February 25, 2016 VIA ELECTRONIC FILING Marlene H. Dortch, - - PDF document

February 25, 2016 VIA ELECTRONIC FILING Marlene H. Dortch, Secretary Federal Communications Commission 445 12th Street, S.W., Room TW-B204 Washington, DC 20554 Re: WT Docket No. 10-208 WC Docket No. 10-90 Madam Secretary: In accordance with Section 1.1206 of the Commission’s rules, 47 C.F.R. § 1.1206, we hereby provide you with notice of an oral ex parte presentation in connection with the above- captioned proceedings. On February 23, 2015, Grant Spellmeyer and undersigned counsel, on behalf of United States Cellular Corporation (“U.S. Cellular”), along with James Stegeman, Mike Wilson and Luis Rodriguez of CostQuest Associates, met with James Schlichting, Margaret Wiener, Sue McNeil, Audra Hale-Maddox, Chris Helzer, Eliot Maenner, Gary Michaels, Thuy Tran, Paroma Sanyal, Matthew Pearl, and Rita Cookmeyer. We discussed the state of the telecommunications industry, including the expansion of wireless-only households and the rapid expansion of wireless-only users among younger demographic groups and low-income households. We also introduced two papers from CostQuest Associates, one on how to measure coverage and service quality in rural areas, and another on the cost of operating and maintaining rural wireless networks. A copy of our slide presentation, along with the CostQuest papers, is enclosed. In addition, we have enclosed a copy of written testimony filed by LeRoy T. Carlson, Jr. with the

• Hon. Marlene H. Dortch

February 25, 2016 Page 2 U.S. Senate Subcommittee on Communications, Technology, Innovation, and the Internet on February 4, 2016, providing U.S. Cellular’s position on, among other things, the need for the FCC to increase its commitment to mobile broadband networks in rural America. Respectfully submitted, David LaFuria Counsel for United States Cellular Corporation cc: James Schlichting Margaret Wiener Gary Michaels Sue McNeil Audra Hale-Maddox Eliot Maenner Chris Helzer Thuy Tran Paroma Sanyal Matthew Pearl Rita Cookmeyer James Stegeman Mike Wilson Luis Rodriguez

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Universal Service Reform – Mobility Fund

(WT Docket No. 10-208)

Connect America Fund

(WC Docket No. 10-90)

Ex Parte Presentation of United States Cellular Corporation

February 23, 2016

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

U.S. Cellular

• Wireless service in nearly 200 markets across 24 states
• Majority of geography served is rural – large regional

clusters

• An ETC in 14 states.
• Success in serving many rural/remote areas largely

attributable to the federal USF program.

2

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Benefits of Wireless

• Health and Safety, E-911, FirstNet.
• Economic Development, IoT, Agriculture, Rural Employment.
• Multiplier Effects – one wireless job supports six more; each dollar

invested results in $2.32 of economic activity.

• Attracts and keeps talented people in rural areas.
• Benefits flowing from 4G/5G available only in areas having access to

high-quality mobile broadband coverage.

3

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Wireless-only households are now the norm, especially among low-income populations and the young

4

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Smartphone, But No Broadband At Home, Has Nearly Doubled In Just Two Years

5

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Wireless is the Future

6

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Data and Methodologies that Systematically Overstate Mobile Broadband Deployment Have Led Some to Conclude That the Job is Largely Done

• Government claims that 98% of Americans have access

to 4G LTE service.

• Further Notice relied on claims from largest carriers

focused on urban areas to justify 99.5% mobile broadband coverage figure.

• Advertising data provided in National Broadband Map
• verstates coverage.

7

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Data and Methodologies that Systematically Overstate Mobile Broadband Deployment Have Led Some to Conclude That the Job is Largely Done

• Coverage at a centroid point incorrectly assumes both coverage and speed

threshold are met throughout the Census Block.

• Coverage data appears to depict homogenous speeds that do not

accurately capture wide variances in throughput speed between cell tower and cell edge.

• “As the Commission has found in the past, the methodology and data used

to report this coverage has the potential to overstate that coverage. Additionally, the data do not expressly account for factors such as signal strength, bit rate, or in-building coverage, and may convey a false sense of consistency of speeds across geographic areas and service providers.” See, 2016 Broadband Progress Report, at Para 112.

8

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

CostQuest Analysis

Sources: 2014 Mobile Pulse & CostQuest Associates Analysis

9

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Percentage Of Field Tests Meeting Advertised 10/1 Throughput Speed:

Sources: 2014 Mobile Pulse & CostQuest Associates Analysis

State Measurement Unit Carrier X Mobile Pulse Carrier Z Mobile Pulse SD Roads 23.2% 57.1% SD Population 31.5% 58.9% SD Business Firms 28.8% 63.0% CO Roads 6.8% 18.1% CO Population 4.9% 18.6% CO Business Firms 5.8% 17.5% ID Roads 27.6% 31.7% ID Population 27.7% 24.5% ID Business Firms 24.6% 25.4% WI Roads 55.3% 60.6% WI Population 59.1% 55.4% WI Business Firms 62.5% 56.1% WY Roads 11.6% 35.4% WY Population 10.1% 25.6% WY Business Firms 8.9% 23.7% TOTAL Roads 33.0% 44.5% TOTAL Population 34.3% 32.3% TOTAL Business Firms 31.9% 32.3%

10

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Comparing the National Broadband Map to a Drive Test Yields Very Different Results

Source: 2014 CostQuest Analysis of National Broadband Map and Mobile Pulse Data

11

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

These Findings Should Give the Commission Pause – Using Some

• f the Drive Test Resources Deployed for Auction 901 Verification

to Audit the NBM Would Be a Useful Exercise.

Source: 2014 CostQuest Analysis of National Broadband Map and Mobile Pulse Data

12

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

13

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

CostQuest O&M Paper

• At a typical cell site, achieving positive cash flow by year five

requires approximately 900 subscribers at $56.00 of ARPU.

• If ARPU is reduced by just $4.00, positive cash flow is delayed

until year ten.

• A cell site needs at least 250 subscribers to cover opex.
• Roaming is no longer a stable revenue source.

14

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

CostQuest O&M Paper

15

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Other Data Reveal Substantial Portions of Rural America Lack Access.

• 18th Mobile Competition Report:

– 25% of road miles and 50% of square miles in the US do not have coverage by two or more carriers. – As of July 2015, while more than 90 percent of the U.S. population lived in census blocks with coverage by at least four providers, these census blocks accounted for only approximately 31 percent of the total land area of the United States, and approximately 55 percent of U.S. road miles.

• 2016 Broadband Progress Report: 87% of rural Americans lack access to mobile

broadband at 10 Mbps/1 Mbps. 16

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

CDMA/GSM Incompatibility is a Public Safety Issue

• 25% of road miles and 50% of square miles lack coverage by both

GSM and CDMA networks (Eighteenth Mobile Competition Report).

• A person with a CDMA-only phone cannot complete a call when

they are in an area served only by GSM, and vice-versa. As a result, the current reality in rural areas is a patchwork quilt of coverage by incompatible technologies, frustrating the goal of seamless access.

• For public safety, it is critical that rural Americans have access to

wireless networks capable of connecting both kinds of devices, just as those who live in cities do.

17

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Statutory Directive: To Provide Rural Citizens With Access To Services That Are Reasonably Comparable In Price And Quality To Those Provided In Urban Areas.

• The FCC must first learn how much area must be served, how much

speeds need to be improved in such areas, and what it will cost.

• Only then can the program be properly sized and an annual budget

determined.

• The NBM and subsequent 477 data do not provide the FCC with accurate

data that allows reasonable conclusions as to what needs to be done and what it will cost to make services in rural areas reasonably comparable.

• Sizing the program first (e.g., $400 M), without even estimating the annual

cost of providing reasonably comparable access, contravenes Congressional directive to take meaningful action to close the divide.

18

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Statutory Directive: To Provide Rural Citizens With Access To Services That Are Reasonably Comparable In Price And In Quality To Those Provided In Urban Areas.

• The FCC has already invested more than $50 Billion in wireline broadband,

with plans for, $20 Billion more over the next five years, compared to approximately $12/$2.5 Billion for wireless.

19

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

CAF Phase II Must be Distributed in a Competitively Neutral Manner

• Applicants capable of meeting the CAF Phase I

requirements must be able to access CAF II support on a level playing field.

• Proposals for tiered bidding that prefer fiber over more

efficient technologies disserve many without access, reduce marketplace competitors, and must be rejected.

• Section 214 authorizes participation of only qualified

common carrier ETCs.

20

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Alternatives for Mobility Fund Phase II

• MFII support should focus more on roads and agricultural areas, to deliver

the benefits of mobility, rather than focusing on coverage at residences.

• This is consistent with the FCC’s finding that fixed and mobile broadband

are complimentary services.

• Rather than conduct a reverse auction, which proved to be an inefficient

means of distributing support in Auction 901, the Commission should consider alternatives that increase targeted investment and leverage program funds, and seek comment on further alternative methods of distributing support.

• Accountability measures, including demonstrations of how support is used

consistent with the statute, must be a part of any MF program.

21

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Roadmap for Mobility Fund Phase II

• Seek further comment on, (a) the current state of mobile broadband

deployment in rural America, and (b) additional methods of obtaining better data on the coverage, quality and performance characteristics of mobile broadband.

• Declare now, and in the next 706 proceeding, that any area lacking actual

access to 10/1 speed is unserved by mobile broadband. This will harmonize MFII and 706 program goals.

• Seek comment on how to configure MFII to drive deployment of

ubiquitous 10/1 speeds across rural America, subject to the funding constraints currently placed upon the program.

22

8300 Greensboro Drive Suite 1200 McLean, VA 22102 www.fcclaw.com (703) 584-8678

Alternative Proposal: Federal-State Broadband Grant Program

• Set up a grant program for carriers to apply for infrastructure funds, anywhere in

rural areas that require investment.

• Provide a “carrot” to incentivize states to invest and create program leverage.1
• Let carriers identify areas of poor coverage to state regulators, using a single

streamlined federal rule for making awards.

• FCC allocates funds among states and ensures program accountability to Congress.
• We have attached an explanatory summary and a proposed rule for discussion.

1 “Nevertheless, the FCC may not simply assume that the states will act on their own to preserve and advance universal service. It remains

• bligated to create some inducement—a “carrot” or a “stick,” for example, or simply a binding cooperative agreement with the states—for

the states to assist in implementing the goals of universal service. For example, the FCC might condition a state's receipt of federal funds upon the development of an adequate state program, an approach the FCC at oral argument conceded was possible.” Qwest Corp. v. FCC, 258 F.3d 1191, 1204 (10th Cir., 2001).

23

1 Proposal for a State Block Grant Program for Mobility Fund Phase II Summary The FCC will:

• Invite states to accept a block grant of Mobility Fund Phase II support, using a

formula to target support to areas requiring additional investment.

• Incentivize states to add funding from state universal service mechanisms.
• Provide a Mobility Phase II rule, to be implemented by states.

Because state public utility commissions are closer to consumers, they are more likely to target support to areas that need investment, and more responsive to the needs of rural citizens. Allocation Among the States There will be three separate allocations: Lower 48, excluding Tribal Areas Alaska, Hawaii, and other Insular Areas Tribal Areas in the Lower 48 Allocation method: Allocate to census blocks unserved by 10/1 service (from the 2015 Broadband Progress Report). Areas should not be disqualified due to claimed coverage by any carrier unless supported by drive test data similar to that required of Auction 901/902 participants. Incentive for States to Participate The FCC will provide matching funds of up to 50% to states willing to contribute to a mobility fund. For example, if the FCC allocates $1.00 and a state contributes $0.50, the FCC will increase the federal funding amount to $1.50, making $2.00 available for the state mobility fund. The FCC will provide a mandatory mobile broadband rule for states to implement. No new rulemaking needed, only program administration and oversight.

2 States declining to participate in Mobility Fund II receive federal funding through a reverse auction using Auction 901 procedures, but would not receive matching funds. Unused or forfeited funds will be carried forward and used to increase the following year’s budget. Program Structure Recipients must be designated as ETCs pursuant to Section 214 of the Act. The application for funding must be self-scoring to simplify review. States will review applications, verify compliance, certify (and decertify) ETCs. Funds must be used for capital expenditures to build or upgrade facilities, or to fund

• ngoing operations.

Facilities must be used to provide mobile broadband service consistent with then- existing FCC requirements. Program size: $500 Million in annual support, plus up to $500 million set aside for

• perating expenses for legacy ETCs demonstrating need-based support for operating

expenses of existing facilities, plus up to $250 million of matching support for states

• pting in.

Possible state matching funds: $250 Million. Total possible program leverage: $750 Million. Likely federal program size: Less than $1 Billion. To ensure certainty of funding, legacy high-cost support must continue without phase- down until first date that funding awards are made to new program awardees.

MOBILE BROADBAND PROGRAM PROPOSED FEDERAL RULE

• A. Designated ETCs may apply to the state commission for grants to fund the construction of

facilities capable of mobile broadband service, to areas unserved or underserved by broadband in the state.

• B. A project must meet the following requirements to be eligible for a grant award:

(1) Support mobile broadband service at speeds and service levels consistent with FCC’s federal universal service requirements to households and businesses in the proposed project area. (2) Support access to emergency 911 services.

• C. Contents of grant applications. An application for support shall include:

(1) A proposal to build or upgrade facilities to serve an area where the applicant is designated as an ETC; (2) A detailed build plan setting forth a description of the facilities to be deployed, including all costs of constructing facilities; (3) A map showing where service and/or coverage will be provided; this requirement can be met by providing a coverage map generated using a radio frequency propagation tool generally used in the wireless industry; (4) An estimate of the number of road miles and square miles to be covered; (5) The amount of support requested; (6) A description of the technology to be deployed, including data throughput speeds to be delivered to customers; (7) A demonstration that the area to be served is an area unserved by broadband or an area underserved by broadband; and (8) An exhibit providing the proposed application score, including supporting documentation.

2

• D. Applicants shall self-score their applications. Points will be awarded as follows:

(1) Serving roads unserved by any other commercial mobile wireless carrier - __ points per mile. (2) Serving roads unserved by another commercial mobile wireless carrier using compatible technology - ___ points per mile. (3) Serving areas unserved by any other commercial mobile wireless carrier - __ points per square mile. (4) Serving areas unserved by another commercial mobile wireless carrier using compatible technology - __ points per square mile. (5) Serving roads at a cost of less than $___ per mile - __ points. (6) Serving proposed grant area at cost of less than $___ per square mile - __ points.

• E. The applicant must make the following commitments:

(1) To offer service at reasonably comparable rates for comparable services in urban areas. (2) To provide service for at least five years following project completion. (3) To complete funded projects within two years from the date of commission approval; (4) To provide an annual progress report to ensure that all grant funds are being used for the purpose intended. (5) To respond to commission inquiries regarding service-related complaints and commit to resolve service-related complaints in a reasonable manner. (6) To allow collocation on reasonable terms by other providers of commercial mobile wireless service or any public safety network, and to abide by the FCC's collocation requirements for awardees under the federal universal service program.

• F. Procedure for awarding support from the broadband fund:

(1) On or before March 1 of each year, the commission shall open a thirty (30) day window for filing applications for broadband program support for the following calendar year.

3 (2) The commission shall review and score all qualified applications. (3) On or before July 1, the commission shall make initial awards consistent with funds available. (4) On or before September 1, the commission shall make final awards and submit them to the FCC for funding. (5) On or before December 31, the FCC shall confirm awards and announce its intention to disburse funds, or notify states of any awards that it will not fund and the reasons therefor. (6) The FCC and state commissions shall, within sixty (60) days, resolve any disagreements concerning funding commitments.

• G. Conditions for disbursement of awarded funds:

(1) Within thirty (30) days after project completion, the awardee shall submit a report demonstrating that the project as completed meets the coverage requirements set forth in the application, including a certification from an officer or director that all program requirements have been met. (2) The administrator shall disburse 50% of the award when made, and 50% of the award forty-five (45) days after submission of an acceptable project completion report under subparagraph (1) above. The commission may, within thirty (30) days after submission, suspend payment by the administrator and order additional information to be provided. (3) Any applicant found to have willfully misrepresented information in an application is found to have used support unlawfully or fails to meet the commitments set forth in the application, shall refund all award funds immediately and shall be subject to having its ETC designation revoked.

• 1
• UnitedStatesCellularCorp.

OngoingSupportforOperations

&MaintenanceofRuralMobileNetworks

February22,2016

• CostQuestAssociates(CQA)EconomicResearch&Analysis
• CostQuest Associates (CQA)

Economic Research & Analysis

• For further information or if

there are any questions or concerns contact: research@costquest.com

• 2
• TableofContents

Introduction..................................................................................................................................................3 ConsumerExperienceandtheImportanceofMaintainingNetworks.........................................................3 FinancialsofaRuralMobileNetwork...........................................................................................................4 FinancialComponents ...............................................................................................................................4 DifferencesinCostsofMaintainingRuralNetworksversusTypicalNetworks........................................5 TheBusinessCase.....................................................................................................................................6 BusinessCaseResults ............................................................................................................................7 EconomicViabilityandFundSizing .............................................................................................................10

• 3
• Introduction

CostQuestpresentsthisbriefanalyticspaper,examiningthecoststructureofaruralwirelesscarrier,to informtherecordwithdataandfindingsthatcanprovideafoundationforaneconomicallyrational fundingprogramforFCC’sMobilityFundPhaseII.CostQuest’sprimarygoalistodevelop,throughaseries

• finternal,governmentbased,andproprietarydatasources,ameanstoquantifythefullfinancialpicture
• fmobilewirelessserviceinruralareas.

WhetherMobilityFundPhaseIIsupportisdistributedviaamodel,areverseauctionasusedinAuction 901,oracombinationofboth,theprogramwillrelyonbotha“cost/financialmodel”anda“support model,”eitherexplicitorimplicit.Inamodeldistributionprogram,suchastheCACMbeingusedfor PriceCaplandlinecarriers,thecost/financialmodelidentifiesareasrequiringfundingandthepotential fundingrequirements.Indeterminingthemodeldistribution,theCACMsupportmodeltakesthecosts andthenappliespolicydecisions,suchasnotfundingcompetitiveareas,toarriveatthefinal disbursementamounts.Inareverseauction,thecost/financialmodelcanhelpidentifythepotentialsize

• fthefund,thetargetedareas,andpotentiallythereserveprice.Additionally,forsuccessfulbidders,a

cost/financialmodeltypicallyformsthebackboneofthebidssubmitted.Onthesupportsideforreverse auctions,thesupportmodelcapturespolicydecisionsthatcoverwhoiseligible,whatareasareeligible, andtheregulatoryrequirementsimposedonfundrecipients. Nomatterhowtheultimatefundissetup,understandingthecost/financialpictureofruralarea deploymentisanimportantfirststeptodevelopingasuccessfulMobilityFundPhaseIIprogram.As such,weprovideadetailedviewofthefinancialpicturetoservearuralareawithmobilewirelessservice.

ConsumerExperienceandtheImportanceofMaintainingNetworks

Forwirelesscarriers(justasforanytelecommunicationprovider),theconsumerexperiencedetermines thecompanies’ultimatelevelofsuccess.Inthisworldofmobility,consumersatisfactionisdriven primarilybytheconsistency,quality,andreliabilityofvoiceanddatacoverageandthroughput. Maintenanceandoperation(collectivelyreferredtoas“Opex”)ofthenetwork(e.g.,cellsites,switching, backhaul)isthereforeparamounttotheconsumerexperienceandultimatelytheprovider’slevelof success.Inmoreheavilypopulatedareas,opexislesscostlypersubscriberandoftentimesoffsetbythe revenuesreceivedfromthelargeravailablecustomerbase(i.e.,greaterpopulationdensity),althoughcell siterentals,backhaulandmaintenancecostscanbehigher.Inmorerurallocations,whereopexcostper subscriberisoftenprohibitive,carriersmayreacheconomicviabilityusingexplicitsubsidies(universal servicesupport)orimplicitsubsidies(transferringfundsfromprofitabletounprofitableareas) (collectively,“Subsidies”).Insuchareas,insufficientfundingcausescarrierstocutbackonthecell density,performancemetrics,upgrades,functionalityandmaintenance,allofwhichareneededtomake ruralnetworksreasonablycomparableinqualitytothoseinurbanareas.Allofthesefactorsmeanthe consumerexperiencecanbequitedifferent,especiallyifSubsidiesaremissingand/orreduced.Inour

• 4
• view,expandingthepublicrecordonthefinancesofruraldeploymentiscriticaltomakingtheoptimal

policychoices.

FinancialsofaRuralMobileNetwork

ThroughtheMobilityFundPhaseIAuction,theFCCclearlyrecognizedinitialdeploymentcostsasa hurdletogettingmobileservicetoruralareas.ThesecostsweresupportedthroughtheAuction’sone timeinfusionoffunds–tothetuneof$300milliontoserve12.9%1oftheeligibleunservedroadmilesin theU.S.Deploymentoffacilitiesinthesesparselypopulated,variedterrainareasisafinancialchallenge, andmayhaveneverhappenedwithoutfundingsupport.Thatbeingsaid,perhapsmoreofaneconomic challengeisoperatingandmaintainingthosenetworksoverlongperiodsoftime. Inthefollowing,weexplorethefinancialpictureofdeployingandmaintainingmobilewirelessservicein ruralareas.

FinancialComponents

Asafirststepinidentifyingareasofthecountrythatmayrequiresometypeofexternalfunding(e.g., ConnectAmericaFund)tomakemobilenetworkdeploymentviable,itisnecessarytounderstandthe costofmobilenetworkdeployments.Forthepurposesofthispaper,we’llfocusonthreekeycost categories:InitialInvestment,OperationsandMaintenance(hereafterreferredtoas“Opex”),and Maintenance/ReplacementInvestment(hereafterreferredtoas“MaintenanceCapex”).This MaintenanceCapexcategoryofcostisoftenoverlooked,butmustbeconsideredwhenattemptingto identifytheforwardlookingcostsandthefundingneededtosupportthosecosts. InitialCapexcapturestheupfrontcapitalexpensesrequiredtobuildcellsites,increaseswitchingcapacity andtheedgecorethataggregatesandmanagestraffic,providesfunctionalitiestotheenduser,facilitates callcompletionanddatapackethandling,routingandbilling,andmore.Thecellsiteswillinclude: towers,antennas,amplifiers,cabling,power,RANelectronics,shelters,transportelectronics,andother miscellaneousgeardeployedtomeetexpectedtrafficdemandsanduserneeds. Opexwillincludethecosttorunandmaintainthecellsites,edgecoreandnetwork,managetraffic, supportthecustomer,operatestores,handlebilling,andotheraspectsofrunningamobilitycompany. ThetypicalOpexcostsforamobilecarriercanbebrokendownasfollows:CustomerOperations (CustOps),RoamingExpense(Roaming),BadDebt,SalesandMarketing,Advertising,NetworkOperations

• 1Onecouldinferfromthisthatitwouldhavetakenatleast$2.3Billiontoserve100%oftheeligibleunservedroad

miles.However,thiswouldappeartobeaconservativelowerboundsincethe$300Millionwasawardedtothe lowercostareas.Keepinmind,thisvalueissimplythecosttoexpandservicetocurrentlyunservedareasanddoes notincludethesupportthatmayberequiredincurrentlyservedareasthatweretheresultofpriorUSFfunding.

• 5
• (NetOp)2,CellSiteOperations(CellSiteOp)3,HandsetSubsidies,andGeneralandAdministrativecosts

(G&A)whichcanincludebilling,ITdepartments,HR,etc. MaintenanceCapexcapturestheongoingcapitalinvestmentsrequiredtomaintainaviablemobile network.Overtime,newtechnologiesandtechniquestomanagethenetwork(e.g.,SoftwareDefined Networks)emerge,anddemandchanges(e.g.,newdevices,newapplications,anddramaticgrowthin datatrafficoverthelastnumberofyears).Continualchangesrequireongoingnetworkinvestments, coveringitemssuchascarrieraugmentation,resectorization,deploymentofeachsucceedingnew generationofmobiletechnology(e.g.,HSPA,HSPA+,LTE,LTEAdvanced),powerbackupsystems,shiftsto Ethernetbackhaul,andmore.Inaddition,aswithanycapitaldeployment,thereareequipmentfailures requiringcapitalforreplacements. Allthreeofthesecostcategories,InitialCapex,Opex,andMaintenanceCapex,mustbeconsideredbya carrierastheyplanforinitialdeployment,maintainingdeployment,and/orexpandingdeployment. Additionally,carriersmustconsidertheoftenprohibitivelyhighcostoffiberforbackhaulinruralareas.

DifferencesinCostsofMaintainingRuralNetworksversusTypicalNetworks

Operationalandmaintenancecostsforwirelesscarriersinruralareasobviouslycanvarygreatlyfromthe costsinmetroareas,justastheydoforwirelinecarriers.Intuitively,itismoreexpensivetoprovide service,onapersubscriberbasis,inlessdenselypopulatedareas.Maintainingruralsitesmayalsobe moreexpensiveasmanyoftheutilityandotherservicesandamenitiesarenotasreadilyavailableasthey areinmetropolitanareas.Ruralsitesareofteninremoteareasrequiringmoretraveltimetoreach,via unpavedroadsandvaryingtopography;but,theyrequiremaintenanceneverthelessandoftentimesthis expenseisnoteconomicallyfeasibleorcommerciallyviablewithouttheaidofSubsidyfundinggiventhe lownumberofsubscribersusingtheruralsite.Ruralsitesoftenrequiremoreexpensivebackhaul arrangements(higherleasedlinecostsoramicrowaveinvestmentamortization).Also,whilecollocation atmostruralsitesshouldcostless,itismorelikelyinthemostruralareasand,certainly,theunserved ruralareasthattherewillbefew,ornoopportunitiestocolocateonexistingstructures(likelyresultingin increasedtowercapexamortization). UniversalServiceFundinghashistoricallysupportedserviceareasthatarenotcommercially/economically viable.However,commercialviability,ortheantithesisofit,canbedifficulttoquantifybecauselegacy mobilityfundingiscloudedwithinterandintracompanycrosssubsidization.Thismakesthejobof isolatingtruecommercialviabilityinruralareasadifficultone. Itisclearthathighinitialinvestmentisnottheonlybarrierthatkeepscarriersfromservinghighcostrural markets.Carrierslookatthetotalexpectednetcashflows,includingmaintenanceandotheroperating

• 2NetworkOperationscostsincludecallandpackethandling,911facilitationandothercostsassociatedwiththe

edgecoreandnationalnetwork.

3SiteOperationscostsincludetowerlease,backhaul,utilitiesandcellsitemaintenance.

• 6
• expenses,overanextendedtimeperiodindecidingwhethertoextend/continueserviceinagivenarea.

ThelessdenseareaswithinmarketsareoftenleftunservedorunderservedbecauseoftheongoingOPEX asmuchasforthehighhurdleofInitialCapex.IfUniversalServicefunding,crosssubsidizationorother sourcesofsupportwerenotavailable,whatareaswouldbedeemedviablemarketsforarational executiveofamobilecarrierservingruralareas?

TheBusinessCase

ToexaminethefinancialpictureofdeployingserviceinruralAmerica,CostQuestconstructedafinancial modelusingafiveyearcashflowbusinesscaseanalysis.Themodellookedattheannualcashflowin(via revenues)andtheannualcashflowout(viathethreetypeofcostsdiscussedabove)overafiveyear period.Thecashflowineachyearwasthendiscountedbacktotoday’sdollarvalue,whatistypically referredtoasthenetpresentvalueorNPV.ApositiveNPVindicateseconomicviabilitywhileanegative NPVindicatestheopposite. Inthemodel,inputsforinitialcapex,opex,andmaintenancecapexwerederivedfromCostQuest’s knowledgeoftheindustry,proprietaryfinancialdata,publicfinancialinformation,andothersources. Togglesweresettoprovidethediscountrate,assumedtakeratesandARPU(AverageRevenuePer User4),andthetermofthestudy.Finally,populationcountswithintheservingfootprintoftypicalcell siteswerevariedtoprovideanindepthunderstandingoftheeconomicallyviablebreakpointsfor deployment. Insettingupthebusinesscase,weusedthefollowinginputs:

• MonthlyARPU:
• $56.21
• ValueexcludesUSFandETCderivedrevenue
• Valueincludesroamingrevenue
• MonthlySubscriberoperationalcosts:

$35.68

• MonthlyHandsetSubsidyandInventorycosts: $9.24
• MonthlyCellSiteoperationalcosts:
• $4,662.44
• Initialcellsitecapex:
• $290,000.00
• Initialcorecapex:
• $145,000.00(unitizedtopercellsite)
• Maintenancecapex:
• 42%ofinitialcapexover5years
• 4Whilethereareanumberofwaystoattributerevenuetocellsites,forthemodelingpurposesinthispaperwe

attributedtherevenueofuserstotheirhomebasecellsite.

• 7
• BusinessCaseResults

1,000SubscriberSnapshot Fromthefiveyearbusinesscasemodel,anexampleofthecarriercostbreakdownforacellsite supporting1,000subscribersisshowninthechartbelow.5Thecomponentsshowninredrepresentcosts associatedwithdeployingandmaintainingthecellsite.

• Whilemuchofthecostdisplayedaboveisdrivenbysubscribercounts(i.e.,volumesensitiveto

subscribers),asizeableportionofthecosts,thosenotedinredtext,arefixedcostsassociatedwiththe cellsite.Considerthecellsite,thebackhaulcosts,thepowerconsumption,andcellsitemaintenance. Allofthesecostsarerelativelyfixedoncethecellsiteisinplace.Aslongasthecarriercan“amortize” thesefixedcostsoverenoughsubscribers,thesite,oracollectionofsitesinacommunityoreventhe company,canbeeconomicallyviable.TheissuethatarisesinruralAmericaisthatthesubscribercounts

• 5Thesecostsarebasedona5yearbusinesscaseforatypicalsitesupporting1,000subscribers.

CustOps 4% Roaming 6% BadDebt 4% G&A 12% SalesandMarketing 12% Advertising 5% NetOps 7% HandsetSubsidy 13%

MTCECapex 5% InitialCapexat Core 8% IntialCapexatCell 16% CellSiteOps 8%

TypicalRuralCellSiteFinancials AtCellSiteSupporting1,000Subscribers

• 8
• servedbyatypicalcellsitecanbeinsufficientforacarriertodefraythesefixedcosts.Assuch,the

expectedorrealizedfinanciallossmaywarrantsubsidizationthroughgovernmentfundingprograms.6 ImpactofSupportedSubscriberCounts Usingthefinancialmodel,wecanexaminetheimpactofthefixedcostsofcellsitedeploymentand maintenance.Inthechartbelow,thecashflowanalysisshowsthereisaclearpersubscribercost advantagetoplacingsitesindenseareas,notjustforthebenefitinthemultiplesorARPU,butforthe spreadingoftheinitialcellsiteandswitchcapexcostsacrossmanysubscribers.Italsohighlightsthe impactofCellSiteoperationscostsandMaintenanceCapex,whicharethe2ndand3rdhighestcostsfor siteswith250subscribers,nexttotheinitialcapexdeployment.Whilethisrecognitionofdensityissues couldgowithoutsaying,wefeelit’simportanttoinformtherecordastotheextentofthedifferences betweentheassumedsubscriberdensitytiers.

• EconomicViability

Asecondstepinthebusinesscaseanalysis,andperhapsmoreinformative,iswhentheCashFlow AnalysisisdrawnouttoitsconclusiontoseetheNetPresentValueoftheexpectedcashflowsoverthe

• 6Inourfinancialmodel,werecognizebothroamingrevenueandroamingexpense.However,ourvaluesrepresent

anaverageanddonotcapturetheearlyindustryphenomenonofdeployingsitessolelyfortheroamingrevenue. Withindustryconsolidationthesetypesofsiteshavebeenlargelyabsorbedintothelargercarriers. CellSiteNetOps MaintenanceCapex

• 9
• lengthoftheexpectedbusinesscase.7Thepointatwhichatypicalmobilecarrierbreaksevenonthe

scaleofisperhapsthemostsalientfigure. Ouranalysis,usingthesameinputs,showsthat,inordertoseeapositivepresentvalueofnetcashflow

• natoweroverthefiveyearperiod,thetypicalcarrierwouldneedtohave,onaverage,approximately

900activesubscribersassociatedwiththecellsite.YoucanalsoseethatthemonthlycashflowOpex breakeven(definedasmonthlylevelizedrevenueminusthemonthlylevelizedoperationalcosts)isclose to275subscribersandthemonthlycashflowOpexplusMaintenanceCapexbreakeven(definedas monthlylevelizedrevenueminusboththemonthlylevelizedoperationalcostsandthemaintenance capex)isaround450subscribers.Notethattheanalysisassumestheactivesubscribercountsrepresent thelevelizedvalueoverthefiveyearbusinesscase.8

• Thechartaboveisbasedonastaticsetofinputassumptionsnotedearlier.Itisimportanttounderstand

theimpactofchangesinassumptions.Lookingatsensitivities,ifweadjusttheARPUdownjust6.5%,to $52.50(potentiallythevalueselectedasthebenchmark,ortheresultofcompetitionontheARPUvalues intothefuture),weseethepicturechangesquiteabit.ThefiveyearbusinesscaseNPVbreakevenjumps

• 7Weuse5yearsinthispapertolineupwiththeexpecteddurationoftheMobilityPhaseIIfunding.

8Fornewsites,subscriberadoptiontypicallyincreasesovertime.Assuch,the5year,endperiodcountofactive

subscriberswillbehigherthanthelevelizedvaluepresentedabove. 5Year Break EvenPoint OpexBreak EvenPoint Opex+Mtce BreakEvenPoint

• 10
• toaround1,200averageactivesubscriberspercellsite,theOpexbreakevenincreasesto~350

subscribers,andtheOpexplusMaintenanceCapexbreakevenincreasesto~550subscribers.

• Akeydriverinthebusinesscasenotedaboveisthecountofsubscribersassociatedwithacellsite.In

effect,thegreaterthenumberofsubscribersacarriercansignupinanareainwhichthesubscriberslive, workandtravel,themoreviableacellsite(orcommunity)deploymentinthatareabecomes.Assuch, theeffectivetakerateacarriercanachieveinanareaiskeytothebusinesscase,andultimatelytothe determinationofwherefundingisrequiredinordertomaketheareaeconomicallyviabletoacarrier.

EconomicViabilityandFundSizing

Asshownaboveinthefinancialmodeling,theviabilityofanareaisdrivenbythenumberofsubscribers

• necanobtain,theexpectedARPU,theInitialCapex,theopex,andtheMaintenanceCapex.Insimple

terms,whentheCEOofamobilecompanylookstoexpand/maintainserviceinanarea,he/shewill reviewthebusinesscase.Ifthebusinesscasedoesnotwarrantsufficientreturns(withsolid assumptions),thentheserviceareawillnotbeexpandedor,conversely,servicewillbeterminated.And, aspricingcompetitionheatsup,thepressuretounderstandwheretheeconomicallyunviableareasare, andhowtoaddressthemwillbecomevital.

Opex+Mtce BreakEvenPoint OpexBreak EvenPoint 5YearBreak EvenPoint

1 | P a g e

Mobile Voice and Broadband Coverage: An analysis of sources, measures and

reporting methods

December 19, 2014

CostQuest Associates (CQA) Economic Research & Analysis

CostQuest Associates (CQA) Economic Research & Analysis

For further information or if there are any questions or concerns contact: research@costquest.com

2 | P a g e

Table of Contents

Executive Summary ....................................................................................................................... 4 Introduction ................................................................................................................................... 6 Section 1: Performance Measures for the Goal of Universal Coverage of Mobile Voice and Broadband ..................................................................................................................................... 8 Carrier Obligation Requirements ................................................................................................ 9 Subscriber Experience Measures ................................................................................................. 9 Network Service Quality Measures ........................................................................................... 11 Mobile Voice and Broadband Service Availability ..................................................................... 12 Presence ............................................................................................................................... 13 Geographic Unit of Measure ................................................................................................. 14 Geographic Area to Determine Service Availability Within ................................................... 15 Summary of Coverage Criteria: ................................................................................................. 16 Section 2: Sources of Voice and Broadband Coverage Information ................................................ 17 Voice and Broadband Service Coverage Data Sources ............................................................... 17 Mosaik Solutions (Mosaik) ................................................................................................... 18 National Broadband Map (NBM) .......................................................................................... 19 Mobile Pulse ......................................................................................................................... 21 FCC Form 477 Data ............................................................................................................. 22 Coverage Availability Source Summary ................................................................................ 23 Comparison of coverage sources and measures for sample states ............................................... 24 Analytical Data Creation/Source/Methodology .................................................................... 24 Review of Potential Coverage Measures and Currently Available Data .................................. 42 Conclusion: .................................................................................................................................. 44 Addendum A ............................................................................................................................... 47

3 | P a g e

Table of Figures

Figure 1: Hierarchy of Coverage Summary ..................................................................................... 9 Figure 2: South Dakota Verizon Retail Outlet Trade Area Analysis .............................................. 10 Figure 3: South Dakota AT&T Retail Outlet Trade Area Analysis ................................................ 11 Figure 4: Basis for the Performance Measurement Framework ...................................................... 16 Figure 5-Mosaik Custom Marketing Map; from http://www.mosaik.com/showcase/portfolio/ ... 18 Figure 6: Sample Map from the National Broadband Map web site (25Mbps availability). ............. 19 Figure 7: Summarized Sources of Service Coverage ...................................................................... 24 Figure 8: South Dakota AT&T Wireless Coverage Comparison .................................................... 26 Figure 9: South Dakota AT&T Wireless Coverage Comparison .................................................... 27 Figure 10: Tabular Coverage Data ................................................................................................ 27 Figure 11: Impact of NBM Coverage Stats using CB .csv versus .shp Files .................................... 29 Figure 12: South Dakota Mobile Pulse Maximum Download Speed ............................................. 30 Figure 13: Road to Population Coverage ...................................................................................... 31 Figure 14: National Broadband Coverage Results ......................................................................... 32 Figure 15: Carrier X NBM Coverage ............................................................................................ 33 Figure 16: Carrier Z NBM Coverage ............................................................................................ 34 Figure 17: Road Coverage by Density ........................................................................................... 35 Figure 18: Road Coverage by Type of Road .................................................................................. 36 Figure 19: Estimated Mobile Wireless Broadband Coverage by Census Block ............................... 37 Figure 20: LTE Handset Mobile Pulse Test Results for Top 2 National Carriers ............................ 38 Figure 21: Percentage of field tests conducted with LTE capable handsets performed in the NBM LTE footprints of AT&T and Verizon .......................................................................................... 39 Figure 22: Carrier X Pop Coverage NBM vs Drive Test ................................................................ 40 Figure 23: Carrier Z Pop Coverage NBM vs Drive Test ................................................................ 41 Figure 24: LTE Handsets Mobile Pulse Test Results Top 2 National Carriers Successful Test: 768kb Download/200kb Upload ............................................................................................................ 42 Figure 25: Coverage Measures Table ............................................................................................ 43 Figure 26- Coverage Definition Image .......................................................................................... 45 Figure 27: Coverage Measures Table ............................................................................................ 46

4 | P a g e

Executive Summary

Investment in our nation’s mobile wireless ecosystem continues to drive benefits to our citizens, from public safety, to health care, to education, to economic development to business and personal

• efficiencies. It is critical to our nation’s well-being that every citizen have access to high-quality mobile

voice coverage, as well as mobile broadband services. Much has been said about the growth of mobile coverage. Indeed, carrier advertising maps make clear that LTE service is available in more places every year. However, while the advertised maps appear to represent near uniform coverage, what a subscriber experiences at a given location, at a given point in time may differ from this advertised uniformity. Sometimes coverage gaps may be short lived. Other times they may be stable and an artifact of an underlying network issue. As we debate mobile voice and broadband coverage, it is critical to understand the relevance of advertising coverage maps relative to what a subscriber actually experiences. Congress expressed a desire for Americans living in rural and high-cost areas to have access to facilities and services that are reasonably comparable to those available in urban areas. In order to determine whether this goal is being met for mobile services, it is important for policy makers to come to agreement on how to measure whether the benefits of mobility and mobile broadband are available where all Americans live, work and travel. Coverage One needs to define what coverage means before determining if there is a presence of coverage at a specific location. What determines presence of coverage? What is the measure of presence? What is the geographic unit for measuring presence? In what geographic area are you determining successful presence? Presence of coverage can be considered an amalgamation of obligations (legal and regulatory), subscriber experience, network service quality and service availability. All of these attributes are examined herein. Available Data This paper presents a deeper examination of data currently available and proposes a basic framework for performance measurement and for more clearly defining the presence of coverage. However, our work has led us to conclude that there is currently no single data source that can be used to accurately identify mobile wireless coverage for FCC purposes. We have looked at sample data sets from,

• The National Broadband Map
• Mosaik Coverage Right
• Mobile Pulse drive tested and crowd-sourced data
• FCC Form 477 data

Each data source plays a vital role in understanding mobile voice and broadband coverage, but none

• n their own may provide enough context to inform policy. Our analysis of the data concludes that

multiple data sources contributes to a better understanding, but relying on a single source may lead

5 | P a g e to an unclear view. On the reporting side, how you define and develop your measurement and what you chose to measure in each area are all critical inputs into answering the question of what is covered. This analysis begins with the proposition that decisions regarding the goals of mobile voice and broadband networks need to be established, and then performance measures, tests and appropriate reports should be established. Only with a clear linkage from the goal to a measure can the policy debate shift from “is it covered?” to “is the level of mobile voice and broadband coverage fulfilling the FCC goal of providing broadband where Americans live, work and travel?”

6 | P a g e

Introduction

The USF/ICC Transformation Order established a performance goal to “Ensure Universal Availability

• f Mobile Voice and Broadband Where Americans Live, Work, or Travel.”1 This goal was “designed

to help ensure that all Americans in all parts of the nation, including those in rural, insular, and high- cost areas, have access to affordable technologies that will empower them to learn, work, create, and innovate.”2 In the Transformation Order, the FCC declined to adopt performance measures for this “coverage” goal, but directed the Wireless Telecommunications Bureau to develop one or more appropriate means of measuring.3 Since the time of the Transformation Order;

• There have been a number of changes in the deployment of mobile voice and broadband

networks;

• There have been changes in data available to analyze these networks; and,
• There have been changes in the analytical methods to review these data.

Yet, we are not aware of any Commission actions to define the performance measures for the “coverage” goal. As such, it will be difficult to demonstrate whether the stated goals of the Transformation Order are being achieved, and ultimately, what level of CAF funding is required and where it should be invested, until:

• The performance measures are articulated to support the coverage goal,
• There is buy-in on the applicability of those measures, and
• There is clarity and transparency in how the measures are developed and monitored.

In the recent FCC Notice, the FCC is examining data and considering new rules for Mobility Fund Phase II. Given the apparent ambiguity that exists between the stated goal and its accomplishment, the purpose of this paper is to expand the public record in regard to the performance measurement framework and examine data that could be used to review progress toward the accomplishment of the FCC’s performance goal (i.e., coverage in the context of the FCC’s Mobility Fund Phase II rulemaking). As part of our analysis, we address key aspects and questions from the FCC Notice, including:

• Shifting the measurement of coverage to a population basis rather than road basis.
• Measuring coverage including statements that,
• 99.5% of the US population is covered by some form of mobile broadband4

1 USF/ICC Transformation Order, 26 FCC Rcd at 17682, para. 53. 2 Ibid. 3 USF/ICC Transformation Order, 26 FCC Rcd at 17682, para. 54 4 “According to some sources, nearly 99.5 percent of the U.S. population today (and the road miles associated

with that population) is covered by some form of mobile broadband technology.” See, FCC’s Connect America Fund Omnibus Order and FNPRM, para. 238.

7 | P a g e

• Verizon LTE covers 303 million people in the U.S. (95% of American Pops)
• AT&T LTE covers 280 million people in the U.S. (88% of American Pops)
• Defining the areas of coverage and whether targeting funding to preserve and extend service

in those areas that will not be served by the market without government support5, including exclusion of areas served by Verizon or AT&T 4G LTE service6, will preserve existing service in those situations where the network of a mobile provider covers both eligible and ineligible areas so as to promote the preservation of service in the portion that does not overlap Our research is developed from various sources which are further explained throughout the analysis. They consist of a combination of both public and proprietary data which we have analyzed and melded together to provide additional information for the public record. The first section of this paper examines potential performance measures. In the second section, the data products that can be useful in measuring achievement of the goals are reviewed. And while we compare and contrast data products that are useful in measuring achievement of the goals, we make no claims on which one may be better, more correct, or more accurate. Rather, our purpose is to point

• ut what the data represents, dissimilarities between the data, and the complexity of answering a

difficult question (what is covered?) with each one.

5 Ibid, para. 239. 6Inviting the comparison in terms of service provided by Verizon and AT&T seems to setup a framework in

which any examination of Verizon and AT&T coverage may be viewed as a criticism of the carriers. That is specifically not the intent of our analysis. Rather AT&T and Verizon 4G LTE service must be examined to analyze the potential incomparability in terms of each of the potential coverage measurement criteria.

8 | P a g e

Section 1: Performance Measures for the Goal of Universal Coverage of Mobile Voice and Broadband

Section IV of the Transformation Order articulated a number of performance goals. However, with regard to mobile voice and broadband networks, the definition of performance goals and the corresponding measures were deferred to later action. 7 While we recognize that translating a performance goal into one or more performance measures is complex and rarely done in one draft, if a performance goal is not reflected back into a set of agreed upon measures, there is a low likelihood that the goal will be achieved. Because the Transformation Order declined to articulate a performance measure for mobile voice and broadband networks, this paper will start with the performance goal and then suggest a performance

• framework. The performance framework represents objective criteria of that goal. We are not

suggesting these are the only criteria by which performance should be measured, but they are criteria which have been advocated by parties, and for which methods are available to measure. The following, based in part on FCC Orders, represent measurable criteria which could be considered in determining whether the performance goals are met. The following criteria could help define when an area is “Covered”8 for Mobility Fund Phase II purposes, as well as what could be required of recipients of Mobility Fund support who are expanding Coverage. 1) Obligation Requirements: This captures a series of regulatory obligations that are required of network providers. These may be mandated by receipt of FCC funding or they may be mandated by other actions. 2) Subscriber Experience: This defines coverage based on how the end user perceives the performance and value of the network. The next two are somewhat intertwined given that availability is linked with quality. 3) Network Service Quality: This captures a series of large scale measures of the technical performance of the network. While an individual’s download or upload speed can be an

• utcome of many factors, service quality measures reflect isolated tests against specific

network functions. These measures are generated at the network level rather than the individual subscriber. 4) Service Availability. This measures the presence, or absence, of mobile broadband and voice networks at a particular speed threshold. In summary, to determine if the goal of Coverage has been achieved, the concepts of Coverage need to be clearly defined. And once defined, how Coverage is measured needs to be clearly laid out.

7USF/ICC Transformation Order, 26 FCC Rcd at 17682, para.54. 8 We refer to Coverage as a term the FCC needs to define. For the remainder of this paper “Covered” will no

longer be wrapped in quotes.

9 | P a g e

Figure 1: Hierarchy of Coverage Summary

The next sections of this paper will discuss each of these areas as well as potential measurement approaches.

Carrier Obligation Requirements

Reviewing the Transformation Order, prior USF and CETC requirements, and the requirements of Mobility Fund Phase I recipients, the following is a potential list of obligation metrics.

• For facilities constructed with Mobility Fund Phase 1 support, provide collocation for other

providers on newly constructed towers

• Offer Lifeline service in an area and provide evidence of compliance with all applicable federal

and state rules

• File periodic service improvement plans that demonstrate how the network operates or

planning for network upgrades

• Comply with the Commission’s voice and data roaming requirements on networks that are

built through Mobility Fund support

• Provide voice at reasonably comparable rates to urban areas
• Provide broadband at reasonably comparable rates to urban areas
• Design rates with capacity utilization limits comparable to usage limits in urban areas

Verification of carrier obligations could be an important factor in validating stated Coverage

• information. In addition, these criteria could be used to determine if an area is eligible for funding.

Subscriber Experience Measures

Measuring subscriber experience may be an important dimension of understanding progress toward defining an area as Covered and monitoring deployments that are funded. Some potential measures to consider are listed below.

• Bona-fide complaints of wireless service via FCC Form 2000B submissions (or similar) by

wireless customers.

Coverage Obligation Requirements Subscriber Experience Network ServiceQuality Service Availability

10 | P a g e

• Excessive roaming charges; this may be manifest in consumer complaints or it could also be

analyzed from wireless provider billing records comparing the roaming charges of certain classes of customers (urban versus rural) to understand how impacted one class is by roaming. If a particular class seems to be more likely to be charged roaming fees, it may imply something about the availability of a particular carrier’s offering.

• Proximity to retail support; this captures the concept that some carriers choose to serve

customers who “live, work and travel” in the area while other carriers have deployed service in an area only to serve customers who are passing through (e.g., highway coverage). One could infer that a carrier is providing a different subscriber experience if it has a retail/service center within a reasonable driving distance for the customers in the area. The illustrations below summarize an analysis of drive time to the nearest retail/service site based on data we downloaded for the South Dakota retail/service sites from both AT&T’s and Verizon’s (VZN) websites9.

Figure 2: South Dakota Verizon Retail Outlet Trade Area Analysis

9 Trade Area Analysis performed with ESRI Network Analyst, 10.2.1. All classes of Streetmap North

America roads were used; detailed polygon output is shown.

11 | P a g e

Source: National Broadband Map & CostQuest Associates Analysis Figure 3: South Dakota AT&T Retail Outlet Trade Area Analysis Source: National Broadband Map & CostQuest Associates Analysis

While we cannot attest to the completeness of the AT&T and Verizon retail/support locations, they do represent what is advertised to consumers on the carriers’ websites. If we assume the data is complete, the images do raise the questions of what is a reasonable subscriber expectation of travel time to reach a representative to buy services or discuss billing, technical support or service quality and whether lack of local support should be considered as a component of community Coverage.

Network Service Quality Measures

The FCC has released four annual reports measuring fixed broadband.10 Over time, the reports have grown in subjects covered and in the granularity and depth of metrics. The Commission’s team behind

10 The initial Measuring Broadband America Report on Fixed Broadband was published in August 2011, and

presented the first broad-scale study of directly measured consumer broadband performance throughout the United States. This effort was followed approximately one year later by a second Report, released in July

12 | P a g e this report has shown dogged determination in examining data sources, listening to comments and improving the deliverable each year. Although measurement of mobile voice and broadband networks’ service quality likely represents a significant increase in complexity over fixed wireline, the FCC could utilize data captured in both its

• wn and commercial mobile testing applications as one source of service quality measurements.

While we understand the complexity of this undertaking, the service quality data are becoming

• available. However, what doesn’t seem to exist is consensus on the appropriate network tests to use,

the method to test, the appropriate measures, and the interpretation of the results. As a potential starting point, mobile voice and broadband network measures could be expressed in terms of defined technical criteria listed below. These measures are intended to be consistent with

• ther measures, but may also represent a more technical testing platform consistent with the

Commission’s Report on Consumer Wireline Broadband Performance in the US11.

• Bandwidth by technology (in addition to speed and latency)
• Packet loss
• Packet jitter
• Failed call/session attempts by technology
• Voice
• Data
• Dropped call % by technology
• Voice
• Data
• Target Technology Connection rate
• Percent success of 4G phone connecting with a 4G network
• Percent success of 3G phone connecting with a 3G network

Mobile Voice and Broadband Service Availability

Consumers are bombarded by claims regarding mobile voice and broadband network availability. Most mobile carriers provide a service availability locator feature and users tend to look for bars on their mobile devices in an attempt to determine the adequacy of the service they have purchased. Mobile carriers have long had proprietary network performance data available to aid their network

• management. In some areas, there are independent performance metrics available to analyze network

performance on a location specific basis. And, the Department of Commerce’s NTIA collects broadband availability through its National Broadband Map. Yet, despite this data, we see gaps in the public record on what “voice and broadband network availability for where Americans live, work and travel” means or implies. As such, care and caution

2012, a third Report released in February 2013, and now this Report.”; A Report on Consumer Wireline Broadband Performance in the U.S., http://www.fcc.gov/reports/measuring-broadband-america-2014

11 The 2014 wireline testing report is available at http://www.fcc.gov/reports/measuring-broadband-america-

• 2014. Results of the mobile testing do not seem to be available, but the data dictionary is available at

https://github.com/FCC/mobile-mba-androidapp/wiki/Data-Representation

13 | P a g e should be used in measuring any concept of availability based on any source of service availability

• information. That is, before we can start to measure service availability we first must understand what

service availability means. At that point, we can then assess whether the considered data can measure Coverage from the aspect of service availability. As a first step to define service availability, we believe the following questions can help frame the definition of “availability” by looking at it as a collection of piece-parts: Presence, Measure, Unit and Area.

• What determines presence?
• Are there any specific quality measures?
• Where is presence measured: In-buildings, on-roads?
• The presence of what: Voice, Broadband, both?
• What is the measure of presence?
• What is considered mobile Broadband: 4G LTE, 4G or a speed threshold?
• In testing, do we need 100% successful connections at a specific speed to conclude

adequate availability?

• If speed is the threshold, what speed is required and what determines speed: advertised

minimum or maximum, cell edge drive test, other?

• What is the geographic unit for measuring presence?
• Area, Roads, Structures, or Population as measured by household locations
• If using a boundary area such as Census Block of Tract, what determines

whether or not an area is covered? Is it the centroid of a Block or Tract or is it a matter of having just the edge or any protion of the area covered?

• In what geographic area are you determining successful presence?
• Census block, tract, designated place, Metropolitan Statistical Area, etc.
• And given the area selected, what count/summary/test of the units in the area defines

“service availability” (e.g., what % of the units in the area) In the material to follow, we review each of the four key piece-parts of service availability: Presence, Measure, Unit and Area. Presence In addition to the quality issues addressed earlier, we consider presence, including a list of service items that the FCC has identified in its various orders and notices with respect to the Mobility Fund.

• Availability of Mobile Voice where Americans Live
• Availability of Mobile Voice where Americans Work
• Availability of Mobile Voice where Americans Travel
• Availability of Mobile Broadband where Americans Live
• Availability of Mobile Broadband where Americans Work
• Availability of Mobile Broadband where Americans Travel

14 | P a g e Measure of Presence We list a number of items that the FCC has identified in its various orders and notices regarding the Mobility Fund to measure the presence of voice and broadband service. 12

• Broadband speed
• Broadband service that meets performance metrics for actual speeds rather than

“advertised” or “up to” metrics

• Actual speed and latency measured on each carrier’s access network from the end-user

interface to the nearest Internet access point

• For the rural experiments, requirements of 4 Mbps download and 1 Mbps upload
• In the Mobility Fund Phase I Auction, requirements of 768 kbps download and 200

kbps upload at the cell edge

• In the Form 477 data collection process, minimum advertised speeds are requested
• Broadband latency: Low latency to enable use of real-time applications such as VoIP
• Broadband capacity: Usage limits comparable to usage limits in urban areas
• Broadband availability measured at the end user

What is considered Broadband -- Access to the speeds observed in the urban areas As stated in the FCC’s Transformation Order as part of the “Principles and Goals”, support is intended to “…ensure universal availability of modern networks capable of providing advanced mobile voice and broadband service.”13 Additionally, the Order states that funded areas should have access to the speeds observed in the urban areas. However, the Notice14 asks if the presence of 4G LTE should be sufficient, or should a different deployment standard be used. One could ask how a static technology definition (i.e., 4G LTE) ensures universal availability of modern networks with access to speeds

• bserved in urban areas in a mobile network environment that is constantly evolving. There are

inherent issues that come with defining broadband in the form of a static technology choice whose capabilities are dependent upon the spectrum available and used, backhaul sizing and availability. Geographic Unit of Measure The Commission established as a goal the universal availability of “mobile networks capable of delivering mobile broadband and voice service in areas where Americans live, work, or travel.”15 Given this, what should be the geographic unit measured – roads, business locations, or the households in which the population lives, etc? Or, if agricultural needs are to be incorporated as suggested in the recent Deere & Company filing,16 does the unit of measure need to extend beyond roads (e.g., specified buffer about the roads) and into “croplands” and other demand points beyond roads and structures? At this point, we believe it’s worth examining how moving from road miles, as used in the Mobility Fund Phase I Auctions for non-tribal lands, to the proposed population approach in the FCC’s Notice

12 USF/ICC Transformation Order, 26 FCC Rcd at 17682, para. 74-114, 358-453 13 Ibid, para. 17. 14 FCC Report and Order, Declaratory Ruling, Order, Memorandum Opinion and Order, Seventh Order on

Reconsideration, and Further Notice of Proposed Rulemaking, released June 10, 2014, para. 238-239.

15 USF/ICC Transformation Order, 26 FCC Rcd at 17682, para. 53. 16 Deere and Company Reply Comments. See http://apps.fcc.gov/ecfs/comment/view?id=6018329462

15 | P a g e conforms with the concept of measuring “mobile networks capable of delivering mobile broadband and voice service in areas where Americans live, work, or travel.”17 In concert with the selection of the geographic unit of measurement is the need to line up the appropriate measure of presence. For example, if the geographic unit selected is population within households, which may infer in-building coverage, are availability measures that are based on on-road availability an appropriate measure of presence? Geographic Area to Determine Service Availability Within The Notice repeatedly refers to coverage in an area, yet does not define areas eligible for funding and, we assume, in determining how coverage is measured post receipt of funding. Paragraph 239 of the Notice states “…preserving and extending service in those areas that will not be served by the market without governmental support” and “expanding access to 4G LTE in those areas that the market will not serve”. In paragraph 242, the Notice states ”…we propose to identify areas eligible for support, i.e., areas where neither Verizon nor AT&T provide 4G LTE but also seek comment below on whether this standard will preserve existing service in those situations where the network of a mobile provider covers both eligible and ineligible areas. We also propose to identify eligible areas using the most recently available data for this purpose as reported on Form 477.” It appears from these statements that the FCC is viewing Coverage as a measure of service availability throughout a collection of geographic units. To get a sense of how areas have been determined, we can look to prior FCC Orders. For example, in the Mobility Fund Phase I Auction, the centroid of census blocks had to be unserved for the area to be eligible, while a recipient of funds had to deploy to at least 75% of the unserved roads in an auctioned census tract. Further, in choosing the geographic area for Mobility Fund Phase I auctions, the FCC indicated that “…census blocks are on average far smaller than the average area covered by a single cell tower, which is likely to be the minimum incremental geographic area of expanded coverage.”18 One could add to this, that the service area of a single tower is simply a piece part of providing coverage to where people live, work and travel in their community. As such, should the coverage area for Mobility Fund Phase II be defined in terms of a community and should service availability be determined by the ability to serve the majority of the roads in the community at a specified minimum speed? As an additional consideration in defining the appropriate geographic area to be used in determining service Coverage, if a provider covers the highways through a town and some of the secondary roads but not the town itself, are the citizens of the town, the places they work in the town and the roads and parks that they travel on and to considered covered? If they attempt to buy mobile service but the closest retail store is over 2 hours away, is that area considered a covered area?

17 USF/ICC Transformation Order, 26 FCC Rcd at 17682, para. 53. 18FCC Order, In the Matter of GCI Communication Corp, Waiver of Section 54.1007(a) of the Commission’s Rules.

Released November 21, 2013, para. 5.

16 | P a g e

Summary of Coverage Criteria:

From the criteria above, the following list summarizes the items that could be considered in defining coverage and could provide the basis for the performance measurement framework.

Figure 4: Basis for the Performance Measurement Framework Source: CostQuest Associates Analysis

ID CoverageMeasure 1 AvailabilityofMobileVoicewhereAmericansLive 2 AvailabilityofMobileVoicewhereAmericansWork 3 AvailabilityofMobileVoicewhereAmericansTravel 4 AvailabilityofMobileBroadbandwhereAmericansLive(Populationmeasures) 5 AvailabilityofMobileBroadbandwhereAmericansWork(businesslocationmeasures) 6 AvailabilityofMobileBroabandwhereAmericansTravel(roadmeasures) 7 Signalmeasuredat: 8 InBuilding 9 OnRoad 10 Broadbandmeasuredon 11 Speed:Broadbandthatmeetsperformancemetricsforactualspeedsratherthan “advertised”or“upto”metrics 12 Ruralexperiments,landlineandnonedgemobilityrequirementsof4mbdownloadand 1mbupload 13 Mobilityauctionrequirementsof768kbdownloadand200kbupload 14 Latency:lowlatencytoenableuseofrealtimeapplications,suchasVoIP 15 CoveragemeasuredonDrivingTestResults 16 ActualspeedandlatencybemeasuredoneachETC’saccessnetworkfromtheenduser interfacetothenearestInternetaccesspoint 17 Geographicunitofmeasurementselectedandidentifiable:Roads,Pop,Structures,etc.. 18 GegraphicAreaformeasuringsuccessinidentifiablesuccessdefined 17 Bandwidthbytechnology 18 Packetloss 19 Packetjitter 20 Failedcall/sessionattemptsbytechnology 21 Voice 22 Data 23 Droppedcall%bytechnology 24 Voice 25 Data 26 TargetTechnologyConnectionrate 27 Percentsuccessa4GphoneconnectswithLTE 28 Percentsuccessa3Gphoneconnectswith3G 29 Voiceatreasonablycomparablerates 30 Broadbandatreasonablycomparablerates 31 Capacity:Usagelimitscomparabletousagelimitsinurbanareas 32 ServiceImprovementplansfiled 33 OfferLifeline 34 Providecollocationforotherprovidersonnewlyconstructedtowers 35 ComplywiththeCommission’svoiceanddataroamingrequirementsonnetworksthatare builtthroughMobilityFundsupport 36 Bonafideconsumercomplaintsper1000customers 37 Averagebilledroamingcharges 38 Proximitytolocationbasedretail/support ObligationRequirements NetworkServiceQuality SubscriberExperience ServiceAvailability

17 | P a g e

Section 2: Sources of Voice and Broadband Coverage Information

At this time, CostQuest Associates’ working assumption is that to determine the amount of Mobility Fund Phase II support required, and where those funds should be targeted, a dataset which purports to represent mobile voice and broadband Coverage should provide indications on provider obligations, subscriber experience, and a single measure for each area tested on, (1) the presence or absence of quality mobile voice and broadband service (2) to a specified geographic unit (3) at a required speed threshold (4) over a specified geographic area. It should be available (or can be developed) for all regions of the United States and it should be linked back to a documented, repeatable method of

• production. Given the complexity, this may require the creation of a new data source or the

amalgamation of multiple sources. As a start to identifying the potential dataset to determine Coverage, there are several sources of data which have been mentioned by parties that can be assessed for analytical suitability. We review those that have either been mentioned in the FCC’s Notice or identified in the National Broadband Map efforts. Each source, to be discussed below, has a different target requirement and production mechanism. It may be produced by a service provider, a State Broadband Initiative grantee, or a third party. It may be derived from crowd sourced data that could be targeted to a point in time, a particular provider, a particular device or a particular geographic area. It may or it may not provide distinctions to demonstrate differences in voice and broadband service availability in-building or on a road.

Voice and Broadband Service Coverage Data Sources

CostQuest Associates reviewed several potential data sources in regard to their potential use for assessing Mobility Fund Phase II Coverage. The data sources were as follows:

• Mosaik Coverage Right (Furnished September 2014, formally known as American Roamer)
• National Broadband Map (NBM) - the NBM was created by the National

Telecommunications and Information Administration (NTIA), in collaboration with the Federal Communications Commission (FCC), and in partnership with 50 states, five territories and the District of Columbia. The NBM is part of NTIA's State Broadband Initiative.

• Field Tests – Conducted by companies such as Mobile Pulse, who use crowd sourcing as their

primary method of data collection. Other vendors such as QoS or RootMetrics provide similar services related to testing mobile networks. There is some disagreement among the firms as to the suitability of any specific test for a particular purpose.

• FCC Form 477: The information developed by service providers, collected and used by the

FCC to comply with statutory requirements and as an aid in developing and revising policy.

18 | P a g e Mosaik Solutions (Mosaik) Mosaik is a long standing industry data source originally created to assist carriers in identifying potential roaming partners. An FCC report cited the company’s wireless network coverage data as the industry standard for coverage information.19 Source Mosaik collects coverage data from participating providers on a periodic basis for each specific

• technology. We are unsure what happens in the case where a provider does not submit information

to Mosaik. Services Represented From our understanding, Mosaik data represents a carrier’s advertised coverage in terms of an advertised platform such as LTE or HSPA+. Both voice and broadband data layers are available. A sample map from their site is shown below.

Figure 5-Mosaik Custom Marketing Map; from http://www.mosaik.com/showcase/portfolio/

19 FCC Connect America Fund Report and Order and Further Notice of Proposed Rulemaking, released November 18,

2011; para. 334.

19 | P a g e Data Derivation U.S. Cellular, a participating contributor to Mosaik, has informed CostQuest that based on their participation with Mosaik there is nothing specific in Mosaik’s data request. It is simply a request of coverage information. We are not in possession of a specific data request and are not aware of any specific frequency, speed or performance criteria in the data collection. Additionally, we are unaware of any documentation describing the methodology by which the data are assembled, normalized, updated and corrected. Data Availability The data appears to be available for most of the United States and internationally. Licensing Mosaik solutions data are available as a licensed product for a fee. Presumably, licensing restricts usage and distribution rights. National Broadband Map (NBM) As a result of the National Broadband Plan, the NBM was created by the National Telecommunications and Information Administration (NTIA), in collaboration with the Federal Communications Commission (FCC), and in partnership with 50 states, five territories and the District of Columbia to collect national broadband deployment. The NBM is part of NTIA's State Broadband Initiative (SBI). The NBM is updated approximately every six months and was first published on February 17, 2011. A sample map from the National Broadband Map site is shown below.

Figure 6: Sample Map from the National Broadband Map web site (25Mbps availability). Source: http://www.broadbandmap.gov/technology

20 | P a g e Source State entities (grantees) collect coverage data from mobility providers bi-annually. The data are sourced from providers or derived using other sampling methods as developed and verified by the

• grantee. Data are submitted from grantee to FCC/NTIA who review data and post to the NBM.

Data are disseminated in either geographic format (ESRI Shape files) or tabular format (CSV files). Coverage represented The data captures the broadband coverage and the maximum advertised speed up and down, by frequency and by technology. Data collection represents the grantees view of coverage as of a specified date. Services Represented Only broadband data are collected. There is no information specific to voice networks. Data Derivation Grantees collect data two times each year. Many grantees survey service providers for the

• information. Some service providers request review of grantee information prior to submission to

NTIA/FCC. The data collection was authorized in 2009 under the Broadband Technologies Opportunities Program (BTOP) program20. Mobile broadband coverage data is expected to reflect the following characteristics.

• All map areas must be closed, non-overlapping polygons with a single, unique identifier.
• Any variation in any of the required fields necessitates the creation of a separate closed, non-
• verlapping polygon.
• In the area covered by each polygon, subscribers must have broadband service with the speed

characteristics shown in the data record 95% of the time to within 50 feet of the polygon’s boundary.

• The technology of transmission is a categorical, specified value (e.g., 80 represents all mobile

broadband).

• The speed tiers should be entered as integers (NBM uses a categorical speed schema; e.g.,

category 7 represents 10 to 25 Mbps)

• The data must be expressed using the WGS 1984 geographic coordinate system.
• Submissions must be accompanied by metadata or a plain text ‘‘readme’’ file that contains a

comprehensive explanation of the methodology employed to generate the map layer including any necessary assumptions and an assessment of the accuracy of the finished product. As required above, each grantee publishes a methodology document describing how the data were developed.

20 See Broadband Technology Opportunities Program (BTOP) at http://www2.ntia.doc.gov/about,

21 | P a g e Data Availability NBM data are available in either whole Census block format (text file) or in a geographic format closer to the original submission from the grantee File Geodatabase21 file format. The State Broadband Initiative will cease to collect data after October 2014. Presumably data from the FCC Form 477 will be used as the replacement data source for the National Broadband Map. Licensing We are unaware of any restrictions on the use of the data. Mobile Pulse Mobile Pulse is an independent entity that has been retained by a number of the grantees working on the National Broadband Map State Broadband Initiative (SBI) data collection efforts to undertake end-user testing of mobile broadband coverage. Source Mobile Pulse uses crowdsourcing via smartphone apps to record and report on end-user test results on speed, latency, failures and other metrics listed below. Services Represented Mobile Pulse can provide information for voice or broadband networks. Data Derivation Data are derived from a crowd source application which runs on iOS and Android phones. Users volunteer the use of their smartphones. The application runs a specific set of tests over a predetermined frequency as the user moves about. A sample of the fields captured by the Mobile Pulse data follows: addressaddress latencyaverage addresscity latencystatus addresscountry downloadexecutionTime addresscounty downloadfileSize addressstate downloadspeed addresszip downloadmessage deviceaccuracy downloadstatus devicealtitude latitude devicenetop longitude devicenetType IsWifi devicenetworkType uploadexecutionTime deviceprovider uploadfileSize deviceroaming uploadspeed devicesignalStrength uploadmessage deviceSimop uploadstatus

21A File Geodatabase is a container for geographic data. It is a format maintained by ESRI.

22 | P a g e latencymin Indoor VS Outdoor Analysis latencymax Data Availability Data are provided as a text file; each record represents a location tested and test results. Records are time stamped such that temporal as well as geographic factors can be examined. Licensing Mobile Pulse data are available as a licensed product for a fee. Presumably licensing restricts distribution rights. FCC Form 477 Data FCC Form 477 has been collected twice yearly by the FCC since 2000. The reporting mechanism seeks information on mobile and fixed network availability and deployment. It also records information on subscribership. In 2014, the Form 477 process was expanded to start the collection

• f data similar to that currently captured in the SBI data.

Source Carriers submit information to the FCC. The FCC processes submissions and supplies non- confidential information for public availability. Services Represented Information on voice and broadband networks is submitted to the FCC. Information is available for both network availability and network deployment. Deployment is a facilities based view for broadband and voice networks. Data Derivation With respect to mobile broadband networks, FCC form 477 requires the carrier to ascertain the accuracy of the submitted information and describe the methods by which deployment information was developed. Carriers are responsible for their submission although some have used outsource agents for assistance. The broadband coverage specification is similar to the SBI specification. The boundary resolution is 100 meters and the carrier is responsible for describing the method by which the coverage files are generated.

• All map areas must be closed, non-overlapping polygons with a single, unique identifier
• Any variation in any of the required fields necessitates the creation of a separate polygon

showing the relevant coverage

• The shapefile must have an assigned projection with an accompanying .prj file.
• The shapefile must use un-projected (geographic) WGS84 geographic coordinate system.
• The coverage boundaries should have a resolution of 100 meters (approximately three arc-

seconds) or better. An arc-second represents the distance of latitude or longitude traversed on the earth's surface while traveling one second (1/3600th of a degree). See

23 | P a g e http://www.esri.com/news/arcuser/0400/wdside.html. Three arc-seconds is a common resolution of terrain databases. See USGS Standards for Digital Elevation Models, Part 1- General, at 1-2, 1-4, http://nationalmap.gov/standards/pdf/1DEM0897.PDF.]

• The shapefile should be submitted as a WinZip archive.
• In addition to the shapefile, each zip must include metadata or a plain text “readme” file that

contains a comprehensive explanation of the methodology employed to generate the map layer including any necessary assumptions and an assessment of the accuracy of the finished product. The 477 coverage is the only source that requires corporate certification upon submission to FCC. Data Availability The first data collection will be completed in the latter half of 2014. The availability of the data for use in any service availability determinations is currently unknown. Licensing We are not aware of any license but the FCC has not clarified how confidential information will be treated for public review. Coverage Availability Source Summary This table summarizes reviewed sources of service coverage. Mosaik Solutions National Broadband Map (SBI) Mobile Pulse FCC Form 477 Source Carrier request of ‘coverage’ Grantee request

• f broadband

coverage with respect to speed, technology, frequency per NOFA standards Crowdsource App Carrier Submission to FCC Services Represented Voice and/or broadband Broadband only Voice and/or broadband Voice and/or broadband Data Derivation As produced by carrier response for coverage, unknown process by Mosaik Grantee process, documented in methodology. Methodology available for public review. Licensed application. Presume tests are proprietary Carrier submission per FCC specifications and corporate certification Data Availability US and International, unknown how non-responders are handled All Grantee areas, SBI program ceases 10/2014. Purchase license by area Service areas of all broadband and voice providers. Not currently available.

24 | P a g e Mosaik Solutions National Broadband Map (SBI) Mobile Pulse FCC Form 477 Speed Portrayal Makes no speed claim Maximum Advertised Speed in an area As measured during testing Minimum Advertised speed in an area Wireless Communication Technology Portrayal Described as LTE, HSPA, UMTS, etc. Technologies are described as mobile wireless, coverage distinct

• n spectra used

Characterizes the type of wireless communication network and handset capability Technologies are described as mobile wireless, coverage distinct

• n spectra used

Voice and/ or Data information Provides voice and data information Broadband capable layers

• nly

Tests used are for data only Information on both voice and broadband deployment. Licensing Fee based No license Fee based No license but FCC has not clarified how confidential information will be treated for public review.

Figure 7: Summarized Sources of Service Coverage Source: CostQuest Associates Analysis

Comparison of coverage sources and measures for sample states

In the following sections, we present a comparison of the different potential coverage sources (Mosaik, National Broadband Map, and Mobile Pulse) for a sample of five states, unless noted. Before we begin our analysis of the data sources, a few notes:

• The new FCC Form 477 data, while stated by the FCC as the preferred source for determining

coverage,22 is not currently available. The first data collection is underway. There is no indication when the data will be available nor what will be made available (confidential submissions).

• Given this project’s constraints of data, time and budget, we limited our analysis to 5 states.

The states were selected, in part, based on the availability and extent of Mobile Pulse’s drive test data. Analytical Data Creation/Source/Methodology Geographic Units In developing measures of voice and broadband network coverage, the first question will be what geographic unit is going to be used. Is it covered residences, covered population, or covered roads

22Report and Order, Declaratory Ruling, Order, Memorandum Opinion and Order, Seventh Order on Reconsideration,

and Further Notice of Proposed Rulemaking, FCC 14-54, released June 10, 2014, at para. 241 (“Seventh Order”).

25 | P a g e that are being measured? In other words, one must define what unit will be used to determine what and how you are going to measure coverage. As a basis for our analysis, two data sets were developed to represent geographic units to be measured as Covered. Our targets of Coverage come from road centerlines. Through a GIS23 process, points were created every 100 meters along all TIGER road paths with MTFCCs of S1100, S1200, S1400, S1500, S1640 and S1740 (the same road types used as the basis of the Mobility Phase I CAF Auctions). The vintage

• f the TIGER roads was Census 2010. If a road segment was less than 100 meters in length, a point

was created at the midpoint of the segment. The length of the segment was then apportioned equally to the number of road points found on the segment. For segments that split census blocks, two points were created at each location along the segment and the corresponding distance halved. For our analysis of population and firm coverage, we started with Census 2010 Population and housing units by census block, then adjusted them for Census Population and Housing Units 2012 estimates at the County level. To this, we added in the summarized business firm counts at the census block, based on 2012 GeoResults Business location data. Using the Census Block counts, the population, housing unit and business firm data were assumed to be equally apportioned to the road footage in the associated census block. Based upon a point in polygon analysis, if the road point and its associated length was considered covered, the apportioned population and business firms were considered covered as well24. Coverage/Availability Data Sets Given that the FCC has noted that areas covered by AT&T and Verizon 4G LTE may be excluded from funding,25 in our analysis we focus on AT&T and Verizon coverage areas as depicted by each of the data sets for the five sample states. Mosaik Our primary review of Mosaik coverage attributes is based on the summary of nationwide Mosaik data from the FCC’s 16th Mobile Competition report. However, to get a sense of how Mosaik compares to the NBM data, we obtained Mosaik data for South Dakota (SD) and Wisconsin (WI) on August 30th 2014. For the comparative analysis, we used the LTE layers of AT&T and Verizon. Mosaik Compared To National Broadband Map The Mosaik data is supplied in terms of technology layers—LTE, HSPA+, EVDO, etc. We are not aware of any speed attribute or frequency used in deriving coverage information. This means

23 PostGIS, version 9.3. 24 In future iterations of this analysis, we could move to our location based database that contains the latitude

and longitude of housing units and business firms.

25 Seventh Order, supra, at para. 241 (“We propose to identify areas eligible for support, i.e., areas where neither

Verizon nor AT&T provide 4G LTE, but also seek comment below on whether this standard will preserve existing service in those situations where the network of a mobile provider covers both eligible and ineligible areas.”)

26 | P a g e comparison across data sources, such as National Broadband Map, is especially difficult because, by definition, what is covered is going to be different in each data source. This doesn’t imply that any source of coverage information is wrong, but what each may be portraying is different. From a visual standpoint the following series of maps shows South Dakota. Green represents where Mosaik and National Broadband Map overlap; the overlapping NBM coverage is transtech 80, maximum advertised speed of 10-25 Mbps.

Figure 8: South Dakota AT&T Wireless Coverage Comparison Sources: Mosaik, National Broadband Map, & CostQuest Associates Analysis

As shown below, the Verizon Mosaik coverage appears more consistent with the National Broadband

• Map. Given the same color scheme, Verizon appears to have only marginal differences between that

submitted to Mosaik and the National Broadband Map.

27 | P a g e

Figure 9: South Dakota AT&T Wireless Coverage Comparison Sources: Mosaik, National Broadband Map, & CostQuest Associates Analysis

In total for the two states, Mosaik indicates 45.4% more road miles covered for AT&T and 3.6% more for Verizon, as shown in the table below.

Figure 10: Tabular Coverage Data Sources: Mosaik, National Broadband Map, & CostQuest Associates Analysis

In summary, there could be any number of reasons to explain the incompatibility between the NBM data and Mosaik data (e.g., speed, technology, vintage, roaming and billing arrangements, etc.). Verizon appears more consistent between the two data products than AT&T (which in part could be that Verizon completed its rollout of LTE earlier than AT&T). Our intent is not to claim that any State Measurement Unit AT&TMosaik Coverage AT&TNBM Coverage Difference VZNMosaik Coverage VZNNBM Coverage Difference SD Roads 37.6% 26.6% 41.21% 84.8% 80.8% 5.0% WI Roads 24.6% 15.8% 55.23% 94.8% 93.0% 1.9% TOTAL Roads 32.1% 22.1% 45.42% 89.0% 85.9% 3.6%

28 | P a g e portrayal is more accurate than another. Rather we wish to point out that these two sources of data produce different characterizations of coverage. National Broadband Map (NBM) Coverage The geographic area coverage files of AT&T and Verizon were derived from the December 2013 (Round 9) update to the NBM. Shapefiles were downloaded from www.broadbandmap.gov/data- download (date). In the analysis, technology of transtech 80 (10-25 Mbps) was considered.26 To estimate the 4G/LTE coverage, we filtered the data where the Maximum Advertised Download speed was greater than or equal to 10-25 Mb.27 We used this NTIA category band based on our knowledge of the NBM collection process as well as industry advertising that places 4G LTE download speeds into a category greater than 10 Mbps.28 As noted, we used the shapefiles since they provide a view of coverage likely more consistent with the depictions of provider advertised coverage rather than the corresponding tabular census block .csv files

• n the NBM site. As we understand, the tabular Census Block .csv files indicate the census block is

covered whenever any portion of the census block is covered. As such, specifically with respect to mobile coverage a census block .csv based approach may tend to overstate coverage results. In order to quantify this potential overstatement, we tested the coverage using the tabular census block .csv files versus the shape files available on the NBM site for the 5 states in our analysis. In the image below, we compare the Road, Population and Firm coverage between each dataset within the specified location density groups. As can be seen, the use of tabular census block files for determining mobility coverage may lead to a greater overstatement of coverage as areas become less dense.29 The vertical axis on the following chart compares NBM coverage measurements using census blocks to NBM coverage measurements using .shp files at various residential and business location densities per square mile. For example, in areas with less than 5 business locations per square mile, the number

• f census blocks covered using .csv files is 122% of those using .shp files.

26 As an alternative tabular census block files could be downloaded where coverage is indicated for an entire

census block. We elected to use the provided shape files for this analysis as we wanted to remove the clouding issue of conversion of coverage contours into census blocks.

27 Based on our work on the SBI projects, it is our understanding that mobile providers have been advised to

use Speed Tier 7 for the maximum advertised download speed for 4G LTE and Speed Tier 6 as the typical download speed for 4G LTE.

28 http://business.verizonwireless.com/content/dam/b2b/resources/LTE_Benefits_Overview_TDM.pdf 29 This overstatement is likely less about the use of a CSV file than a SHP file rather than just saying any

amount of mobile coverage will cause the entire demand (roads, structure, demand points) in that census block to be called covered.

29 | P a g e

Figure 11: Impact of NBM Coverage Stats using CB .csv versus .shp Files Sources: National Broadband Map & CostQuest Associates Analysis

Mobile Pulse Coverage The field test data was purchased from Mobile Pulse on August 22, 2014. Tests are conducted for success in regard to latency, download and upload. In addition, achieved download and upload speeds were analyzed. The data is recorded with the latitude and longitude of where the test took

• place. In the section below, where we contrast Mobile Pulse coverage to National Broadband Map

coverage on download speed, the Mobile Pulse point records were converted to a speed polygon. For expediency of comparing the road test data to the demand data at road points, we rasterized the Mobile Pulse recorded speeds into 500 meter grid cells. The average download speed within each grid was derived and translated into an NTIA category. A polygon was then developed from each carrier’s speed raster. Grids with common speed attributes were combined together into a common polygon layer.30 One important note on Mobile Pulse road test data is that every road, and every road segment, may not be tested. As such, in our comparisons of Mobile Pulse to the other coverage layers, we can only draw preliminary conclusions for the roads/segments tested. However, we believe the road tests are extensive and cover a good portion of the state, as seen in the following image that displays the road test coverage of all carriers in South Dakota.

30Renderingpolygonsorshapesintoviewableimages.

30 | P a g e a

Figure 12: South Dakota Mobile Pulse Maximum Download Speed Sources: Mobile Pulse & CostQuest Associates Analysis

We did not stratify the raw data to study temporal shifts. (Note: The map above reflects tests that were conducted on “all” road types.) Investigation of Geographic Unit The Notice states “For example, the proposed rules now focus on coverage of population rather than road miles.” 31 In addition, paragraph 238 of the Notice focuses solely on population coverage statistics. As the FCC considers moving to population coverage as a measure to determine whether an area is covered by mobile broadband and voice services, consider the following image which compares AT&T and VZN 4G LTE coverage by roads and by population in different density bands for the five sample states. From the image, it is apparent that in urban areas “covering population” is nearly equivalent to “covering roads” (nearly 100%). As such, using covered population in an urban as a

31 Seventh Order, supra, at para. 248, n. 451.

31 | P a g e means of measuring mobility Coverage “where people live, work and travel” appears

• acceptable. However, as we move to lower density bands, it becomes less apparent that population

is an adequate proxy to determine where Americans live, work and travel and could lead to biased conclusions that an area is covered when, in fact, many of the road miles and areas where Americans live, work and travel are not covered.

Figure 13: Road to Population Coverage Source: CostQuest Associates Analysis

DataSourceDifferencesResults In the material that follows, we investigate the differences in road, firm and population coverage in the reviewed data products. National Broadband Map Coverage Results Based on a review of the NBM data for five states, we found the following for AT&T and Verizon coverage.32 Given that our goal is to compare the unit of measure and not to critique the carriers’ coverage, we will not utilize the carrier’s name in the charts and tables below. As the reader will note

32As noted above, the sole reason for using AT&T and Verizon in our analyses is because of the FCC

statement in the Seventh Order at para. 241, proposing that areas eligible for mobility support should be those areas where neither Verizon nor AT&T provide 4G LTE service.

32 | P a g e in the table and charts below, while business firms and population coverages within a state for a carrier are similar, the road coverage is different.

Figure 14: National Broadband Coverage Results Source: National Broadband Map

State MeasurementUnit CarrierX NBM CarrierZ NBM SD Roads 15.8% 93.0% SD Population 17.4% 97.3% SD BusinessFirms 17.6% 98.1% CO Roads 29.9% 66.6% CO Population 86.5% 97.6% CO BusinessFirms 85.8% 97.0% ID Roads 32.9% 57.8% ID Population 85.0% 95.8% ID BusinessFirms 83.5% 95.6% WI Roads 26.6% 80.8% WI Population 64.0% 93.8% WI BusinessFirms 59.9% 92.9% WY Roads 13.4% 49.2% WY Population 39.2% 94.2% WY BusinessFirms 37.1% 93.7% TOTAL Roads 24.3% 68.0% TOTAL Population 71.0% 95.7% TOTAL BusinessFirms 69.5% 95.3%

33 | P a g e

Figure 15: Carrier X NBM Coverage Source: National Broadband Map

34 | P a g e

Figure 16: Carrier Z NBM Coverage Source: National Broadband Map

If we assume roads are to be used as the geographic unit to measure coverage, trends in NBM data become more apparent. In the image below, we review the NBM data to analyze variances in road coverage based upon the density of the area.

35 | P a g e

Figure 17: Road Coverage by Density Sources: National Broadband Map & CostQuest Associates Analysis

In the image below, we review the NBM data of the five sample states to analyze variances in road coverage based upon the density of the area.

36 | P a g e

Figure 18: Road Coverage by Type of Road Source: National Broadband Map & CostQuest Associates Analysis

These images of road coverage demonstrate a greater likelihood for the carriers to cover higher density areas and primary roads. If this is the case, it should be noted that a pure road coverage metric may need to be adjusted based upon where the road is; in other words are particular classes of roads in particular places more important to cover or is it total road miles covered that are important? Mosaik We can get a view of the differences in population, area and road coverage from nationwide Mosaik data using tables presented in the FCC’s 16th Mobile Competition Report. In the table below, extracted from Table 9 of the report, the nationwide coverage based on Mosaik is 99.5% population coverage (as reported in the June 10, 2014 Further Notice of Proposed Rulemaking), 67.8% of area, and 91%

• f roads. Note that similar to our findings for the NBM data, the road coverage drops from population
• coverage. This data also shows that the road coverage drops from 91% to 76.9% when excluding

areas served by only 1 provider.

37 | P a g e Estimated Mobile Wireless Broadband Coverage by Census Block, Oct. 201233

Number of Providers with Coverage in a Block Number

• f Blocks

(Thousands) POPs Contained in Those Blocks (Thousands) %

• f

Total US POPs Square Miles Contained in Those Blocks (Thousands) %

• f

Total US Square Miles Road Miles Contained in Those Blocks (Thousands) %

• f

Total US Road Miles 1 or more 10,708 311,025 99.5% 2,577 67.8% 6,209 91.0% 2 or more 9,889 305,590 97.8% 1,950 51.3% 5,245 76.9% 3 or more 7,954 286,121 91.6% 1,070 28.1% 3,570 52.3% 4 or more 5,977 256,191 82.0% 521 13.7% 2,252 33.0% 5 or more 4,222 215,375 68.9% 228 6.0% 1,353 19.8%

Figure 19: Estimated Mobile Wireless Broadband Coverage by Census Block Source: National Broadband Map & CostQuest Associates Analysis

In reviewing these published results, it is important to ask how coverage metrics are derived. As we noted earlier in the review of the NBM data, the use of the tabular census block .csv file may bias the coverage statistics upwards. In reviewing derivation of the table shown above from the FCC’s 16th report, the FCC used a census block aggregation approach that counted the geographic and demographic units of a census block as “covered” if the centroid point of the census block was covered34 by a Mosaik polygon. In other words, the attributes of the census block were covered if the centroid of the block was covered. This is in contrast to the demand or road point intersection approach used in our derivation of NBM statistics. We cannot comment at this time on the bias in assumed coverage that a block centroid method approach may introduce. Investigation of Quality of Service To get a view of the quality of the service available, we summarize the Mobile Pulse road test results for the top two national carriers. The charts capture successful latency test, successful download connection successful upload connection, the ability to get a 4 Mbps download and 1 Mbps upload, and the ability to get a 768 kbps download and 200 kbps upload test. Based upon our analysis, the network performance tests based on Mobile Pulse tests using LTE capable handsets indicates that for the top two national carriers, 5.4% of the latency tests fail, 3.4% of the download connection tests fail, 11.3% of the upload connection tests fail, 83.6% of the tests were able to achieve a 768kpbs download and 200kbps upload connection, and 59.7% of the tests were able to achieve a 4Mbps download and 1Mbps upload. These results are illustrated in the figure below.

33 Includes Federal lands. Commission estimates based on census block analysis of Mosaik CoverageRight

coverage maps, Oct. 2012. The estimates include coverage by all EVDO, EVDO Rev. A, HSPA/UMTS/WCDMA, HSPA+, LTE, and mobile WiMAX networks. Population data are from the 2010 Census, and square miles include the United States and Puerto Rico.

34 From the FCC’s 16th Annual Wireless Report in FCC WT Docket No. 11-186, Annual Report and Analysis of

Competitive Market Conditions With Respect to Mobile Wireless, Including Commercial Mobile Services, released March 21, 2014, at para. 43.

38 | P a g e

Figure 20: LTE Handset Mobile Pulse Test Results for Top 2 National Carriers Sources: Mobile Pulse & CostQuest Associates Analysis

Comparison of Quality of Service to Availability of Service To understand what the presence of service means in comparison to quality and what is measured in the data products, a comparative analysis of the data products is useful. In the table below, we capture the percentage of field tests conducted with LTE capable handsets performed in the NBM LTE (NTIA layers >= 10 Mbps) footprints of AT&T and Verizon in which the Mobile Pulse speed tests meet the download advertised speed of the carrier (i.e., 10 Mbps).

Latency_Success 94.6% Download_Success 96.6% Upload_Success 88.7% DL_UL_768_200 83.6% DL_UL_4M_1M 59.7% Download_AvgSpeed 8,772

• 5StateAverages

39 | P a g e

Figure 21: Percentage of field tests conducted with LTE capable handsets performed in the NBM LTE footprints of AT&T and Verizon Sources: Mobile Pulse & CostQuest Associates Analysis

An interesting trend appears between the maximum advertised download speeds recorded or inferred

• nto the top two national carrier NBM layers and how often users can attain those speeds. This trend

could be an artifact of our interpretation of the speeds from the NBM, the manner of testing speed, or that the advertised speeds are aggressive in certain areas, or a combination of the three. The trend appears across roads, business firms and population.35 The comparison is shown in the following figures which compares the NBM stated 4G LTE road coverage for the two top national carriers to the ability to get a successful 10 Mbps field test in those 4G LTE areas (note the ideal bar for the field test data would be 100% - that is, the user is able to

• btain the advertised speed in all portions of the carrier’s 4G LTE service area). It is important to note

that having 4G LTE coverage does not necessarily imply that a subscriber will have access to a speed

• f 10 Mbps. However, at this point our analysis focuses on NBM coverage which should be capable
• f yielding 10 Mbps versus field test results at a point in time on a particular LTE capable handset.

35 Other states having performed similar tests note similar results. See West Virginia Geological Survey

(http://www.broadband.wv.gov/Strategic_Planning/Documents/Region%202%20Broadband%20Fieldwork. pdf)

State MeasurementUnit CarrierX MobilePulse CarrierZ MobilePulse SD Roads 23.2% 57.1% SD Population 31.5% 58.9% SD BusinessFirms 28.8% 63.0% CO Roads 6.8% 18.1% CO Population 4.9% 18.6% CO BusinessFirms 5.8% 17.5% ID Roads 27.6% 31.7% ID Population 27.7% 24.5% ID BusinessFirms 24.6% 25.4% WI Roads 55.3% 60.6% WI Population 59.1% 55.4% WI BusinessFirms 62.5% 56.1% WY Roads 11.6% 35.4% WY Population 10.1% 25.6% WY BusinessFirms 8.9% 23.7% TOTAL Roads 33.0% 44.5% TOTAL Population 34.3% 32.3% TOTAL BusinessFirms 31.9% 32.3%

40 | P a g e

Figure 22: Carrier X Pop Coverage NBM vs Drive Test Sources: National Broadband Map, Mobile Pulse & CostQuest Associates Analysis

41 | P a g e

Figure 23: Carrier Z Pop Coverage NBM vs Drive Test Sources: National Broadband Map, Mobile Pulse & CostQuest Associates Analysis

The Location of the Quality Test The Mobile Pulse data can also be used to look at both indoor and outdoor testing results. In the images below, we look at the ability to successfully obtain a 768kbps/200kbps session36. In general, these data show drops in the success rate once the test moves indoors.

36756kbps/200kbpsisconsiderablylowerthantheFederalCommunicationsCommission’s4Mbps/1Mbpsgoal.

42 | P a g e

Figure 24: LTE Handsets Mobile Pulse Test Results Top 2 National Carriers Successful Test: 768kb Download/200kb Upload Source: Mobile Pulse

Review of Potential Coverage Measures and Currently Available Data In the previous sections we presented measures that could be considered in defining Coverage. In this section, we reviewed the various data products that have been considered for determining

• coverage. In the table below, we compare the two and indicate where we see intersections. From

this early view, it would appear that: a) A clear definition of what Coverage means is required, b) As no definition is agreed upon, there is likely no single data source that can be used to identify Coverage, c) Multiple sources may be required to determine Coverage, and d) It would be prudent to adjust the 477 collection process to more closely align with the ultimate definition of Coverage.

43 | P a g e

Figure 25: Coverage Measures Table Source: CostQuest Associates Analysis

44 | P a g e

Conclusion:

Before any conclusions are stated, we would like to point out a number of issues

• Given the sample size of this analysis, the results shown are tentative. Specifically, two

national carriers, Verizon and AT&T, were used for comparison in this sample. These carrier results are not an attempt to depict any service coverage problems, but were used at this time because the FCC has indicated that eligible areas for CAF Mobility Funding Phase II may be those areas where there is no AT&T/Verizon 4G LTE coverage.

• These comparisons are intended as an example of analyses that can be performed more so

than an indication of evidence of a problem. The analyses indicate potential measures against a coverage performance metric. Clearly more review and analysis should be done prior to reaching final conclusions.

• The Field Test data that is available today should not be used to draw conclusive results.

Rather, it should simply be used as a tentative indication that any particular coverage product may not be conclusive on capturing the full scale of what is considered Coverage. It may also indicate that further refinement of field test software and field test methods should be revised. As described in this paper there are a number of ways that measures against a goal of universal accessibility can be evaluated. From our review, it is not clear that the FCC’s stated performance goals for mobile broadband have been accomplished. However, our analysis indicates that there are information sources and methods available to improve the assessment of mobile broadband availability. At a fundamental level, it seems that there have been no clear definitions of performance measures and reporting of those measures. In short, Coverage has not been clearly defined. Moreover, the relevant data sources have not been defined nor have the analytical methods to be used been determined. As the job of measuring the coverage of mobile broadband and voice networks appears to be incomplete, we suggest an examination along two lines. First, there needs to be a relationship established between the stated goal and the measures to be used to determine if the goal is being met. The figure below, using the preliminary measures proposed in this paper, demonstrates this.

45 | P a g e

Figure 26- Coverage Definition Image Sources: CostQuest Associates Analysis

Second, there needs to be an examination of all data sources and methods to understand what performance measure or measures each source can provide and the strengths and weaknesses of each in supporting the FCC goals.

Coverage>>

ServiceAvailability

Presence

Inbuilding,OnRoad

Measure

Speedvalues %success

Unit

Roads,Pop,Locations

Areaforsuccess

CB,CT,county,CDP,other Testofsuccess

SubscriberExperience

RetailSupport Roamingcharges BonafideComplaints

NetworkServiceQuality

Bandwidthbytechnology

PacketLoss PacketJitter

Failedcall/session

Voice Data

Droppedcall

Voice Data

Targetconnectionrate

4Gto4G 3Gto3G

ObligationRequirements

Offercolocationonnewstructures Offerroaming SIPs Voiceatreasonablerates Broadbandatreasonablerates Capacitycomparabletourbanareas OfferLifeline

46 | P a g e

Figure 27: Coverage Measures Table Source: CostQuest Associates Analysis

With that information in hand, answering what is covered can be objectively defined. Then, a discussion can move into funding determinations.

ID CoverageMeasure Mosaik NBM 477 DriveTest 1 AvailabilityofMobileVoicewhereAmericansLive 2 AvailabilityofMobileVoicewhereAmericansWork 3 AvailabilityofMobileVoicewhereAmericansTravel Advertised 4 AvailabilityofMobileBroadbandwhereAmericansLive(Populationmeasures)

• 5

AvailabilityofMobileBroadbandwhereAmericansWork(businesslocationmeasures)

• 6

AvailabilityofMobileBroabandwhereAmericansTravel(roadmeasures) Advertised 95% likely95%

• 7

Signalmeasuredat: 8 InBuilding

• 9

OnRoad Likely

• 10

Broadbandmeasuredon 11 Speed:Broadbandthatmeetsperformancemetricsforactualspeedsratherthan “advertised”or“upto”metrics

• 12

Ruralexperiments,landlineandnonedgemobilityrequirementsof4mbdownloadand 1mbupload 13 Mobilityauctionrequirementsof768kbdownloadand200kbupload 14 Latency:lowlatencytoenableuseofrealtimeapplications,suchasVoIP

• 15

CoveragemeasuredonDrivingTestResults

• 16

ActualspeedandlatencybemeasuredoneachETC’saccessnetworkfromtheenduser interfacetothenearestInternetaccesspoint

• 17

Geographicunitofmeasurementselectedandidentifiable:Roads,Pop,Structures,etc.. supports supports supports supports 18 GegraphicAreaformeasuringsuccessinidentifiablesuccessdefined supports supports supports supports 17 Bandwidthbytechnology 18 Packetloss 19 Packetjitter 20 Failedcall/sessionattemptsbytechnology 21 Voice 22 Data 23 Droppedcall%bytechnology 24 Voice 25 Data 26 TargetTechnologyConnectionrate 27 Percentsuccessa4GphoneconnectswithLTE

• 28

Percentsuccessa3Gphoneconnectswith3G

• 29

Voiceatreasonablycomparablerates 30 Broadbandatreasonablycomparablerates 31 Capacity:Usagelimitscomparabletousagelimitsinurbanareas 32 ServiceImprovementplansfiled 33 OfferLifeline 34 Providecollocationforotherprovidersonnewlyconstructedtowers 35 ComplywiththeCommission’svoiceanddataroamingrequirementsonnetworksthatare builtthroughMobilityFundsupport 36 Bonafideconsumercomplaintsper1000customers 37 Averagebilledroamingcharges 38 Proximitytolocationbasedretail/support Avaialble/ProposedCoverageDatasets ObligationRequirements NetworkServiceQuality SubscriberExperience ServiceAvailability

47 | P a g e

AddendumA

February 22, 2016 In the time since this paper was written, the FCC has released the first publicly available data associated with its new Form 477 data collection program. There’s no doubt that changes in availability and speeds have occurred over this time. However, given some limitations in this initial release of the Form 477 data, we are unable to yet conduct a comparative study between this data and prior data from the National Broadband Map, Mosaik and mobile testing data. The Form 477 data does not show presence of speed, there is no documentation explaining data sources, assumptions and validation methods with the data output, and we have not yet seen a second release of the data set from the FCC. Once these issues are resolved CostQuest will consider updating the analysis to be inclusive of this From 477 data. We have included a map (below) showing 4G/LTE unserved roads across the U.S. according to the initial From 477 data release.

WRITTEN STATEMENT

• f

LeROY T. CARLSON, JR. CHAIRMAN, UNITED STATES CELLULAR CORPORATION before the SUBCOMMITTEE ON COMMUNICATIONS, TECHNOLOGY, INNOVATION, AND THE INTERNET U.S. SENATE COMMITTEE ON COMMERCE, SCIENCE AND TRANSPORTATION FEBRUARY 4, 2016

2 Chairman Wicker, Ranking Member Schatz, and members of the Subcommittee, my name is LeRoy T. Carlson, Jr., and I am Chairman of United States Cellular Corporation. Thank you for the opportunity to discuss the need for mobile broadband in our nation’s rural areas and the important role that the Federal Universal Service Fund can play to address this need. Introduction. U.S. Cellular provides wireless service in nearly 200 markets across 24 states located in regional clusters across the country, including many of the states represented on this Committee such as Missouri, New Hampshire, Nebraska, Kansas, West Virginia, Wisconsin, and

• Washington. The overwhelming majority of the geography we serve is rural in character. We

have participated in the FCC’s universal service program for many years, using support to construct and operate network facilities in small towns and on rural roads that would not

• therwise receive service, because they would never prove to be economically feasible without

assistance. In each of our company’s previous appearances before this Committee to discuss universal service, we have made the point that Congress directed the FCC to ensure that rural citizens have access to modern telecommunications and information services that are reasonably comparable to those available in urban areas.1 Based on our deep experience in rural America, we have concluded that the current and proposed Mobility Fund mechanism lacks the necessary size and focus to ensure that rural communities have timely access to high-quality mobile broadband services needed to compete,

1 See, 47 C.F.R. § 254(b)(3).

3 here in the United States and around the world, for jobs and economic opportunities. We fear that policy makers have grossly underestimated the amount of work that remains to be done in rural America before mobile broadband can be deemed comparable to what exists in our nation’s urban areas. As explained below, we urge the Committee to direct the FCC to develop a more accurate picture of mobile coverage and mobile broadband availability in rural America, and to estimate how much it will cost to bring mobile broadband networks in rural America up to the reasonably comparable standard that Congress set. Once these tasks are done, Congress can make the policy choice as to how best to complete the task. Today, my testimony touches upon three things: (1) The critical role that mobile broadband plays in enabling public safety, education, and our rapidly expanding information economy; (2) the insufficiency of mobile broadband deployment in rural America today to meet stated goals; and (3) the need to make smart and creative policy choices to allocate and target scarce federal universal service funds to rural and high-cost areas to maximize the value of such investments in extending the reach of mobile broadband service.

• 1. The Rise of Mobile Broadband as an Enabler of Public Safety,

Education and Economic Development. In the 1980s, experts projected that there could be 800,000 mobile phones in use by

• 2000. They came up short by 10,000,000. Today there are over 350 million mobile wireless

subscriptions in the US. In 1984, the first commercial cell phone sold for $3,995.00. Today, there are more types of mobile wireless devices than I can list, capable of performing thousands upon thousands of tasks, at a small fraction of the 1984 price, with many having far more computing power than Apollo 11. However, looking back at how we have benefited from

4 mobile services dramatically undersells the future that consumers will enjoy, because we’re just getting started. At a time when consumer preferences are rapidly shifting to mobile broadband, policymakers must refocus universal service mechanisms to ensure that citizens in rural areas have access to high-quality service. For example, the Pew Research Center reports that adults living in households with a cellphone but no landline, and the number of households that rely solely on a smartphone for broadband have increased dramatically:2

2 See, http://www.pewresearch.org/fact-tank/2015/01/07/pew-research-will-call-more-

cellphones-in-2015/; and http://www.pewinternet.org/2015/12/21/home-broadband-2015/. Another barometer of consumer preference is mobile ad revenue, as evidenced in Facebook’s most recent quarterly report: “Mobile ad revenue reached $4.5 billion, up 81% year---over--- year, and is now 80% of total ad revenue.” See also, http://files.shareholder.com/downloads/AMDA-NJ5DZ/1421180082x0x872005/02B28FAD- 354C-4CA0-8CDE-3ADB6F8A4734/Q415_and_FY_2015_Earnings_Call_Transcript.pdf.

5 These compelling statistics gain further meaning when you consider just a few of the many benefits that mobile services provide: Public Safety. The ability to use 911/E-911/Text-to-911 depends 100% on high quality coverage, to fully enable location-based services.3 When disaster strikes, first responders depend on mobile wireless and broadband networks, which are the first to return to service. The value and utility of FirstNet, our nation’s mobile broadband public safety network, increase every time a new cell tower is constructed, as it provides a place to locate critical public safety communications equipment. Health Care. Mobile devices and applications capable of diagnosing, monitoring and treating various conditions are exploding into the marketplace and revolutionizing health care.4 These advances improve patient outcomes, and increase efficient delivery of services. It is now possible for a diabetic patient to continuously monitor, store, and transmit glucose levels to health care providers through a mobile device.5 Mobile video conferencing is increasingly important to emergency medical services and in delivering health care to remote areas where facilities are not easily accessible. These applications are but a small fraction of the incredible health care tools enabled by mobile broadband. The Internet of Things. Soon, almost any object will be capable of connecting to the Internet. Gartner expects 21 billion devices to be deployed by 2020.6 According to General Electric, the Industrial Internet, defined as the combination

• f Big Data and the Internet of Things, may be responsible for $15 trillion (not a

typo) of worldwide GDP by 2030.7 Most of these connected devices, numbering

3 As of November, 2015, the FCC estimates that 70% of 911 calls are placed from wireless

phones, and that percentage is growing. See, https://www.fcc.gov/consumers/guides/911- wireless-services.

4 A list of mobile medical applications can be found at:

http://www.fda.gov/MedicalDevices/DigitalHealth/MobileMedicalApplications/ucm368743.ht m

5 http://www.dexcom.com/g5-mobile-cgm. Someday soon, patients may wear a contact lens

that constantly measures glucose level through tears, transmitting the data to attending

• physicians. See, https://verily.com/.

6 See, http://www.gartner.com/newsroom/id/3165317. 7 See, http://www.ge.com/digital/sites/default/files/industrial-internet-insights-report.pdf

6 in the tens of billions, will need the flexibility that mobile wireless provides. The amount of data flowing through mobile broadband networks will dwarf what we see today. Cisco predicts that, between 2014-2019, U.S. mobile data traffic will rise seven-fold, driven by four billion new mobile connections, a 2.5X increase in throughput speeds, and mobile video traffic reaching 72% of all traffic.8

• Education. Students are increasingly using mobile devices to access learning

materials, do homework, create presentations, and communicate with teachers. Students with connectivity throughout the community are more likely to meet educational goals, especially in an age where learning through the Internet is essential.

• Agriculture. Connected tractors, irrigation systems, livestock management,

commodity tracking, and many more applications depend upon mobile wireless connectivity. Low-income households. For households that cannot afford to purchase a desktop computer and subscribe to both mobile and fixed networks, a single mobile device is capable of meeting voice communications and Internet needs. If the Committee takes nothing else away from these examples of how mobile wireless is enriching our lives, it should be this: None of the benefits described above will adequately benefit rural Americans unless high-quality mobile broadband coverage is available everywhere they live, work, and travel. In areas where emergency calls cannot connect, or where medical devices cannot transmit data, lives will be lost. In areas where tablets and laptops don’t work, educational

• pportunities will be foreclosed. The enormous power of the Internet of Things cannot be fully

realized without ubiquitous mobile broadband. As Deere & Company has previously noted to the FCC, a lack of connectivity on our nation’s farmlands costs productivity and wastes water

8 See, Cisco VNI Global Mobile Data Traffic Forecast, 2014–2019, accessed at:

http://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index- vni/white_paper_c11-520862.pdf.

7 and fertilizer.9 The lack of mobile broadband denies low-income households the opportunity to fully participate in our nation’s economy. Mobile broadband, which didn’t exist thirty years ago, and was considered a luxury item just ten years ago, is now an essential part of our lives. I cannot emphasize enough how important it is for Congress and the FCC to foster development of robust mobile broadband networks in rural areas. We are in just the second inning of a huge revolution in how Americans live their lives, a revolution that may never come to rural Americans who live in areas where it is too expensive to make a business case to build and upgrade networks. We at U.S. Cellular provide our customers with access to the applications they use, because we enable all of them. If coverage is weak or throughput is slow, devices will not work as designed. We note that new investments in mobile broadband infrastructure each year will have multiplier effects, creating jobs and stimulating economic growth.10 One wireless industry job supports over six additional jobs in the economy, almost one and one half times higher than that of the U.S. manufacturing sector.11 Each dollar of investment in wireless results in $2.32 of economic activity.12 In our experience, rural areas continue to support a tremendous amount

• f manufacturing, as well as a growing distributed service economy (for example, call centers

9 See, http://apps.fcc.gov/ecfs/document/view?id=7521752479. 10 See, http://www2.deloitte.com/content/dam/Deloitte/us/Documents/technology-media-

telecommunications/us-tmt-impactof-4g-060612.pdf.

11 See, Coleman Bazelon and Giulia McHenry, Mobile Broadband Spectrum, A Vital Resource for

the U.S. Economy, at pp. 19-20 (May 11, 2015), available at:

http://www.brattle.com/system/publications/pdfs/000/005/168/original/Mobile_Broadband_Spectrum _-_A_Valuable_Resource_for_the_American_Economy_Bazelon_McHenry_051115.pdf?1431372403.

12 Id.

8 and medical clinics). We hear directly from our employees and customers that managers and educated professionals no longer consider rural areas that lack high-quality mobile wireless services to be attractive to locate to, or to stay in. I’m sure members of this subcommittee have heard the same thing from their rural constituents. This is not just anecdotal evidence. Rural areas have large gaps with urban areas, which gaps need to be closed. Data from the Department of Agriculture reveals that “2010-2014 is the first period of overall population decline on record for rural America as a whole.” 13 The same report shows employment growth since the 2008 recession heavily skewing in favor of

• ur nation’s urban areas and a persistent rural/urban educational attainment gap:

13 See, USDA, Rural America at a Glance, 2015 Edition, accessed at:

http://www.ers.usda.gov/media/1952235/eib145.pdf (revised Jan. 2016).

9 One of the best ways to stimulate economic activity, attract talented people to areas needing an employment boost, and to increase educational opportunities, is to build mobile broadband infrastructure. It is therefore vital for policymakers to have accurate data about the state of mobile deployment in rural America. As a Committee that is forward-looking, I urge you to consider the essential role that mobile broadband services will play in the future, and to ensure that the universal service program provides sufficient resources to realize that future in rural areas.

• 2. Mobile Broadband Deployment in Rural America is Insufficient.

Let me continue by acknowledging that we are well aware of the misleading claim that the job of providing mobile broadband to rural America is largely finished.14 When the FCC proposed Phase II of its Mobility Fund in 2014, it stated, “According to some sources, nearly

14 See, http://www.theverge.com/2015/3/23/8273759/obama-administration-passes-goal-lte-

for-98-percent-of-americans.

10 99.5 percent of the U.S. population today (and the road miles associated with that population) is covered by some form of mobile broadband technology.”15 That statistic cannot be right. Based on our experience, the state of mobile broadband is nowhere near developed enough to conclude that rural Americans have access to a strong 4G LTE signal throughout the area where they live, work, and travel. In a recent letter to the FCC, Senator Manchin astutely called out problems with available mapping resources, stating “the reality in my state is far different than what the maps indicate.”16 Senator Manchin’s experience is far from an isolated case and I’m sure each of you know from personal experience in your own states that mobile broadband coverage with a strong signal is far from complete and dead zones remain to be covered. In testing our networks, and those of our competition, we can confirm that the National Broadband Map and

• ther publicly available mapping resources significantly overstate where rural citizens can

actually use their devices to access rapid mobile broadband service, especially on rural secondary roads and in agricultural areas.17

15 See, Connect America Fund, Report and Order, Declaratory Ruling, Order, Memorandum

Opinion and Order, Seventh Order on Reconsideration, and Further Notice of Proposed Rulemaking, FCC 14-54, 29 FCC Rcd 7051, 7127 (2014) (“Further Notice”).

16 See, Letter from Hon. Joe Manchin, III to Hon. Thomas Wheeler, September 22, 2015, at

http://www.manchin.senate.gov/public/index.cfm?a=files.serve&File_id=D660F970-2859- 46B3-8145-CFE461A47719.

17 For example, we’ve heard directly from Senator Tester that he can’t get any signal on and

around his working farm in Montana, and from Senator Brown that southeastern Ohio lacks coverage.

11 In its recently released Eighteenth Mobile Competition Report, the FCC states that 25%

• f road miles and 50% of square miles in the US do not have coverage by two or more carriers,

and concedes that its data sources likely overstate coverage.18 This is significant because there continue to be two incompatible wireless network technologies in use today – the GSM standard and its 3G successors, used by AT&T, T-Mobile, and a number of other carriers, and the CDMA standard, used by Verizon, Sprint, U.S. Cellular, and a number of other carriers. A person with a CDMA-only phone cannot complete a call when they are in an area served only by GSM, and vice-versa. As a result, the current reality in rural areas is a patchwork quilt of coverage by incompatible technologies, frustrating the goal of seamless access. Accordingly, for public safety, it is critical that rural Americans have access to wireless networks capable of connecting both kinds of devices, just as those who live in cities do. In the run up to the FCC’s 2011 Connect America Fund reforms, we warned of universal service mechanisms that pick a single winner in the auction room rather than allowing consumers to pick winners in the market. By limiting support to a single carrier, the current mechanism is promoting service by one carrier and one technology, thus limiting consumer choice in many areas that would otherwise support competition, and requiring additional

• regulation. We urge the Committee to encourage the FCC to adopt universal service

mechanisms that direct support to high-cost rural areas without picking a winner in advance.

18 See, Implementation of Section 6002(b) of the Omnibus Budget Reconciliation Act of 1993,

Eighteenth Report, FCC 15-1487 (Dec. 23, 2015) at p. 28, Chart III.A.3 (“Eighteenth Mobile Competition Report”).

12 Last year, we inaugurated new coverage and mobile broadband service in Paw Paw, West Virginia, a town of 500, a project that would not have been possible without the federal universal service program.19 There are many more towns similar to Paw Paw that we would like to serve or upgrade, if support mechanisms provide us with a reasonable opportunity to

• succeed. It is low population density and traffic levels that make new construction infeasible

and make necessary an effective universal service mechanism. Today mobile broadband coverage and throughput speeds in rural America must receive a grade of “Incomplete.” Using the “reasonably comparable” standard set by Congress in 1996, anyone telling you that rural Americans have access to mobile broadband networks that are reasonably comparable to those in urban areas has not taken a drive across this great nation. And that’s not a surprise - no carrier can be expected to invest unless there’s at least the possibility of earning a return. If it could be done, we wouldn’t need a universal service mechanism because it would have happened already. In sum, we cannot base critical policy choices on conflicting data and maps that the government admits overstate coverage. We must have accurate data in order to target funds where they are needed.

19 See, http://www.morganmessenger.com/news/2015-11-

18/Front_Page/Paw_Paw_welcomes_arrival_of_cell_service.html.

13

• 3. Allocating Scarce Federal Universal Service Funds Effectively

Requires Smart Policy Choices. Over the years, we have consistently advocated for a robust federal universal service fund that provides rural consumers with access to both mobile and fixed networks. We believe the FCC’s historical allocation of support to wireless networks has been insufficient to close up coverage gaps and deliver mobile broadband to many areas. As shown in the chart below, between 1999 and 2014 the FCC allocated over $50 billion in support to fixed networks and less than $12 billion to mobile networks.20 Over the next five years, fixed networks are projected to receive $22.5 billion in federal funding, while mobile networks are projected to receive $2.5 billion, a disparity in the universal service mechanism going forward of nearly 90/10. 21 With wireless consumers nationwide now contributing nearly half of the total federal Universal Service Fund of $9 billion (which includes E-Rate, Lifeline, Connect America Fund,

20 Source: Federal-State Joint Board Monitoring Reports, at

https://www.fcc.gov/general/federal-state-joint-board-monitoring-reports.

21 The fixed network allocation is estimated by summing Connect America Fund support with

projected support for rate of return carriers. The mobile network allocation derives from the FCC’s Further Notice, supra.

$0 $20 $40 $60 $80 Funding for Fixed Networks Funding for Mobile Networks

FCC Universal Service Allocation

1999-2014 2015-2020

Billions

14 Mobility Fund, and Rural Health Care)22 the proposed funding for mobile broadband does not accurately reflect consumer usage, preferences, and infrastructure needs in rural areas. Given rapidly expanding demand for high-quality coverage and fast broadband connections, the current level of funding shortchanges rural Americans who increasingly rely on mobile services. Nor does the FCC’s proposed budget account for investments that mobile wireless carriers have made over the years. Many carriers, including U.S. Cellular, have used support to build towers in areas so remote that revenues are insufficient to meet ongoing operating expenses and to earn a reasonable return. These investments were made with the understanding that support for ongoing operations would be made available, either in the

• riginal fund, or in Mobility Fund Phase II.

Although the FCC proposed to use at least some of Mobility Fund Phase II support to cover operating expenses on towers, it recently proposed to change course based on “substantial marketplace developments,” nothing more than fallacious claims by some carriers that the job of covering rural America is largely done.23 This course change may prove to be catastrophic for rural citizens in small communities, which often do not generate enough revenue to meet a tower’s operating expenses. In addition to our experience and the weight of data, I am troubled by these FCC claims

• f substantial marketplace developments leading toward a conclusion that a Mobility Fund of

22 The most recently available FCC report from 2011 containing assessable carrier revenues for

universal service can be accessed at: https://transition.fcc.gov/Bureaus/Common_Carrier/Reports/FCC- State_Link/IAD/quarterly_roll-upsasof050112.pdf .

23 See, Further Notice, supra, FCC Rcd at 7126-29.

15 less than $400 million annually may be appropriate. When it comes to broadband, I agree that we as a nation should be setting big and audacious goals and working toward them.24 In last week’s 2016 Broadband Progress Report, the FCC reported that 87% of rural Americans lack access to mobile broadband at 10 Mbps/1 Mbps: Because the above data for LTE technology is based largely on advertised coverage at a single point within a census block, I don’t agree that the job of populating rural areas with LTE technology is largely done. That said, if the data on 10/1 availability is even in the ballpark, it is beyond dispute that the job of getting to an adequate level in rural America is only beginning. Because the big carriers continue to provide their customers with access to many rural areas by using the networks of rural carriers, it is fair to conclude that the future of 10/1 Mbps service depends on a universal service policy that encourages rural carriers to invest, as well as an FCC spectrum policy that ensures rural carriers have access to sufficient bandwidth to deliver speeds of the future. The critical role of universal service is to ensure that broadband technologies being deployed and commonly used in urban areas are made available to our rural communities in a timely manner. This is no different than any other infrastructure, whether it be roads, electricity, or water.

24 See, Separate Statement of Commissioner Jessica Rosenworcel, at:

http://transition.fcc.gov/Daily_Releases/Daily_Business/2016/db0129/FCC-16-6A5.pdf.

16 My takeaway from the past several years of uncertainty is that the FCC has not devoted sufficient attention to determining how best to maintain the investments that have already been made, how much it will cost to fill in slow broadband zones and dead zones, and what it will cost to deliver 5G services, and more, to rural citizens in the coming years. The Mobility Fund Phase I auction left many areas still without coverage, and bidders forfeited back to the FCC nearly 25% of the $300 million in original funding, for a variety of reasons. The Commission has yet to act on our petition to distribute forfeited support to “next in line” bidders who could move quickly to build towers in many states that need investment. Moreover, the amount projected for Mobility Fund Phase II is insufficient to do the job on a reasonable schedule. In sum, the Mobility Fund program has not fulfilled the goal of fostering, “an environment in which the widest possible variety of new technologies can grow and flourish.”25 We also believe that the reverse auction approach for distributing mobile broadband support did not produce equitable results across the nation. Because reverse auctions allocate funds first to areas that cost less to serve, mountainous parts of the country are served last, or not served at all if funding runs out. I predict that with a reverse auction mechanism, many of you on this committee representing mountainous regions will never see your states receive meaningful assistance, even though the rhetoric of the program gives you false hope. And, we can assure members of the Committee representing flatter states that, based on our experience, the program is insufficient in those areas as well.

25 See, https://www.fcc.gov/news-events/blog/2015/08/03/leading-towards-next-generation-

5g-mobile-services.

17 In recognition of the fact that the fund is finite and consumer willingness to fund programs is an important factor, we suggest that the FCC solicit new ideas for how to leverage existing federal funds, in combination with state universal service mechanisms, and private investment, to provide an incentive for competitors to invest and improve service. Several states, such as for example, Nebraska, Colorado and New Mexico, have begun developing their

• wn broadband universal service mechanisms, any of which could be trialed in a pilot program,

something the FCC has recently done in the fixed service arena. We suggest that the FCC consider a grant program in which the combined federal and state support funds could be used in a targeted way to address those areas most in need of mobile broadband coverage. States may be in the best position of all to know what is adequately covered and what is not. States that have been shortchanged by the legacy program (paying into the fund far more than they have drawn out for mobile voice, let alone mobile broadband coverage) and are willing to contribute state funds to the mechanism, should be given an opportunity to access some level of support, especially where the need for expanded coverage has been established. Equitable distribution of funding will likely not occur if the fund is administered at the federal level in an auction format, which disfavors the highest cost rural areas. Separately, Congress can make all universal service fund support go farther by passing legislation to exclude universal service support from taxable income, similar to funds provided under the American Recovery and Reinvestment Act. By excluding support from taxation, we will be able to use 100% of the support received for investments in rural areas and not just the net amount after taxes.

18 Concluding Remarks Just last month, Verizon announced an intent to begin limited deployments of 5G technology as early as 2017, technology that will provide speeds perhaps 50 times faster than 4G.26 National carriers will continue to focus on urban areas, and they will invest billions upgrading networks to 5G. But make no mistake, these investments will take priority over building new coverage and upgrading rural areas that make less economic sense. In sum, if we fail to foster robust mobile broadband networks in rural areas, they will likely never have access to the amazing things described above. Having studied this industry for many years, I’m humble enough to know that this task is easier said than done, in part because in a nation of entrepreneurs and risk takers and innovation, if there were a business plan to cover all of rural America, the free market would have done it long ago. Making rural infrastructure reasonably comparable is a big and multi- faceted task, as evidenced by the enormous efforts the FCC has made in over twenty years since the 1996 Act. This year, we celebrate the sixtieth anniversary of Eisenhower administration’s enduring achievement, the federal interstate highway system. My sense is that broadband networks will be as important to our nation’s success in the next sixty years as our interstate highway system has been over the past sixty. Just as our highway needs have expanded, so too will our broadband needs, and it will be up to this Committee to give the FCC proper direction to ensure

26 See, http://www.pcworld.com/article/3025461/mobile/verizon-vows-to-build-the-first-5g-

network-in-the-us.html.

19 that rural Americans fully participate in modern life and remain comparable with their urban counterparts.