in downloads decrease sales, we should see that whenever we have fluctuations in downloads, we would have fluctuations in sales," Oberholzer-Gee says. "That's not what we've found." Sales of top-selling albums such as the 8 Mile soundtrack, for example, did not decrease after several downloading spikes caused by factors such as Internet congestion and increased uploading from German students on vacation, according to the report. (Fourteen percent of music downloads occur in Germany; more than half take place outside the U.S.)
The research also supports the idea that most people download music that they wouldn't buy anyway. And, says Oberholzer-Gee, "the Internet is more like radio than we thought. People listen to two or three songs, and if they like it, they go out and buy the CD."
The record industry rejected the report immediately. "It flies in the face of reality," says an anti-piracy lawyer at one of the major labels. "All you have to do is ask a few college students to find out that they're buying less music." College students, in fact, have become a prime industry target: Eighty-nine of the alleged violators in the most recent round of cases were caught on university networks. Twenty-one schools ranging in prestige from Georgetown and Stanford to California State, Northridge, were caught in the net. And according to the Recording Industry Association of America, the strategy will continue: Every few months, about 500 cases will be filed, with people randomly pulled from services such as Kazaa. "We're not necessarily targeting university networks, but we want to send a stronger message," says Stan Pierre-Louis, the RIAA's executive vice president of legal affairs. "Everyone will face consequences if they violate copyright."
Two of the researchers who have done their own studies on file-sharing also have criticized the report. Josh Bernoff, an analyst at Forrester, a technology research firm, and Stan Liebowitz, a University of Texas economist, say that the Harvard-UNC study is flawed partly because it focuses too narrowly on the holiday season. Gift-buying in late December skews sales upward, they argue, undermining the damage that downloading might do.
But, says Oberholzer-Gee, "we excluded the holiday season and didn't find a different result." And while critics continue to question the study's methodology, Oberholzer-Gee says that the conclusions are irrefutable. "We did all the tests we could to make sure it was robust, and it is," he says. "What I find bizarre is that this is five years after file-sharing started, and we're the first people to do a sensible study. Wouldn't most industries do this before starting lawsuits? This should have been done years ago."