Lawrence Spiwak offers an important caution about OECD Broadband Rankings and analysis of broadband and wireless.

Absent a better quality of analysis, U.S. policymakers cannot rely upon the OECD’s research as justification for additional regulation. Let me provide just a few examples. Twice a year, the OECD publishes rankings of per-capita broadband penetration among its member countries. Though the United States was at the top of these rankings back in the early 1990s, when the Internet was in its infancy, it has fallen to the rank of 15th.

Needless to say, this statistic has been bemoaned widely by politicians on both sides of the aisle as evidence of some sort of policy failure. However, the problem is not with our vibrant ICT sector, but with the way the OECD calculates the index and how others interpret it. Specifically, the OECD simply adds fixed-line connections by both businesses and households and then divides by population. The more people live in the typical home, the lower your rank.

Consider the following thought experiment: Assume every home and every business in the OECD had broadband in some sort of “nirvana” scenario. Though logic would dictate that all member countries should be tied for first place as we would be “equal” under the OECD’s pedantic methodology, the United States actually would fall to a rank of 20th. That is five places “worse” than where we are now – with no room for improvement because there already would be 100 percent broadband penetration.

[...]

Finally, we have the OECD’s recent attempt to show the United States has the highest mobile rates in the world. Under the best of circumstances, formulating a price index that can be used meaningfully to compare mobile prices across countries is exceedingly difficult. The OECD approach was to set up three “baskets” of high, medium and low customers in each country and only use, once again, prices from a single U.S. carrier. However, not only did the OECD fail to accurately model usage levels for U.S. consumers (specifically understating actual usage by a factor of 2 to 5), but it ignored essentially all of the pricing options from multiple providers available to consumers today.

Also: See George Ou’s previous work on Measured Broadband Versus Advertised Broadband Rankings.  Real world measurements tell a very different story from estimates.