Buyers may try to motivate their sellers to make relationship-specific investments to reduce the probability that the design of the goods they procure is defective. In some countries, courts examine how much real authority the seller had in performing the work to assign liability for a design failure. I show that this courts' approach induces the sellers to under-invest and the buyers to under-specify the design of the goods. I find that this approach can also make it harder to sustain optimal relational contracting, leading to the conclusion that it cannot be justified on efficiency grounds.