How much to get ConnectED?

Warning: this post contains irresponsible speculation which leads to unpleasant projections.  Read at your own risk.

According to numerous news reports, "administration officials" have said the new ConnectED program might temporarily add "no more than 40 cents" to monthly phone bills.  Let's take a look.

OK, first: 40 cents per what?  The news reports all say 40 cents per bill.  Does that mean if my family is on a single cell phone account, we pay 40 cents, but if we have 3 accounts and get 3 bills, our cost will be $1.20/month?  Will a business with a hundred phone lines also be paying an extra 40 cents/month?  Or do they really mean 40 cents per line?

Since a flat 40 cents/month would require a complete change in the way contributions are calculated, I'll bet what they mean is that they'll increase the USF contribution factor and the result will be a 40 cent increase for a typical consumer.  That would be a very substantial percentage increase.  Like maybe 25%.  We'd be talking about an increase in the contribution factor from 16% to 20%.

And how much does it add up to?  At this point, any calculations would be wild speculation.  So of course you know I'm going to speculate wildly.

Let's start with a "per line" flat fee of 40 cents.  According to the ITU, the U.S. has 145,875,000 landlines and 290,304,000 cell phone subscriptions.  [Am I the only one who's surprised to find that we have twice as many cell phones as landlines, and that we have about 1 cell phone per man, woman and child?]  So that's about 436 million lines.  At 40 cents per month, that would be about $2.1 billion per year.  Dang, that doesn't seem like much.  Maybe they really are talking about 40 cents per line.  [I'm guessing the ITU leaves out VoIP lines, but I'm guessing that's chicken feed.]

What about 40 cents per bill?  Well, now we're moving from rough estimate to pulling numbers out of my a...er...out of the air, because I have no idea how many phone lines there are on the average bill.  I'd guess the median is probably 1, so let's say that the giant pool of singletons drags the average down to 2 lines/bill.  That means we get about $1 billion/year.

What about a 25% increase in the contribution factor?  That's easy.  The USF disbursed $8.7 billion in 2012, so a 25% increase gets us an extra $2.2 billion.  But that 25% figure was a total guess; I wouldn't be surprised to learn the actual increase would be as little as 15% or as much as 40%.  That would mean somewhere between $1.3 billion and $3.5 billion in additional funding.

How much do we need?  I did some back-of-the-napkin calculations [I have a high-tech napkin that runs Excel].  Under my most optimistic scenario, it will require $733 million/year in E-Rate funding to run 100 Mbps circuits to 99% of the schools in the country, or $1.2 billion to reach 1 Gbps.  But that estimate is unrealistically optimistic, because it assumes that all the schools are within 10 miles of a carrier's existing Metro Ethernet node, and uses a price that's on the low end.  Realistically, I don't see how $2 billion/year to provides E-Rate funding for 1 Gbps to 99% of the schools in the country.  And that's just to have a connection to nothing.  If you have one location, then you need to pay an Internet port charge.  If you have multiple locations, you'd have to get a bigger circuit at your hub site to handle incoming WAN connections, and then get an Internet connection from the hub.

What about putting wireless into buildings?  I have a good data point there: a client with about 1,000 students that will be putting in a forklift upgrade to their wireless infrastructure.  Their cost?  $93,000, or about $93/student.  I expect that average cost/student to be higher nationwide, because: 1) this district is in an area where data network gear is below the national average; 2) the tech director at this school knows his stuff, and is quite parsimonious, so he found a solution that was much cheaper than what the normal tech director could find; and 3) they had a top-notch E-Rate consultant, who minimized the inevitable increase in cost due to the 470 process.  I think $100/student is a very optimistic estimate on the installation cost of a wireless infrastructure. There are about 55 million students in the U.S., so we'll need more than $5.5 billion to install wireless for every student in the country.  That means something like $3.5 billion in E-Rate funding.  If we can spread that out over 3 years, we'd need $1.1 billion/year.  I'm not going to estimate the cost of maintenance.

So my optimistic estimate is that we'll need about $3 billion per year to cover this initiative.  That means increasing the size of the fund by 34%, which means increasing the contribution factor by 34%, which means that a phone bill with a current USF charge of $1.16 would go up $0.40.  And that doesn't address the existing funding shortage.  I hear that we'll pay for some of it with savings from reform, but I don't think those reforms are going to bring in more than a couple hundred million, so we can use those savings to cover rounding errors.

One final observation: the administration is saying that the increase would be temporary.  Let me think, when was the last time a charge was temporarily added to phone bills?  I think that would be the Federal Telecommunications Excise Tax, which comes from a bill passed in 1932* and re-authorized 29 times.  The increased contribution factor, on the other hand, would never have to be re-authorized.  The Russians have a saying for this:  There is nothing as permanent as "temporary."


*For telecom history geeks: I sometimes see claims that the tax came from a bill to pay for the Spanish-American War, but that tax was repealed, as was the telephone excise tax imposed to pay for WWI.  The 1932 tax, passed to cover the revenue lost due to the Depression, stuck.  I guess it was re-authorized to pay for WWII, and by 1945, it had become a habit.