Thursday, March 28, 2019

Lies, Damned Lies and Statistics

Nothing remains unchanged in the modern communications business, so we should not be surprised at how much, and how fast, “average” fixed network internet access speeds increased in the U.S. market in 2018.

To be sure, the actual value or meaning of most cross-country internet access statistics is nuanced and perhaps somewhat questionable.

“Do counties with mostly 25 Mbps broadband connections fare better economically than counties with mostly 10 Mbps broadband connections?” George Ford, Phoenix Center chief economist rhetorically asks. “I find no evidence of such an effect here, at least with respect to the growth in jobs, personal income, or labor earnings between 2013 and 2015.”


“Broadband (and higher speed broadband) is not randomly distributed across geography, but rather is deployed in areas where the ratio of demand to costs is favorable, complicating the task of discovering broadband’s influence on economic outcomes,” Ford notes.


To cite just one example, “population density in counties with predominately 25 Mbps service averages 603 persons per square mile, but only 32 persons-per-square-mile for counties with predominately 10 Mbps broadband service,” Ford notes.

According to measurements by Ookla, fixed network internet access speeds increased by nearly 40 percent in 2018 alone, largely on the speed boosts instituted by cable TV providers.


But that is not the only surprising statistic one might encounter.


A 2014 study of broadband adoption in the United States and European Union contained some surprising comparisons. On several key measures--including faster speeds, 4G service, fiber to home and cable modem service, the United States had a large lead over EU nations as a whole.


By 2017, the take rates for fixed network internet access remained at 75 percent, though coverage (ability to buy) was at 97 percent.


Perhaps significantly, a European Commission report recently said that “although fixed broadband is available to 97 percent of EU homes, 25 percent of homes do not have a subscription.”


One apparent reason is that although subscription growth “was very strong until 2009, but has slowed down in the last few years, partially due to fixed-mobile substitution," according to Techknowledge.


In 2017, about 15 percent of EU homes bought internet access at speeds of at least 100 Mbps. In 2016--a year earlier--about 23 percent of U.S. homes were buying service of at least 100 Mbps.


By the end of 2018 the average U.S. fixed network internet access speed was 95 Mbps, according to Ookla meaning half the customers had speeds faster, while half had lower speeds.  

The EC study also highlighted the importance of taxes when comparing retail prices for services such as linear video subscriptions. The posted retail prices are one thing; the actual costs--including all taxes--are something else.


As the study illustrates, even if posted retail prices for U.S. cable TV video are higher than Denmark’s prices, for example, the net cost of service in Denmark is higher, because of higher taxes.


The larger point is that although most casual observers would assume the EU has higher percentages of internet access connections running at 100 Mbps or faster, that is not true. Nor is it the case that customers actually buy services at such speeds, even when available.


It is good public policy to make quality broadband service available. It is a separate matter whether consumers want such services, how much they are willing to pay for it and--most importantly--what economic and social value can be wrung from such usage.


It is difficult to impossible to confirm that speed increases beyond 100 Mbps (or any other minimum speed) actually produce quantifiable economic or social outcomes.

No comments:

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...