In his recent OPEC 2.0 op-ed, Columbia University law professor Tim Wu offered his vision of a broadband policy by declaring the broadband market a “bandwidth cartel” that has gouged the public like the energy market. To remedy the situation, Wu advocated much more facilities-based competition, particularly through municipally-provided fiber-optic Internet service and called on the government to open up wireless radio spectrum to “liberate us from wires, cables, and rising prices”. While this bash-the-corporation rhetoric may have some populist appeal, Wu’s analysis is both factually and logically flawed.
Wu began by claiming that Americans spend almost as much on bandwidth as they do on energy. But even if we ignore the fact that Mr. Wu conveniently lumps in cable television entertainment costs as a “bandwidth” expenditure, his comparison still doesn’t pass the smell test.
In fact, broadband service performance has gone up significantly while prices have declined. Wired long distance fees have almost become nonexistent because unlimited nationwide phone service costs as little as $40 a month. Wireless prices have remained relatively constant but minutes have multiplied. And broadband prices have fallen while performance has increased. Yet in recent years, gasolineconsumption for the average family has gone down while prices have more than doubled. Indeed, the Bureau of Labor Statistics shows that from 1999 to 2007 electricity prices rose 60% and gasoline prices 138%, but telephone service prices fell 2% with Internet service (including broadband) prices down 21%.
Wu extends his tortured comparison with the oil industry by arguing that phone and cable companies are artificially limiting the supply of broadband to jack up prices and extract maximum profit like OPEC. But the data shows just the opposite as mobile phone minutes increased 12-fold from 1999 to 2007 according to a CTIA wireless industry survey and broadband speeds have increased many times during that period of time. So while the supply of oil has been limited, the supply of wireless and wired broadband services have expanded.
Not be deterred by the facts, Wu then argues for municipal or community owned broadband fiber networks as a way to neutralize this imaginary bandwidth cartel. It’s one thing for communities to provide their own broadband when they have no commercial broadband providers, but it’s quite another thing to publicly subsidize overbuilding. In fact, recent history suggests that such overbuilding is not a panacea. Ironically, Wu cites Utah as one of his examples of municipal fiber, but Utah’s UTOPIA and iProvo fiber broadband projects have been anything but a success. UTOPIA is facing serious shortfalls and asking their 11 member communities to increase their sales tax pledge of $202 million over 20 years to $504 million over 33 years. So in order for a small percentage of the population to benefit from fiber, the local population as a whole must subsidize those fiber customers to the tune of half a billion dollars—even as most of the area’s residents can already subscribe to broadband services provided by cable or telecom companies.
Despite this massive tax payer subsidy, UTOPIA is considering a one-time connection fee that averages $2,300 for all of its current and future subscribers. Financial failure is common in public fiber projects and the other Utah fiber project in Provo is no exception. Provo was recently forced to sell its iProvo municipal fiber network to a private company Broadweave Networks under a plan where Broadweave assumes the cities bond payments and takes ownership of the network.
Even the municipal fiber implementation in Burlington Vermont, touted by many muni-broadband proponents as the most successful, charges $72 per month for 8 Mbps symmetrical fiber service. That’s substantially slower and more expensive than, for example, Verizon’s $65 per month “FiOS” 20 Mbps symmetrical fiber service. Compared to Verizon’s FiOS service, Burlington’s municipal fiber service is nearly three times the price for each megabit per second. While the residents of Burlington got their fiber well ahead of Verizon FiOS customers, they’re stuck with a more expensive and slower service. It wouldn’t be surprising if Fairpoint, the incumbent telecom broadband provider in Burlington had second thoughts about deploying faster broadband service to Burlington because of the municipal competition.
Not content with simply wanting more municipally-provided “overbuilding”, Wu then proposes a new approach to wireless spectrum policy where anyone who follows a few basic rules to avoid interference would be granted nonexclusive use of the airwaves. While expanding spectrum use to digital white space would help eliminate unused spectrum, Wu seems to suggest a more far-reaching strategy where all spectrum should be open to anyone who wants to use it, including licensed auctioned spectrum. Yet in the past, giving away TV spectrum for free has been associated with inefficient use whereas commercially licensed spectrum purchased through auction has not. Achieving Wu’s vision would require replacing every radio device with new devices that even Wu himself admits have not been perfected yet. A wholesale change to a wide open spectrum regime would require reneging on all existing spectrum licenses sold, and a huge a leap of faith that technology will eventually allow radios to work in this new and untested environment. With ever improving spectral efficiency leading to the rapid decline of cellular minute prices and US wireless Internet adoption growing rapidly, would such a radical untested scheme be warranted?
The real reason Wu advocates muni-fiber and open spectrum is because he believes that these municipal fiber or wireless overbuild projects will somehow result in lower costs to consumers. But most of these wireless overbuild projects lay out redundant broadband connections to communities that already have broadband. In doing so, even Wu must surely acknowledge that the result will be higher overall telecommunications expenditures. After all, adding a third wire in a community (municipal fiber to supplement existing telephone and cable broadband), or a new wireless system, is certainly more expensive than just having the existing two. Someone has to pay for the third redundant system. The only way these additional costs could possibly lead to lower prices for consumers is if the profits of companies selling services on the first two wires (cable and telephone companies) were monopolistic. If this were the case, the added competition, while increasing overall costs to society, would reduce profits hopefully enough to outweigh the increased costs (but not so much to limit further investment). Yet, there is no evidence that broadband profit levels are excessive.
So what is the state of broadband in the United States? The truth is mixed because there are many positive things happening, but also areas that need serious improvement. On the bright side, high-speed wired and wireless broadband deployment is growing rapidly and prices are dropping. Moreover, as ITIF recently documented in a study comparing broadband policies among nine nations, while the United States lags behind Korea and Japan in fiber broadband build out, we are ahead of Europe. Verizon now serves 10 million homes with its fiber broadband service which is 622 Mbps of Internet bandwidth shared between 32 homes. At present time, Verizon limits burst speeds to 50 Mbps but they can easily set it higher. AT&T is laying Fiber To The Node (FTTN) which allows existing copper phone cables to have 25 Mbps, or soon 50 Mbps, of dedicated capacity for each home shared between Internet and IPTV service. At this point AT&T has only allocated 10 Mbps to Internet service but they can in the future dynamically allocate up to 50 Mbps for Internet such that each HDTV channel the customer watches would lose 8 Mbps out of that 25 or 50 Mbps budget. Cable companies like Comcast, fearing the competition from the telecoms, have announced aggressive plans to deploy their next generation (DOCSIS 3.0) cable broadband technology which will have 80 to 160 Mbps of the last-mile capacity shared between a neighborhood of 100 to 400 people.
Notwithstanding this progress, there are some Americas that still can’t get access to broadband, most of whom live in rural areas. Moreover, many Americans are not fully digitally literate and/or can’t afford computers. Indeed, if the United States had the same rate of computer ownership as many leading broadband countries our OECD broadband rank would be significantly higher. These are areas where government can take an active role in increasing broadband penetration. We can implement demand-side programs that promote computer literacy and computer ownership. We can expand the universal service fund beyond telephony to include broadband service and make broadband more affordable to low income households. We can promote and fund public-private partnerships like Connected Nation to increase broadband penetration in rural areas and support innovative approaches like reverse auctions to get broadband to places lacking it today. Arguing for public subsidies of even more broadband providers in places that already have two or more will simply divert funding and attention from these critical challenges.