Why municipal fiber hasn’t succeeded
The following paper was authored by Dr. Robert D. Atkinson (ITIF) and George Ou (while at ITIF), and originally published at Chaffee Fiber Optics (page 17) on March 2, 2009. Article was updated by George Ou on 3/11/2010.
With the United States falling in broadband rankings and trailing many other countries in broadband deployment for a variety of reasons[i] [UPDATE 3/11/2010 - the rankings are extremely dubious], some communities have lost faith in the private sector and have begun to look to community-based alternatives. The promise of affordable ubiquitous fiber broadband service and the potential economic development opportunities that come with it are so enticing that a number of cities are considering the municipal fiber route. But despite the promises of municipal fiber, the actual success rate of these community fiber projects has been lukewarm at best and in many cases a failure at worst. This article explains why fiber is not necessarily the only technology to focus on, how faster speeds are evolving, and why municipal fiber over-build projects are economically inefficient, and why municipal fiber hasn’t succeeded in many cases. Finally it presents a policy framework for thinking about this.
Broadband technology has evolved rapidly within the last 10 years. First generation broadband started as a copper based technology either through copper phone wires from the phone company or copper coax from the cable TV company. Next generation broadband uses three technologies: Fiber to the Home (FTTH), Fiber to the Node (FTTN), and DOCSIS 3.0. All three technologies use a combination of fiber and copper technology to bring broadband into the home. FTTH brings fiber all the way to the home but it uses copper coax cables for the final 30 meters within the home. FTTN is a Digital Subscriber Line (DSL) technology that brings fiber to within 1,000 meters of the home and uses existing copper phone wiring to get to the home. DOCSIS 3.0 is a cable broadband technology that brings fiber to a node which splits into 50 to 1000 homes and uses existing copper coax cables originally deployed for cable TV service.
Figure 1 – How broadband technologies compare under load
Figure 1 compares the potential bandwidth performance of broadband technologies on the access portion of the network. It assumes 10 homes sharing an FTTH node and 150 homes sharing a DOCSIS 3.0 node. Also note that the total capacity numbers are based on “signaling rates” and actual performance will be 10 to 20 percent lower due to overhead. “Average user activity” indicates the average per subscriber bandwidth consumption and not the percentage of activity on the entire network. So if 20% of the cable broadband subscribers are using the network and each person has an average activity level of 10%, the average user activity per subscriber is 2%.
The capacity and performance advantages of running fiber to the home are undeniable, but the high deployment costs of FTTH are often difficult to justify so long as the other two technologies which leverage existing plant continue to offer “good enough” performance to meet consumer demand. Some ardent FTTH proponents have questioned why we are investing more money in copper technologies like FTTN and DOCSIS 3.0 that may or may not be obsolete in 5 or 10 years when FTTH is far more future proof. The reality is that staying with existing copper phone and coax wiring is like squeezing another few years out of an old car by spending a little money on tune-ups. Since technology always drops dramatically in price, boosting speed while staying with existing copper allows network operators to hold off on FTTH technology until it is cheaper in the future.
Furthermore, there is no reason to believe that DSL and cable technology based on existing copper wiring have reached a ceiling in performance. Long term research in Dynamic Spectrum Management (DSM) is promising up to 1,000 megabits of dedicated symmetric capacity using existing copper phone wiring[ii]. Whether or not DSL technology can improve quickly enough to keep up with market demand or if the telecoms will be forced by market pressure to extend fiber all the way to the home remains to be seen. DOCSIS 3.0 technology will continue to improve incrementally based on market demand and it could have a practical ceiling of 4,000 Mbps shared between as few as 20 homes[iii]. Even under extremely heavy average user activity levels of 25 percent per subscriber, each DOCSIS 3.0 subscriber could still receive 800 Mbps. These upgrades to FTTN and DOCSIS 3.0 don’t come free, but they are smaller incremental which are easier to finance.
Higher Speed Broadband is Coming
The broadband market has grown at a rapid pace over the last 10 years. At the turn of the decade, only 4.4% of all households subscribed to broadband[iv]. By the end of 2008, household penetration had reached 59%[v] and next generation broadband offerings from 20 Mbps to 50 Mbps had grown at a rapid pace. Verizon passed 11.9 million homes with FTTH service and AT&T passed 17 million homes with FTTN service. By the end of 2009, Comcast will have passed over 30 million homes with DOCSIS 3.0 service.
We are now in midst of an intense market competition between cable and telephone companies for the lucrative “triple play” market which encompasses TV, phone, and Internet service. This competition started when the cable companies started having success with their digital voice phone services which is taking away the market share of traditional phone services. Cable companies with first generation DOCSIS broadband are already effective triple play providers while phone companies with basic DSL service are only capable of providing phone and slower broadband service. The three major Incumbent Local Exchange Carriers (ILEC) telephone companies are deploying FTTH or FTTN technology to be able to compete in the TV market and offer more competitive broadband services. This in turn is pressuring cable companies to deploy DOCSIS 3.0 technology to remain competitive. Because of this intense competition, it is only a matter of time before the 94 million broadband homes covered by ILECs today will have access to either FTTH or FTTN broadband. The cable broadband footprint in the United States as of September 2008 was 119.8 million homes passed out of 124.6 million homes passed by cable video service[vi], and most of those broadband enabled homes will eventually be converted to DOCSIS 3.0 to compete with the ILECs. This does not mean that next generation broadband, or even basic broadband, is everywhere. In fact, we estimate that about 19,000 communities lack broadband altogether. But in most communities broadband is available.
Competition as the Answer?
One of the leading rationales used by supporters of municipal broadband networks (either wireless or wired) is that a publicly subsidized (whether publicly or privately owned) additional network will boost competition, driving down prices and making it easier for residents to afford broadband.[vii]
Yet, surprisingly very few advocates of municipal fiber networks make the argument that cities need to invest to an additional electric wire network to the home, or a second gas pipe network to the home. Indeed, most homes have just one electricity wire, one water pipe, one gas pipe, and one sewage line because building a duplicative “pipe” for any of these services would cost an enormous amount of money, significantly outweighing any consumer benefits from more competition.
The economics of broadband are no different. Building a duplicative network costs a large amount of money and often provides no better service, only more choice in the service. Broadband markets are different than existing single-pipe network infrastructures like electricity in that in most communities there are two existing providers, cable and telephone. But the implications are still the same: subsidizing a municipal fiber over-builder will lead to a waste of societal resources.
The impacts of this will be felt both within the community and outside. To understand why, consider that by definition an additional new network will mean fewer subscribers for existing providers[viii]. Even if some of the lost revenue from the fewer subscribers goes directly to lower profits, it is unlikely that all of the loss will, with the result that the provider will have to raise prices (or at least not reduce them as much as they would otherwise). This means that municipal fiber overbuilding projects will result in higher prices for broadband consumers outside the community deploying fiber who subscribe to a service from a competitor in the community. Because the existing companies now are getting less money, but their costs have not come down a commensurate amount, they have to make up that loss somewhere, and it will come from higher rates. Fiber overbuilding also means that there will be marginally less investment in next generation networks by incumbents outside the community because there will be less overall revenue to support that investment[ix]. In this sense, there is a negative externality to society as a whole from communities investing in municipal fiber overbuild projects.
But many community leaders might respond that their responsibility is to their residents, not to residents outside their community. Toward this end they will often argue that they need to invest in municipal networks in order to spur competition and lower prices. But this notion overlooks the fact that pricing plans for TV and broadband services are regional in nature and not street- or community-based. In other words, because pricing plans are regional, customers are not given higher monthly bills or inferior support because they don’t have any other broadband provider. For competition to work there does not need to be a competitor on every street, in every neighborhood, or even in every city. As long as companies do not engage in price discrimination (which to date they have not), as long as there are two competitors in at least a moderate portion of the region the companies are serving, then the pressure to compete against each other there will discipline prices in all areas a provider serves.
The effect on pricing for the municipal fiber network itself depends on a number of factors, including take up rates and levels of subsidy. Even under ideal circumstances, three facilities based competitors in the market makes survival very challenging because the overall broadband market as of 2008 is approximately 59% of all households. Assuming that all the broadband players are equally competitive, that’s roughly 20% market uptake for cable, telecom, and municipal fiber provider each. Twenty percent market share simply isn’t enough to sustain a healthy and cost effective business model because network operators typically have to build out to roughly 90% of any region to attain those levels of uptake. This is extremely hard for network operators especially fiber operators because they have to bear the cost of building out new infrastructure to 5 homes just to reach 1 subscriber. Municipal fiber projects throughout the country have not been very successful for this very reason. If municipal fiber projects can’t get enough subscribers to cover their costs, prices could still stay low but only if the project is subsidized from other sources, such as general fund revenues.
The Municipal Fiber Experience
While there have been some limited success stories of communities that have tried municipal fiber, there have been many failures, for precisely the reasons related to economics described above. In Utah, Provo County’s “iProvo” municipal fiber service was such a financial failure that they were forced to give the network to private network operator Broadweave Networks where Broadweave would resume the city’s bond payments. Eleven other counties in Utah which formed a municipal fiber coalition called “UTOPIA” which was on the verge of financial collapse in 2008 and they were forced to extend their $202 million dollar 20-year sales tax pledge to $504 million over 33 years[x]. If that wasn’t bad enough, UTOPIA began asking their customers to pay co-op fees of $1,100 to $3,500. But even with the additional sales tax pledges from the UTOPIA communities, there is no assurance that the project will ultimately survive.
The lessons learned in Utah is that projected uptake models and deployment plans don’t always come to fruition, and when that happens the consequence is failure. For UTOPIA, the project was projected to reach 35% uptake rates by February 2008 but the reality was less than 17% uptake. UTOPIA had also hoped for 17% uptake from lucrative business customers but the reality was only 2 to 3 percent. Provo County’s iProvo was hoping for 10,000 subscribers by July 2006 with the assumption that 75% of those customers would subscribe to lucrative triple play services, but the reality was 10,000 customers in late 2007 with only 17% of those customers subscribing to triple play[xi]. Many consumers were quite happy to subscribe to existing broadband cable or telecom providers. The consistent theme in Utah was an overestimation of the uptake rates and the underestimation of competition from incumbent cable operator Comcast and telecom operator Qwest which led to consistent underperformance.
Exacerbating the situation for many municipal fiber projects like UTOPIA and iProvo is the wholesale business model which has proven to be very inefficient. These networks were originally designed to provide wholesale services to companies that would resell retail services (e.g., data and video) Broadweave Networks has already vowed to switch iProvo to a more efficient retail model where it operates the network and sells broadband service directly to consumers rather than rely on an intermediate service provider. Burlington Vermont’s former general manager Dr. Timothy Nulty concurs with this conclusion when he stated that wholesale fiber was “a recipe for financial failure”[xii].
In Burlington Vermont, Burlington Telecom, which is considered the most successful municipal fiber operation in the nation [UPDATE 3/11/2010 - Burlington Telecom is now officially a financial disaster], offers a lower performing product at a higher price than the nation’s largest FTTH provider Verizon. When comparing standalone pricing, Burlington Telecom’s 8 Mbps symmetrical fiber service costs $72[xiii] per month compared to Verizon’s 20 Mbps symmetrical service at $70[xiv] per month. This price difference can be attributed to scaling efficiencies. Large network operators like Verizon operate nationwide networks allowing them to peer with other large Internet networks at no additional cost. Centralized network operations centers and all the experience of installing fiber in previous communities eliminates redundant infrastructure and training which translates to lower costs when deploying fiber to new communities.
Even with generous government incentives, spurring fiber deployment through tax incentives will remain difficult because the cost of deploying fiber to the home per subscriber can easily be more than $4,000[xv] and that takes a tremendous amount of external incentives to get any company to assume that kind of risk. Even Verizon with its ability to absorb the near term losses and lower FTTH deployment costs due to a high percentage of aerial fiber deployment faced immense pressure from their shareholders for their decision to deploy FTTH. Qwest has stated that over 75% of their cabling is underground whereas Verizon is just the opposite[xvi]. This is a crucial distinction because a study in San Francisco showed that underground fiber costs 6.69 times more than aerial fiber to deploy.[xvii] That could easily mean that Qwest’s FTTH deployment costs are a few times higher than Verizon. The end result many of these projects are not viable without government subsidies.
Policies Regarding Muni Fiber
While municipal fiber overbuilding almost always represents a waste of community and broader societal resources, there may be times when municipal fiber provision may be appropriate. Cities rightly seeking to have faster broadband networks, however, should consider municipal provision as a last resort, rather than a first one. Assuming that a community has at least one broadband provider, their first step should be to try to incentivize their existing cable or phone operators to reach a certain broadband performance milestone, and not some specific technology such as FTTH. They should do this by engaging in a public private partnership to fund at least some of the costs of expanding and upgrading the existing networks. They should also reduce regulatory barriers such as right of way fees.
If municipalities try to work in good faith with incumbent providers to have them upgrade their networks and the operators refuse to move forward with reasonable and timely performance milestones then cities should consider going the municipal fiber route, assuming that they believe the economics of doing so are viable. However, cities that are seriously considering municipal fiber should also factor in accelerated deployment and aggressive competition from cable and phone companies in any viability analysis.
The situation is different for areas without any broadband providers. There municipal fiber provision can make more economic sense, not just for them but for society as a whole. But in these cases a key factor is whether the municipality has the technical wherewithal to build, manage and upgrade networks and do so efficiently.
Over the years, ITIF has been a most vocal proponent of faster broadband and we have just released a new report “The Need for Speed: The Importance of Next-Generation Broadband Networks” stressing the need for next generation broadband[xviii]. However, we advocate a more pragmatic, technology neutral approach where companies achieve the required level of performance (currently defined by a minimum of 20 Mbps downstream). While we would ultimately prefer to see more FTTH deployment because it brings us into the gigabit era and beyond, we recognize that consumer demand may not always be high enough to make it economically feasible today. What’s more important is reaching next generation broadband speeds today with any feasible technology and cross tomorrow’s milestones when we reach that bridge.
Robert D. Atkinson, Daniel K. Correa and Julie A. Hedlund, “Explaining International Broadband Leadership”, (Washington, D.C.: Information Technology Innovation Foundation, May 2008) <http://www.itif.org/index.php?id=142> (accessed March 12, 2009).
David Orenstein, “Pioneer of digital subscriber line wins prestigious fellowship”, (Stanford News Service, September 2006) <http://news-service.stanford.edu/news/2006/september27/cioffi-092706.html> (accessed March 12, 2009).
OECD, “OECD broadband portal”, n.d. <http://www.oecd.org/document/54/0,3343,en_2649_34225_38690102_1_1_1_1,00.html> (accessed March 12, 2009).
IDC, “IDC 2008 U.S. Consumer Panel Broadband Survey”, (IDC, August 2008) <http://www.idc.com/getdoc.jsp?containerId=213724> (accessed March 12, 2009).
NCTA, “industry data”, n.d. <http://www.ncta.com/Statistics.aspx> (accessed March 12, 2009).
Ford models how reduced market size reduces the number of profitable providers. George S. Ford, Competition After Unbundling: Entry, Industry Structure, and Convergence, 59 Fed. Comm. L.J. 331, 332-67 (2007).
Verizon’s FIOS strategy requires considerable capital. Comcast’s recently announced DOCSIS 3.0 investment is estimated to cost less, but will still be in the billions of dollars. Whether such high-speed networks will be rolled out in most places, though, remains to be seen.
Grace Leong and Joe Pyrah, “The Case for UTOPIA and iProvo: Double down or cut bait?” (Daily Herald, April 2008) <http://www.heraldextra.com/content/view/263223/18/> (accessed March 12, 2009).
Grace Leong and Joe Pyrah, “The Case for UTOPIA and iProvo: Double down or cut bait?” (Daily Herald, April 2008) <http://www.heraldextra.com/content/view/263223/18/> (accessed March 12, 2009).
Ed Gubbins, “Bell tolls for wholesale-only muni fiber”, (Telephony Online, May 2008) <http://telephonyonline.com/fttp/news/telecom_bell_tolls_wholesaleonly/index.html> (accessed March 12, 2009).
Burlington Telecom, “Residential Services,” n.d. <http://www.burlingtontelecom.net/residential/internet/> (accessed on March 3,2009).
Verizon, “Plans and Prices,” n.d. <http://www22.verizon.com/content/consumerfios/packages+and+prices/packages+and+prices.htm> (accessed on March 3, 2009).
Larry Dignan, “Qwest CTO on FTTP, bandwidth caps and integrated services”, (ZDNet, August 2008) <http://blogs.zdnet.com/BTL/?p=9781> (accessed March 12, 2009).
CTC Communications, “Fiber Optics for Government and Public Broadband: A Feasibility Study”, (San Francisco Government, January 2007) <http://www.sfgov.org/site/uploadedfiles/dtis/tech_connect/SFFiberFeasibility.pdf> (accessed March 12, 2009).
Stephen Ezell, Robert Atkinson, Daniel Castro and George Ou, The Need for Speed: The Importance of Next-Generation Broadband Networks, (Washington, D.C.: Information Technology Innovation Foundation, March 2009) <http://www.itif.org/index.php?id=231> (accessed March 12, 2009).