Published on (
 See this if you're having trouble printing code examples

Network Effects: Stan Liebowitz and the MP3 Debate

by Miriam Rainsford

Editor's Note: This is an opinion piece presented by Miriam Rainsford, who is a known proponent of P2P file sharing. Miriam is a musician herself and brings that point of view to this ongoing discussion about the damage that file sharing may be causing to the artists and their labels. Miriam is not a lawyer. She presents her views here, however, as if she were arguing a case in court. This is a writing device and not intended to be viewed as a true legal argument. Whether you're inclined to agree or disagree with the opinions in this article, I think you'll find this a passionate counterpoint to some of the conclusions presented by the report from the respected Stan Liebowitz.

The past month has been one in which the attention of the music industry has focused with renewed interest on digital downloads, beginning with the success of Apple's iTunes Music Store and finishing on a sour note with the RIAA's announcement that they will now be pursuing lawsuits against individual file-sharers. In light of these developments, it is worth re-examining Stan Liebowitz's analysis of the impact of MP3 downloads on the music industry. Can we in fact ascertain whether actual damage has been done through the rise of peer-to-peer file sharing systems, or is the record industry simply afraid of change and too slow on the uptake of new formats?

Stan Liebowitz, professor of Economics at the University of Dallas, Texas, became a controversial figure during the Microsoft anti-trust trial, when he testified that "network effects," the phenomenon where market saturation determines consumer choice, were more likely to have been responsible for Microsofts' dominant position than anti-competitive practises.

He later found himself to be an unlikely ally of the file-sharing community, when he released a study commissioned by the Cato Institute asserting that file-sharing "should be causing the industry a lot of harm. But we're not seeing it". However, with his latest paper, "Will MP3 Downloads Annihilate the Record Industry? The Evidence So Far", he concludes that we are beginning to see an undetermined, yet detrimental effect of file sharing on music sales. On examination I am concerned by several obvious flaws in Liebowitz's logic and would suggest a revision: that there is as yet no clear evidence that can prove beyond reasonable doubt that MP3 downloads are solely responsible for the current downturn in CD sales.

I cannot categorically say that MP3 downloading has no effect on CD sales. Nor do I support the RIAA's claims that file-sharing is "killing the music industry." What I wish to emphasize, through analysis of Liebowitz's paper, is that we lack the proof to convict beyond reasonable doubt. And if we cannot convict MP3 downloads of this "crime" against the music industry with the same burden of proof that we would expect of a trial in a court of law, then it is unethical and indeed unconstitutional to pursue lawsuits against individuals for a crime which cannot even be established to exist.

If there is no body, one cannot be convicted of Murder One. Yet Jesse Jordan and the Princeton Four were charged with "killing the music industry," and gave their life savings in settlement fines, money which was to have paid for their studies. Perhaps it would be more accurate to say that the RIAA are obtaining wrongful convictions in their drive to sue students and now even children for compensation against monies that they cannot even prove they have lost.

A Closer Look at Liebowitz

And so to Exhibit One (PDF), the expert analysis of Prof. Liebowitz. Even from the start, concern should be raised regarding the sources Liebowitz has used to obtain his data. Rather than commissioning an independent survey, Liebowitz takes his data straight from the RIAA (p.4), an institution which he himself admits manipulates sales figures to show "the picture they wish to portray about the conditions of the industry." However, Liebowitz does present the various biases that may be read into this data.

Liebowitz attempts in his work to take an impartial position, his stated preference as an academic being "to let the data tell... what's actually happening." But despite these attempts, and his healthy skepticism of industry hype, he cannot escape letting his personal opinions show below the surface, and through the channels who commission his work. In letting the data "speak for itself," Liebowitz opens his statistics to many possible interpretations, yet he chooses to dismiss these quite plausible alternatives one by one through methods which are somewhat less than rigorous.

These arguments would be better strengthened by using a combination of both pure data and opinion polls, in the manner in which Tempo, an independent industry analysis group, has demonstrated [see link below]. This gives a more accurate picture of where sales may have been lost -- or indeed, gained -- rather than the conjecture which Liebowitz employs.

The RIAA's figures present a clear decline in sales over the period 1999 - 2002. Yet Liebowitz fails to mention Tempo's analysis demonstrating a growth in CD sales through MP3 file sharing, as users are introduced to new repertoire and eventually purchase the CD. Indeed CD sales did rise during Napster's first year, a phenomenon Liebowitz does attribute to increased exposure through downloading. It's noteworthy that his study does not include statistics on "legitimate" MP3 purchases, focusing only on the drop in CD sales during this period. Given that he later refers to "media change" and "librarying" (where consumers duplicate purchases in a new format) as being a source of revenue during the introduction of cassette tapes, it seems implausible that he could fail to incorporate data collected through the many licensed MP3 download services.

One of the key arguments in favour of MP3 file sharing is that such librarying does occur, when P2P users download new songs, decided that they like them and then buy the album. Dave Rowntree, drummer with UK "Britpop" band Blur, confirmed this in an interview with the BBC's Breakfast program. He claims that figures on the effect of music "piracy" given by the BPI (British Phonographic Institute, the British equivalent of the RIAA) show a definite bias, and that independent studies support the theory that downloaders, like home tapers before them, are using the service on a "try before you buy" basis. He feels that it is inaccurate for industry associations to tar downloaders with the same brush as professional CD duplicators, and that while allegations of links to organised crime may be true for professional "piracy" rings, it is somewhat far-fetched to apply the same claims to students swapping files.

Perhaps it is the RIAA themselves, through their latest offensive against music lovers, who are driving their own audience away. Any parent knows that the harder one continues to push and push in an effort to discipline an unruly child, the more they will rebel. It is better to invest trust, in the knowledge that it will produce tenfold returns in mutual trust and respect for the parent.

The RIAA might do well to take a lesson from child psychology in its dealings with customers. Such trust was demonstrated by Apple in the launch of iTunes, a system with minimal DRM which until recently preserved the ability to share files across a network. Customers came in their droves to buy 99 cent downloads from the Apple musical jukebox, but their confidence suffered a blow when the networking facility was removed in iTunes 4.0.1. And by punishing and pushing their customers instead of respecting their trust, Apple then suffered in precisely the way they wished to avoid--a hack was quickly compiled which restored file sharing to the iTunes platform.

If Steve Jobs had left iTunes the way it was, he may have continued to enjoy the respect of file-sharers who had turned to legitimate purchases in order to enjoy a high level of service and respect. Given that the service enjoyed one million sales in its first week of use, and continues to attract sales, it would be hard to say that consumers are unwilling to purchase music in a downloadable format when the price is right and the level of user trust is seen to be reasonable.

The Album is not Dead

Ted Cohen, senior vice president of New Media for EMI, brokers of the iTunes deal, attributes iTunes' simplicity as a key factor in its success where others have failed. The system is integrated into the iTunes player, and is easy to use: "It really is actually the first application that is easier than Limewire or KaZaA". If it is easier to use a licensed system than to share unlicensed files, perhaps such positive reinforcement is a better avenue for the record labels to pursue, rather than continuing to provoke the anger of their own customers.

I asked Cohen about the implications that digital downloads may be having on album sales, as the Red Hot Chilli Peppers and Metallica both have shunned iTunes in the opinion that it is "destroying the album." Cohen points to figures on Coldplay's album and single sales from the first few weeks of iTunes' availability; "10 of the top 50 tracks were Coldplay. But the album wasn't in the top 50. The third week, the album moved into the top 10, and half of the tracks moved out of the top 50. So people saw how many good songs there were, and that its more than "there's only two good songs on the album" routine, and they started saying, "I'm going to buy the album instead of buying the single." And so the album went top 5, and the single tracks moved out of the top 25."

iTunes' capacity for tracking customer purchases is providing a more accurate picture of what the listening public want, and demonstrates clearly that the album is not dead. In fact, Cohen is exploring the possible extension of the iTunes service to include playlist compilations and richer media including background information on the band and their discography. He describes the current service as a musical "Quick Mart" : "People don't spend a lot of time in it because its pretty much, 'here are the top 100 songs, which one do you want to buy?'" By extending the service to a fuller shopping experience, Cohen and Jobs will be creating further reasons for users to choose a licensed service.

Back to Liebowitz

Returning to Liebowitz, other possible causes are given for the current drop in album sales which may fit with the popularity of the iTunes service. Listening to music takes time "and higher incomes do not necessarily lead to greater amounts of free time"(p.14). Assessing possible changes in entertainment preferences may help to establish whether our modern, busy lives could be contributory, as we may be choosing to spend that precious time remaining meeting friends, rather than listening to music at home alone. Unfortunately Liebowitz cannot share any conclusions about this hypothesis as his assessment relies purely on spending data.

Another suggestion which Liebowitz explores is that, given a limited income, consumers might be purchasing alternative items such as DVDs or videogames. "It is frequently pointed out, for example, that DVDs cost about the same as CDs, yet contain both pictures and sound, giving them an advantage that explains the shift away from CD sales" (p.20). But Liebowitz chooses to dismiss this possibility by comparing box office sales with CD sales, despite the obvious differences between cinema-going and home viewing.

It emerges from this study that, while Liebowitz might be a respected economist and academic, his logic leaves a lot to be desired. Attending entertainment events cannot be considered as parallel to DVD or CD purchases. Even using Liebowitz's own argument, with the pressures of modern life, we have increasingly less time to listen, and so it follows that purchasing the DVD might be favoured over time-consuming visits to the cinema.

Also, since Liebowitz fails to compare statistics of cinema attendance with purchase of the same title, he cannot accurately assess whether librarying is taking place alongside cinema-going, or whether DVDs are being purchased as an alternative. It would be better to collect such data through an independent market survey questionnaire. Indeed when Liebowitz quotes from a study by Veronis, Suhler & Associates on time spent by individuals on entertainment, his argument begins to hold water:

Published estimates indicate that listening to recorded music took approximately 45 minutes of a person's time per day, whereas going to movies took 2 minutes, watching prerecorded movies took 9 minutes, and playing videogames took 7 minutes. I do not believe it would be reasonable to argue, at the low time- levels of usage for these non- music activities, that changes in movie attendance, DVD usage, or videogames usage, for the population as a whole, could be responsible for more than a small portion of the changes in album sales discussed below (p.21).

By far the most interesting part of Liebowitz's study was the section "Has the Music Changed?" (p22), where Liebowitz considers the question "if mp3s are not hurting the record industry, what is?" and the possibility which is being argued by many professional musicians, that too much attention is being focused on teen-pop bands and the lowest common denominator at the expense of musical quality. While such a focus is often explained as a form of rationalisation in current economic conditions, it is widely believed to have backfired as sugar-sweet pop caters to the teen audience but is driving the adult listener away.

Liebowitz's analysis of this critical question again disappoints, as his method involves a comparison of concert revenues and record sales in a similar manner to his comparison of cinema-going and DVD purchase. He notes that "Interestingly, the years 2000 and 2001 had the largest real increases in concert revenues, the same time the record industry was experiencing unusually large decreases in revenues" (p.21). This is quite plausible when one stops to consider that those bands who are popular as live acts are rarely the same bands who feature in the Top 10.

In fact, many of the sugar-sweet teen groups are incapable of actual live performance, as their singing is usually manufactured through Antares Auto-Tune. Hearing the Spice Girls attempt to sing without accompaniment in a live television interview is enormously revealing. If a survey of concert attendance was to be conducted, it might in fact demonstrate that customers are finding a new interest in live music as an escape from the pap fed to them from the Top 40 charts. Or even that music downloaders are inspired to attend concerts of their favourite bands after hearing rare live MP3s circulated on the Web. There are, as ever, many possible explanations which point to the role of MP3 file sharing in promoting live and recorded music sales, none of which can realistically be found through examination of data alone.

Incomplete data also causes Liebowitz's arguments to suffer. While he makes a reasonable analysis of the decline in radio listening, and concludes that this fails to support disinterest in Top 40 music as the decline occurs across the board of musical styles and age groups, he has not taken into account changes in format of radio listening from FM radio to streaming media. Liebowitz skips digital format statistics entirely, which is an appalling oversight given that his study was written in the month that iTunes was launched. Perhaps his source is at fault - the RIAA would hardly wish to release figures supporting MP3 as a promotional tool, but they do have access to statistics on purchased MP3s from their own member services. Licensed digital media formats ought to play a vital role in any study which attempts to assess whether MP3 will "annihilate the music industry".

More convincing are the statistics Liebowitz quotes from the "Fine Report" given by the RIAA in the Napster case. Bias must obviously be kept in mind, but the report does indeed seem to suggest that the age groups 15-19 and 20-24 show a decline in CD sales in correlation with increased mp3 usage. Yet various factors should be born in mind, not the least of which being the lack of inclusion of licensed MP3 download statistics. Rising university fees and living costs over the last few years should also be examined, as students may lack the income to purchase the same amount of CDs as in previous years. This could support their apparent increase in MP3 usage. But is this really "killing the record industry?" If a student no longer has the income to purchase CDs, sharing an MP3 file cannot reasonably be considered for statistical purposes as a lost sale.

In considering the Fine Report, Liebowitz cautions the reader "to avoid placing too much weight on this analysis since these data are based on surveys, which are often quite unreliable" (p.24). I would suggest that it is in fact his own aversion to surveys which have caused holes to appear in his logic. But again, even the Fine Report does not present a full enough picture of patterns of student spending, presenting only two factors and omitting opinion polls. Returning to the analogy of criminal law, which is increasingly becoming the domain of file-sharing cases, one cannot convict beyond reasonable doubt without considering all the evidence. Insubstantial evidence is often good reason for a judge to throw a murder case out of court. And so, if we are to attempt to try MP3 for "murder" of an industry, we need to have at our fingertips every detail of forensic evidence.


Stan Liebowitz's study is indeed a thorough examination of the situation, perhaps more so than any other presently available. But despite the strength of his conclusions, when one examines the deep flaws in his logic, and the incomplete or biased nature of the evidence he presents, it is impossible to consider his study as admissible evidence in the trial of MP3 file sharing.

Particularly disturbing is his omission of iTunes and other licensed digital services, which may well give a fuller picture of consumer behaviour. Market factors are subject to sudden change and chaotic forces, and it is not normally considered reasonable to assess an economy to be in decline after only two or three years data. It is simply far too early to tell whether the drop in record sales over the period 1999-2002 is simply a blip on the screen, or evidence of further recession to come. Nor is it possible to assess reliably, without survey data, whether consumers are indeed copying instead of purchasing, or whether other factors, such as lack of income and the narrowing of musical style in teen-pop chart records have caused people to look elsewhere for their sources of entertainment.

If, as can be seen from this analysis, the evidence is indeed inconclusive, then it is unjust for the RIAA to move to pursue lawsuits against individuals without burden of proof. Instead, the industry should be taking a leaf out of iTunes' book and building on the success of fully-featured download services, which can offer better facilities, greater quality and more reliable connections as an incentive for their prodigal children to return to the fold.


Miriam Rainsford is a composer, singer and songwriter in classical, electroacoustic and underground dance music.

Return to the

Copyright © 2009 O'Reilly Media, Inc.