Advisory Futarchy

Me April 2:

Most CEOs are not at all eager for futarchy to advise on the decisions they feel they “own”. But it should be easier to get CEO support for futarchy advising decisions that they often cannot make on their own, but must instead make together with boards and investors. For example, futarchy could advise whether to keep or dump the CEO. Another promising application is decisions on when and how to raise more capital.

Another way to make decision markets more appealing to executives wary of ceding control is to make markets advisory. One need be less confident in the value or reliability of markets that only make suggestions. Decision makers could initially mostly ignore such advice, and then slowly put more weight on them as the proved themselves via concrete track records.

Yes, advisory markets do require that the decisions to be advised be made visible to market participants, who need to know which options will be considered, what outcome metrics are seen as relevant, and roughly when those decisions will be made.

Such markets would be seen as less valuable than markets which directly decide, and would thus justify smaller budgets and attention, at least at first. And it would be a bit harder to judge how much to reward those who propose options that markets approve.

But the biggest problem is likely to be decision selection biases. However, those can be mostly eliminated if decision makers will agree to two key conditions:

A) Declare at time T1 that the decision time T5 will fall within a particular period [T3,T6], with price averages over [T2,T5] offered as advice. The shorter is duration T6-T3 and the longer is T2-T1 the better, though 48 and 24 hours seem sufficient.

B) Allow and encourage key decision makers and advisors who have access to inside info relevant to the decision to trade in the markets during [T1,T5]. Or at least allow associates of theirs to whom they can reveal such info.

Decision selection bias is caused by fears of traders at T4 that the decision will be made at T5 based on info unavailable them at T4. Including key informed parties should ensure that they reveal their info to the markets at some point, subsidizing to ensure these markers are thicker during [T1,T2] than [T2,T5] should induce all to reveal what they know by T2, and a short duration [T2,T6] should insure that little new info appears during that period. (Manipulation attempts would thicken markets during [T2,T6]; make the extra [T1,T2] subsidy big enough to compensate.)

Yes, another approach is to get executives to make decisions randomly some small X% of the time, and make decision markets condition on such random decisions. Alas, I fear that will be seen as a much bigger ask.