“Supply Responses to Digital Distribution: Recorded Music and Live Performances,” J. Mortimer, C. Nosko & A. Sorenson

Ah, good old music piracy. Incredibly, it’s been eleven years since Napster Beta came out. During this period, album sales have fallen a substantial amount in the US, among other countries. Presumably these two things are related, though the most careful paper on the topic (Oberholzer-Gee and Strumpf’s 2004 JPE) used quite detailed individual-level data and found no effect of internet downloads on sales of individual albums. They tossed this fact in as an aside, but I found it quite persuasive: the most popular genres for illicit mp3s saw almost exactly the same decline in sales as genres for which mp3 downloading was less common.

In any case, if it is true that Napster and the like substituted for sales, might bands have responded by touring more often? Mortimer et al gather comprehensive album sales and concert touring data for the US from 1995 to the mid-2000s. Since 1999, total concerts performed, and total ticket sales, have increased. This occurred even though the price of concerts has increased much faster than inflation. The effect was more pronounced in cities with higher broadband penetration. At the artist-level, concert revenues for the top 50 concert touring artists each year has fallen slightly since 1999, whereas revenues have increased monotonically as the rank of the band in terms of concert revenues increases; that is, revenues fell for superstar artists, went up slightly for moderately popular touring bands, and went up substantially for less popular artists. The same is true of album sales during this period: sales decreased substantially for top-selling bands, less so for less-popular bands. Data from an internet database of music releases also suggests that the number of new artists, and the number of CDs released each year, increased substantially during this period, but I fear the authors may be misreading this evidence. There are something like 30000 albums “released” (as in, available somewhere for sale) in the US each year, though many of these sell only tiny numbers. If the music database shows a doubling of new artists CD information, then since the database obviously isn’t tracking every CD, we may just be capturing an lowering cutoff for how successful a CD needs to be for someone to bother to enter it into the MusicBrainz database.

The sum of evidence is convincing that during the past decade, musicians made up for lost sales revenue by touring more often. I don’t know that Napster/mp3 services are necessarily the reason: bands also offered samples online, Youtube music videos were watched by huge numbers of people, music blogs made it easier to find new tunes, Pandora and the like broke the stranglehold of radio, etc. I also wonder why the authors don’t use data on downloads directly. Is it in fact true that downloads are more weighted toward smaller bands than in-store sales? My intuition says yes, but I have seen anecdotal evidence suggesting that there is no difference between the two. What is the magnitude of mp3 downloading during this period? Has it been continuously increasing?

More importantly, though, no one has yet written the important question when it comes to music piracy: was it welfare increasing or decreasing? In the static sense – i.e., no impact on innovation – it is necessarily good for welfare, since the number of tunes supplied by assumption stays the same, and the price falls to marginal cost (to zero), hence deadweight loss is reduced and welfare increases. Dynamically, though, if band entry slowed, then the effect of welfare is ambiguous. The evidence, given caveats above, seems to suggests that band entry did not decline. Therefore, welfare has increased. Essentially, you are being asked whether loosening copyright on music would be welfare increasing, and anyone who has read the political economy history here surely has the prior of “yes, it would.”

That said, testing the proper counterfactual, “the internet exists with all its other effects, but there is no such thing as mp3 sharing,” is a difficult job. I have some very preliminary ideas in this direction, so hopefully I’ll be able to share something here next year.

http://mortimer.fas.harvard.edu/concerts_01oct2010.pdf (WP version)


3 thoughts on ““Supply Responses to Digital Distribution: Recorded Music and Live Performances,” J. Mortimer, C. Nosko & A. Sorenson

  1. Tobias says:

    Very interesting indeed.

    There’s a very nice overview and critique of a lot of the empirical literature by Rufus Pollock, which you can find here: http://www.rufuspollock.org/economics/p2p_summary.html. I find the Blackburn results to be particularly interesting but haven’t looked at it closely enough to know how severe the econometric problems in the study really are.

    Moreover, the Oberholzer-Gee&Strumpf paper has come under quite severe attack by Stan Liebowitz. His reply is well worth reading: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1014399

  2. Tobias says:

    Don’t get me wrong: I agree that whether piracy is welfare enhancing or not is the crucial question from a policy point of view and I personally suspect that piracy has indeed been beneficial.

    That doesn’t mean that as a matter of positive economics the question of whether piracy has lead to a decline in music sales isn’t also of interest. As you rightly point out, what matters for welfare isn’t sales, it’s production. Thus, thinking that piracy has lead to a decline in sales but that it has been hugely beneficial aren’t mutual contradictory statements. To put it another way: “Piracy has not harmed even music labels” is a very different statement from “Piracy has been bad for music labels, but so what?”

    Now, I never knew Liebowitz had such strong ties to industry and it’s been ages since I read both the Oberholzer-Gee&Strumpf paper and Liebowitz’s reply, but upon rereading the latter, his criticism doesn’t strike me as obvious hackery. Whatever one may think of whatever else he’s written (and I’ve read precisely zilch), his criticism of O&S seems pretty reasonable to me.

  3. davidclowery says:

    Hi. I know you wrote this a long time ago, but as a long time independent musician, studio complex owner, and husband of a veteran concert promoter, it makes me laugh that you harvard guys can’t seem to find any effect of piracy on sales and revenues. or the misery it’s cause songwriters, musicians and anyone else that works for the artists. No one working in the biz thinks that piracy is beneficial. I would be glad to introduce you to hundreds of people that will give you there personal story.

    One of the main things you are misinterpreting is the rise in Live Music Revenue.

    Artists began to tour more as recorded revenues dropped. This probably increased artists revenue in the short term, but as any veteran concert promoter will tell you eventually each artist saturates the market with live shows.

    My wife and I predicted in the early 2000’s that average attendance for each individual artist was gonna fall. More artists touring, touring more often, longer tours, with lower attendance. This is exactly what has happened. So what you end up with is a headline number that says something like Global Live Revenues up. But if you dig into what’s happening to each artist? you see a nasty malthusian trajectory. The artists are netting much less. You can look up any of this data with a pollstar account.

    What you won’t see is the details of the contracts. But here is also something else dramatic. The last 5 years has become a buyers market for concert promoters. My wife has gotten to the point that she no longer offers guarantees, unless she is forced to. instead she just offers a 50/50 or 60/40 split of revenues. Most of her friends have similar stories.

    As a former Quant I probably appreciate more than you the subsidy she is receiving by dumping all the risk onto the artist.

    Live revenue will never replace recorded music revenue.

    Plus have you ever thought about this? the performer and the songwriter are often not the same person. how would a non-touring songwriter like brian wilson be fairly compensated?

    Another massive trove of data is downstream at the studio end. I can tell you just from my studio records (100s of projects over nearly 2 decades) recording budgets have fallen dramatically and length of projects are about 30% of what they once were.

    It’s not just my studio complex. there are dozens of legacy studio complexes in the country that would be happy to share their data.

    I also teach entertainment business finance and econ at UGA. I’m working on a similar study but I’d be glad to point you towards some good sources of aggregate data.

    Those of us out there still working in the business really scratch are head when see academics tossing around data from a 2004 study. That was the dark ages of file sharing and unlicensed downloads. now you can buy an Android app to your phone that searches cyberlockers and immediately downloads billboards hot 100 to your phone.

    I sometimes wonder if you guys just don’t really want to see the truth.

    Also record labels are dead. it’s like railing against the soviet union. it’s a little passe.

Comments are closed.

%d bloggers like this: