Subject: Re: Tom W. Bell paper
From: <stephen@xemacs.org>
Date: Sat, 2 Sep 2006 16:24:37 +0900

Don Marti writes:

 > So you invent a lead-to-gold machine and start
 > buying lead and selling gold.  Just thinking common
 > sense here, you'd want people to be able to do that
 > even without a SPEX.  I understand the point of
 > not allowing people to make a trade that breaches
 > their duty of confidentiality to an employer, but

Huh?  That is not what regulation of insider trading is about.
Confidentiality is dealt with by contract law, it is not an issue for
securities markets.  The problem with insider trading is that what
should be a bet is in fact known to be a sure thing by the
counterparty.  They win, you lose, you have no chance.  The optimal
thing for you to do is leave the market.  If everybody leaves the
market, an extreme case which can occur with plausible parameter
configurations, then the developer is left with zero opportunity for
funding via these markets.

 > not the point of calling something "insider" trading
 > when it's information that the trader discovered or
 > invented independently.  Don't we want the laws to
 > provide incentives for people to discover and create? 

Yes.  The point of talking about insider trading is two-fold.  First,
it requires legal changes before it will be practical for inventors to
trade in claims.  This is a substantial hurdle.  Second, even with the
legal changes, these markets may suffer from a severe "market for
lemons" problem, which will result in underfunding of desirable
projects.

> > Whenever one sells "invention of a public key algorithm" shares,
 > > people will assess that he knows something they don't.  This will
 > > force the price down, perhaps to zero if he is a recognized expert in
 > > the field.  So "players" in the crypto field will engage dummies, etc.,
 > > to hide the identity of the actual traders.  This is not going to be a
 > > very transparent market.
 > 
 > So you'll have an incentive to use a discreet broker
 > who also handles the accounts of many people on the
 > other side of the deal from you.  Seems doable enough.

Sure.  But lack of transparency increases the risk, and reduces the
informational content of the market---but that's exactly the opposite
of what these markets are supposed to be about.

I went and took a closer look at Bell and his paper.  First off, he's
at George Mason; these guys are by and large so right-wing that even
Chicago economists consider them wackos.  Of course, James Buchanen
who recently won the Nobel Memorial Prize in Economics (deservedly) is
one of those wackos---So this doesn't necessary affect the quality of
the research, but you should remember that Bell is very likely
politically, not academically, motivated.

In fact, he more or less admits this.  The paper *assumes* benefits to
science and technology development, and then goes on to propose that
in order to achieve these benefits we "must" free these markets from
the burden of government regulation.

This agenda suggests two issues, which he does not address.  First,
there are already many play-money and toy real-money prediction
markets which do function very well as information aggregators.  One
wonders, however, if they would do so well if there was "real money"
to be made.  With more "reputation" than money at stake, people will
do a good job of playing according to the spirit of the game.  But
when money comes to the fore, some players (who care little about
their scientific reputation) will start taking the letter very
seriously, especially if it offers loopholes they can exploit.  There
is no evidence on offer that I can see that says these markets will do
a better job of anything if they have larger turnover.

Second, as far as I can tell from background reading of Bell's
citations to economists (the George Mason website is down or
something, I couldn't access any of his or his colleagues' other
papers), not to mention his own Table 1, nobody in the field takes the
"capital provision" role of prediction markets seriously.  It's all
about information aggregation.  The people who know the subject best
make money by guessing the tenor (not the content) of future news and
betting accordingly, and price becomes a (good) estimate of
probability.  Nobody talks about the benefits to society that come
from rich traders.

That doesn't mean a capital provision role can't happen, but it does
mean that at present advocating it is rather like investing in
swampland in Death Valley.

 > The exchange would have to cover the drafting of the
 > initial spec for the market, but some markets would be
 > launched once and never see a claim.  (The inventor
 > could be required to fund the testing of his or her
 > own invention, as claimants to The Amazing Randi's
 > prize are now.)

It doesn't matter who pays, it's still a transactions cost that
inhibits trade.

 > > You're also ignoring the costs of the
 > > expert consultants who will be demanded by investors for every new
 > > contract, and who will have to be on retainer (or salary) to evaluate
 > > every piece of technical news.  There are also substantial costs of
 > > making new markets per se.  The majority of these markets will be
 > > horrendously illiquid, which will greatly increase risk, and
 > > implicitly transactions costs.  Not to mention the opportunity cost of
 > > all those smart people doing well- and regularly-paid referee and
 > > consulting work rather than risky R&D.

 > Yes, which is the same problem we have now with the
 > patent system.   An advantage of the SPEX system here
 > is that multiple exchanges could compete to lower
 > costs, while a Patent Office has to be a government
 > monopoly.

But it's *not* the same problem.  Currently, each currently traded
company (ie, I'm talking about additional stock issues, not VC
funding) specializing in, say, hard drive technology must devote some
of its expertise that could go into R&D to dealing with patent issues.
That's more or less a wash, it will be true in the SPEX system (for
dealing with claim issues) too.  But in the current system the
investors generally get to hear glowing predictions of future
developments and announcements of actual patents.  They do not get to
see actual R&D interim results or the patent applications and evaluate
them, so they need pay no experts on patent applications.  But in the
SPEX system, not only do investors need expertise on the specific
claim, they need it on a daily basis (or at least every time "news"
occurs---but that assumes they can recognize important "news" without
their technical consultants).  This is a potentially big transactions
cost (depending on the content of the claim) that the patent system
does not have to pay at all.  The reason is that we know what a patent
does, so once you've learned that, you don't need an outside expert.
But these claims are free-form, and while there are likely to be
standard forms, there are also likely to be lots of them, and for new
ones to arise with some frequency.

These trends can be easily seen in the markets for financial
derivatives, and they're very expensive in absolute terms; it's just
that when a market sees $100 billion or so in annual transactions, a
tiny spread means you can afford thousands of analysts.

As for being a government monopoly being a "disadvantage" of a Patent
Office, yes, and no.  The patent system turns on the award of
monopolies, and that has to be a monopoly.  The SPEX system need not
be a monopoly because there are no monopolies---in fact, the market is
zero sum.  But that's a *big* problem for our purposes, because we
want to be able to extract funding for R&D.  But there's no net
surplus to be extracted.

To see why this matters, consider the following.  Suppose you want
money *now*.  You have web 3.0 in development, so you set up a claim
on web 3.0.  To get money, you must *sell* claims, right?  You are now
liable to pay somebody a lot of money if web 3.0 actually works.
That's OK, you're going to make much more profit than that, and in the
end your VC is backing this play, it's no more your money in this
scenario than in the patents world.  Problem is, *I* invent web 3.0
*first* and get the lion's share of the market.  Now how do you plan
to pay off those claims?  Of course, actually the VC pays, but since
"IPOs" in these markets typically trade at a small fraction of the
claim value, the loss is a multiple of the amount of research they
were able to fund through it, while if they succeed, the profit is a
fraction of what they could have made.  This just doesn't make sense
unless VCs (or inventors) are *extremely* cash constrained now but
expect to be very rich in the future regardless of the success of the
project.

Alternatively, you could write the claim as "don marti invents web 3.0
on or before Dec. 31 2010", so you wouldn't have to pay in the event I
get there first, but you know what?  That's not very much a prediction
market, much more like a junk bond.  Ie, this kind of claim doesn't
really fit Bell's definition of prediction market, as it's really a
security (the holder gets an interest in the success of a particular
project).  So you see we already have very similar securities but few
R&D projects are funded that way.  Furthermore, the presence of the
patent system only makes this kind of funding more attractive, because
it increases the payoff when you "win".  Without the patent system,
you should see less.

So if we get rid of the patent system, prediction exchanges are not a
way to replace whatever incentives patents provide for R&D.

Steve