Electronic Comment Filing System

Before the FEDERAL COMMUNICATIONS COMMISSION Washington, D.C. 20554 In the Matter of ) ) Connect America Fund ) WC Docket No. 10-90 ) A National Broadband Plan for our Future ) GN Docket No. 09-51 ) Establishing Just and Reasonable Rates for Local ) WC Docket No. 07-135 Exchange Carriers ) ) High-Cost Universal Service Support ) WC Docket No. 05-337 ) Developing an Unified Intercarrier Compensation ) CC Docket No. 01-92 Regime ) ) Federal-State Joint Board on Universal Service ) CC Docket No. 96-45 ) Lifeline and Link-Up ) WC Docket No. 03-109 ) Universal Service Reform – Mobility Fund ) WT Docket No. 10-208 COMMENTS OF ACCIPITER COMMUNICATIONS INC. Accipiter Communications Inc. 2238 W. Lone Cactus Dr. Suite 100 Phoenix, AZ 85027 (623) 455-4500 January 18, 2012 TABLE OF CONTENTS I.  INTRODUCTION............................................................................................................ 1  A.  The company has made recent investments to reach previously unserved populations with advanced telecommunications services. ..................................... 5  B.  Accipiter’s assigned study area contains some of the highest-cost-to-serve locations in the nation. ........................................................................................... 6  C.  The deployment of telecommunications facilities in the last five years has required investment in facilities which support broadband services. .................... 6  D.  Accipiter’s shareholders and principals, the ACC, and the RUS all relied upon the statutory predictability and sufficiency of USF being available to sustain investments to high-cost rural customers who would otherwise remain unserved. .................................................................................................... 6  E.  The long-term growth potential of Accipiter’s study area provides financial benefits which offset the need for USF support to serve high-cost subscribers.............................................................................................................. 7  II.  COMMENTS ON ELIMINATING SUPPORT FOR AREAS WITH AN UNSUBSIDIZED COMPETITOR. ................................................................................ 8  A.  Adjusting support in study areas with less than 100% of competitive overlap fails to recognize an ETC’s obligations in its study area. ......................... 8  B.  Unsubsidized competition should be qualified. ................................................... 12  C.  The FCC’s methods for determining the extent of competition are flawed ........ 14  III.  COMMENTS REGARDING LIMITS ON REIMBURSABLE CAPITAL AND OPERATING COSTS FOR RATE-OF-RETURN CARRIERS ..................... 15  A.  The proposed methodology should be reconstructed to consider only costs categories where limitations are needed. ............................................................. 17  B.  The proposed methodology fails to accomplish the Commission’s objective of comparing similarly-situated companies. ........................................ 19  C.  The complexity of the proposed methodology and its proposed implementation fails to provide “predictability” to USF. .................................... 27  D.  The proposed methodology should not limit costs that are in-place due to decisions made in the past according to the former USF rules. ........................... 30  E.  Accipiter’s history and cost structure highlights a number of issues the FCC methodology inappropriately fails to consider. ........................................... 30  F.  The limitation methodology for HCLS support must be properly constructed before limitations impacting ICLS can be considered. .................... 36  IV.  CONCLUSION .............................................................................................................. 36  SUMMARY Accipiter Communications Inc. (“Accipiter”) submits these comments to express grave concerns over a number of elements in the FCC order that could result in: 1. A severe financial impact on the company that would endanger the only source of telecommunications services to a number of rural Arizona residents and businesses. 2. Uncertainty in future funding that would prohibit Accipiter from making investments which leverage its existing network to extend advanced telecommunications services to rural Arizona consumers who are still unserved. 3. Policies which fail to protect consumers with carrier-of-last-resort (“CoLR”) obligations and in fact promote behavior that can leave consumers victims of unregulated monopolies while stripping CoLR providers of economies of scale which are necessary to serve consumers in high-cost areas. Accipiter is a relatively small ILEC with very limited resources and is sensitive to the FCC’s limitations on corporate operations expense. Accipiter does not have the manpower or budget to provide full comment on all aspects of the FNPRM nor does Accipiter believe it would be a wise stewardship of its resources to do so. However, there are two elements of the FNPRM that could have a grave impact on Accipiter’s operations and its ability to serve rural customers in its study area. These two elements are: • Eliminating support for areas with less than 100 percent overlap of an unsubsidized competitor, and • The limitations placed upon costs by the FCC’s proposed regression methodology. Accipiter’s comments focus on the impact of these proposals on the company. Accipiter is also concerned about many other elements of the FNPRM and expresses its support for additional comments filed by the company’s consultants Moss-Adams, LLC. Accipiter expects its rate-of-return peers and rural industry associations (including NECA, NTCA, OPASTCO, and WTA) will also be filing comments expressing concerns over this rulemaking. Accipiter hopes that the FCC fully considers the serious potential impact of the proposed rules. Rate-of-return carriers are the primary stewards of advanced telecommunications services for the hardest-to- serve consumers in the United States. These rural consumers provide many essential goods and services that benefit the entire nation. It would be tragic for the FCC to move forward in rule- making while ignoring the concerns expressed by the rural telecommunications industry. Before the FEDERAL COMMUNICATIONS COMMISSION Washington, D.C. 20554 In the Matter of ) ) Connect America Fund ) WC Docket No. 10-90 ) A National Broadband Plan for our Future ) GN Docket No. 09-51 ) Establishing Just and Reasonable Rates for Local ) WC Docket No. 07-135 Exchange Carriers ) ) High-Cost Universal Service Support ) WC Docket No. 05-337 ) Developing an Unified Intercarrier Compensation ) CC Docket No. 01-92 Regime ) ) Federal-State Joint Board on Universal Service ) CC Docket No. 96-45 ) Lifeline and Link-Up ) WC Docket No. 03-109 ) Universal Service Reform – Mobility Fund ) WT Docket No. 10-208 COMMENTS OF ACCIPITER COMMUNICATIONS INC. The Federal Communications Commission’s (Commission or FCC) Report and Order and Further Notice of Proposed Rulemaking in the above-captioned proceeding requests comment on proposed rules intended to modify the existing rules on the Universal Service Fund (USF) and Intercarrier Compensation (ICC).1 Accipiter Communications Inc. (“Accipiter”) submits these comments for the FCC’s consideration. 1 Connect America Fund, A National Broadband Plan for Our Future, Establishing Just and Reasonable Rates for Local Exchange Carriers, High Cost Universal Service Support, Developing an Unified Intercarrier Compensation Regime, Federal-State Joint Board on Universal Service, Lifeline and Link-Up, Universal Service Reform – Mobility Fund, Report and Order and Further Notice of Proposed Rulemaking, FCC 11-161 (rel. Nov. 18, 2011) (“FNPRM”). 2 I. INTRODUCTION Accipiter is an Incumbent Local Exchange Carrier holding a Certificate of Convenience and Necessity (“CC&N”) and ETC status granted by the Arizona Corporation Commission (“ACC” or “Arizona Commission”) to provide local telephone service to a study area of 1,010 square miles northwest of Phoenix, Arizona. There are approximately 4,600 inhabited residences within the study area. Most of the study area is very sparsely populated. As of November 2011 Accipiter serves 875 regulated loops (718 residential and 157 businesses). Essentially all of these lines may be reached with Accipiter’s high speed Internet services. Of these, 356 lines are served by fiber-fed digital loop carrier systems and the remaining lines are served through fiber-to-the- home (“FTTH”) facilities. Accipiter was incorporated in 1995 and in that year was granted a Certificate of Convenience and Necessity (“CC&N”) by the ACC to serve portions of Maricopa and Yavapai counties in Arizona.2 The original Accipiter service territory encompassed approximately 650 square miles and 115 occupied residences. As the Arizona Commission found “many residents of the amended proposed service area support Accipiter’s efforts to expand telephone service in the area. We heard from a number of residents of the enormous construction charges USW [US West, subsequently Qwest] has quoted to extend service to remote locations in the area. One resident commented that USW would charge $20,000 to extend service 100 yards to his property. Another spoke of a $200,000 charge to receive service from USW.”3 2 In The Matter of the Application of Accipiter Communications Inc. for a Certificate of Convenience and Necessity Authorizing the Construction and Operation of a Public Utility Telephone System in Portions of Maricopa and Yavapai Counties, Arizona, Decision No. 59346 (Docket U-2847A-95-0026). (“Accipiter CCN Order”). 3 At the time the Certificate was granted, USW “provided telephone service to approximately 22 customers (“the existing subscribers”) over approximately 30 access lines”. Accipiter CCN Order at 3. 3 As is true in many rural service areas, the Accipiter’s service proposal was made possible by a combination of USF support provided under the FCC’s Universal Service programs and low interest loans for rural telecommunications development provided by the Department of Agriculture Rural Utilities Service (RUS). As a result of the support provided by these programs, Accipiter was able to deploy service charging local tariff rates of $16.78 per month for residential service and $35.78 for business service. As the Arizona Commission noted, prior to the Accipiter proposal, residents in the area were quoted prices of tens or even hundreds of thousands of dollars to pay for the extension of ordinary telephone service to their homes. The Commission’s Universal Service program and the Agriculture Department’s RUS lending program working together as intended by Congress placed these rural residents on par with residents of urban areas. In 2003 Accipiter filed with the ACC an application to expand its service territory by serving an area where no ILEC or ETC had previously been designated. Although portions of the area were within the incorporated limits of the Arizona towns of Buckeye and Surprise, residents living within the proposed boundary had no regulated telephone or terrestrial broadband service available. A limited number of residents in this area could receive weak cellular signals and could make wireless phone calls of inconsistent quality from their home. The Buckeye extension application was granted by the Arizona Commission on January 14, 2005.4 In granting the extension, the Commission noted that: As a rural carrier, Accipiter’s ability to serve high-cost areas that would otherwise remain unserved for many years promotes the public interest by enabling rural customers the opportunity to 4 In the Matter of the Application of Accipiter Communications, Inc., for an Extension of its Existing Certificate of Convenience and Necessity, Docket No. T-02847A-03-0655, Decision No. 67675. (“Buckeye Extension Order”). 4 receive voice and data service, including calling to the Phoenix Metro calling area. The benefit of extending telecommunications services to rural areas is more than a hypothetical possibility. At the hearing, an existing Accipiter customer in the Lake Pleasant exchange, Mr. Joe Hull, offered public comment in support of the company’s application. Mr. Hull stated that he resides in the Castle/Hot Springs area north of Lake Pleasant, along with approximately 40 other families. Despite the lack of any paved roads in the area, Mr. Hull indicated that Accipiter provides Castle/Hot Springs residents with local calling to the Phoenix Metro area as well as high speed internet service. Mr. Hull claims that Accipiter has consistently delivered on its promises to area residents in providing telecommunications services.5 The ACC also took explicit notice of the importance of the FCC’s Universal Service policies as well as RUS funding in making such rural service possible. “Staff witness Boyles also testified that granting Accipiter ILEC status for the proposed extension area provides a benefit to potential customers due to Accipiter’s status as a rural carrier. Rural carriers depend on Federal Universal Service Funds (“FUSF”) to compensate for the difference in costs incurred to serve high-cost rural customers and revenues received from such customers for service (citation omitted). As a rural carrier receiving funding from RUS, Accipiter would therefore not be permitted to charge customers for construction costs incurred by the company to extend service.”6 As noted previously, the FCC’s Universal Service programs and the RUS telecommunications lending program working in concert as intended by Congress place rural customers in essentially the same posture as urban customers who do not have to pay special and extraordinary charges to have a telephone company extend its network to serve them. Although much of the Accipiter service territory was, and continues to be, very sparsely populated, for approximately the past seven years portions of the service territory have been the focus of a number of significant development efforts. These development efforts peaked prior to 5 Id. at 9-10. 6 Id. at 8. 5 the 2008 real estate crash and since that time modest development has resumed. While the advent of suburban-density homes is changing the rural character for approximately five square miles of the company’s 1,010 square mile study area, Accipiter notes three important aspects of suburban density development which are relevant. First, these developments emerged in what was once vacant and remote desert land. Constructing facilities to reach these developments before having the ability to add customers proves costly, yet Accipiter shoulders the carrier-of-last-resort obligation to serve the businesses and residents emerging there. Second, the densities and scale occurring in these developments offer efficiencies and economies which could lower Accipiter’s dependence upon USF support while still assuring that Accipiter’s rural customers have access to advanced telecommunications services. Third, in each development in the study area Accipiter has encountered anticompetitive CLEC behavior which serves to deny residents and businesses a choice of telecommunications providers as well as access to the regulated telephone carrier. Accipiter’s inability to serve these developments further harms rural customers by stripping the study area of the economies of scale critical for supporting consumers in rural densities. Further, exclusion of the ILEC allows the CLEC to selectively monopolize the most profitable customers while skirting regulatory obligations to serve customers that the CLEC finds unattractive. The background recounted here is critical to understand some unique elements of Accipiter’s history and study area which are relevant context for its comments on the FNPRM. A. The company has made recent investments to reach previously unserved populations with advanced telecommunications services. Almost 70% of the company’s total plant investment has been made in the past five years while expanding services into an area of rural Arizona which was previously unserved. Since the 6 initial network investment in 1996 the company has initiated service to approximately 150 customers who didn’t before have access to regulated telephone or broadband services. B. Accipiter’s assigned study area contains some of the highest-cost-to-serve locations in the nation. There is a reason that several of the establishments in Accipiter’s service area lacked service before Accipiter’s deployment7. Accipiter’s customer base resides in some very remote and sparsely populated areas of Arizona. Extending services to these populations requires difficult and costly network construction through rocky and mountainous terrain. C. The deployment of telecommunications facilities in the last five years has required investment in facilities which support broadband services. From a practical standpoint a company making a long-term investment in telecommunications plant gives strong consideration to infrastructure which will support the consumers’ needs for the service life of the plant to be constructed. Certainly in the last five years telecommunications plant investments have required deployment of broadband capable infrastructure. This was reflected in the rhetoric and policies of Accipiter’s lender (the Rural Utilities Service8) as well as the Arizona Commission9. 7 In fact, it is noteworthy that there are still approximately 129 unserved residents within the service area. 8 See, e.g., Agriculture Secretary Vilsack Announces Funding to Expand and Improve Broadband Services in Rural Areas, News Release No. 0485.11, available at http://www.usda.gov/wps/portal/usda/usdahome?contentid=2011/11/0485.xml&navid=NEWS_R ELEASE&navtype=RT&parentnav=LATEST_RELEASES&edeployment_action=retrieveconte nt (Nov. 14, 2011). 9 See, e.g., Letter from ACC Chairman Kris Mayes to Arizona Telecommunications Companies, available at http://www.azcc.gov/commissioners/Mayes/Stimulus/Telecom/Mayes%20Letter%20to%20Telec om%20Cos.pdf (Apr. 3, 2009) 7 D. Accipiter’s shareholders and principals, the ACC, and the RUS all relied upon the statutory predictability and sufficiency of USF being available to sustain investments to high-cost rural customers who would otherwise remain unserved. The universal service program has always had the objective of making comparable services available to all American consumers at reasonably comparable prices. The various stakeholders involved in Accipiter’s regulatory approvals, financing, and network investment decisions relied upon the statutory principles of predictability and sustainability in USF in moving forward with extending services in Accipiter’s study area. E. The long-term growth potential of Accipiter’s study area provides financial benefits which offset the need for USF support to serve high-cost subscribers. Once the prospects for suburban density growth began to become a reality in Accipiter’s study area, the company aggressively pursued the opportunity, realizing that the economies of scale and efficiencies to be obtained in suburban densities would provide low-cost offsets to Accipiter’s high-cost rural customer base. With these low-cost customers averaged into Accipiter’s total study area costs, the company continues to lower its dependence upon USF support mechanisms.10 The company’s best offset to USF dependence is to leverage existing investment in its suburban densities and continue adding low-cost customers. In Accipiter’s short history the company, its state regulatory body, and its lender all relied upon the FCC’s statutory obligations to provide high-cost universal service support that is predictable and sufficient. While following the rules of the program, the company extended 10 Of course, the company also has the carrier-of-last-resort obligation to serve homes and businesses that are constructed within these subdivisions. 8 advanced telecommunications services to previously unserved populations. The FCC now proposes certain changes in the universal service program which would undermine the company’s prior investment. The result is an unnecessary and unproductive confiscation of the company’s assets and an endangerment of the sole source of advanced telecommunications services for a number of Arizona consumers. The proposed FCC rules also fail to recognize the unique challenges in serving rural and unserved populations. Additionally, the proposed rules fail to recognize the true factors that drive costs for the highest-cost-to-serve companies in the nation. For these reasons the FCC must delay implementation of certain aspects of the proposed rules so they can be reconstructed in the proper context. Only after that can the FCC’s revised rules can be properly vetted and moved forward into implementation. II. COMMENTS ON ELIMINATING SUPPORT FOR AREAS WITH AN UNSUBSIDIZED COMPETITOR. A. Adjusting support in study areas with less than 100% of competitive overlap fails to recognize an ETC’s obligations in its study area. If an unsubsidized CLEC has chosen to provide service within a portion of an ILEC’s study area, it follows that the CLEC’s services in that particular portion of the study area are economically feasible without requiring USF support. The FCC’s proposed rules seem to imply an extended logic that suggests that the ILEC therefore should also be able to economically support services in the area of overlap without requiring any support. This faulty logic fails to recognize that the ILEC must also serve the larger and more sparsely populated area in its territory and has different regulatory obligations for the area, different market dynamics, and different benefits from providing service in the area of overlap. An ILEC that is an ETC carries a number of regulatory obligations that are not imposed upon a CLEC. Most notable is an ILEC’s carrier-of-last-resort obligations. A CLEC has the 9 competitive advantage of being able to pick and choose which areas and customers it would like to serve, the ILEC has more strict oversight requiring extension of services regardless of consumer income, demographics, revenue potential, population densities, and other factors. The ILEC’s regulatory obligations represent a cost that the CLEC does not carry but an ILEC does, not only for the area of overlap but the entirety of the ILEC’s study area. A CLEC that chooses to compete in a limited area without incurring the regulatory obligation to serve either the area selected nor the entirety of the ILEC’s study area is not harmed by the presence of the ILEC’s USF support (if the CLEC did perceive such harm it would most likely choose not to compete in the first place). Further, a downward adjustment of the ILEC’s revenues would impose harm upon the consumers in the ILEC’s study area who depend upon the regulatory protections provided by the state and federal tariff and terms-of-service rules applying specifically to the ILEC. While one may argue that it is poor stewardship of the universal service funding to allow an ILEC to receive support while a CLEC serves the same area without support, this argument loses its credibility when the area of overlap is a subset of the ILEC’s total study area. Universal service support for a rate-of-return ILEC is not calculated on a per-subscriber basis, but rather is a function of the total costs to serve the entirety of a study area. Thus the ILEC’s customers in the area of competitive overlap probably have densities and economies of scale which are lower than the ILEC’s average cost per subscriber and therefore lower the ILEC’s dependence upon USF for the highest cost subscribers in the study area. This provides an internal subsidy mechanism that helps support high-cost customers before additional USF support is required. In fact when viewed in this context perhaps the FCC’s updated universal service goals would be better served if a CLEC was prohibited from competing in a subset of an ILEC’s study area and 10 thereby depriving the ILEC of that portion of the ILEC customer base that actually lessens its dependence upon the USF fund. This is not the only difference in market dynamics that must be considered between ILECs and CLECs. Most frequently the facilities-based CLECs present in the higher density portions of ILECs’ study areas are cable television providers. These cable providers established their presence in the market with video offerings and high-speed internet services. With their backbone networks supported by the revenues generated by video and broadband services, the incremental investments required to deploy competitive voice service required very little revenue to support. While Accipiter recognizes that an ILEC has the opportunity to compete with its own video and broadband services, it is important to recognize that an ILEC’s obligations to a study area promote a different market dynamic than the CLEC’s opportunity to selectively choose its markets. A cable CLEC that comes from the background of providing video services as its primary revenue stream will require a certain amount of subscriber base to gain the economies of scale necessary to support investment in video headend equipment. A specific level of per-subscriber revenue would also be required to ensure the cable CLEC is profitable. Since the cable industry is driven mostly by unregulated free-market dynamics the cable companies generally seek markets with adequate densities and economies of scale and that can support sufficient revenues per subscriber. Contrast this with an ILEC having regulatory obligations for a specific study area. The ILEC must serve regardless of densities and economies of scale and further must provide services at publicly scrutinized tariff rates. Fortunately the universal service high-cost program has traditionally provided the support necessary for the ILEC to meet its service obligations. 11 Now consider the situation where a CLEC leverages its larger economies of scale and higher per- subscriber revenues to make incremental investments whereby it may enter a smaller-sized ILEC’s study area and offer competitive services. Unless the CLEC is unwilling to leverage its existing economies of scale and revenues to extend services into the entirety of the ILEC’s study area, it would be inappropriate to dilute the ILECs support and thus endanger the support required to serve areas undesirable to serve by the CLEC. It is important to reiterate the differences in benefits of serving an area subject to competition. In any area of overlap where an ILEC competes with a CLEC, both are competing for customers. More specifically, both are competing to leverage their existing network investment to add customer revenues. In areas of competition, the incremental investment to add a customer is most likely small and thus very beneficial to a carrier’s cost recovery structure when the customer revenue is added for a relatively small incremental investment. For the unsubsidized CLEC, the financial benefit adds profit to its company, thus the CLEC is the sole financial beneficiary of adding the new customer. For the USF-supported ILEC, the financial benefit is imputed into the cost study and used to offset the USF revenues required to serve the high-cost customers in the study area that the CLEC refuses to serve. Thus the high-cost customers in the ILEC study area are financial beneficiaries of the ILEC’s new customer, as well as all participants in the universal service fund. This point is made particularly poignant in light of the fact that the FCC’s USF reform has imposed new per-loop caps on USF support and has additional per-loop caps proposed. With these caps in place an ILEC’s incentive to add customers in its competitive footprint is even greater, as an added regulated loop will add clearance under the FCC’s per loop caps. On the contrary an ILEC losing support because of a CLEC’s partial presence in the ILEC study area 12 would result in the reduction of supported lines in the highest density areas and thus create higher support per line requirements in the lower density areas, potentially leaving the ILEC and its high-costs customers victims of the per loop caps. B. Unsubsidized competition should be qualified. There are three subdivisions within Accipiter’s study area where a CLEC established preferred-provider agreements with the developer which create significant market barriers for Accipiter’s entry into the market. Accipiter contends that these preferred-provider agreements, which are a nationwide trend and not just unique to Accipiter’s study area, serve as a form of subsidy that victimize consumers while avoiding the regulatory obligations normally imposed upon a sole-source telecommunications provider. The FCC must qualify “unsubsidized competition” so as to avoid promoting monopolistic behavior that skirts regulation and harms consumers. The basic idea of a preferred provider agreement establishes an arrangement between a “preferred” telecommunications carrier and the developer of a new subdivision. The carrier will agree to provide voice, video, and broadband services to all homeowners and businesses in the subdivision as they become occupied. In return the developer will provide the carrier with exclusive rights to market and sell telecommunications services on the developers’ property. Since the developer controls the marketing and sales process for new homes in the development, the preferred carrier is virtually guaranteed to have the new homeowner signed up for service when the home closing documents are signed. Often the preferred provider agreements provide financial terms where the developer provides cash payments to the carrier to address the carrier’s costs of initial construction before new home move-ins begin and customer revenues begin flowing. The carrier will then return 13 financial compensation back to the developer over time, usually on terms that are incentivized so that the developer has more to gain if the carrier gains higher market penetration of telecommunications services sold in the development. Sometimes these agreements involve an assessment within the development’s homeowners association (HOA) fees where the carrier receives a portion of the HOA payment whether that particular homeowner subscribes to the carrier’s services or not. These arrangements have also involved restrictive covenants that make it more difficult for consumers to access the services of a competitor. It is notable that these arrangements are established upon the initiation of the HOA, at the point in time when the developer controls the homeowners association. It is obvious that the preferred provider agreement has a chilling effect on competition. Indeed, that is one of the objectives of these agreements. When the development involves rural areas where costs are high to construct facilities to the new subdivision, such an anticompetitive arrangement which excludes the ILEC can be particularly concerning. When the ILEC is deprived of the high-density establishments in its study area, the ILEC is deprived of densities and economies of scale that could help lower its dependence upon USF and thus further secure the long-term assurances for sustaining service to rural customers. Further, the consumers in a new subdivision subject to a CLEC as the preferred provider are unknowingly and possibly illegally deprived of a choice in telephone, video, and broadband services, while also losing the regulatory protections provided through ILEC/ETC/carrier-of-last-resort status. 14 Accipiter has concerns regarding the appropriateness of these agreements11 and believes the FCC should conduct its own inquiry on how such agreements affect administration of its universal service support programs and potentially harm consumers. Specifically within the context of the current rulemaking the FCC should determine if the presence of such an agreement should cause the Commission to disregard the presence of a “competitor” in determining whether such a competitor should affect the USF support available to the ILEC/ETC and whether such agreements are harmful to the consumer. C. The FCC’s methods for determining the extent of competition are flawed Accipiter emphasizes that it would be inappropriate for the FCC to limit support in a study area where the extent of unsubsidized competition is anything less than 100 percent. However, for purposes of measuring 100 percent competition in USF supported study areas, the FCC’s proposed method has the following flaws: 1. The Tele Atlas tool the FCC uses to define study area boundaries is based upon algorithms that can be highly inaccurate in depicting a rural study area. Exhibit 1 portrays Accipiter’s actual 1,010 square mile study area compared with the Tele Atlas 36 square mile depiction of Accipiter’s study area. 2. Extracting census blocks whose centroids reside within the boundaries of the Tele Atlas defined study area boundary introduces additional error into the analysis, as census blocks aren’t particularly coterminous with study area boundaries. 11 See, e.g., Letter from Phillip K. Sotel, Accipiter general counsel to U.S. Department of Justice, available at http://www.justice.gov/atr/public/workshops/telecom2007/submissions/227981.htm (Nov. 27, 2007) 15 3. The State Broadband Initiative data is subject to inaccuracies, allowing for incorrectly stating the extent of census blocks served by wireline providers and failure to determine if the competing provider actually offers regulated voice services. 4. It appears the FCC would measure and report the extent of competition by the percentage of population where competition is assumed to exist compared to the total population in the estimated study area. This would fail to account for populations that the ILEC may not serve. Given that the FCC’s proposed methodology is based upon a sequence of steps all of which introduce their own margin of error, the errors are cumulative and the final result is highly likely to be flawed. Certainly a carrier identified with 100 percent competition in its service area should be given adequate time and opportunity for rebuttal. The FCC should also carefully consider the additional burden placed upon a carrier filing a rebuttal required only because of the FCC’s flawed approach to determining the extent of competition. III. COMMENTS REGARDING LIMITS ON REIMBURSABLE CAPITAL AND OPERATING COSTS FOR RATE-OF-RETURN CARRIERS The FCC’s proposed methodology for imposing limits upon reimbursable capital and operating costs fails to adequately measure and predict costs for similarly-situated companies and therefore in its current structure cannot be implemented without compromising the status of telecommunications services to rural Americans. Because of the dramatic impact the proposed model can have on rural carriers and the unmeasured combined impact of other USF reforms, the proposed quantile regression model must be redesigned and subject to further public scrutiny before any new limitations on capital and operating costs are implemented. 16 The FCC requires that the methodology for limiting payments from HCLS compare costs of similarly-situated companies12. In the current proposed quantile regression model, similarly- situated companies are companies who are similar in the value of their independent variables. As Accipiter demonstrates below, the selection and quantification of independent variables fails to identify the real variables which define the costs of a rural carrier. Therefore the model does not and cannot compare costs of similarly-situated companies. Instead the model produces arbitrary results which can impose limits upon a rural carrier who has properly and efficiently accomplished universal service objectives within the standing rules of the program and the letter and spirit of the Communications Act. Accipiter also notes below that the design of the model introduces so many interacting variables that it would be impossible for a rural carrier to forecast USF support within a reasonable margin of error. Because of this lack of predictability a rural carrier will not be able to make decisions regarding investments to serve and sustain rural customers. Not only does this fail to achieve the statutory obligation that USF be predictable, it also produces a result that is unnecessarily detrimental to the support and advancement of broadband services to rural America. The implementation of the proposed methodology would impose limits upon carriers for decisions that were made under the former USF rules. Accipiter believes this retroactive application of the proposed rules is confiscatory. Further, the post hoc limitation of revenues that are needed to support investments which already provide telephone and broadband services to rural consumers could result in the elimination of services in places where USF has already supported investments by ILECs pursuing the program’s existing objectives to serve rural 12 FNPRM at ¶ 217. 17 customers. It would be poor stewardship of USF funds if a rule change resulted in the abandonment of that which had already been accomplished and services that are being provided today. While Accipiter’s comments here offer potential improvements to the FCC methodology, Accipiter reemphasizes that the model used to create the proposed cost limits is so severely flawed that it would require more than a few simple adjustments to be ready for implementation. The model must be significantly redesigned and resubmitted for public comment before it can be used to determine the revenues of rural carriers and services extended to rural America. A. The proposed methodology should be reconstructed to consider only costs categories where limitations are needed. The FCC has already approved a $250 per-line monthly cap on total high cost support to implement fiscal limits on universal service support.13 Implementation of this cap was delayed six months after approval so companies could “make operational changes, engage in discussion with their current lenders, and bring any unique circumstances to the Commission’s attention through the waiver process.”14 This $250 per-line monthly cap serves the purpose of limiting overall support. Presumably a company above this cap would need to restructure operations and/or obtain limited duration waivers15 so that eventually the company could operate below the cap and operate within USF’s newly defined fiscal limits. It is important to realize the purpose and effect of the $250 per-line monthly cap when considering the need for additional caps upon USF recipients. The $250 per line cap represents a 13 Id. at ¶ 274. 14 Id. at ¶ 279. 15 Id. at ¶ 278. 18 global cap above which a rural study area costs will not be supported. The cap also effectively “redlines” the highest cost areas of the rural U.S. by providing that landline service will not be supported for customers in those areas. The cap encourages rural carriers not to undertake service to such areas in the first place. Beyond the global $250 cap the individual caps placed upon specific capital and operating costs serve as additional limitations to assure that “companies do not receive more support than necessary to serve their communities.”16 Because the global cap is intended to address overall fiscal prudence, the individual caps should only address specific costs where there is an incentive to be imprudent. Further, the individual cost caps should consider the interplay between different cost categories to avoid penalizing a higher investment in one cost category to produce lower costs in another category. The FCC’s methodology fails to recognize this and needs to be reconsidered in the proper context. An example is provided by the costs associated with Cable & Wire Facilities (C&WF). There is limited incentive for a carrier to overspend in this category of plant.17 To the extent a carrier spends more in C&WF than its similarly-situated peers, the extra costs would most likely be justified because of a reduction in another cost category. For example, spending more in C&WF often reduces costs for the ongoing maintenance of that facility. Additionally, an extra expenditure on C&WF can reduce costs related to Central Office Equipment (COE). 16 Id. at ¶ 210 17 Importantly, there is limited legitimate incentive for a carrier to overspend in C&WF. One could imagine a scenario, however, where a carrier used USF to improperly overpay for construction and thus funnel USF monies to a contractor. Fraudulent and abusive behavior cannot be tolerated in the USF program, thus one possible appropriate use of the FCC regression model is to identify carriers that significant outliers in comparison with their similarly-situated peers and trigger an audit to determine if the USF funds are being used legally. 19 The FCC’s proposed model produces individual caps which are independently calculated on different cost categories without considering the interaction between the cost categories. So a carrier which spent more on C&WF could be limited in that category without getting consideration for the cost reductions realized in other cost categories such as COE or maintenance. The FCC’s overreach in applying cost category limitations is manifested in the fact that almost 40 percent of the companies in the FCC data set are limited in at least one cost category even though the costs of many of these companies may be reasonable overall.18 If the proposed model is structured to limit only the top 10 percent of companies in a particular cost category, it makes no sense that the FCC’s ultimate result would impose revenue limitations on almost 40 percent of all the companies in the FCC data set. The FCC should redesign its model to select fewer individual cost categories subject to limits. The limited cost categories should only be those where incentives to overspend may exist. Further, cost categories should only be limited where management changes can result in a change in the cost structure. In this regard the FCC might consider operating cost categories primarily driven by labor costs, general support expenses, and benefits. Accipiter notes that costs related to plant expenditures do not lend themselves to a carrier’s incentive to overspend and generally cannot be restructured to meet newly imposed limitations – these plant cost categories should be excluded from the FCC’s methodology. Accipiter notes than the FCC’s methodology chose to avoid establishing a limitation on costs related to rents because “the regression fit is so poor.”19 It is curious that the rents category would be relevant enough for the FCC to consider for its limitation model, yet the only criteria 18 FNPRM at ¶ 398 19 Id., Appendix H at ¶ 17 20 mentioned for excluding this category is the arbitrary criteria that the FCC’s model cannot produce an acceptable output. This indicates that the proposed regression model has become the tail that wags the dog – a flawed approach which also indicates the extent of design flaws in the model. B. The proposed methodology fails to accomplish the Commission’s objective of comparing similarly-situated companies. Within the proposed quantile regression model, similarly-situated companies are those who are similar in the value of their independent variables. Therefore one must examine the independent variables upon which the model was built to determine if they adequately reflect the cost drivers of a rural carrier. Accipiter notes that the challenge in building reliable regression models is that they must be built upon adequate independent variables that truly predict the desired level of the selected dependent variable. 1. The measurement of independent variables introduces too much error into the data set upon which the model was built. The measurement of independent variables in the FCC methodology begins with the study area determined by the Tele Atlas Telecommunications Suite 2010.6.20 Accipiter’s understanding is that the Tele Atlas boundaries are drawn based upon algorithms. Accipiter has found these boundaries to be notably incorrect (by a factor of close to 30:1) for Accipiter’s study area as depicted in Exhibit 1. Accipiter believes Tele Atlas’ algorithmic approach to defining study area boundaries is certain to have a margin of error which probably increases as the study area becomes more rural. Accipiter notes that the FCC acknowledges the Tele Atlas margin of 20 Id., Appendix H at n. 10. 21 error in the discussion on competitive overlap21 but fails to mention that error in its discussion of the regression methodology. Accipiter further notes that it has highlighted this error with the FCC on three separate occasions.22 The Tele Atlas depiction of a carrier’s study area is then used to query census blocks whose centroid is contained within the depicted study area boundaries.23 Census data from these census blocks was used to define the value for ten of the eleven independent variables, the exception being the independent variable of loops. Since census blocks have their own designated boundaries that are not necessarily coterminous with a study area boundary (either a real or algorithmically estimated study area boundary), there will naturally be a new margin of error introduced where census block data is queried in the manner proposed by the FCC. Thus the data collected from census blocks assigned to a carrier will contain values (households, land area, etc) that is not within the carrier’s study area. Likewise the data will exclude values within the carrier’s study area.24 Again Accipiter suggests that the margin of error created here could be more significant in more rural study areas where census blocks tend to be larger and less 21 Id. at ¶ 1066 22 See Letter from Patrick Sherrill, CEO, Accipiter to Sharon E. Gillett, Chief, Wireline Competition Bureau, FCC (Dec. 21, 2010) (submitted pursuant to protective order); Letter from David Cosson, Counsel to Accipiter Communications Inc. to Marlene H. Dortch, Secretary, FCC, CC Docket No. 96-45, App. A (filed Mar. 11, 2011); and Letter from Patrick Sherrill, CEO, Accipiter to The Honorable Fred Upton, Chairman, Committee on Energy and Commerce, U.S. House of Representatives, et al. [cc to The Honorable Julius Genachowski, Chairman, FCC], (August. 12, 2011) 23 FNPRM, Appendix H at n. 36. 24 As an example, Accipiter’s attempt to duplicate the FCC’s process on the company’s actual 1,010 square mile study area yielded results which included census blocks containing 103 square miles of area outside the company study area. Likewise the results omitted 94 square miles of area that is within the company’s study area. The magnitude of these errors can be concealed within the net effect, which only overstates the company’s square mileage by nine square miles. 22 populated. However, smaller and more populated blocks adjacent to a study area boundary are a concern also, as a small error of boundary overlap could significantly influence the value of an independent variable for a carrier and thus skew the results of the FCC model. Nine of the independent variables are subject to the Census Bureau’s urban categorization assignment (either non-urbanized area, urbanized cluster, or urbanized area). When the FCC collected the data used to build its model the urban categorization assignment was not yet complete for the 2010 census data, so the FCC used the categorization from 2000.25 This means the model fails to fully recognize the impact of growth that occurred between the years 2000 and 2010. While it may seem simple enough to wait until the Census Bureau provides the categorization data and then update the model, the problem is that the updated data can dramatically change not only the results of the model but the model’s very structure as well. Further, since the census is only performed every ten years, it is unclear how the FCC will adjust the regression model in subsequent years while waiting for updated census data from the 2020 census. If interim estimates are used, additional error will be introduced into the model’s inputs. If updates are deferred until 2020, the FCC model will fail to recognize the change in a carrier’s costs structure that can be created by growth in the interim period between censuses. The FCC must address this issue before the model can be validated. To summarize, the collection of values which define each carrier’s independent variable values is based upon an algorithmically-generated study area boundary (introducing the first margin of error) and is then used to gather data from census blocks whose centroid’s are contained in that boundary (introducing the second margin of error) and then is tabulated based upon the most recent census data available at the time the analysis is performed (introducing the 25 FNPRM, Appendix H at n. 34. 23 third margin of error). Since the FCC’s regression model was built upon these inputs which are subject to significant skew, the model cannot be trusted as reliable. Even if an independent variable within the model is proven statistically reliable it is tested using a data set that is flawed. Before the FCC model can be properly vetted, it must be reconstructed using more accurate data. Accipiter notes that actual study area boundaries should be on file with all the respective State/Tribal regulatory authorities. Even with this correction Accipiter questions whether the model output will be valid and suggests that the FCC (once the revised model is completed and properly vetted) consider a comment period to permit carriers to correct their individual independent variable data. 2. The independent variables chosen for the model are not good predictors of a carrier’s costs. As noted previously, the selection of independent variables is critical in the design of the model because they define what constitutes similarly-situated peers. Thus each independent variable should be judged upon how well it determines the comparability of costs among carriers. Accipiter comments on each individual independent variable in the following discussion. Accipiter also offers alternative variables for consideration in the model. However, Accipiter reiterates that the inherent flaws in the FCC model must be addressed and the model redesigned before implementation can be considered. The FCC should not consider Accipiter’s suggestion of alternative independent variables as an endorsement of the model in its current form. a. Loops Accipiter acknowledges that loops are a relevant predictor of costs and should remain in the model. However, Accipiter believes the change in loops over time should also be considered, 24 as carriers that are growing (or carriers that are losing loops) are prone to have unique cost structures. Accipiter also notes that the model fails to consider a company’s level of broadband penetration. Because the purpose of USF reform is to promote broadband deployment in rural areas it would be appropriate to include an independent variable measuring broadband customers. While recognizing that the cost study separates specific investments for broadband, presumably the costs of a company that has incurred additional regulated network costs to support and implement broadband should not be compared with the costs of a company which has a lower cost structure because it has not deployed broadband. b. Housing units While Accipiter recognizes housing units as a potential measure of scale, some carriers may not serve all of the housing units contained in their study area. Accipiter assumes this is a minor portion of carriers and therefore can be corrected individually without having a significant impact on the formula created by the regression model. However, the FCC should determine the extent of this error and recalibrate the model accordingly. Accipiter notes that the correction of data for unserved households becomes more complex when the correction requires delineation by urban categorization. c. Land Area The FCC utilized Land Area as a variable for costs related to terrain. While the FCC’s logic concludes that “the more land area that a carrier has in its territory the more expensive it is to serve,”26 the FCC also recognized “the possibility that large swaths of land in a study area may 26 Id., Appendix H at ¶ 27 25 have absolutely no homes or businesses.”27 Thus this variable cannot be a good measurement to adequately compare costs of similarly-situated companies and therefore cannot be included in the model in its current form. Perhaps the FCC could produce a better model by limiting land area to that of only populated census blocks, since that would more accurately reflect the land area served within a study area. Accipiter stresses, however, that accounting only for populated census blocks could fail to consider the costs of extending network across unpopulated census blocks to reach remote populations. Further, it is widely recognized that a major barrier to the deployment of broadband is the transport costs associated with connecting a rural area to the Internet backbone. That is why stimulus funding was specifically directed to target these costs. Yet the FCC’s approach fails to adequately account for these costs. d. Percent Water The FCC concludes that Percent Water serves as an indicator of terrain-driven costs “because roads are typically routed around such water, so the natural pathways for the carrier’s cabling are longer than they otherwise would be.”28 The FCC failed to include any other terrain variables which also require longer cable routes such as mountains, canyons, wilderness preserves, or public lands, just to name a few examples. It is remarkable that the FCC would include Percent Water as a terrain predicting variable while failing to address why other obvious terrain variables that drive costs are excluded. Even if the model statistically validates Percent Water as an independent variable, the exclusion of other terrain driven variables will cause the model to fail to produce results which accurately compare similarly-situated companies. e. Number of Census Blocks 27 Id., Appendix H at ¶ 25. 28 Id. at ¶ 27. 26 The FCC utilizes the Number of Census Blocks as an independent variable depicting density. Accipiter is troubled by the use of this variable, as it does not appear to have any real relationship to density, particularly given that census blocks vary significantly in size. This variation is greater for the more rural census blocks such as those found in rural study areas. The FCC rejected the other proposed independent variable for density, Weighted Housing Unit Density, because it was statistically significant for only one dependent variable. Because this variable was rejected from the model, Accipiter will forgo in-depth commentary on the flaws of this measurement of density. Instead Accipiter simply notes that Weighted Housing Unit Density overstates a company’s density by weighting the density too highly towards the company’s more dense census blocks and thus fails to be a relevant variable for determining similarly-situated companies. Accipiter can identify two independent variables which are relevant for densities which drive a rural carrier’s costs. First, a carrier’s network density can be measured by loops per route mile of cable. RUS borrowers already report this figure annually to RUS. Other carriers should be able to extract this number from plant records. Thus this variable could easily be made available and should be a better variable to be used in predicting costs for similarly-situated companies. Second, the density of a company’s network footprint can be approximated by taking the number of loops and dividing by the land area of populated census blocks. The land area of populated census blocks should be available from the data set already extracted by the FCC. Accounting for land area in only those census blocks which are populated will address the FCC’s concerns over the simple density calculation by not considering large swaths of unserved and unpopulated areas inside a company’s study area. However, it should be noted that this proposed 27 independent variable will still fail to account for costs related to reaching across unpopulated census block to reach distant populations as well as the costs related to connecting the rural network to the Internet backbone. f. The delineation of Housing Units, Land Area, and Census Blocks by their urban categorization. The FCC does not completely explain the purpose for the delineation of Housing Units, Land Area, and Census Block independent variables by their urban categorization. Presumably this is included as an additional proxy for density. Accipiter believes there should be a more appropriate measure of density (such as loops divided by the area of populated census blocks) that would statistically fit the model and more appropriately describe similarly-situated companies. The introduction of a single alternate density measurement could simplify the model by reducing the number of independent variables required for calculation as well as reduce the margin of error imputed into the model as described above. C. The complexity of the proposed methodology and its proposed implementation fails to provide “predictability” to USF. The FCC states its proposed methodology “will inject greater predictability into the current HCLS mechanism, as companies will have more certainty of support if they manage their costs to be in alignment with their similarly-situated peers.”29 As Accipiter has demonstrated above, the model fails to identify similarly-situated peers and as such achieves the exact opposite of what the FCC has stated. Because the independent variables upon which the model is based are so poorly measured, subject to such a high margin of error, and ignore more relevant factors that drive a rural carrier’s costs, the FCC model can limit a carrier’s support for arbitrary and 29 Id. at ¶ 221. 28 illegitimate reasons.30 Rural carriers cannot be expected to serve rural customers and achieve the Congressional purpose of the USF program when they are subject to such fundamentally flawed and unpredictable results. The FCC states that the proposed methodology will be updated annually and delegates to the Wireline Competition Bureaus the authority to update the model.31 More clarity is needed as to how the Bureau would implement this authority. When the regression model is calculated anew each year based upon new input data, will certain independent variables be eliminated and/or added based upon the whims of the Bureau staff or the results of the revised model? How will census data be updated in the period between decennial censuses to account for the fact that a carrier’s costs structure can change based upon study area growth? Inexplicably, the FCC delegated authority to the Bureau without first answering critical questions relevant to the model’s predictability. Even if answers to these questions are provided, the variability in model results from year to year is likely to be unpredictable and significant. The FCC claims predictability under the proposed methodology is no different than the current method of calculating HCLS in which a carrier receives support based on where its own cost per loop falls relative to a national average – a measurement that changes from year to year.32 In fact, the national average cost per loop is predictable within a certain margin of error, and carriers are able to make investment decisions based upon this reasonable predictability and margin of error. This is not at all comparable to the FCC’s proposed methodology, which 30 It is also noteworthy that the flawed model can also fail to predict and limit a carrier’s wasteful spending. 31 FNPRM at ¶ 217. 32 Id. at ¶ 220. 29 utilizes inputs of highly inaccurate data placed into a poorly designed model which produces a formula limiting certain costs recoveries based upon arbitrary assignments of similarly-situated companies. This faulty procedure will be duplicated annually with no defined limits on what new changes in formulation may or may not be introduced. It is not at all apparent that the FCC even fully understood the design and complexities inherent in the proposed model when it authorized its pending implementation. If the FCC did understand these issues, it certainly would not suggest that predictability is no different than predicting the impact of the NACPL. The FCC has a statutory obligation to make sure that USF funding is predictable so that rural carriers can assure their consumers sustainable access to universal services. While the FCC highlights that USF is meant for the benefit of the consumer, it fails to recognize that consumers are benefited by the rural carrier that serves them. Thus insuring that a rural carrier can anticipate that it will recover its costs for serving rural customers in its assigned study area is an inseparable link to the FCC’s statutory obligations. The FCC’s arbitrary, poorly design, and overly complex regression model does not provide the necessary predictability to meet its statutory obligation. While the model requires significant redesign before it can be adequately evaluated, Accipiter notes that the FCC’s desired outcome could possibly be better achieved by adjusting the cost recovery limitations to only limit costs which are an inordinate amount above the 90th percentile (or other percentile) prediction. If one considers the theoretical possibility of a near perfect predictor model and a group of similarly-situated companies which have all made prudent network investments, establishing limits and the predicted 90th percentile will inevitably cap a number of carriers, even though all carriers may have made prudent investments. The point is that it does not matter how well carriers manage their costs, there will always be a 90th 30 percentile group of companies subject to a cap. When that 90th percentile capped group is not significantly astray of the distribution of the remaining similarly-situated carriers, it is arbitrary to cap them. The FCC should be mainly concerned about extreme outliers, and should consider imposing caps only upon carriers whose costs are unreasonable and are an inordinate amount above the quantile limits established under a more properly constructed model. D. The proposed methodology should not limit costs that are in-place due to decisions made in the past according to the former USF rules. The FCC states that the proposed methodology will “help to identify those study areas where past investments may have been excessive and caps their reimbursement”33. A carrier that made an investment to serve rural customers in the past under the standing rules of the USF program should not suffer a reduction in the revenues supporting that investment because of rule changes enacted after the investment decisions were made. Such action by the FCC is confiscatory, antithetical to the USF goal of sustaining service for rural customers, and a precedent which left unchecked will serve to create gross unpredictability for any carrier trying to serve rural America. E. Accipiter’s history and cost structure highlights a number of issues the FCC methodology inappropriately fails to consider. Accipiter’s study area cost per loop is the seventh highest among the 720 companies analyzed by the FCC. While on the surface it would seem logical that the company would be subject to certain costs limitations, a more in-depth look reveals that there are unique elements of the company’s history, study area, and growth pattern that justify Accipiter’s cost structure. Accipiter is a good example of where the FCC model fails to adequately recognize true cost 33 Id. at n. 351. 31 drivers and thus establishes limits based upon faulty criteria. Implementing the FCC’s methodology as currently proposed could result in the unnecessary loss of service to a number of Arizona homes and businesses which do not have any other options for telecommunications services. The company’s history recounted in the introduction to these comments outlines the challenges the company faced in serving one of the most sparsely populated and high-cost service territories in the nation. One item worth expanding upon is that in order to serve previously unserved high cost rural populations, Accipiter’s shareholders made equity contributions into the company to cover cash flow deficits created by the start-up operational costs and costs incurred in advance of the recovery of revenues. The FCC seems to have lost sight of the fact that the former USF rules did contain financial limitations which had to be integrated into the business decision-making process. Accipiter’s shareholders made their commitment to serve rural customers in the study area in good faith based on the understanding that the FCC would honor its statutory duty to promote and expand universal service. The FCC must consider the interplay of its statutory obligations regarding USF and the role of the program as a key aspect of Accipiter’s loan contract with the United States of America. Fulfillment of this loan contract required shareholder equity contributions and currently restricts distributions of equity back to the shareholders. The FCC cannot forgo its statutory obligations in a way which destroys Accipiter’s shareholder equity and threatens default on the company’s loan obligations. The key to understanding how the FCC’s methodology fails to recognize the factors that drive Accipiter’s costs is to compare Accipiter’s true cost drivers with the independent variables used by the FCC in the proposed limitation formula. Accipiter discusses these drivers below – most of which were previously presented to the FCC in Accipiter’s response to an FCC data 32 request which sought information regarding the drivers of Accipiter’s costs.34 Given Accipiter’s time and financial investment in responding to this data request, the company was disappointed and surprised to see that the FCC failed to address any of the company’s cost factors in the FCC’s proposed methodology. While the issues addressed below offer some options for different independent variables which could potentially improve the regression model, Accipiter reiterates that the model must be rebuilt and resubmitted for public commentary, not slightly modified and implemented. Further, not all of the cost factors below would be appropriate as independent variables and would possibly be a poor statistical fit for modeling based on the entire population of high-cost USF participants. Regardless, each cost factor discussed below should be considered as criteria by which the FCC would measure waivers of the per-line limitations imposed by the FCC’s regression model. 1. Accipiter’s study area is among the lowest density study areas in the nation. Accipiter’s study area, much like the entire state of Arizona, is similar to a checkerboard with large sections of unpopulated public lands. Thus, the populated portions of Accipiter’s study area lie in clusters that are significant distances apart. As Accipiter had previously noted to the FCC, the company serves seven geographically distinct population centers.35 The presence of public lands and widely dispersed population centers is a characteristic more prevalent in the West, and the FCC should consider this factor to avoid unfairly biasing its model 34 Letter from Patrick Sherrill, CEO, Accipiter to Sharon E. Gillett, Chief, Wireline Competition Bureau, FCC (Dec. 21, 2010) (submitted pursuant to protective order). 35 Id. 33 toward carriers in the East. Accipiter has suggested considering subscribers per route mile and loops per populated census block as more relevant cost predictors related to a carrier’s density. 2. Accipiter has incurred higher costs than the average rate-of-return carrier because of recent investments in plant to reach populations that were previously unserved. Almost 70 percent of the company’s total plant investment has been made in the past five years while expanding services into an area of rural Arizona which was previously unserved. Since the initial network investment in 1996 the company has initiated service to approximately 150 customers who did not previously have access to regulated telephone or broadband services. The newness of Accipiter’s investments required higher costs than the typical rate-of-return carrier. When extending services to unserved populations there is no existing plant to be upgraded, so the construction does not benefit from previously established rights-of-way or the leverage provided by previously constructed facilities. Also, materials and construction unit prices typically grow higher over time, so costs for newly constructed facilities will generally be higher than facilities built longer ago. Accipiter suggests the FCC develop an independent variable which predicts costs based upon age of plant. Also relevant to the age of the investment is the extent of broadband capability offered by the plant which was built. Certainly in the past five years carriers which were constructing network facilities were planning for broadband to be the primary long-term requirement of the customer base. Since plant investments can have an aggregate service life of 20 years or more, carriers should have been considering deployment of robust broadband infrastructure. Certainly this was the case for Accipiter, which has been deploying FTTH facilities in its study area since 34 2004. The costs of deploying FTTH are marginally higher than copper-based DSL36 but are justified provided that FTTH assures a long-term revenue stream for future broadband services and incurs lower costs for maintenance over the life of the plant. It would be unreasonable for the FCC to fail to consider the broadband capabilities of the existing network as a factor relevant to costs. If the network is fully capable of providing broadband service, the carrier has used the existing USF mechanisms to accomplish the new objectives of the USF program. Support should not be limited under this scenario – it should continue in order to sustain services that are already provided. Accipiter reiterates that the FCC should consider an independent variable which measures the extent of a carrier’s broadband capabilities. 3. Portions of Accipiter’s study area are extremely mountainous and rocky. Accipiter’s study area in central Arizona contains significant areas of mountainous and rocky terrain which makes network construction extraordinarily expensive. The FCC model’s assertion that percent water area serves as an adequate predictor of terrain driven costs fails to properly consider Accipiter’s terrain. The factor of mountainous terrain and lower percentages of water area are likely more prevalent in the West. If the FCC does not redesign the model to account for true terrain drivers the model will most likely show an unfavorable bias against carriers like Accipiter serving mountainous areas in the West. 4. Accipiter’s Arizona study area requires higher than normal costs to account for the regulatory environment. 36 Some would argue that copper based DSL networks can be more expensive to construct and maintain due to fluctuations in copper pricing and the all-too-frequent theft of copper facilities in some areas. 35 As mentioned above, the State of Arizona and Accipiter’s study area in particular is checkerboarded with significant areas of publicly owned and managed lands. Not only does the presence of public lands create longer routes to serve subscribers, it also requires costly permitting and right-of-way procurement. Permitting on public lands often involves environmental studies and archeological surveys which are costly to perform and usually result in restrictions and/or modifications in construction methods which, in turn, create still more costs for the project. Accipiter has completed a project in the recent past where the costs for permitting and rights-of-way exceeded the costs of construction of the project. Due to the significant cost additions that can be added for carriers that are required to construct on public lands, the FCC must consider this factor to truly measure similarly-situated companies. Again, this factor, if left unaddressed, could pose another unfair penalty on carriers serving customers in the western United States. 5. Accipiter’s costs structure is highly unusual because of the growth of population in the study area and the subscriber base. Accipiter’s study area loops have increased over 350 percent since the end of 2006. Over the same period rural carriers loop counts have declined over 15 percent. Obviously Accipiter is experiencing a much different growth dynamic than the typical rate-of-return carrier. Planning for pending short-term growth creates higher per unit costs which are quickly averaged downward once the planned growth is realized. While the FCC model considers loops and housing units as measures of cost driven by a company’s scale, the model fails to consider the dynamic and direction of costs. The normal pattern among rate-of-return carriers is negative loop growth, which means the model will fail to adequately address Accipiter’s growth rate and thus unfairly deem the company a cost outlier that should be subject to limitations. 36 6. Accipiter has a significant number of unserved households in its study area. Accipiter has documented above the company’s history and current situation with regard to subdivisions where preferred provider agreements exist.37 The market monopolization accomplished through these agreements has prevented Accipiter from expanding its network footprint to a number of homes in its study area. Accipiter assumes the standard for most study areas is that a home in the study area will be passed by the ILEC. Therefore Accipiter’s study area would be an outlier in that a significant number of homes in the higher density portion of its study area are not passed and thus no additional costs were incurred for these homes (nor did Accipiter receive the higher revenues and margins available in these areas). The FCC should consider allowing for corrections to the independent variables accordingly. Accipiter notes, however, that bizarre results could occur by submitting a correction for Accipiter’s housing units as the coefficients related to housing units are sometimes negative. This means that removing a housing unit from a company’s study area (a move that logically would reduce costs) would actually raise the limit generated by the FCC formula. Accipiter points to this bizarre result as another indicator that the FCC model is flawed and must be redesigned. F. The limitation methodology for HCLS support must be properly constructed before limitations impacting ICLS can be considered. Accipiter’s comments clearly demonstrate that the FCC has much work to do to correct the proposed methodology for HCLS cost limits. The applications of limits for ICLS support must be deferred until the FCC can bring clarity and sanity to the HCLS limitations. 37 See infra Section II.B. 37 IV. CONCLUSION There are two elements of the FNPRM that could have a grave impact on Accipiter’s operations and its ability to serve rural customers in its study area: • Eliminating support for areas with less than 100 percent overlap of an unsubsidized competitor, and • The limitations placed upon costs by the FCC’s proposed regression methodology. These two elements of the proposed rulemaking are poorly conceived, confiscatory toward prior investments, and introduce unnecessary complexity and unpredictability for carriers serving rural customers. The result of implementing these elements as proposed will be detrimental to rural consumers and poor stewardship of the universal service fund. The FCC should not impose any elimination or adjustments of universal service support for areas with less than 100 percent overlap of an unsubsidized competitor. The FCC must delay the implementation of the proposed regression methodology so the model can be reconstructed upon better data and assumptions and re-vetted through the public rulemaking process. Respectfully Submitted /s/ Patrick Sherrill Patrick Sherrill President and Chief Executive Officer Accipiter Communications Inc. January 18, 2012 EXHIBIT 1