Reducing incentive constraints in bidimensional screening

This paper studies screening problems with quasilinear preferences, where agents' private information is two-dimensional and the allocation instrument is one-dimensional. We define a preorder to compare types based on their marginal valuation to the instrument, which facilitates the reduction of incentive compatibility constraints that must be checked. With this approach, the discretized problem becomes computationally tractable. As an application, we numerically solve a problem introduced by Lewis & Sappington (1988)