Bundling and Nonlinear Pricing in Telecommunications
Open Access
- Author:
- Luo, Yao
- Graduate Program:
- Economics
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- April 18, 2013
- Committee Members:
- Isabelle Perrigne, Dissertation Advisor/Co-Advisor
Paul L E Grieco, Committee Chair/Co-Chair
Jonathan W Eaton, Committee Member
Fenghua Song, Committee Member
Quang Vuong, Special Member - Keywords:
- Bundling
Nonlinear Pricing
Telecommunications
Multidimensional Screening
Asymmetric Information - Abstract:
- This dissertation consists of three chapters. Chapter 1: Nonlinear Pricing with Product Customization This paper proposes a method to incorporate product customization and unobserved heterogeneity in structural analysis of nonlinear pricing data. My model is adapted from Maskin and Riley (1984). In equilibrium, average price per unit decreases with an index of quality/quantity, which aggregates observed attributes and unobserved heterogeneity of the product. I study the identification of the aggregation parameters and tariff function from data on payments and consumption quantities. My novel aggregation method leads to a problem reminiscent of transformation models. I then exploit both the firm's and consumer's first-order conditions to identify the model primitives, which are the consumer's utility function, taste distribution, distribution of product unobserved heterogeneity and firm's cost function. I also develop a semiparametric estimation method to recover these primitives. The estimation method is applied to data from a major mobile service provider in Asia. My empirical results support the model. Counterfactual experiments show that, with the same level of complexity, incremental discounts seem to be more preferable as it better approximates the second-best, leading to higher firm profit and consumer surplus relative to all-units discounts and quantity forcing. Chapter 2: Multiproduct Nonlinear Pricing: Mobile Voice Service and SMS, with I. Perrigne and Q. Vuong This paper studies multiproduct nonlinear pricing in the cellular phone industry with voice and message services. The model derives from Armstrong (1996) in which the unknown types of consumers are aggregated while the firm designs an optimal cost-based tariff. As usual in multidimensional screening problems, there is pooling at equilibrium. Moreover, given that the consumers add a large number of unobserved extra features, we introduce two terms of unobserved heterogeneity for voice and message add-ons. The model defines two one-to-one mappings between the unknown aggregate type to the cost and the cost to the payment. We then study the identification of the model primitives. Under some identifying assumptions such as a parameterization of the cost function, we show that the primitives of the model (the aggregate type density, the indirect utility and the joint density of unobserved heterogeneity) are identified from observables. The empirical results support the model and display an important heterogeneity in types and unobserved heterogeneity. The cost of asymmetric information for the firm is assessed. Chapter 3: Bundling and Nonlinear Pricing in Telecommunications This paper studies bundling and price discrimination by a multiproduct firm selling internet and phone services in an imperfect information setting. I derive the optimal selling mechanism, and provide primitive conditions under which different bundling strategies arise, such as component pricing, pure bundling, semi-mixed bundling and mixed bundling. I show that the model structure is nonparametrically identified. I then propose a three-step semiparametric estimation procedure involving a new regression spline estimator under both monotonicity and bound restrictions. An illustration on China Telecom data shows that mixed bundling is beneficial to both the firm and the consumer relative to component pricing.