Genericity and Robustness of Full Surplus Extraction
研究了在通用类型空间上，允许完全剩余提取的先验分布是否普遍存在，发现这类先验在拓扑意义上是普遍的，并证明了Crémer-McLean机制的稳健性。
We study whether priors that admit full surplus extraction (FSE) are generic, an issue that becomes a gauge to evaluate the validity of the current mechanism design paradigm. We consider the space of priors on the universal type space, and thereby relax the assumption of a fixed finite number of types made by Crémer and McLean (1988). We show that FSE priors are topologically generic, contrary to the result of Heifetz and Neeman (2006) that FSE is generically impossible, both geometrically and measure-theoretically. Instead of using the BDP approach or convex combinations of priors adopted in Heifetz and Neeman (2006), we prove our genericity results by showing a robustness property of Crémer–McLean mechanisms