Elephants Businessman
Eelnosiva/Getty Images

If you think you know about Spotify's impact on piracy and music purchases, you may need to think again.

The working paper -- a collaboration between the National Bureau of Economic Research and the Institute for Prospective Technological Studies (a division of the Europen Commission's Joint Research Center -- was written by Luis Aguiar (of IPTS) and Joel Waldfogel (of the NBER and the University of Minnesota), received a great amount of attention after an article appeared on Slate on October 30. (Because it's a working paper, it has not been peer reviewed and hasn't been published.) It searches for a relationship between streaming and music industry revenues by examining Spotify streams, music piracy and music sales across multiple countries.

The coverage was heavy but misleading. Slate's title claimed the study "says Taylor Swift is wrong about Spotify," although the paper doesn't address the things about Spotify Swift opposes. The article was echoed by Business Insider. Re/code was more balanced: "How Can Spotify Shrink and Grow the Music Business at the Same Time?" The Atlantic proclaimed: "Spotify Isn't Killing Record Sales."

Get beyond the paper's abstract and you might find the headlines and their corresponding articles represented the paper's nuance and sometimes conflicting information. At his blog Rockonomic, David Touve of the University of Virginia -- and longtime watcher of the music industry -- questions the abstract's explanation that Spotify streams have a neutral effect on music sales.

Take the main statistic in the abstract: 137 Spotify streams displace one track sale. Aguiar and Waldfogel call this a "revenue-neutral" combination of streaming royalties and lost track sales. But is it?

The researchers' own numbers don't seem to support this explanation, according to Touve. At a royalty rate of 0.71 cents per stream (an average of the range of royalties presented on page 29; Touve uses 0.72 cents), 137 streams generate royalties of 0.986 cents. Those incremental 137 streams displace a track sale that returns, on average, 0.822 cents for record labels. The net result is a gain of about 15 cents. That's hardly neutral.

Touve raises another red flag regarding the study's focus solely on Spotify. It's hard to say exactly what influences track sales. What about YouTube and Vevo? What about Deezer, which is not popular in the United States but had 1.36 million subscribers in France at the end of  2013? They could be affecting track sales, too. And who's to say royalties from other interactive streaming services wouldn't help alleviate or overcome the lost royalties from displaced track sales?

"In other words, we don’t know which portion of any lost revenue in the sales market could be attributed to Spotify, exclusively," Touve writes. "And yet, the... paper essentially treats the situation as if the entirety of any decrease in unit sales can be explained by adoption of Spotify alone."

There are other reasons to avoid headlines and articles that trade nuance for certainty. The study was unable to differentiate between Spotify's free and premium tiers. It's possible free streams affect track sales differently than streams by subscribers. What's more, it's possible -- as some labels allege -- that Spotify's unlimited free tier, supported by advertisements and meant to encourage listeners to upgrade to the premium service, discourage listeners from becoming subscribers. (Spotify would undoubtedly argue differently.)

This gets close to the heart of Swift's arguments against Spotify: she doesn't believe the free tier adequately compensates creators and says it cheapens the value of music. (Her label boss, Scott Borchetta, often wears a jacket with "Music Has Value" stitched on a left sleeve.) "Music is art, and art is important and rare. Important, rare things are valuable. Valuable things should be paid for," she wrote in a Wall Street Journal op-ed in July 2014. But Swift never argued Spotify reduces total music industry revenues.

As is common in economics papers, Aguiar and Waldfogel couch their findings in layers of additional explanation and admittances of uncertainty. Studies like this rarely have the certainty implied by a news headline. Reader beware.