Welfare Trap
A welfare trap is a structural condition where taking a job leaves a benefit recipient financially worse off than remaining unemployed. It is created by means-tested welfare programs — programs that cut off benefits once income exceeds a threshold.
The mechanism: new wages exceed the eligibility threshold and benefits stop. But the wages don't cover the same costs as benefits did, and employment adds new costs (transport, childcare, clothing). The net result is negative: work pays less than not working.
This is a perverse incentive embedded in the policy itself. The person isn't making an irrational choice — the structure makes employment economically irrational.
Why It Persists
Means-testing exists because it concentrates resources on those most in need, which is a reasonable goal. The trap is a side effect of the implementation: the cliff edge. Benefits drop to zero at the threshold rather than tapering. Any policy that has a cliff will have a trap.
The trap also has second-order effects: fewer people enter the workforce → lower economic output → slower growth → more people on the poverty margin. It harms both recipients and the broader economy.
Exits from the Trap
Gradual phase-out — taper benefits continuously as income rises. No cliff edge, but still a marginal disincentive. Most common reform.
Universal benefits — provide some benefits (healthcare, childcare, education) to everyone regardless of income. Removes the cliff for those categories.
Universal basic income (UBI) — a fixed unconditional payment to all citizens. The only design that fully eliminates welfare traps because earned income supplements the base rather than replacing it. First proposed in the 18th century; tested locally in several countries but not yet implemented at national scale with sufficient evidence.
Connections
- incentive-superpower — the trap is a perverse incentive; welfare design is an incentive design problem
- second-order-thinking — the trap is what happens when policy optimizes for the first-order effect (directing aid to the poor) without modeling the second-order effect (disincentivizing work)