Robust Procurement: Bayesian Design under Worst-Case Approval Constraints

We study optimal procurement when a Bayesian designer must obtain approval from a non-Bayesian authority that shares the designer's objective but is uncertain about the value of the good and the supplier's cost.

The designer uses a conjectured model to compute expected payoffs but is constrained to select among mechanisms delivering the largest payoff guarantee to the authority.

This robustness requirement reshapes the tradeoff between efficiency and rent extraction: it reduces procurement from sellers with intermediate costs but may increase it from those with a high cost.

When the good is sold in a market, we show that quantity regulation dominates price regulation if markups under the conjectured model are large, whereas price regulation dominates when demand uncertainty is substantial.