No economist has ever used a metric such as "jobs created or saved" to measure employment statistics, though that is the "metric" being peddled to the American people as a serious indicator of how well the econonomic stimulus plans have worked. No economist appears to use the metric because there is no way to collect the fundamental statistics that would be required to make such a measurement. 

Now the Federal Communications Commission has decided to redefine "broadband" when it collects its statistics. To be sure, there is nothing wrong with periodically revising the scale or scope of entities from whom statistics are collected. Nor is there any problem with periodically revising the quantitative measures in line with changed historical circumstances. 

So the FCC (News - Alert) has decided to define "broadband" as "speeds of at least 4 Mbps and actual upload speeds of at least 1 Mbps." Many would note that the old metric of 200 kbps was too low, in fact absurdly low. That is both correct and yet misleading. "Broadband" has been defined by international telecommunications bodies, for decades, as being any speed at or above 1.5 Mbps. Engineering textbooks, telecommunication textbooks and standards bodies have called bandwidth at or above 1.5 Mbps as "broadband," while bandwidth at or below 64 kbps is "narrowband" and any speeds in between 64 kbps and 1.5 Mbps are "wideband."

You immediately see the problem. Cable operators market speeds much more in the range of 50 Mbps as "wideband," though the standards bodies use a different definition altogether. 

Here's the other problem: we now have a set of data collection practices and definitions that instantly make all third-generation wireless broadband networks, and all small business T1 services, "not broadband," by definition. 

This defies logic, good sense and reality. Using the FCC's definition, nearly all small businesses no longer have broadband. Neither do iPhones. 

In fact, that has been a key sticking point for wireless service providers in dealing with any formal definitions for what "broadband" speeds are, or should be proposed for the "National Broadband Plan."

Wireless networks, for physical reasons, will always have less theoretical capacity, and likely always will have less "lit" or "available" bandwidth than fiber networks. Setting a single numerical value for broadband, as well as targets, therefore is of huge importance for wireless providers who fully expect to be providing the majority of broadband connections in several, or a few years, depending on which forecast one wishes to use. 

A "4 Mbps" definition fairly immediately makes T1 service and 3G wireless broadband "not broadband." This might be an annoyance at the least, but could be a major business factor if wireless providers are, for example, not allowed to bid for universal service funds, for example, which require a  "4 Mbps" threshold. 

"It is by this benchmark that we find that broadband remains unavailable to approximately 14 to 24 million Americans," the FCC says. What the FCC did not say is that it also has made 3G and small-business T1 services "not broadband."

By redefining "broadband" in this way, the FCC is free to conclude that "broadband is not being deployed to all Americans in a reasonable and timely fashion." Nobody would disagree that there are some households without any wired access, or without wired access capable of reasonable speeds. 

Executives at firms that provide satellite broadband will, once again, be found muttering quietly that nearly every isolated U.S. household can be reached, today, by at least two different satellite providers. If the real intent was merely to get 4 Mbps capability to every household, then we effectively already are there. 

HughesNet already sells a 5 Mbps package, for example. 

That is not to say consumers will always be happy with any of their providers. A Los Angeles iPhone user I recently talked to "will not answer the phone because the calls drop so frequently." That isn't to say this user has any intention of ditching the iPhone. It is to say consumers will have complaints, some of them serious, about virtually every communications service, from time to time, or even "quite often."

But it is a bit disingenuous to set a "4 Mbps" standard, claim the country has a gap, and not qualify that statement at all times by saying the claim applies only to fixed terrestrial connections, not to mobile or satellite services. That speaks to the coverage issue.

But as the MIT (News - Alert) researchers also have pointed out, even the commonly-experienced end-user speeds might be as much as twice the rate many believe is the case.

Recent tests conducted by researchers at the Massachusetts Institute of Technology (MIT) say that previous estimates of U.S. broadband speed may have underestimated just how fast our national networks really are. In March, the FCC said that the broadband network was only half as fast as advertised.

However, the MIT study found that those measurements didn’t fully measure the speed of the “access network,” which Internet service providers (ISPs) control. For example, using the best method, Ookla (News - Alert)/Speedtest, current typical speeds are 7.7 Mbps, not 3.8 Mbps.

According to the study, a simple figure for broadband speed isn’t sufficient to understand the quality of the nation’s digital infrastructure, and it’s just as affected by a user’s computer and the location of servers being accessed as it is by the ISP.

That's a bit akin to attributing all of an Apple iPhone's dropped call performance to AT&T (News - Alert), and attributing zero to the iPhone's design, to the extent that the device itself can cause dropped calls.

Worse are the policy or economic distortions which will occur as broadband policies are applied uniformly over a range of technologies with distinct features and capabilities. We saw a preview of such issues in the funding of "broadband stimulus" projects, where various contestants sparred over uniform definition that did not take account of the different network characteristics.

Mobile 3G networks cannot match the raw speed of cable or DSL networks, but provide mobility, faster time to market and lower investment cost. Satellite networks have propagation delay constraints as well as bandwidth constraints, but are ubiquitous and "lowest cost" in terms of getting service deployed immediately, to very-isolated locations. 

In a world where a majority of connections are of the mobile type, it will not make sense to keep focusing exclusively on fixed wireline access, rather than the total fabric of broadband connections, many of which will be optimized for different use cases. 

The ironic result of redefining "broadband" is that the FCC now has dropped 3G and T1 services from consideration as "broadband" access services. Real live people and users might find that incongruous. 

Aside from that, how does the FCC avoid the charge that it has simply "created a new problem" so it can prove a solution is needed? Policymakers virtually everywhere want the fastest-possible broadband deployment and features. 

End users tend to say they are happy with the speeds they've already got, and their buying patterns suggest that is the case. Demand for 50Mbps or 100 Mbps services is not strong. 

Network service providers are happy to supply higher bandwidth so long as they can do so profitably. 

In the end, progress is not possible unless policy, investment and demand align. 

Nor is the decision to apply the "4 Mbps" definition immediately likely to assuage some thinking that the FCC is, among other things, closing the door to wireless service providers for at least some USF funding, as 3G networks will not likely meet the minimum threshold to qualify. Of course, rural fixed line providers will like that just fine, as they have been complaining that wireless providers should not get funding.

Defining wireless service providers out of eligibility might not have been the intent, or not the main intent of the definitional change. There are reasons aplenty for wanting a number consistent with the stated objectives of the national broadband plan. 

The issue is really the possible sense that politics is interferring with dispassionate analysis of the data. "Jobs created and saved" is a non-standard, impossible to verify political category, not a standard econometric measure. Some observers are going to think a political measure is being applied in the broadband arena as well, as the definition will have political ramifications.