Subject: Re: idea futures markets
From: <stephen@xemacs.org>
Date: Fri, 8 Sep 2006 16:17:31 +0900

Kragen Javier Sitaker writes:

 > (In case you're not familiar with the idea, one way to explain it is
 > as follows: you make a public offer of $10 for an option, redeemable
 > in 60 days, to sell some not-yet-written piece of free software to the
 > option seller for $100.

I gotta wonder why VCs aren't doing this already.

 > Stephen Turnbull writes:

 > > Think about it.  Why would any VC give you money to bet on the market?
 > > "*You* are a developer, go develop, dammit!  *I* am the VC, and *I* do
 > > the betting and get the returns to supplying capital and taking
 > > risks."
 > 
 > So far, there aren't many examples of ways for developers to get paid
 > for something that isn't developed yet

Yup.  That's not the way things work.  It's "what have you done for me
lately?"  Even the NSF post-funds research---you have to ante up good
research, then they give you money to do better research (or supervise
others to do it for you).

 > Ben Tilly wrote:
 > > What Stephen is envisioning is something like a traditional futures
 > > market, where someone sells a contract saying, "I will pay you X at
 > > time Y if Z happens."  The seller of that contract gets money now, in
 > > return for a potential liability later.  The contract is not covered,
 > > but if the author of the contract doesn't pay up, they can face
 > > penalties (up to and including jail).
 > 
 > Is that right, Stephen?  That's definitely not what Bell was
 > proposing, but I agree with you that it wouldn't work, for all the
 > reasons you described, and also for the reason that in the markets
 > Bell is talking about, the contracts are sold only by the bank, not by
 > the traders in the market --- so nobody gets any money up front.

Well, if no money changes hands until the event actually happens, that
puts paid to the whole idea.  It should, however, be possible to sell
appropriate options in the OTC market, or to find someone who's got
the other side of the market and swap tickets for real money now.

 > Stephen Turnbull wrote:

 > > SPexes as described in Bell's paper are zero-sum, and cannot fund
 > > development per se, except if they redistribute profits from people
 > > less interested in investing in R&D to those more interested in
 > > investing in R&D.  If this consistently happens on average, the former
 > > will leave the market.  Uh-oh.
 > > ...
 > > The market for lemons problem == insider trading.
 > 
 > Strictly speaking, the market-for-lemons problem is something that
 > might or might not result from insider trading, in the way you
 > describe very clearly in the above-quoted paragraph.  If it were
 > possible to use the market for hedging, it wouldn't.

No.  Insider trading means that you can't use the market for hedging,
really, unless you're way, way risk averse.  Consider the following
little story (from my dissertation):

1.  I flip a coin, hiding it on my palm so that I can see it, while
    you cannot.

2.  I say, "I'll bet you at 1:1 odds that the coin turned up heads.
    How much are you in for?"

Now, the objective probability is 1/2, and that's what you'd bet on
before the toss---but you now know (with as much confidence as you
have that I'm not totally crazed) that the actual probability that I
saw "tails" is zero.

This is the extreme lemons market; total failure.

 > > Your Transmeta scenario has a small boo-boo, and a big conceptual
 > > problem.  The boo-boo is that the perfect hedge requires buying $1 of
 > > stay-hot claims per $1 of hot-CPU expense, not 1 per 2.

 > I don't know much about derivatives, so I may have erred in my
 > terminology; what I meant by "hedge that risk completely" was that
 > they would no longer care whether Transmeta succeeded or failed,

That's the right terminology.  I may have mistaken the way the market
works---it's much easier to calculate a hedge if the way the market
works is "pay $P no matter what, receive $1 if Event happens, receive
$0" if Event doesn't happen".  In that case, since the payment of $P
is certain, you can ignore it in calculating the hedge.

If the contract is "receive $1 if Event happens, pay $P if it
doesn't," then the calculation of the hedge depends on P.  Suppose you
lose $1 if Event happens.  Then a perfect hedge satisfies

		 -$1 + $1*X = -$P*X  or  X = 1/(1+P)

Note that there's no reason why a rational trader would refuse P > 1
(eg, if Prob(Event) = .99, P = 1.01 seems a real bargain to me!) so
you can't interpret P as an estimate of probability, as is usually
done.  That's why I assumed you always pay $P.

 > Maybe you're right, although of course the free-software developers
 > aren't at much risk betting that their own projects will succeed ---
 > that's the whole information-asymmetry problem you're objecting to :)

Of course they are at risk.  That's why we're having this
discussion---we want to bet with Other People's Money[tm], thus
mitigating the risk.

 > > If you ship web 3.0 and cause those things to happen, you'll stand to
 > > make enough money that your piddling prediction market investment is
 > > meaningless.  If you don't, you lose development costs (which you'll
 > > lose anyways).
 > 
 > The standard lament about open-source market failings is that you can
 > ship web 3.0 and not make a dime.  Certainly CERN hasn't made anything
 > from web 1.0.  Again, the idea is to provide a non-copyright way to
 > make that money.

Copyright isn't the worst of all possible worlds---there are always
ASPs.  I don't think it's an accident that asps are snakes. :-)

 > Of course, nobody will want to use an alternative that doesn't work,
 > but this is begging the question; the question at issue in the
 > above-quoted text is whether it will fail to work because nobody will
 > put in money.  You can't answer that by observing that nobody will put
 > in money if it fails to work.

Of course you can.  That's called "equilibrium coordination failure",
and it happens all the time.  Mere fear that that equilibrium will be
realized may be enough to block entry.  To make this "work", we need
to destabilize that equilibrium so that people will coordinate on the
good equilibrium.

Steve