A report on music piracy by the European Commission's Joint Research Centre has ruffled some feathers in the music business. "Digital Music Consumption on the Internet: Evidence from Clickstream Data” also mischaracterizes the conclusions of two of the academic papers it cites.
The IFPI issued a press release this week denouncing the study that found a file sharing has a slightly positive relationship between visits to illegal download sites and legal download stores. IFPI chief executive Francis Moore said the study "contains significant flaws" and cited, among other things, its use of traffic to legal sales sites rather than actual sales data. Spotify's Will Page cites three problems. First, the lack of sales data mentioned by Moore. Second, illegal downloaders are not a uniform group. Page says as many as 44.8% of file-sharers in the U.K. do not buy any music. Third, the JRC study took too narrow of a view by focusing just on download revenue. According to the IFPI, Page writes, subscription services and ad-supported streaming sites account for 30% of recorded music revenue in Europe.
There have been numerous studies that have concluded file sharing hurts recorded music sales. The JRC study is different because it focuses only on file-sharing's effect on digital purchases. While it does attempt to further the boundaries of research in piracy's impact on digital sales, the JRC study would have you believe its conclusions are shared by other studies "that show illegal music downloads have little to no affect on legal digital sales." While there is some truth to that statement, it's not completely accurate. One study acknowledges piracy can help an individual artist while hurting the larger industry, but the study did not look specifically at digital sales. The other study was based on a survey that asked consumers only about CD sales, not digital sales. 
One paper cited is "Profit Leak? Pre-Release File-Sharing and the Music Industry," by Robert G. Hammond of North Carolina State University. This study used sales data from Nielsen SoundScan and data from a file-sharing website to estimate the effect of pre-release file-sharing on the sale of individual albums. The results showed that pre-release piracy had virtually no effect at the artist level.
But there's more to "Profit Leak?" given that it's cited in the JRC study. "Profit Leak?" also notes that file-sharing has "harmful effects on the music as a whole." It examines a very specific slice of consumer behavior -- piracy of pre-release album leaks. The paper lacks any conclusions on piracy's impact on legal streaming services (this is certainly relevant in today's market). And despite what the JRC study would have you believe, "Profit Leak?" examined album sales data and did not focus on digital sales data.
An April 2012 paper (translated into English as "The Impact of Piracy on the Purchase and Legal Download: A Comparison of Four Cultural Sectors") is also cited. The study surveyed the cultural practices and purchasing behavior of over 2,000 people. Rather than rely too much on Google Translate, I will say what can be easily seen in a translation of the paper's summary: legal downloading and piracy appear to be complements rather than substitutes.
A few problems with the JRC's characterization of "Impact" can easily be found using Google Translate and a quick scan, however. First, the survey on which the study is based took place in late 2008. The music marketplace was a vastly different place four-and-a-half years ago. Consumers had fewer legal music options and the country's Hadopi anti-piracy law had not yet been adopted. Second, the study examined piracy's effect on only CD sales, not digital download sales. The survey asked consumers only about their purchases of CDs. Thus, the JRC study was wrong to state that "Impact" found that illegal music downloads have little to no effect on legal digital sales. Third, "Impact" found that piracy helped sales of video games, not music CDs.
The story on file-sharing's impact on physical sales is pretty much written. Numerous studies have found that file-sharing hurts CD sales. More research needs to be done on file-sharing's impact on digital sales and -- if streaming companies are willing to share their data with researchers -- legal streaming services, too.