X

Is the U.S. really so terribly behind in broadband?

Critics say one way to find out is to change the way federal regulators measure broadband penetration rates. Some refinements may soon be on the way.

Anne Broache Staff Writer, CNET News.com
Anne Broache
covers Capitol Hill goings-on and technology policy from Washington, D.C.
Anne Broache
5 min read

It has been a sob story for American politicians, technology companies, and advocacy groups for at least half a decade: the United States, which developed the Internet, is consistently trounced by the likes of Korea and Iceland in how widely it delivers broadband access to its citizens.

But do the latest statistics from the often cited Organization for Economic Cooperation and Development (OECD), which ranked the United States 15th out of 30 countries for broadband penetration rates in 2006, tell the whole story?

Maybe not. Which is one reason, it seems, that the Federal Communications Commission is coming up with a new approach that may change the way broadband penetration is evaluated.

One argument is that a truly accurate picture of the U.S. broadband marketplace--and, in turn, the process of achieving "universal" coverage--won't be possible until better data is available. Yet the FCC continues to work with data that deems 200 kilobits per second (Kbps) service "high speed" and to consider such access to be widely available even in ZIP codes that may, in reality, have only one connection.

On Wednesday, this may start to change. That's when the FCC is planning to consider--and most likely adopt--changes to the way it measures "reasonable and timely deployment of advanced services to all Americans," as required by federal law. (A Democratic bill approved by a U.S. Senate committee last year may have helped to apply additional pressure.)

At the moment, Internet service providers are required to fill out forms about their service offerings and submit them to the FCC semiannually. From those forms, the FCC produces reports that attempt to reflect on the state of Internet access availability, and in recent years, they have claimed to document "significant and steady progress" in broadband availability nationwide.

But critics charge that the information the FCC collects is not granular enough to be useful. Judging from their statements in recent months, there's consensus among the five commission members, too, that some refinements are necessary.

"By improving our data collection, we will be able to identify more precisely those areas of the country where additional broadband deployment is needed," FCC Chairman Kevin Martin said last spring, when the agency began soliciting public comments on how it should proceed.

How significant those changes will be--and what happens next--remains to be seen.

Cable and telephone companies have generally been more hostile to major changes in the FCC's process. One major clash, for instance, concerns whether broadband providers should be required to do more reporting about the price of their offerings.

Liberal consumer advocacy groups like Free Press, Consumers Union, and the Consumer Federation of America argue that would be a productive step--and that the FCC should, at a minimum, survey Americans about how much they're paying for Internet service--to compare price and service menus across different locales.

But major cable and telephone companies strongly oppose the idea. They contend that no meaningful price comparison could be achieved through government collection of that data because a wide variety of factors--including the term of contract, bundling with other services, and promotions--influence an individual's bill and because there would inevitably be a lag time in publishing those statistics anyway.

Tracking broadband speed
The FCC is famously quiet about its policy plans before formally adopting them, but Martin previewed some of them during a briefing with reporters earlier this month.

First, Martin said, the FCC plans to "move away" from one often criticized component of its data collection process: a requirement that Internet service providers report all of the ZIP codes in which they have as few as one broadband connection. The FCC has used that data to suggest in recent reports that there are broadband connections in every state but no "high speed" connections in 12 percent of all U.S. ZIP codes.

As the FCC itself has acknowledged, that technique has the potential to significantly overstate broadband availability. For instance, in a rural area, one or two customers who happen to live close enough to telephone or cable facilities could get service, but people located miles or even blocks away may not.

It's not clear exactly what a new proposal would entail, although the commission said it's weighing whether to ask providers to submit customer counts and number of houses served along with the ZIP code data. Free Press, Consumers Union, and the Consumer Federation of America have suggested that the FCC should ask for the number of subscribers not just within a particular ZIP code, but in an even more granular way--say, using the ZIP code plus its four-digit extension, or through census blocks, which California state officials have used.

The cable industry says it supports the nine-digit ZIP code idea, but there's considerable resistance from Internet service providers like AT&T and Verizon, who suggested the requirement would be burdensome, costly, and potentially "competitively sensitive"--that is, allowing rivals to target their operations based on where they see more or fewer customers.

Martin said the commission also plans to overhaul the way it accounts for broadband speeds offered by various providers. Right now, it lumps those service offerings into six big buckets: 1) exceeding 200Kbps in the download or upload direction; 2) greater than 200Kbps and less than 2.5 Megabits per second (Mbps) in the faster direction; 3) between 2.5 and 10Mbps in the faster direction; 4) between 10 and 25Mbps in the faster direction; 5) between 25 and 100Mbps in the faster direction; and 6) 100Mbps or higher in the faster direction.

The trouble with that approach, critics say, is it inherently raises the possibility that the number of faster connections could be overstated and the number of slower connections could be understated.

By Martin's description earlier this month, the FCC is planning to break down the lower-end categories into six smaller tiers: 200 to 768Kbps at the low end, 768Kbps to 1.5Mbps for "basic" service, 1.5 to 3Mbps for "high speed" service, 3 to 6Mbps for "robust" service, and a 6 to 10Mbps for "premium" service. He didn't say, however, whether the FCC plans to collect information about those speed offerings on anything but a nationwide basis, which is currently the procedure.

None of the Internet service providers that filed comments with the FCC recommended chopping the broadband tiers into as many pieces as Martin suggested. The cable industry, for example, argued against changing the tiers at all, saying it would make comparisons over time more difficult. (That position isn't exactly surprising, since cable's typical 6Mbps offerings already fit neatly into a category that also includes speeds above that level.)

The FCC is also considering collecting more information about wireless broadband and Internet phone subscribership and use, although it's unclear how those issues will be incorporated into the order it's expected to consider Wednesday. More details may be left for future rulemaking jaunts: the FCC, as is typical, is also expected to ask for additional comments.