Mortgage prices are determined according to two major alternative plans: One is the lock plan (LP), in which the price is determined at time of application, and the other is the no-lock plan (NLP), in which the price is determined when the application process ends and mortgage commitment is issued. It is widely believed that the difference between the plans is mainly due to risk shifting between borrowers and lenders. By constructing a queueing model to represent the mortgage system and analyzing the optimal pricing policy for the two plans, the authors show, however, that even if risk factors are controlled, there exist inherent differences between the plans. In particular, when the optimal policy is utilized, the NLP consistently yields higher revenues that the LP. The ability of lenders to regulate the queue size by varying prices, combined with their ability to change the price spread while customers are already in the queue, generates those extra revenues. The relevance of these results to policy application is discussed.