Economic behaviour and decision making: theories of two-sided markets, multiproduct pricing and weighting for cumulative prospect theory

Gao, Ming (2010) Economic behaviour and decision making: theories of two-sided markets, multiproduct pricing and weighting for cumulative prospect theory. Doctoral thesis, University of London: London Business School.

Abstract

This thesis studies the microeconomic theories of two-sided markets, multiproduct pricing, and decision making in risky choice situations. In the first part of the thesis, we focus on a special kind of two-sided markets, where participants can act on both the buying side and the selling side of the market, which we call "mixed two-sided markets". The literature on two-sided markets has assumed that buyers cannot sell and sellers cannot buy. In real life, however, many markets are mixed, with examples ranging from telecommunication to stock exchange. We provide a general model for mixed two-sided markets. We observe that in practice, platforms in such markets often use a "hybrid" bundling strategy: A common membership fee gives access to both buying and selling services, while the individual transaction fees are separated. The main impact of this strategy is what we call the "two-part-tariff effect": Imposing a small bundled membership fee on top of any transaction fees always leads to zero first-order losses in the demand of consumers who use both services, thus enabling the platform to extract more surplus from them. When this positive effect dominates the losses in demand from single-service users, hybrid bundling dominates unbundled sales. We present general conditions that guarantee such an outcome. In the second part of the thesis, we show that the two-part-tariff effect still applies when such tariffs are used in a more general context of multiproduct pricing. We consider a monopolist provider of n (> 1) products who uses two-part tariffs consisting of a membership fee that is common to all consumers, and separate prices for different product bundles. We show that the change in demand for any bundle of k Epsilon [1, n] products caused by imposing an extra membership fee on top of any separate pricing strategy is proportional to the membership fee to the power of k. Therefore a small extra membership fee has no first-order impact on the demand for any multiproduct bundles, which enables the firm to extract more consumer surplus. When this positive effect dominates the loss of single-product demand, two-part tariff dominates separate pricing. We present conditions that guarantee such an outcome, which generalize McAfee, McMillan and Whinston (1989)'s result from two products to multiple products. The two-part-tariff effect provides a new multiproduct perspective for the wide application of two-part tariffs, complementary to the classical "single-product" efficiency-related explanation. Our results suggest that two-part tariffs can achieve multidimensional price discrimination and should be subject to similar regulatory scrutiny as other discriminatory strategies, such as mixed bundling. The theories discussed in the first two parts address market situations where participants face deterministic decision problems. However, many if not most decision making processes involve uncertainty. The third part of the thesis focuses on people's economic behavior in risky choice situations. In this context, the cumulative prospect theory (CPT) by Tversky and Kahneman (1992) has been accepted as one of the best descriptive models that reconcile, within one unified model, the major phenomena that violate standard utility models. However, the inverse S-shaped weighting of cumulative probabilities posited in CPT causes difficulties in preference representation, which hinders its application in wider situations of risky choice. We propose a simplified weighting function for CPT, the (Beta, c) model, which plays a similar role in models with risky choice as that of the quasihyperbolic discounting function in models with intertemporal choice. The (Beta, c) model has a weighting function that is linear with slope smaller than 1 on the open interval (0, 1), jumps down to 0 at end point 0, and jumps up to 1 at end point 1. It achieves highly tractable utility representation for CPT whilst preserving the basic tenets of CPT. It by construction can explain all four major phenomena of risky choice violating the standard model that CPT was developed to reconcile, including reference dependence and certainty effect. It also allows Bayesian updating (with distortions at certainty) which CPT cannot accommodate. We systematically examine the (Beta, c) representation of discrete and continuous lotteries, and provide four applications which illustrate that the (Beta, c) model is a useful work horse to analyze implications of preferences exhibiting certainty effect and reference dependence in standard models. More detailed critical discussions are provided in each part of the thesis.