From: Robin Hanson (rhanson@gmu.edu)
Date: Mon Sep 25 2000 - 10:47:33 MDT
At 11:09 PM 9/24/2000 -0700, hal@finney.org wrote:
>I've belatedly had a chance to read Robin's futarchy paper,
>http://hanson.gmu.edu/futarchy.pdf or .ps. ... a fascinating idea, ...
Thanks!
>I had a few concerns which may not have been clearly addressed. The first
>is, what evidence is there that futures market predictions can work over
>the time scale envisioned for futarchy? The paper talks about 50 year
>time horizons. But most futures markets only predict prices a few months
>out. Even the long-term LEAPS options are only good for a year or two.
Stock markets implicitly forecast out to infinity. That is, they estimate
a weighted average of future profits, weighted by risk-rated interest rates.
They arguably do worse estimating long-term outcomes than short come ones,
but presumably any system of estimation would share that feature. My claim
is not that speculative markets do very well at everything, just that they
on average do better than the other institutions we know of.
>Along these lines, can any conceivable system estimate GDP over such a
>long time frame? Robin's specific proposal is to average log of GDP over
>the next few decades, discounting by a few percent a year. What if we
>combine this with Robin's estimates from his paper on long-term growth,
>... We could see the economy doubling over periods
>of weeks, and later even hours. I think with Robin's formula you'd
>get a divergent series; at a minimum, the sum would be dominated by the
>ultra-explosive growth we hit during the Singularity.
If you take the log of an exponential, you get a line. And then if you
exponentially discount that line, you get a finite result. So any finite
set of speedups in the exponential growth rate of the economy gives a
convergent, not divergent, result for the value of my first cut
proposed measure of national welfare.
It is quite true, however, that to the extent one takes the possibility
of such transitions seriously, and takes seriously the idea that policy
decisions might influence when such a transition would occur, that my
proposal would give an immense weight to preferring policies which are
likely to make such transitions happen earlier. It would be very hard
for other considerations to be heard over this.
But I think this is as it should be. The benefits of citizens being
ridiculously rich just a year sooner are so enormous that it should
get a very large weight. Of course it wouldn't be worth, say, losing
half the population to a sudden death, but my proposed measure gets
that right too I think.
>I was somewhat confused by the futarchy descriptions of the legal
>structure as being composed of nested policies, with more fundamental
>policies authorizing less fundamental ones. This does describe the
>current regime to some extent, but how would it work with futarchy?
>Would more-fundamental policies constrain less-fundamental ones? How
>would this be enforced, by the courts?
Any proposal for a fundamental "rule of government" requires that one
assume some structure which enforces that rule. Technically, there
needs to be some way to complain that the rule has been violated, and
some process that responds to such complaints. Ideally you just
complain to the citizens via free speech, who become so outraged that
they revolt or throw the bastards out. Some sort of court system
can often effectively economize on citizen attention.
>The other point that seemed troublesome to me was the tremendous
>complexity of current laws. ... Digital Millennium Copyright Act,
>... 4570 lines long. ... I don't see how such a complex document
>could successfully make it through a futarchy process. ...
>Currently a process exists in which various competing interests
>influence legislatures to try to get language way they want it.
>At the other extreme you could let anyone propose any law, but there
>would be so many possible variations on even a simple 5 or 10 page
>law that it's hard to see how the market could work. Every different
>clause or section could be reworded, added or removed. You'd have a
>combinatorial explosion of possible documents.
There are costs to getting speculators to pay attention to any give
proposal, and the ideal would be to have those costs borne by the
person or group that makes a proposal. So if there is a status
quo copyright act, and you want to add two words to line 3289, you
have to pay a "bank" to create the relevant assets and a "market"
to publish offers to trade them, and you have to get someone to
continue to put up significant offers to trade which imply that this
proposal would clearly improve expected national welfare. You might
also have to pay a "proposal fee" to the government, which is
intended to account for the costs imposed on others which must
track
If your proposal says that it only applies to the status quo as of
Sept 15, and your proposal says it will be decided over the week
from Sept 15 to Sept 22, then if the status quo is changed on Sept 17,
your proposal is invalid and can't be considered. If the chances of
a proposal being approved are too low, its proponents won't bother
to pay the cost to propose it.
>Robin's document is somewhat sparse in its description of these
>procedural issues. He suggests using the futarchy market itself to
>evaluate different ways of handling these matters. This is a consistent
>viewpoint but it makes it harder to evaluate the system because we can't
>know in advance how the specifics would work. I'd like to see futarchy
>details fleshed out a little more as a straw man proposal.
I suppose I could think more about giving a specific example of this:
>There are some conservative rules regarding what is a valid market,
>when such a market clearly estimates something, and how long the veto
>delay is. These rules say who can propose policies, who can bet on
>them, what assets and market places count for determining prices, and
>how long and distinctly prices must differ in order to say that one
>price is ``clearly" above another.
Unfortunately, the feedback I get from economist friends is that I've
already gone too far in the engineering direction to be publishable
in a good economics journal. Going farther in that direction causes
even more troubles.
Thanks again for your thoughtful comments!
Robin Hanson rhanson@gmu.edu http://hanson.gmu.edu
Asst. Prof. Economics, George Mason University
MSN 1D3, Carow Hall, Fairfax VA 22030-4444
703-993-2326 FAX: 703-993-2323
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:31:11 MST