Network growth: Theory and evidence from the mobile telephone industry

Firms in mobile telephone markets charge one another for delivering calls to subscribers. These so-called termination charges are controversial from a policy perspective since they are fairly high and stable. We present a model of consumer and firm behavior in mobile markets in order to identify the role of termination charges in determining the market equilibrium. We analyze three segments of the mobile market: termination, calling, and subscribing. Our model predicts a “waterbed effect”, that is, high termination rates will be associated with low subscription prices, if preferences are the primary source of variation in termination rates. If costs are the main driver of termination rates our model predicts a “tide” hypothesis in which high termination rates exist alongside high subscription prices. We test these and other predictions from our model using international data on mobile subscriptions per person. We find results which are broadly consistent with our model. More specifically, we find evidence that mobile termination rates are positive and significantly related to mobile phone adoption. This result is robust to the inclusion of a variety of other structural, institutional, demographic, and income controls. We also find that competition, internet subscriptions, and a free press are positively associated with mobile phone adoption while fixed termination rates and inequality slow the adoption of mobile technologies. ∗ Authors may be contacted by email. Cunningham: [email protected]. We would like to thank Pedro Pereira, Elena Argentesi, Nathan Goldstein, an anonymous referee and conference participants at the 6th Annual International Industrial Organization Conference, the 2008 SIDE-ISLE conference, and the 2009 Georgetown Center for Business and Public Policy Wireless Technologies Conference for many helpful comments. We obtained our termination market data from Arbinet’s Voice on Thexchange Marketplace. Hundreds of Member carriers trade billions of minutes a year on this system. We maintain responsibility for any errors or omissions.