Take it or leave it: take-up, optimal transfer programs, and monitoring

This paper studies the optimal income redistribution and optimal monitoring when disability benefits are intended for disabled people but when some able agents with high distaste for work mimic them (type II errors). Labor supply responses are at the extensive margin and endogenous take-up costs may burden disabled recipients (because of either a reputational externality caused by cheaters or a snowball effect). Under paternalistic utilitarian preferences that do not compensate for distaste for work, inactive disabled recipients should obtain strictly lower consumption than disabled workers. The cost of monitoring supports adoption of an Earned Income Tax Credit. However, and surprisingly, with or without take-up costs, even if perfect monitoring is costless, it proves optimal to have type II errors. These results are robust to a utilitarian criterion. The paper provides numerical simulations calibrated on U.S. data.