1 Toward a National Broadband Policy

I commend the NTIA for its initiative in addressing the need for a national policy to promote innovation and investment in broadband access and applications. Since the passage of the Telecommunications Act of 1996, the Internet has become a dominant feature of the communications landscape. At the time, the Internet was in its infancy; some wondered if it was any more than a passing fancy, so it is not surprising that the Act was concerned almost solely with voice-grade communications. In less than six years since the passage of the Act, the Internet has become one of the most significant and revolutionary technological changes of human history. Thus, there can be no doubt about the power of digital convergence to accelerate technological innovation, and the potential benefits of broadband access to further stimulate productivity and economic growth; improve education and access to information; and increase a community through connectivity.

While few could foresee the Internet explosion at the time, Congress did recognize, in generic terms, the importance of public policies to promote the development of advanced telecommunications services, which surely would include broadband access. In Section 706 of the Act, Congress instructed that:

The Commission and each State commission with regulatory jurisdiction over telecommunications services shall encourage the deployment on a reasonable and timely basis of advanced telecommunications capability to all Americans (including, in particular, elementary and secondary schools and classrooms) by utilizing, in a manner consistent with the public interest, convenience, and necessity, price cap regulation, regulatory forbearance, measures that promote competition in the local telecommunications market, or other regulating methods that remove barriers to infrastructure investment.[1]

A reasonable reading of this provision of the Act suggests that very different regulatory polices toward advanced services—especially broadband access—should have been implemented by the FCC and state commissions. Unfortunately, that has not been the case: in their implementation of the Act, the Federal Communications Commission (FCC) and state commissions have developed regulatory policies that are one-sided: incumbent local exchange carriers (ILEC’s) are heavily regulated, while their competitors are not, whether “competitive” local exchange carriers (CLEC’s), cable operators, inter-exchange carriers (IXC’s), mobile carriers, satellite carriers, stationary wireless carriers, or any other mode of communications or type of service provider. Rather than “removing barriers to infrastructure investment” in broadband networks, regulators appear to have been erecting them. For that reason, broadband access is not developing at the rate it could be.

Fortunately, there is growing recognition of the need for major changes in our nation’s public policies, and the NTIA can and should play an important role in that process. As it undertakes that effort, I strongly support the “Guideposts for Broadband Policy Development” enumerated by NTIA Administrator Nancy Victory:

· facilitating deployment of new technologies by eliminating any roadblocks;

· promoting efficient facilities investment to gain the network reliability and security advantages of a diversity of facilities-based competitors;

· promoting competition in a technology-neutral way and being mindful that the market “might not always work as well or at the same pace in all areas.”[2]

Likewise, the leadership of the FCC has acknowledged the need for major policy changes. Chairman Powell has noted that development of a national broadband policy is necessary to correct what thus far had been government policy of “lurching and reacting” to unanswered questions about broadband.[3] Commissioner Abernathy has urged that policy-makers to learn from experience in the wireless and long distance service markets-that relying on market forces as much as possible offers the ‘”best means of delivering innovative services and lower prices to consumers.” She also cautioned against the “risks associated with too much regulation,” noting that the FCC lost sight of the “danger of over-regulation” in its efforts to implement the Telecommunications Act of 1996. She expressed the FCC’s intent “to restore the incentives for facilities-based investment that Congress intended… This means a shift away from policies that actively encourage resale as a long-term business strategy and force the unbundling of virtually every network element at rates based on TELRIC.”[4]

It should be understood, though, that regulatory changes will not come easily. Many firms benefit from regulatory policies that hamper their competitors. During the past six years, CLEC’s, IXC’s and cable companies have been strident advocates of regulations that apply asymmetrically to only one class of service providers, the ILEC’s. No matter that those regulations hinder innovation and investment in broadband infrastructure. Moreover, state commissions have, in some cases, gone even beyond the FCC in adopting regulations that increase obstacles and reduce incentives for investment in broadband access.

Thus, the NTIA has a critical role to play in advocating and organizing changes in public policy. Many of the necessary changes can be accomplished through administrative proceedings; in some cases, though, legislative changes will be required. In either case, the NTIA can and should be a voice for change in regulatory, tax and right-of-way policies, and by encouraging government agencies to "lead by example" in their own use of broadband services, through procurement practices.[5] In much the same way that the government has been a key customer for other new technologies (and sponsored the early development of the Internet), the government can demonstrate the efficacy of broadband applications and thereby increase demand for more rapid investment in broadband networks.

2 Defining Broadband Access

2.1 Broadband and Digital Convergence

Both wireline and wireless networks were designed and built to carry analog traffic (voice, audio or audio-video). As the use of PCs for Internet and remote Local Area Network (LAN) access increased, end users added terminal equipment to move data over those voice networks (hence, modems to convert digital signals to analog signals, and Internet connections via “dial-up access”). This represented the first stage in the development of data networks. We are now well into the next stage: with digital convergence, carriers will need to substantially upgrade the existing infrastructure to carry voice, data and video. The expensive process of upgrading analog networks (copper twisted pair or coaxial) to provide digital access is well along, but the cost of upgrading increases markedly as one moves from the dense core of those networks in the major cities to the less dense peripheries in rural areas.

The fundamental change in these developments is from analog to digital and circuit- to packet-switched networks. This shift not only increases access speeds (typically from analog rates of 28-56 kbps to digital rates of 256 kbps –1.5 Mbps), but, even more importantly, “digital convergence” facilitates intermodal competition (i.e. competition among services provided over different technologies), and interconnection of and interoperability across modes. No wonder Chairman Powell has lamented “pervasive references to broadband as ‘a simple incremental advance from telephone service.’"[6]

Digital convergence also represents a technological paradigm shift, in that the rate of technological change (e.g., the rate at which bandwidth increases) will occur much faster than it did in analog networks. As this paradigm shift occurs, telecommunications will come much closer to following Moore’s law,[7] since microelectronics (and opto-electronics) will drive technological change in digital networks. This will be a major benefit: consider, for example, how long it took to evolve from 300 baud or .3 kb modems to 56 kb modems on the one hand versus the much faster rate of change from OC-3 to OC-192 and beyond. These differential rates of change flow directly from the inherent differences in analog versus digital technologies.

Thus, the fundamental distinction that should be made in defining “broadband” access is NOT transmission speed, but class of technology. On one side are legacy analog systems that deliver audio, video and voice over wires or airwaves. Even though those networks can be used to send data in digital form, they were not designed to do so. On the other side are networks that provide access by means of “data-rate, always-on, digital packet” transmission.

Thus, to capture this paradigm shift in communications and to form the basis for public policies that will promote wider and more rapid deployment, broadband access should be defined in these terms: “any network or technology that is built or modified to carry digital data traffic and provide end users with always-on access to one or more data networks.” [8] In short hand, “broadband” equals “digital data,” where data can be used to carry an enormous range of information—words, numbers, voice, audio, pictures, video, etc. The distinguishing characteristic of digital data networks is that they enable digital devices to speak to each other in their own language.

This definition also captures the fundamental difference between users adding equipment (e.g., a modem) to allow digital devices to communicate over analog networks and modifying or building networks that are digital. In the latter case, the incentive—or disincentive—effects of public policies on technological innovation and network investment become crucial factors in the rate of deployment and adoption. In the days of analog modems, it was expenditures by consumers that determined the rate of Internet access penetration, given a ubiquitous analog network. Today, and more so in the future, investment by carriers and service providers in expanding and developing new digital networks—by whatever technology—will determine the availability of broadband access.

As a practical matter, this definition of broadband implies access speeds equal to or greater than 256 kb downstream, the minimum speed for most cable modem and DSL users. However, this definition will not be static with respect to bandwidth: as computer processing speeds increase, larger storage capacities decrease in cost (e.g., server farms, hard-drives, RW-DVDs) and higher-bandwidth applications spread (video email, video telephony), broadband will be continuously redefined at higher speeds. At some point, we will no doubt distinguish the first generation of broadband access from the next generation.

According to this definition, one analyst estimates that about 10% of American households (10.85 million households, by end of 2001) use broadband access to the Internet and other networks (e.g., enterprise LANs for work-at-home). Of those with broadband access, 58% are using cable modem, 37% are using DSL, and 5% are using another technology (wireless, satellite). Penetration rates are expected to increase rapidly, to 35%, or 41 million households by 2005, with market shares of 53% cable modem, 35% DSL, 9% satellite and 3% optical.[9] Other estimates of broadband access penetration and modal shares are shown in Table 1.

Consistent with the focus of the NTIA notice, the measurements in Table 1 focuses on broadband access services for the mass market. It does not include the wide range of broadband access available to large businesses. Large businesses use high capacity services whose speeds can far exceed current cable modem and DSL speeds. These services (ranging from DS-1 to OC3+) are available through multiple competitors in urban areas throughout the country.[10]

Table 1: Estimates of U.S. Broadband Access

Penetration Rates and Modal Shares

Investment Firm	Broadband Subscribers, 2000	Broadband Subscribers, 2005
BMO Nesbitt Burns[11]	Cable modem: 70% DSL: 30% Other: excluded	Cable modem: 63% DSL: 37% Other: excluded
Jefferies & Company[12]	Cable modem: 61% DSL: 37% Other: 2%	Cable modem: 47% DSL: 44% Other: 9%
Salomon SmithBarney[13]	Cable modem: 71% DSL: 29% Other: 0%	Cable modem: 59% DSL: 34% Other: 7%
Lehman Brothers[14]	Cable modem: 67% DSL: 33% Other: excluded	Cable modem: 64% DSL: 36% Other: excluded

2.2 Broadband and Wireless Networks

Although many observers focus on broadband access over wireline networks—cable and DSL—there is every reason to believe that broadband access will also be realized over upgraded existing and newly built wireless networks as well. This has enormous implications for public policy: it means that (1) rational spectrum allocation and use policies are critical; and (2) policies that facilitate intermodal competition between wireline and wireless networks will best promote innovation and investment in broadband access facilities.

There are three major classes of wireless broadband access networks emerging: mobile, fixed and satellite. Like wireless telephone networks, both cellular and PCS mobile telephone networks were built for voice communications. The original cellular networks were analog (1G), and have been or are being converted to digital (2G), while PCS networks were digital from the start. In both cases, though, mobile networks have had only limited data capabilities, as anyone who has tried to use a mobile phone for Internet access well knows. There are two significant developments, though, that will change this markedly, namely 2.5G (general packet radio services, or GPRS) and 3G broadband digital data networks.

GPRS has already been deployed in Europe:

“The number of always-on mobile Internet users in Western Europe will grow to 110 million in 2006, from just a few million this year… One in three Western Europeans will use the latest mobile phone services technology… Business travelers will be the first to use the faster always-on connections that are offered by GPRS packet-switched technology.”[15]

GPRS services will soon be offered in the U.S., followed soon thereafter by 3G:

“In the United States, carriers have been given the flexibility to choose which technology to use to deploy voice, as well as advanced mobile data, services. The two largest mobile telephone carriers that currently use CDMA as their 2G technology, Verizon Wireless and Sprint PCS, announced in early 2001 that they plan to roll out cdma2000 1X as the first phase of their 3G technology rollout during 2001…The major GSM and TDMA carriers in the United States, AT&T Wireless, VoiceStream, and Cingular Wireless, are taking a different migration path to 3G technology. All three carriers plan to deploy GPRS technology during 2001,which is expected to raise data transfer speeds to between 25 and 144 kbps.”[16]

Moreover, a new class of service provider is emerging for mobile broadband access, those deploying wireless local area network (WLAN) technology:

“Fast access to the Internet, at speeds 100 times greater than over a GSM phone, will soon be a reality for mobile workers, according to a new report, from Analysys. Public wireless local area network (WLAN) services enable users to connect laptops and PDAs to their Internet service providers or company intranets at speeds of up to 11Mbit/s… such services are now becoming available at airports, hotels and cafes in countries such as Austria, Germany, Norway and Sweden.”[17]

In addition to mobile wireless networks, there will be major developments in fixed wireless technologies for broadband access, using a host of alternatives, including LMDS, MMDS and WCS. Even though initial efforts in fixed wireless were not successful, there is growing evidence that further technological advances are in the offing:

"…there are currently over 210,000 subscribers to broadband fixed wireless services throughout the world, including both enterprise and residential customers. While the [Broadband Fixed Wireless Access] BFWA market has suffered somewhat, by 2005 service provider revenues from BFWA are expected to increase by 10 times its current level….’By circumventing the costs and time associated with laying expensive fiber, broadband fixed wireless technology offers an excellent means by which to capitalize on the vast potential of the broadband market,’ said Becky Diercks, director of In-Stat's Wireless Group.”[18]

“Wireless broadband operator Tele2 is close to achieving its planned target of 45 percent population coverage of the U.K. by the end of the year, and is also aiming for 65 percent coverage by the end of 2003. The carrier… offers wireless broadband services at a range of up to nine miles from a base station, at speeds of up to 2 megabits per second (Mbps).”[19]

“There is a growing opportunity for next-gen, fixed-wireless equipment vendors to quickly gain market share… Sprint and AT&T both recently put residential and small-business fixed wireless initiatives on hold due to difficulties with developing a viable business model. This has provided next-gen vendors with an opportunity to meet a rising demand for these solutions, thus establishing market leadership.”[20]

In addition to these terrestrial wireless developments, satellite communications service providers (e.g., DirectPC) now offer Internet access and pending network upgrades will substantially improve the quality of broadband access and services. For example, Hughes Network Services plans to have its “Spaceway” system operating in 18 months. The system will consist of three satellites providing coverage in North America and delivering high-bandwidth services to residential and business customers.[21] Industry analysts believe that “Satellite offerings should become increasingly visible over the next 12-18 months, at first competing effectively in markets underserved by cable and xDSL and, over time, as part of a bundled video offer with strong appeal for certain customer segments….”[22]

2.3 Next Generation Broadband

As exciting as these developments in broadband access technologies may be, they are just the first stage. In each of these modes of broadband access, bandwidth will increase substantially, by an order of magnitude over first-generation broadband. Whereas access speeds in the analog access world was measured in tens of kilobits per second (i.e, 9.6-56 kbps), the current generation of broadband access is measured in hundreds of kilobits per second (i.e., 256-1,544 kbps). The next generation of broadband access will be measured in the thousands of kilobits, i.e., megabits. These speeds will be needed to support bandwidth intensive applications such as online gaming, video-on-demand and streaming video.[23]

Until a substantial number of subscribers have adopted first-generation broadband, the development of broadband applications will not develop sufficiently to create the demand for even higher bandwidth access or applications. Given the substantial investment required to implement next-generation services, current adoption is critically important. For example, one analyst estimates that the cost to implement fiber-to-the-home, which will pave the way for next-generation applications offered by the ILEC’s, will be approximately $5,000 per subscriber assuming a 50% penetration rate. This estimate increases to over $9,000 if the penetration is 25%.[24] Thus, it is crucial to adopt and implement public policies that clear away the regulatory obstacles and disincentives that are inhibiting innovation and investment in the current generation of broadband access technologies.

2.4 Implications of Broadband Definition

Defining broadband as digital data access is critical for regulatory policy: it compels us to draw a sharp distinction between voice-grade, dial-up analog circuit and data-rate, always-on, digital packet access, because the worst policy is one that intentionally or unintentionally applies analog voice regulation to the digital data services.

This technology-neutral definition of broadband will promote both intra- and intermodal competition. “The convergent nature of broadband will permit, if not foster, industry convergence and consolidation across traditional industry lines—cable television and telephone services are viewed today as separate markets, but the distinction will make less sense over time. Convergence is a potential enabler of competition…”[25]

Defining broadband as digital data access is also consistent with the NRC’s recommendation that “Broadband services should have sufficient performance—and wide enough penetration of service reaching that performance level—to encourage the deployment of new applications.”[26] As the NRC notes, this is critical to innovation because network access and applications development are interconnected in “chicken-and-egg” fashion:

“an application will not be made available until a critical fraction of subscribers receives a high enough level of performance to support it, yet service providers will not deploy higher-performance broadband until there is sufficient demand for it. The performance of a broadband service, therefore, [must] be good enough and improve sufficiently to facilitate this cycle and not impede it.”[27]

Thus, investments must be made in broadband deployment now to get a critical mass of broadband subscribers. A critical mass of broadband access subscribers is necessary to justify investment in broadband applications, which in turn generate the demand for next generation broadband access. These critical masses cannot be reached if regulations impede the current deployment of broadband.

3 Primary Policy Goals & Objectives

3.1 Promoting Intermodal Competition

One of the reasons why broadband has such enormous potential for being the engine of the next wave of innovation, productivity and economic growth is that there are so many different technologies for realizing its potential. As acknowledged by the National Research Council report, “popular accounts tend to focus on which technology or players are “ahead” in broadband deployment, broadband is not a horse race between technologies, with an eventual winner.”[28] Even so, there is most definitely a race underway among broadband technologies, but there is no finish line to that race; rather, it is a perpetual race and will have multiple winners. In other words, this perpetual technology race among modes of communications that are using and will use competing technologies to provide broadband access to end users, over digitized copper, coaxial or fiber optic cables, or over terrestrial or extraterrestrial wireless networks. The long-term outcome of this perpetual technology race will be diversity in technology options, because of geographic diversity; incremental investments in existing infrastructure; continued exploitation of technology skills across modes; and varying levels of technology maturity.

For this reason, public policies that promote intermodal competition are absolutely crucial to the rapid and widespread deployment of broadband access. The critical policy for promoting intermodal competition is regulatory symmetry, i.e., reducing the regulation of ILEC’s, by far the most highly regulated of all intermodal competitors. Promoting intermodal competition would stimulate innovation and investment in existing and new telecom network infrastructures, including telephone, cable, mobile wireless, stationary wireless and satellite.

Experience in surface freight transportation demonstrates the benefits of promoting intermodal competition. Prior to 1980, transport industries were regulated on the basis of modal competition, causing massive inefficiencies (e.g., empty backhauls in trucking, misallocation of traffic by mode) and financial failures (i.e., bankrupt railroads). The Staggers and Motor Carrier Reform Acts of 1980 promoted intermodal competition, leading to enormous gains in efficiency and productivity in freight transportation.[29]

3.2 Promoting Innovation by Adopting Technology Neutrality Policies

FCC Chairman Powell has noted that the Commission needs to work hard to remain "technology agnostic" so that it doesn't promote or discourage the deployment of any broadband technologies over others. Mr. Powell has acknowledged that the FCC “runs the risk” of preferring one technology over another “thereby drying up innovation and investment in a host” of other possible solutions.[30] Unfortunately, both the FCC’s and some states’ policies appeared to have singled out one class of service providers (ILEC’s), and, thereby, the technology they deploy (DSL), for regulation. All other actual and potential providers of broadband access and, thereby, all other broadband access technologies, are virtually unregulated. So, whether intentionally or not, current policies are not remotely technology neutral.

Technology neutrality is an important policy objective because it would promote a rich array of interconnected competing and complementary networks, ensuring the adoption and deployment of appropriate technologies, depending on location, applications and other factors. Neutrality would also promote technology competition to improve existing technologies and develop new ones, including technologies not yet imagined.

Finally, any policy that attempts to mandate deployment of a particular broadband access technology by a particular class of service providers (e.g., DSL by ILEC’s) will be counter-productive because it will cause inefficient use of that technology (e.g., wireline over wireless in rural areas) and inhibit technological innovation and the adoption of superior technologies (e.g., requiring DSL deployment specifically will slow the development of wireless broadband access technologies).

3.3 Promoting Investment and Facilities-Based Competition

Facilities-based competition ensures robustness and redundancy and protects against network breakdowns and outages. Thus, one of the key recommendations of the National Research Council is that U.S. broadband “Policies should favor facilities-based competition over mandated unbundling... Increasing the extent of competition through facilities ownership (and voluntary arrangements to open facilities) rather than relying on regulation that mandates unbundling…”[31]

As the NRC Report emphasizes, policies that promote facilities-based competition, rather than unbundling, have substantial benefits. They (1) reduce the need for persistent regulatory intervention; (2) permit the natural (i.e., competition-shaped) character of broadband service and industry structure to be discerned; (3) promote technological diversity; (4) avoid deterring competitors from investing in their own facilities; (5) remove disincentives to new investment by incumbents; (6) avoid costs and complications of coordination between incumbents and competitors; and (7) facilitate technical optimization of total bandwidth.[32]

So, facilities-based competition should be a high priority policy objective, but it should definitely not be limited to “same technology” or intramodal competition. Given actual and potential developments in broadband access across multiple technologies, we should remove policy obstacles and disincentives to investment in any technology, thereby promoting facilities-based competition across those technologies.

3.4 Promoting Widespread Deployment of Broadband Access

“Universal” broadband access is an important long term objective, but attempts to reach this objective in the short-to intermediate-run by “forcing” deployment, especially if targeted at one class of service providers, will be counter-productive. Rather, widespread broadband access can best be achieved through intermodal, facilities-based competition, which will stimulate the use of appropriate technologies under different circumstances (e.g., cable modems or DSL in cities and suburbs, WLANs on college campuses and office parks, satellite in rural areas).

The worst possible policy would be one that extends the traditional regulatory regime of analog voice communications to data services and broadband access, however noble the motivation may be. Attempting to achieve some kind of “universal broadband service” by regulating one class of service providers–ILEC’s–would substantially reduce their incentive to invest in infrastructure. That, in turn, would reduce the rate of infrastructure investment by their intermodal competitors, since a major stimulus for deploying broadband is meeting competition.

Thus, I strongly concur with the NRC recommendation:

“[Because] Some forms of [government] intervention to expand access... may affect private investment decisions, it should be undertaken with great care in this nascent area in order to avoid unintended consequences.”[33] [We should] “defer development of a universal services policy for broadband access until the nature of broadband services, pace of development, distribution of access and social significance become clearer.”[34]

At the same time, it may be desirable to provide public funding for broadband access in school libraries, senior centers and other public access points, so that individuals without a computer or desire for broadband access at home can gain broadband access in other convenient locations. Promoting broadband access in public places (e.g., schools, libraries, senior centers) through public funding will enable access by lower income or lower use households. Such support is currently being provided through the federal government's e-rate program, which committed nearly $6 billion between 1998 and 2000 to schools and libraries for the implementation of advanced services.[35] Additional targeted government subsidy programs may well be useful in further meeting the need for public broadband access and stimulating demand for development of broadband applications. Any such program, however, should be funded through general revenue sources or, at the least, through a tax that is technology- and competitively-neutral.

3.5 Eliminating Regulatory Obstacles and Disincentives

As noted in the introductory section, there is a large “disconnect” between our policy objectives and our policies toward broadband access. In an age of digital convergence, too many of our policies are geared for a voice world. I agree completely, therefore, with the assessment of the National Research Council:

“The present policy framework for broadband, which revolves around the Telecommunications Act of 1996, is problematic and unsuited in several respects to the new era of broadband services… the central role of the Internet in the communications landscape was not fully anticipated… the Telecommunications Act of 1996 devotes much of its attention to the voice telephony market and maintains distinct rules for the various communications networks (telephone, cable, cellular, broadcasting, and so on).”[36]

Thus, “problematic and unsuited” regulation is a major inhibitor of investment in broadband access networks. While less regulation is not a policy objective per se, it is the best means of achieving other policy objectives. Unfortunately, due to the long history of telephone regulation, and specific provisions of the Telecommunications Act of 1996, there has been a strong tendency to extend regulation from voice-analog services into broadband access services.

Hence, while I agree with the thought underlying the National Research Council’s recommendation to “defer new regulation in the early stages,”[37] it is not sufficient to merely defer new regulation—it is imperative that we repeal existing regulations that have been wrongly applied to broadband access services and—unless removed—will inhibit and distort innovation and investment in broadband access networks and services. Moreover, unless and until we decrease regulatory obstacles to facilities investment and intermodal competition in the current generation of broadband access, we will not get to the next generation of data access (fiber-to-the-home, broadband wireless). Slowing down investment in the current generation of broadband access will impede the development of the next generation.

What is especially harmful about existing regulation is that it is so highly asymmetric: for all practical purposes, only one set of service providers and, hence, one type of broadband technology is regulated, namely ILEC’s and DSL broadband access service. Other providers of broadband access are barely regulated, or not at all. That disparity in regulatory treatment of direct competitors in the market for broadband access services distorts competition and technological choices.

In assessing the weight that should be given to reducing regulation of broadband access, it should be noted that regulation is particularly harmful when applied to high technology industries, i.e., those in which technological innovation is the driving force for investment and deployment. Rapid advances in CPUs, PCs and other digital devices occurred because those “markets for innovation” were unconstrained by regulation. As such, chip manufacturers and PC manufacturers had every incentive to produce the fastest technology available. The net result of the competitive market is that consumers can now purchase a variety of PCs for less than $600 that have capabilities that far exceed most business computer systems a decade ago. Given the potential rate of technological change and the dramatic increases in intermodal competition, regulation of broadband services would be especially harmful because of its long-term dynamic effects on the “market for innovation.”

4 Disincentives for Investment in Broadband Access

4.1 Promoting Investment in Broadband Access Facilities

As discussed in Section 2, there are many different technologies for providing broadband access, and Section 3 explained why a national broadband policy should be technology neutral and should promote facilities-based intermodal competition. Unfortunately, current policies do neither. Even worse, there is a very real threat of policies—especially state regulation of ILEC’s—taking a turn for the worse. The prices for UNE-P (unbundled network elements-platform) are already below cost, but some states are considering lowering them even further. While the FCC has found that packet switching and DSL facilities needs to be unbundled in only limited circumstances, one state has, and other states are considering, requiring additional unbundling of advanced services. So, while public policies should be moving in one direction to achieve broadband policy objectives, they are actually moving in the opposite direction, toward even greater bias against DSL technology and even less incentive for innovation and investment in broadband access. It is imperative that NTIA marshal its resources to reverse this trend.

Unfortunately, there is a strong misperception that regulation is not hindering investment in broadband. Defenders of current regulatory policy cite the enormous investments ILEC’s have made in deploying DSL. So, for example, the FCC has argued that:

“Notwithstanding the fact that the incumbents have been on notice that they could be required to unbundle facilities used to provide advanced services, the incumbents have announced aggressive rollout plans for xDSL service. In fact, a recent financial analyst’s report indicates that advanced data services currently comprise an average of 9.9 percent of the revenues of the BOCs and GTE… We find these statistics to be significant because they demonstrate that the development of competition, and the threat of losing revenue and customers to carriers offering advanced services, provides a powerful incentive for carriers to invest.[38]

That is false logic for three main reasons. First, given the clear directive of Section 706 of the Telecom Act, it was reasonable for ILEC’s to assume—and make capital budgeting decisions based on that assumption—that regulators would not require mandatory unbundling or TELRIC pricing of DSL equipment. Given recent regulatory developments, particularly at the state level, that is no longer the case.

Second, the initial upgrades from an analog network to a digital network can be made relatively easily and inexpensively. The cost of that upgrade goes up dramatically, however, as one moves to the edges of the network. Thus, the ILEC’s have made the less expensive upgrades to provide broadband access on a substantial share of their networks; the question now, though, is whether they have sufficient incentives for the additional investments to push the digital upgrade further out into their networks. Given regulatory indisincentives, that is by no means assured.

Third, there has been a decided shift in capital markets, from emphasizing growth to corporate cash flow and earnings:

Ernst & Young reports many analysts in the fixed-line telecom market have altered their valuation strategy to focus heavily on free cash flows. Non-financial indicators of growth largely have been discarded as performance indicators, and analysts now are focusing on incremental achievements rather than long-term growth projections.[39]

Not surprisingly, this change in financial performance metrics already may be affecting investment:

“We believe ILECs in general are not being as aggressive as they were last year towards DSL deployment. At the present time, the investment community is focused on EPS and positive cash flow in determining stock valuations rather than growth in subscribers and revenues. In general, it takes two years for an ILEC to become cash flow positive on a DSL subscriber. Hence, slower subscriber growth improves near-term EPS and cash flow.”[40]

There can be little doubt that negative regulatory decisions, and growing uncertainty about even more unfavorable regulatory decisions, are harming ILEC investment incentives:

“RBOCs… are the major providers of residential high-speed Internet access via DSL in the U.S… but penetration rates are low relative to cable companies… due to… unfavourable regulatory decisions with respect to wholesale DSL services that continue to inhibit deployment.”[41]

“Cable modem’s advantage today is that it does not have to share or un-bundle its networks as do the ILECs. Lack of regulation provides a clear advantage [for cable] in service deployment.”[42]

“While regulatory developments continue [to] favor cable MSOs, the constraints on RBOCs are increasing. Line sharing with other competitive local exchange carriers (CLECs) has been required for the Bells… Moreover, the establishment of separate subsidiaries for DSL operations has been required.”[43]

Even if investment disincentives only reduce investment at the margin, they can substantially slow deployment and adoption because of the effect on (1) competitive dynamics and (2) network interdependencies between broadband availability and applications development (“the chicken and egg problem”). Thus, in the remainder of this section, we will review the disincentive effects of specific regulatory policies that are hampering investment in broadband access and must be changed to realize our national policy objectives.

4.2 Disincentive Effects of Regulated Rates for Interconnection, Resale & UNE’s

As a theoretical proposition, setting prices of unbundled network elements (UNEs) at TELRIC can facilitate entry and promote investment in facilities-based competition. As a practical matter, it has done anything but that. The predominant use of TELRIC has NOT been in the pricing of UNEs, but in the pricing of UNE-P, which has nothing to do with unbundling and everything to do with providing a wholesale price arbitrage opportunity for entrants. Consequently, UNE-P has become a major impediment to infrastructure investment and facilities-based competition.

As applied by state commissions, TELRIC costs have been systematically under-estimated (see 4.3.), so UNE prices are typically well below true economic costs. The problem has been exacerbated by numerous “compromises” in which ILEC’s “voluntarily” lower UNE prices to gain regulatory approval on unrelated matters (e.g., merger or §271 approval). Moreover, because some states have set UNE prices even further below costs than others, there is a growing tendency to hold up the lowest UNE prices in an ILEC region as the standard for UNE prices in other states, which only spreads and increases the harm of poor regulatory decisions.

Thus, the financial evidence indicates that UNE prices are below cost, in fact, “UNE prices are at a deep discount to Regional Bell’s costs, as reflected on their financial statements.”[44] If the trend toward lower UNE prices, and more extensive unbundling requirements continues (e.g., DSL unbundling), the harm will grow exponentially: ILEC’s will not be able to tolerate the much larger losses (due to UNE prices below costs) if the quantity purchased increases substantially.[45]

If ILEC losses due to higher “sales” of UNE-P at prices below costs, that will assuredly reduce their incentives and ability to attract capital to invest in network upgrades, including broadband. Moreover, pricing UNE-P below costs reduces incentives for all infrastructure owners to invest, by setting an artificially low “cost” for non-facilities based competitors. An MSO considering investments in plant upgrades to provide cable telephony faces competition from a CLEC or reseller using UNE-P, which reduces expected revenues and therefore makes the investment that much less likely.

4.3 Disincentive Effects of TELRIC

To the extent that TELRIC provides an accurate estimate of the actual economic cost of building a network, and to the extent that TELRIC-based prices provide for recovery of ACTUAL costs, TELRIC is a useful tool for establishing UNE prices. In many jurisdictions, though, TELRIC has not been implemented in a way that fully compensates ILEC’s for their costs. TELRIC estimates are based on complex cost models with a large number of assumptions and inputs. Unrealistic and inconsistent assumptions and inputs have resulted in unrealistically low TELRIC estimates.

There is also a fundamental flaw in the application of TELRIC costs in determining UNE prices (in addition to the biases below). Even though the TELRIC cost models adopted by most states use excessively long depreciation periods, there is typically no requirement that competitors make commitments on the duration of their UNE purchases. So, an ILEC may have to make very long-term investment commitments to provide UNE’s to CLEC’s, but the CLEC’s can buy those UNE’s for a short period of time, then switch over to their own facilities (or lease facilities from another CLEC), stranding the ILEC’s investment.

But the biggest problem with TELRIC pricing is that, even if it is conceptually sound for pricing network elements, it is not being used mainly for that purpose: its main application is in the pricing of network services—UNE-P—for which it is not intended and for which it is conceptually wrong. The Telecom Act provided two different pricing mechanisms for good reason: a resale discount is the appropriate method for pricing services; correctly estimated TELRIC is correct for pricing elements.

“UNEP is physically similar to resale. In each case, the CLEC uses the ILEC network to provide service to the end-user and essentially limits its own functions to marketing, inputting the order into the ILEC’s systems, and billing.”[46]

“UNEP can be more economic, where the customer’s retail bill is high enough. Thus, CLEC’s have generally preferred UNEP to resale as an entry mechanism, where they have felt entry was economic at all. But they have generally limited themselves to targeting states in which UNEP prices are low and then cherry-picking customers within those states.”[47]

Not surprisingly, local competitors are now arguing that state commissions should mandate unbundling even where the FCC does not. In Texas, for example, CLEC’s and resellers have petitioned the PUC to mandate unbundling of local switching in major metropolitan areas, even though the FCC has found that it is not required. It is ironic that competitors seek “unbundling” when they are not even buying unbundled switching. Rather, they seek to maintain the existing price arbitrage opportunity, of having both a resale discount and a UNE-P wholesale price available.

In addition, those same applicants are attempting to ratchet down the UNE-P price by recalculating TELRIC, based on the premise that the costs of “best available technology” have decreased since the currently used TELRIC costs were estimated. But it is completely inappropriate to periodically reapply TELRIC as they request. As estimated in Texas and every other jurisdiction, TELRIC is based on the unrealistic assumption that the entire incumbent network is replaced with a single-vintage of best available technology. Reapplying TELRIC every few years is directly at odds with that assumption and the long depreciation lives used in previous TELRIC estimates.

Because telecom is a network industry characterized by large-scale durable assets and rapid technological change, re-applying TELRIC periodically would put TELRIC on a declining cost trajectory that is not achievable, chilling investments from all providers. That downward spiral would have a disastrous effect: “If [there were] radical reductions in the price of UNE-P, two things would happen. CLEC’s would find UNE-P entry economic and would begin to enter the market very actively. The RBOC’s, in turn, would quickly become uneconomic, as they would be forced to serve customers at prices that are at an 80%-90% discount from the cost on their financial books.”[48]

It would be even more inappropriate to apply TELRIC to new investments used to provide new network capability, such as broadband. By its nature, unbundling reduces incentives for investment, but that disincentive effect is increased exponentially when rapid technological change can cause early technological obsolescence.[49] Consider the effect of requiring Intel to unbundle its manufacturing plants and price those unbundled elements at TELRIC. Even worse, imagine requiring Intel to sell its Pentium 4 chips to its competitors at downward-biased TELRIC prices—which is the correct analogy to UNE-P pricing of DSL. Can anyone imagine that Intel would continue to spend such a large share of its revenues on R&D, or make even riskier investments in new semiconductor manufacturing facilities? Of course not.

4.4 Disincentive Effects of Uncertainty of Investment Returns

As a matter of economic principles and empirical observation, there can be no doubt that increasing the risks and uncertainties associated with investments decreases incentives to invest. This is especially true of large-scale investments in durable assets, such as investments to extend DSL capabilities into wireline networks.

Even without required unbundling, there is a great deal of risk associated with the substantial investments required to extend and enhance broadband availability (estimated at over $10 billion[50]). These risks stem from both the supply and demand side of the business. For example, on the supply side, ILEC’s face challenges in conditioning lines, deploying equipment in outside plant, and managing customer acquisition costs.[51] On the demand side, ILEC’s face risks associated with customer take-rates, customer churn and price stability. These “normal” risks of providing broadband service are reflected in the fact that at approximately 30% of broadband subscribers, DSL is significantly behind cable modem service in market penetration.

Adding regulatory requirements that increase the cost for the incumbent and/or artificially reduce the cost to competitors will dampen ILEC investment in DSL facilities.

§	Title	Responds to NTIA Questions:	Page
1	Toward a National Broadband Policy		1
2 2.1 2.2 2.3 2.4	Defining Broadband Access Broadband and Digital Convergence Broadband and Wireless Networks Next Generation Broadband Implications of Broadband Definition	B & J	3
3 3.1 3.2 3.3 3.4 3.5	Primary Policy Goals & Objectives Promoting Intermodal Competition Technology Neutrality Policies Facilities-Based Competition Widespread Deployment of Broadband Access Eliminating Regulatory Obstacles	A & D	10
4 4.1 4.2 4.3 4.4 4.5	Disincentives for Investment in Broadband Access Need for Investment in Broadband Access Disincentive Effects of Regulated Rates for Interconnection, Resale and UNE’s Disincentive Effects of TELRIC Prices Disincentive Effects of Investment Returns Disincentive Effects of Retail Price Regulations	C, E & G E.1. E.4 E.2 & 3	15
5	Regulatory Policies for Broadband Access & Services	F, K & N	24
6 6.1 6.2	Other Public Policies to Promote Broadband Access Tax Policies Right-of-Way Policies	L & M	28