Re: Social Policy Bonds -- Attn: Robin Hanson

From: Robin Hanson (
Date: Wed Jan 24 2001 - 08:44:05 MST

Chris Hibbert wrote:
> > I assume that [do-gooders] do sometimes change their non-bond-market
> > behavior in ways that helps the social policy, but this doesn't help
> > them unless they also have private info about whether they will help,
> > and buy more bonds when they privately know they will help.
>I'm still not getting it, Robin. My supposition was that some do-gooders
>would decide whether or not to help depending on whether they were able to
>buy enough bonds to repay the effort. This doesn't sound like your "unless
>they also have private info about whether they will help", which seems much
>more passive. You seem to be drawing causality the other direction: if
>they know they will help address the policy issue, then they buy more bonds.

They need to have private info about whether they will buy enough bonds.

Now this private info can come naturally in the process of trying to
privately buy bonds. If the people on the other side of each trade do not
know whether this buyer is the do-gooder or a mere speculator, if the buyer
is a do-gooder, she can gain a little on the trade. This gain to the
do-gooder is a loss to the seller.

Let's consider an example. You want to reduce the crime rate in some region
and sell a billion dollars in bonds at an auction. Each bond pays $1 if
the crime rate falls below some threshold, and nothing otherwise. If the
crime boss does nothing, there is a 20% chance of crime falling enough.
If he foregoes $100M in profits he can increase this to a 40% chance.
Now if the crime boss held all billion bonds in his hands, he would max
his profits by reducing crime. This would cost him $100M but gain him
(.4-.2)*$1B = $200M in expected bond value. If he held half of the bonds
he would be just indifferent between reducing crime or not doing so.

Now imagine that the bonds go up for sale in a (non-package) one-time
auction. If speculators thought there was a probability P that the crime
boss would gain 75% of the bonds, and thus want to reduce crime, they would
bid $P for each bond, and so the crime boss would profit by the amount
(.4 - (.4*P + .2*(1-P)))*75%*$1B - $.1B. This is negative if P > 1/3.
Thus to profit from this approach the crime boss needs to create the
impression that there is less than a 1/3 chance of reducing crime.

>This is also supposed to be one of the strong points of Idea Futures. IF
>provides incentive for effort to be placed in areas where someone has
>invested. Some polymaths can decide where to spend their investigation
>time based on which market they have collected sufficient coupons to make a
>sufficient return on.

The concept of idea futures is to create incentives to produce information,
not to create incentives to make good policy happen. (That's the concept
of futarchy :-).

Robin Hanson
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 2b30 : Mon May 28 2001 - 09:56:24 MDT