32 Pages Posted: 1 Apr 2013 Last revised: 31 Mar 2017
Date Written: February 2013
Objective: The paper analyzes recent trends and dynamics that affect the relative availability, utilization, quality, and value of American broadband networks versus those in other OECD nations. Most of the literature on this subject was developed five to ten years ago, prior to the deployment of DOCSIS 3, LTE, VDSL, and the smartphone, and during a period in which fiber deployment proceeded at a relative slow rate in the U.S. due to overhang of the fiber bubble of the late 1990s. Consequently, many analysts have accepted a version of conventional wisdom according to which the United States offers “second rate broadband” at high prices. While this was a defensible position in the late 2000s, the most recent data shows that it is no longer the case. If the United States is improving its position relative to OECD competitors on the important metrics, it follows that its policy framework is fundamentally sound.
Method: The paper examines data drawn from the most comprehensive and current sources: OECD surveys, FCC surveys and testing, the National Broadband Map, the Berkman Center analysis, ITU Surveys, Akamai, SamKnows, and Netindex performance data, and up-to-data industry analyst data on subscriber churn, fiber deployment, and provider profitability. It employs regression analysis to estimate the correlations between broadband adoption, price, and computer ownership as well as survey data to identify barriers to adoption. The paper develops trend lines on the relative ranking of the U. S. and OECD nations with respect to the key metrics.
Novelty: Traditional analysis of broadband performance tends to be error prone, as the method typically used combines two estimates: 1) the availability of service tiers throughout the nation; and 2) the distribution of customers across available plans. This method assumes that plans actually provide the level of performance suggested by advertising, and that consumers are aware of the current speeds advertised for the plans they choose. As measurement indicates that these assumptions are faulty in many nations, the paper pursues a different avenue and examines speeds on the basis of measurement data collected by Akamai, Netindex, and SamKnows. It finds the Akamai data most useful and suggests these data are a powerful resource for policy analysis that has largely been overlooked.
Results: The paper finds the downward trend in the adoption of high performance broadband plans and in overall broadband performance that characterized American broadband service in the late 2000s began to reverse in 2010. In one measurement, the U. S. ranked 22nd overall in the average speed of shared IP connections in Q4 2009, but ranked 8th by the same measurement in Q3 2012. It finds that the adoption of broadband by computer-owning American households exceeds the OECD mean, and that the deployment of DOCSIS 3 and LTE is stronger in the U. S. than in other nations.
The paper finds that the prime benefit of facilities-based competition is the creation of market dynamics in which service providers are able to compete on the basis of performance and coverage as well as price and customer service. It notes that the European Commission is developing a revised policy framework in order to create incentives for performance-based competition between providers.
Keywords: broadband, deployment, adoption, performance, price, international, trends
JEL Classification: L96, L98
Suggested Citation: Suggested Citation
Bennett, Richard and Atkinson, Robert D. and Stewart, Luke A., The Whole Picture: Where America’s Broadband Networks Really Stand (February 2013). TPRC 41: The 41st Research Conference on Communication, Information and Internet Policy. Available at SSRN: https://ssrn.com/abstract=2240589 or http://dx.doi.org/10.2139/ssrn.2240589
By Daniel Lyons