A r c h i v e d  I n f o r m a t i o n

Report to the Federal Communications Commission (FCC)
by
The E-Rate Implementation Working Group (Working Group)
composed of
U.S. Department of Education
Institute of Museum and Library Services
National Telecommunications and Information Administration
Rural Utilities Service
Education and Library Networks Coalition

July 31, 1997

VII. Aggregate Discount Rates

The E-Rate Ruling establishes a matrix for calculating an individual discount rate for each eligible school and public library in the nation, based on a specified poverty measure and its classification as urban or rural (¶520). As discussed at pages 7-8, however, most procurements eligible for E-rate support will likely involve contracts covering multiple schools and/or libraries. Depending on the services ordered and the method of payment under such a contract, it may be appropriate to apply the individual discount rates, to calculate an aggregate discount rate for the entire contract or to use a combination of the two methods. In calculating any such aggregate discount rates, it must be remembered that the Snowe-Rockefeller-Exon-Kerrey Amendment and the E-Rate Ruling specifically identify schools and libraries with high enrollment of students in poverty or in rural areas as those that should most benefit from USF support. The benefit of their special discounts was not intended to be accorded, as the result of some mathematical calculation, to schools and libraries not entitled to them. Accordingly, although the calculation of aggregate discount rates should be as simple and flexible as possible, applicants that calculate an aggregate rate should still be required to strive to ensure that each covered school and library receives the full benefit of the discount to which it is entitled.

The FCC has recognized the impact of aggregation on discount rates and has established averaging as an alternative methodology for calculating aggregate discounts for applications covering multiple schools and/or libraries (¶¶523-24, 528, 569, 576). ¹³ Under the E-Rate Ruling, applicants that are school districts, library systems or States are to strive to ensure that each school and library covered by their applications receives the full benefit of the discount to which it is entitled (¶¶523-24).
Service providers to consortia are to keep careful records, maintained on a reasonable basis of approximation, and publicly available, of the allocation of the cost of shared facilities (¶576).

The Working Group recommends that the methodology and application process in the E-Rate Ruling be further clarified to ensure that they achieve the desired goal of targeting poor and rural schools for special discounts in a minimally burdensome manner, and that the same methodology and process apply both to higher-level governance units for schools and libraries and to consortia. The choice of methodology should depend on the allocability of the services to individual schools and/or libraries and whether billing for the services is central or distributed, not on the legal status of the applicant.

When a single application is filed for a contract covering multiple schools and libraries that will pay their own bills directly, there should be no need to calculate an aggregate discount rate ¹⁴ The applicable individual discount rate should be applied to each bill. The applicant should still strive to ensure that each school or library receives the full benefit of its own discount, and the service provider should retain its record-keeping responsibilities with regard to cost allocation.¹⁵ When the applicant pays the contract charges, however, an aggregate discount rate would be calculated for the funding request. Whether individual schools and/or libraries reimburse the applicant for contract charges paid at that level, or the applicant absorbs the full cost itself, calculation of an aggregate discount rate should presumably be designed to yield the same dollar amount of overall USF support as would have occurred in the distributed-billing configuration. The "perfect" solution for calculation of an aggregate discount rate for central billing would thus require measuring the actual services received by each school and library (including metering of services with charges based on the amount of usage), and then weighting the individual discount rates by the measurements of services actually received. However, sophisticated subsystems would be needed to meter the distributed usage of centrally acquired telecommunications services, and the "perfect" methodology would not allow service acquirers and providers to know the applicable discount rate until after the services had already been provided. At the other extreme, a simple (unweighted) average of the applicable individual school and library discount rates would cause high-volume users (likely users with large populations and/or ample resources of their own) or applicants acting on their behalf to benefit or suffer from the rates of the low-volume users. A population-weighted average of the individual school and library discount rates might reflect potential distribution of the centrally billed services, other things being equal; but the availability of other resources may well be a more significant determinant of that distribution than population. The Working Group believes that the appropriate methodology for the calculation of discount rates for contracts involving central billing for services provided to multiple schools and/or libraries is to average the individual discount rates for those users weighted by the projected allocation of directly allocable services and the projected distribution of nonallocable common or shared services. The applicant should make the calculation in the first instance. Since SLC should retain its ultimate responsibility for the validity of discount rates, the applicant should be required to maintain its calculation work papers for SLC, which would also be available for audit and for public inspection. The FCC seems to have allowed some flexibility in the allocation of costs within consortia and to have balanced that flexibility with a similar requirement with respect to work papers (¶569). This solution would be an expanded version of that allocation system.¹⁶ The Working Group recommends that the following principles be adhered to in the calculation of weighted averages for applications involving multiple schools and/or libraries and centralized billing:

For those services that can be directly attributed to an individual school or library, the discount level for that school must be directly applied to that service. Service providers should be required by regulation and/or by applicants' RFP/descriptions of services to provide detailed breakdowns of projected allocable non-shared costs by individual school and/or library to facilitate this attribution.

For services that are "shared" or "common" to multiple schools and/or libraries (that is, cannot be broken down by user and allocated directly), the applicant will need to determine a rational cost-allocation method. In determining such a method, the applicant should have great flexibility in determining the appropriate methodology for projecting allocation of eligible services covered by the applications that cannot be so broken down and directly allocated. In appropriate circumstances, such methods could include (but are not limited to) the number of networked computers in each school or library divided by the total number of networked computers in the school district or library system. If the applicant can demonstrate that factors other than population affecting allocation of services do not vary substantially, another possible methodology would be to base the allocation on population, either by calculating an average of the individual discount rates weighted by population (the population of individual schools divided by the total district population or the population of library service areas divided by the total population served by the library system) or by calculating an area-wide poverty rate and then applying the special discounts for rural areas. ¹⁷ Appendix A contains two tables illustrating the different methods of calculating aggregate discount rates for central billing. The first table contrasts simple, population-weighted, and service-weighted averaging when services are unevenly distributed and shows the economic incentive that service weighting creates for more even distribution of services. The second table shows the similarity of population-weighted averages and area-wide calculations when they are evenly distributed.

The "work papers" that the applicant should maintain for SLC to show the methodology used to arrive at a discount level should also be available for auditing purposes and to the public. Those work papers should contain not only the applicant's actual calculations, but also a short explanation in layman's language of the rationale for calculating the aggregate discount rate in that manner, including how it assures each school and library the full benefit of the discount to which it is entitled. Finally, the applicant must certify that the discount rate has been calculated according to the principles outlined above. ¹⁸

Because the applicant and the service provider have the relevant information concerning allocation of services, and we expect most applications to involve central billing and at least some common or shared services that cannot be broken down by individual user and allocated directly, we recommend that the applicant rather than SLC calculate the discount rate(s) for the contract in the first instance, although SLC would retain its ultimate responsibility for the validity of the discount rate(s). Service providers should be relieved of the responsibility to allocate the cost of shared services when they cannot do so. To guide applicants in their responsibility to calculate their own discount rates, the Working Group recommends that SLC create a list of individual discount rates for every school and library for which the necessary data is publicly available and post that list on the website.

Service providers have told members of the Working Group that, since discount rates affect only the timing of the payments that they ultimately receive rather than the total, ¹⁹ they do not need to know the rate(s) before responding to a posting. Moreover, projecting allocations of services before the specific services have even been identified is particularly prone to error. Subject to further confirmation of service providers' needs, the Working Group recommends that applicants (or the eligible parties actually paying the bills) calculate their discount rates in the first instance in their funding requests rather than in their initial applications. Moving the discount rate calculation to the funding request would also eliminate the necessity for applicants representing multiple schools and libraries that will be billed separately either to calculate a pro forma aggregate discount rate that will never be used, or to submit to SLC a possibly very long list of individual schools and libraries and their individual discount rates (possibly thousands in the case of a State applicant). In a data warehouse, however, the information on individual schools and libraries would automatically be integrated into the application by reference to their unique identifiers.

Since the FCC's requests for recommendations were limited to the application process, and most members of the Working Group are not sufficiently familiar with the disbursement processes currently used for USF support, we have not attempted to deal with the consequences that should flow from differences between actual distributions of services, as indicated by detailed bills from service providers, and their projected distributions for purposes of calculating aggregate discount rates. In the absence of such differences, application of the previously projected aggregate rate should yield the same mathematical result, at least before any rounding, as the application of individual rates to all directly allocable services broken out in the detailed bills. If the differences between actual and projected distributions in services result in higher aggregate discount rates, the applicants will presumably file amended funding requests. In determining the consequences of differences potentially resulting in lower discount rates, the FCC should, of course, protect the integrity of the E-rate and the targeted nature of its high discounts. When the FCC makes that determination, however, the Working Group does urge it to keep in mind the highly leveraged nature of the E-rate for high-discount applicants, and the potentially great financial impact on them from relatively small absolute changes in the calculated discount rates which they may have used to commit themselves to substantial contractual obligations.

¹³ See also §§54.505(b).
¹⁴ This could happen, for example, in the case of a consortium contract or a State bid list that individual lower-level units have the option of using. The Working Group recommends that such master contracts should be subject to competitive bidding through posting only before signing by the consortium or State. When a school, library or higher-level entity eligible to use the master contract elects to obtain eligible services under the contract, it should be allowed to file a combined application and funding request without a second posting. The original applicant could still have the option of filing an aggregate funding request to earmark funds for the contract as soon as possible, and combined application/funding requests from individual schools and libraries or higher-level entities would then be credited against this amount.
¹⁵ This would be the case if the individual bills were based on the services actually used by each school and/or library.
¹⁶ Calculation of weighted averages will often result in aggregate discount rates that are not on the matrix in the E-Rate Ruling. Some service providers have informed members of the Working Group that their legacy computerized billing systems will only accommodate a limited number of discount levels. We recommend that applicants round their aggregate discount rates to the nearest five percent to accommodate the service providers.
¹⁷ An area-wide poverty calculation is not technically an average but a ratio. In doing an area-wide discount-rate calculation using such a ratio, the applicant should not be able to consider itself entirely rural unless that is actually the case. To ensure that the special discounts for rural schools and libraries are not extended to urban areas, the Working Group recommends that separate aggregate discount rates be calculated for rural and urban entities, and that the two rates in turn be averaged on a population-weighted basis.
¹⁸ Because of the importance of ensuring that schools and libraries do in fact get the full benefit of the discounts to which they are entitled, the Working Group recommends that SLC develop a computer program to identify funding requests in which the aggregate discount rate calculated by the applicant exceeds certain parameters, such as its relationship to the applicable individual discount rates. In such cases, but without delaying the application process, SLC should review the applicant's work papers to verify the calculation before funds are actually disbursed from the Universal Service Fund, rather than waiting for the possibility that it might be reviewed after the fact in a random audit.
¹⁹ In paragraph 51 of the July 18 Ruling, the FCC established a 40-day time limit for payment from the time of notification by the eligible party paying the bill.

[ VI. Interim Requirements ]					[ VIII. Approval of Technology Plans ]

This page last updated 8/10/97 (pjh).