The New America Foundation’s (NAF’s) “Cost of Connectivity” (0) report – now in its third year – has been widely publicized (1) by consumer rights warriors (2) and Internet policy wonks working for American media (3) and public interest groups (4), but its flaws are so severe (5) that it adds little substance (6) to the discussion of broadband and Internet policy (7). The 2014 edition of the NAF report improves on previous versions by dropping the pretense of comparing 200-channel US triple play packages (which combine cable TV with broadband and telephone service) to 50 channel bundles in other nations that are confined to free-to-air content, admitting that it’s not meaningful to compare such bundles.
Despite this correction, the report’s primary defect remains: it ignores thorough and methodical data collection on broadband price, performance, and usage in favor of an odd collection of advertising claims that bear little resemblance to actual broadband conditions in the US and our competitor nations.
ISP costs are high in the US because of population distribution, and the large carriers who serve cities also serve exurbs where costs are higher. Only Canada has a more dispersed population than the US (8) within the Group of 7 (G7), and our ISPs charge uniform prices across their entire footprints.
It’s also well known that advertising claims in Europe and Asia grossly inflate the actual speeds of advertised plans, while American broadband advertisements understate actual speeds; the SamKnows surveys conducted in the US and Europe show that American broadband packages outperform their claims (10) while European plans deliver less than 76% of promised speeds (11).
This is not a new finding: the UK has been publishing SamKnows data since 2003 (12), long before NAF began publishing its advertising surveys; the EU has published two European SamKnows surveys, and the US has published four. Japan is reluctant to publish speed tests, but independent researchers such as Toshiya Jitsuzumi have shown that actual speeds in Japan are no more than 35% of advertised ones (13). Korea claims to provide 100 Mbps and 1 Gigabit/second connections to most of that nation’s area, but measured average peak connection speeds in Korea (14) are only 68.5 Mbps.
Another flaw of the NAF report is its focus on select US cities. Local surveys are much less accurate than nationwide surveys with respect to price. The EU contracts with Van Dijk consultants (15) to see how its prices compare to those in other regions, and their study says the US offers lower prices than the EU (16)for speeds up to 12 Mbps.
Point Topic also performs global price surveys (18). City-level surveys, like that of NAF tell us very little, even if the data were accurate.
And the NAF data is not remotely accurate. NAF includes some plans that are not widely available in the surveyed areas and excludes others. For example, it lists Webpass and Astound as leading wireline service providers in San Francisco (despite limited service areas) but excludes Paxio and Sonic, the two providers who offer the best ratios of speed and price. NAF includes a $77.45 package from Astound providing 110 Mbps and a $30 package from Webpass providing 200 Mbps, but Sonic offers a $40 package providing 1000 Mbps (19) (and telephone service!) and Paxio offers a $108.50 plan for 1000 Mbps (20) in the San Francisco Bay Area.
If these plans were included, privately owned Paxio would beat taxpayer-subsidized Lafayette, LA and Bristol, VA on the NAF speed leaders chart (21) (see page 12) and privately owned Sonic would beat Google Fiber and heavily subsidized Chattanooga, TN for the top spot in the US.
That’s a very serious omission. NAF may very well claim that it excluded Sonic and Paxio because they only provide service in the Greater San Francisco Bay Area rather than the city proper (they want to serve the city, but the Planning Commission forbids it (22)), but then they would also have to explain why they surveyed the 14th and 24th largest cities in the US rather than four of the top six Primary Statistical Areas (23). That explanation would need to be very good to overcome the factual errors Roslyn Layton has found in previous NAF surveys (24).
Despite these flaws – dubious methodology combined with a less-than-diligent collection process – the Cost of Connectivity report resonates with writers eager to explain why its counter-factual claims must be true. David Lazarus of the L. A. Times claims (25) that open access regimes provide the highest broadband speeds, for example, despite the fact that the highest broadband speeds in most nations are only found on networks exempt from open access rules (26), such as the cable and fiber networks in Japan, Korea, Germany, and the UK. But let’s not let mere facts get in the way of a good story.
Honest surveys of broadband prices, speeds, and usage show that we pay somewhat higher prices than citizens of several other nations. However, the different prices we face are the result of our heavy usage (more than any other nation but Korea), our vast, sparsely populated rural areas, and our dearth of cost-hiding subsidies. That’s to be expected. ISPs who serve limited areas or who are heavily subsidized can overcome these effects, at least for a short while, but you can’t wire a large nation by excluding the hard-to-serve areas that drive up ISP costs.
But that’s a “dog bites man” story, not an outrage. The US should strive to be number one on the list of nations that provide and use the most advanced Internet applications and services (which we are), but we should leave the number one spot on the list of lightly-used broadband networks serving small areas to those in need of a boost to their national self-esteem.