Endogenous Vindication: Reputation and Effort in Expert Advice

Abstract

An expert's advice is often evaluated only when a client acts, and the quality of that evaluation depends on implementation effort. We study repeated advice by an expert whose fixed ability is unknown to both the expert and the public. A favorable recommendation may induce a short-lived client to undertake a costly project and choose effort. Higher reputation elicits greater effort, making outcomes more informative about ability. The expert therefore controls whether a reputation-dependent performance test occurs, while the client controls its precision. With diminishing career returns to reputation, the expert may withhold favorable advice even when implementation creates positive surplus. Monotone exposure follows from broad smooth primitives when implementation requires sufficiently high reputation, and from an explicit open family of quadratic environments. Advice then follows a reputation cutoff; stronger career concerns weakly raise it and can stop testing earlier. A low-ability expert eventually enters an absorbing distrust region, while a high-ability expert is permanently distrusted with positive probability and otherwise fully vindicated. Individually more informative tests can therefore coexist with less learning overall because reputational incentives reduce the number of tests conducted.