Effort without Evidence

Abstract

I study how paternalistic concern for descendants' success affects learning about the need for costly effort. Each generation can use a demanding method or experiment with a less demanding alternative whose effectiveness is uncertain. Parents who underweight descendants' effort costs undervalue the savings that discovery would provide. On an explicit region of priors and preferences, paternalism removes a materially valuable experiment even though the relevant private record is truthfully transmitted. Relative to an aligned equilibrium attaining the material full-record optimum, the paternalistic equilibrium has greater long-run effort and observed success, lower discounted material welfare, and a less informative terminal public signal. Each sender obtains its preferred full-record continuation, so the comparison survives communication deviations within the same physical environment. I also study concealment as an incentive to experiment. A continuous-effort extension establishes that concealment and finite evidence production can coexist in an infinite-horizon equilibrium. Its welfare gain over babbling is distinct from the loss relative to alignment in the main result.