Fabricating a Broadband Crisis? More Evidence on the Misleading Inferences from OECD Rankings

Download report: Misleading_Inferences_from_OECD_Rankings_PhoenixCenter_72010.pdf

The late Senator Daniel Patrick Moynihan was famous for observing that “Everyone is entitled to his own opinion, but not his own facts.” If only current policymakers heeded such wisdom when it comes to citing international rankings of broadband penetration as a justification for aggressive public policy interventions.

Indeed, there are many policymakers (and policy peddlers) in this country—including current Federal Communications Commission Chairman Julius Genachowski—that (apparently) believe that the U.S. is “falling behind” in broadband adoption and, therefore, aggressive regulatory intervention is required to remove this blight from our national reputation. This belief is derived largely from data on broadband connections collected and reported by the Organization for Economic and Cooperative Development (“OECD”). Every six months, the OECD releases its data on per-capita broadband connections for its thirty member countries, and these countries are listed in descending order based on per-capita connections.  This practice presents the data in terms of a rank, and that rank has (regrettably) become the standard by which to judge the successes and failures of broadband policy in this and other countries.

In numerous papers, I have debunked the idea that the OECD rankings can be used as a measure of relative performance. One reason per-capita connections are an invalid measure of-capita connections are an invalid measure of broadband penetration is that each country has its own unique maximum value for the measure (all share zero as the minimum).  In other words, if in every OECD country every household and business had broadband (the “Broadband Nirvana”), you would still observe large differences in their per-capita subscription rates.  As such, each country’s per-capita subscription rate has its own scale, and consequently, comparing per-capita connections presents the quintessential apples-to-oranges problem. Moreover, in this Nirvana, the U.S. ranks 20th, five spots below its present position.  Consequently, if near ubiquitous adoption across the OECD is the expected outcome (even for just the more developed economies), then the U.S. will always have a middling rank. 

Also, as I have argued before, when interpreting rank it is essential to first establish an expectation of rank.  Without a meaningful expectation, it is impossible to say whether our observed rank is too high, too low, or just right.  In PERSPECTIVE NO. 08-03, Broadband Expectations and the Convergence of Ranks, I provide compelling evidence that the U.S. is meeting expectations on broadband connections per capita (a rank close to 15th is expected even with good performance).That is, there is no “broadband crisis.”