Skip to main content

Incentive Design

Why do shared systems get captured by corridor-seekers rather than contributors — and what design prevents it?

Coordination stays healthy when the work earns more than the corridor to the work.

Every shared system is vulnerable to the same structural failure. Value concentrates around a scarce choke point — an authority, a certification, a platform, a committee seat. Those who do the work rarely invest in controlling the choke point. Those who want the corridor do.

Over time, corridor-seekers capture the corridor. The original signal is displaced. This is not cynicism. It is structure.

The fix is not to remove authority. It is to make authority structurally expensive to hold for the wrong reasons — and to make contributing the primary path to influence.

The Corruption Vector​

The sequence is always the same:

  • Signal is valuable — people gather around it.
  • Scale needs coordination — management structures appear.
  • Management structures attract people who want the role, not the signal.
  • Defending the role displaces serving the signal.
  • The institution protects itself and calls that protection governance.

This is a vicious feedback loop: the setpoint shifts from "produce the signal" to "hold the position." Once that shift happens, the loop reinforces itself. Every cycle makes capture cheaper and contribution less rewarded.

The incentive layer is where you interrupt that loop.

Reputation Currency​

Cash is easy to concentrate. Reputation is harder.

When reputation is the primary currency of influence — rather than formal position, certification, or financial stake — three things change:

  • Influence must be earned repeatedly, in public, against a standard others can inspect.
  • Past performance is visible and persistent; a bad actor cannot easily reset the ledger.
  • The cost of maintaining a false reputation exceeds the cost of doing the work.

Design conditions that make reputation visible:

  • Track records are open. Decisions, commitments, and outcomes are on the record. Not claims — outcomes.
  • Reputation transfers slowly. Influence earned in one domain does not automatically transfer to unrelated decisions. Each claim to authority requires fresh demonstration.
  • The community is the verifier. Reputation is not assigned by the institution; it is assigned by peers who have direct experience of the work.

Graduated Sanctions​

The failure mode of most sanction systems is binary: nothing until something so large that enforcement becomes political. Then either the sanction is never applied (the violation normalises) or the sanction is applied inconsistently (trust collapses).

Graduated sanctions interrupt both failure modes. [source: Ostrom, Governing the Commons, 1990]

Level 1 — First violation. Small, public, expected. The norm is restated. The message is: we noticed, we care, we are watching.

Level 2 — Repeat violation. Meaningful penalty. Proportionate to the resource drawn. Not devastating, but impossible to ignore.

Level 3 — Persistent violation. Escalating loss of privileges. Access reduced. Voice reduced. The corridor shrinks.

Level 4 — Expulsion. Final and rare. Reserved for systematic violation after escalation has failed. Expulsion is not punishment — it is boundary enforcement.

The key design principle: sanctions scale with repetition, not severity alone. A serious one-time breach is different from a pattern of small violations. Systems that treat them the same fail to deter the pattern.

Making Capture Expensive​

Capture is cheap when:

  • Authority is concentrated and scarce — worth more to control than the work it coordinates.
  • The corridor is opaque — you cannot see who holds it or why.
  • Reputation is assigned by the institution — the institution controls who gets credit.
  • Sanctions are symbolic — the cost of violating norms is lower than the benefit.

Make capture expensive by reversing each condition:

Distribute authority. No single node holds exclusive routing power. Each group governs its own domain. Coordination happens through shared protocols — auditable, modifiable, not owned. [source: Ostrom, Governing the Commons, 1990 — nested enterprises principle]

Make the corridor visible. Who holds influence, how they got it, what decisions they made, what outcomes followed. Opacity is the corridor-seeker's primary asset. Transparency removes it.

Let reputation set by peers. Track records are public. The community verifies. No institution gets to assign credit for work it did not witness.

Price violations accurately. When violations are costless, every actor who can violate without sanction will eventually do so. Graduated sanctions that escalate make the full cost of repeated violation legible.

The Integrity Test​

The Integrity Test asks: does this system replenish what it draws?

Apply it to incentive structures:

  • Does this reward signal the work or proximity to the decision about the work?
  • Does this structure make it easier to contribute than to capture?
  • When influence accumulates, does it reflect performance or tenure?
  • If you removed the formal authority structure tomorrow, would coordination improve, stay the same, or collapse? If it collapses, the authority was doing real work. If it improves, the authority was friction.

Practice​

The moves a practitioner takes when designing an incentive system:

  • Map the choke points first. Where does authority concentrate? Where does the corridor appear?
  • Ask who holds influence and why. Tenure, title, or demonstrated work?
  • Design track records before designing rewards. Visible outcome history is the foundation.
  • Write the sanction ladder before you need it. Graduated sanctions lose legitimacy when written in reaction to a crisis.
  • Test for capture: if the people most incentivised to hold a role also benefit most from blocking newcomers, the design is already compromised.
  • Check whether the institution assigns credit for work it did not witness. If yes, the reputation system is captured.

Checks​

Signs the system is working:

  • New contributors gain influence faster than long-tenured members who stopped contributing.
  • Level 1 sanctions fire regularly — violations are caught early, before they normalise.
  • The community assigns credit; the institution only records it.

Signs it is drifting:

  • The same people hold influence across multiple decisions with no fresh demonstration required.
  • Violations are handled privately; no escalation history is visible.
  • Compliance becomes the goal of governance, not coordination.

Measure by asking: who got influence this month, and what work earned it?

Context​

  • The Commons — The design principles for self-governing shared resources; goodwill as the commons
  • Feedback Loops — The corruption vector is a vicious loop; incentive design is how you re-engineer the setpoint
  • Governance — Protocol-over-authority as the institutional layer of the same argument
  • Governance Problems — The specific failure modes that broken incentive design produces
  • Scoreboard Reality — How to measure whether contribution is being rewarded over corridor-holding

Questions​

Which is harder to design — a reputation system that resists gaming, or a sanction system that escalates?

  • At what organisation size does peer-assigned reputation become too slow or too noisy to function as the primary currency of influence?
  • What is the minimum coordination protocol that makes authority distributed without making decisions impossible?
  • When a graduated sanction system is first introduced into an existing community, what determines whether members accept it as legitimate or resist it as control?