Welfare Effects of Paid for Prioritization Services: A Matching Model with Non-Uniform Quality of Service

Abstract

We develop a stylized two-sided model of the local broadband Internet market where prices are used to facilitate “interactions” between the Broadband Internet Service Provider’s (BISP) subscribers and Content and Application Service Providers (CASPS), as well as the priority (i.e., speed) of those interactions. The model is used to evaluate the welfare and distributional effects of several “paid for prioritization” services offered by a BISP. The evaluation is conducted across a variety of economic environments to test the robustness of the welfare effects. The environments are generated by defining parameters that determine whether CASP valuations for interactions are either greater than or less than comparable valuations of subscribers, whether the BISP operates in a high cost or low cost environment, and whether linear or nonlinear pricing strategies are used on the BISP platform. Analysis demonstrates that in the majority of the examined economic environments, the BISP’s resources are more efficiently allocated when it is allowed to charge a uniform (non-discriminatory) price for priority on both the subscriber and CASP sides of the market. The efficiency gain is most pronounced when CASP valuations for interactions are high relative to subscriber valuations, when priority service is priced on an interaction basis, and when total costs are low. In a non-linear “data cap” pricing environment in which CASPs are only able to obtain priority service on an all or nothing basis for each of their subscriber interactions, priority pricing does not enhance market efficiency unless CASP values exceed subscriber values and total costs are low.