Perverse Incentives

mechanism

Perverse incentives arise when a system rewards behavior that undermines its larger goal. They endure because each participant captures an immediate, legible gain while the resulting harm is delayed, dispersed, or difficult to assign.

Pay people for each fragment of skull they find, and they may smash intact skulls into pieces. The metric records a spectacular rise in “discoveries” even as the objects worth discovering are destroyed.

E2

Private wins, collective failure

The system wants an outcome but pays for a proxy: skull pieces instead of preserved skulls, bugs fixed instead of reliable software, conferences held instead of credible scholarship. Once the proxy carries the reward, participants optimize it—even when doing so reduces the real value it was meant to represent. The arrangement survives because benefits arrive locally and immediately, while reputational or institutional damage remains fuzzy. To diagnose it, apply Understanding Incentives: ignore the declared purpose and ask what action actually earns money, status, or a favorable count.

E1 E2

Where it shows up

The bug factory

A programmer paid per bug fixed can increase earnings by creating more bugs. The reward turns preventable failure into inventory.

E2

Conferences nobody wants to stop

A fake conference can satisfy several scorecards at once: organizers collect fees, students gain résumé lines, and colleges report international activity. No participant needs the conference to be intellectually valuable for the transaction to look successful.

E1

A bad result is not enough

Not every failure proves a perverse incentive. The mechanism applies when the rewarded behavior itself damages the larger goal. Nor does it require cartoonish dishonesty: people may simply pursue the legitimate local gains available to them while diffuse costs fall elsewhere.

E1 E2

Audit the unit of reward

Tomorrow, take one metric used to allocate money, status, or credit in your work. Write down the cheapest way to increase it without improving the real outcome. If that shortcut causes harm, redesign the reward around the outcome—or add a cost for the damage—using Incentive Alignment as the test.

E1 E2

Episodes that teach this