Abstract
Strausz (2017) claims that crowdfunding implements the optimal mechanism design outcome in an environment with entrepreneurial moral hazard and private cost information. Unfortunately, his analysis, solution and claim depend critically on imposing an untenable condition (29) that he had explicitly discarded from his weak feasibility concept.1 This condition is essentially equivalent to ex post participation. We explain why it is inconsistent to assume consumers can opt out after learning the entrepreneur's cost structure in a model of fraud. We then study weak feasibility without the corresponding ex post participation constraint. We provide a simple example of a crowdfunding design that raises profit and welfare by tolerating some fraud risk. This shows how cross-subsidizing between cost states relaxes the most restrictive moral hazard constraints and generates better outcomes. We then characterize the optimal mechanism in the case of one consumer and two cost states. In general, this must hide information, including prices, from consumers. So crowdfunding cannot implement these optima.