The EFF Offers a Better Way Forward on File Sharing

An Metet anmetet at
Wed Feb 25 22:00:28 PST 2004

Invisiblog is a service which uses anonymous remailers and gpg-signed
messages to allow bloggers to make their contributions anonymously.
The following essay is taken from the Unlimited Freedom blog: .
It discusses the EFF white paper on file sharing at: .

Thanks to Derek Slater for his pointer to an interesting new proposal
for handling music file sharing.

I've criticized the EFF in the past, but their new white paper is
definitely worth a read. "A Better Way Forward: Voluntary Collective
Licensing of Music File Sharing" has some great ideas and would be a
terrific solution to the file sharing problem if it could be made to
work. I do see some problems, but perhaps they can be solved, and in
any case the concept looks highly promising.

Voluntary Collective Licensing (VCL), as the EFF explains, is the system
used by radio broadcasters and other public performers to compensate
the music copyright owners. The broadcasters pay a fee to ASCAP and
other collection societies to gain the rights to play the music;
ASCAP then divides it up and pays it out to the artists. Having just
a few collective organizations like ASCAP makes it easy for the radio
broadcasters to license all the music they play.

The same idea in the P2P world would mean that users would pay a fixed
fee to the VCL agency, which the EFF suggests might be a modest five
dollars a month. This would then give them the rights to freely download
and share all the songs belonging to the artists (or record companies)
which had joined the VCL collective. The fees would be divided among
the artists using a combination of something like Nielsen ratings
plus monitoring public file sharing activity. Users would be free of
the threat of lawsuits or of the record companies using the various
technical countermeasures which are continually being developed and
deployed. And the record companies would be making money from the P2P
filesharing phenomenon which shows no sign of stopping.

My initial reaction to this proposal doesn't even need to get past the
first word. It's voluntary? I'm for it.

Look up at the top of this blog and you see what I'm about. Voluntary
arrangements are the essence of freedom as I see it. This is why I support
trusted computing and DRM, as well as encryption and anonymity. All of
these are technologies which people can use voluntarily to interact with
each other in new and powerful ways. Giving people access to the maximal
set of options in their lives is part of what I call Unlimited Freedom.

And the EFF proposal truly is voluntary. Artists would only join the
collective if they chose to do so; but staying outside would force the
artists to pursue expensive legal and technical warfare against file
sharers. Users would only sign up and pay the fees if they wanted to; but
refusing to pay would expose them to the hassles and hardship of illegal
downloading (not to mention the ethical conflicts). The EFF points out
that current anti-trust laws might interfere with the formation of a VCL
agency, but from the libertarian perspective I am of course entirely in
favor of weakening antitrust.

So before going on to criticize the proposal, I want to emphasize again
how pleased I am that the EFF has come up with an idea which respects
principles of freedom and choice for all parties. It makes me proud that
I have joined EFF and have supported it with contributions over the years.

However, while the goals of the VCL are laudable and the methods are
highly ethical, I do see practical difficulties which the EFF seems
overly optimistic about. These are fundamentally due to the loss of
price information which is such a crucial part of a normal market
structure. With music downloads disconnected from costs, signals are
no longer automatically provided to the producers to show how consumer
demand is being distributed. While the plan tries to provide proxies
and other mechanisms to provide this information, it is extremely hard
to do so reliably.

More concretely, there are three obvious problems with the idea. First,
people are still going to freeload, and it's possible that the greater
availability of legal P2P filesharing will be exploitable by illegal file
traders to make their jobs that much easier. Second, setting the size
of the per-month fees is going to be a difficult and risky calculation,
and it's all too likely that the artists will initially err on the side
of caution, setting them too high for the plan to succeed. And third,
finding a reliable way to divide up the proceeds which accurately reflects
listener preferences will be difficult, since cheaters will have financial
motivation to distort the results.

As far as illegal file sharing, the EFF very optimistically and
idealistically predicts that if the license fee is in the modest range
they suggest, $5/month, users will want to be honest and enjoy the ethical
and practical gains from getting their music legally and honorably. I
have real admiration for the EFF's high opinion of the file sharing
community. The EFF has worked with members of this group for years,
and they should have a good sense of how file traders feel.

But I'm sure that even the EFF would admit that there might well be a
significant percentage of people who will refuse to pay any amount for
the privilege of downloading music. I have certainly seen such sentiments
expressed in various online forums. Many people have come of age with
the expectation that music ought to be free, because that is all they
have known. They see the creation of music as an inexhaustable fountain
from which they can drink as often and as long as they wish. Now that
they have gotten used to having their music for free, they won't want
to go back to paying for it.

The existence of such people is going to make the whole plan more
expensive and difficult, in several ways. For one thing, the only thing
that will motivate people to stay with the program and not freeload
will be some kind of negative consequence if they leave. So the record
companies may have to continue their program of lawsuits and technical
countermeasures, even once the VCL plan is in place. A major potential
cost savings thus does not materialize. And even if the percentage of
freeloaders is small, the expense of the lawsuits may have to continue
to be large, in part because locating such people may be more difficult
now that they can hide among a much larger number of legal file sharers.

If the freeloaders can't be effectively deterred, then their existence
will serve as a constant temptation for legal members to quit the
program. We have all seen situations where a law which is ignored with
impunity by a minority eventually becomes less respected by everyone. The
same kind of gradual breakdown would be a constant threat to a VCL
program. It's not like radio, where there are only a few hundred or
thousand broadcasters. We're talking millions or potentially billions
of people now, and the enforcement lessons from radio are not applicable.

Another problem with the EFF plan is setting the size of the VCL fees. I
give the EFF enormous credit for resisting the temptation to put the
government in control of this aspect of the program. Politicizing this
issue would be disastrous. Letting the creators of the music (or their
agents) decide how much they are willing to let it go for is the only
free, fair and economically sound possibility.

Nevertheless, there are going to be significant practical difficulties
in getting a reasonable licensing fee. First, from the theoretical side,
it's not even clear what kind of fee we would see, if the VCL agency had
perfect knowledge. What would maximize their profits? As they increase
their fees, revenue per subscriber goes up, but they lose subscribers,
and face the costs described above to try to coax people back into the
program. But as further argued above, those enforcement costs may not
be able to be reduced below a threshold value, since people can always
leave the program and potentially still get access to music. It's a
complicated tradeoff.

It's often the practice, with new business models, to set initial fees
on the high side in order to reduce risk. Then the fees can be gradually
lowered as market demand increases. This provides for a gradual and
manageable growth rate. The VCL agency probably can't realistically sign
up and manage 100 million subscribers the first year. They might prefer
to set the fees high enough so that they have only a million subscribers
that first year, and then bring them down gradually. This will also give
them a sense of the shape of the demand curve, how many more subscribers
they get for a particular decrease in licensing fees.

For these reasons, I'd expect the fees to be quite a bit higher than the
EFF estimate of five dollars a month, especially at first. Hopefully they
would come down within a few years, and maybe the EFF goal would not be
too far off, eventually. But initially the plan will face criticism and
disillusionment if the fees are considerably higher than were initially
promised. As far as likely long-term licensing rates, I'd like to see
some analysis and modeling of this tradeoff which could shed more light
on where a good price point is likely to be found.

The third problem I listed is the difficulty of accurately figuring
out how the proceeds should be distributed to the artists. The EFF
report suggests copying the Nielsen ratings and having randomly selected
families allow their downloads to be monitored. This isn't a bad idea,
but the problem is that there are far more songs in common use than
TV shows. At any given time there are probably no more than 200 shows
tracked by the Nielsen ratings. But there must be tens or even hundreds
of thousands of different songs being downloaded in a typical day. That
means that the statistical precision necessary to fairly divide up the
revenues is going to be far more difficult to meet than for TV shows. The
truth is that Nielsen ratings have larger error bars than many people
realize, and once you divide the measurements a hundredfold or worse,
the statistical errors are likely to swamp the measured results. The
noise will overwhelm the signal.

The other idea the EFF suggests is to monitor file sharing activity,
as is done already by companies like Big Champagne. The problem with
this is that at present, no one has much incentive to try to manipulate
this data. But in a VCL system, these measurements will determine the
financial success or failure of artists and record companies. If the
system can be gamed, it will be. So far I haven't seen a proposal that
would both respect user privacy and simultaneously prevent people from
creating bogus download requests or using other tricks to inflate the
numbers for a desired artist or company. The EFF's confidence that these
two methods will work is not well justified at present.

In summary, the EFF's Voluntary Collective License proposal is innovative,
ethical, and idealistic. It respects the freedom of all parties involved
and tries to offer a path which everyone can walk to their mutual
benefit. There are still significant practical problems to be overcome,
but the goal is important and desirable enough that it is worthwhile
putting in effort to see if this system can succeed. Even if it ultimately
fails, VCL should be encouraged as a new, voluntary and creative option
for people to approach the difficult issues of file sharing today.

More information about the cypherpunks-legacy mailing list