Even a casual look at the supposed costs of universal 100 Mbps broadband cited in the FCC’s National Broadband Plan brings laughter and outrage in the industry – but not to FCC chairman Julius Genachowski, who is smart enough to know better, nor to Blair Levin, the former investment analyst who oversaw the plan’s creation and has since decamped.
The industry (which actually builds networks) says 95 percent of the whole country can get 100 Mbps for well under $200 billion. We agree. The National Broadband Plan says the total cost would be $670 billion, of which $320 billion would have to come from government.
The plan’s cost estimates for other bandwidth specs is equally out-of-kilter.
There are 115 million households in the United States (not 130 million as stated in the model documentation, done by those highly paid PhDs at FCC contractor CostQuest). There are 2.6 million miles of paved roads. If every mile required new fiber to deploy 100 Mbps (and that’s a gross over-estimate) that $670 billion (in the model’s backup documentation, but not cited in the National Broadband Plan) translates to an outrageous deployment cost of $257,000 per mile. The NBP itself, in text and graphics, cites a merely laughable $144,000.
This magazine has detailed more than 450 FTTH builds. Verizon costs out at about $40,000 per mile or $700 a household including drops and tax savings. That’s estimated from Verizon’s SEC filings. The others, mainly rural, seem to be in the $60,000-$90,000 range and about $2500 a household. In other words, real-life experience comes in well below the model’s estimates.
Reality check: There are 130 million addresses in the U.S., if one includes businesses. Of course, many are unoccupied these days and close to 30 million of them are already passed by telco fiber alone – a fact the model ignores. DOCSIS3-capable fiber already deployed by cable companies and passing all but 6 percent of households could serve 50 Mbps with DOCSIS node splitting and similar tricks, and without necessarily laying new fiber, says DOCSIS- and DSL-guru David Burstein.
I’ve talked about this before, in my past few Broadband Hawk columns in the magazine itself, and in some industry talks. But as time goes on, my laughter has turned to outrage:
- Outrage that Congress is buying into this nonsense as it considers revamping the Universal Service Fund.
- Outrage that FCC officialdom keeps repeating the numbers although industry interests have submitted dozens of formal briefs pulling apart the FCC’s model.
- Outrage that the economic model the FCC used – CostProLoop from CostQuest Associates – has been used by only four U.S. companies in 30 states, all for designing network layouts for networks already decided upon – and none of those networks were for 100 Mbps or even for FTTH! One of those companies, BellSouth, no longer exists as an independent entity.
- Outrage that while the FCC crows about its openness, the model’s exact equations are proprietary – although it is fairly easy to reverse-engineer the basics for fiber-to-the-home deployments.
- Outrage that the key base assumptions, as stated in the FCC’s own documentation, are grotesque fiction.
For example:
The FCC says 4 Mbps download, 1 Mbps upload is a good future target for rural America. But Akamai, which does more store-and-forward services than anyone and obsessively measures network speeds, says nonrural America’s average real bandwidth is already almost exactly 4/1 Mbps!
Key for the USF debate: The NBP text and the model’s backup documentation says the hardest-to-serve 7 million households would cost $23.5 million more to serve than is available from investors, at just 4/1 Mbps and that the toughest 1.1 million households would cost more than $13 billion of that. The accompanying exhibit illustration of the model run in the NBP agrees on the latter but not on the total – the remaining 6 million homes would cost only about $5 billion to serve, not $10 billion.
The model (using a regression on Pew Foundation data) assumes rural take rates should be calculated at the same, lower figures as urban take rates (where there is more consumer choice). But Michael Render of RVA LLC, based on his actual polling of FTTH deployers, cross-checked by their equipment purchases, pegs urban rates at more than 30 percent and rural at more than 50 percent. Real data. Why believe a speculative calculation based on the merest gossamer web of facts?
The model refuses to assume that bandwidth demand will continue to rise. That’s absurd. To be precise, the model documentation states, “While end-users are likely to demand more speed over time, the evolution of that demand is uncertain. Given current trends, building a futureproof network immediately is likely more expensive than paying for future upgrades.” That’s nuts, too. And it proves the lie that the NBP is technology-neutral. That statement is, by the way, backed up by zero calculation in the model itself.
That the model engages in fiction about costs is bad enough. But there’s absolutely no estimate about new revenue streams that might be enabled. In fact, by restricting calculations to triple-play, it assumes revenue will decline, not rise. Tell THAT to telcos deploying fiber now!
The model is restricted to an 11.25 percent internal rate of return, which of course also inflates costs. Yes, that’s barely above the cited coupon rate now for telco investments in totally new networks – if investors will bite at all – but that IRR provides a nice cushion indeed in the real world, where networks are being expanded by bootstrapping from anchor tenants.
Says the documentation: “The model and its associated documentation provide an unprecedented level of transparency.” I think the above-stated facts reduce that statement to the level of, well, bullshit. The documentation also says the model “should spur debate.”
Indeed it should. But at the FCC itself, the debate seems over and the minds seem closed.
FCC Commissioner Genachowski and I both spoke at OPASTCO’s Seattle meeting in late July. He stuck to the script. I didn’t . The video of both talks is at
Also on that page is a video of Blair Levin justifying the plan he oversaw – and video of me grilling him for 12 minutes, at the Occam User Group conference in Washington in June. He makes some important points. But he’s lost on costs.
Watch the videos – and decide who has a better grasp of the numbers.
