Incentive Alignment
A financial instrument that creates new perverse incentives is worse than no instrument at all. This chapter stress-tests every proposed defense mechanism against three principles from mechanism design theory — then introduces five new instruments designed from first principles.
Three Tests Every Instrument Must Pass
From Hurwicz, Maskin, and Myerson's mechanism design theory: an incentive-compatible mechanism makes honest behavior the profit-maximizing strategy. These three tests operationalize that principle.
Honest behavior must dominate dishonest behavior
The mechanism must make manipulation either structurally unprofitable or exponentially expensive. This is the Myerson-Hurwicz-Maskin foundation: incentive compatibility means truth-telling is the profit-maximizing strategy.
Test: Does a manipulator lose money in expectation, even when their trade is directionally correct?
No new perverse incentives
The instrument must not create new attack vectors that didn't exist before. Credit rating agencies failed this test — the "issuer pays" model created incentive to inflate ratings. CDS failed this test — they created incentive to push companies into bankruptcy.
Test: Does the instrument create any scenario where a participant profits from the system failing?
Self-enforcing over oracle-dependent
Mechanisms that require someone to judge "did manipulation occur?" inherit the incentive alignment problem of the judge. The best instruments are self-enforcing — like LMSR's exponential cost function, which makes manipulation expensive without any human ruling.
Test: Does the mechanism work if every oracle/judge is corrupt?
Grading the Original Five
Each instrument from Chapter V, strawmanned and steelmanned. Click to expand.
“The instruments that score highest are self-enforcing — they make manipulation expensive through market mechanics, not through human judgment about whether manipulation occurred. The oracle problem is the graveyard of clever financial instruments.”
Five New Mechanisms
Designed from first principles to pass the three incentive alignment tests. These prioritize self-enforcing mechanics over oracle-dependent adjudication.
The Full Ranking
| Instrument | Grade | Category | Self-Enforcing |
|---|---|---|---|
| Soulbound Reputation Tokens | A | Infrastructure primitive | Yes |
| Credibility-Weighted Markets | A- | Market design | Yes |
| Credibility-Linked Bonding Curves | A- | Market design | Yes |
| Conviction Futures | B+ | Commitment device | Yes |
| Retroactive Slippage Penalties | B+ | Deterrence | No |
| Reputation Escrows | B+ | Commitment device | Yes |
| Anti-Slander Bonds | B | Cost-raising | No |
| Adversarial Prediction Bonds | B | Crowd enforcement | Yes |
| Parametric Insurance | B- | Hedging | No |
| Reputation CDS | C+ | Hedging | No |
The Pattern
A clear hierarchy emerges. The instruments that score highest share two properties: they are self-enforcing (no oracle required) and they make manipulation expensive through market mechanics rather than adjudication.
Soulbound Reputation Tokens score highest because they are pure infrastructure — a primitive that makes every other instrument more effective. You cannot buy, sell, or fake a track record. The only path to credibility is demonstrated accuracy over time. This is the foundation layer.
Credibility-Weighted Markets and Credibility-Linked Bonding Curves score next because they translate reputation into market mechanics. Bot armies become structurally unprofitable. The cost of manipulation scales exponentially with the number of accounts required.
The instruments that score lowest — Reputation CDS and Parametric Insurance — share a fatal dependency on oracles. Someone must decide whether an event was “disinformation” or “genuine scandal,” and that judge becomes the highest-value corruption target in the system. The 2008 credit rating agency failure is the cautionary tale: when the judge is corruptible, the entire mechanism collapses.
The lesson: build the reputation layer first. Everything else — bonds, insurance, market weighting, conviction futures — works better when verifiable track records exist. Without that foundation, every instrument degenerates into an argument about who gets to define “truth.”
“The best financial instruments don't require honest judges. They make dishonesty structurally unprofitable. Prediction markets need the same principle: not better referees, but a game where cheating costs more than winning.”