Comment on "Ironing, sweeping, and multidimensional screening''

In their study of price discrimination for a monopolist selling heterogeneous products to consumers having private information about their own multidimensional types, Rochet and Chon\'e (1998) discovered a new form of screening in which consumers with intermediate types are bunched together into isochoice groups of various dimensions incentivized to purchase the same product. They analyzed a particular example involving customer types distributed uniformly over the unit square. For this example, we prove that their proposed solution cannot be correct, and explain how it can be corrected.