← Patterns

Moral Hazard

Pattern · economics

Core Concept

Moral hazard is when someone takes on more risk because they know they won’t bear the full consequences. The insulation from consequences changes behavior — people act differently when they’re spending someone else’s money, driving someone else’s car, or borrowing against someone else’s future.

The classic example is insurance: you drive less carefully when the insurer pays for the crash. But the pattern is everywhere — bailouts, guarantees, subsidies, and any system where risk and reward are decoupled.

Why It Matters

Moral hazard isn’t just “people are irresponsible.” It’s a structural incentive problem. Rational actors should take more risk when they’re insulated from consequences — that’s the individually optimal play. The damage is systemic, not individual.

This is what makes it hard to fix: the behavior is rational at the individual level but destructive at the system level. You can’t solve it with appeals to responsibility. You solve it by realigning incentives — making the risk-taker bear more of the consequences.

Where It Shows Up

• Insurance: Deductibles and copays exist specifically to counter moral hazard — they force the insured to bear some cost
• Banking: “Too big to fail” banks take excessive risk because they expect government bailouts. The 2008 crisis was a moral hazard cascade.
• Student loans: Government-backed loans with potential forgiveness → borrowers borrow more, universities charge more. See Debt-Forgiveness-Economics
• Agent design: An AI agent with no cost signal will over-generate, over-explore, and over-build. See Effectiveness-Over-Efficiency, Autonomy-Through-Constraints
• Management: A manager who never lets reports fail creates moral hazard — the team takes reckless shortcuts because the manager always catches it

The Defense

1. Skin in the game. The person taking the risk must bear some of the downside.
2. Counter-metrics. For every metric you reward, track the cost metric too. See Goodharts-Law.
3. Structural constraints over trust. Don’t rely on people to self-regulate when incentives point the other way. See Autonomy-Through-Constraints.

Related Patterns

• Incentives-Drive-Behavior — the general case. Moral hazard is what happens when incentives reward risk-taking without consequences.
• Managers-Mind — people model your incentive structure better than you do. If you signal that failure has no consequences, they’ll act accordingly.
• Goodharts-Law — moral hazard is a specific instance of Goodhart’s: the metric (risk-adjusted return) gets gamed when consequences are removed.
• Autonomy-Through-Constraints — the design pattern that counters moral hazard in agent systems.

Cross-References

• Debt-Forgiveness-Economics — moral hazard applied to student loan policy
• Incentives-Drive-Behavior — the foundational incentives pattern
• Goodharts-Law — metric gaming as a related failure mode