The down low on downloads

Ever since the slide in album sales began about four years ago, the music industry has been arguing that file sharing is to blame. Where’s the evidence to back that up?

Koleman Strumpf, an economics professor at UNC-Chapel Hill, decided to see if there really is any such evidence. He looked at server logs spanning four months in late 2002 of a major peer-to-peer network. There were about 2 million downloads, which Strumpf says amounts to about .01 percent of everything downloaded in the world during that period–a huge number for a single study, he says. “The idea is, if we can link every downloaded file to an actual song on an album, then we can say what’s the effect on record sales for that particular album.”

Strumpf and his co-author at the Harvard Business School, Felix Oberholzer, are the first to attempt a study of this kind. “It’s astonishing given all the media attention and interest in this question that nobody’s done a study like this before,” says Strumpf. They plan to release the results on his Web site (www.unc.edu/~cigar ) later this month.

Preliminary results show that file swapping’s impact on the industry has been grossly exaggerated. “It turns out that the effect is somewhat negative, but it’s very, very small. To give you an idea of the scale, it takes about 500 downloads to lead to one less album sale.”

That’s big news. “The record industry has consistently said, ‘We know file sharing is the reason,’” Strumpf says. “My research so far is saying that’s not true. It’s a contributing factor but it’s not the primary reason.”

People who download might buy less music, but correlation is not causation, Strumpf says. “It’s what an economist would call a selection effect. People who don’t buy things are also attracted to downloading, because it’s a free way to get the goods.”

Mostly, people download the big hits. Eminem’s song “Lose Yourself” from the 8-Mile soundtrack was popular during the months the researchers looked at–that song alone accounted for 1 percent of all downloads. “It’s pretty clear that labels understand this, because if you look at Billboard charts, two or three of the albums in the top ten are compilations of singles.”

Perhaps the reason major labels understand this so well is that they’ve been mining file-sharing data for years. A company called Big Champagne rates peer-to-peer files according to popularity, sort of an online Nielsen ratings. Their weekly list of “Top Swaps” has become the new metric of the music industry, replacing the Sound Scan ratings that came from record store sales. The syndicated programming arm of broadcast mega-giant Clear Channel announced a partnership with Big Champagne last year.

“The service is valuable to radio programmers,” reads the press release announcing the deal, “because they will be able to obtain immediate, market-specific reports about what passionate music listeners want. Record labels also use the information to gauge the popularity of their artists and songs online.” Remember, these are the same “passionate music listeners” that are being sued by the Recording Industry Association of America.

Top Swap hits fall along the lines you might expect: Big singles from mainstream artists like Britney Spears, OutKast, Alicia Keys–the same songs that get played to death on the radio.

What’s surprising is who’s doing the downloading. A separate study released last month by the Pew Internet & American Life Project demonstrates that the stereotype of the white middle-class college student isn’t entirely accurate. In a nationwide phone survey, 14 percent of all adults said that they download music; broken down by demographics, the picture gets interesting: One-quarter of African Americans surveyed said they download music, compared to 20 percent of Hispanics and 13 percent of whites. Even more surprising, the highest number of those who admitted to downloading made the least amount of money–22 percent of downloaders make less than $30,000 a year. People without a high school education were more likely to say they download (24 percent) versus 13 percent of those with some college education, and only 11 percent of those with a degree.

“That’s actually been the general trend for some time,” says Pew research specialist Mary Madden, “that those with lower levels of education and incomes have been somewhat more likely to download.”

The study, which surveyed 1,358 Internet users last November and December, also found that music industry lawsuits have successfully put a dip in the downloading trend. A previous Pew study done last spring recorded that 29 percent of all adults said they downloaded. File sharing applications Grokster and KaZaa also have reported sharp declines since then.

Madden says there’s no way to be sure if the dip in illegal downloading will last, especially in light of the recent U.S. Court of Appeals ruling that Internet service providers do not have to provide information about computer users in response to RIAA subpoenas. “Certainly the potential is there for some users to feel an increased sense of confidence, since the procedure the RIAA has to go through to determine the identity of users is a bit more complicated and cumbersome,” Madden says. Meanwhile, those who can afford it will likely be attracted to legal, paid subscriber services like iTunes. “Our research has suggested that the paid online music market is growing and will continue to grow.”

Strumpf thinks there are a lot of problems with phone surveys like the one done by Pew. First, they ask people to self-report their own illegal activity. “It’s hard for me to take that data at face value,” he says. Another big hole in the study, which Madden acknowledges, is that Pew wasn’t able to survey anyone under 18–kids tend to be proficient at downloading music, and cutting them out of the picture might also skew other aspects of the demographics.

Because nearly all downloads are major-label hits, the effect of downloading on independent labels is an open-ended question. Strumpf says the people he’s talked to at indie labels have wildly diverging viewpoints. Some think file sharing will prove to be the panacea for distribution and marketing, while others think it’s the kiss of death because their profit margins are so narrow to begin with. But since very few indie-label songs are downloaded, his study didn’t register much information at all about the economic impact on those labels.

Strumpf, whose own tastes run toward independents, says it’ll be difficult for a study like his to measure their financial prospects. “The kind of albums that are put out by indie labels are not economically very important,” he says. “I know that must sound like a terrible statement. Believe me, if you look at my music collection, I’ve come to the conclusion that most of the music I listen to is economically irrelevant.”

Links