Sequential Nonlinear Pricing of Experience Goods with Network Effects

We study a nonlinear pricing problem of experience goods in dynamic environments where consumers' private information evolves stochastically over time, and their present types and/or quitting costs are affected by their past consumptions. Our model sheds some light on the pricing of addictive goods. Also, we provide a new ironing technique to meet the dynamic implementability condition. In standard contract model, to guarantee the static implementability condition, the principal needs to set an allocation monotonic in the agent's type. If the contract obtained from the relaxed problem is not everywhere monotonic, a horizontal ironing technique is adopted to revise it and a contract with bunching intervals is thus obtained. In our dynamic setup, however, monotonic condition is not sufficient for implementability. The slope of allocation with respect to type needs to be larger than a positive number. So, we use a sloping ironing technique to obtain a perfect sorting contract. Moreover, consumers in this paper are assumed to be distributed over a network. In contrast to the existing literatures of network game, we focus on a principal-agent model on the network/graph. We construct a new measure for identifying the key player with respect to quality. This measure takes into account both the prominence of actors inside a network and asymmetric information between the principal and agents