|Looking at the US 706 Broadband Report|
|Monday, 26 July 2010 15:52|
(Folks. Consider this an unfinished draft. I wanted to get the basic facts out. I posted it as is pending improvements. Ideas welcome.) I discovered several anomalies in the details of the 706 report, none of which invalidate the conclusion. Mark Wigfield is right "in terms of giving us a snapshot of availability in the US, and zooming in as much as possible to the local level, this 706 Report is a giant step forward from past reports."
The FCC 's main conclusion is the U.S. is doing a terrible job reaching the last 5-10% of homes. Britain has been at 99% for years, including Scottish islands and Welsh highlands, proving what's possible. Anyone who knows the data - especially the near freeze at most carriers the last several years - acknowledges the U.S. is doing a terrible job connecting the last ten million or so homes.
Lobbyists screaming "but we're good to the other 90% or 95%" haven't read the law and probably haven't bothered to look at the last few years. Section 706 requires the FCC to report "the availability of advanced telecommunications capability to all Americans." In elementary school they taught me all is 100%, not 95%. Not even a $2M/year D.C. silver tongue can convince me otherwise.
All the public evidence suggests that carriers serving 60-80% of American homes virtually stopped extending their deployment early in 2009 or before. Carrier after carrier cancelled orders for equipment. They thought it would be stupid to spend their own money when the government might pay in the stimulus and now USF. (They usually ignore the clauses in the stimulus bill that money cannot be spent where the carrier would already be building.) It's becoming far worse in 2010 as they smell more money from "USF Reform."
There are several items important to anyone who wants to understand how to bring broadband to those without. They imply the real figure for unserved could be 60-100% higher than the 5% figure the broadband plan developed.
We won't know until we get good maps from each state. The issues:
* The Black Swan Unserved 1-4+% of U.S. homes are not served because of uncommon issues not captured in the FCC analysis. Some people within range of the exchange cannot get DSL because their line goes to an old field terminal that's hard to upgrade, requires a tech to go the building to replace pairgain equipment, or just have rotten old lines (Brooklyn). I was amazed a while back to discover that 10% of both New York and Chicago couldn't get DSL for reason like that despite the short loops in dense cities. Cable possibly has similar problems of an occasional inadequate CMTS, etc. The 99.5% of homes covered at Comcast and Cox sounds impressive unless you are one of the other 500,000 homes. The cablecos claim they count as "unserved" the homes in-district that are in low density pockets. I bet that in fact some homes that they won't serve without a $12,000 or $18,000 charge are not counted. No one really knows but the carriers and their agents.
* These 1-5M homes are "Black Swans" that didn't change the count significantly when 25M homes were unserved. With fewer unserved today, these homes are 15% to perhaps 30% of the problem.
* There's no reason not to include this data because most carriers have it available in an on-line. "Pre-qualification database." It's expensive to sign up a customer, send them equipment, and then have it not work. To prevent problems they've done extensive testing to identify these anomalies. The carriers in turn give access to the records of the pre-qual to literally thousands of outsiders in real time.
* You're often offered broadband when you buy a new computer in a store, order on the phone to Dell, or go to certain sites on the Internet. They in turn check your address against the prequal database. Best Buy, Dell, and other sales agents don't want angry customers. In the early DSL Hell days, the carrier records were in terrible shape. Dell considered stopping selling DSL to new computer customers because they had too many problems. Since then the databases have become far more accurate.
* These millions of homes have been ignored. Most of the D.C. experts don't even know they exist. One hundred thousand or so are in Time Warner territory. Michael Copps' FCC shouldn't have approved the Time Warner spinoff with the $B tax break without requiring service for at least some of the Black Swans. The stimulus did nothing for them because it was focused on projects covering hundreds and thousands of homes. None of the proposed "investment incentives" are likely to "incent" the companies to reach them. A direct policy, such as traditional rules that telcos repair bad lines, is required. So might a targeted incentive, like $600 to the company for every previously "unservable" customer connected. That would more than cover the cost of sending a tech to do the line to a different copper pair that can offer DSL.
* That's why getting the facts right on who is unserved matters. It turns out the biggest problem for the stimulus is they set a system designed for projects for thousands and tens of thousands. What nobody realized at the time was that there were fewer homes involved and they were overwhelmingly scattered.
* There's no reason the carriers can't provide pre-qual data to the mappers. Connect Kentucky told me a while back they were using it. NTIA and the FCC need to confirm with each state they are making sure to include these homes. To make good policy, NTIA should also ensure the state mapmakers get good information about any area they discover is unserved. In particular, an unserved area that has cable TV needs different policies than one that only has a telco and many homes are too far for DSL. Individual homes and very small clusters need yet a third set of requirements.
2) According to the data in the report appendix, they counted no unserved homes in New York, Massachusetts, and Vermont. I know that's not true. I have a friend in a small community in upstate New York where they can't get anything. Vermont estimates they have 10% unserved and a telco identified 300 of them in two counties they considered serving with stimulus money. The Massachusetts Broadband proposal for funding in round 1 included 3,000 unserved homes.
That can be explained by how they analyzed the data. The FCC analysis was based information from the companies. (Form 477). Because it wasn't designed to do this count, they had to make some assumptions. If 1% of the homes in a county subscribed to broadband, they assumed the entire county was served. There certainly are some cases where half a county is unserved but enough people in the served areas subscribe to be more than 1%.
The total missed is not enough to change the conclusion but probably is substantial enough to matter. NTIA has allocated $100M for mapping and I hope they get things right.
3) The figure derived by the FCC from the 477 forms is much higher than the analysis the plan did working from available industry data. The plan came to the conclusion only 5% were unserved. The FCC 477 calculations were almost 8%. The team that worked for the plan included Jim Stegeman and other respected experts but the 477 are probably better source material. This is why the official announcement said 14M to 24M people can only get satellite. The lower figure if the broadband plan, the higher the 477 forms. I believe neither of the figures includes the black swans.
So the actual number could be 50% higher than the plan estimate, or even double. No way to tell until we get good state maps.
4) The July 2010 report said these are current figures but they actually are from December 2008. If more than an insignificant number of homes were reached in those 18 months, the unserved would be substantially fewer. During this year and a half $2B of stimulus money was allocated, other RUS programs were large, and presumably some of the companies continued the gradual expansion they had previously announced.
I believe that 2009 was the worst year for new deployments in the U.S. since the beginning of broadband last century, but I have no firm data. 2010 is not doing much better. I have no good estimate of how much higher the unserved figure is in the report compared to what it should be.
It's incomprehensible to me that in such a dynamic field multibillion dollar policies are being decided on obsolete data so I assume the FCC has an estimate. If not, something is wrong.