Funding the Public
Telecommunications
Infrastructure

Historically, the social policy goal of universal service has been served through a system of massive intraindustry transfers of funds, with the primary flows being from business and toll service users to basic residential subscribers, and from urban to rural exchanges. Before divestiture, these transfers were accomplished largely within the corporate umbrella of AT&T. Since then they have been maintained by a complex, gerrymandered set of rules governing transfers among carriers. This system, rooted more in political compromise than economic logic, was sustainable when there was a single monopoly network. However, competition is the natural enemy of cross-subsidy and the emergence and continued growth of competitive suppliers makes it doubtful that current mechanisms for managing these transfers can be maintained much longer.

The growth of competitive suppliers means that, for policy purposes, the working definition of the telecommunications infrastructure-which in the past has largely reflected concerns with monopoly providers of common carrier services-must be expanded to incorporate the facilities and functionalities of new, increasingly diverse and competitive industry players. Most components of the traditional public telephone network either are, or soon will be, competitively supplied. For example, terminal equipment, toll services, and business services are all available from increasingly sophisticated competitive providers, and competition in residential services appears imminent.

The definition of infrastructure should further be refined to distinguish between what might be called the "core network" and "on/off ramps." For the most part, the core network is the interoffice portion of the infrastructure. By and large, new and advanced services reflect and depend on the increasingly sophisticated hardware and software components of the core network. Access lines are the on/off ramps to the core network and the two meet at points of interconnection, which traditionally have been end office switches. However, as fiber and other broadband media continue to make inroads into neighborhoods and business districts, and as existing residential broadband media such as cable television begin to offer switched services, the points of interconnection will shift as well. For the most part policy concerns with network functionality relate to the competency of the core network, while access concerns relate to the provisioning of residential on/off ramps.1

The task of listing goals and objectives that should be served by the telecommunications infrastructure is complicated by the fact that different people and groups look for different things in a telecommunications infrastructure. Nevertheless, there are a number of high-level goals on which there is broad agreement, and these are useful for focusing our analysis of the current funding systems and the alternative we propose. At a minimum, the infrastructure should serve the following set of policy goals. As the discussion shows, however, there is considerable room for disagreement over what level of public commitment they imply.

Universal service. Basic telecommunications services should be both universally available and (nearly) universally subscribed to. Universal service is a long-standing goal of U.S. telecommunications policy and is supported by both equity and efficiency arguments. Universal service serves equity objectives because without the availability of at least a certain minimum level of telecommunications services effective participation in modern society is not possible. It might also be argued that policies promoting universal service serve efficiency as well, because while the value of the network to all of its users increases with the level of societal participation, individual decisions to participate reflect personal benefits only.

Participatory fairness. Long ago we passed the point at which access to telecommunications services became absolutely necessary to effective participation in the economy and in society more generally. As more advanced services become widely used, the level of services necessary to effective participation will advance as well. Of course, the quality of service that should be universally supplied is a controversial issue that will have to be resolved through the political process.

Contributions to economic productivity. Ongoing advances in basic communication technologies are creating new services with the potential to improve productivity in other sectors of the economy, as well as offer direct value to consumers. Further economic gains are anticipated from the use of the telecommunications infrastructure to deliver new and current public services.

It is important that the infrastructure incorporates these technologies as it evolves and that it makes generally available the services the technologies make possible. Strong network externalities in telecommunications services imply that investment driven by market incentives alone is unlikely to push technology and infrastructure capabilities ahead as rapidly as society desires.

Reliability. It is important that public network services, however sophisticated, be reliably provided.

Participatory government. A widely accessible, advanced public network infrastructure could make more direct connections between the electorate and government possible. Distance would no longer affect the ease and convenience of participating in government affairs. We are already witnessing the primitive beginnings of this process in the use of computer bulletin boards by the Clinton Administration.

New public services and electronic entitlements. Advancing technological capabilities are almost certain to make possible the electronic provision of new public services that currently are beyond our capabilities. Possibilities include public bulletin boards and database access, public services directories, medical advisory services, various types of stand alone library and educational services, as well as advanced technologies for classroom instruction. Along with these capabilities we are likely to see the emergence of a political demand that they be made available as publicly supported resources analogous to public parks and libraries.

While the properties of an infrastructure funding mechanism cannot be divorced entirely from the objectives the infrastructure is to serve, it is still useful to list properties of a desirable funding mechanism. We begin with the following list.

Contributory and distributive fairness. The current system has a large number of internal transfers, both among industry segments and among groups of consumers. Political viability requires that recipients of the transfers be truly deserving and that the levies required to maintain these transfers be fairly distributed. A funding system will lose political support if there is a widespread belief that some are unfairly profiting at the expense of others or that some are unfairly able to avoid contributory obligations that then fall more heavily on others.

Political sustainability. Any new infrastructure funding mechanism must represent a politically stable coalition of interests. While not sufficient by themselves, contributory and distributive fairness are both necessary to political sustainability.

Promotion of economic efficiency. Both static and dynamic efficiency concerns must be addressed. To promote static efficiency-the efficiency with which current services are provided-it is important that the infrastructure funding mechanism not create purely financial incentives to engage in activities that function solely to avoid participating in intrasystem transfers. To serve the goal of dynamic efficiency, it is critical that the funding mechanism not discourage investments in and experiments with new technologies and services. Equally important is that funds in the promotion of new technologies and services be deployed to compensate for deficiencies in the market process.

Viewed from the national level, the current infrastructure has been financed and continues to be serviced largely from U.S. telecommunications industry revenues.2 Viewed from the levels of its various parts, however, we see that this apparent self-sufficiency at the national level masks a complex set of intraindustry transfers sustained through regulation and law. Focusing on within-industry transfers, we see that there are three primary types of support mechanisms. (1) Funds transferred from toll services to local services, both at the federal and state levels.3 (2) Transfers from low-cost local exchanges to high-cost local exchanges, which are primarily transfers from more densely populated urban exchanges to less densely populated rural exchanges. (3) Differential allocation of local service costs through rates charged to different customer classes (for example, higher local service rates for business than residential customers). Less obvious is rate averaging in which high-cost customers and low-cost customers are charged a common rate. Also included in this category are social support services such as lifeline programs that provide service to low income customers at below marginal cost rates, 911 emergency services, and services for the hearing impaired, the costs of which are commonly covered through charges on local phone bills.

In combination, these mechanisms for reallocating resources within the industry have worked tolerably well to ensure universal service up to now. Given the record of stagnant or declining investment in the public network, however, it is hard to argue that the funding system has worked well to promote investment in advanced network capabilities. While there may be considerable debate on this point in the academic literature, the need for more infrastructure investment and more widely available advanced services is increasingly popular on the political agenda, as evidenced by the many network infrastructure initiatives pending in the states and at the federal level. In any case, it is clear that the current system is beginning to crumble under the pressure of the competitive forces unleashed by technological advances and open-entry regulatory policies.

For example, the growth of competitive access providers (CAPs) has made it possible for large toll customers to directly connect to toll service providers to avoid paying the usage-based contributions to within-system transfers that traditionally have been collected through the fees charged by local exchange carriers (LECs) for access services. As long as only LECs and their access customers have to make mandatory contributions to maintaining universal service obligations, true competition on the merits of access services and its attendant efficiencies is not achievable. Furthermore, it is becoming increasingly clear that the inequities inherent in the current system of access charges is undermining the sense of political legitimacy that sustains at least this component of intraindustry transfers.

Similarly, as genuine local exchange competition develops, regulated LECs in the larger metropolitan areas will find it increasingly difficult to maintain their contributions to high cost areas, or, in the longer run, to support the (free) default capacity relied on for system back-up by competitive network providers. Finally, in both large and small markets, the locally averaged rates and pricing of social services below cost put in place through the political regulatory process are likely to become increasingly unsustainable, as alternative service providers pick off the customers with the largest price-cost differentials. This is already happening in the case of business and toll-access services and is likely to become an issue in residential local services as well in the not too distant future. Two-way cable television networks and new wireless Personal Communication Networks (PCNs), both of which are relatively inexpensive to provide, are on the horizon. In both large and small exchanges, this type of competition will make it increasingly difficult for incumbent LECs to continue to provide public services such as Lifeline and emergency services below cost. In smaller markets, it threatens the viability of local exchange carriers left to serve a dwindling number of high-cost customers. Clearly, given current trends, this system of internal transfers is not sustainable for long. But a return to the constraints and rigidities that held this system together in the past is neither conceivable nor desirable.

As an important component of an alternative funding mechanism, we are proposing Value-Added Service Surcharges (VASS) to either augment or totally replace funds generated through the current set of levies that transfer funds from toll services to local services, and from low-cost exchanges to high-cost exchanges. As a public policy tool, the VASS mechanism could serve to augment current transfer mechanisms if policy makers wished to speed up infrastructure development and investment, or it could eventually replace the current mechanisms entirely. The VASS is a sustainable, equitable, and relatively efficient mechanism for funding intraindustry transfers, even in the face of widespread competitive entry. The VASS proposal is described in the next section.

It is important to note that while VASS differs from the current system in many respects, it does not break with the U.S. tradition of relying on intraindustry transfers funded by levies on services. Thus it differs from the Japanese approach which draws on general tax revenues to fund the support of certain services and customer classes. While support from general revenues has attractive efficiency properties, a shift to this type of system in the U.S. is not politically feasible in the foreseeable future.

To the extent that social distributional and infrastructure objectives currently met through local rate averaging cannot be satisfied with VASS-generated disbursements, we recommend that state or local authorities pursue one or both of the following approaches for adjusting local rates: (1) Efficient Component Pricing (ECP) principles applied to LEC bottleneck monopoly services, as recently articulated by Baumol and Sidak.4 (2) Make contributions to local rate averaging or infrastructure/service objectives specific to certain classes of customers, regardless of who serves them. ECP and class-specific customer obligations are described below.

Simply described, VASS is a set of surcharges calculated as a fixed percentage of gross revenues that potentially could be collected from all businesses selling value-added telecommunications services. Value-added services include the services of all service providers interconnecting with the public switched network, with the exception of local loop services provided to residences by state certified common carriers with provider of last resort obligations.5 As the benefited service, the local residential loop or primary access line is exempted from VASS surcharges. The size of the fixed percentage VASS charge would be adjusted annually to reflect anticipated needs for the coming year and projected revenues for the contributing services. Judged by the criteria set out above for evaluating funding mechanisms, VASS has a number of important advantages over the current system of fixed charges per minute-of-use (MOU) on toll services, the most important being greater economic efficiency and contributory fairness.

To see the nature of the efficiency advantages of a value-added surcharge over fixed MOU levies, consider the effects of a value-added surcharge and a fixed-per-unit levy, both collected on the same good. Both the fixed-per-unit levy and the value-added surcharge would affect economic decisions by altering marginal revenue-output relationships as perceived by sellers. The effect in both cases is to raise price and to reduce sales and consumer surplus. However, it is straightforward to show that a value-added surcharge or tax (VAT) has a smaller effect on marginal revenue and price than a revenue-equivalent fixed per unit tax. For a seller, a fixed-per-unit levy is equivalent to a downward shift in its demand and marginal revenue schedules by the full amount of the tax. By contrast, a fixed percentage value-added surcharge shifts a seller's demand and marginal revenue schedules downward by equal percentages. As long as demand curves are downward sloping, marginal revenue must be less than price, so that a value-added tax reduces marginal revenue by less than the amount of the surcharge extracted from the sales price. Because a value-added surcharge reduces a seller's marginal revenue by less than a revenue-equivalent fixed-per-unit levy, less output is sacrificed and more surplus remains in the hands of consumers.6 Put another way, for equivalent reductions in consumer surplus due to induced price increases, a value-added surcharge generates more revenue than a fixed-per-unit levy.

As an effective tax on output, a VASS distorts economic activity much less relative to the current system which taxes productive inputs (e.g., toll- and carrier-access services). This is perhaps the most important efficiency advantage of a system of ad valorem surcharges relative to the current system of fixed per unit levies; it eliminates the incentive to inefficiently bypass LEC access facilities to avoid the MOU charges levied on LEC access services. With the same value-added surcharge applied to LEC and CAP access services, competitive advantage will be determined by economic efficiency-based cost advantages alone.

The political legitimacy of the current system as an equitable mechanism for funding telecommunications infrastructure investments is being seriously eroded by the fact that, increasingly, telecommunications winners are separated from telecommunications losers by avoiding the contributions that all are supposed to be making to infrastructure support. The main problem with the current system is that, for the most part, the largest users, who appear most able to carry the load of transfer payments, have been the most successful at avoiding them by purchasing access services from CAPs. As technology and competition expand the class of customers able to bypass LEC access facilities, the burdens of supporting universal service goals will be borne increasingly on the shoulders of small users. Continuation of this trend will make the current system simply unsustainable in the long run. If VASS is applied to all services, or to end services universally used, bypass will no longer function as a mechanism for avoiding universal service contributions.

In addition, the current system of targeting an inelastic, service-like toll as a primary source of transfer funds can be very regressive, especially for certain customer classes like the rural and poor, both of which spend proportionally much more of their disposable income on toll service than other demographic groups. Broadening the base of contributing services to include nonessential, value-added services can make the system more progressive.

The efficiency advantages of VASS as a mechanism for raising revenue for infrastructure investments and to support basic services will also contribute towards meeting national goals for economic growth. The contributory equity advantages of VASS also make it attractive in terms of meeting the public policy objectives that the telecommunications infrastructure should be serving. Since a more efficient collection mechanism reduces the opportunity cost of funds raised, the funding for more advanced infrastructure services, including the most basic level of service universally available, can and should be provided.

Finally, if VASS revenues are used to finance infrastructure improvements necessary for advanced services, exclusion of residential loop services from VASS levies ensures that the telecommunications users who benefit most are the ones who pay. This goes to the heart of a major industry controversy. As the public network is upgraded to provide more and more advanced services, some subscribers either do not want or cannot afford them. Excluding basic residential loop services from VASS levies ensures that consumers who are threatened by increasing costs-because they do not see any direct consumption benefits in enhanced services-do not have to help pay for them.

The extent to which the theoretical advantages of VASS described above are realized in practice will depend on how it is implemented. Here we briefly discuss three alternative approaches to implementation: (1) selective targeting of one or a few final (as opposed to intermediate) telecommunications services; (2) general application to all but the benefited class(es) of service, with each service provider subtracting payments to other VASS target services purchased as inputs before determining its own obligations; and, (3) general application to all but benefited services with no deductions for payments for VASS-targeted telecommunications inputs.

The selective targeting alternative is similar to Ameritech's recent proposal that transfer payments currently funded through levies on its switched carrier-access services provided to toll carriers instead be recovered by "bulk billing" those same carriers in proportion to their shares of interstate toll revenues.7 Ameritech's bulk billing proposal is a logical first step towards a full VASS system as proposed here. Ameritech's plan is equivalent to a value-added surcharge on toll services alone.

Since the size of a VASS-funded revenue pool would be determined by policy needs, as is the pool of funds collected under the current system, switching from the current system of per-minute access charges to a bulk billing arrangement results in little change in the total bill to carriers. Nevertheless, both the short and long term economic welfare gains could be substantial. As we have explained elsewhere,8 toll service is an attractive target for a VASS because it is applied to a final service rather than an input and, because toll demand is inelastic, the effect of the surcharge on price has little effect on total surplus. In addition, because it would apply to all toll services regardless of service type, it would have little if any effect on the relative prices of toll service providers. A VASS on toll services would also allow the relative prices of different services to reflect relative costs in the same proportions as would be generated by an untaxed competitive market. In the past, the per-minute charges on switched carrier-access services severely distorted relative prices resulting in uneconomic substitution towards so-called dedicated access arrangements. What's more, dedicated access arrangements, used in lieu of switched arrangements to avoid paying the policy levies on LEC access services, are generally less efficient and more expensive to provide.

While a VASS on toll services would be a good way to initiate the program, it is desirable that over time the VASS be extended to cover other service providers that interconnect to, and thereby benefit from, the Public Switched Telephone Network. Obvious candidates to bring under the umbrella of VASS would be cellular telephone network operators and electronic information service providers (e.g., Prodigy and CompuServe). Broadening the base of services that contribute to the total subsidy amount would proportionately reduce the burden on toll services providers and their customers, and would allow the relative prices of toll and other services to more faithfully reflect relative costs.

From a theoretical perspective, a generally applied VASS would be most efficient if the VASS contributions of each service provider were calculated after payments for other telecommunication services were subtracted out. Otherwise, some of the efficiency advantages of a VASS over fixed per unit charges would be sacrificed due to the effect of taxes on the prices of final services and the possibility that service providers might merge with telecommunications input suppliers to avoid the double taxation. However, no method of taxation is perfect. The advantage of a general VASS-where telecommunications input purchases are subtracted out-over a generally applied VASS-where they are not-is less clear because the procedures and bureaucracy required to manage the necessary system of accounts could be cumbersome, expensive, and generate inefficiencies of its own. Further study is required to determine which of these three approaches offers the greatest benefits.

A VASS mechanism would be significantly easier to operate than the current system's tedious and relatively expensive per-minute measurement and billing process. Another source of efficiency for the VASS mechanism is its relatively low administration, enforcement, and monitoring costs. Other consumption-based levies on end-user services might well be as efficient as a VASS, but charging carriers and not end users will greatly reduce the number of entities involved and the direct costs of administration. Furthermore, a VASS applied to business revenues is like a broad-based retail sales tax, with similar economic efficiency. By placing the VASS on revenues-which are generally publicly reported to meet the requirements of regulators and tax and securities laws-costs of monitoring, enforcement, and auditing are reduced.9

Both federal and state regulators should participate in implementing VASS nationwide. The FCC should take the lead, initially by reforming the current interstate toll to a local transfer mechanism based on toll usage charges applied to interexchange carriers' switched traffic. The model for the initial reform program should be Ameritech's "bulk billing" proposal filed on March 1, 1993, at the FCC.

Federal guidelines and basic principles describing the VASS mechanism could then be drafted and state participation in similar programs encouraged. The differing financial and information requirements likely to exist at the state level and the varying economic development goals of individual state governments require that local and state authorities be responsible for implementing VASS in their jurisdictions, in ways consistent with broad guidelines established by the FCC. The model for such federal and state cooperation has been established previously in various FCC initiatives, including the "Link-up America" lifeline assistance program. This program was set up by the FCC and implemented in different ways by individual states in accordance with federal guidelines.

VASS would be implemented through a regulatory process distinct from the normal federal and state government budget processes. It would be couched in terms of a telecommunications service tariff rate (but in percentage terms) for interconnection to the PSTN. Beginning with toll service providers, VASS could eventually be applied to the revenues of all carriers interconnecting with the PSTN, local or toll, including resellers, data networks, enhanced service providers, cellular and other wireless services, and other "value-added" network operators.

Revenues for the services subject to VASS would have to be reported by the carriers selling them. The records of LECs providing interconnection to the PSTN could be used to identify carriers with contributory responsibilities and help with the verification of their reported revenues. Annual independent audits of the billing mechanism, revenue collections, and funds distributed would be conducted and the results reported to the regulatory authorities overseeing and monitoring the VASS program. VASS monies could be disbursed by an industry revenue pool administration authority (e.g., NECA), and funds would be distributed to carriers based on the needs identified in setting the overall VASS sum to be collected.

VASS funds might first be targeted to the regulated LECs currently receiving the toll to local transfer to cover a portion of their residential access line costs. But over time the class of qualified recipients might be broadened to include any facilities-based carrier providing basic local exchange telephone service subject to the same obligations and oversight as incumbent LECs, providing requisite cost of service thresholds are satisfied.

The most difficult part of administering the VASS will be ensuring that the funds disbursed are used to expand and modernize the regulated LEC PSTN infrastructure. On the other hand, as with most potential complaints concerning implementation of the VASS mechanism, these problems already exist with the current intraindustry transfer mechanisms. While the specific application of VASS funds is probably best left to the recipient carriers, as has been the case with funds allocated in the past, regulatory oversight will be necessary to ensure that funds are in fact invested in the local network for residential access and not otherwise used to gain competitive advantage in unregulated markets.

A potential disadvantage of a VASS-type mechanism for infrastructure financing is that, if administered at the state or federal level, it may not be sufficiently responsive to variations in telecommunications policy goals among local communities. For example, a community served by a rural LEC may want to reduce residential rates charged to farm families by more than would be possible with VASS-funded disbursements alone. One solution would be to build the necessary contributions into charges for the residual monopoly services (e.g., certain switching services) of regulated LECs. Baumol and Sidak show that it is possible to grant competitors access to LEC natural monopoly facilities and still maintain the contributions embedded in charges for services provided with LEC facilities that still have natural monopoly status.10 Competition and continued contribution are both possible if competitive providers are allowed to bid for the use of LEC monopoly facilities with the stipulation that bid prices cover the sum of the cost of the monopoly element and its contribution to infrastructure maintenance or public services. Baumol and Sidak's plan goes under the name of Efficient Components Pricing, or ECP.

In the long run, the residual monopoly elements of the current system are almost certain to disappear as new technologies erode LEC cost advantages. However, there is no reason, in principle, why obligations cannot be tied to specific classes of customers and/or services in such a way that firms competing for these customers would do so contingent on assuming their contributory obligations.11

The current system of access charges used to fund the intraindustry transfers supporting universal service goals is crumbling rapidly. The largest toll customers have been able to avoid the universal service contributions implicit in these charges by turning to CAPs, whose services may be profitable even when their underlying costs exceed those of the incumbent LECs with whom they compete. These inefficiencies and inequities are eroding the economic rationale and the political legitimacy of the current system. As an alternative we have proposed VASS (Value-Added Service Surcharges), a system of levies collected as a fixed and common percent of revenues for selected nonbasic services. Compared with the current system, VASS would have the efficiency advantages of: (1) Leaving the relative prices of contributing services virtually the same as relative costs would dictate in competitive markets not subject to such levies. (2) Minimizing deadweight surplus losses due to surcharge-induced price increases. (3) Placing the primary burden for funding infrastructure enhancements on those benefiting the most.

While a VASS might initially be levied only on the toll services targeted by the current system of access charges, over time it could expand the range of contributing services to minimize distortions of the relative prices of contributing and noncontributing services. Theoretically, the most efficient form of a VASS would allow carriers to deduct any VASS levies built into prices they pay for the telecommunications services they purchase as inputs. Whether the benefits of this approach outweigh the costs of a more cumbersome collection and oversight apparatus is open to question.

Ideally, VASS would be implemented through a coordinated and complementary set of state and federal policies. To provide greater flexibility for meeting various local needs and goals, VASS might be supplemented with efficient components' pricing or customer-specific levies.

1. In rural areas the provisioning of business on/off ramps and core network components might also be issues.

2. REA funds, which have been important to the modernization of rural service, are an exception.

3. This represents the lion's share of the intraindustry flow of transfer payments at risk due to competitive supply of toll services. This flow has been estimated at by Monson and Rohlfs at approximately $20 billion annually. See C. S. Monson and J. H. Rohlfs, "The $20 Billion Impact of Local Competition in Telecommunications," Strategic Policy Research, July 16, 1993.

4. W. J. Baumol and G. J. Sidak, Toward Competition in Local Telephony, Cambridge: MIT Press, 1993.

5. The nonbenefited services of LECs would also be included in the value-added services category. Also, a number of definitional issues regarding residential loop services must still be resolved. For example, what level of functionality should be associated with basic service? And should second and additional access lines for a single residence be subject to the VASS?

6. See J. E. Stiglitz, The Economics of the Public Sector, New York: W. W. Norton & Company, 1986, p. 361, for a more extensive version of this analysis.

7. See the FCC filings in Petition for Declaratory Ruling and Related Waivers to Establish a New Regulatory Model For The Ameritech Region, March 1, 1993, and more specifically, Steven S. Wildman and Bruce L. Egan, "Promoting Local Telephone Competition Through Access Charge Reform," in Ameritech's supplemental filing of April 16, 1993.

8. ibid. Wildman and Egan.

9. Even in the case of a VASS based on market shares measured in physical units (e.g., subscriber access lines of installed network capacity), as long as the data required to administer the system is standardized and periodically reported it should be relatively easy to audit. In practice, the unit measured would have to be one that all affected carriers would be required to report.

10. Baumol and Sidak, op. cit.

11. For additional discussion of customer-specific obligations and problems that arise with various proposals for targeting transfer payments to specific customer classes, see J. C. Panzar and S. S. Wildman, "Competition in the Local Exchange: Appropriate Policies to Maintain Universal Service in Rural Areas," September 1993.