CA AB BL LE E H HS SI I P PR RI IC CI IN NG G C  

Contents

Introduction   3

  Methodology  5

  Data  6

  Analysis  8

  Conclusion  9

  References  11

  Da an n S S. C Cr ry yn ns s 2 2 D  

Introduction

The last mile connection is the bottleneck preventing the ubiquitous availability of broad-band. Consumers are currently presented with four last mile solutions: cable, satellite, DSL, and fixed wireless. Of these alternatives, only cable and DSL are widespread and priced relatively competitively. Echostar, a leading consumer satellite provider, recently stopped selling broadband services because the product was unprofitable (Barnako, 2002). High-speed Internet access (HSI) provided by cable companies has emerged as the leading broadband solution. High-speed Internet penetration of high-speed Internet ready households passed is currently around 9.3%, versus about 7.8% for the RBOCs with DSL (Zia, 2002). The problem faced by the cable company is how to price the service and further drive penetration. The goal for the cable company is to maintain average revenue per user (ARPU), while driving incremental penetration. Because cable is a shared bandwidth medium, current users can experience a decrease in their quality of service as the cable company adds more subscribers to a node. The speed and bandwidth needs of consumers vary widely. An emerging cable modem standard, DOCSIS 1.1 (Data over Cable Service Interface Specification; CableLabs, 2002), will allow cable companies to offer ‘dynamic throttling’ of speed and different bandwidth capabilities. Cable companies are planning to use DOCSIS 1.1 to offer tiered pricing. The higher priced tier would be offered a guaranteed quality of service, while the lower priced tier would offer a lower speed service.

Because access to DSL is only available to consumers who live within two to three miles of a local telephone company's central office switching facility, cable companies effectively possess a monopoly on broadband internet access in many areas of the country. In areas closer to a central

Da an n S S. . C Cr ry yn ns s - 3 3 - - D -

office, there is an effective duopoly (cable and DSL) on high-speed Internet access. The academic literature has examined the revenue-maximizing pricing of a congestible resource in a monopolistic and duopolistic framework. The purpose of this paper is to compare the revenue-maximizing pricing scheme(-s) suggested by theory against the pricing scheme(-s) proposed by the cable companies.

The literature, in general, suggests that a revenue-maximizing strategy will at a minimum reduce the “Tragedy of the Commons”. Congestion externalities arise when users of a shared resource take into account their own costs and benefits from usage, but ignore the congestion, delay, or exclusion costs that they impose on other users (Falkner et. al., 2000; MacKie-Mason & Varian, 1995). Studies of pricing congestible resources show that economists tend to favor pricing mechanisms, as opposed to rationing or quota systems, as a means of alleviating congestion (Falkner et. al., 2000). Ideally, the pricing mechanism selected should internalize the externality created by the congestion caused on a network.

Historically, internet access within the US has been priced using at a flat rate or a usage based scheme. Mason (2001) shows that “a monopolist (weakly) prefers to use a two-part tariff compared to a fixed (non-usage based) price”. Following the influential work of A.C. Pigou in 1920, a two-part tariff allows the monopolist to price discriminate more effectively, extracting more surplus from its consumers (i.e. third-degree price discrimination). Duopolists also consider the ability to price discriminate when selecting a pricing scheme (i.e. second-degree price discrimination). However, duopolists, unlike monopolists, m ust take into account the increased competition created by more price points. Mason (2001) finds that it is always a best-response for a firm to use a two-part tariff when its rival is using a flat rate. However, both firms earn lower

Da an n S S. . C Cr ry yn ns s - 4 4 - - D -