Subject: Re: How FSBs and liberal economists can be allies [was: universities struggling to avoid making money]
From: "Stephen J. Turnbull" <>
Date: Sat, 26 Oct 2002 18:12:53 +0900

Jeez, this was long.  I've cut it by about 50%, but it's _still_
_long_.  I've rearranged it to present the economic arguments first,
followed by the political/philosphical/rhetorical strategy stuff.

Besides chopping up context, my statements are _really_ terse.  Please
interpret this as my respect for other readers, rather than any
disrespect for Ben or his ideas.  I've tried hard not to take his
words out of context, with a fair amount of success (I hope) because
much is clarifying my own post.  My apologies for remaining failures.

Executive Summary:

1.  The unknown rights-holder ("10 man-hours of work for a transaction
worth a few dollars") problem is debunked from pragmatic and
theoretical viewpoints.  (First 100 lines.)

2.  Various and sundry points debated with Ben.  (Sorry, but that's the
most specific statement that's still accurate.)

***** Economics *****

>>>>> "Ben" == Benjamin J Tilly <" <>> writes:

    Ben> And indeed, [free contracting] is the standard economic
    Ben> answer to how to balance concerns.  But that doesn't work
    Ben> when it takes 10 man-hours _just_ to figure out who the right
    Ben> person is to try to negotiate a contract with!

Do you suppose that that 10 man-hours is a _permanent_ feature of the
IP system?  Heavens, no!  Consider Kinko's "Professor Publishing,"
which performs _exactly_ this service for professors who wish to
create a unique compilation of articles for a class.  I bet Kinko's
has it down to about 15 minutes for a whole one-page reading list,
mostly low-skill keypunching and envelope-stuffing.

Of course the "technology" used by Kinko's to serve me at very low
cost will not serve Kinko's to solve the "unknown owner" problem, but
to insist it's insoluble at higher, but still reasonable, cost is
untenable.  Professor Publishing didn't exist 15 years ago, although
both Kinko's and fairly cheap Xeroxing did.  With near certainty, 15
years from now, _if_ this is a big enough problem to be worth solving,
there will be similarly efficient technical solutions to the problem.
And of course it will just get easier and cheaper on average as the
"bit trail" gets broader as the works in question become digital.

Or consider the _exactly analogous_ "title insurance" problem in real
estate.  Something similar could be set up for "old" IP.  The
transactions costs are proportionately higher compared to the value of
the individual transaction, it is true, but probable losses are by the
same token much lower.  Premia would probably be reasonable.  More to
the point, they'd be line items in the proposal to the NEH.

Of course, the transition period involves (possibly large in
aggregate) losses from the "unregistered old IP" problem.  But that
doesn't justify implementing the "wrong" IP system permanently!

At best, it justifies an implementation where "old" copyrights that
are not registered with an appropriate public access, publically
funded (no proprietary Westlaw, please) database by 11:59 pm December
31, 2005 UTC will revert to a system for assessing royalties _ex post_
on a scale fixed by legislation (ie, not allowing unregistered works
to extort royalties ex post from compilation editors), and thence to
public domain at their existing expiration.  It is the owner's
responsibility to file the registration.  Until that date, newly
registered "old" works are subject to an arbitration system, with the
legislated scale to be given strong weight in the assessment of "ex
post" royalties.

I'm not saying that this solves all the problems, but the ease with
which I came up with the three proposals above makes me wonder whether
the radical, permanent changes often demanded by FSBers are really
necessary.  And I'm _dead certain_ that Friedman & Co. are well within
the bounds of reason to say the burden of proof is still on us to show
that "free contracting" isn't enough.

    Ben> I disagree that [transaction cost] is a second-order issue
    Ben> when you spend 10 man-hours of work for a transaction worth
    Ben> at most a few dollars.

First, you're out of context: I explained under what circumstances it
would be second-order, and why that is the right way to think about it
if we want to get and keep Friedman & Co. on our side.

Second, if the example transaction is worth a few dollars, and the
transaction cost is ten man-hours, you don't need the former Senator
from Wisconsin to give it a Golden Fleece Award.  Some things just
aren't worth doing is the obvious answer in many such circumstances.

But obviously you misspoke yourself.  What you presumably mean is that
the _social_ value is "high," but the compilation project only has a
few dollars per work budgeted for compensating owners and transaction
costs.  Thus a net-positive-value task is blocked.  The "rights"
answers to that are

(a) this is a valuable property, and you are trying to cheat the
    current owner (who should get up to `social value - transactions
    cost', depending on bargaining skills), or

(b) the NEH underfunded the project; too bad.

These are valid economic answers as well, but the Jeffersonian
economist sees a third way, namely adjusting the property rights
regime to reduce transactions cost, as you propose.

But the project is obviously socially marginal: you and I have already
spent 10 man-hours on this thread; if the copyright is worth less than
this thread, who cares?!  Is it really worth a revolution in the
country's and the world's IP regime to make sure that some oddball's
flames to U. S. Grant get digitized for the sake of "completeness"?

Even if you add up _one million_ such $100 wins, you're still an order
of magnitude away from "real money".  Can't the whole "Grant Corpus"
project wait until Kinko Copyright Insurance gets its act together?

The only rhetorical question in the preceding 2 paragraphs is marked
with`?!'.  I don't mean to imply I know the answers to the others, if
rephrased for reasonable examples, only that the implication of your
example is by no means obvious.  My responses would be, though, to any
professional microeconomist.  (May you never run into George Stigler!)
We need to deal with these arguments; they will be made.

"Stephen J. Turnbull" <> wrote:

    >> [H]istorically new technologies like software itself [...] 
    >> argue for _weakening_ IP.

    Ben> You speak as someone too familiar with software development.

I'm specifically referring to the fact that software can be developed
and disseminated without being associated with large concentrations of
capital, viz the GNU system.  Furthermore, the "firmware-ization" of
much hardware may hold similar promise for traditional goods.  This
implies that the franchising of monopolies, justifiable only for the
purpose of compensating capital for innovation, loses (some of) its
_economic_ justification.  Thus, weakening is appropriate.

    Ben> [New technology] has increased [...]  the promise of
    Ben> pervasive enforcement measures that people can imagine would
    Ben> be virtually unseen.

This is not a _justification_ for strengthened IP.  It merely suggests
that some of the technical and ethical difficulties would be reduced
_if_ there were another justification for strengthened IP (eg, the
alleged natural right of creators).

    Ben> You sometimes guarantee freedom best by simply saying that
    Ben> each person owns the right to exercise that freedom, rather
    Ben> than by trying to cut it up into many ownerships and then
    Ben> convince people to sell it to the right persons at the right
    Ben> time...

But that's the beauty of the property rights scheme.  The invisible
hand _works_.  There's no need to "convince," just leave them alone.
They'll figure it out for themselves, and cheaply, too.

At least, it works better than the commons.  The Enclosure Movement
proved that.  Yes, the Enclosure Movement screwed the smallholders and
peasants, and most never recovered.  But there's no question that it
made life much better for pretty much everybody else by making food
production dramatically cheaper.  The dire implication for FSBs is
obvious, but ... omelets and eggs, you know.

Now you're going to say "but software is different" because it's
non-rival and generates network externalities, unlike pastureland.
Yes---but we have only two decades real experience with _any_ good
like software.  The other non-rival goods we know about are the
so-called public goods that "justify" monopoly government.  It seems
to me _more_ likely that free software will suffer from _both_ the
tragedy of the commons _and_ the free rider problem than that it will
suffer from _neither_.  Of course, either problem by itself is enough
to suggest IP as a "second-best" solution (ie, not as good as possible
to a benevolent omniscient central planner, but the best possible in
view of the economic incentive constraints).

    Ben> [The argument from "blocking innovations"] is an argument
    Ben> against patents, not copyright.  Patents have exactly this
    Ben> effect.  But re-inventing a way to do something that you have
    Ben> seen work is far easier than figuring out how it could work
    Ben> in the first place.  Therefore when only copyright is on the
    Ben> table, all innovations are enabling.

Only if you use the definition of "innovation" used by the PTO.

Economically "innovation" is basically equivalent to "marketing a new
product", and the compilations and comprehensive corpuses that you
"can't get out of your head" count as innovations under that
definition, although by definition they themselves contain nothing
new.  These are blocked by the "component innovations" (individual
shorter works).  [Note: I don't mean to deny the usefulness of the
definitions of "innovation" used by the PTO or by FSBers.  I am saying
that when talking to economists, we should use their lingo.]

    >> FSB is historically unprecedented because software makes it
    >> possible for uncoordinated individuals to have big economic
    >> effects.  But this is not recognized by either economists,
    >> lawyers, or political scientists to date.

    Ben> I find that there are more historical precedents than you are
    Ben> admitting.

Your analogies are inexact in important ways.  This is not the time or
place to go into detail.  The most important is that U.S.-sized impact
in any given traditional market, _including_ those with sizable
network externalities, requires a proportional-to-U.S.-sized labor
force.  But there is only one Linus, and maybe a few thousand in total
who got the famous letter from Red Hat, and there's no reason to
suppose that getting a China-sized effect out of Linux requires any
new Linux developers (except for I18N, which will also be a fixed cost
unrelated to the population of China), just more "traditional"
communication and distribution networks in China.

In this sense, the only software-like goods I know of are pure
consumption goods: music and books.  These do not generate the same
kind of economic leverage that a productive asset like software does.
And if you try to turn books into a productive asset, you get
"education".  Oops....

    Ben> Furthermore if I am right that networks are a good place to
    Ben> look at for models of how software works,

You're almost certainly not.  If you were, the same system of property
rights that works tolerably well under the circumstances for telephone
calls and railroad trips would work for software.

    Ben> I believe that there is a close parallel with free software.
    Ben> Making it free to get and modify the software drops
    Ben> transaction costs.

Unlikely.  Transactions costs drop, yes.  But compare academic
publishing, a much more realistic analogy than commodity services like
telephone communications and rail transport.  Software is not a
commodity, by definition.

    Ben> But that [free software] can be a viable choice says that
    Ben> economists should think again, long and hard, about the
    Ben> importance of transaction costs.

Who says it is viable?  Tom Lord claims to be starving.
and sourceXchange are both defunct.  Richard Stallman wants free
software to be tax-supported.  Red Hat and SuSE are based partly on
reducing the transactions costs of software that's already 100% free.
And Perry Metzger calls it a "lifestyle business"!

Note that _I_ do not deny its viability.  However, I just don't see
enough evidence for viability to justify demanding a paradigm shift
from Nobel-Prize-winning economists.

***** Political Philosophy *****

    >> (1) Friedman & Co. are _liberals_ first, economists second, and

    Ben> The word "liberal" is used in so many ways

The British one.  "Libertarian" in American parlance is close enough.

    Ben> You sometimes guarantee freedom best by simply saying that
    Ben> each person owns the right to exercise that freedom[...].

Freedom unlimited by others' rights degenerates to might makes right.
Cf. Karel van Wolferen, _The Enigma of Japanese Power_.  It can work,
as Japan proves, but most Japanese are _significantly_ less free in
many ways that are important to Anglos.  I think it would chafe
especially on those who espouse the FSB ethos.

    Ben> [I]f there is a natural right to intellectual property, then
    Ben> [you] have the right to [...] dispose of it in a way that
    Ben> supports Free Software if you can find a way to do that.

We really don't want FSBs to depend on noblesse oblige; there's damn
little noblesse, and less oblige, in the business world.  If there is
a natural right, I see no excuse for (corporate) businesses to give up
the option---they should decide whether to exercise it on a case by
case basis, instead.  Then free software simply becomes a pole of a
continuum of acceptable licenses for business---surely not the default.

    Ben> Secondly what I consider important are people who do not know
    Ben> what belief to hold.  There are reasonable arguments for
    Ben> treating IP as property, and someone who has just been
    Ben> exposed to them is unlikely to react well to your casually
    Ben> dismissing them.

I don't dismiss arguments, based on reason or otherwise, when I run
into someone who holds them.  Eg, my whole purpose here is to focus on
the "social construct" rationale that you and I share, although I lean
far more to the "IP is OK" side than you do.  (I have hope you'll
convince me of my error, but ... well, please try harder!)

But on FSB, I see no present benefit to discussing the natural rights
theory of IP beyond noting that we cannot refute people who believe
it.  On FSB, few, if any, hold that theory.  That theory, and the
Gordon Gecko argument it may conceal, are all that I dismissed.

    Ben> Instead you should point out that there are reasonable
    Ben> arguments for not treating IP as property,

But there are _no such arguments_ if IP is a natural right, to go
along with life, liberty, and the pursuit of toys to die for.  All we
can do is "damage limitation" with arguments like a "moral obligation
to put that IP to socially beneficial use" and the "transaction cost
economies of open source."

    Ben> The problem is that the status quo is a compromise between
    Ben> two extremes.

They are not "extremes".  Calling the principles that the status quo
compromises "extremes" compounds the persuasion problem.  They are
different dimensions, not bounds along a single dimension.  They might
be social goals (securing rewards to the productive vs. distributing
gains that derive from being part of a society widely in that society)
to be balanced.  Alternatively, incompatible rights one or both of
which must be compromised (preferably with the rights-holders'
consent).  Either way, gains from negotiation are much easier to
realize in a multidimensional framework.

    Ben> As
    Ben> [Jefferson] shows, the other extreme [from the natural rights of
    Ben> inventors position] has had competent representation in the past.

Which "other extreme" (ie, dimension)?  There are several.  There is
Proudhon: property is theft.  There is Marx: property belongs to
society.  There is Jefferson: property is a social construct, based on
occupation of the property.  There are others.

Unfortunately, Jefferson's argument against property in inventions is
vulnerable to the same ASP loophole that the GPL is.  (Actually, any
"process technology" demonstrates this loophole, but ASPs make it
glaringly obvious.)  Due to the Internet, it _is_ possible to "occupy
and not relinquish" certain works of software _while exploiting them
commercially_, in a way it is not possible to "occupy" a book or a
hardware invention.  Then using exactly his analogy we may generalize
to perpetual property in such software, just as property in land is
perpetual.  Extension to all software and further to all intellectual
capital is not automatic, but it is no longer so implausible as
Jefferson would have you believe.

    >> [H]ow do you propose to "address" the moral argument for
    >> perpetual IP?

    Ben> By pointing out the moral argument against binding each
    Ben> person with a thousand chains in perpetuity.

Specious. That is an argument against _all_ property.  Eg, in land.
Title to land must involve 100 million "bricks in The Wall" in the
U.S. alone.

You have to come back to the non-rivalry of intellectual assets, which
is an economic, not a moral, argument.  As such it cannot prevail
against a natural right of creators.

The only way to counter a right is with another right.  Ie, Stallman's
"right to share."

    Ben> Why can't both [Stallman and Justice Thompson[1]] be right?
    Ben> Each has laid out an argument saying that there is a
    Ben> legitimate interest in having things be this way.  Both
    Ben> arguments can be correct - the legitimate interest exists.

I'm drawing a distinction between rights and interests.  (You don't
seem to, or perhaps you don't admit that it matters.)  Rights are
either compatible, in which case there is no problem, or they
conflict.  The latter is the case here.  Both Stallman and Thompson
insist on the primacy of the right they assert---both cannot be
correct in that insistence.

    Ben> I know of no group that usually succeeds in figuring out what
    Ben> others care about, and darned few of them who figure out what
    Ben> they themselves care about.

FSB has an excellent track record IMO.  That's the only possible
reason why I'm tolerated here.  :->

[1] The author of the dissenting opinion in Wheaton vs. Peters,,
the strongest statement of the natural rights position I know on the net.

Institute of Policy and Planning Sciences
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
 My nostalgia for Icon makes me forget about any of the bad things.  I don't
have much nostalgia for Perl, so its faults I remember.  Scott Gilbert