Mar 29 2010
A Perspective from the Chair: Complete Version
One of my goals is to formulate the best environment for people and businesses to thrive in this region (1). To sustain a successful business and learning environment, we must make sure affordable, next-generation broadband access is available to all residents and businesses of all sizes. Our neighbors in other smaller cities and regions have capitalized on this proposition and are investing in stronger technology systems to attract people and businesses. Maintaining the status quo regarding Internet speed and capacity could mean being left behind. To suggest that Seattle has neglected a plan to deploy high speed fiber is somewhat naïve and inaccurate. There are clearly reasons why no major city has deployed a municipally owned fiber system and why Clarksville, Tennessee; Lafayette, Louisiana; and Monticello, Minnesota are cited as the examples of cities that have deployed it. There is a reason why San Francisco, Portland and St. Paul have not launched a city-owned system. Irrespective of these facts, I agree that an opportunity for national leadership on this issue presents itself.
I have played major roles in new product and service deployment so this is the type of opportunity that I embrace. But here, there is no room for impetuousness because our citizens deserve better; this is not a lab experiment or classroom exercise. We can cite numerous examples of failed or troubled municipally-owned fiber networks.
Let me share some details about the policy issues related to providing public high-speed broadband access for all residents in Seattle. While past Councilmembers have concerned themselves with this issue, I am not sure whether they had the data that we now have, or the wherewithal to drill deep into what this may now mean or cost; or in contrast, what it may mean or cost not to have full broadband access.
In short, Seattle is a technology and business leader. To continue that leadership and remain a renowned hub of innovation, Seattle should have a network that provides residents with download speeds of 20 Megabits-per-second (Mbps) to 100 Mbps and upload speeds of 10 Mbps to 50 Mbps. The network infrastructure should be capable of offering symmetrical download and upload speeds of 100 Mbps by 2020. This type of high-speed Internet would advance our access to medical monitoring tools, distant learning tools, and enhance running a small business from your home.
Why? To a practical user, what does a 50Mbps or 100 Mbps connection mean? If you just surf web pages, upload some photos, watch streaming videos or download music, then the bandwidth capacity from today’s cable and DSL services are fine; 1 to 6 Mbps. However, web services are rapidly increasing and the “New Internet” will require higher bandwidth for its applications and services. Internet video accounts for one-third of all consumer traffic and is expected to account for 60% of all consumer traffic by 2013. The culmination of all video modes (television cable, video on demand, Internet streaming videos, and peer-to-peer) will comprise 91% of global consumer traffic by 2013. Additionally, many companies are now offering cloud computing services or software as service (SAS) applications that run on the Internet instead of being installed on your computer. Without actually rebuilding the Internet system to improve efficiencies and securities in the overall Internet system, a host of these video and cloud applications for personal and business use will require greater bandwidth than today’s speeds. The use of gesture-based computer interfaces, as seen in the movie “Minority Report” are in research and development and fiber can facilitate bringing such sophisticated applications to your home or business.
On February 16, 2010, the Federal Communications Commission (FCC) finally put a jolt in America’s broadband position by calling for 100 Mbps Internet Access by 2020. In comparison, Japan averages 61 Mbps and the U.S. averages 4.8 Mbps. Average cost for 1Mbps is $.27 for Japan and $3.33 for the U.S.
What this means is that we cannot stay on the same track regarding Internet capacity and speed. We can lose any advantages that we now have. If we believe this premise, as I do, then the delivery of these speeds should be comparable to other cities around the world that offer these speeds at affordable prices.
Who should build it? I have approached this question by first determining whether the private market participants would even be willing or capable of meeting the demand in terms of speed, pricing, and service. If we agree there is a need that should be met, then our private market system is designed to take advantage of the market opportunity. Strategically, the City should aggressively incentivize the private market providers to enter. This approach smartly uses the healthy market forces of competition to drive better service and pricing points. It allows the core competencies of technology companies to penetrate an emerging market and compete for our business. These market forces can be lost in a system that lends itself to a monopoly and in particular, one owned by the government. This strategy is important because we must first understand what market forces are in play which discourages private companies from entering. For example, many smaller cable/content providers who have chosen to get out of that line of business cite as a reason their lack of negotiating power to compete against larger, better funded cable/content providers.
However, if the private market cannot or chooses not to meet this emerging need and a record is established that demonstrates this conclusion, the city should aggressively fill this void by lawfully investing into this opportunity as cities have done in other jurisdictions throughout the country. I use the term “aggressively,” because the city should not condone any tactical delays caused by incumbent providers.
I must note however, that a business case should first be made prior to market entry. This business case should include, among other things, a demonstration of the private market’s unwillingness to enter the market; an analysis of the customer take-rate that will be necessary to repay the revenue bonds necessary to build the network; and whether a limited market trial should be established prior to city-wide deployment. Since fiber is capable of providing cable television, Internet and landline phone services, a decision as to what line of business the city should offer should also be established. Comcast and Verizon for example, offer bundled services referred to as “triple plays” and Seattle should decide whether it would also provide a triple play. As stated below, an “Internet only” play, for a variety of reasons, makes better sense from the Chair’s perspective.
What role should the City play? There are a few options the City can pursue. The City can stimulate the market by building fiber-to-the-neighborhood (FTTN) which entails approximately 800 miles of new fiber strands at a cost of roughly $150 million. The City already has approximately 450-500 miles in place, 200 of which connect 10 high schools, 9 middle schools and eventually 70 elementary schools. In this scenario, the City would attract private companies to complete the “last mile” to the premise and would pay back the bonds that paid for this construction through access fees accessed to the private company. Of course we would first have to validate the private companies’ interest prior to building the FTTN and prior to selling revenue bonds.
The other option is for the City to actually build a fiber-to-the-premise (FTTP) network, running fiber all the way to the home or building. In the most recent financial model, to equip all homes and businesses with fiber, the investment needed could easily be upwards of $700 to $800 million dollars. A 2007 study estimates the time of construction to occur over 18-months for engineering, construction, and setting up operation facilities. The various financial models are discussed further below, analyzing the benefits and risks to the City.
What is my bottom line? I believe we should immediately determine whether we have exhausted all measures of collaboration with current service providers. The City of Seattle has established monopolies in providing power, water and solid waste services as examples. Establishing another city-controlled monopoly, even if using a wholesale model, requires careful examination and justification. If we have a record that the private market is unwilling to meet the needs of our city and clearly understand the business basis for that decision, then Seattle should build a fiber network under the conditions described in this memorandum.
Are Seattle’s current needs met?
On a daily basis, our citizens demand public safety, safe roads and transit, environmental sustainability. Are they demanding this kind of fiber and capacity? Well, many are and many are not. Here is what we do know: people want a system that drives lower prices for the services they now use and they want a system that allows Seattle to remain affordable and capable of retaining and attracting businesses that can compete on the global market. They also want their kids to have the best educational opportunities available.
Here is what we have now. Our current infrastructure to deliver Internet, cable television, and phone/voice within the 84 square miles in Seattle is a hybrid system consisting of a fiber backbone with copper running the last few miles to the home or premise. Copper refers to twisted pair telephone lines as well as coaxial cable. That is most likely what you have now.
The major wired service providers in Seattle are Comcast, Qwest, and Broadstripe. The City has non-exclusive contracts with Comcast and Broadstripe, meaning any company can actually enter the Seattle market and provide service. Between the three service providers, the range of speed and prices for residents are $27.95 for 512 Kilobits-per-second (Kbps) and $54.95 for 15 Megabits-per-second (Mbps) download and 3Mbps upload. In contrast, for a small business seeking download speeds of at least 20 Mbps and at least 5Mbps upload, the Comcast business class costs $189.95 and Verizon offers 20 Mbps and 20 Mbps upload (more upload) for $139.99.
It should be noted that our current copper configuration is relatively limited. For example, in order to accommodate higher speeds, most of us would require an upgrade to the Data Over Cable Service Interface Specification 3.0 (DOCSIS 3.0) for our cable modem. However, even with DOCSIS 3.0, the fiber/coax combination would technically max out at 152 Mbps download and 108 Mbps upload. The problem is that these speeds are based on the best channel configuration for data and most service providers deliver video, Internet, and voice on the same line source, so these speeds would not be achieved because of the multiple data layers. With fiber to home, the technical ceiling is 10 Gigabits-per-second (Gbps) or 10,000 Mbps.
Google also recently announced (2/10/2010) they want to deliver 1 Gbps (1000 Mbps) fiber-to-the-home connections. The benchmark for next generation broadband bandwidth is a symmetrical 100 Mbps download and 100 Mbps upload connection.
In this era, technology and next generation applications and services are comparable to the carrot and stick analogy. Advanced technological innovations like two-way classrooms, where one classroom can interact with another classroom 3000 miles away are being discussed and, as described above, gesture-based computing has already become a reality. Cities will compete for businesses and people. Seattle can lose opportunities by not offering next-generation broadband speeds at competitive prices.
If service providers do not provide next generation broadband speeds at competitive prices, the most important reason for a City to build a fiber network is the strategic advantage that residents and businesses will have compared to other cities in terms of competition, innovation, and high-speed fiber capacity. Providing affordable high-speed Internet for all 70,000 school children will improve their education; a student without Internet access will be at a disadvantage. Because of a competitive market on an open access network with multiple service providers, prices will decrease from current prices.
It must be noted: a survey conducted in 2008 showed that 75% of Seattleites believe the City has a major role in ensuring that students, parents, and teachers have high-speed Internet access. The survey also indicated that 49% of the respondents said the City should build a publicly-financed communications network.
Other significant benefits. Data shows that a $500 million plus, fiber-to-the-premise investment in Seattle will lead to 10,000 direct and indirect jobs. The environmental monetary savings for Seattle will total $324 million in terms of offsetting vehicle expenses, traffic congestion, electricity, and improving teleconferencing. The monetary health care benefits total $602 million from reduced costs and enhancement of quality healthcare. Medical cost savings associated with cardiovascular and diabetes total $500 million. The Seattle City Council announced on February 22, 2010, its desire for Seattle to be a carbon neutral City by 2030. A major piece of the formula to achieve carbon neutrality will be the environmental benefits from a fiber-to-the-premise network. The reduction of vehicle miles from teleconferencing could lead to an annual reduction of 535 million kilograms of CO2 emissions, which is equivalent to removing 102,294 vehicles from the road (conversion from http://www.epa.gov/RDEE/energy-resources/calculator.html). Note: the methodology for measuring the exact carbon credits as a result of a fiber-to-the-premise system is difficult because of unclear baselines, quantifying the secondary effects, and the lack of current protocols for measuring credits of a fiber-to-the-premise network.
The issue of a municipally owned fiber system is not as simple as determining whether Seattle should build one; it also requires an examination of what they would actually own and who their customers would be.
Let me describe the different fiber network scenarios:
Option 1, Fiber-to-the-neighborhood (FTTN): The City builds and controls a fiber-to-the-neighborhood network in Seattle. The initial $150 million dollar capital to build the fiber-to-the-neighborhood network attracts investments from other companies to build the “last mile”. The customer at the premise or business can also contract with a company to build the last mile to their home. Private companies will be selected to operate the network and offer services to these residences and businesses and will pay the City an access fee to use the City’s neighborhood fiber network. These access fees will pay back the $150 million dollar revenue bonds. In a fiber-to-the-neighborhood model, the following market penetration percentages must be met for success: Residential Market: 21 percent cable television, 54 percent Internet, 15 percent telephone. Business Market: 21 percent cable television, 36 percent Internet, and 27 percent telephone. Average market share required is 30%.
Option 2, Fiber-to-the-premise (Wholesale model): The City builds and controls a fiber-to-the-premise network in Seattle. The capital investment will total $450 – $500 million dollars with a recent study mentioning an $800 million dollar price tag. The City will run 1600-1800 new miles of fiber in the City. Because of the elimination of the upfront capital costs, companies would be enticed to provide service on the City network and pay the City an access fee. The City does not generate content, but instead provides the pipe. Revenues from the access fee will pay back the revenue bonds. This is called the wholesale network model. Market share required to pay back the revenue bonds is 30%.
Option 3, Municipal Fiber-to-the-premise (Retail Model): The City builds, controls, and sells retail services to customers. This is the direct retail model and would be a new Utility similar to Seattle City Light. 2,000 customers would need to be added per month to meet the market share of 24%. This Retail model needs a smaller amount of market penetration to pay back the revenue bonds.
Option 4, Municipal Fiber-to-the-premise (Hybrid Model): This model combines the retail and wholesale models. The City builds, controls, and sells retail services for seven years. This period would require the City to capture a significant market share to generate enough revenue to pay back the revenue bonds. After seven years or when the City repays the bonds, the fiber network would be opened up for all retailers. This model allows the city to meet its bond obligations and then open access after seven to ten years would meet the innovation and price competition criteria. When the City opens up the network after repaying the bonds, it has the option to remain as a retailer and compete with other service providers and play a major role in the digital divide issue. The market share required for this model is projected to be 24%.
Comparing the 4 options: Of the three models: retail model, wholesale model, and hybrid model, the retail model would require the biggest bond, $478 million. The hybrid model is $473 million and the wholesale model is $459 million. With the desired market share met, the breakeven point for the retail and hybrid model will be 14 years. For the wholesale model, 18 years. The retail model would require the most employees at 190; 161 for the hybrid model; and 33 employees for the wholesale model.
The Finances & the Risks.
The financial projections for a fiber-to-the-neighborhood network are highly speculative and there are no proven municipal examples in the U.S. Of the 57 public providers encompassing 85 cities in North America, not one example is a major metropolitan city. How does the City guarantee that companies are willing to pay access fees to use the City’s network, meet the market penetration rates, and manage the network effectively?
If the City only invested $150 million in a fiber-to-the-neighborhood network and no private partners come forward, the City would then have to sell the FTTN infrastructure to an entity willing to operate and maintain it or complete the fiber-to-the-premise build out at a cost of an additional $350-$650 million. With this complete fiber build out, the City would become the open-access provider and depend on companies to provide service and pay access fees to use the network in order to pay back the revenue bonds. The City also has the option of becoming the Internet service provider. If the City was unable to find partners to build the last mile in the fiber-to-the-neighborhood scenario, the City would clearly lose money if they were required to sell the FTTN system. Because of this, I believe the City should be prepared to commit itself from the outset to the finances that would be required for a fiber-to-the-premise network.
The retail model assumes the City will build and operate the network. My recommendation is for the City to only provide broadband service. But let’s examine all the major operating and maintenance costs associated with a telecommunication network (Internet, video, and voice). It would be critical for the City to provide excellent customer service to its new customers. At a 40% market penetration rate, 203 employees would be required to take orders, answer general customer questions, answer technical questions, billing, operate a main office, maintain and monitor equipment to run the network. Fiber Installations at the home or business would most likely be performed by contractors. The City would also have to pay for the connection to the Internet backbone. After ten years, it is estimated to cost $8 million per year for this connection to the Internet. If the City ventured into the cable programming business, the City would need to purchase programming from the National Cable Television Cooperative (exact costs are not available). Debt financing would be used to complete the operation costs. The financial models suggest a municipal revenue bond at a 5.2% interest rate for debt financing. To reiterate, all financial models show that the bond will fund 100% of the fiber-to-the-premise business with no required cash from City.
Problems with the Wholesale Model: Data from 2007 identifies several issues with the wholesale model. Because service providers will be paying the City an access fee to use the fiber-to-the-premise network, one significant issue could arise as to whether they would cherry pick the customer base. The numbers from that study suggested that the City needed to collect at least $40 for access to residential customers to pay back the FTTP network. The retailer would therefore target customers willing to pay $75 per month on services. Households that spend $40 or $50 per month would likely be excluded from the fiber network since no retailer would choose to serve them.
For small business customers at $75 per month and $150 per month for large businesses, the same vulnerability to cherry picking exist. In most markets, roughly 20% of customers buy the most expensive full package of products and a cherry picker will target that 20% of the market and if they do, well, there will not be enough sales to sustain Seattle as the network owner.
A 30% overall market penetration is required for the $459 million wholesale network bond. It will be absolutely critical that the market penetration percentages be met. If the service providers are charged a connection fee of $6 per-month-per-customer to pay back the revenue bonds, the City would receive a cash balance of $4.1 million by the end of the 20 year term. On the other hand, if the City only charged $5 per-month-per-customer, the City would see a negative cash balance of $32.9 million.
In the wholesale model, the City would build the core fiber-to-the-premise network and sell access to the network to private retail providers. The biggest key for success in a wholesale model is the speed at which private providers can add customers to the network. To be safe, the City should set as a goal, 6 quality retail providers offering service on the network. In order to maintain the cash balance, retailers must add 1,500 customers per month for several years to be successful. On paper, the wholesale model is the model for generating innovation, competitive pricing and programming, but has the greatest risk for the City. If the retailers on the City’s network fail, the City will lose money. Because no major commitments to capital were made by the retailer, the retailer might choose to abandon the market if their position in the market was not successful.
The Retail Model: The retail model in which the City builds and sells services on the fiber-to-the-premise network presents different risks than the wholesale model. Data shows that the City would need to provide a 15-20% discount from today’s incumbent prices for all services in order to achieve the needed market share for success. The savings are projected to be $2 billion over 20 years for the local economy. Data also shows that the City should only compete for services for residential customers and small business customers. The data recommends that the City not compete for large business customers because of the complexity of large business networks. Fiber access or the option for fiber is already available for the downtown buildings. One great cost benefit for the City to build and own a fiber-to-the-premise network is access to its own network. All City locations (schools, libraries, community centers, recreation centers, fire stations, and substations) would be connected. The City would bill itself and the revenue would stay in the City’s budget.
In order to support the debt, the City would need to add customers quickly. One of the significant capital costs with a fiber-to-the-premise network is that a large percentage of capital costs are at the customer end. The total costs at the customer end are approximately half the capital costs of building the entire network. The costs pay for the fiber drop to the building and installation of an optical network terminal (ONT). The optical network terminal installed at the home or building transitions the fiber optic line to the coaxial or twisted pair lines in the building. To build the customer base, the City should use a sign-up list during construction to meet the goal of adding 2,000 customers per month. By adding 2,000 customers per month, the City would reach 40% market penetration after five years. Attracting small businesses would require a door-to-door approach to educate and market the new City service. Businesses often put a greater value on reliability than price for telecommunication networks, so market penetration is usually slower for businesses.
The End Game:
Because of the technological capabilities, innovation, economic and educational development that could result from building a fiber-to-the-premise network, no one debates the merits of such a network. The elephant in the room is the cost and how to go about building one where incumbent providers exist. Both retail and wholesale models have pros and cons and would generate enough cash to pay back the debt over 20 years depending on market penetration. From a financial perspective, the retail model is a better model, generating $270 million in cash over 20 years at 30% market share. The wholesale model would only generate $7 million. At 40% market share, the retail model would generate $488 million. Because to date, no private company has decided to build and operate such a network in Seattle, the retail model does present itself as an option for the City of Seattle.
If the City was the service provider, data from 2007 suggests that it would offer across the board 15-17% discounts on all telecommunication services. On the other hand, a wholesale model with multiple service providers providing a menu of different service packages in Internet, cable TV, and voice would bring more innovation to the City.
Competitor Reactions: I would anticipate the fact that if the City became the one provider of fiber service, it would provoke litigation from the incumbent carrier similar to what has occurred in other cities. But the telecommunication service industry is not a robust industry of retailers providing Internet, cable TV, and voice. For the City to enter the telecommunication service arena, we would be competing against companies with more experience and strategic plans that have examined decades of consumer behavior. This is a huge undertaking and of the three fiber services, a strong argument could be made that the City should only enter the Internet business.
As you can see, like all major capital investments, this is a very complex issue. I will continue my dialog with Mayor McGinn, Bill Schrier (Director of the Department of Information Technology), the Citizen’s Telecommunication and Technology Advisory Board, Seattle City Light, and people in the community seeking a better broadband system. It is vital that Seattle City Light and the Department of Information Technology work together on this issue because of its synergy with Smart Grid. As you may recall, I requested Seattle City Light to report back to Council by the middle of this year on a plan that positions Seattle to develop a Smart Grid. In the next six-eight months, I will work with the Mayor to reach an agreement and be in a position for the City to offer next generation broadband access to all residents at affordable competitive prices.