Full surplus extraction in mechanism design with information disclosure

I study mechanism design settings with quasi-linear utility where the principal can provide agents with additional private information about their valuations beyond the private information they hold at the outset. I demonstrate that the principal can design information and a mechanism so as to fully extract the complete information first-best surplus if agents’ ex ante information only affects their beliefs about, yet not their valuations. Otherwise, the result holds if each agent’s initial private beliefs satisfy a spanning condition.