In last week's State of the Union address, President Barack Obama highlighted government programs "rebuilding for the 21st century." Among the investments in vital public infrastructure he mentioned -- roads, bridges, rail -- he promised "high-speed wireless coverage to 98 percent of all Americans." The president was referring to mobile broadband Internet and similar services advertised as "3G" or "4G" -- low-cost ways to help make basic broadband near universally available.
Unfortunately, that's not where government financing has been going so far. The U.S. Department of Agriculture, responsible for $7.2 billion in stimulus funds put towards bringing broadband Internet to underserved and rural areas, has already blown the majority of that money on the rollout of "superfast" Internet via fiber-optic cable -- joining a global bandwagon of speed-obsessed governments sinking billions into extending a whole new communications infrastructure across their respective countries.
Superfast Internet connectivity is often defined as delivering content at 50 megabits per second or faster. That makes it about 12 times faster than the Federal Communication Commission's standard for basic broadband. There is more than one way to deliver superfast connectivity -- cable television wires can manage 50 megabits per second speeds, for example. But for really fast Internet, fiber-optic cable is technically unbeatable -- it can deliver speeds twice as fast as cable and up. Why settle for a Ferrari when a Bugatti Veyron can go twice as fast?
The Agriculture Department is hardly the first government agency, in the United States or elsewhere, to fall for fiber. Last year, the Federal Communications Commission set the goal of outfitting 100 million households with 100-megabits-per-second Internet access. The European Commission approved spending nearly $2.5 billion worth of member government subsidies -- along with $3.2 billion of its own funds -- on rolling out fiber across the continent, with the aim of having half of European households online at speeds of over 100 megabits per second by 2020. Meanwhile, Australia has really gone all out -- it is investing $23 billion in its National Broadband Network, which includes connecting almost all Australians' homes with fiber.
The premise for giving so much cash to the global trench-digging industry is governments' belief that superfast broadband willbring substantial social and productivity benefits. Unfortunately the evidence for these benefits is awfully thin. All else equal, faster is surely better, when it comes to the Internet at least. But faster technologies don't always triumph; think of passenger hovercraft, maglev trains, and suspersonic airliners. These technologies didn't fail because they weren't superior, but because the demand wasn't there or was insufficient to justify the cost. And that is the concern with fiber -- the costs of rollout look particularly high, while the benefits in terms of new applications look limited indeed.
Compare the cost of fiber to the two previous iterations of Internet connection technology. The first was dial-up, which cost less than $200 per home. The next was a digital subscriber line (DSL), with a total cost per subscriber of about $150. The fiber upgrade is different from these in that it requires building an entirely new network rather than retrofitting existing ones -- a difference that is reflected in fiber's price tag. Verizon's costs for rolling out a fiber-to-the-home network in the United States are in the region of $2,750 per home connected.
Given widespread availability of basic broadband (averaging 88 percent coverage in the 34 member countries of the OECD), to believe that the investment in fiber is worthwhile, you have to think there will be considerable benefits from superfast applications that basic broadband cannot deliver. Yet, today, there are virtually no services that can only be delivered over fiber-based broadband.
Some commentators have argued that fiber will enable "smart grids" that allow electricity consumption to be smoothed, reducing peak demand and in turn the need for new power plants. But in Italy, 30 million smart meters were installed between 2001 and 2005 delivering just such benefits; they needed kilobit-, not megabit-, per-second speeds to operate, and were connected using existing copper or mobile networks, not fiber ones.
Those touting health-care applications point to remote consultation: doctors in their offices diagnosing and prescribing treatments for patients who remain in the comfort of their own homes. Advocates note the positive results of trials of tele-homecare for the elderly and U.S. military veterans. But those trials, too, did not require fiber-grade speeds -- it was possible to get the medical benefits in question using basic broadband or even dial-up. As for education, it may be that certain university lectures will be dependent on very high resolution video requiring superfast home connections, but the vast majority of educational material can be -- and in many cases already is -- delivered perfectly well without it.
Fiber advocates also argue that the technology will enable more people to telecommute to their jobs, with associated benefits for traffic congestion and the environment. Data from a U.S. survey sponsored by a fiber industry lobby group suggests households early to adopt fiber do telecommute more -- one extra day per month. But while fiber may make this easier, it is neither necessary nor sufficient. Some Scandinavian countries already had 17 percent of the population telecommuting in 2000, well before superfast broadband made its debut. Conversely, in South Korea, which has led global fiber rollout for some time, less than 1 percent of the workforce telecommutes.
What South Korea's lead in fiber may help explain, however, is the fact that around 1 million of the country's citizens (by the government's estimate) are addicted to online gaming. Indeed, where fiber rollout has occurred to date, the evidence is of few new speed-dependent applications, but more intensive Internet use for gaming and watching television. This is the most plausible benefit of fiber over basic broadband going forward -- but it is an odd justification for public subsidy. And even here, the benefits of fiber are marginal to most consumers. It is true that fiber can provide on-demand high-definition television, but so can the sort of cable services to which 96 percent of households in the United States can already subscribe, at an estimated cost of just $100 per home.
If fiber isn't all it's cracked up to be, that might explain why few people with access to it seem willing to pay for it. In six out of nine European countries where fiber to the home is available, the price that providers can charge for fiber broadband is the same or even less than that for digital subscriber line or cable broadband services. Even so, in Europe as of July 2009, less than 16 percent of homes wired for fiber were connected to it. In Britain in particular, a service offering up to 50 megabits per second available to around half of the country's households since the end of 2008 had only 74,000 subscribers at the end of June 2010. Adoption of new services is rarely immediate, of course, but these figures suggest it could be a very long time before consumers are as enamored of fiber as their governments appear to be.
The future may hold a bevy of powerful applications that require superfast broadband. These applications may indeed carry significant social benefits. If so, the case for subsidising fiber rollout will become more compelling. But the "killer application" for superfast may not appear at all. And when there is no apparent need to hasten into investments in a technology with uncertain demand, the best answer is to wait -- after all, a seat on the fiber bandwagon is awfully expensive.