This is the final exam for Incentives. It pulls together everything: the precise definition (anything that changes the payoff of a behavior), extrinsic rewards versus intrinsic drives, the way people unconsciously rationalize whatever pays them (incentive-caused bias), the cobra effect and the proxy problem, Goodhart’s law, the surprising backfire where paying for something gets you less of it (crowding-out), and the design principles that point a reward at the real outcome. Take your time and reason each one through; several look easy until you spot the trap — blaming character instead of the reward, trusting a proxy, or assuming money always motivates.
How this exam works
Read carefully — this exam is final. Each question appears one at a time. Once you submit an answer it is locked for good: there’s no going back, no retry, and no restart. Your score is hidden until the end, where you’ll see a pass/fail verdict. The pass mark is 70%. A few questions ask you to select all correct answers.
What is the cleanest definition of an incentive?
Select an answer to continue.
Course Recap
Big picture
Incentives, in one picture
- Incentives
- What an incentive is
- Anything that changes a behavior’s payoff — extrinsic (money, status, penalties) and intrinsic (meaning, enjoyment); behavior follows the reward, not the stated value
- Incentive-caused bias
- People sincerely rationalize whatever pays them — "hard to make a man understand something when his salary depends on not understanding it"; watch the principal–agent gap
- Perverse incentives
- Reward a proxy, get the proxy: cobra effect, Hanoi rats, Goodhart’s law — when a measure becomes a target it stops being a good measure
- Intrinsic vs extrinsic
- Paying for something can crowd out the motive that drove it — "a fine is a price"; the daycare that got more lateness, not less
- Designing good incentives
- Reward the outcome not the proxy; align agent with principal; pre-mortem the gaming; pair metrics; mind crowding-out
- What an incentive is
Key takeaways
Incentives are the most reliable lever for explaining and predicting behavior: to know what someone will do, find the reward, not the stated intention — people and systems drift toward whatever is actually paid for. An incentive is anything that changes a behavior’s payoff, extrinsic or intrinsic. Beware incentive-caused bias — people don’t just respond to incentives, they sincerely believe whatever serves them — and the principal–agent gap it opens. The most expensive mistake is a perverse incentive: reward a proxy instead of the goal and you manufacture the gaming you feared (the cobra effect; Goodhart’s law: when a measure becomes a target, it ceases to be a good measure). And money is not magic — paying for something can crowd out the intrinsic or moral motive that was driving it (a fine is a price). So design deliberately: reward the real outcome, align the agent with the principal, pair metrics with guardrails, pre-mortem the gaming, and remember that the fix is never “find more virtuous people” — it’s aim the reward at the thing you actually want.