Why are there so few choices for broadband in many markets? Is Internet service a natural monopoly? Will it always be?
(This article covers wired connection options: DSL, cable, and fiber, mostly. In an upcoming article I’ll discuss wireless connection options, and positive government regulation, too. Really: Some government regulation encourages competition.)
Wiring a city is a natural monopoly. (I’m going to ignore the libertarian arguments against this term; feel free to discuss.) The costs to string power, telephone wire, cable, or fiber to a neighborhood are high. For a new subdivision or apartment building, these costs are ultimately borne by the homeowner/renter. Everyone pretends “the developer” pays them, but they’re all passed along. Adding new services, like fiber Internet, to existing housing is even more expensive, and adding buried services can be two to three times more expensive.
Unless a buildout is likely to be profitable, no company will spend the money to install it. This is a classic natural monopoly, as compared to a government-granted monopoly.
And not all wire is equal: Cable Internet is much faster than DSL, because coaxial cable carries a much broader band of signal than phone lines. In a nutshell, that’s why many addresses only have a single affordable high-speed option: Cable Internet is fast enough to carry Netflix, DSL isn’t.
(I’m using the 2015 FCC definition of broadband: 25-plus Mbps down, 3-plus Mbps up.)
Scale and density count, too. South Korea plans to connect every household with 1Gbps or faster broadband, which is a lot easier to do for 50 million people living in an area making up 1% of the landmass of the lower 48 U.S. states. Also, half the population lives around Seoul. A lot of “American broadband speed is awful” reporting ignores this basic reality.
Considering those figures, is there another way to increase competition for wired services?
One way is faster DSL, which is on the horizon. Moore’s Law is widely cited for faster computing, but it also holds for communications over existing channels. Those new, smarter, faster chips power cable modems, cellphones, DSL, Wi-Fi, and a dozen other amazing new communications methods, not just the latest overclocked gaming PC.
Case in point for DSL: The industry is trialing G.Fast and related technologies, using new chips supporting 500 Mbps to 1 Gbps of download speed over existing phone lines. But that blazing-fast DSL is for short distances: 50 to 250 meters, as compared to the 3,000-12,000 feet (1,000 to 4,000 meters) between you and the Central Office (where the DSL equipment usually lives.)
Because they’re short-haul, G.Fast-type speeds require outside plant DSLAMs, pedestal- or pole-mounted gear on every street or two, thence connected via fiber to the Central Office. (There’s also bonding and vectoring, where two or more phone lines combine for faster service, but that’s still only 12 or so Mbps of download speed and not very common.)
These faster DSL options require serious capital to extend the life of the local loop, the copper wires between your house and the rest of the network, with G.Fast gear, and then to get new DSL modems out to subscribers.
Will AT&T or Verizon build out G.Fast? The crystal ball is hazy. Verizon is acting as if they wish to abandon (or at least stop spending money on) their telephone line infrastructure. In multiple cases, they’ve declared phone lines to be unrepairable, and sent fixed-cellular stations as replacement. Both AT&T and Verizon are instead spending capital on their U-Verse and FiOS offerings. In many cities, they’re not replacing existing phone wiring except where required by regulation.
Expect these shenanigans to become more common as the maintenance cost of 50- to 75-year-old telephone cable starts to climb, and telcos expect more profit in fiber.
Can other companies use G.Fast? Probably not.
G.Fast is only practical for the wire-owners, the local telephone company. Competitive DSL companies (Covad, Megapath, etc.) rent phone lines from the local telco, all the way back to their own rack in the Central Office (CO); rolling out G.Fast would require equipment in hundreds of neighborhoods, on poles or pads. That’s an impractical level of investment for these companies, versus likely uptake. Don’t expect faster DSL from them with G.Fast — but stay tuned; technology is still improving.
On the other hand, if G.Fast is successful for the incumbent carriers, expect cable companies to drop Internet prices quickly.
Meanwhile, Google Fiber installation has slowed, as the actual costs of running fiber to every address in a municipality is high. Even the small Google Fiber rollouts have seen cable Internet rates fall in the affected markets, often before service is even available in an area. Competition works, opening up monopolies and lowering prices.
In the next article, I’ll discuss how governments, especially local governments, really can lead or get out of the way, encouraging actual competition for wired high-speed Internet services — even when their own anticompetitive (and even socialist) urges kick in. Also, a wildcard: an X-Prize for DSL.