Responses to Questions Posed by the US Department of Commerce's NTIA for Comments on Broadband Policy


Responses by Dr. Robert B. Cohen,

Visiting Fellow, Economic Strategy Institute, Washington, DC.

Dr. Cohen has studied on the impact of broadband and the Internet on the US economy, including an early analysis of the impact of broadband that was done for Senators Burns and Gore. He has prepared several forecasts of the growth of the Internet for private firms and government. Dr. Cohen is also a past president of the Forecasters Club of New York and the President of Cohen Communications Group in New York.

The opinions expressed here do not necessarily reflect those of the Economic Strategy Institute.


A. What should be the primary policy considerations in formulating broadband policy for the country? Please discuss the relative importance of the following: access for all; facilities-based competition; minimal regulation; technological neutrality; intra-modal competition; inter-modal competition; and any other policy consideration.


The key consideration for broadband policy should be how to facilitate broadband deployment and taken into account some major transformations in the telecommunications and equipment industries as a result of the "dot-com recession."


Investors put $200 billion into new telecom entities during the 1997-2000 period and much of this investment has imploded. The picture for the next few years seems unclear. Capital spending is being cut and several large service providers, including Global Crossing, still risk bankruptcy. Sales of equipment vendors are down dramatically, with Lucent expecting to sell about 25% of what it did last year.


In this environment, getting the industry restarted or rejuvenated should be the primary policy goal. Telecom services and equipment sales fueled the great expansion of the 1990s and have had a key impact during the downturn. Regulatory barriers are one of the issues that should be addressed but the key question is creating an environment with incentives. This certainly suggests that the old policies may need to be jettisoned and new ones are needed.


What might be some approaches:


1.      Previous efforts have focused on helping the new entrants and promoting competition. The ways of estimating whether competition was increasing were cumbersome and difficult to describe with readily available data. To stimulate new investment, it may be easiest to measure the end result -- is there more competition? Setting certain goals for competition and service deployment might do this. If these broader goals were not met, bigger firms would need to compensate new competitors each year by contributing to a deployment fund, called the "Broadband Infrastructure Fund." New competitors could use this fund to offset the costs of interconnection or to deploy bypass networks. Regulators could use market share in specific service markets -- local service, long distance to see if the incumbents had not done enough for the market to remain competitive.

2.      There has been little done to stimulate new investment. This might be addressed by shortening the time for telecom firms to write off new investments to the present year. This would reflect the major changes in technology and equipment. In addition, when large service providers need to write off an older legacy network or parts of a legacy network, they should be given more incentives to do so as rapidly as possible. Many of these service providers face a major dilemma in moving to all-optical networks and coming up with the large investments to do so. These networks would greatly enhance service providers' abilities to offer new services and bring dramatic innovations to customers. If DOC supported some additional type of capital write-off of legacy networks, it would not only enhance the performance of existing service providers but also help underpin the growth of equipment vendors.

3.      To improve the expansion of broadband into rural areas, a policy that linked economic development to infrastructure deployment might be helpful. For instance, many states have developed fiber loops that pass through or near many areas that need more economic development. If these fiber backbones could be linked into commercial networks for a limited period of time, they might enhance the ability of these rural areas to attract new businesses. In many states, including Iowa and New York, these fiber backbones were built for state purposes, but releasing them from regulatory purgatory might let states put them up for commercial use or develop a way to have commercial service providers bargain for their use. If service providers could offset some of the funds they might pay into the "Broadband Infrastructure Fund" by investing in upgrading or enhancing these state-based fiber loops, that might provide one incentive for growth.


B. How should broadband services be defined? Please discuss (1) what criteria should be used to determine whether a facility or service has sufficient transmission capacity to be classified as "broadband;" (2) how the definition should evolve over time; and (3) the policy implications of how the term is defined.


At the present time, there is no clear definition of broadband by the US government. The FCC has at times stated that 128 kbps and faster is a good definition of broadband. Industry observers, including the head of Intel, Craig Barrett, noted that a better definition is faster than 100-120 Mbps. That is more than a 1000 times disagreement!


If DOC wants to set a broadband definition, it should probably identify both a speed that truly brings broadband applications to the end user. For companies, this probably means a 100 Mbps speed or greater. For consumers, there may be a need to establish an initial goal that is more limited but that pushes the ability to gain more substantial applications. If consumers are using video, they probably want to have 5 Mbps or more as a goal for the next year to two. As newer networks are deployed, this goal should probably be scaled up.


The key point may not be the creation of facilities or services with adequate capacity, but the shift over to new networks that can deliver this capacity. The dilemma for service providers is how to create high-speed networks with a viable business plan. The answer to this question is likely to remain unclear until more firms begin to deploy such networks and offer new services to businesses and consumers. If policies promote the shift to high speed networks, that would be a big plus. We can already see from the university community that many universities are reaching the point where they need 1000 Mbps (1 Gbps) links to the Internet because of substantial video use by students. Corporations have links of this size already and lower prices will help move them along. But the issue of helping service providers get to the "networks of the future" looms large.


US policy should set broadband in larger terms. To retain the old FCC definition is no longer realistic. Some achievable level of access speed or backbone speed for networks might be tracked by sampling a few networks and larger businesses. An achievable level for many consumers, using an evolving goal, might also be set. But the level should not be set so low that it is meaningless. Canada and other nations (South Korea) are far ahead of the US in providing broadband access. We should not continue to fall behind in this area.


C. Several studies indicate that the rate of deployment of broadband services is equal to or greater than the deployment rates for other technologies. What is the current status of (1) supply and (2) demand of broadband services in the United States? When addressing supply, please discuss current deployment rates and any regulatory policies impeding supply. When addressing demand, please discuss both actual take rates and any evidence of unserved demand. Please also address potential underlying causes of low subscribership rates, such as current economic conditions, price, cost-structure, impediments to the development of broadband content, or any other factor. To what extent has the growth in competition for broadband and other services been slowed by the existing rates and rate structures for regulated telecommunications services?


On the rate of broadband deployment, yes it is ahead of deployment rates for other technologies. But several ILECs/RBOCs have recently announced cutbacks in their deployment efforts. This will lower the deployment rates in the future. In part, these cutbacks reflect a low level of demand among consumers and the higher monthly subscription prices for DSL. While the US would not want to directly subsidize service providers to deploy broadband, as is the policy in Canada, there may be some steps that would help.


I believe that the economic slowdown will slow the deployment of broadband for two reasons. First, consumers will hesitate to move to broadband because the cost is still about double the cost of subscribing to an established ISP (AOL, AT&T, etc.) and far more than a free Internet email connection. Second, service providers are finding that they don't make enough money on broadband connections to warrant a bigger push to get consumers to adopt it. The market will probably take longer to expand than many observers had hoped.


I am not sure that the rates and rate structures for regulated service have been the key factors acting as a drag on broadband deployment. Rather, I think the lack of business and consumer applications has fostered a slow growth of demand.


Certainly, lower prices for broadband would speed the uptake, but this would mean a shift in technologies in the core network (providing higher speed overall and easier management of the networks) or a change in technologies that would improve the delivery of broadband. Research will bring us these new technologies in the next few years, but in my opinion the main barrier will remain how rapidly service providers can shift away from legacy networks and establish seamless high-speed optical networks. Speeding up this process requires enhancing the ability of all service providers to write off these networks and to invest in new ones. 


An additional dilemma is that cable providers/MSOs may become the main providers of broadband to the home in the next few years. The current US regulatory structure does not deal with this type of development. The government should begin to address how existing policy might be modified for ILECs, CLECs, and other entities, if this trend emerges. In addition, there is a good chance that with new investments by Vivendi, foreign service providers could become a more significant force in broadband delivery. US regulations have limited foreign participation in our media and services sector, but the environment will change irrespective of these barriers. US regulations may need to be reconsidered in light of these developments.


D. Should government adopt as a goal "access for all" to broadband service? What would be the costs of such a goal? What policy initiatives, if any, should be considered to achieve that goal? Are there areas or persons that are unlikely to be served through marketplace forces?


Yes, this should be a long-term goal. But there could be a series of steps that address demonstrating the steps that are being taken to achieve this goal and what help the government can offer local service providers to reach them.


Given that the marketplace may not provide solutions that promote "access for all," there is probably a need to recognize this and to establish a series of policy incentives to help achieve this goal. I mentioned one possible incentive -- offsetting contributions to a "Broadband Infrastructure Fund" by investments in rural broadband deployment. Others might include a more rapid write-off of capital expenditures in rural areas and incentives to deploy more advanced applications sooner in rural areas. 


E. Do the interconnection, unbundling, and resale requirements of the Telecommunications Act of 1996 reduce incumbent local exchange carriers' (ILECs') incentives to invest in broadband facilities and services?


1. Are their (there sp?) investment disincentives attributable to the regulated rates for interconnection, unbundled network elements, and resold services?


Yes, these regulations have reduced ILECs' incentives to invest and have created an environment where it is cumbersome to compute the proper costs without calling upon a wide range of economists who usually have widely differing opinions about how to evaluate the costs and prices of network elements and services.


On the other hand, ILECs have also slowed the pace of investment to see how much new competition interconnection would create. This has resulted in a kind of waiting game, with businesses and customers not obtaining the kind of network services and access speeds they might prefer.


2. To what extent are those disincentives due to ILECs' uncertainties about their ability to recover the added network costs needed to accommodate potential requests from competitors? What are the magnitude of those additional costs? What mechanisms could be used to share the risks of those costs efficiently and equitably among ILECs, competitors, or users?


I don't believe that the additional costs to accommodate competitors' requests are sizable. I think that the ILECs are more concerned about the competition that easier access to the network will create. I don't believe that sharing the risks of the new costs is a major stumbling block to progress. Now that many CLECs have disappeared or are financially weaker than they were last year, the ILECs have little to worry about in most parts of the nation.


3. To what extent are the returns on ILECs' investments in new infrastructure uncertain? Is the uncertainty of gaining an adequate return on each infrastructure improvement (attributable in part to other firms' ability to use those facilities to offer competing services) significant enough to deter investment?


I don't think this is a major issue in the current environment. Certainly, ILECs are unsure about returns on many new infrastructure investments. But this is because new applications that run on the infrastructure don't have well defined markets. In addition, the ILECs and the rest of the telecom industry is in a significant downturn and capital investments will not resume in earnest until the economy revives.


4. What are the principal strengths and weaknesses of the FCC's total element long run incremental cost (TELRIC)(1) methodology? What changes could be made to render TELRIC an effective deterrent to the exercise of market power and conducive to efficient infrastructure investment? Would it be possible to construct an alternative methodology that would not depend on cost information controlled by regulated firms?


I think the TELRIC should be scrapped. It has not worked well. It is cumbersome to measure. It is not clear to the public what it will achieve.


More objective and easily measured yardsticks for competition in service markets should replace TELRIC. These yardsticks should use data that is easy to obtain.


The main goal, as I have stressed above, would be to stimulate new investment and to facilitate the move to new network technologies (all optical networks). My opinion is that TELRIC has not helped spur new investment. In addition, the main new investment should be in replacing older legacy networks. TELRIC focused on access points rather than network investments. This was a major shortcoming. You can't focus on access points and their cost without getting to the economics of the underlying networks themselves. This is perhaps the greatest shortcoming of the TELRIC effort, because it focused on network unbundling while the real issue was network upgrading.


F. Some have suggested that a regulatory dividing line should be drawn between legacy "non-broadband" facilities and/or services and new "broadband" facilities and/or services. Is this a feasible approach? If so, how would it work?


I don't agree with the line of reasoning here. I think that there should possibly be a line between legacy facilities and new broadband ones if there could be a policy to help get rid of the legacy facilities. My opinion is that new technologies and new innovations are pushing service providers of all types to get rid of legacy networks and facilities. If a policy could support this effort, then it might prove helpful. But merely to discriminate between the two when a large economic force is pressing service providers to move along to new networks doesn't seem to be very constructive.  


1. What effects would changes in the regulatory structure for broadband services and facilities have on regulation and competition with respect to voice telephone and other non-broadband services?


2. If ILECs deploy broadband services using a mixture of new and old facilities, will competitors be able to use the older shared facilities that they previously had access to?


3. If ILECs deploy broadband facilities to replace portions of their existing copper plant, will the displaced copper plant give competitors a viable opportunity to offer alternative services? What would be the annual costs to the ILEC (or to a purchaser of the displaced copper plant) of a continuing obligation to maintain that plant?


I think this assumes that the costs of older plant would be lower than the costs of new plant so that a purchaser of a displaced copper plant could compete with a newer technology network. I don't believe that the economics would move this way. It would be more likely that competitors would find their costs of maintaining legacy plant is too high. Of course, this could create problems for businesses and consumers that want to remain on the legacy network.


4. What regulations, if any, should apply to new broadband facilities and/or services to ensure a competitive marketplace?


The telecom industry is facing a major investment challenge. I don't believe that applying regulations that have been followed in the past to broadband facilities will insure a competitive marketplace.


We most likely need a new set of regulations that provide a more open environment for service providers to make the investments in new technologies. That should be the key goal of any new regulatory efforts.


Given experiences overseas, the history of broadband deployment does not appear to have resulted in any approaches that facilitate the kind of rapid investment that is needed. In addition, the fact that new players from beyond the traditional regulatory framework will very likely be major service providers in a broadband environment creates challenges for establishing the regulatory framework. I don't believe it is easy to address this question at the present time. We need more experience with broadband networks to help develop this framework.  


G. To what extent have competitive firms deployed their own (a) transport, (b) switching, and (c) loop facilities? Are those investments limited to particular areas of the country or to particular portions of communities and metropolitan areas? What market characteristics must exist for competitors to make facilities-based investments? Do competitors have the ability to deploy their facilities in ways that minimize costs and facilitate efficient network design?


There are a number of competitive regional service providers that have deployed their own transport, switching and loop facilities. These are usually regionally focused with networks that serve business customers. While I have not spoken to some of these carriers for about six months, they are probably facing much of the financial distress that has affected the rest of the industry and have not been able to expand services or infrastructure as much as they had planned. They have tried to deploy facilities that minimize costs and facilitate efficient network design. But I am not sure that they will survive in the current economic climate.


An interesting effort by the US government would be to review the status and challenges facing some of these regional service providers. Many of them deployed the most advanced networks that they could find. It would be useful to know if many have been able to retain their business customers and survive during the current downturn.


H. What cable companies are currently conducting trials to evaluate giving multiple Internet service providers access to broadband cable modem services? Describe the terms and conditions of ISP access in such trials. What technical, administrative, and operational considerations must be addressed to accommodate multiple ISP access? How can cable firms manage the increased traffic load on their shared distribution systems caused by multiple ISPs?


I do not know much about the current trials to provide multiple ISPs with access to broadband cable modem services. I believe that this kind of peering with cable networks could be fraught with the same type of peering (or interconnection) problems that are obvious in the rest of the Internet. To manage these types of connections well, there would need to be a way to manage traffic flows through the points where such peering takes place so there is not service disruption or congestion. This takes a fair amount of engineering skill and investment in sophisticated equipment. Because cable companies have not been significant ISPs in the recent past, I have not examined how they are managing this type of peering. I have studied peering among the major ISPs in some detail.


I. What problems have companies experienced in deploying broadband services via wireless and satellite? What regulatory changes would facilitate further growth in such services? Is available spectrum adequate or inadequate? What additional spectrum allocations, if any, are needed?


There are major problems due to the lack of adequate capacity in the backbone networks of wireless companies. I believe that not very many companies will be moving to deploy broadband services because of the sheer cost of upgrading networks and developing good applications. Regulatory changes that spurred the new investment would be helpful. Spectrum to offer real broadband applications is short and not easy to assemble for all parts of the US market.


Part of the problem is that technologies to provide broadband services are not available. The recent efforts by NTT's DoCoMo to provide handsets for broadband applications in Japan has shown how early we are in the development of this market. It is likely to take longer to get such Third Generation (3G) systems in place. Additional spectrum allocations for wireless companies would help.


J. How should the broadband product market be defined? What policy initiatives would best promote intra-modal and inter-modal broadband competition?




K. Would it be appropriate to establish a single regulatory regime for all broadband services? Are there differences in particular broadband network architectures (e.g., differences between cable television networks and traditional telephone networks) that warrant regulatory differences? What would be the essential elements of a unified broadband regulatory regime?


It would probably be appropriate to establish a single regulatory regime for broadband services, but this would bring television broadcasts under the new authority if they were carried on broadband networks. The convergence of networks is undermining what we previously assumed to be well formulated regulatory regimes.


There probably will not be significant differences between broadband network architectures, but the applications they are providing could differ widely so that the distribution end-points may look quite different on a largely broadcast network and one that is more interactive (carrying two-way Internet communications). There is also a significant difference because cable networks are more outside traditional regulatory regimes so it may be easier for them to develop advanced broadband networks than the ILECs or CLECs. This would mean that cable entities act as the main innovative group for broadband communications and meld some of the old legacy telecom networks into their spheres.


I think that the government should assess what is needed in a unified regulatory regime. I believe that because of the history of the US telecom and cable industries, one may have an easier time deploying broadband and that the two industries will consolidate, perhaps rapidly, into a single communications service providing industry. I don't believe that establishing a unified regulatory regime in this context offers much for any of the players. A transitional unified regime might make some sense, but it should not impose additional barriers to innovation and investment.


Some of the elements in a transitional unified regime might be to consider how to insure some pattern of competition, but the yardsticks for measuring competition may need to be changed. The future could easily have an environment where three or four players offer broadband voice, data and video services in the US. This could easily end up being one or two players. In this context, what the US policy should be is a very different question than it was when there were 6 or more ILECs/RBOCs and several hundred CLECs. In fact, CLECs would probably disappear and many ILECs would consolidate into a smaller group of service providers. Cable operators might buy RBOCs. In addition, other service providers (Microsoft, AOL) might influence how networks connect to consumers and businesses.  


L. Are there local issues affecting broadband deployment that should be addressed by federal policies? Please provide specific information or examples regarding these problems. Should fees for rights of way and street access reflect costs in addition to the direct administrative costs to the municipalities affected? To what extent do state laws and regulations limit municipalities' ability to establish nondiscriminatory charges for carriers' use of public rights-of-way? Please discuss the most appropriate relationship between federal, state, and local governments to ensure minimal regulation while removing disincentives or barriers to broadband deployment.


I think that many local issues have acted as a major issue to deploying new networks and innovative technologies. A federal review of these issues and the development of an approach that would deal with state laws that affect municipal treatment of carriers would be positive developments.


While I believe it would be helpful to have a more appropriate relationship between federal, state and local governments to help promote broadband deployment, I have not studied this area in detail and cannot comment on this aspect of the question.


M. Are there impediments to federal lands and buildings that thwart broadband deployment? Please provide specific data. What changes, if any, may be necessary to give service providers greater access to federal property?


I am not enough of an expert in this area to comment on these questions.


N. With respect to any proposed regulatory changes suggested in response to the above questions, can those changes be made under existing authority or is legislation required?


I believe that a number of the new regulatory changes I have suggested would require new legislation. It is possible that some could be done under existing authority.