Eight false assumptions on which national broadband plans are based
Broadband is good, the nation is a positive notion and most good things need a plan. If we put these three things together the result is a national broadband plan. Is this still desirable or is it like a Hawaiian pizza – several good things in a combination that does not quite work?
The most ambitious effort at building a nationalized network has been in Australia which by 2020 had finished its National Broadband Network (NBN) considerably over budget, substantially late and with much reduced specifications. Its history has been well documented and at the over 8 million premises uptake level of early 2021 for a budget of $51 billion (Australian), the cost was over $6,000 per connection. Most of the premises served are in urban areas where the same (or better) connectivity would, was (or is) available without any public investment. Nevertheless, parts of suburban Sydney do not even have 25 megabit per second (mpbs) service and justification of the huge cost of the project requires attempting to quantify purely social benefits.
The Australian plan was to build a network under federal government control that covered the entire country, with a mix of technologies. Other countries have less ambitious plans like the US federal and state programs that subsidize connections in outlying areas. At least, these plans leave large areas open for private investment and market competition in the provision of access services.
It is worth looking at the assumptions that can be used to justify government investment, either directly or through subsidies in the broadband market.
Virtually everybody is in fact able to get internet service, if necessary from a satellite provider. Starlink has pre-launched a world-wide uncapped service at $100 (US) per month and in the rural US several companies provide service at $50 per month. In many areas, fixed-wireless or cellular service is a feasible alternative and this can be improved through the further use of unlicensed spectrum as there is usually little spectrum congestion in rural areas. Unless excessively restrained by regulation and/or the availability of spectrum, this will be provided by the private sector.
Generally, the cost of living in outlying areas is much lower than in cities although many services and products are not available for instore retail. There is no reasonable expectation to have the same products available (fresh artisinal croissants, for example) in Paris as in the last village of North Dakota and this is also true for broadband services.
An upcoming paper by Roslyn Layton of Strand Consult and myself, examines how the bulk of the traffic and middle-mile costs of rural broadband providers in areas covered by US federal subsidies, can be attributed to streaming video, gaming and software updates for subscription services provided by Netflix, Google, Amazon, Disney and Microsoft. Fast and uncapped rural connections contribute to the already huge revenues of these firms, at the expense of the broadband providers, users with modest requirements and the taxpayers who subsidize the connections. Services such as tele-medicine and e-government require connections with high capacity only for very short periods of time.
Broadband is a general purpose technology but that does not make it an infrastructure where there are externalities that can be easily captured by collective or public investment. It is not what economists call a public good – the best example of which is perhaps the lighthouse. There is of course a correlation between broadband deployment and economic development, just as there is a correlation between sales of electric power or fancy coffees and economic development. Yet, everyone understands that coffee consumption is a consequence of economic growth and electricity consumption is necessary (and somewhat commensurate) but, like the coffee, not causal.
Is there any other magic through which dropping broadband from a plane unto rural villages will create economic growth? Well, it might (especially in the 2020s) entice white-collar workers from the city to relocate, driving up rural property prices and simply displacing economic activity from the city to a rural area. Natchez (MI) has started to pay relocation bonuses to individuals resettling in that US city, which is surely a more efficient way to achieve the same. This leads directly into the next assumption.
This assumption locates the broadband plan at national or state/province level and conflates two entire different issues: the putative “right” to enjoy the same services as urbanites and the economic growth argument. If the economic growth argument was water-tight, the local community would surely invest (and be able to borrow to invest) in this wonderful technology and no decisions about this would have to be taken at a higher level. That is, if a public investment in broadband really did make sense, it would happen at the lowest possible level of administration. Recent research in the EU, on the other hand, has shown that local authorities are often hostile and the takeup can be very low – 2.8% of homes served in and Italian public broadband expansion project.
The US has a lot of rural broadband providers that take advantage of federal subsidies to provide service in their communities. Elsewhere, the picture is less small is beautiful and more too big to fail. In the UK, former incumbent BT was awarded most of the contracts for fibre expansion under the BDUK program and a similar situation prevailed in Italy. The EU routinely approves state aid for broadband projects which at least acknowledges that this is in fact industrial policy but denying that it amounts to the old-fashioned picking of winners. In fact, that is precisely what it is – governments are taking general tax revenue and pouring it into specific industries.
Although national and regional broadband projects do not necessarily advocate wireline solutions, they often do so as the fiber or cable line is now seen as the gold standard of connectivity. It certainly is the gold standard for consumers who get through between 20 and 100 gigabytes per day of content (as I do) but six months ago, I lived and functioned perfectly on an LTE connection in a house where there was no wireline service at all. Wireless technologies are perfectly adequate for providing baseline connectivity to households in areas where the private sector finds it uneconomic to lay fiber or cable.
Broadband commissions and authorities, like universities, have many well-meaning and clever people working in them. Like King Canute (the Great), it does not follow that they can control the monumental forces of this world – one of which is technological progress, and another is the market. Australia, with its completed NBN (consisting of as much fiber as possible) now has one of the highest wireless-only household percentages in the OECD. Only the market will determine what the appropriate technologies are for meeting very dynamic demand. This demand is driven, nowadays, by a wealthy and diverse content industry that might find its needs better accommodated by 5G wireless that incorporates (yes, it is back!) one-to-many broadcast transmissions.
The use of fixed-wireless broadband access in peri-urban areas in countries including the US and South Africa is a case in point. This form of internet access can accommodate households up to 10 or 20 km from the base station and the economies depend very much on demand. Which areas can efficiently be server by this technology depends, like all internet access technologies, on how the relatively high fixed and low marginal cost can be balanced by different technologies to provide last-mile access. Medium-term planning by officials is simply unlikely to discover the best way of doing this.
Investment tends to sound so good when politicians are selling it. The counterfactual is assumed to be the usual frittering away of taxpayers’ dollars. From a political point of view, it is quite convenient as delivery is always a few years’ off which provides a lot of time and wriggle room. There are of course better and worse ways of directing state investment in broadband and general telecommunications infrastructure. Australia has not only the colossal and unwieldy NBN but also a transparent and quite efficient mobile blackspots program (MBSP) which allocates federal funding from a fixed budget to competing projects, based on user-reported mobile network reception problems.
Other projects remind one of former Alaska governer Sarah Palin’s “bridge to nowhere” at Gravina Island, projected cost $400 million. The bridge was cancelled but the upgrade of the connecting “road to nowhere” was completed for $25 million. Broadband projects have the interesting feature that their completion and level of use is not as easily observed by the general tax-paying public. Consider three more projects in brief.
€2.6 billion has been budgeted to achieve “100%” high-speed broadband coverage in Ireland. This is a wholesale-access network to be constructed in areas which have been identified based on somewhat wobbly criteria – presumable in 100% of them. The country has only around 5 million people so the cost is around a very hefty €500 per Irishman to cover around 0.5 million premises at a cost of roughly €5000 per location. One has to ask how many locations would have perhaps preferred to receive the money rather than the broadband access.
The best question about the FCC’s 2010 plan is how many people are able to name its achievements. The steps that the FCC took under chairman Pai to simplify local authority rules, fees and obligations when considering applications for mobile towers probably did more to expand mobile broadband than the 2010 plan, at essentially zero cost.
The first national broadband plan in The Philippines was launched in 2007 by a previous administration and subsequently cancelled because of allegations of corruption. Its 2017 plan is an amalgam of different goals, including connecting government offices, subsidies for the development of local content and stimulation of demand. This includes subsidies to operators for in underserved areas. Four years into the new program, its budget keeps growing and some progress has been made to establish a government-owned fiber optic backbone network. Otherwise, progress has been slow despite the plan’s laudable goals of simplifying regulation for private operators.
While even Zimbabwe is now on the national broadband plan bandwagon, the following questions remain.
- Are any government broadband plans cost-effective at all?
- What are the market failures being addressed?
- If the real goal is a transfer payment from urban to rural areas, is it necessary to do this through the mechanism of broadband?
- What lessons have other countries learned from the Australian experience?
- Are broadband plans not simply a disguised subsidy to the telecommunications sector?
- What are the zero-cost options for enabling better rural broadband, e.g. simplifying planning permissions and making spectrum available?
Policy makers who sell national broadband plans to the public are unlikely to understand either the fundamentals of the industry or the dynamics of the digital marketplace (which by now includes huge profits in the digital content industry). They are peddling an expensive policy response to conditions which will (where they even exist) be resolved more quickly and less expensively by supply- and demand-side signals interacting naturally.