The manipulation problem, in short, is when a speculator privately benefits from a particular decision D and hence is willing to buy excessive conditional stock SD so that D is enforced. I’ve heard Robin Hanson explaining that other speculators will react to counter manipulation but I’m not sure I get how it’s supposed to work in practice:
Suppose that the optimal decision is D∗ and denote E(X) as the expected value of one conditional stock SX (given that X is enforced). The rational, risk-neutral & disinterested speculators will pay at most E(D∗) for each SD∗ .
If a speculator receives a private benefit V when D is enforced, she’d be willing to pay V+nE(D) for n conditional stocks SD, or Vn+E(D) for each SD. It seems to me that given finite stocks and a large enough V, the conditional price of D might outweigh that of the optimal decision D∗ (i.e.Vn+E(D)>E(D∗))
A counter to my description is that speculators can also short-sell, but short-selling (at least currently) involves borrowing existing stocks and selling them, whereas in futarchy the manipulator can simply refuse to lend stocks to short-sellers. In addition, if only one speculator receives the private benefit V, then short-selling SD to other investors would only gain E(D) in revenue, which undermines the incentive of short-selling in the first place.
Is my analysis wrong? Or is there some other mechanism to counter market manipulation?
[Question] Manipulation resistance of futarchy
The manipulation problem, in short, is when a speculator privately benefits from a particular decision D and hence is willing to buy excessive conditional stock SD so that D is enforced. I’ve heard Robin Hanson explaining that other speculators will react to counter manipulation but I’m not sure I get how it’s supposed to work in practice:
Suppose that the optimal decision is D∗ and denote E(X) as the expected value of one conditional stock SX (given that X is enforced). The rational, risk-neutral & disinterested speculators will pay at most E(D∗) for each SD∗ .
If a speculator receives a private benefit V when D is enforced, she’d be willing to pay V+nE(D) for n conditional stocks SD, or Vn+E(D) for each SD. It seems to me that given finite stocks and a large enough V, the conditional price of D might outweigh that of the optimal decision D∗ (i.e.Vn+E(D)>E(D∗))
A counter to my description is that speculators can also short-sell, but short-selling (at least currently) involves borrowing existing stocks and selling them, whereas in futarchy the manipulator can simply refuse to lend stocks to short-sellers. In addition, if only one speculator receives the private benefit V, then short-selling SD to other investors would only gain E(D) in revenue, which undermines the incentive of short-selling in the first place.
Is my analysis wrong? Or is there some other mechanism to counter market manipulation?