Jon Brodkin at Ars Technica and Brian Fung at The Switch have posts featuring a New America Foundation study, The Cost of Connectivity 2013, comparing international prices and speeds of broadband. As I told Fung when he asked for my assessment of the study, I was left wondering whether lower prices in some European and Asian cities arise from more competition in those cities or unacknowledged tax benefits and consumer subsidies that bring the price of, say, a local fiber network down.
The report raised a few more questions in my mind, however, that I’ll outline here.
The NAF report concludes that US consumers would see lower prices if there was more competition. Or, as Brodkin says,
What’s the takeaway from all this data? It’s not a surprising one: lack of competition makes for bad choices.
I don’t disagree with the sentiment but the report makes no mention of competition data that would tend to support their broad conclusion. How many wireline competitors are there in Paris, Seoul, NYC, and Prague? Is there correlation between more competitors and lower (quality-adjusted) prices? NAF never tells us.
I raise this because the US actually has more market fragmentation (measured by HHI, an established tool used by antitrust agencies) for wireless carriers than many European countries, yet higher prices. If the cause of relatively higher prices is lack of competition, per NAF, wouldn’t we expect lower advertised prices in the US for wireless subscriptions because there is more competition? The fact that the US has higher prices indicates there are other factors besides number of competitors that drive price.
This lack of discussion of competition data is the major gap of the NAF study. Despite concluding that lack of competition is the problem in the US, the authors seem uninterested in rigorously examining the state of competition in the cities and countries they highlight (but perhaps this will be taken up in their promised forthcoming full report).
Another problem is that the report mostly consists of documenting advertised download speeds. While sensible since it’s easily available, this reliance on advertised speeds warrants a warning. The FCC publishes an annual report comparing international broadband offerings and noted in its 2012 report that advertised speeds are a troublesome metric that often misrepresent what consumers actually see. Recently in the UK, for instance, broadband packages with an advertised speed of 24 Mbps featured an actual speed around 5 Mbps for the typical customer. (Recent truth-in-advertising reforms have made this problem less likely–but only in the UK.) Different countries have different methodologies and advertising standards, and the US carriers tend to have more “honest” advertised speeds. Unlike the FCC, which carefully notes issues with these sorts of measurements, NAF makes no mention of possible discrepancies despite letting advertised speeds do a lot of the work that leads to their conclusion.
My final dispute is with the inclusion of “triple play” subscriptions (combination voice, Internet, television service). Documenting the price for this bundle is next to worthless because the quality of the television package is a substantial reason for buying. For example, what do we learn from the fact that you can get phone service, 20 Mbps Internet speeds, and a television package in Riga, Latvia for $22? Is the $99 price in San Francisco high because of a lack of competition or because the television package is so much better than the channels offered in Riga? Or because of regulatory distortions in US television policy? (Retrans, anyone?) Without some measure of quality-adjusted price, the triple play comparisons are just noise.