Network Growth: Theory and Evidence from the Mobile Telephony Industry

Transcription

1 Network Growth: Theory and Evidence from the Mobile Telephony Industry Peter Alexander Adam Candeub 1 Brendan Cunningham 2 1 Michigan State University 2 US Naval Academy Wireless Technologies Conference 2009 Georgetown CBPP 1 / 12

2 Background and Policy Network Interconnection Firms offer consumers a networked good / service in variety of settings. Banking - automated teller machines, Internet Service Providers - computers, Telecommunications - fixed line or mobile phones. Consumers derive utility from intra-firm networking and inter-firm networking. Appropriate pricing of interconnection? Relationship of interconnection pricing to access prices? 2 / 12

3 Background and Policy Regulatory Considerations Inter-firm information transmission has low marginal cost. Interconnection prices, or termination rates, fairly high. Our sample, median of 15 cents / minute, international calling. Regulatory concern: termination bottleneck generating inefficiently high monopoly prices for interconnection and consumer access? Alternatively, profitable interconnection may encourage network expansion and investment. 3 / 12

4 Model Foundation Assumptions Assume there are three segments of mobile market: termination, calling, subscription. Sequential three-stage process in mobile market: 1 extensive firms price subscriptions and consumers choose a mobile provider 2 intensive consumers choose volume of calls, firms price minutes, 3 termination firms choose interconnection prices Model yields predictions for equilibrium network size and the relationship between termination rates and subscriptions. 4 / 12

5 Termination Rates Equilibrium Proposition Equilibrium termination rates are a constant markup over the marginal cost of terminating calls. Termination rates increase with the marginal cost of termination and decrease in the elasticity of demand for termination services. Equilibrium termination profits per subscriber decrease with both the marginal cost of termination and the elasticity of demand for termination. 5 / 12

6 Calling Call Supply c = c 0 + αβt c + (1 α)t f is the composite marginal cost. Proposition In equilibrium the price of a minute of calling is a constant markup over the total marginal cost of originating a call. Consumer utility from calling and a firm s profits from the calling activity of a single subscriber are both decreasing in the total marginal cost of originating calls. 6 / 12

7 Subscriptions Subscription Demand n mobile firms horizontally differentiated (located on a Salop circle with mass of L consumers). Consumers compare subscription price, calling utility, and differentiated product features and subscribe to firm yielding greatest overall utility. Mobile firms compete with nearest neighbor in terms of product features. Prices and product features yield a mass of subscribers for a mobile firm. 7 / 12

8 Subscriptions Equilibrium Mobile Network Market Size Proposition The market-wide mass of consumers subscribing to mobile services is increasing in the number of mobile operators and per-subscriber profits from calling and termination. The number of subscribers decreases with the total marginal cost of originating calls, the marginal cost of adding a subscriber to a network, and the utility from the outside option. 8 / 12

9 Hypotheses Waterbed Effect Termination rates negatively related to subscription prices and positively related to total network size. Sufficient conditions: (1) variation in termination rates across markets must be driven by differences in elasticities of demand for termination services (2) a small fraction of all originating calls are mobile-to-mobile and terminate on a competitor s network. 9 / 12

10 Hypotheses Tide and Countervailing Effects Termination rates positively related to subscription prices and negatively related to network size. Sufficient conditions: 1) significant cross-mobile network traffic, 2) marginal cost of termination primary source of variation Third possibility: no relationship between termination and network size. 10 / 12

11 Empirical Results Cross-Country OLS Regression Statistically significant results: Larger mobile networks emerge with higher mobile termination rates (waterbed), more mobile operators, internet penetration, press freedom, and income. Smaller mobile networks with higher fixed termination rates (tide), income inequality. Level of income mitigates impact of inequality. No statistically significant results for population density, population age, language fractionalization, government ownership, party paying regime. 11 / 12

12 12 / 12 Summary Presented a model of termination, call, and subscription pricing in mobile markets. Model predicts possible existence of waterbed and tide effects. Empirical evidence from cross-section of countries provides evidence in support of waterbed hypothesis. Results are relevant for other interconnection markets. Network neutrality may impact the pricing of broadband subscriptions.