[dvd-discuss] Thoughts on Posner and Landes (was Re: Eldred v. AshcroftAccepted ...)

["Arnold G. Reinhold" <reinhold(at)world.std.com>]

At 5:47 PM -0500 2/20/02, Scott A Crosby wrote:
>...
>Isn't 'progress' usually considered fulfilled when you maximize the
>economic output from a work. When the most money is made on a work? In
>that case, as per Posner and Ladner (I believe), it may be that that is
>true with these obscenely restrictive laws.
>
...


At microlenz' suggestion, I've read the Posner and Landes paper*  and 
I have lots of comments.  To begin, there is a lot in Posner and 
Landes that we can use to support our arguments.  For example, Posner 
and Landes:

   o agrees that more copyright protection increases the expense of 
creating new works.

   o warns against the danger of copyright protection leading to rent 
seeking (this is a technical term in economics referring to excess 
profits, not what you pay for your apartment).

   o emphasizes the importance of fair use.

   o urges economic analysis be performed in considering questions 
about application of copyright.

Proponents of stronger copyright protection cite Posner and Landes to 
claim that a reduction in the cost of copying should be matched by an 
increase in the level of copyright protection to maximize social 
good. Posner and Landes is taken particularly seriously because it 
appears to prove this mathematically.  Since everyone seems to agree 
that the Internet lowers the cost of copying, the proponents say any 
measure that increase copyright protection is justified.

The problem with the math in Posner and Landes is not that there are 
errors, but that it doesn't say anything quantitative at all.  The 
analysis is equivalent to saying lack of exercise is bad for you and 
too much exercise is bad for you, so there must be an optimal amount 
of exercise.  While that argument is commonsense, it can also be 
dressed up in a rigorous mathematical exposition (with a few 
assumptions, of course). But all that the math will not tell you how 
much exercise to do nor how to do it (running vs swimming vs golf vs 
lifting the channel-changer).

There is the old joke about the physicist who is hired to advise the 
dairy industry. After months of work he is ready to present his 
recommendations to the assembled industry leaders.  He gets up, walks 
to the blackboard and begins by saying "Assume a spherical cow." 
Posner and Landes is full of spherical cow assumptions. Perhaps the 
biggest is representing the entire corpus of copyright protection 
level by a single real variable, z in their notation. Other aspects 
of the creation, distribution and consumption of literary works are 
similarly condensed into a single variable or function.

Once set up in this manner and with the right assumptions on the 
slope of the functions, the result that z should go up as the copying 
costs go down is almost tautological. It adds nothing to the common 
sense analysis, logically identical to the exercise example, that 
when copying is very expensive, you don't need copyright protection 
and when copying is free, most people will do it rather than buy from 
the author, so there must be a an optimum that shifts with the cost 
of copying. Looked at more closely there are serious problems, 
however.

1.  Posner and Landes say nothing about which copyright protections 
should be increased.  It's all lumped into the spherical cow z. There 
are many aspects of copyright protection, including:

    o Duration of copyright protection
    o Level of penalties (actual damages, statutory damages, criminal)
    o Coverage
    o Requirements for renewal
    o Requirements for marking
    o Enforcement (e.g. FBI copyright budget level)
    o Extent of exemptions (fair use, compulsory licensing, etc.)
    o Standards of proof
    o Procedural issues (standing to sue, right to injunctive relief, etc.)
    o Ability to enforce rights in other countries
    o Assignment of court costs
    o Anti-circumvention laws


Many of these (duration, statutory damages) are already quantified. 
Others could be quantified (e.g. mean time to obtain a copyright 
injunction, mean cost of obtaining an injunction, mean damage award, 
FBI copyright expenditures in constant dollars, percentage of world 
population/GDP in countries that have signed a copyright treaty). It 
should be possible to build a mathematical model and evaluate the 
economic effectiveness of different protection measures. Posner and 
Landes would approve of a more fine grained analysis of these 
measures.

2. We have to look more critically at the notion that computers and 
the Internet merely reduces the cost-of-copying term in the Posner 
and Landes equations. The Internet is a broad technological 
revolution, much more so than other improvements in copying 
technology such as the Xerox machine or photo-offset printing. It is 
vital to realize that computer technology and the Internet affect 
*all* terms in the Posner and Landes equations, not just the cost of 
copying. This means that the effect of this technology on the 
copyright balance is not so obvious and cannot be deduced by the 
simple analysis of Posner and Landes.

For example, cost of authorship is also significantly reduced. 
Traditional typesetting is a thing of the past. Word processors are 
nearly universal and publishers expect manuscripts in machine 
readable form.  Scientific journals expect camera ready copy. 
Writers, editors, and production staff communicate via e-mail. 
Digital technology is also transforming motion picture production. 
Anything a director can imagine can now be rendered as convincing 
moving imagery.  Film itself is about to be supplanted by reusable 
digital media. Digital editing will soon be able to correct errors 
that, in the past, would have required reshooting a scene.

New technology reduces the cost of advertising. Consumers can obtain 
reviews and word of mouth recommendations almost instantly. Motion 
picture trailers -- "previews of coming attractions" are now 
downloaded over the Internet by avid fans.

The risk of publishing, which is a major justification for copyright 
protection, has been reduce by lower editing, production costs and 
increased advertising efficiency. On-demand short run printing 
significantly reduces up-front investment and inventory carrying 
costs, while at the same time meaning that books no longer have to go 
out of print.

In addition some aspects of copyright protection are enhanced by the 
new technology. The same search engines that consumers use to find 
unauthorized copies on-line can be used by publishers to track down 
these sources. Cease and desist letters (e.g. under 17 USC 512) can 
be and are generated automatically and delivered in seconds via 
e-mail.  E-mail can also be used to collect tips on unauthorized 
copying and distribution inexpensively from anywhere in he world, 
with rewards for information paid electronically.

The new technology also affects the costs of distribution. Motion 
pictures can now be distributed to theaters by fibre optic cables 
instead of bulky reels of film. Not only does this save money, but it 
will allow simultaneous release and shorten the window of opportunity 
for pirates. Movies realize a substantial of their revenue in the 
first few weekends. In colonial times, it took longer than that to 
ship a book from Philadelphia to Boston

3.  Posner and Landes is weak on the question of what would happen if 
there were no copyright protection. I am not suggesting that there 
should be none, but an understanding of what such a world would be 
like is vital for Posner and Landes' argument to hold. This is the 
analog of the "no exercise is bad for you" datum. If a copyright-free 
world works, then the case that there is some optimal level of 
copyright has to be proven by a different argument. One must show the 
the social benefit (W in Posner and Landes' notation) is greater with 
copyright than without it.

Posner and Landes dismisses the fact that works were created before 
the existence of copyright, citing the difficulty of copying prior to 
printing and the monopoly on printing in England prior to the first 
copyright law. They ignore the rest of the world, however. Copyright 
protection has become widespread only in the last century and many 
cultures have vibrant literary traditions despite its absence.

There are also some particular examples in living memory.

o  When the technology for recording a musical performance first was 
developed with the player piano, courts in England and the U.S. 
rejected the notion that piano rolls were subject to copyright.  Yet 
the player piano industry thrived until the invention of the 
phonograph.

o  In the early years of the computer industry, the legal status 
software copyright was in question (because software can be 
considered functional) and few attempts were made to enforce software 
copyrights, yet vast amounts of software was written. Arguably, 
stronger copyright and patent protection have slowed development in 
this field. Seen any killer apps lately?

o There is a whole class of works that are deemed functional and do 
not receive copyright or patent protection, such as building designs. 
Yet these fields thrive.

o  Some functional areas have received protection only recently, e.g. 
integrated circuit masks and ship designs, but they were filled with 
creative work before being so protected.

o  There is also the likelihood that very different forms of 
expression might flourish in a copyright-free world. A good example 
is the Talmud, which combines the work of numerous authors, 
representing hundreds of years of collective dialog, on a single 
page.  See http://www.ucalgary.ca/~elsegal/TalmudPage.html The 
Talmud's format anticipates the Internet and is arguably a better way 
of encapsulating knowledge than the shelf filled with volumes by 
individual authors that copyright laws demand.

o  My favorite example of a thriving copyright-free community is the 
law itself. Laws, legal decisions and forms, (and briefs ??) are not 
subject to copyright. Lawyers lift text from other lawyers with 
abandon.  Imagine the chaos that would ensue if all this material 
were subject to the strict copyright protection that many in the 
legal community wish to impose on the rest of society.

One can also question Posner and Landes' assumption that creation of 
new works to society is driven by expected market value. Maslow's 
Hierarchy of Needs 
http://chiron.valdosta.edu/whuitt/col/regsys/maslow.html is widely 
accepted by social scientists and is a better explainator of authors' 
behavior than Adam Smith. A proof case is poetry.  The Writers Market 
lists numerous outlets for publishing poetry. The payment offered is 
almost invariably one or two free copies of the volume that includes 
the authors poem. That's it. Yet these publications receive a flood 
of submissions.

Self help
In the absence of copyright, authors and publishers would take 
measures to protect their investment in creating works.  They might 
alter their distribution patterns, for example. Movies might be 
introduced simultaneously throughout the world, at least in the 
original language. DVD's might be held back a few months to insure 
that maximum revenue was derived from theatrical release.  Fixed 
media might be replaced entirely by direct transmission to home 
screens that were designed to preclude copying.  (Any studio concerns 
about losing post theatrical sale revenue should be tempered by the 
fact that the studios fought tooth and nail against this market 
developing in the first place during the Sony Betamax litigation. 
They would no doubt would have cited Posner and Landes if it had been 
written. Only a 5-4 Supreme Court majority saved them from 
themselves.)

Another important form of self help is increasing the information 
content (entropy) of works to make copying more difficult. DVDs can 
contain more background material, music videos, out-takes, 
biographical material on artists, etc. Eventually full three 
dimensional views can be provided allowing viewers to choose camera 
angles and even enter the action.  Increasing the content to 
compensate for the rise in consumer available bandwidth is a 
potential counter to the predictions of doom from Hollywood.

Similar techniques could be employed by the recording industry. Fans 
would buy a package that included music videos, fan material, videos 
of recording sessions, discounts to concerts, access to special fan 
fan sites  and even personalized messages generated by synthesizers 
programmed to the recording artists voice. In such a market place MP3 
files would be free advertising.

A free market in protection technology would develop.  The Content 
Scrambling System used in DVDs was broken by a high school student. 
Left on their own, market forces would insist on much better 
cryptographic designs. Economic doctrine says this process should be 
allowed to find an equilibrium and not be truncated by government 
fiat. Consumers could choose between works sold in formats that 
prevented copying and other works that were could be more freely 
copied.  Posner and Landes' analysis suggests the later set will be 
smaller or of poorer quality. Why not test that theory in the 
marketplace? The market would tell us if the value of the new works 
incentivized by copy protection was worth the cost.

4. Copyright threshold

There are already areas where copyright protection exists in law but 
is completely unenforceable.  Consider the question of jokes. An 
original joke is clearly a literary work and subject to copyright. 
However, with the advent of e-mail, good new jokes are copied and 
spread around the world within hours of first being told in 
performance. Yet comedians are still working (a few even making a 
living) and new jokes are still being told.

Posner and Landes does not distinguish between commercial copying and 
peer-to-peer transmission. Society has long tolerated non-commercial 
peer-to-peer transmission of works. People have always borrowed 
books. The shape of the modern book was dictated by the size of 
middle age saddle bags. Ben Franklin himself invented the public 
literary to institutionalize this practice.  Office walls are covered 
with xeroxed clippings of cartoons and articles.

Again there is an entropy/enforceability  tradeoff. One practical and 
compelling justification for fair use is that it avoids embarrassment 
from wide spread flouting of the law. The situation goes beyond fair 
use. One good example is the song "Happy Birthday to You." It is 
copyright 1935 by Mildred J. Hill and Patty Hill and controlled by 
ASCAP.  http://www.ascap.com/press/songs-122799.html  Anyone who has 
hired a performer to sing this tune at a child's birthday is 
presumably subject to $50,000 in statutory damages.

One could make a mathematical argument very similar to Posner and 
Landes' that there is an optimal threshold where cost of enforcement 
balances the gain to rights holders in short works. It may be, for 
example, that society will soon be unable to protect recorded 
performances of short musical works, no matter what laws are passed. 
The recording industry will either have to find something of greater 
entropy to sell or disappear.  If the choice is that or severe 
restrictions on new technologies' ability to let people exchange 
information on a peer to peer basis, with all the economic benefits 
that will bring, it is may be acceptable price. After all the 
recording industry has only existed since 1886. Like all industries, 
it must adapt to technological change or die.

5. Copyright may have been out balance before.

The PL paper assumes that the copyright balance was about right, but 
use only the most qualitative arguments to support that claim. It is 
entirely possible that the level of copyright protection was too high 
and the new technology has corrected that balance. Nothing in Posner 
and Landes speaks to this question. Even more likely is that the 
individual components of copyright protection are poorly balanced and 
need to be redressed in light of economic changes:

o  Works have a shorter economic life than, say 20 years ago.
o  There is greater variety in types of work (movies, songs, books, 
newspaper articles, ephemera,  computer software) yet one size is 
supposed to fit all.
o Greater social and economic mobility means there is increased 
danger of copyright orphans -- works where the cost of finding the 
copyright owner exceeds their worth.  Publishers merge or go out of 
business. Author die and their heirs can be difficult to locate 50 
years later.

6. Increasing cost of maintaining the copyright status quo

Attempts to maintain the copyright status quo in the face of 
technological change will require ever more Draconian measures:

Restrictions on publishing research results
Consumer bandwidth limits
Bans on unbreakable encryption
Compensatory taxes (q.v. American Home Recording Act)
Requirements for technical protection measures in all new computers
Bans on upgrading old computers
Requirements that personal computers be sealed against tampering 
(q.v. scanner laws)
Confiscation of older computers
Restriction on availability of computer development software (q.v. 
chemical supply regulations)
Making large portions of the population felons

Such measures will burden the courts with the enforcement of 
unpopular laws and carry a social cost that is not figured into the 
Posner and Landes equations.

Conclusion

There is no support in Posner and Landes for expanded copyright laws 
that goes beyond the plain English argument that since copying is 
easier, copyright holders should get better protection. But at least 
the plain English version allows the rejoinders "but why this 
particular measure?" and "don't they have too much already?" 
Translating broad, qualitative arguments into mathematical equations 
should carry no more weight than translating them into Greek. 
Unfortunately, integral signs do seem to impress the uninitiated.


Arnold Reinhold


* "An Economic Analysis of Copyright Law," William M. Landes and 
Richard A. Posner, Journal of Legal Studies, Vol. XVIII p. 325 (June 
1989), University of Chicago Press