Government funding of broadband networks is once again garnering discussion. FCC Chairman Tom Wheeler’s Feb. 19 announcement on Net Neutrality indicates that the Commission may consider Federal pre-emption of state laws restricting municipal communications networks. On the tail of that announcement comes a report from the Government Accountability Office providing some anecdotal evidence of the way small businesses use ten different federally subsidized or municipally run networks. Unfortunately, it is difficult to draw any general conclusions from the report, as it draws on only a handful of interviews with users of a select few networks – a limitation the report itself acknowledges.
I have no beef with the report itself – this GAO report, like all GAO reports I’ve read, is precise, well-written, and self-aware, though I must say I question the value of a study with such a small sample size and so little information about the networks discussed. My concern is that many are misinterpreting the report as offering broad evidence of the success of government run or funded networks. As the report says, “the results of [the GAO’s] interviews cannot be projected to all service providers and small businesses because they were selected using a nonprobabilty approach, they illustrate a range of possible views and experiences.” In the end, all we can draw from the report is that subsidized networks will generally be slightly cheaper and somewhat faster than similarly situated unsubsidized networks, which hopefully shouldn’t come as much of a surprise. We should be even less surprised considering these funding mechanisms are designed to go to areas where service is lacking – I would certainly hope that pouring millions of dollars into a small, rural community would help leap-frog to advanced technology! It is important to consider what hidden issues can remain unexamined with these types of informational interviews.
First of all, the report offers very little information on the networks the GAO studied – both the publicly funded networks and the private networks it compared them to, preventing the reader from evaluating the comparison. To what extent are the services provided by muni and publicly funded networks the same? Are they serving similar populations (in terms of income and geography) and thus facing similar externally imposed costs? Without this type of comparison it’s hard to know if the advantage GAO says these networks have are from the fact that they are municipal or from some other factor. The GAO offers that they “collected unbundled, month-to-month pricing when available,” but, considering, for example, how much of networks’ costs come from programming fees, it is very difficult to draw meaningful conclusions without a rigorous “apples-to-apples” comparison.
Second, the report dangerously conflates federally funded networks with municipal networks that are operated by a local government. These are actually very different types of projects. Although both are funded, at least up-front, with an injection of public money, certainly some municipal networks (e.g., Burlington, VT, UTOPIA in Utah) have generally not been very successful, in part because they sought to overbuild networks where there were already existing network providers. Federal funding, on the other hand, can make sense if done right and targets those communities and areas that have no wired broadband. To lump the two together and discuss them as if interchangeable is a mistake.
Even some of the municipal networks discussed in the report are facing trouble. Fibernet in Monticello, MN, for example, is struggling with significant financial difficulties, and the city is now facing a suit from bondholders. Similarly, the MINET network in Oregon has had financial trouble that has led to disagreements between the served communities. Municipalities sometimes mistakenly believe that a fiber network is simple infrastructure, like a utility. These cities are then surprised by unexpected costs, both upfront and ongoing, and low up-take rates due to citizens either uninterested or already satisfied with existing services. The politics of risking public debt aside, municipalities are generally not well suited to operating these complex services, especially where providers already exist.
Most municipal networks that fail are over-building in an already served area. This same problem plagues at least some of the investments made with federal funding, as the history of Department of Agriculture RUS loans in the last decade demonstrated. To come into an area and over-build with a completely duplicative network makes little economic sense. If a population is adequately served, which, to be fair, is not always the case, a new network is not only a waste of the capital required to build the initial infrastructure, but sucks capital out of existing networks that could fuel upgrades. The priority for scarce federal dollars should be to serve unserved areas. When the report simply interviews small businesses that use subsidized networks, these substantial costs are ignored.
There is often a strong tension between extending broadband access to underserved areas and building in population-dense areas where it is already economically viable to operate a network. Even where federal funding is used, operators will want to continue to see a profit after the initial seed money is spent. This often results in funding recipients pushing up to the edge of the rules tied to the federal money be spent on “unserved or underserved” populations. This is often where existing providers are already able to provide sufficient broadband without risking public capital.
There are certainly areas where federal funding makes sense. The report describes a few conditions that greatly increase the cost of deployment, in other words where subsidization would make the most sense. These include: low population, difficult terrain, and frequent natural disasters. We should take the GAO report at its word and keep our subsidize limited to these sorts of areas where Americans are genuinely underserved.