“…Thus, the proposed mechanism does not convey any information to the agent. However, since the optimal mechanism we derive is EPIC, it also constitutes a solution to the informed-principal game in which the principal proposes a mechanism only after observing her private signal: this game has a separating equilibrium in which the principal proposes one mechanism for each signal realization and, thus, the agent perfectly learns the signal from the 28 Mechanism design with alternative verification technologies or evidence is studied in Green and Laffont (1986), Bull and Watson (2004), Glazer and Rubinstein (2004), Deneckere and Severinov (2008), Kartik and Tercieux (2012), Hart, Kremer, and Perry (2017), Mylovanov and Zapechelnyuk (2017), Ball andKattwinkel (2019), Ben-Porath, Dekel, andLipman (2019), Epitropou and Vohra (2019), and Koessler and Perez-Richet (2019), as well as Rappoport (2020). More recently, Silva (2021) studies a model with an analogous informational setup where the principal's verification is costless but yields imperfect results.…”