telecommunications reform in germany · comparison of the united states ... need to reform the...

Bonn, Germany 20 November 1997

American Institute for Contemporary German Studies The Johns Hopkins University

TELECOMMUNICATIONS REFORM IN GERMANY: LESSONS AND PRIORITIES

Conference Report

Conference Report

TELECOMMUNICATIONSREFORM IN GERMANY:

LESSONS AND PRIORITIES

Bonn, Germany20 November 1997

American Institute forContemporary German Studies

The Johns Hopkins University

ii

The American Institute for Contemporary German Studies (AICGS) is a center foradvanced research, study, and discussion on the politics, culture, and society of theFederal Republic of Germany. Established in 1983 and affiliated with The JohnsHopkins University but governed by its own Board of Trustees, AICGS is a privatelyincorporated institute dedicated to independent, critical, and comprehensive analysisand assessment of current German issues. Its goals are to help develop a new generationof American scholars with a thorough understanding of contemporary Germany, deepenAmerican knowledge and understanding of current German developments, contributeto American policy analysis of problems relating to Germany, and promoteinterdisciplinary and comparative research on Germany.

Executive Director: Jackson JanesResearch Director: Carl LankowskiBoard of Trustees, Cochair: Steven MullerBoard of Trustees, Cochair: Harry J. Gray

The views expressed in this publication are those of the author(s) alone. They do notnecessarily reflect the views of the American Institute for Contemporary GermanStudies.

©1998 by the American Institute for Contemporary German StudiesISBN 0-941441-37-7

This AICGS Conference Report paper is made possible through grants from the GermanProgram for Transatlantic Relations, AT&T and o.tel.o communications GmbH.Additional copies are available at $5.00 each to cover postage and processing from theAmerican Institute for Contemporary German Studies, Suite 420, 1400 16th Street, NW,Washington, D.C. 20036-2217. Telephone 202/332-9312, Fax 202/265-9531, E-mail:[email protected], Web: http://www.jhu.edu/~aicgsdoc

iii

C O N T E N T S

Foreword..................................................................................................v

INTERCONNECTION AND UNBUNDLINGPOLICY IN NORTH AMERICA

Richard Simnett..........................................................................1

INDEPENDENCE AND THE REGULATORYARRANGEMENT ISSUES IN INSTITUTIONALDESIGN

Richard J. Schultz......................................................................12

LONG RUN INCREMENTAL COSTS AND THEREGULATION OF INTERCONNECTIONCHARGES IN THE UK

Geoffrey Myers..........................................................................22

COSTING AND PRICING OF INTERCONNECTIONSERVICES IN A LIBERALIZED EUROPEANTELECOMMUNICATIONS MARKET

Günter Knieps ..........................................................................51

COSTING AND PRICING OF INTERCONNECTIONCHARGES IN THE U.S.: LESSONS FOR GERMANY?

Ingo Vogelsang...........................................................................74

INFRASTRUCTURE COMPETITION ANDLOCAL-LOOP UNBUNDLING

Martin Cave & Peter Crowther................................................102

UNIVERSAL SERVICE OBLIGATIONS:COMPARISON OF THE UNITED STATESWITH THE EUROPEAN UNION

Dr. Barbara A. Cherry............................................................113

iv

RECENT DEVELOPMENTS IN THE REGULATION OFINTERNATIONAL TELECOMMUNICATIONS

Dr. Andrea Huber..................................................................130

Conference Agenda............................................................................150

v

F O R E W O R D

As part of its focus on comparative public policy issues of importanceto the United States and Germany, AICGS has been monitoring thedevelopment of telecommunications reform in Germany before and afterthe passing of new telecommunication legislation in the summer of 1996and continuing through the establishment of the Regulatory Authority onTelecommunications, effective January of 1998. While the challengesGermany has faced in this fast-changing sector have been significant, thetelecommunications maker in Germany has adapted quickly toglobalized competition, assessing experiences of other countries,including the United States.

In order to shed light on the lessons of telecommunications reform,AICGS organized a conference on November 20, 1997 in Bonn duringwhich several presentations were made on the central issues revolvingaround telecommunication regulatory arrangements on both sides of theAtlantic. The conference was attended by telecommunications policyofficials and industry executives as well as research experts from theUnited States and Europe. The result of this conference wereincorporated into this publication.

We are grateful to the eight authors who provided us with theirassessments of the agenda in the ongoing developments intelecommunications reform. We also wish to express our deepappreciation to Prof. Juergen Mueller of the Berlin School of Economics(FHW) who provided both the intellectual leadership and theorganizational assistance in making both the conference and thepublication possible.

This publication was supported by a grant from the German Programfor Transatlantic Relations and the German Marshall Fund of the UnitedStates.

Jackson JanesExecutive Director December 1998

INTERCONNECTION AND UNBUNDLING POLICY IN NORTHAMERICA

Richard Simnett

1. INTRODUCTION

This short paper is intended to highlight some of the key elements ofregulatory policy in the U.S. and Canada in order to provide a perspective forEuropean regulators and other interested parties. I have drawn upon severalregulatory documents and rate filings to show that the context of rate setting inboth the U.S. and Canada is quite different from that in Europe, and that, unlessthe differences are understood, apparently similar policies (from verbaldescriptions) can turn out to have radically different consequences.

2. THE POLICY CONTEXT IN U.S. AND CANADA

The U.S. and Canadian telecommunications markets have been dominatedby government-regulated private monopolies for nearly all of the twentiethcentury. Similar regulated rate structures evolved, with long distance chargessubstantially higher than costs providing the funds for carriers to sustainresidential basic service rates at rates below their accounting costs, in manycases including unlimited free local calling. Business rates are higher, generallycovering their accounting costs. Both the U.S. and Canadian authorities haveintroduced competition into the telephone industry, and have responded indifferent ways to the needs for reform. Both sets of regulators recognize theneed to reform the distorted price structure so that efficient price signals are sentto entrants, and incumbents have a chance to compete. Residential raterebalancing to bring rates more into line with economic costs (as required forequitable treatment of the regulated firm in a competitive market) has been verycautiously approached with extended (five years or more) transition periods forany substantial change in residential line rates.

The policy context also differs from what Europeans might assume. TheU.S. Telecommunications Act of 1996 imposes obligations on incumbent localexchange carriers (ILECs) specifically, and on Bell Operating Companies(BOCs) as a special class within the ILECs. ILECs must offer unbundlednetwork elements, wholesale services for reseller competitors to use, andordinary retail and long distance carrier access services. Currently there areseveral different costing and pricing standards in effect for these differentservices, creating potential arbitrage opportunities. The U.S. legislative focuson ILECs is different from one focusing on dominant firms with market power

Telecommunications Reform in Germany

2

and the ability to abuse it, and Federal Communications Commission (FCC)proceedings have not yet reached the issue of under what conditions ILECsmight be deregulated. One state commission (Colorado) notified entrants thatthey too would become incumbents a number of years after they commencedservice. Entrants are almost wholly unregulated.

In Canada regulators have focused instead on the essential facilitiesdoctrine, drawn from antitrust law, and have decided to impose only thoserequirements which they believe will encourage long term facilities-basedcompetition and whose benefits will outweigh their costs. Some of theseobligations are to be applied to entrants too, since they may attain localdominance or might be able to leverage dominance of one relationship intoexcess profits in another. In particular, a customer selecting carrier A as hisprovider of access lines and local calling gives Carrier A the opportunity to sethigh call termination charges for callers (and their carriers) trying to reach thatcustomer. The Canadian Radio-Television and TelecommunicationsCommission (CRTC) has imposed interconnection obligations on entrants, andalso retained its powers to regulate these rates, terms and conditions. Canadianentrants are also required not to discriminate among other carriers forinterconnection and long distance services.

3. UNBUNDLED RATE STRUCTURES

In both the U.S. and Canada there is a substantial degree of geographicalrate variation, with access line charges varying inversely to the density of thearea in which the service is provided. It costs much more to serve areas with lownumbers of access lines per square mile than dense urban and suburban areas orhigh density buildings (and this is so whether average historic or incrementalforward-looking cost measures are used). Some of these cost differences havebeen passed on in retail rates, but social policies have kept residential ratedifferences within bounds, given universal telephone service goals. The costdifferences have been more fully passed on in the rates for unbundled accesslines in the U.S.

The initial FCC order on unbundling contains a table of upper bound ratesfor loops in each state, but readers who assume that this actually indicates therates which have gone into effect following state commission proceedings willhave made a substantial mistake. This is shown in the table below. The tablecontains unbundled two wire access line rates of large local exchange carriersin the U.S., excluding BellSouth, SNET and Sprint, and as can be seen the ratescan be as low as U.S.$3.72 per month in Illinois and as high as U.S.$135 permonth in Hawaii. In general the rates based on state cost studies are higher than

Lessons and Priorities

3

the FCC proxy cost model estimates. Most of these rates have been calculatedbased on incremental cost studies.

Table 1. Unbundled Access Line Rates in the USA, 1997

State FCC State Approved Rateupperbound (2wire loop)

Alabama 17.25 28.15 GTEArizona 12.85 21.76Arkansas 21.18 19.25, 32.50, 73.05 Varies by densityCalifornia 11.10 16.81 GTEColorado 14.97 17.00-82.00,

average 20.65 Varies by densityDelaware 13.24 10.07, 13.13, 16.67 Varies by density,

under appealDC 10.81 10.81Florida 13.68 20.00 GTEHawaii 15.27 14.55 Oahu, 25.38 Maui,

28.78 Kauai 10.88 Hawaii,43.84 Lanai 135.20 Molokai

Illinois 13.12 3.72, 10.02, 11.53 Varies by density;Ameritech rates

Indiana 13.29 12.19 Ameritech ratesIowa 15.94 28.12 GTE

12.72 USWKansas 19.85 19.65, 26.55, 70.30 Varies by densityKentucky 16.70 19.65 GTEMaine 18.69 17.53 BAMaryland 13.36 11.87, 12.09,

16.13, 19.38 Varies by densityMassachusetts 9.83 7.54, 14.11, Varies by density; TELRIC

16.12, 20.04 study filed 14 Feb 1997Michigan 15.27 9.31, 11.84, 14.67 Varies by densityMinnesota 14.81 28.60 GTE

12.03 USWMissouri 18.32 10.50/11.54,

16.92/ 19.78, Varies by density27.63/ 32.07 and interconnector

Montana 25.18 27.41New Hampshire 16.00 12.67, 15.59, 23.00 Varies by densityNew Jersey 12.47 11.95, 16.02, 20.98,

16.21 average Varies by densityNew Mexico 18.66 19.49, 21.30, 26.74,

21.21 average Varies by densityNew York 11.75 12.49, 19.24 Varies by density,

PUC order April 97Ohio 15.73 15.73 GTE, 8.36

11.68, 13.73 Ameritech Varies by densityOklahoma 17.63 20.70, 27.75, 49.30 Varies by densityOregon 15.44 15.00 GTE

16.00 USWPennsylvania 12.30 11.52, 12.71,

16.12, 23.11 Varies by density


4

Rhode Island 11.48 17.53 BATexas 15.49 25.49/30.00 GTE Varies by line

15.00/ 17.00 SBC quality (signal loss)Utah 15.12 22.97Vermont 20.13 17.53 BAVirginia 14.13 10.16 GTE, 9.52

13.31, 19.54 BA Varies by densityWashington 13.37 13.62 GTE

11.33 USWW. Virginia 19.25 14.49, 22.04, 43.44,

24.58 average Varies by densityWisconsin 15.94 8.10, 12.80, 13.84 Varies by density;

Ameritech rates

Additional charges are needed for interconnecting carriers to actually use theseloops: there are non-recurring charges per order and per line, charges for centraloffice space, line terminations, and links to these lines, among others.

Other U.S. Services for CarriersThe U.S. has several different rate regimes which to European eyes may fall

into the area of unbundling or network interconnection. Any retail serviceoffered is subject to resale by other carriers, and in addition, wholesale servicediscounts for resellers apply to any LEC retail services. There has been muchregulatory dispute about the proper level of these discounts, which are supposedto be calculated to reflect the avoided cost of retailing by the incumbent.Entrants, by and large, have found that the discounts are insufficient for them tomake a profit in competition with the incumbents’ retail services, and not muchresale has taken place. The wholesale rate structure is thus pegged to the retailrate structure, with all of its cross-subsidies and non-economic features intact.

The third set of services for carriers is carrier access, offered for longdistance carriers to complete their services by using local carrier networks.These rates were based on 25 percent of the total fully distributed historic costof providing local exchange service before price caps were applied, and includethe toll subsidy to local service (or the old implicit universal service fund andrecovery of residential access deficits) by being set at rates well aboveincremental costs. The FCC has reduced the economic distortion this involvesby two separate rebalancing proceedings over the years. The first was thecreation of subscriber line charges (flat rate per month, paid by the retailcustomer) which were increased over a number of years and capped at $3.50 permonth for residential customers to avoid “rate shock.” These are to be increasedfor lines beyond the first line per household, and other measures are also to betaken to rebalance access charges. Second, flat rate presubscribed line charges(to be paid by the long distance carriers) are also to be introduced over a number


5

of years. Gradualism in adjusting rate structures is thus the order of the day.Most of the implicit subsidies are going to be made explicit by the new universalservice fund and access charge regimes.

Canadian Services for InterconnectionDeaveraging of unbundled access line rates is also found in Canada. The

CRTC Decision 97-08 provides a table of unbundled loop rates in Appendix 1.The monthly and non-recurring charges are presented below, but there areadditional charges for space, links and so on.

Table 2. Unbundled Loop Rates in Canada

Company Monthly Rate, Non-recurring Non-recurring$C Charge per order Charge per loop

BCTel 17.40-32.35 166.00 82.50Bell Canada 24.30-53.65 152.00 171.00Island Tel 17.30-34.35 140.00 114.00MTS 13.90-26.55 98.25 141.00MT&T 20.40-38.20 145.00 118.00NBTel 36.45-48.05 145.00 132.00NewTel 34.70-48.85 153.00 76.50TCI 16.00-23.55 27.75 79.25

Canadian unbundling policy contains a feature not found in the U.S. TheCRTC has found that there is a presumption in favor of competitors buildingtheir own networks so that competition will reach all parts of the business. Theyalso recognize that there are areas (particularly of low density) where there islittle likelihood of competing infrastructures, and in these areas the incumbents’network access lines may properly be treated as “essential facilities.” The testapplied is whether other carriers could build their own facilities to compete. Inhigher density areas it is clear that rivals can build, and are, particularly to servebusiness customers and high density housing developments. Incumbents’networks are thus not, strictly speaking, essential facilities. Nevertheless, theCRTC imposed a duty to offer these for sale as unbundled lines but for only afive-year transitional period to allow time for entrants to build their ownnetworks without complete reliance on the incumbent. These are the cheaperrates from each company in the table above.

The CRTC considered the introduction of wholesale services, butconcluded that there would be little if any cost saving from selling to carriersrather than to retail customers and did not order that they be introduced. They


6

did order that local services (including term and volume discounts) be subjectto resale, with the restriction that cross-subsidized residential service may onlybe resold to residential customers. The CRTC also found that access by entrantsto the incumbents’ operations systems (for service provisioning, networkoperations and so on) would not be in the public interest given that its costswould be high. Access to these systems is required in the U.S. by FCC order.

Canadian toll services will continue to subsidize local services through anexplicitly identified “contribution” charge. This is basically a flat-rate monthlycharge per trunk connected to the local network, levied to support the localnetwork and separate from incremental-cost-based charges for the use of thelocal network. Canadian regulators have also taken steps to rebalance rates,with an approved annual increase of C$2 per month per line for a number ofyears. This is similar to the U.S. FCC’s gradual introduction and increase ofnew residential flat rate charges, and quite different from the Germanregulator’s disallowance of continued long distance service support of lowmonthly line charges by interconnection rates.

4. RETAIL RATE STRUCTURES

Both the U.S. and Canadian retail rate structures contain a substantialdegree of cross-subsidy, both from long distance services to local service, andfrom local business rates (which are generally at or above cost) to residentialservice which generally is priced below cost. Long distance rates arethemselves geographically averaged within U.S. states and on all interstateservices, but volume, term and contract discounts are generally available.

The tables below are from the most recent FCC reference book ontelephone service, issued in March 1997 and reporting the results of October1995 surveys. Since that time no major changes in local rates have been made.Table 3 shows the company-wide averages for major companies, while tables 4and 5 show basic residential and business flat rate service charges.

Table 3 shows, by comparing the residential and business rates with eachother and the average, the degree of cross-subsidy sent to residential service: thenational average is more than $15 per month lower than business rates, and thisis before usage charges are considered.

Table 3. Average Basic Rates by Company, October 1995

Ameritech Bell Bell-South NYNEX Pacific SBC U.S. West Inde- TotalAtlantic Bell pendents

Residential $16.29 $16.66 $17.44 $21.35 $14.37 $14.93 $17.51 $16.77 $16.93Single-Line


7

Business $32.72 $31.60 $38.12 $37.20 $24.85 $27.57 $36.00 $31.79 $32.76Multi-Line

Business - Key $34.16 $30.52 $55.89 $37.28 $25.31 $33.70 $42.45 $36.94 $37.72Multi-Line

Business - PBX $35.35 $36.23 $62.89 $40.80 $25.31 $38.29 $45.81 $43.41 $41.89Weighted

AveragesFor access line $22.38 $22.36 $28.66 $28.35 $18.33 $21.15 $24.70 $22.09 $23.81Additional

for touch-tone $0.77 $1.40 $0.98 $0.88 $0.00 $0.33 $0.23 $0.62 $0.66

Tables 4 and 5 show the results by the 100 sampled cities reported by theFCC, and while none of these is a rural area they give some idea of thegeographic variation in retail rates for access lines. This is also in markedcontrast to the situation in Europe where national rates available to bothbusiness and residential customers are more common. Both of these implicitdifferences in background situation need to be borne in mind when assessingNorth American policy statements or literature for possible application toEurope. In particular, the idea that non-discrimination in rates requires the samerate for all customers cannot be supported from North American practice. Acontrasting definition has gained some currency with regulators: that it isdiscriminatory to treat differently situated customers the same, as well assending economically inefficient market signals to potential entrants. The issueof discrimination becomes one of showing cost differences or satisfying theBurden Test: that a lower rate for some set of customers produces more profitthan higher rates, and thus enables lower rates than would otherwise be possiblefor the apparently discriminated-against class of customers.

Table 4. Residential Rates in the U.S., October 1995

State City Generally available Generally availableunlimited calling rate, touchtone rate, withrotary dial inside wire maintenance

AL Huntsville $23.01 $24.51AK Anchorage $14.47 $15.97AZ Tucson $19.17 $20.42AR West Memphis $28.42 $30.67

Pine Bluff $21.96 $24.21CA Anaheim $15.49 $15.99

Bakersfield $15.49 $15.99Fresno $15.49 $15.99Long Beach $23.53 $24.48Los Angeles $16.97 $17.47Oakland $16.60 $17.10Salinas $16.38 $16.88San Bernadino $23.21 $24.16San Diego $15.49 $15.99


8

San Francisco $15.49 $15.99San Jose $16.23 $16.73

CO Boulder $21.61 $23.56Col. Springs $19.78 $21.73Denver $21.21 $23.16

CT Ansonia $18.70 $20.05Norwalk $17.60 $18.95

DC Washington $19.50 $21.45FL Miami $16.96 $19.46

Tampa $17.61 $18.61W. Palm Beach $15.65 $18.15

GA Albany $19.15 $22.45Atlanta $23.32 $26.62

HI Honolulu $19.58 $22.73IL Chicago* $17.21 $19.71

Decatur $20.19 $22.69Rock Island $20.76 $23.26

IN Indianapolis $19.77 $21.77Terre Haute $22.13 $24.43

IA Fort Dodge $14.03 $15.28KY Louisville $22.65 $25.65LA Baton Rouge $20.99 $23.08

New Orleans $20.02 $22.11ME Portland $17.99 $18.74MD Baltimore $24.88 $25.73MA Boston $22.01 $23.94

Hyannis $22.01 $23.94Springfield $22.01 $23.94

MI Detroit $16.71 $21.39Grand Rapids $15.40 $20.08Saginaw $16.19 $20.87

MN Detroit Lakes $18.57 $21.72Minneapolis $20.39 $23.54

MI Pascagoula $24.93 $28.23MO Kansas City $18.12 $21.12

Mexico $16.91 $19.91St. Louis $18.18 $21.18

MT Butte $18.22 $20.17NE Grand Island $21.81 $23.76NJ Phillipsburg $11.96 $14.20NM Alamogordo $20.77 $22.72NY Binghamton $26.22 $26.74NY Buffalo $30.88 $31.40

Massena $23.60 $24.12New York City* $25.28 $25.80Ogdensburg $24.26 $24.78Rochester $16.81 $18.80

NC Raleigh $17.21 $20.21Rockingham $15.67 $18.67

OH Canton $19.18 $21.98Cincinnati $20.28 $22.60Cleveland $19.18 $21.98Columbus $19.18 $21.98Toledo $19.18 $21.98

OR Corvallis $18.84 $18.84


9

Portland $22.12 $22.12PA Allentown $16.62 $18.81

Ellwood City $15.75 $17.94Johnstown $20.05 $23.05New Castle $14.00 $16.19Philadelphia $18.87 $21.06Pittsburgh $17.71 $19.90Scranton $16.62 $18.81

RI Providence $23.47 $24.42SC Beaufort $19.76 $21.01TN Memphis $18.66 $21.41

Nashville $17.72 $20.47TX Brownsville $15.05 $17.73

Corpus Christi $15.59 $18.27Dallas $17.66 $20.34Fort Worth $16.46 $19.14Houston $18.15 $20.83San Antonio $16.29 $18.97

UT Logan $15.96 $17.91VA Richmond $23.96 $24.81

Smithfield $14.64 $18.14WA Everett $18.91 $20.41

Seattle $16.45 $18.40WV Huntington $27.15 $29.50WI Milwaukee* $15.90 $18.40

Racine* $15.86 $18.36

* The measured service rate plus 100 five minute, same zone, business day calls is shown because unlimited

local service is not offered.

Source: FCC Reference Book, March 1997

Table 5: Business Rates in the U.S., October 1995

State City Single PBX Line Single Business PBX LineBusiness Line With wire With wireLine maintenance maintenance

AL Huntsville $56.01 $73.60 $57.51 $75.10AK Anchorage $31.05 $41.13 $32.55 $42.63AZ Tuscon $41.69 $59.98 $43.69 $61.98AR West Memphis $53.42 $65.68 $56.17 $68.43

Pine Bluff $40.72 $51.01 $43.47 $53.76CA Anaheim** $30.35 $32.04 $31.35 $33.04

Bakersfield** $30.98 $32.67 $31.98 $33.67Fresno** $30.98 $32.67 $31.98 $33.67Long Beach** $43.93 $52.24 $45.88 $54.19Los Angeles** $33.25 $35.10 $34.25 $36.10Oakland** $32.53 $34.33 $33.53 $35.33Salinas** $32.87 $34.65 $33.87 $35.65San Bernadino** $43.34 $51.54 $45.29 $53.49San Diego** $30.35 $32.04 $31.35 $33.04San Francisco** $32.53 $34.33 $33.53 $35.33San Jose** $31.80 $33.57 $32.80 $34.57


10

CO Boulder $47.26 $58.73 $50.01 $61.48Col. Springs $43.82 $54.56 $46.57 $57.31Denver $46.79 $58.23 $49.54 $60.98

CT Ansonia $43.70 $49.53 $46.35 $52.18Norwalk $40.86 $46.69 $43.51 $49.34

DC Washington** $33.61 $34.21 $36.61 $39.27FL Miami $40.74 $81.59 $43.24 $84.09

Tampa $37.83 $65.38 $38.83 $66.38W. Palm Beach $37.43 $75.16 $39.93 $77.66

GA Albany $39.67 $63.03 $41.67 $67.33Atlanta $58.76 $93.28 $60.76 $97.60

HI Honolulu $44.40 $69.15 $46.15 $73.06IL Chicago** $32.04 $32.75 $33.54 $34.25

Decatur** $35.98 $36.69 $38.48 $39.19Rock Island** $36.55 $37.26 $39.05 $39.76

IN Indianapolis $55.92 $62.49 $57.42 $63.99Terre Haute $44.16 $65.88 $45.66 $68.98

IA Fort Dodge $22.42 $37.29 $23.27 $38.14KY Louisville $61.12 $93.47 $63.12 $98.56LA Baton Rouge $47.68 $73.56 $49.68 $75.74

New Orleans $48.17 $73.83 $50.17 $76.01ME Portland $38.63 $62.35 $40.58 $64.30MD Baltimore** $43.50 $46.12 $44.50 $47.12MA Boston** $42.81 $47.99 $44.76 $51.40

Hyannis $46.91 $72.87 $48.86 $76.28Springfield** $38.92 $44.09 $40.87 $47.50

MI Detroit** $37.30 $40.66 $39.55 $45.41Grand Rapids** $35.57 $38.79 $37.82 $43.54Saginaw** $37.47 $40.69 $39.42 $45.14

MN Detroit Lakes $42.40 $51.63 $44.40 $55.83Minneapolis $54.96 $64.91 $56.96 $69.11

MI Pascagoula $56.21 $82.55 $58.21 $87.55MO Kansas City $45.48 $60.66 $48.73 $63.91

Mexico $35.55 $47.93 $38.80 $51.18MO St. Louis $45.15 $60.22 $48.40 $63.47MT Butte $43.82 $54.63 $46.57 $57.38NE Grand Island $47.80 $68.17 $49.80 $70.17NJ Phillipsburg** $27.60 $27.42 $28.55 $30.36NM Alamogordo $56.15 $68.13 $58.90 $70.88NY Binghamton** $49.99 $53.21 $55.70 $60.29

Buffalo** $50.87 $54.16 $56.58 $61.26Massena** $49.70 $52.90 $55.41 $59.99New York** $51.23 $54.54 $56.94 $61.65Ogdensburg** $51.08 $54.37 $56.79 $61.46Rochester** $48.44 $54.39 $50.94 $60.76

NC Raleigh $41.53 $72.75 $44.03 $76.75Rockingham $36.09 $62.75 $38.59 $66.75

OH Canton** $44.25 $54.26 $46.25 $56.26Cincinnati $52.99 $70.26 $55.49 $74.45Cleveland** $43.22 $53.23 $45.22 $55.23Columbus** $43.22 $53.23 $45.22 $55.23Toledo** $44.25 $54.26 $46.25 $56.26

OR Corvallis $37.10 $45.33 $39.10 $47.33Portland $42.84 $51.79 $44.84 $53.79

PA Allentown** $37.28 $38.77 $38.23 $41.52


11

Ellwood City** $36.54 $40.22 $38.04 $43.62Johnstown** $37.28 $41.99 $38.78 $45.97New Castle** $38.84 $40.22 $40.34 $43.62Philadelphia** $30.64 $32.02 $32.14 $35.42Pittsburgh** $31.67 $33.05 $33.17 $36.45Scranton** $36.00 $37.49 $37.50 $40.79

RI Providence** $43.59 $45.17 $45.54 $47.12SC Beaufort $38.04 $66.61 $39.29 $67.86TN Memphis $54.84 $92.81 $56.09 $97.21

Nashville $52.30 $88.46 $53.55 $92.71TX Brownsville $31.13 $47.76 $33.88 $51.81

Corpus Christi $31.74 $48.36 $34.49 $52.41Dallas $38.35 $60.02 $41.10 $64.07Fort Worth $34.39 $53.00 $37.14 $57.05Houston $41.22 $65.41 $43.97 $69.46San Antonio $34.03 $52.45 $36.78 $56.50

UT Logan $32.13 $58.35 $34.88 $61.10VA Richmond $75.13 $119.35 $78.13 $124.26

Smithfield $29.97 $60.76 $31.72 $64.51WA Everett $39.92 $59.48 $41.67 $61.23

Seattle $37.23 $55.38 $39.23 $57.38WV Huntington $73.37 $114.59 $76.37 $120.34WI Milwaukee** $37.50 $39.05 $39.25 $40.80

Racine** $39.14 $40.68 $40.64 $42.18

** The measured service rate plus 200 five minute, same zone, business day calls is shown because unlimitedlocal service is not offered.

Source: FCC Reference Book, March 1997


12

INDEPENDENCE AND THE REGULATORY ARRANGEMENTISSUES IN INSTITUTIONAL DESIGN

Richard J. Schultz

All countries that seek to transform their telecommunications industry fromthe traditional monopolistic structure to one based on competitive markets mustconfront complex questions about the appropriate forms of alternative publiccontrol institutions to guide and supervise the restructuring. This is particularlytrue in the case of countries which have traditionally relied on public ownershipof telecommunications enterprises to accomplish public policy objectives, butis no less the case for countries, such as Canada and the United States, whosetradition has been to permit private ownership of the telecommunicationssystem subject to public regulation. In the latter case traditional regulation,which was premised on monopoly provision, has had to be revisited so as toensure that the goals and instruments of the regulatory system are appropriatefor contemporary needs. As a result in these countries there have beensubstantial revisions made in recent years to the basic regulatory statutes and setof institutions and instruments.

The design of telecommunications regulatory institutions is a complexundertaking that must respect national administrative and legal cultures whilesimultaneously responding to domestic and increasingly international pressuresas a result of the General Agreement on Trade in Services and the creation of theWorld Trade Organization. This paper will concentrate on only one set of issuesassociated with the nature of the regulatory system, although it is one that isparamount for all countries. In particular, the paper adopts as its organizingtheme the comment from one of the participants, Robert Crandall of TheBrookings Institution, at the Washington workshop on “TelecommunicationsReform In Germany: Lessons from North America” that “there is no such thingas an independent regulatory agency.” While this is undoubtedly true in thesense that no regulatory agency is completely or totally independent of politicalcontrol, it nevertheless ignores the importance that both traditionally andcurrently is placed on the concept and belief that regulatory agencies must havesome degree of independence if they are to perform the functions assigned tothem. As a generalization, however, Crandall’s statement raises someimportant issues that go the heart of the current attention being paid to thedesign of regulatory arrangements by individual countries.

The issue of regulatory independence can be summed up in threeinterrelated questions: Independence for whom? Independence from whom?Independence for what purposes?


13

Inasmuch as both Canada and the United States have experience, asindicated, with independent regulatory agencies, this paper will review thedebates and answers to these questions based on that experience.

1. INDEPENDENCE FOR WHOM?

This is the easiest of the three questions to answer. Members of theregulatory agency, or in the case of the recently-created British agencies suchas the Office of Telecommunications (OFTEL), the Director-General, areappointed for fixed terms such as five or seven years and as such are giventenure during that period. The significance of this tenure is that they serveduring “good behavior” and not at the pleasure of those who appointed them(Sommer 1987). Such tenure is presumed to give them independence to makeregulatory determinations free of concern for their positions, although thespecified nature of their tenure may, as some have suggested, make themsomewhat more responsive, nearer the end of their term, to the interests of thosewho appointed them. In the case of the American independent agencies, such asthe Federal Communications Commission, the independence of the regulatorsis supposedly reinforced by the bi-partisan composition of the Commission andthe staggered terms of the Commissioners. As a sign of both the possibleturbulence within regulatory arrangements that will be discussed below and theconditional nature of the independence of the regulators, the CanadianGovernment limited its most recent appointments to the Canadian Radio-Television and Telecommunications Commission (CRTC) to three year termsinstead of the statutory and traditional term of five years. Presumably theobjective of such shortened terms was to remind such appointees that they werenot to be too independent but were to be held to a short leash.

2. INDEPENDENCE FROM WHOM?

This question is also relatively easy to answer. Regulators were givenindependence through security of tenure for a set period in order to providethem with a degree of protection from political interference, legislative orexecutive (Cushman 1941; Willis 1941). Here it is important to recall thecircumstances under which the first independent regulatory agencies werecreated in North America, whether it be the Interstate Commerce Commission,which was the first American national independent agency created in 1887, orthe Board of Railway Commissioners in Canada, which was created in 1903 andassumed responsibility for telephone regulation in 1906. In both countries,regulation was a relatively novel function of government and the concern,


14

backed in the United States by the “no takings” constitutional protection in theFifth Amendment, was that private investors should be protected from arbitraryor political interference. If governments were to assume the power to controlsome aspects of corporate behavior, they should be insulated against politicalopportunism. In contemporary language, corporate interests were insistent onthe need for regulatory commitment and a relatively accurate understanding ofthe future policy environment (Levy & Spiller 1996). It should also be recalledthat independent, specialized agencies were created in part because of thepresumed inadequacies of existing public institutions such as the courts andexecutive or public service institutions for the performance of the newregulatory functions of government.

In the United States, the concern for regulatory independence wasreinforced by the fear that the president should not be able to control the newagencies; hence the requirements that appointments had to be bipartisan innature. This meant that the president could only nominate a bare majority of thecommissioners on any designated independent regulatory agency from hisparty while the remainder came from the other party. Furthermoreappointments were subject to Senate confirmation. These provisions were notmeant to produce “partisan” regulators but merely to prevent either executive orlegislative control.

In Canada the concern that regulators should be insulated and protectedfrom political interference in the course of their decisionmaking led thegovernment approximately twenty years ago to issue instructions prohibitingpolitical authorities, specifically cabinet ministers, from contacting regulatorsduring the exercise of their regulatory responsibilities for individual cases(Kane 1980). The prime minister of the day likened regulatory independence tojudicial independence and made the penalty for breaching this conventionsimilar to that for approaching a judge during the exercise of a judicial function,namely resignation from the Cabinet. The situation in the United States is lessclear-cut inasmuch as both members of the executive and particularly Congresshave over the years insisted on their right to discuss even current cases beforeregulatory agencies with regulators (Krislov & Musolf 1964).

Given Robert Crandall’s comment cited above that “there is no such thingas an independent regulatory agency,” it is worth noting the qualifications onregulatory independence that do exist. In both Canada and the United States, thebudgets of the regulatory agencies are subject to political scrutiny and controlin ways that those of traditional judiciary are not. In the former this has been animportant tool to control some aspects of regulatory behavior while in theUnited States the appropriations process is an ongoing and direct instrument forpolitical control of the agencies, especially at the macro level of the agency but


15

also with respect to individual policy initiatives and on occasion even specificdecisions (United States Senate Committee on Government Operations 1977;Davidson & Vietor 1985). Certainly budget approval is a way in both countriesto “send messages” to the regulatory agencies.

In the United States, the congressional oversight function is also animportant check on the independence of the agencies. More importantly overthe past two decades there have been repeated attempts to subject regulatoryagency rule-making to both legislative vetos or executive direction (Cutler &Johnson 1975; Kaiser 1980; Saks 1984). Another check on the independence ofAmerican regulatory agencies is the fact that the chair of each agency serves inhis or her capacity as chair but not, it is important to note, as member of theagency at the pleasure of the president. Selection of the chair, or alternativelyremoval, is deemed to be an important instrument for sending policy signals toagencies.

In Canada, one of the most important limitations on regulatoryindependence is the provision for appeals to the Cabinet against regulatorydecisions; in addition, in some cases the Cabinet may on its own initiativereview such decisions (Schultz 1977). There is no standardized systeminasmuch as in some cases the Cabinet is limited to sending decisions back forregulatory review or setting them aside while in others, notably in thetelecommunications sector, the Cabinet can change or vary the decision inquestion. We shall return to this political override power below because forsome it constitutes a denial of the independence of the regulatory agency whilemore importantly for others the process of political appeals which are highlyconfidential in nature would appear to violate the transparency requirements ofthe telecommunications annex of the General Agreement on Trade in Services.

Although there is no political appeal mechanism per se available to eitherthe Executive or Congress in the United States, Congress does have the powerthrough legislative action to override individual regulatory decisions. Over thepast two decades such a power has been used. Perhaps more significantly theinvocation of the threat to do so, as for example in the case of the FCC’s initialattempt to introduce a subscriber line charge in the early 1980s, can be aneffective instrument to communicate congressional wishes to independentagencies.

As the preceding suggests it is important to emphasize that regulatoryagencies in Canada and the United States do not have complete or absoluteindependence from political authorities. In other words, to expand onCrandall’s comment, “there is no such thing as an absolutely independentregulatory agency.” That said, the relative independence of such agenciesshould be recognized both conceptually and empirically. Regulators in the


16

exercise of their regulatory decisionmaking, or more precisely for the exerciseof that power, have judicial-like security and cannot be removed from theiroffices. Such security of tenure is what distinguishes so-called regulatoryagencies from other executive bodies in both countries whose members serve atpleasure and thus can be removed for whatever reason or impulse that those whomade the appointments in the first instance choose to invoke to justify suchactions.

It is worth noting that in both countries prior to recent decadestelecommunications regulatory authorities were, and were seen to be,independent. In the case of Canada, for example, notwithstanding the broad-based nature of the power of Cabinet to overturn individual regulatorydecisions, no such action was taken between 1906 and 1973. Politicalauthorities in both countries appeared to value the importance of the “arm’s-length” relationship that existed between them and their regulatory agencies. Itis only in the last two or three decades in both Canada and the United States thatdebates have emerged over the proper scope of regulatory independence and tounderstand why we must turn to our third question.

3. INDEPENDENCE FOR WHAT PURPOSES?

No one would challenge the argument that there is today a continuingdebate over the merits and appropriate extent of regulatory independence. Onthe one hand, there are those who argue that regulators must be given themaximum amount of independence possible in order to give service providersand investors, whether domestic or foreign, assurance that they will not besubject to arbitrary and/or unpredictable political interference. Advocates ofthis position seek, if anything, to enhance regulatory independence. On theother hand there are those within individual countries who maintain thattelecommunications regulation is far too important, given the ongoingtelecommunications and information revolution, to be left to the control ofarm’s length, independent regulatory authorities. Their prescription is tosubject regulatory authorities to more and more political constraints andcontrols. At its heart this is a debate, however, not about independence per sebut of the fundamental purposes of regulatory agencies. This is a debate thatrages in both Canada and the United States where there has been a traditionalreliance on independent agencies as well as in those countries which arecreating such agencies to cope with the imperatives of sectoral restructuring andinternational trade constraints.

To appreciate the controversies that have ensued it is important tounderstand the traditional purposes of independent regulatory agencies in


17

comparison with contemporary rationales for such agencies. Traditionally inCanada and the United States, telecommunications regulation, like itscounterparts in other sectors, was narrowly construed (McManus 1973;Armstrong & Nelles 1986). The emphasis in institutional design debates was onimpartiality, non-partisanship and the maximum avoidance of politics possibleto minimize corruption, abuse and favoritism. Regulators had a reactive,proscriptive function. They were to be society’s economic policemen to correctunacceptable behavior by regulated firms. They were not to be sectoralmanagers. Rather they were largely irrelevant, and to a large extentuncontroversial, because the private firms were the primary decisionmakersboth for the design of the system, its development or extension and for suchthings as the pricing and costing of services. Regulators were largelyadjudicators of disputes enjoined to balance the interests and demands oftelecommunications firms on the one hand and their subscribers on the other.

Starting in the 1930s, particularly with the American New Deal, a newconception of the role of regulatory agencies was adopted in some sectors, ofwhich transportation is perhaps the most striking example. The state’s role, andthat of its primary institutional agent, the independent regulatory agency, aseconomic policeman was no longer perceived in some circles to be sufficient.The reactive, case-specific, corrective focus was found to be wanting in the faceof economic turbulence, or partisan claims of such turbulence. The result was adebate about the appropriate role of regulation and regulatory agencies. Thedebate about independent agencies as the “headless fourth branch ofgovernment” has its roots in this period in the United States. In Canada therewas far less controversy in part because there traditionally had been anacceptance of a more activist economic role for the state. This was manifestedboth in the employment of public enterprises in major sectors as well as a moreprominent role for regulatory agencies.

In any event the function of regulation in both countries was to be one thatstressed not reaction and correction for unacceptable economic behavior bymonopolies but anticipation and direction by state actors at the sectoral level(Jaffe 1956). Regulation in particular was to be less concerned with “policing”individual firms rather than promoting and indeed planning the activities ofboth firms and sectors (Landis 1938; Schultz & Alexandroff 1985; Eisner1993). As importantly in the aspirations of American proponents and in theactual design of Canadian decisionmakers, the independent, specialized,impartial regulatory agency was to be the central state actor to fulfill thepromotional and planning roles of regulation.

It is important to note that in both countries the debate just summarilydescribed did not extend to the telecommunications sector, for reasons that we


18

need not address here. In both countries well into the 1960s, although there wereperiodic disputes about the performance of the telecommunications regulator,more so in the United States than in Canada, overall the agencies did not becomeembroiled in the extensive conflicts typical, for example, of the transportationsector. In large part, in the United States this was because the courts, through theanti-trust system, played a major preemptive role. In Canada, telecommunica-tions was unique among industrial sectors in not being given a governmentalembrace, at least not by the federal government, until the late 1960s largelybecause of the divided jurisdiction for the sector combined with the success ofthe private sector in meeting Canada’s telecommunications service needswithout overt, extensive political direction.

There is no need here to recount the turbulence that has embroiled thetelecommunications sector throughout the world. For our purposes thesignificance rests in the implications for regulation and regulatory agencies. Asa result of the turbulence of restructuring, regulatory agencies in thetelecommunications sector throughout the world today are expected to be“agents of change.” Regulators have been, as indicated with few nationalexceptions, assigned the role to help make (in some cases to be the primaryactor) the transition from monopoly to competition. At the same time, it isimportant to emphasize, regulators are expected to protect and promote thetraditional social interests such as consumer protection and universal service.

This new positive role for regulatory agencies carries with it considerabletension and conflicts. Those who are forced to change may not appreciate eitherthe speed or direction of change. Others seek to benefit from the changes or tochannel them in favorable directions. A central controversy characteristic ofalmost all countries is to what extent regulators as appointed public officials,with specified degree of tenure, should be the primary agents of such change;hence the ongoing debates over regulatory independence. It is important to notethat this is an unresolved set of conflicts and that no country has really gotten itright—at least not to the complete satisfaction of the central participants.

Recent experience in Canada and the United States can be cited usefully toillustrate the tensions and the ongoing search for new solutions to demarcate theappropriate limits of regulatory independence and the concomitant legitimatepolitical controls. In Canada, which has seen fundamental conflicts between theregulatory agency and political authorities over the past decade or so, newlegislation has been enacted (Schultz forthcoming). This legislation, however,takes a rather indirect route to establishing the primacy of political authoritiesin goal-setting and limiting regulators to a goal implementation role. Theproblem springs from the fact that rather than making the hard choices in thelegislative process between conflicting and ambiguous policy goals, Canada


19

opted for a rather open-ended, multiple statement of legislative objectives thatoffers the regulator little in the way of concrete policy guidance. The result inthe first instance is that the regulator becomes responsible for making what areessentially and unavoidably political compromises and tradeoffs; hence theredress to indirect political controls such as policy directives and politicalappeals against regulatory decisions, especially after the fact. As indicated,these appeals, although now subject to some procedural constraints such as duenotice to all parties, are still subject to a non-transparent political process at thehighest political level. As such it is arguable that the appeal mechanism is incontravention of the GATS transparency requirement. In addition, the recentshort term appointments of a number of regulators suggests that Canada’spolitical authorities have decided that one solution to the question of regulatoryindependence is a temporal short leash for the regulators. The alwaysimpending need for reappointment may concentrate the mind of the regulatorsas to who their masters are.

In the United States, reflecting their different political traditional andinstitutional system, Congress and the Administration have taken a differentroute in the 1996 Telecommunications Act. This legislation reflects the lack oftrust that has ensued over the years between both members of Congress and theExecutive Branch as well as between the FCC and Congress, and the fear on thepart of various parties that an independent regulator may not respect the variouscommitments and agreements made by the members of the legislative branch.These have resulted in one of the most detailed pieces of legislation in recentyears. The FCC is subject to detailed legislative instructions not only on theoutcomes to be pursued but the actual regulatory process to be employed,including rigidly specified deadlines. Congress, to ensure regulatorycommitment and respect, has legislated for itself a role in continuingsurveillance and participation in implementation. The irony, of course, is thatsome of the parties have employed the judicial process to derail some of thehard-fought compromises that Congress and the FCC had sought to implement(Bartlett 1997).

To return once again to Robert Crandall’s comment, it is true that there isno such thing as, and one could add, should not be, a totally independentregulatory agency. Political authorities must make the difficult political choicesin order to direct effectively the regulators in their concrete decisionmaking.They cannot simply pass laws that in effect say “here is a problem, or complexset of problems, you deal with it.” To do so is an abdication of responsibility andwhen it happens, as it does all too frequently, one should not be surprised thata political maelstrom ensues with the regulatory agency at the eye of the storm.Of course, not all policy requirements can be anticipated and not all ambiguities


20

and inconsistencies avoided. Consequently, secondary instruments, subject toeffective public scrutiny through legislative and disclosure requirements suchas policy directives (or ordinances as they are called in the Germantelecommunications legislation) are an appropriate political controlmechanism. With these two prior constraints on regulatory behavior, regulatoryindependence for individual decisions, free from direct political control orinterference, is both appropriate and indeed necessary for contemporarytelecommunications restructuring. Finding that balance between politicalcontrol over major policy determinations and regulatory independence forindividual decisions and applications is perhaps one of the most difficultchallenges confronting all countries today.

REFERENCES

Armstrong, C. & Nelles, H. Monopoly’s Moment: The Organization and Regulation ofCanadian Utilities 1830-1930 (Philadelphia, Temple University Press), 1986.

Bartlett, H. H. The Public Interest and the Introduction of Competition into LocalTelephone Networks, Commlaw Conspectus (Summer), 1997.

Cushman, R. E. The Independent Regulatory Commissions (New York, OxfordUniversity Press), 1941.

Cutler, L. & Johnson, D. R. Regulation and the Political Process, Yale Law Journal,1941, 84.

Davidson, D. & Vietor, R. Economics and Politics of Deregulation: The Issue ofTelephone Access Charges, Journal of Policy Analysis and Management, 5 (Fall),1985.

Eisner, M. A. Regulatory Politics in Transition (Baltimore, The Johns HopkinsUniversity Press), 1993.

Jaffe, L. L. The Independent Agency: A New Scapegoat, Yale Law Journal, 65 (June),1956.

Kaiser, F. M. Congressional Action to Overturn Agency Rules: Alternatives to theLegislative Veto, Administrative Law Review, 32 (4), 1980.

Kane, T. G. Consumers and the Regulators (Montreal, Institute for Research on PublicPolicy), 1980.


21

Krislov, S. & Musolf, L. D. (eds.) The Politics of Regulation (Boston, Houghton MifflinCo.), 1964.

Landis, J. The Administrative Process (New Haven, Yale University Press), 1938.

Levy, B. & Spiller, P. T. (eds.) Regulations, Institutions and Commitment (New York,Cambridge University Press), 1996.

McManus, J. C. Federal Regulation of Telecommunications in Canada, in: H. E. English(Ed) Telecommunications for Canada (Toronto, Methuen), 1973.

Saks, M. Holding the Independent Agencies Accountable: Legislative Veto of AgencyRules, Administrative Law Review, 36 (1), 1984.

Schultz, R. J. Regulatory Agencies in the Canadian Political System, in: K. Kernaghan(Ed) Public Administration in Canada 3rd Ed. (Toronto, Methuen), 1977.

Schultz, R. J. Still Standing: The CRTC 1976-1996, in: G. Doern, et al. (Eds.) CanadianRegulatory Institutions: Globalization, Choices and Change (Toronto, Universityof Toronto Press), (forthcoming) .

Schultz, R. J. & Alexandroff, A. Economic Regulation and the Federal System(Toronto, University of Toronto Press), 1985.

Sommer, S. Independent Agencies as Article One Tribunals: Foundations of a Theoryof Independence, Administrative Law Review, 39 (1), 1987.

United States Senate Committee on Government Operations, Congressional Oversightof Regulatory Agencies Vol II, Ist Session 95th Congress (Washington, U.S.Printing Office), 1977.

Willis, J. (ed.) Canadian Boards at Work (Toronto, Macmillan), 1941.


22

LONG RUN INCREMENTAL COSTS AND THE REGULATION OFINTERCONNECTION CHARGES IN THE UK1

Geoffrey Myers

SUMMARY

Since October 1, 1997 the regulation of British Telecommunications’ (BT)interconnection charges has moved into a new phase. The network chargecontrols involve several key changes in the regulatory regime: the move awayfrom detailed and intrusive regulation (e.g., annual determinations) to allow BTto set its own interconnection charges within appropriate limits; relating thetype of charge control to the competitiveness of the service; the move to RPI-X for non-competitive services to sharpen incentives for BT to reduce itsnetwork costs.

Another key difference in the network charge controls is the change in thecost base from historic cost accounting fully allocated costs (HCA FAC) toforward looking long run incremental costs (LRIC) plus equal proportionatemark-ups. Estimates of incremental costs have been derived from a processinvolving three essential elements:

(a) top-down model (developed by BT);(b) bottom-up model (developed by an industry working group);(c) reconciliation of top-down and bottom-up outputs and the productionof hybrid results - the best available measure of BT’s relevant incurredincremental and common costs.

The top-down and bottom-up models have different strengths andweaknesses. The top-down model has significant articulation especially ofoperating and indirect costs, but includes any inefficiencies that are present inBT’s operations and suffers from a lack of transparency, because it containscommercially sensitive data and cost functions. The bottom-up model hasclarity of the key cost drivers especially for capital costs, but has no “audit trail”and could omit relevant costs. The reconciliation enables interconnectingoperators to have confidence that costs have not been artificially inflated andgives BT the comfort that the cost base can be reconciled back to its accounts.

The charges based on incremental costs now in place in the UK can becompared to other measures of costs, such as fully allocated costs, and also tocharges in other countries. BT’s interconnection charges are comfortably thelowest in Europe and are comparable to the charges for network components of


23

the lowest charging Local Exchange Carriers in the U.S. (and substantiallylower, once access charges are included in U.S. interconnection charges).

1. INTRODUCTION

The regulation of interconnection charges in the UK has recently movedinto its third distinct phase since the privatization of British Telecommunica-tions in 1984 with the introduction of the “network charge controls” on October1, 1997. An essential element of the network charge controls is the introductionof long run incremental cost (LRIC) as the cost base for interconnectioncharges. This paper addresses the following topics:

- the background to the regulation of interconnection charges in the UK;

- the changes in the regime that have resulted from the introduction ofnetwork charge controls;

- the derivation of estimates of long run incremental cost, using top-downand bottom-up models; and

- comparisons of the resulting costs and interconnection charges with thosederived from other cost bases and with costs and charges in the rest ofEurope and the U.S.

The paper discusses the arrangements for regulation of BT’sinterconnection charges, but does not address local loop unbundling (alsoknown as Direct Access to the Local Loop, DACL), since BT is not required tosupply DACL in the UK.

2. REGULATION OF INTERCONNECTION CHARGES

There have been three distinct phases of the U.K. Office ofTelecommunications’ (OFTEL) regulation of BT’s interconnection charges:

(1) up to 1994, regulatory intervention only in the event of a failure to reachagreement in commercial negotiations;

(2) introduction in 1995 of Standard Charges—automatic annualdeterminations by OFTEL of the vast majority of BT’s interconnection charges;and


24

(3) from October 1, 1997 the introduction of network charge controls.

Phase 1: Determination in Case of Failure to AgreeIn the initial years following privatization the only fixed wire competitor to

BT was Mercury. OFTEL was only required to make determinations relating tointerconnection if called upon by one of the parties, in the case of a failure toreach an agreement. In practice, BT and Mercury never reached agreement oninterconnection charges and OFTEL was called upon to make determinations in1985 and 1993 (indexation of the 1985 prices was used to set charges in theintervening years). The cost base used for such determinations was fullyallocated costs (FAC) using the historic cost accounting (HCA) convention. Incontrast to charges, in general, agreement was reached between theinterconnecting operators on a range of other matters, such as points ofinterconnection, technical interfaces etc.

Phase 2: Standard ChargesFollowing the duopoly review (see DTI (1991)), the UK government

decided to operate a far more liberalized licensing regime and a large number ofnew entrants came into the industry operating in a wide range of markets. Theoperators fall into a number of broad categories:

Operators with both local and long-distance networks, such as CWCand NTL.

Cable operators, providing both telecoms and cable televisionnetworks in more than one hundred franchise areas in the UK. As atJuly 1997 cable operators had passed approximately 9.4 millionpremises (about 40%) and provided almost 2.5 million residential lines(11%) and 0.3 million business lines (4%). By the year 2000 it isexpected that about 70% of UK households will be able to take servicefrom a cable operator.

Radio fixed access providers, offering direct access to customers.Ionica launched service late in 1996. By 2001 it is expected to havecoverage of 75% of the UK population. Other operators with licencesfor fixed radio access include Atlantic Telecom and, in rural areas, BTand RadioTel.

Other local access providers in particular regions, such as ScottishTelecom and Torch.


25

High bandwidth networks aimed at large business customers, such asCOLT and MFS in the City of London.

Operators with long distance networks, such as Energis and Racal.

Providers of international services, both international facilitieslicensees (almost 50 licences issued since liberalization in 1996) andinternational simple resellers.

With the proliferation of operators requiring interconnection with BT, itwas felt that a more regular determination of BT’s interconnection charges wasrequired to avoid repeated and uncoordinated requests for determinations fromOther Licensed Operators (OLOs). This precipitated the introduction of theStandard Services regime, under which BT was required to offer to OLOs arange of services at charges determined annually by OFTEL. The list ofStandard Services covered a wide variety of interconnection services, such asinland conveyance and transit, international conveyance, interconnectioncircuits, directory assistance, operator assistance, emergency operator, and datamanagement amendments. Interconnecting operators paid only for theelements of the network that they used and charges did not vary with end-use.Determinations occurred automatically without the trigger of a request from anOLO or BT. First, OFTEL made a determination of Interim Charges, based ona forecast of the HCA fully allocated costs of services for the coming financialyear. During the year interconnect payments were based on these interimcharges. The charges were adjusted retrospectively (including interestpayments) following the Final Charges determination, which was based on theout-turn HCA fully allocated costs after the financial year end.

The other key element of the ICAS regime was the introduction ofAccounting Separation (AS), the production by BT of audited regulatoryaccounts for separated businesses (Access, Network,2 Retail Systems Business,Supplemental Services, Apparatus Supply, and Residual) in accordance with apublished documentation of allocation principles and methodology—see BT(1997c,d). The AS accounts have provided a more soundly based and rigorousdivision of costs between retail and network; they facilitate the identification ofcross-subsidies between BT businesses; and they provide confirmation of non-discrimination in interconnection charges, i.e., that BT Retail pays transfercharges to BT Network on the same basis as interconnecting operators.


26

Traffic Sensitive CostsFollowing the Duopoly Review, Access Deficit Contributions (ADCs)

were introduced into interconnection charges to “compensate” BT for its accessdeficit, arising from the explicit regulatory constraints on rebalancing (a sub-cap of RPI+2% on the exchange line rental within the overall retail basket).ADCs can also be seen as derived in a broad sense from the principles of theEfficient Component Pricing Rule (ECPR)3, though with some significantdifferences (see Armstrong and Doyle (1994)). For example, the size of theADC increased with BT’s profitability of the end-use call for which theinterconnection service was used. But ADCs were associated with a complexseries of arrangements including ADC waivers (related to market sharethresholds) and ‘reverse ADCs’ (for competing local access providers).

In February 1996 the sub-cap on the exchange line rental was removedleading to the abolition of the system of ADCs. BT, therefore, has freedom torebalance between line rental and call charges as it sees fit, subject to the overallprice cap. With the abolition of ADCs, interconnection charges relate only totraffic sensitive costs of the network—all non-traffic sensitive costs, includingthe local loop and the line card in the local exchange, are recoverable by BTfrom retail prices.

3. NETWORK CHARGE CONTROLS

Some aspects of the Standard Charges regime remain importantunderpinnings of the network charge controls, such as the list of StandardServices, non-discrimination between operators (though differential chargingby end-use is possible), and separated accounts. But there have been a numberof important changes to the regulation of BT’s interconnection charges:

- deregulation: degree of regulation appropriate for each of BT’sinterconnection services depends upon the competitiveness of the market inwhich it competes;

- incentive regulation with the move from determinations (rate of returnregulation) to charge caps for those services that require continuingregulation of charges;

- change in the cost base from fully allocated costs (FAC), using historiccost accounting (HCA), to forward looking long run incremental costs(LRIC);


27

- BT given flexibility to set charges, but with a responsibility to demonstratethat charges are reasonably derived from costs (e.g., between incrementalcost floors and stand-alone cost ceilings); and

- charges known contemporaneously with the removal of retrospectivedeterminations.

Structure of the ControlsBT faces competition from OLOs in a number of interconnection markets.

Such competition will increase over the coming years with the entry of newoperators and the further establishment of existing players. OFTEL carried outan assessment of the competitiveness of the markets in which BT’s servicescompete. It defined three categories—competitive, prospectively competitive(i.e., expected to become competitive during the control period, 1997-2001),and non-competitive—with a different regulatory treatment for each category,as shown in Table 1.

Table 1: Categories of Interconnection Services

Classification Charge control treatment

Competitive No charge controls

Prospectively competitive RPI+0% individual (“safeguard”) caps(Cap on each individual charge)

Non-competitive RPI-8% basket control(3 separate baskets: Call termination basket;General network basket; and Interconnect specific basket)

Very few services were considered competitive by the start of thecontrols—only operator assistance services provided by BT to other operatorson an agency basis and new services, which will be initially treated ascompetitive. Inter-tandem or long distance conveyance was classified asprospectively competitive because of the competition that BT will increasinglyface from other operators acting as “carrier’s carriers,” such as Mercury,Energis and Racal (though these operators also have retail customers).International Direct Dial (IDD) conveyance was also prospectivelycompetitive, given the liberalization of international facilities by the UKGovernment. To include the charges for these services in the baskets would beundesirable, because BT would tend to focus price cuts on these services, whereit faces increasing competition, rather than on the non-competitive services.


28

The use of safeguard caps allows these services to be excluded from basketcontrol, but with a fallback control on prices to protect customers. However, theexpectation is that competition will force BT to set prices below the limit set bythe safeguard cap, i.e., that the safeguard caps will become non-bindingconstraints. If, during the control period, the market for a service is found to becompetitive, the safeguard cap will be removed.

The non-competitive services are those associated with the local exchange:conveyance from (to) the local exchange to (from) the customer, i.e., calltermination (origination), and conveyance from (to) a tandem switch to (from)the local exchange. The diagram at Annex C illustrates the network componentsused in BT’s main inland conveyance interconnection services and the type ofcharge control they face. A full list of interconnection services and theirregulatory treatment under the network charge controls is set out in OFTEL(1997c).

Call TerminationOne of the new aspects in the classification of services is the distinction

between call termination and call origination. Both are considered non-competitive, but call termination gives rise to particular regulatory concernbecause it may be a long-term bottle-neck and so has been placed in its ownseparate basket. The distinction between termination and origination arisesfrom the externality associated with the former, which arises because it is thecaller (and her operator) that pays the call termination charge, whereas thechoice of network and influence over the call termination charge is exercised bythe call recipient. Operators face weak incentives to minimize call terminationcosts, because they are not borne by their own customers. They also haveincentives to inflate termination charges, because this may serve to raise rivals’costs when competing in retail markets. Call termination is defined from thelowest switch in the hierarchy at which interconnection can occur to thecustomer, i.e., from the local exchange (the interconnection service, localexchange segment). The term “call termination” is therefore restricted to thebottle-neck—there may be competition between operators for delivery to thelocal exchange, but competition in call termination cannot be expected (unlessthe customer takes access lines from more than one operator).

A related issue is the appropriate level of the charge to be paid by BT toOLOs for call termination on OLO networks. OFTEL’s close interest in chargesfor call termination on OLO networks arises for two reasons:

- to protect against abuse of market power and exploitation of customers;and


29

- to ensure no distortion of competition between operators.A full analysis of the difficult issue of call termination payments between

operators competing in retail markets is beyond the scope of this paper—forOFTEL’s views, see Annex C of OFTEL (1997c). The agreement reachedbetween BT and OLOs specifies (in accordance with OFTEL’s analysis) thatthe charges levied by OLOs will be based on BT’s charges for delivering calls:local exchange segment and single tandem. Therefore, the starting charges andthe floors and ceilings for BT’s services, derived from the calculation ofincremental costs, will influence not only the charges paid by OLOs to BT, butalso the reciprocal charges paid by BT to OLOs.

Role of Long Run Incremental CostsThe objective of moving to LRIC is to provide improved signals for entry,

exit and investment. The “forward looking” element of LRIC is the use ofCurrent Cost Accounting (CCA) for asset valuation, i.e., using the replacementcosts of assets rather than their historic cost at time of purchase. In particular,replacement should generally be assumed to be by the Modern Equivalent Asset(MEA), the lowest cost, proven technology that performs the same function asthe asset under consideration (e.g., all switches valued as if digital, even if someanalogue switches remain in place4).

The methodology developed in the UK to calculate LRIC for BT and theresulting cost estimates are discussed below. In the context of the networkcharge controls there are two distinct uses of the LRIC information. The firstuse is to set the starting charges (or initial levels) of the network baskets.Second, since over time BT will be able to vary the charges of individualservices, the initial assessment of the legitimacy of the cost orientation of thecharge set by BT will be whether it lies between the floor based on LRIC and theceiling based on stand-alone cost (SAC).

Level of Charges: Mark-upsOFTEL has set the starting charges for all of the services in the call

termination and general network baskets at long run incremental cost plus equalmark-up. These charges will apply from October 1, 1997 until BT announceschanges in network charges, giving 90 days’ notice. In using LRIC as the costbasis to set the level of charges, OFTEL considers that it is necessary to add amark-up. The common costs between the access network and the conveyancenetwork are a relevant element of BT’s total costs and it should be given theopportunity to recover those costs. The key question is how the burden ofrecovery of the common costs should fall upon access and conveyance.


30

In extensive consultation OFTEL gave consideration to a number ofdifferent possible mark-up regimes, such as the Efficient Component PricingRule and Ramsey mark-ups, but concluded that whilst each of these regimes hassome theoretical attractions they would be problematic to implement and couldhave undesirable implications.5 The highly pragmatic approach of equalproportionate mark-ups was chosen, in which the common costs areapportioned for recovery from access (through the line rental and other retailprices) and conveyance (through interconnection charges, including BTRetail’s transfer charges) in proportion to the incremental cost of access andconveyance. This simple rule can be illustrated using incremental cost figuresfor interim 1996/97 (i.e., the first six months of the financial year, April-September 1996), shown in Table 2.

Table 2 Equal Proportionate Mark-ups: Figures for Interim 1996/97

£ million Incremental Cost Common Cost Cost including (multiplied by two to before mark-up apportioned mark-up give full year estimates)

Inland Conveyance 996 104 1,100

Inland Private Circuit(conveyance related costs) 318 32 350

Access 3,220 334 3,554

Common Costs betweenConveyance and Access 470

Total 5,004 470 5,004

Note: The equal proportionate mark-up rate is 10.3%.

LRIC plus the equal mark-up for inland conveyance is the cost base for thelevel of the baskets but, in principle, BT has freedom to set the structure ofprices within the basket subject to meeting the RPI-8% constraint and controlson anti-competitive behavior.

Structure of Charges: Floors and CeilingsThe key cost concepts in the network charge control regime will be

incremental cost and stand-alone cost. In the first instance, whether or not acharge is excessively low and potentially anti-competitive, or excessively highand exploiting customers will be judged relative to floors and ceilings. There isan argument that anti-competitive behavior can be expected where the firm setsprices generating revenues that are insufficient to cover incremental costs. Onlyif revenues exceed the incremental cost of the service is the firm’s profitability


31

improved by providing the service. The stand-alone cost (of an efficientoperator) can be argued to set the upper bound on the charge for a service,because a price above this level can only be maintained by the incumbent if themarket is uncompetitive—otherwise the price could be undercut even by asingle product firm, i.e., one that benefits from no economies of scope.

OFTEL does not propose to implement floors and ceilings mechanically—the key is the effect of the charge in the relevant market. Floors and ceilingsprovide a useful general guideline when examining whether a charge can beexpected to have anti-competitive effects or exploit customers, but there will beexceptions. These may be related to the particular features of the case underinvestigation, but may also reflect different approaches to the analysis of anti-competitive behavior, such as the relevance of net revenue tests to theidentification of predatory behavior—see, for example, Myers (1994). Theparticular construction of floors and ceilings and the relevance of combinatorialtests is discussed below under the definition of the increments and in Annex A.

4. INCREMENTAL COST METHODOLOGY

The robust methodology to calculate incremental costs was developed byOFTEL and operators in the UK industry over a period of more than two years.The heart of the work was a three-pronged approach:

- Top-down model, developed by BT;

- Bottom-up model, developed by an industry working group; and

- Reconciliation of the results of the top-down and bottom-up models.

The top-down and bottom-up models have different strengths andweaknesses and each provides a powerful and, in some cases, very detailedcross-check on the other. The process of developing the two models in parallelover a period led to substantial improvements to both models.6

Key Methodological AssumptionsDefinition of the Increments

The increments could be defined in a number of different ways.Interconnection services would be the most theoretically pure approach,because services are the units purchased by interconnecting operators.However, incremental costs on this basis would be the most complex to derive.Taking components as the increments would be the approach most similar to


32

Total Element Long Run Incremental Cost (TELRIC) in the U.S. The approachtaken in the UK has been to define the increments in a more aggregated manner,because of its pragmatic advantages. Hence, the increments assumed in themethodology are the whole of the conveyance (or core network)—trafficsensitive costs, both switching and transmission—and the whole of the accessnetwork—non-traffic sensitive costs, both local loop and line card. Threeelements of cost are identified: the incremental cost of conveyance, theincremental cost of access, and the common costs between conveyance andaccess. The incremental cost of conveyance is then broken down into the costsof the network components; the costs of interconnection services are derivedusing routing factors applied to the component costs. The characteristics of thisapproach to defining the increments in contrast to services or components isdiscussed in Annex A.

Definition of the Long RunThe long run was taken to be the period over which all assets are replaced.

It is assumed that there are no sunk costs in the long run.

Degree of Network OptimizationThere are two main options: “scorched earth,” under which an attempt is

made to optimize the number and location of the incumbent’s switching nodes;and “scorched node,” under which the network is optimized, but taking thenodes as given. In principle, the scorched earth approach is desirable, if theintention is to measure the fully efficient level of costs. However, the optimalnumber and location of nodes are matters of considerable controversy.Therefore, the scorched node approach has significant pragmatic advantages,because it reduces the modeling complexities, even though it will lead to anoverstatement of efficiently incurred costs to the extent that the incumbent’snodes are not optimally situated.

5. TOP-DOWN MODEL

The top-down model starts from the cost information captured in BT’scurrent cost accounting systems and processes it to remove those costs that arenot incremental. The top-down model analyses cost information in some 485cost categories derived from BT’s Accounting Separation system (whichcarries out the split into network and retail costs). Each cost category has anassociated cost-volume relationship that shows the way in which that type ofcost would decline in the long run with reductions in the volume of the


33

identified cost driver. The operation of cost-volume relationships is illustratedin the diagram at Annex D.

The top-down model is structured to derive a “decremental” (or avoidable)cost.7 The definition of the increment implies the volume of the cost driverassociated with the increment. The level of cost resulting from the removal ofthat volume can be read off the cost-volume relationship, starting from therecorded CCA cost and cost driver volume level. This cost is subtracted fromthe recorded cost to give the contribution to the incremental cost arising fromthe cost category under consideration. The total incremental cost is derived bysumming over all the cost categories.

BT has derived the cost-volume relationships from a variety of sources:e.g., based on engineering studies, past experience, statistical relationships. Fulldetails of the top-down methodology and descriptions of the cost-volumerelationships in the model are set out in BT (1997a,b). Recent top-down modelresults have been audited to a “fairly presents” standard in accordance with thedocumentation of the model set out in these documents.

Depreciation in the top-down model uses the principle of Financial CapitalMaintenance (FCM). It is the sum of Operating Capital Maintenance (OCM)depreciation, computed on a straight line basis, and a holding loss (or gain)reflecting changes in the price of the modern equivalent asset during the year.The derivation of operating costs has significant articulation and typicallyinvolves several levels of cost-volume relationships (e.g., the cost of personneldepends on the number of employees, which in turn depends upon the volumeof calls and lines). Further details of the top-down model are set out in Annex B.

6. BOTTOM-UP MODEL

The bottom-up model involves a collection of engineering and othermodels that explicitly build up the annual costs of switches and transmissionfrom the dimensioning capacities required to serve busy hour demand. Giventhe unimportance of the costs of access in UK interconnection charges (theironly impact is on the equal mark-up calculation), the bottom-up model focuseson deriving the incremental cost of conveyance. The strength of the model is indeveloping the costs of capital equipment. Operating costs, however, are notbuilt up from explicit cost drivers, because of the difficulty and complexity ofthe modeling that would be required—instead, operating costs are computed asa proportion of capital costs by asset type, calibrated using information from arange of operators.

The bottom-up model comprises the following constituent models:


34

- Network model, which derives the costs of local switches,concentrators and tandem switches and presents the costs of switchesand transmission in pence per busy hour minute and pence per minuteterms;

- Transmission model, which builds up the costs of remote-local, local-tandem and inter-tandem transmission per 2 Mbit/s, using provisioningrules;

- Erlangs per circuit model, which uses provisioning rules to derive theutilization of transmission circuits; and

- Economic depreciation model, which produces profiles of capital cost(depreciation plus return on capital employed) for switches,transmission electronics, fibre cable, duct and the major indirect assets(accommodation, computers and motor vehicles) over their economicasset lives—these are used in the network model to convert theinvestment costs into annual capital costs.

A detailed description of the bottom-up model, including print-outs ofmodel workings and results, is set out in OFTEL (1997a).8 A further descriptionof the modeling is included at Annex B.

7. RECONCILIATION

From the outset OFTEL regarded the reconciliation as an essential elementof the methodology. The top-down and bottom-up models each have strengthsand weaknesses. For example, the top-down has the advantage of an “audittrail” and significant articulation, especially of operating costs, but it suffersfrom a lack of transparency, because it contains information that iscommercially confidential and reflects any inefficiencies in BT’s networkoperations. The bottom-up model is transparent and has clarity in the derivationof capital costs, but is relatively weak in developing operating costs and mayexclude relevant costs (e.g., those associated with more indirect functions, suchas personnel and finance).

The object was not to choose one model over the other but to draw on thestrengths of both to derive hybrid results, the best available measure ofincremental costs. The concept of the hybrid results is that they could becalculated using either model by applying a specified set of adjustments. Thisgives BT the comfort that the hybrid figures can be traced back to its accounting


35

information. OLOs can see how the figures were derived and have theconfidence that costs are not artificially inflated. Of course, both BT and theOLOs may disagree that the hybrid figures are the most appropriate measure ofincremental costs, preferring the top-down or bottom-up model respectively.However, both will know how and why the costs were derived.

The reconciliation process identified the differences in the results of the twomodels, explained the reasons for those differences and so led to thespecification of the hybrid adjustments. The results of the models differed for avariety of reasons, such as differences in:

- input costs

- parameter assumptions (e.g., the cost of capital)

- annualization of capital costs (the top-down model uses straight linedepreciation plus holding loss, whereas the bottom-up model useseconomic depreciation)9

- modeling approaches (e.g., the bottom-up model assumes point-to-pointtransmission, whereas BT’s transmission network uses intermediatemultiplexing)10

- routing assumptions (e.g., the bottom-up model assumes no direct routesbetween local switches, although such routes do exist in BT’s network)

- equipment utilization and efficiency

Details of the reconciliations for the 1993/94 and 1994/95 models are setout in NERA (1996b,c). Switching costs were reconciled at a disaggregatedlevel. Transmission costs were reconciled at a more aggregated level, becauseof the difficulties of disaggregated comparisons due to the differentclassifications of transmission (junction and trunk in the top-down model;remote-local, local-tandem and inter-tandem in the bottom-up model) and thedifferent modeling approaches (intermediate multiplexed versus point-to-point).

OFTEL made a decision to base the starting charges for the network chargecaps on BT’s incurred incremental costs, not the costs of an efficient operator(with the degree of inefficiency reflected in the value of X). This allowed thetop-down model after the hybrid adjustments, which were derived following thereconciliation, to be used to generate incremental cost information in the years


36

following the reconciliation work. OFTEL used the incremental cost data for1995/96 and interim 1996/97 as the basis for deriving the starting charges (byforecasting costs forward to 1997/98).

8. COST AND PRICE COMPARISONS

BT has been required in recent years to produce both HCA and CCA fullyallocated costs Accounting Separation financial statements. These can becompared to the latest available audited incremental cost figures (for interim1996/97). Table 3 shows this comparison for the local exchange segment, theinterconnection service used for call termination or call origination (see AnnexC).

As can be seen, the LRIC figure, including mark-up, for the local exchangesegment is almost identical to the CCA FAC figure, although there are someminor differences at the level of component costs. HCA FAC is significantlyhigher than LRIC plus mark-up. The main reason is the difference in the cost ofthe local exchange, arising from the gap between asset valuations under HCAand CCA.

Table 3: LRIC, HCA and CCA Costs of Local ExchangeSegment in Interim 1996/97

Component/ Routing factor LRIC LRIC HCA FAC CCA FACService for local exchange plus

segment mark-up

Local exchange 1.00 0.189p 0.215p 0.274p 0.210p

Junction link 0.66 0.056p 0.059p 0.070p 0.058p

Junction length 10.11 0.0053p 0.0056p 0.0062p 0.0063pLocal exchangesegment 0.279p 0.312p 0.384p 0.313p

% difference -11% 23% 0%

Despite the similarity between LRIC plus mark-up and CCA FAC, the greatadvantage in deriving LRIC information is that it allows floors and ceilings tobe computed. It will be essential to have this kind of information in the networkcharge controls, since it will be BT that sets the charges for individual services,subject to the charge caps and controls on abuse of a dominant position. Table4 shows the floors and ceilings derived for the local exchange segment.


37

Table 4: Floors and Ceilings for Local ExchangeSegment in Interim 1996/97

Component/ Routing factor for Floor (LRIC) CeilingService local exchange segment

Local exchange 1.00 0.189p 0.289p

Junction link 0.66 0.056p 0.069p

Junction length 10.11 0.0053p 0.0067p

Local exchange segment 0.279p 0.403p

% difference compared to floor 44%

International ComparisonsComparisons between interconnection charges can be problematic for a

number of reasons. It may be difficult to ensure that a strict like-for-likecomparison has been carried out, because interconnection services in differentcountries may not be strictly analogous, or because the structure of chargesdiffers somewhat (e.g., different time of day structures; per set-up as well as perminute charges in some countries). Depending upon the purpose, differenttypes of comparison may be preferable, e.g., constructing an average chargefrom a basket of services or comparing charges for specific services.Furthermore, the comparisons vary with the particular exchange rates used toconvert into the same currency.

The attempt in this section is not to provide a comprehensive series ofcomparisons, but to illustrate the relative level of UK interconnection chargesagainst charges in the rest of Europe and the U.S. Set out below are somecomparisons, which will not capture all dimensions of variation betweencharges in different countries. For example, in some countries interconnectswitch ports are charged on a rental, not per minute basis—to compute anaverage overall pence per minute charge, a particular number of minutes of useof the switch port needs to be assumed. Also, although the effect of differentpeak-off peak charge ratios can be captured to some extent by computing anaverage charge across times of day, the definition and duration of the peakdiffers between countries and this may itself reflect differences in callingpatterns.

European UnionBT’s interconnection charges are comfortably the lowest in Europe. Table

5 shows charges in seven European Union countries for the three maininterconnection services (or equivalent)—local exchange segment, singletandem and double tandem. Single tandem is likely to be the most commonly


38

purchased service. In this table the charges are averaged across times of day(making a consistent assumption about the proportion of traffic in peak periods,namely 50 percent). The charges shown are for the incumbents in UK,Denmark, Germany, France, Sweden, the Netherlands, and Italy.

Table 5: Comparison of EU Interconnection Charges

pence per minute, Local exchange Single tandem Double tandem11

average across times of day segment

BT (October 1, 1997) 0.33 0.47 0.76

Tele Danmark 0.45 0.85 1.02

Deutsche Telekom 0.53 0.88 1.22

France Télécom12 0.60 1.25 1.71

Telia 0.77 0.97 1.33

KPN n/a 1.16 1.45

Telecom Italia 0.96 1.57 n/a

Notes: Exchange rates used to pence per minute are purchasing power parities for 1996—seeOECD (1997).

Deutsche Telekom’s interconnection services are defined by the distance oftransmission, not by the network elements used in interconnection (e.g., for thesame distance there is the same charge whether the point of interconnection isat a local switch or a tandem switch). However, there is likely to be a correlationbetween the transmission distance and the number of switching stages, soDeutsche Telekom services have been mapped onto the services shown in Table5 as follows: City Zone = local exchange segment, Regio 50 = single tandem,Regio 200 and Fernzone = double tandem.

Tele Danmark, Telia and KPN have two-part interconnection charges: perset-up and per minute. Two-part charges have been offered in Denmark for fouryears or more; in Sweden and the Netherlands they are more recentdevelopments. Tele Danmark charges per set-up, but KPN’s and Telia’scharges are levied only on successful calls. The figures in Table 5 make noadjustment for this factor, so Tele Danmark’s charges are understated, becausepayments will also be made on unsuccessful call attempts (except when due tonetwork failure). This factor is likely to understate charges by about 10 percentat peak times and 15 percent at off-peak times. Telia’s charges may contain asmall relative overstatement for a different reason—Telia’s charges include thecost of interconnection circuits, whereas in other countries there are typicallyseparate connection and rental charges for such circuits.


39

Table 5 understates the difference between the charges that will actually bepaid in UK and France, because in addition to the charge shown,interconnecting operators in France will also have to pay universal servicecontributions and make access deficit payments, amounting to about 0.18 penceper minute (1.8 centimes per minute13) in 1998. On the other hand, the size of thedifference in single tandem charges could be exaggerated, because in France thesingle tandem service gives the interconnecting operator access to about 2msubscribers, more than double the number in the UK (because BT has a largernumber of tandem switches that are, on average, smaller than FranceTélécom’s) .

As yet KPN does not offer interconnection at local exchanges. HenceKPN’s interconnection service “Local Terminating Access” corresponds in theUK to single tandem, not local exchange segment. Neither access charges noruniversal service contributions will be paid on call termination in theNetherlands.14 The prices shown are preliminary—they could be subject toretrospective downward adjustment following the completion of theinvestigations by OPTA, the Dutch regulator. In a similar vein, the charges ofTelecom Italia have yet to be approved by the national regulator.

The interconnection charges in these countries can also be compared to the“best current practice” benchmark ranges recommended by the EuropeanCommission (1997). As shown in Table 6, most but not all of the charges fallwithin the recommended ranges. However, whilst BT’s charges are at thebottom end of the ranges, the charges in other EU countries tend to be close tothe upper bound.

Table 6: Comparison of EU Interconnection Chargeswith DGXIII Recommended Ranges

ECUs/100 per minute, Local Metropolitan National

peak rate charges (> 200 km)DGXIII - lower bound 0.60 0.90 1.50

- upper bound 1.00 1.80 2.60

BT (1 October 1997) 0.64 0.91 1.74

Tele Danmark 0.97 1.81 2.21

Deutsche Telekom 1.00 1.70 2.60

France Télécom 1.00 2.12 2.93

Telia 1.66 2.12 2.93

KPN n/a 1.99 2.50

Telecom Italia 1.53 2.50 n/a

Notes: Exchange rates used to pence per minute are current exchange rates (in November 1997).


40

The basis of the comparison shown in Table 6 differs from Table 5 for threereasons (following the approach used by the Commission):

- peak rate charges are used, not average charges across times of day;

- conversion into a common currency uses current exchange rates, notpurchasing power parity rates; and

- the “national” interconnection service is specified as the service withdistance greater than 200 km, not the average across double tandemservices for BT and Deutsche Telekom.The charges that fall within the recommended ranges (or are very close to

the upper bound) are shown emboldened in Table 6. Telia’s charges for singletandem and double tandem are shown in italics, because these charges would bevery close to or just below the upper bound of the recommended ranges ifpurchasing power parity exchange rates were used instead of current exchangerates. On the other hand, Telecom Italia’s charges would look significantlyhigher if purchasing power parity exchange rates were used.

United StatesCharges in the U.S. can differ greatly between local exchange carriers

(LECs). No attempt is made here to provide a comparison against acomprehensive cross-section of LECs, but Table 7 shows charges for the localexchange for three LECs in Illinois and Massachusetts, two states that havebroadly similar characteristics to the UK. (Illinois is one of the lowest chargingstates in the U.S.) These charges can be compared to BT’s charge for the localexchange component or for the local exchange segment. In the U.S. use ofremote concentrators is generally less common than in the UK and the costs oftransmission between remote concentrators and local switches, where incurred,are included in access network costs (even though such costs are trafficsensitive). On this basis the like-for-like comparison would be with BT’s chargefor the local exchange component.

BT’s charges for the local exchange appear to be lower than current U.S.charges and, even where the local exchange segment is used as the comparator,they are similar to the level of charges of the lowest charging U.S. LECs, if theCCLC (carrier common line charge) and RIC (residual interconnectioncharge) are excluded, as in Table 7. Together these elements of U.S. chargescomprise around 1 cent per minute or more (varying by LEC). There are nocorresponding elements to the CCLC and RIC in the UK, so to the extent thatthese are or continue to be paid,15 BT’s charges are substantially lower. Charges


41

for the local exchange for LECs other than those shown can be in excess of 1cent per minute, even before the CCLC and RIC are added.

Table 7: Comparison of BT’s and Selected U.S.LEC’s Charges for Local Exchange Segment (in cents per minute)

UK Component/ Routing BT U.S. Assumed Ameritech Sprint/ NYNEX/Service factor Starting com- routing (Illinois) MFS C-TEC

charges ponent factor (Illinois) (Mass)cents/min cents/min

Local exchange 1.00 0.298 End 1 0.5 0.4 0.8officelocaltermi-nation

% difference 67% 34% 167%Junction link 0.66 0.082 Tandem 0 0.019

transporttermi-nation

Junction length 10.11 0.0082 Tandem 0 0.0012(km) transport

facilitymileage

PPP 1 0.058Local exchange 0.493 0.5 0.4 0.8segment% difference 1% -19% 62%

Note: Exchange rate used is £1=$1.49, purchasing power parity for 1996—see OECD (1997).

A process of reform of U.S. interconnection charges is underway,following recent Orders by the FCC. Interconnection charges are expected tofall, the RIC is to be removed and the access charge regime is to be reformed.Default price ranges for various elements of interconnection charges, basedupon the TELRIC approach were proposed in FCC (1996). These are lower thanBT’s charge for the local exchange segment but consistent with BT’s charge forthe local exchange component, as shown by the figures in Table 8. It should benoted, however, that the BT figure is an actual charge whereas it remains to beseen how U.S. charges reflecting TELRIC will compare to the FCC defaultprices.


42

Table 8: BT’s Charges and FCC Default Pricesfor Local Exchange (in cents per minute)

UK Service BT floor BT Starting U.S. FCC default(OFTEL charge component price rangeforecast)cents/min cents/min cents/min

Local exchange component 0.261 0.298 End office local termination 0.2 - 0.4% difference compared to component -12% -33% to +34%

Local exchange segment 0.442 0.493

% difference comparedto segment -10% -19% to -59%


43

ANNEX A: DEFINITION OF THE INCREMENT, FLOORS ANDCEILINGS

Since, in the UK methodology, component costs are derived on theassumption that the increment is the whole of the conveyance (or core) network,by construction the sum of the costs of the network components equals theincremental cost of conveyance. One alternative approach would be to takeeach of the components as an increment in its own right (roughly in line with theuse of Total Element Long Run Incremental Cost (TELRIC) in the U.S.). Theincremental cost of conveyance could then be broken down into the incrementalcosts of the component and the common costs between components (alsoreferred to as intra-core common costs). But a question would then arise of howto deal with intra-core common costs when setting the level of charges.16

A consequence of defining the increment as the whole core network is thatthe cost floors thereby derived for components are at least as large as theincremental cost of the components, because they include an element of thecommon costs between components.17 However, component incremental costswould not give the theoretically correct floors, since the units purchased byinterconnecting operators are services, not individual components. Thetheoretically pure approach would be to define interconnection services as theincrements and derive incremental and stand-alone cost on this basis for eachservice and for each combination of services (since combinatorial floor andceiling tests are also relevant). This information would be substantially moredifficult to derive, because it would rely upon the shape of the cost functions forcomponents. Moreover the results could be quite sensitive to the particularmodeling assumptions and simplifications made. Substantial further workwould be needed to develop a model to produce robust results along these lines.

The definition of the increment as the whole conveyance network has theattraction of relative simplicity. The floors and ceilings at component level areonly a means to an end, since the floor and ceiling tests will be carried out atinterconnection service level. The OFTEL approach allows the service floorsand ceilings to be derived in a straightforward manner from the componentfloors and ceilings (using routing factors). Alternative approaches mightrequire the construction of complex matrices showing incremental and stand-alone costs for different combinations of components and services, dependingupon which intra-core or inter-service common costs were relevant to eachcombination. A feature of deriving the floors and ceilings from the definition ofthe increment of the whole core network is that it yields floors and ceilings that,by construction, will satisfy the combinatorial tests (to the level of the corenetwork).


44

ANNEX B:MODELING OF THE LOCAL EXCHANGE SEGMENT

For ease of exposition the descriptions of the top-down and bottom-upmodeling below focus upon the local exchange segment, the interconnectionservice used for call termination or call origination.

Top-down ModelCurrently, BT’s local exchange component covers the costs of both local

switch (processor) and concentrator. The cost-volume relationships for thelocal exchange are based on detailed engineering models for System X andAXE10 exchanges. BT does have other types of local exchange in its network(UXD, TXE), but these are not modern equivalent assets and so are valued as ifthey were System X or AXE10. Local exchanges have two cost drivers: calls(conveyance) and lines (access). This results in a cost function that is a three-dimensional function. Some elements of the exchange vary only with calls,such as the local switch processor. Some vary only with lines, such asmiscellaneous line auxiliaries on the concentrator. Some vary with bothconveyance and access and give rise to common costs, such as concentratorcore equipment, rack equipment and DC power equipment.

The cost function is simplified into two two-dimensional cost-volumerelationships corresponding to the two cost drivers (the function for callsassumes that lines are maintained at existing volumes and vice versa). The cost-volume relationships are derived from engineering models that useprovisioning rules to derive optimal switch configurations for specifiedvolumes of calls and lines. The shape of each function depends upon the natureof the optimal re-dimensioning of the switch as volumes are allowed to vary.The functions derived by BT are piecewise linear (with declining slope) plus afixed cost element.

The other element of the local exchange segment is transmission betweenremote concentrator and local switch (“R-L transmission”). In BT’s currentclassification of transmission, there is no separate component for R-Ltransmission - its costs are averaged with local-tandem transmission in junctiontransmission. The top-down model has separate cost-volume relationships todevelop the capital costs of electronics, cable and duct, as well as a variety ofrelationships relating to operating and more indirect costs.

Bottom-up ModelThe costs of the concentrator and local switch are developed separately.

Dimensioning capacities for each in terms of ports and busy hour call attemptsare derived within the model. The investment cost of this size of switch is split


45

into port, processing, unattributed (i.e., costs that are incremental toconveyance but are not specifically driven either by busy hour erlangs or callattempts), and common costs (relevant to the concentrator, but not the localswitch). The costs per unit of capacity are progressively developed into costsper busy hour minute and ultimately costs per minute, using a number ofparameter assumptions, including erlangs per circuit and the proportion oftraffic that occurs in the busy hour. These calculations and the results for thebottom-up model in 1994/95 are shown in the print-out below.

The costs of R-L transmission are modeled separately from the costs ofother types of transmission (unlike the top-down model). The transmissionmodel builds up the costs of electronics (end equipment and repeaters), cableand duct on the assumption of a point-to-point transmission network, i.e., eachlogical route is assumed to have a dedicated line system and fibre pair.

The electronics cost per 2 Mbit/s is computed, starting from a busy hourtraffic distribution and applying provisioning rules to derive the costminimizing choice of line system size. Three different sizes of line system areconsidered: 34, 140 and 565 Mbit/s. The transmission model derives a mix ofsystem sizes for R-L routes and determines endogenously the utilization ofelectronics.

The duct and cable costs per 2 Mbit/s depend upon the average route length,the amount of duct sharing between logical routes, the amount of physicaldiversity allowed for, and the number of line systems per logical route (derivedin the optimization of electronics). Different sizes of cable are considered andthe lowest cost size is selected, given the input cost assumptions.


46

BO

TT

OM

-UP

NE

TW

OR

K M

OD

EL

RE

SUL

TS

FO

R L

OC

AL

SW

ITC

H A

ND

CO

NC

EN

TR

AT

OR

, 199

4/95

Loc

al s

wit

chIn

vest

men

tE

rlan

gsA

nnua

lisat

ion

Penc

e pe

rPr

opor

tion

Penc

eR

outin

gC

ompo

nent

cost

pe

per

circ

uit

rate

busy

hou

rof

traf

fic

inpe

rfa

ctor

sco

st in

unit

min

ute

busy

hou

rm

inut

epe

nce

per

min

ute

Port

cos

tsT

ande

mfa

cing

2,94

60.

5926

.6%

0.30

9.2

%0.

028

0.9

0.02

5A

cces

sfa

cing

2,94

60.

3526

.6%

0.50

9.3

%0.

047

1.1

0.05

2O

LO

faci

ng2,

946

0.51

26.6

%0.

3411

.2%

0.03

90.

00.

000

Proc

essi

ngco

sts

5.17

28.5

%0.

379.

2%0.

034

1.0

0.03

4T

otal

0.11

1

Con

cent

rato

r (a

vera

ge o

f ho

st a

nd r

emot

e co

ncen

trat

or s

wit

chin

g co

sts)

Inve

stm

ent

Erl

angs

Ann

ualis

atio

nPe

nce

per

Prop

ortio

nPe

nce

Rou

ting

Com

pone

ntco

st p

erpe

r ci

rcui

tra

tebu

sy h

our

of tr

affi

c in

per

fact

ors

cost

inun

itm

inut

ebu

sy h

our

min

ute

penc

e pe

rm

inut

eL

ine

cost

s97

Port

cos

tsL

ocal

sw

itch

faci

ng2,

993

0.35

28.5

%0.

54 9

.3%

0.05

11.

00.

051

Proc

essi

ngco

sts

10.7

328

.5%

0.74

11.0

%0.

082

1.0

0.08

2T

otal

0.13

3

Not

es:

The

uni

ts in

the

inve

stm

ent c

osts

per

uni

t are

bus

y ho

ur e

rlan

gs f

or p

ort c

osts

, bus

y ho

ur c

all a

ttem

pts

for

proc

essi

ng c

osts

, and

line

s fo

r lin

eco

sts

(whi

ch a

re p

art o

f th

e co

sts

of a

cces

s).

For

full

deta

ils o

f th

e ca

lcul

atio

n m

etho

d, s

ee O

FTE

L (

1997

e).


47

Annex C


48

Annex D


49

REFERENCES

Armstrong and Doyle (1994) Access Pricing, Entry and the Baumol-Willig Rule,August 1994

ART (1997) What do new entrants have to pay for interconnection in France, PressRelease, 10 April 1997

BT (1997a) Long Run Incremental Cost Methodology, May 1997

BT (1997b) Long Run Incremental Cost Model: Relationships and Parameters, May1997

BT (1997c) Accounting Documents, July 1997 (HCA) and September 1997 (CCA)

BT (1997d) Detailed Attribution Methods, September 1997

DTI (1991) Competition and Choice: Telecommunications Policy for 1990s,Department of Trade and Industry, HMSO

European Commission (1997) Working Document on Interconnection Pricing in aLiberalized Telecommunications Market, Directorate General XIII, 6 August 1997

FCC (1996) Implementation of the Local Competition Provisions in theTelecommunications Act of 1996, CC Docket No. 96-98, First Report and Order,August 1996

Myers (1994) Predatory Behavior in UK Competition Policy, Office of Fair TradingResearch Paper Number 5, November 1994

NERA (1996a) The Methodology to Calculate Long Run Incremental Costs, March1996

NERA (1996b) Reconciliation and Integration of Top Down and Bottom Up Models ofIncremental Cost, June 1996

NERA (1996c) Reconciliation of Top Down and Bottom Up Models, December 1996

NERA (1997a) An Assessment of the 1995/96 Top Down Model, May 1997

NERA (1997b) An Assessment of the Interim 1996/97 Top Down Model, August 1997

OECD (1997) Main Economic Indicators, Statistics Directorate, February 1997


50

OFTEL (1994) A Framework for Effective Competition, Consultative Document,December 1994

OFTEL (1995a) Effective competition: Framework for Action, Statement, July 1995

OFTEL (1995b) Pricing of Telecommunications Services from 1997, ConsultativeDocument, December 1995

OFTEL (1996a) Pricing of Telecommunications Services from 1997, ConsultativeDocument, March 1996

OFTEL (1996b) Pricing of Telecommunications Services from 1997, Statement, June1996

OFTEL (1996c) Network Charges from 1997, Consultative Document, December 1996

OFTEL (1997a) Long Run Incremental Costs: the Bottom-up Network Model, March1997

OFTEL (1997b) Network Charges from 1997, Consultative Document, May 1997

OFTEL (1997c) Network Charges from 1997, Statement, July 1997

OFTEL (1997d) Guidelines on the Operation of the Network Charge Controls, October1997

OFTEL (1997e) Long Run Incremental Costs: the Bottom-up Network Model, March1997

U.S. Charges in Table 8 Taken from Interconnection Agreements:Ameritech Illinois/AT&T of Illinois, January 14, 1997Ameritech Illinois/TCG Illinois and TCG Chicago, January 7, 1997Central Telephone Company of Illinois/MFS Intelenet of Illinois, September 20, 1996New England Telephone and Telegraph Company/U.S. ONE CommunicationsServices, March 6, 1997


51

COST AND PRICING OF INTERCONNECTION SERVICES IN ALIBERALIZED EUROPEAN TELECOMMUNICATIONS MARKET

Günter Knieps18

INTRODUCTION AND EXECUTIVE SUMMARY

With the beginning of 1998 market entry will be permitted in all subparts oftelecommunications networks, including cable-based infrastructure as well astelephone services, in nearly all European countries. As a consequence,problems of network access as well as network interconnection gain increasingimportance. Global entry deregulation, however, does not automatically implythe abolishment of all sector-specific regulation. In the first place, a long-termproper role of government intervention with respect to technical regulations, forexample the coordination and allocation of radio frequencies, the organizationof number portability, the design of standards etc., remains in order to guaranteean adequate framework for a competitive telecommunications sector.Secondly, politically desired universal service objectives remain to beorganized by entry-compatible instruments—like, for example, a universalservice fund. Thirdly, remaining network-specific market power needs to bedisciplined by regulatory instruments or competition policy respectively.Future regulation dealing with costing and pricing of interconnection servicesis under debate within the individual European Union (EU) countries as well ason the level of the EU as a whole.

The purpose of this paper is to focus on the role of market power whereinterconnection and network access problems are involved.19 The controversyabout the pros and cons of obligations for network interconnection, detailedregulation of tariffs, including the control of the underlying cost conditions, isvery topical all over the world. The recent “Working Document OnInterconnection Pricing In A Liberalized Telecommunications Market,” issuedby the European Commission, Brussels, August 6, 1997 (Working Document),provides a paradigm change in the current debate on costing and pricing ofinterconnection services. Until now the EU regulatory framework forinterconnection remained rather vague, only pointing out the necessity of non-discriminatory, transparent and cost-oriented interconnection tariffs. Incontrast, the Working Document contains very concrete and detailed costingand pricing prescriptions which, after acceptance, may have a strong impact onthe competition process of European telecommunications markets in the future.The purpose of this paper is therefore to provide a critical appraisal of thisWorking Document as well as to develop alternative proposals better suited tobenefit the liberalized European telecommunications market.


52

The paper is organized as follows: Part A sketches the EU history withrespect to open network provision (ONP) policy in order to demonstrate theparadigm shift represented by the recent Working Document. In part B themarket-share concept of market power applied within the InterconnectionDirective (Council Directive 97/33, EC on Interconnection in Telecommunica-tions, June 30, 1997) is criticized as an important source for an oversizedregulatory basis. Instead, the criteria for an economically justifieddisaggregated regulatory framework for the ONP concept are pointed out inorder to localize the proper regulatory basis. It is shown that the regulatory basisshould be restricted to monopolistic bottleneck areas (local cable basednetworks).

Part C provides a detailed analysis of costing and pricing of interconnectionservices within the ONP concept. Section C. I deals with the role of long-runincremental costs (LRIC) in determining interconnection prices. The WorkingDocument assumes that interconnection prices should be equal to LRICbecause this would enable a firm to recover all of its costs related to providinginterconnection. In contrast we indicate that LRIC can only serve as a (long-run) lower boundary. Moreover, mark-ups are necessary to guarantee the long-run survival of the incumbent firm and to provide adequate incentives for futureinvestments. There should be no ex ante regulatory allocation of mark-ups;instead carriers should be allowed to determine market-driven mark-ups inorder to cover the total costs of interconnection.

Section C. II deals with top-down versus bottom-up approaches todetermine long-run incremental costs (LRIC). Whereas the Working Documentis strictly in favor of bottom-up approaches, we conclude that engineering-economic models are inadequate for calculating decision-relevant incrementalcosts for the established carriers. In particular, the various sources which leadto an underestimating of the real incremental costs are laid open. As a prominentexample the Local Exchange Cost Optimization Model (LECOM) will bediscussed. In contrast, top-down approaches based on the cost accountingsystems of established carriers should be applied to derive decision-relevantincremental costs.

Section C. III deals with principles of interconnection pricing, focussing onthe relationship between price-differentiation and price-discrimination. TheWorking Document for example argues that a dominant incumbent operatorwill be in a better position to bear the economic risk of predicting and investingin additional capacity, and subsequently, that it would be inappropriate for newentrants to pay according to capacity-based interconnection charges.Furthermore, the linking of interconnection charges to retail prices would lockthe new entrant into the same retail tariff structures as those chosen by the


53

incumbent. In this section it is shown that the regulatory enforcement of suchpricing rules would entail price discrimination to the disadvantage ofestablished carriers which would subsequently hamper welfare-improvingprice-differentiation. In particular, it is argued that regulatory restrictions withrespect to flexible pricing on interconnection tariffs not only unduly restrict theestablished carrier’s pricing behavior but are also undesired from a welfareeconomic point of view.

Section C. IV provides an economic evaluation of the “best-practiceapproach” taken within the Working Document. “Best current practice”interconnection charges using the three lowest-cost member states are plannedto be progressively reduced in the future, towards a level compatible with abottom-up LRIC based approach. In this section several arguments are givenwhy this “best current practice” approach cannot be supported from theeconomic point of view.20 There is a tendency towards overregulation (e.g.,tariffs for interconnection in long-distance networks should not be regulated,due to the absence of network-specific market power). Large cost differencesfor providing interconnection services may be due to differences in costsbetween different local networks within a country or between countries. “Bestcurrent practice” interconnection charges do not take into account thedifferences between countries in the recovering of subscriber-specific (non-usage-sensitive) costs. Their informational value is also hampered to the extentthat they already reflect rate regulation on the country level. After all, thereference point is chosen inadequately. LRIC must be calculated on the basis ofreal cost data of established carriers (“top-down”). Moreover, LRIC can onlyserve as a (long-run) lower boundary of individual interconnection charges. Inorder to be able to survive, the established carriers must cover the stand-alonecosts of all interconnection services. Remaining market power in monopolisticbottleneck areas should not be regulated by regulation of pricing structures.Instead, the instrument of price-cap regulation should be applied for regulatingthe level of prices.21

The concluding Section C. V points out that the paradigm shift of theWorking Document—which contains very concrete and detailed costing andpricing prescriptions for the member countries—is also reflected within theCommission Recommendation on Interconnection in a liberalized telecommu-nication market, issued by the European Commission, Brussels, October 15,1997. It can be expected that the debate on the proper role of costing and pricingof interconnection services will continue.


54

A. EUROPEAN ONP HISTORY IN TELECOMMUNICATIONS

I. The Concept of ONP in Partially Entry-deregulated MarketsIn 1990 the EU took the first step towards liberalizing telecommunications

services by opening the market for value-added network services (VANS).Voice telephone service as well as public telecommunications infrastructurenetworks still remained legal monopolies owned by state-owned enterprises.Establishment of the internal market for these liberalized services withinEurope required harmonizing conditions for access to and use of publictelecommunications networks and services. The ONP concept was introducedin the Commission’s 1987 Green Paper on Telecommunications Services andgiven substance in Council Directive 90/387/EEC of June 28, 1990 on theestablishment of the internal market for telecommunications services throughthe implementation of open network provision. Subsequent specific Directivesand Recommendations applied the principles of open network provision toleased lines, voice telephony, packet switched date services and integratedservices digital networks (ISDN).22

The purpose of the ONP-policy during the period of partial entryderegulation was to stimulate entry into the VANS market and to ensure “fair”competition between VANS suppliers and the VANS operations of the existingtelecommunications organizations. Article 3 of the Council Directive 90/387/EEC of June 28, 1990 therefore laid down several basic principles ONPconditions must comply with. These principles are as follows:

- conditions must be based on objective criteria;- conditions must be transparent, and published in an appropriate manner;- conditions must guarantee equality of access, and must be non-discriminatory, in accordance with Community law.

Furthermore, it was explicitly stated that ONP conditions must not restrictaccess to public telecommunications networks or public telecommunicationsservices except for reasons based on essential requirements (e.g., security ofnetwork operations, maintenance of network integrity). Focussing on thepreconditions for competition on the VANS market, only a minimally harmonizedoffering of those public telecommunications networks and public telecommunica-tions services identified as being in the European interest were required.

EU’s ONP policy may also have been pursued as an instrument to avoidstructural separation between the VANS activities of the existingtelecommunications organizations and their traditional network activities.23

Since the established carrier was (correctly) considered to be a monopolist on


55

a large part of the market, global regulation of market power was stillconsidered to be necessary, but left to the national regulatory authorities.

II. The Concept of ONP in Globally Entry-deregulated MarketsCommission Directive 96/19/EC (the Full Competition directive) of March

13, 1996 changes the Directive 90/387/EEC by abolishing all legal entrybarriers including free entry into the markets for telecommunications servicesas well as the set-up and provision of telecommunications infrastructurenetworks. Since the future telecommunications infrastructure in Europe isdeveloping towards a set of interconnected networks, owned and operated bymany different organizations, the importance of interconnection is stronglyincreasing. Interconnection may take place among different providers of longdistance networks, among providers of mobile or satellite networks and publiccable-based long distance networks, interconnection also takes place betweenlong-distance telecommunications service providers to local networks etc. Thischanging role of interconnection also led to a revision of ONP-principles. Thebasic philosophy behind the EU ONP policy seems to be that the infrastructureshould be open to all users in the EU, open to any service provider, and open toany provider of elements of the overall infrastructure. The Full CompetitiveDirective (sections 4a-4d) extends ONP-principles to the new fully entry-deregulated environment, focussing on interconnection and public switchednetworks. In addition to the well known criteria of non-discriminatory,reasonable and transparent conditions, the criterion of cost-orientation wasexplicitly introduced. Priority was given to commercial negotiations betweenthe interconnecting parties involved. Council directive 97/33 EC onInterconnection in Telecommunications (the Interconnection Directive) whichwas adopted in June 1997 and is to be implemented into the member states’national laws by December 31, 1997, goes further than the “Full Competition”Directive by introducing a two-tiered approach to ONP-regulation. Providers ofpublic telecommunications networks or public telecommunications serviceswhich are classified as possessing significant market power are subjected tomore restrictive ONP-regulation. This entails the general obligation to providenetwork access (section 4 (2)), the burden of proof that interconnection chargesare cost-based and the possibility of ex ante regulation of interconnectioncharges (section 7 (2)), as well as principles for cost accounting systems(section 7 (5)).

According to the Interconnection Directive, an organization with a marketshare of over 25 percent in a given telecommunications market is considered topossess significant market power (section 4 (3)). Nevertheless, the majorresponsibility with regard to ONP-regulation still has been left in the hands of


56

the national regulatory authorities. National regulatory agencies have theauthority to determine whether an organization has significant market power.According to section 4 (3) they are free to decide whether an organization withmore or less than 25 percent is to be classified as possessing market power.Moreover, principles for interconnection charges and cost accounting systems(section 7), including the control whether tariffs are cost-based etc., areconsidered under the responsibility of the national regulatory authorities. Thus,the Interconnection Directive laid down the general principles of future ONPregulation but left the responsibility for the concrete regulation ofinterconnection to the regulatory authorities of the individual MemberCountries (section 9).

The first (and nearly unchanged) version of the Working Document onInterconnection Pricing in a Liberalized Telecommunications Market, whichwas distributed only one week after the publication of the InterconnectionDirective (see ONP COMMITTEE, ONPCOM 97-29, Brussels, July 7, 1997),leaves one wondering which role EU-ONP policy will play in the future. Incontrast to the Interconnection Directive, the Working Document advocatesvery concrete regulatory measures that will strongly impede the search of theindividual national regulatory authorities for a solution for the remainingregulatory problems with respect to interconnection. Given the extensivedebate on the “correct” interconnection regulation, this EU initiative must beconsidered with caution.

The Working Document contains the following detailed costing and pricingprescriptions for the Member Countries:

(1) Interconnection prices should be calculated according to forwardlooking long-run incremental costs:

“The forward looking long-run incremental cost represents the lowestprice which enables a firm to recover all of its costs related to providinginterconnection and call termination. Any price higher than this will onaverage pay more to owners and investors than they could earn byinvesting in any alternative but similar risky investment.” (WorkingDocument, p. 4)

(2) Long-run incremental costs should be calculated with “bottom up”models:

“This involves building ‘bottom up’ models whereby an economic/engineering model of an efficient network is developed, from which


57

interconnection costs can be calculated by aggregating the costs ofindividual network elements . . .” (Working Document, p. 4)

(3) Reconciliation of “top-down” and “bottom-up” models works:

“Reconciliation of the results of these models with the results obtainedfrom ‘top down’ models...serve to demonstrate that the calculatedinterconnection costs are broadly accurate.” (Working Document, p. 4)

(4) Asymmetric criteria with respect to interconnection pricing:

“Although the dominant incumbent faces mostly one-off capacity costsas a result of terminating interconnected traffic, the individualuncertainty of new entrants about their demand for interconnection,and the informational (transparency) problems and associatedtransaction costs relating to determining each operator’s utilization ofpeak capacity, make it inappropriate for new entrants to payaccording to capacity based interconnection charges.” (WorkingDocument, p. 17)

“Furthermore, the linking of interconnection charges to retail pricestends to lock the new entrant into the same retail tariff structures asthose chosen by the incumbent. This will hinder and in some casesprevent the development by new entrants of innovative retail tariffschemes targeted at different types of user. Thus interconnectioncharges based on retail tariffs may be incompatible with therequirements of Community law.” (Working Document, p. 18)

(5) “Best current practice” interconnection charges:

“Where actual ‘bottom-up’ cost calculations based on forward-lookinglong run incremental costs are not available, interconnection chargesbased on ‘best current practice’ as given in this Recommendationprovide guidance to NRAs when assessing the interconnection chargesproposed by incumbent operators.” (Working Document, p. 13)


58

B. THE APPLICATION OF THE DISAGGREGATEDREGULATORY FRAMEWORK TO THE ONP CONCEPT

The Interconnection Directive introduced the criterion of significantmarket power into the ONP regulatory debate for the first time. It seems obviousthat the criterion of 25 percent market share is rather arbitrarily chosen.However, the national regulatory authorities are currently engaged in the searchfor sound economic criteria to localize network-specific market power as astarting point for remaining sector-specific regulation.

I. The Necessity of a Symmetric Regulatory ApproachSymmetric regulatory conditions should neither advantage nor

disadvantage the former network monopolist. On the one hand, all monopolyprivileges must be abandoned, on the other hand, all one-sided regulatoryobligations (e.g., to cross-subsidize universal services) must end. “In generalterms symmetric regulation means providing all suppliers, incumbents and newentrants alike, a level playing field on which to compete: the same price signals,the same restrictions, and the same obligations . . . But all forms of asymmetricregulation contain an intrinsic bias toward some firms or technologies. . .”24

Even if one accepts criteria like relative market share, financial strength,access to input and service markets etc. as a starting point in order to evaluatethe existence of market power, nevertheless the development of an ex anteregulatory criterion creates a need for a more clear-cut definition of marketpower. This is even more important, because “Vermutungskriterien” on thebasis of market shares can lead to wrong criteria for government intervention inthe telecommunications sector.

Therefore it is important to develop and apply a disaggregated approach ofmarket power regulation. It is necessary to differentiate between those areas inwhich active and potential competition can work and other areas, so-calledmonopolistic bottleneck areas, where a natural monopoly situation (due toeconomies of bundling), in combination with irreversible costs, exists. It can bedemonstrated that the regulation of market power is only justified inmonopolistic bottleneck areas. In all other cases, the existence of active andpotential competition will lead to efficient market results (see figure 1).

The pressure of potential competition can already create incentive for theactive supplier to improve the quality and variety of services as well as toproduce more efficiently. These networks are therefore called contestable.25 Anessential condition for the functioning of potential competition in order todiscipline firms already providing network services is that the incumbent firmsdo not have asymmetric cost advantages in comparison with potential entrants.


59

Figure 1: A Disaggregated Approach to Market Power Regulation

sub-market with sunk costs without sunk costsnatural monopoly (1) (2)(bundling advantages) regulation of market potential competition

power (non-contestable (contestable networks)networks)

no natural monopoly (3) (4)(bundling advantages) competition among competition among

active providers active providers

An important example for contestable networks are the markets fortelecommunications services, which are often provided via service networks.Even the market for public telephone services is contestable, because suppliersof value added services are also prepared to offer telephone services (after legalentry barriers disappear). An important condition, however, is the guarantee ofnumber portability. The term “number portability” means the ability of users oftelecommunications services to retain, at the same location, existingtelecommunications numbers without impairment of quality, reliability orconvenience when switching from one telecommunications carrier to another.Even if market shares of incumbent firms are large, inefficient suppliers wouldthen be immediately confronted with rapidly decreasing market shares. Butcontestable subareas can also be localized in the area of telecommunicationsinfrastructure. The pressure of potential competition in wireless networks, forexample satellite, microwave systems, mobile communication, is guaranteed aslong as symmetric access to complementary inputs, for example right of way,radio spectrum etc., is ensured. More generally, an important condition for theeffectiveness of actual and potential competition is that all (active and potential)suppliers have equal (symmetric) access to the complementary monopolisticbottleneck.

In contrast, in local cable-based networks, where sunk costs are relevant,consumers, who would intrinsically be willing to switch immediately to lesscostly firms, cannot do this. Market entry therefore cannot be expected easily,if sunk costs are sufficiently high and very relevant. Therefore we can concludethat sector-specific ex ante regulatory intervention in order to discipline marketpower can only be justified in non-contestable networks (monopolisticbottleneck areas), i.e., where bundling in combination with irreversible costs isrelevant. Sunk costs are no longer decision-relevant for the incumbentmonopoly, whereas the potential entrant is confronted with the decisionwhether to build network infrastructure and thus spend the irreversible costs.The incumbent firms therefore have lower decision-relevant costs than the


60

potential entrants. This creates scope for strategic behavior of the incumbentfirms, so that inefficient production and monopoly profits will not necessarilyresult in market entry.

The aim of future regulatory policy should, however, not be the globalregulation of markets. Instead, only a disaggregated regulation of non-contestable networks is justified. The aim is then to localize the market powerin monopolistic bottleneck areas and discipline this market power by regulatoryintervention. Asymmetry of market power due to monopolistic bottleneckfacilities, however, does not by itself require asymmetric regulation. Instead,the symmetry principle requires that all firms have access to localtelecommunications networks on terms identical to those of the incumbent(nondiscriminatory access). The symmetry principle demands that onlybottleneck facilities are regulated, irrespective of whether the owner is theincumbent or a newcomer. The disaggregated location of market power issummarized by figure 2.

Figure 2: A Disaggregated Location of Market Power inTelecommunications Systems

Competitive/ Noncontestable contestable (monopolistic bottleneck)

Terminal equipment X ---Telecommunicationsservices (includingvoice telephone services) X ---Satellite/mobilenetworks X ---Long-distance cablebased networks X ---Local cable based networks --- X

I. The Fallacies of Asymmetric RegulationSymmetric regulatory conditions should neither advantage nor

disadvantage the former network monopolists. There is a wide range of possibleasymmetric regulation. Whereas in the past monopolistic carriers wereprotected by legal entry barriers, the regulatory pendulum now seems to swingin the opposite direction. Asymmetric regulation in favor of newcomers ismotivated by the conviction that even after the abolition of the legal monopolythe incumbent carrier would still possess a factual monopoly position on thenetwork infrastructure and the normal voice telephone service. Therefore initialsupport of newcomers—at least for a sufficient transition period—has been


61

recommended recently in national regulatory debates,26 as well as on thesupranational level. For example, the OECD, which argued in favor ofregulatory symmetry (competitive neutrality) in their Working Party onTelecommunication and Information Services Policies in April 1997 (p. 11),27

issued an amendment in September 1997 in favor of asymmetric regulation (p.3): “This does not ignore the need and the importance of having asymmetricalregulation, even in liberalized markets, while effective competition is notwidely established.”28 However, all asymmetric regulation runs the risk of theregulators preserving existing competitors rather than the competition process.

II. Symmetric Regulation of Monopolistic BottlenecksA symmetric framework of regulation can be created by designing a

disaggregated regulatory approach focussing on network-specific marketpower.29 Applying this theory to the ONP concept the following lessons can bedrawn:

(1) Stable, network-specific market power in telecommunicationssystems can only be localized in local cable based networks. In all otherareas, including telephone services, satellite/mobile networks or longdistance cable based networks, active and potential competition doesnot allow for excessive profits. Even if market shares of incumbentfirms are large, inefficient suppliers would lose their customers.

(2) As long as monopolistic bottlenecks are not involved, privatebargaining solutions on interconnection conditions are not onlybeneficial to the carriers themselves, but also improve the marketperformance of the network services provided to the customers.Irrespective of the market size of the carriers involved, inefficientsuppliers of interconnection services are rapidly confronted withstrongly decreasing market shares, due to the pressure of alternative(potential) network service providers.

(3) Regulation of ONP conditions should be strictly limited tomonopolistic bottlenecks. The market power involved in monopolisticbottlenecks may seriously disturb the bargaining over accessconditions.

As an immediate consequence of this disaggregated regulated framework itfollows that the “best current practice” interconnection charges at the nationallevel (Working Document, p. 8) can only be considered as overregulation. Due


62

to the absence of significant market power there simply is no need for regulatoryintervention. Instead, bargaining on interconnection conditions will lead toefficient results. A similar reasoning holds for the competitive parts ofinterconnection on a metropolitan level.

C. COSTING AND PRICING OF INTERCONNECTION SERVICESWITHIN THE ONP CONCEPT

I. The Role of Long-Run Incremental Costs in DeterminingInterconnection Prices

The Working Document argues that the lowest price which enables a firmto recover all of its costs related to providing interconnection is equal to thelong-run incremental costs (LRIC). This cost concept became quite popular intopical regulatory debates as new entrants were arguing for low interconnectioncharges or asymmetric regulation respectively in an attempt to gain startingadvantages. Nevertheless, pricing according to LRIC is not compatible withcompetitive telecommunications markets which require that no specific type offirm, neither incumbents nor entrants, and no specific form of competition,neither infrastructure nor service competition, is given preferential treatment.

It is well known that even after complete entry deregulation economies ofscale and economies of scope create common costs which cannot be directlyattributed to the individual network services. Although activity-based costingcan help to identify the directly attributable costs to specific products it is stillnot possible to declare all costs as incremental costs without applyingeconomically unjustified allocation of common costs. Provided the establishednetwork carrier is determining incremental costs based on decision-orientedaccounting methods it becomes immediately clear that the sum of theincremental costs does not allow survival. In fact the established firm must alsocover its product-group specific costs as well as the firm-specific overheadcosts by means of mark-ups on the LRIC. In order to avoid inefficient bypassactivities of entrants, market-driven mark-ups should be raised by theestablished carriers. An obligation to provide the services according to LRIC,however, would disturb the symmetric treatment of infrastructure owner andservice provider. Incentives to be the owner of infrastructure forinterconnection and network access would disappear, because it would becheaper to use the infrastructure of the competitors and avoid a contribution tothe common costs. Symmetric treatment of owners and users of infrastructuretherefore requires that the stand-alone costs of network infrastructure must becovered.


63

Applying these general principles to the network termination problemsdealt with in the Working Document the following lessons can be drawn:

- The Working Document (at least implicitly) rests on the Total ServiceLong-Run Incremental Costs (TSLRIC). As service call terminationeither the local, the metropolitan or the national level is considered (p.8). Nevertheless, in future competitive telecommunications marketsthe concept of Total Element Long-Run Incremental Costs (TELRIC)becomes the most important incremental cost concept. Theincumbent’s offers to be priced using this methodology generally willbe “network elements” rather than telecommunications services,defined by regulation. TELRIC-based pricing of discrete networkelements or facilities such as local and tandem or transit switchingfunctions or local loops is more economically rational than TSLRIC-based pricing.30

- As pointed out by Haring and Rohlfs there could be separate prices foreach switching occurrence and for inter-office transmission.31 Theprovision of (unbundled) network components is only viable if thestand-alone costs of the different components are covered, becausecross-subsidization among components becomes unstable undercompetition.

- Pricing the (non-traffic-sensitive, subscriber-dedicated) local loopfor interconnection purposes has been explicitly rejected in theWorking Document (pp. 18f.). Nevertheless, not only the incrementalcost of the final link between the customer and the local exchange butalso the stand-alone costs must be covered for the economic position ofthe provider of the local loop to be viable and for economic incentivesfor future investments to exist.

II. Top-down Versus Bottom-up Approaches to Determining Long-runIncremental Costs1. The Obsolescence of Historical Cost Accounting

According to the Working Document (p. 4), identifying the long-runincremental costs (LRIC) of telephone networks, or parts of telephonenetworks, involves building “bottom-up” models whereby an economic/engineering model of an efficient network is developed. The general purposebehind this requirement seems to be that on competitive Europeantelecommunications markets historical cost data—based on the purchase price


64

of durable assets and (historical) accounting depreciation practices—cannotprovide decision-relevant information on “forward-looking” interconnectioncosts, based on the economic value of the network components.

In competitive industries the value of the firms’s productive assets is equalto the discounted (present) value of the anticipated net cash flows earned by theassets over their remaining useful life. These net cash flows are determined bycompetitive market forces and the firm’s actions, but are not influenced by bookasset value. In regulated industries, however, the value of the firm’s assets inplace, the rate base has been strongly influenced by regulated depreciationcharges. Since the regulatory agencies were under political pressure to keepdown the local rates, and therefore also the capital costs of local networks,artificially low depreciation charges and a too long life time were prescribed(not sufficiently taking into account technical progress, changed substitutionpossibilities etc.).

Nevertheless, the Working Document points in the wrong direction asregards the question of how to design and implement the necessary reform.Although it is true that historical cost accounting is obsolete, the reform towardsdecision-relevant costing should still be based on management and financialaccounting data. “Top-down” approaches are based on real costing data,observing the relationship between input-prices, outputs and the costs ofproduction. In contrast, “bottom-up” models develop (pseudo-) cost-data byengineering approaches. After describing the production function fromengineering data, the cost-output relationship is then derived as a result ofassumed global optimization behavior. It shall be shown in the followingsection that, instead of “bottom-up” engineering models, an adequate reform of“top-down” approaches is needed, based on forward-looking cost accountingmethods. Moreover, it should become clear that the concept of an “efficient”network needs much further elaboration than indicated in the WorkingDocument. Whereas current-cost accounting methods take into account, bytheir very nature, the path-dependency of network evolution (as long as it isefficient from a forward-looking perspective) bottom-up models usually ignorethe strategy of successive upgrading of networks.

2. The Fallacies of Bottom-up Approaches in Determining Long-runIncremental Costs

According to the Working Document (p. 4), “bottom-up” models throughwhich an economic/engineering model of an efficient network is developedshould be applied, from which long-run incremental costs of interconnectioncould be calculated. In the following we shall argue that bottom-up approaches


65

are inadequate for determining the long-run incremental costs ofinterconnection services of established carriers.

Bottom-up models are characterized by process analysis in that theemphasis is placed on simulating the production function from engineeringdata. After describing the production function, the cost-output relationship isthen derived as a result of assumed optimization behavior. Instead of realaccounting data, the cost-data developed by engineering models are simulated(pseudo) data with their informational value strongly dependent on the qualityand the characteristics of the underlying process model. Although the processanalysis approach was not very popular for a long time,32 it has also been appliedin the field of telecommunications.33 Gabel and Kennet developed the so-calledLECOM (Local Exchange Cost Optimization Model) in order to generate datato address the issue of economies of scope in local telephone networks. WithLECOM it became possible to solve the problem of selecting the combinationand location of facilities that minimized the costs of satisfying varying levels ofdemand.34 The three types of facilities within the local exchange carrier’snetwork are the local loop, switching and trunking. The local loop is composedof facilities that provide signaling and voice transmission path between a centraloffice and the customer’s station. The central office houses the switchingcomputer that connects a customer’s line to either another customer who isserved by the same switch, or to an interoffice trunk. Calls between centraloffices are carried on trunks. The model takes as data a city’s dimensions andcustomer usage level. LECOM then searches for the technological mix,capacity and location of switches that minimizes the annual cost of production.The location of the switches are optimized by the non-linear optimizationmodel. In principle, there are an infinite number of possible configurations to beconsidered. For each economically and technically feasible combination ofswitches, a certain number of possible iterations are allowed. An iterationinvolves the calculation of the cost of service at one or more alternativelocations for the switches. For each market, and a given level of demand,LECOM evaluates a number of different switch combinations. In other words,LECOM has been designed to develop a green-field approach. Gabel, Kennet(1994, pp. 390ff.) already pointed out important limitations of engineeringoptimization models. In the first place, optimization models typically are notdesigned to quantify the less tangible costs of providing service (administrativecosts). Secondly, LECOM is limited by bounded rationality. Since globaloptimization is not feasible, only a reasonable number of possible solutions areexamined. It is obvious that a great degree of freedom exists in the search for“plausible” solutions. Thirdly, the value of the pseudo-data approach ultimately


66

rests on the quality and completeness of the underlying process models.Measurement errors and behavioral errors still persist, even in the best model.35

Beyond this immanent critique of engineering-economic models the mostimportant point is that bottom-up models are simply the wrong tool for derivingthe LRIC of established carriers. Even if the analysis is based on a “scorchednode” assumption, which implies that the incremental cost estimate reflects thecurrent network topology, bottom-up models—by their very nature—are notable to derive the LRIC of the efficient network of the established carrier. Thereason is the path-dependency of networks. This means that the gradualupgrading is efficient (given the network history), if the additional costs ofupgrading are lower than the costs of building new network facilities. Thisimplies that the economically efficient incremental costs must be calculated onthe basis of the factual costs of the incumbent’s network in place (including itshistory of upgrading). As long as the incremental costs of upgrading of theestablished carrier are lower than the stand-alone costs of a hypothetical newnetwork of an entrant, the required network capacity should be provided by thehistorically grown network of the established carrier. This is true because entrywould replace the service of the incumbent firm over its existing network, notthe service of a hypothetical efficient provider. Path-dependent costs of gradualupgrading are then economically efficient and are also relevant from a forward-looking costing perspective.36 It is simply not in the spirit of the bottom-upoptimization models to take into consideration the network history.

Even under the scorched node assumption engineering models use the highdegree of freedom of simulation models to find cost-minimizing solutions byignoring the historically-grown network infrastructure already in existence.

Beyond this fundamental critique of the usefulness of bottom-upapproaches for determining the LRIC of the established carriers, other points ofcriticism have already been indicated in NERA’s studies for OFTEL.37 Inparticular, the insufficient determination of the factual usage of networkcapacities, and of the factual routing patterns have been stated.

3. The Necessity of Reforming the Top-down Approaches: from HistoricalCost Accounting to Current Cost Accounting

The Working Document suggests that the necessary and overdue departurefrom historical cost accounting (HCA) in a competitive environment can onlybe accomplished by introducing a bottom-up engineering approach. This isparticularly misleading because the necessary reform should still be based ontop-down accounting approaches. In the following we shall argue that atransition from historical cost accounting to forward looking current costaccounting (CCA) is unavoidable. Under competitive conditions the valuation


67

of the assets and the depreciation-charges must reflect their economic values.The true economic value of any productive asset is the discounted present valueof the anticipated stream of net earnings it is capable of producing. Thus theeconomic depreciation of a productive asset during a time period is the decreasein its economic value during this period. It should be noted that historical bookvalues and historical depreciation patterns typically reflect neither capitalmarket valuation of assets in place nor economic depreciation. A transitionfrom historical cost accounting to current cost accounting thus necessarilyposes the problem of phantom costs (“Altlastenproblem”) due to overvaluationof existing network equipment.38 However, phantom costs should not beconfused with economically efficient forward looking costs of upgradingexisting network (path-dependency). A periodical reevaluation of the assets aswell as an adaptation of economic depreciation rates seems unavoidable,especially in such dynamic markets as telecommunications.

There was an extensive debate, especially in Great Britain, to what extenttop-down and bottom-up models could be reconciliated. Although theexpectations were different in the beginning of this costing project, NERAfinally concluded that significant differences between cost estimation ofinterconnection services based on top-down or bottom-up models remained.Given the serious flaws of bottom-up models in determining the decision-relevant forward looking costs of established carriers, this result seems nolonger surprising.

The established carriers thus need to reform their accounting systems,starting with a reevaluation of their assets and introducing economicdepreciation rates. Moreover, modern accounting systems should be developedfurther. Contribution margin analysis serves to determine which servicesshould be offered within a given network, which services should be expandedand which services should be scaled down in the light of the prevailing pricesituation. It is a very flexible tool, as it can apply current replacement costs(“Wiederbeschaffungsneuwerte”). Moreover, contribution margin analysis canbe employed to determine long-run incremental costs by means of a “top-down” approach. Such an approach is methodically accurate if depreciation,imputed interest, and overhead costs of indirect services are established in sucha way that they correspond to the economic depreciation and interest, as well asthe outlays for jointly used services.

III. Principles of Interconnection PricingThe ONP-principles laid down in the Council Directive of 1990 already

required non-discriminatory equality of access conditions. The InterconnectionDirective of June 1997 goes further by introducing a two-tiered approach to


68

ONP regulation. Carriers of public telecommunications networks or publictelecommunications services that are classified as possessing significantmarket power bear the burden of proof that their interconnection charges arecost-based; moreover, there is the possibility of ex ante regulation ofinterconnection charges. According to the Interconnection Directive, the majorresponsibility for ONP-regulation still has been left in the hands of the nationalregulatory authorities. Nevertheless, the Working Document (pp. 16ff.)provides very detailed recommendations with regard to interconnectionpricing. Because of this, its influence on future interconnection pricing policieswithin European telecommunications should not be underestimated. Therefore,the following provides a critical appraisal of the Working Document’sproposals with respect to interconnection pricing.

It is obviously not the purpose of the Working Document to provide athorough analysis of interconnection pricing principles on liberalized Europeantelecommunications markets. Nevertheless, the Working Document (p. 17)makes clear-cut statements in favor of asymmetric interconnection prices to thedisadvantage of established carriers. The Working Document (p. 17) arguesthat a dominant incumbent operator will be in a better position to bear theeconomic risk of predicting and investing in the additional capacity, andsubsequently, that it would be inappropriate for new entrants to pay accordingto capacity-based interconnection charges. Furthermore (p. 18) it is argued thatthe linking of interconnection charges to retail prices would lock the newentrant into the same retail tariff structures as those chosen by the incumbent.The purpose of this section is to show that those views are inadequate from aneconomic point of view. In particular, it is shown that regulatory restrictionswith respect to flexible pricing on interconnection tariffs not only undulyrestrict the established carrier’s pricing behavior but are also undesired from awelfare economic point of view.

1. Price Differentiation Versus Price DiscriminationIt is important to note that strategies of price differentiation are not only

beneficial from a welfare point of view, but are also compatible with theprinciple of non-discriminatory access conditions. Moreover, the principle ofprice differentiation does not contradict the criterion of cost-orientation ofinterconnection prices. In order to cover stand-alone costs of network servicesmarket-driven mark-ups on incremental costs are unavoidable in order to takeinto account product-group-specific as well as enterprise-specific commoncosts.

In order to apply principles of price differentiation it is important todifferentiate between capacity-based (traffic sensitive) costs of shared


69

equipment (e.g., switching equipment), where opportunity costs of usage aredepending on the factual loading of capacity on the one hand, and the costs ofsubscriber-dedicated non-traffi- sensitive local loops (the final link between thecustomer and the local exchange). There is no doubt that the concept ofopportunity costs shall gain particular importance in future interconnectionpricing. Direct rivalry for usage of scarce network capacities requires anefficient allocation procedure so that those demanders with the highestpreference (willingness to pay) for usage can be served. In other words,(marginal) opportunity costs of supplying a single customer with an additionalunit of (scarce) capacity reflect what the next customer (the first one not served)would (nearly) have been willing to pay for that last unit. Since this principle notonly holds for the final service market but also for interconnection inputs, theprices for interconnection services must reflect the scarceness of capacities andtherefore include opportunity costs (the most valuable foregone alternative,including opportunities foregone in the same market).

(a) Symmetric Peak-Load PricingCapacity-based (traffic-sensitive) interconnection costs should be covered

by peak-load pricing reflecting the factual scarceness (opportunity costs) ofcapacity.39 Interconnection services must be considered as derived demand fornon-storable network services. They must therefore also reflect the varyingscarceness of usage-sensitive infrastructure capacity over time. In contrast,uniform, undifferentiated interconnection charges would not be compatiblewith usage-sensitive tariffs for final network services. Moreover, it is importantthat all market participants irrespective of whether they are entrants orincumbents are confronted with the same efficient pricing signals. Thestatement of the Working Document (p. 17) that it would be inappropriate fornew entrants to pay according to capacity-based interconnection chargestherefore points in the wrong direction, because it entails discrimination againstthe owner of the capacity. If, for example, the established carrier’sinterconnection charges were regulated in such a way that entrants had to payless than the factual opportunity costs, incentives for discriminatory behavioron the part of the established carrier, such as using his scarce capacity for hisown purposes, would be created.

As a consequence, asymmetric pricing rules intended to benefit somespecific entrants finally create incentives for discriminatory behavior andsubsequently impede the competition process. As a further consequence the(implicit) linking of interconnection charges to retail prices should not besurprising, although opposed by the Working Document (p. 18) as incompatiblewith the requirements of Community law. As derived demand prices for


70

interconnection services should reflect similar time patterns of capacity-usageas compared with final network services, even the most innovative retailscheme, targeted at different users, must take into account this cost-basedcharacteristic of network usage reflected in peak-load pricing structures.

(b) Symmetric Optional TariffsIt is important to note that strategies of price differentiation usually lead to

an increase in market size and therefore not only improve the economicsituation of the providers but also of the purchasers of interconnection services.It is to be expected that in the liberalized telecommunications markets of thefuture so-called optional tariffs will become important. Given that a uniformtariff for usage-sensitive interconnection services cannot be higher thanmarginal costs (in order to cover fixed costs) it is always possible to improve theindividual welfare of all market participants (“pareto improvement”) by meansof price differentiation.40 Examples are optional two-part tariffs, where a choiceis provided between a low fixed “entry fee” and a high usage charge (for smallinterconnections demanders), or one with a high fixed “entry fee” and a lowusage charge (for large interconnection demanders). Since the small demandershave the option to stick to the uniform tariff or to opt for the two-part tariff (withlow “entry fee”) and the big demanders are preferring the two-part tariff (withhigh “entry fee”), these non-linear pricing schedules are in fact welfareimproving. Moreover, as long as such non-linear pricing schedules are offeredsymmetrically to all market participants, access conditions are non-discriminatory. Pricing regulation which would forbid such pricedifferentiation would therefore be detrimental. It can be expected that the roleof multi-part tariffs shall increase under the future competitive environment.

Although pricing the subscriber-dedicated local loops for interconnectionpurposes has been explicitly rejected in the Working Document (pp. 18 f.), theproblem nevertheless remains of how to cover their costs. In particular, not onlythe incremental costs of the final link between the customer and the localexchange but also the stand-alone costs must be covered for the economicposition of the provider of the local loop to be viable and for economicincentives for future investments to exist. A first-best economic solution wouldrequire a non-traffic-sensitive (flat) subscriber access charge consideringaccess as a separate product with identifiable costs associated causally withproviding it and an identifiable demand for the option to receive or maketelephone calls.41 Nevertheless, to the extent that a rebalancing of tariffs seemsnot politically feasible in the short-run, traffic-based allocation of subscriberline costs can be applied. In a recent article, Henriet and Moulin have developeda theoretical/axiomatic approach to cost-allocation, sharing the connecting


71

costs of subscriber I among all of his correspondents in proportion to theirtraffic with subscriber I.42 Although this is not a first-best solution, at leastincentives for inefficient network bypass are avoided. To the extent thatinterconnection charges entail parts of the costs of the subscriber-dedicatedcomponents of the local loop (for a transition period) non-discriminatory accessconditions at least require application of symmetric interconnection charges(not discriminating between incumbents and entrants).

IV. Evaluation of the Best-practice ApproachThe approach taken within the ONP Committee Working Document is to

use the interconnection charges in the three lowest cost member states as astarting point. “Best current practice” interconnection charges are planned to beprogressively reduced in the future towards a level compatible with a bottom-up LRIC based approach (p. 12). In the following, several arguments areprovided why this “best current practice” approach cannot be supported, neitherfrom the standpoint of established carriers nor from the economic point of view.

- Price-regulation of interconnection charges at the “national level” (asproposed on p. 8, Working Document) is completely superfluous andimpedes the competition process. Tariffs for interconnection in long-distance networks should not be regulated because they cannot beconsidered as monopolistic bottlenecks. In its report on networkcharges of 1997,43 OFTEL already introduced the economically soundprinciple that for competitive services the established carrier will befree to set his own charges.

- The approach taken in the Working Document (pp. 6ff.) is to compareinterconnection charges around the world and to use theinterconnection charges in the three lowest cost member states toderive a set of “best current practice” figures, at which to aim in theshort term. Such a “best current practice” approach cannot besupported from the economic point of view. Large cost differences forproviding interconnection services may be due to differences in costsbetween different local networks within a country or betweencountries. Moreover, costs, even nowadays, are dictated to varyingdegrees by the regulatory provisions of the various regulatoryauthorities by administrative action. Therefore the allocationprocedures, premiums to cover overhead costs, service lives ofequipment, and methods of depreciation vary considerably. Moreover,international cost and price comparisons that ignore the capacity


72

utilization factors in networks do not provide reliable information. It isalso not possible to compare the costs of network usage or networkaccess in different countries without considering the influence of theexisting, heterogeneous network architectures and topologies.

- Prices are distorted by subsidies. To varying degrees, regulatoryauthorities have allowed cross-subsidies to keep local calling chargesas well as monthly subscriber line charges low. “Best current practice”interconnection charges do not take into account the differencesbetween countries in the recovering of subscriber-specific (non-usagesensitive) costs. Their informational value is also questionable becausethey already reflect rate regulation on the country level. In particular,the welfare improving potentials of price-differentiation principles forinterconnection pricing are obviously not exhausted. To the extentthat—to different degrees in different countries—interconnectioncharges still have to cover part of the “access deficits,” pricecomparisons completely fail to provide reliable information on theefficiency behavior of established carriers.

- The Working Document considers the “best practice” approach astransitional, until “bottom-up” LRIC-based costs are available. Animmediate consequence of the results reached in this paper is that the“bottom-up” approach would provide an inadequate reference point.Instead, LRIC must be calculated on the basis of real cost data ofestablished carriers (“top-down”). Moreover, LRIC can only serve as a(long-run) lower boundary of individual interconnection charges. Inorder for the established carrier to be able to survive, the stand-alonecosts of all interconnection services must be covered, includingproduct-group specific costs of all interconnection services.

V. The Working Document as Basis for Commission Recommendation onInterconnection in a Liberalized Telecommunications Market

Although the Commission’s Recommendation44 has given broad support tothe Working Document, some progress on the level of detailedrecommendations can be found within the Recommendation. For example, inmeantime the necessity of mark-ups has been stated (Commission’sRecommendation, p. 9). Interconnection charges which are based on forward-looking long average incremental costs may include justified mark-ups to cover


73

a portion of the forward-looking joint and common costs of an efficientoperator, as would arise under competitive conditions (Commission’sRecommendation, p. 23). Activity-based costing systems, in which costs areallocated to each product and/or service on the basis of the underlying costdrivers and activities of an efficient operator, are recommended in order tominimize the joint and common costs that cannot be directly allocated(Commission’s Recommendation, pp. 24 ff.). Top-down cost accountingsystems based on current costs rather than historic costs are no longer rejected,although strong emphasis is laid on a “top-down” check through reconciliationwith interconnection costs calculated according to “bottom-up” forwardlooking long run average incremental costs (Commission’s Recommendation,p. 9). Unfortunately, similar to the Working Document, strong emphasis is laidon the transitional use of “best current practices” prices during the transition tolong run average incremental cost-based systems.

Nevertheless, it is important to recognize that the paradigm shift of theWorking Document, which contains very concrete and detailed costing andpricing prescriptions for the member countries, is also reflected within theCommission’s Recommendations. Although the Commission’s Recommenda-tions are legally non-binding, their influence should not be underestimated.Firstly, they back up on EU law that requires interconnection prices to be basedon costs. Secondly, the Commission could ultimately take a government to theEuropean Court of Justice if it refused to rein in prices without adequatejustification. Thirdly, the Commission’s Recommendations may be abused bynational regulatory commissions to overregulate the market for interconnectionservices by detailed inadequate costing and pricing prescriptions.


74

COST AND PRICING OF INTERCONNECTION CHARGES IN THEU.S.: LESSONS FOR GERMANY?

Ingo Vogelsang

1. THE INTERCONNECTION PROBLEM: WHY REGULATION?

The largest segments of telecommunications markets are not contestable.In fact, local loops and many other portions of the local network are bottlenecksthat cannot reasonably be duplicated by entrants. Many other parts of thenetwork (network elements) can be duplicated in principle but theynevertheless represent entry barriers because of sunk costs and economies ofscale or because of economies of scope with bottlenecks.45 Also, for cost anddemand reasons, telecommunications competition can only work if allcompetitors can interconnect with each other. Thus, interconnection is the keyto telecommunications competition. This has been expressed in the UK as theany-to-any principle and is incorporated in the 1996 Telecommunications Actsof both the U.S. and Germany.

Largely because of the interconnection problem, competition has notresulted in worldwide deregulation on a larger scale. On the contrary,competition has tended to increase regulation in most countries. Why doesinterconnection have to be regulated? Couldn’t antitrust policy solve theproblem? Doesn’t the example of New Zealand show that almost total relianceon competition policy can replace telecommunications regulation?Interconnection involves essential facilities that incumbent dominant operators(DOs) do not like to offer entrants voluntarily and at competitive rates. This isalso the essence of the New Zealand experience.46 In the U.S. the essentialfacility doctrine in antitrust might have been invoked instead of regulation tosolve the bottleneck problem. However, as explained by Areeda (1990) andHylton (1991), this doctrine does not work well and, for its implementation,actually requires regulation. Regulatory interconnection policy can, inprinciple, help to solve the bottleneck problem.

What should interconnection regulation achieve in a regime that hasabolished legal entry barriers? The first goal is efficient entry, meaning that theefficient market structure in terms of number and size distribution of firmsshould result. At the same time it means efficient investment, neither wastefulduplication nor under-investment. Production should be cost-minimizing andretail prices efficient in the sense that economic profits vanish and retail priceseither approach marginal costs or follow a Ramsey optimal pattern (in case ofeconomies of scale and scope).


75

2. INTERCONNECTION PRICING AND RETAIL PRICES

2.1. ObjectivesInterconnection conditions and prices shape the amount and type of

competition in telecommunications markets• by moving retail prices to efficient levels (toward marginal cost prices or

Ramsey prices)• by providing incentives for the right type and amount of infrastructure

investment.In the presence of network externalities, interconnection is a necessary tool

to achieve these objectives. The conditions of interconnection have to follow adelicate balancing act between providing too much incentive for bypass of theincumbent provider’s network (and thereby too little incentive for competition)and providing too little incentive for investment by the incumbent networkprovider (and thereby too much incentive for competition). If interconnectionconditions are too restrictive and charges too high for potential interconnectorsthere will be too much incentive for bypass of the incumbent provider’snetwork and, at the same time, too little incentive for competitive entry byinterconnectors. If conditions are too liberal and charges too low there will betoo little incentive for investment by the incumbent network provider and byentrants, with too much incentive for non-facilities-based entry byinterconnectors. The goal of the regulator is to find interconnection conditionsand pricing rules that are at the same time approximately optimal andsufficiently robust to perturbations of the relevant business parameters.

2.2. Pricing Rules2.2.1. The Efficient Component Pricing Rule (ECPR)

The interconnection-pricing rule most hotly discussed in the literature isknown as the efficient component pricing rule (ECPR).47 It says the DO shouldcharge an interconnection price equal to the incremental resource costs ofinterconnection plus the so-called “opportunity cost” of interconnection. Thisopportunity cost is the foregone profit contribution of the DO by providinginterconnection to a competitor who might use interconnection to displaceservices provided by the DO. Thus, the ECPR is driven by the DO’s retailprices. If (a) interconnection and final outputs are generated in fixedproportions and if (b) the DO’s and the entrants’ final outputs are perfectsubstitutes and if (c) entrants take the DO’s price of the competing final outputas given, then the opportunity cost is simply the profit contribution or quasi-rentgenerated by the DO’s final output (simple ECPR). Otherwise, the opportunitycost may be a fairly complicated term, reflecting cross elasticities of final


76

demands, technical substitution and types of competition (sophisticatedECPR).

The main peculiarity in approach taken by the proponents of the ECPR isthe assumption that the price for the final output is given (and chosen optimally)and that the only function of competitive entry is to provide part of the networkservice at lower cost than the DO. The ECPR is therefore a partial rule that dealsonly with a specific aspect of network pricing and competition. It hasnevertheless proven to be highly policy-relevant. The reasons are that, with thesimple version of opportunity cost:

• it is easily understood and practiced,• it is usually embraced by incumbents,• it does not require a change in (regulated) prices of final services and doesnot interfere with politically popular cross subsidies.

With the more sophisticated version of opportunity costs the ECPR is alsotheoretically quite attractive but much more demanding on the regulator (seeArmstrong, Doyle and Vickers, 1996).

2.2.2. Ramsey PricesTheoretically optimal interconnection prices can be determined under the

Ramsey pricing approach taken by Laffont and Tirole (1993 and 1994). Thisapproach simultaneously determines optimal interconnection and final goodsprices, and it makes no a priori assumptions about demand relationships,technology and type of competition. Rather, the assumptions vary, as they do inoligopoly models in general. Depending on which assumptions are made, theapproach leads to different results. In general, these results are complex in thatthey have to deal with the DO’s budget constraint, demand relationships, costrelationships and types of competition. This complexity reflects complicatedrelationships that need to be dealt with and is the price to be paid for generalrather than partial optimization.48 Also, the optimal final goods pricesthemselves obey a complicated markup formula. In reality, a regulator cannothope to capture all these effects at the same time.49

2.2.3. Cost-based PricesThe third approach in the literature is to base interconnection prices plainly

on costs. In the form of marginal-cost-pricing the cost-based approach to publicutilities was dominant among economists for the better part of this century. It isthus not surprising that cost-based pricing resurfaces, as public utilities enter thecompetition age. Instead of marginal costs, it is incremental costs that have


77

taken center stage now. Incremental costs equal marginal costs for small outputchanges but may differ substantially from marginal costs if they include largeoutput changes up to entire services or network components. In addition, stand-alone costs play an important role. They are the costs of a single-product entrantfor providing that single service. Under a cost-based approach the stand-alonecosts of a (hypothetical) wholesale network operator would be an upper limitfor interconnection prices charged by an integrated incumbent. This holdsbecause prices above stand-alone costs would be unsustainable under(hypothetical) competitive conditions. Usually, the lower limit would beincremental costs. Otherwise, the interconnection service would be cross-subsidized.50 While stand-alone costs are always an upper bound forcompetitive prices, there usually exist smaller upper bounds that would berelevant. Imposing stand-alone costs as an upper-bound constraint oninterconnection charges is therefore adequate, while setting them at stand-alonecosts would often be too high. For example, a bottleneck is defined by the factthat duplication of the facility by an entrant would not be economical, meaningprecisely that supplying the entrant at stand-alone costs would not be feasible.There are good reasons to price interconnection services close to or at the lowerbound rather than at the upper bound.

As is well known, pricing at incremental costs can only be optimal underspecific conditions. In particular, there should be no regulatory incentiveproblem and the technology should exhibit no economies of scale and scope.Under these assumptions, however, the interconnection problem would betrivial to begin with. There would simply be no bottleneck. So, why have I beenamong those who have advocated basing interconnection pricing onincremental costs with limited markups (usually staying below stand-alonecosts)?51

The first reason is the presumption that economies of scale and scope in thetelecommunications industry are no longer very pronounced. This is the basisfor allowing competition in the first place. Econometric cost estimates are fairlyambivalent on the prevalence of economies of scale and scope intelecommunications networks.52 At the same time economies of scale and scopeappear to prevail at least in parts of networks. However, service-specificeconomies of scale can be captured in the average incremental costs of thoseservices. Thus, all that would be left are true common costs that do not relate toservice-specific scale. As observed by Burton, Kaserman and Mayo (1997), inpractice, the size of such common costs tends to be overestimated.

The second reason is that interconnection provides for particularly largenetwork externalities (conditional upon entry of other firms) that justifyreduced markups of interconnection prices on costs. These network


78

externalities and the accompanying reciprocity of calling have even generatedproponents of a so-called “bill-and-keep” approach under whichinterconnection would be implemented with zero interconnection charges. Iftraffic is symmetric and transaction costs of interconnection charges are high,bill-and-keep may be a desirable approach.

The third reason is that information about the relevant technological anddemand properties necessary to derive Ramsey prices or the ECPR is eitherunavailable or squarely rests with the DO. This makes it virtually impossible todirectly implement the Ramsey approach and the ECPR.53 One could argue thatasymmetric or lacking information can be captured in Bayesian incentiveschemes using subjective probability distributions. However, such schemeslose much of their power if uncertainty parameters are too vague and too many.Without specific information about demand relationships and the state ofcompetition, uniform markups on incremental costs are an appropriate upperbound for interconnection charges. This holds, among others, becausecompetition makes firm demands more elastic.

The fourth reason is that markups over marginal/incremental costs forintermediate inputs create inefficiencies known as double marginalization.Since entrants have their own overhead and other common costs, they have tocharge a markup on top of the interconnection prices they pay. Thus, viablecompetition in the retail market cannot be of the Bertrand type (unless the retailmarket is sufficiently differentiated). Rather, retail competition itself entailsmarkups and quantity adjustments that reduce the optimal interconnectioncharge (possibly below incremental cost).

The fifth reason is that high interconnection charges are the best instrumentfor collusion between competing telecommunications carriers.

Sixth, high interconnection charges would invite possibly inefficientbypass investments by entrants.

Our suggestion therefore has been to impose a burden of proof on the DO.We start out with interconnection prices at average incremental cost of theinterconnection service and let the DO justify any markups over and aboveaverage incremental costs (Arnbak et al., 1994, Mitchell et al. 1995, Mitchelland Vogelsang, forthcoming). In using this approach, we have deliberately notgiven the regulator the task to optimize over the whole set of markets, as doneby Laffont and Tirole (1993 and 1994). At the same time, in contrast to theECPR, we do not take the final goods price as given (by regulation or the DO’smarket power). Rather, in the short and intermediate term we suggest retailprice caps with full rebalancing, while ultimately we want to leave it tocompetition to determine the optimal prices for the final goods. The philosophybehind this approach is that interconnection and final goods can be sufficiently


79

separated so that (for some time) interconnection can be regulated while theretail services produced with interconnection as an input can be left to marketcompetition.

Foreclosure incentives of the DO’s simultaneous pricing in theinterconnection market and in the retail markets can be curtailed with therequirement of imputation. Imputation means that the incumbent may not priceinterconnection (resale, network elements) at a lower price to itself than toothers. While internal prices, in contrast to external transaction prices, do notusually have direct allocative effects (because internal payments cancel eachother out), they can be used as an accounting device to discover cross-subsidies.The imputation requirement shall thus guarantee that the retail stage is notcross-subsidized. Imputation is a minimum principle to make sure that theincumbent uses no exclusionary practices.54 It still leaves the incumbent with allthe benefits from economies of scope, if any, because, under imputation, theincumbent could charge stand-alone costs for the interconnection services andstill beat the competition. For this it is important to remember that the total costof a retail service equals the stand-alone cost of the intermediate input plus theincremental cost of the retail stage. If an entrant has no economies of scope (andno other cost advantage over the incumbent) the entrant could not break even atan interconnection charge equal to the incumbent’s stand-alone cost unless theincumbent makes economic or excess profit.

I conjecture the best practical interconnection pricing rule to be cost-basedbut that the retail prices resulting under competition will ex post yield the ECPR(in the sense that the interconnection charge will equal the incremental cost ofinterconnection plus the foregone profit contribution). This is trivial forhomogeneous Bertrand competition. In cases of heterogeneous goods and othertypes of competition it would mean that the incumbent would expand in theretail market up to the point where the marginal profit contribution from moreretail sales equals that from more sales to interconnectors. Thus, the ECPR isconjectured to appear as an equilibrium result of competition rather than as astarting point of interconnection price setting by an incumbent with marketpower.

2.3. Price CapsSince the determination of costs is tedious and contentious and always lags

behind cost developments, it is advisable to adjust interconnection charges overtime under price-cap formulas rather than through new cost determinations.55

The possibility of rebalancing within the set of interconnection servicesalso potentially reduces the issue that interconnection is unbundled into manyservices with possible economies of scale and scope within the set of


80

interconnection services. For example, economies of scale and scope may beimperfectly captured under the cost-based approach with initially uniformmarkups for common costs. These common-cost issues could have justified aseparate Ramsey pricing approach within this unbundled set. The maincharacteristic of the approach taken here is that this Ramsey pricing problemcan be seen as separate from the full Ramsey pricing problem that involves theinteraction between interconnection prices and final goods prices and that hasbeen treated by Laffont and Tirole (1993, 1994 and 1996). It deservesmentioning that any Ramsey approach that restricts itself to a subset of marketsis theoretically inferior to one that includes more markets. However, this viewdoes not take into consideration incentives for cost reduction and forinformation revelation that are associated with competition, even if it isrestricted to a subset of the markets. The question is if the beneficial effects fromcompetition in such markets can compensate for the mistake of not including allmarkets under a single regulatory constraint relative to the effects frommodeling and estimation errors in solving the Ramsey problem. We areconfident that it can.

3. U.S. PRICING OF INTERCONNECTION

3.1. Network Interconnection in the U.S.Long-distance competition in the U.S. has been highly successful.56

Interconnection with local networks and equal access were at the heart of thissuccess. After establishment of equal access it was the reduction ininterconnection charges that was responsible for most of the price reductionsenjoyed by long-distance customers.

In contrast, local network competition in the U.S. is still in its infancy. It is,however, starting with a number of large, experienced and financially strongnew competitors that include the interexchange carriers (AT&T, MCI, Sprint,and WorldCom), the mobile carriers, cable TV companies, electric utilities, andincumbent local exchange carriers (ILECs) outside their traditional territories.All of these have networks in place that already are complementary to those ofthe ILECs. In addition, there are the small but fast growing competitive accessproviders (CAPs), who own broadband fiber rings in business centers and citiesand most of whom have by now been acquired by long-distance carriers or cablenetworks. These companies have for some time been actively engaged in localnetwork competition, and some of them have big plans in this market. However,competition is actually penetrating very slowly. Part of this slowdown is due tocontinuing disputes about the best interconnection policy, something that hasnot been fully resolved at this time.


81

While local competition has started to emerge in several U.S. states over thepast few years, it was supposed to receive a really big push only very recently,from the U.S. Telecom Act. This act is the first major reform of federaltelecommunications regulation in over sixty years. Unlike the CommunicationsAct of 1934, the U.S. Telecommunications Act of 1996 (hereafter the U.S.Telecom Act) takes competition (rather than regulation) to be the principalmode of governance for telecommunications markets. The main objective ofthe U.S. Telecom Act is to further open telecommunications markets and toprotect competition against the market power of incumbent dominant carriers.In principle, the new act eliminates barriers to entry and provides substantialentry help to the new competitors of the dominant ILECs. However, in line witha famous proposition by Owen and Braeutigam (1978), according to which theadministrative process of regulation slows down change, much less ishappening than was expected at the time of passage of the act. Almost two yearsafter passage of the act, the ILECs still control 98-99 percent of the localnetworks, although local competition through CAPs had already started tenyears earlier.57 In contrast, once allowed, entry into long-distance markets todayhas become much quicker, as shown by GTE’s and SNET’s success in gaining10 percent market share in their regions within less than two years.

A number of the U.S. Telecom Act’s provisions are designed to createopportunities for new local exchange competitors.

First, state and local regulations restricting entry into telecommunicationsmarkets are to be abolished or declared invalid.

Second, the U.S. Telecom Act foresees three entry paths intotelecommunications markets that go along with three types of services providedby one telecommunications carrier for another:

1. The first is full-scale network entry. In this case the any-to-anyprinciple requires that calls generated on one network can be completedon another network. Thus, there has to be interconnection fortransporting and completing calls. Therefore, all ILECs must allowothers to interconnect at all feasible points of their networks. Physicalcollocation is the rule. ILECs and entrants must transport and terminatecalls from each other’s subscribers and compensate each other for thisservice. Pricing is to be at ILEC’s cost in both directions. This path isfairly uncontroversial, except for the possibility of bill-and-keeppricing arrangements, which entrants like but which are vehementlyopposed by ILECs.


82

2. An entrant can build partial networks that lack certain facilities, suchas local loops. For this purpose, ILECs must offer unbundled networkelements that allow entrants to pick and choose. The elements have tobe priced to approximate economic costs with allowance for areasonable return. This path is highly controversial:

• because of the types of elements to be included. For example,unbundled local loops are both quite feasible and in high demand.In contrast, unbundled switching has been hard to define and is notdemanded much. The most controversial issues are the quality ofthe operations support system (OSS) to be supplied with unbundlednetwork elements and the question whether entrants should beallowed to rebundle by building whole networks from combiningunbundled elements. While the FCC expressly obliged ILECs toallow this, a recent federal court decision denied it. If allowed itcould force state regulators to rebalance ILEC retail rate structureswherever they are not cost based.58

• because of the FCC’s pricing requirement that makes entrantsshare in the economies of scale and scope of the ILECs.

3. Entry can occur via resale of the ILEC’s services. Here, entrants canmake use of quantity discounts available under ILEC tariffs and, moreimportantly, of wholesale discounts. ILECs must offer resale of alltheir retail services at wholesale prices with a discount equal to theircost savings from not selling at retail prices themselves. This is thesimple ECPR applied to the wholesale price.

The pricing methods for unbundled network elements and resale are insome conflict if not all retail prices reflect competitive conditions. Thus, thisconflict is one that is most virulent as long as ILECs’ retail prices are regulatedunder an unbalanced rate structure. In all three cases of envisaged entry, ILECsmust offer number portability, allowing a new competitor to take an existingsubscriber with her/his telephone number, and dialing parity, allowingsubscribers of new competitors to dial without special access codes and withoutdelays. Finally, new entrants must be given access to network infrastructuresuch as rights of way, ducts and poles, as well as telephone numbers, databasesand directories.

In order to make these things happen, the telecommunications companiesmust negotiate in good faith. If agreements are not reached quickly, stateregulatory commissions can be asked to mediate or arbitrate. If this doesn’toccur in a timely manner, the FCC becomes the arbiter of last resort. During the


83

first year after passage of the Telecom Act, reaching agreements betweenILECs and new competitors has been contentious and has involved manycomplicated arbitration proceedings. For example, the MassachusettsDepartment of Public Utilities issued an order in the arbitration betweenNYNEX and several new competitors on December 4, 1996, but even sixmonths later none of the interconnection agreements under arbitration had beensigned.

In addition to the provisions cited, the Modified Final Judgment (MFJ) line-of-business restriction barring the regional Bell operating companies (RBOCs)from long-distance services is softened. As a “carrot” to induce the RBOCs toopen their local markets to competition, they must first convince their stateregulatory commissions, the Department of Justice and the FCC that localcompetition is developing in its territory. By November 1997 no RBOC hadbeen able to convince the FCC that its local markets were sufficiently open tocompetition to drop the restriction on long-distance services.

The FCC has devised average long-run incremental costs for entire services(TSLRIC) and for entire elements (TELRIC) as cost bases for interconnectionprices and prices for unbundled network elements. This means that ILECswould receive the costs (inclusive of cost of capital) for these items includingservice-specific or element-specific fixed or setup costs (that the incumbent orentrant would incur). In line with the above arguments, the FCC also providedfor a limited amount of common costs so that effects of economies of scale andscope were largely included. Katz (1997, p. 22) argues that the resultinginterconnection prices might be too high and could encourage inefficientbypass investment by entrants. In contrast, ILECs complained that the resultingprices would not cover their general overhead costs. While state regulators, bya federal court decision, were not obliged to follow the FCC’s pricing rules,many of them did. However, typically, interconnection charges differ fromproxies provided by the FCC.

At this point in time, the U.S. has refrained from using price caps tointerconnection charges in the area of local competition. There is ampleexperience with price caps for long-distance access to local networks.However, this approach has deliberately not yet been extended, for example, tounbundled network elements. This issue is certain to arise, though, when itcomes to renegotiating initial prices.

3.2. Cost Measurement in the U.S.Firms have always measured some form of their costs.59 The measurement

of economic costs of individual services or network components (elements) intelecommunications, however, is a difficult undertaking because economic


84

costs are forward-looking, because of rapid technical progress and because ofeconomies of scale and scope, resulting from the use of long-lived assets. Thereis consensus today that cost measurement requires three sources or techniquesto be used in combination. Bookkeeping is essential for providing most of thequantity and price data for inputs and outputs. Statistical/econometric analysisis important for establishing empirical regularities between variables.Engineering analysis is needed to establish technical relationships that are thebasis for functional forms, for example, for economies of scale and scope andfor deriving forward-looking cost data.

In the U.S. and the UK, cost models are being used to establish universalservice costs, costs of interconnection and retail services and costs of individualnetwork elements and the retail function. Measurement of local network costswas pioneered by Mitchell (1990) and is now done by all large ILECs. Most ofthem have their own cost models which use modules that are often provided byBellcore (until recently the common research unit of the Bell operatingcompanies). The advantage of these firm-specific models is that they can, inprinciple, best reflect local geographic and market conditions. At the same time,they and their data inputs are less open to outside scrutiny than modelsdeveloped and run by independent institutions. However, to the extent thatfirm-specific models are used in regulatory proceedings, they are gettingscrutinized by regulators and adverse parties. For this purpose, other costmodels have been developed on behalf of firms competing with the ILECs (theHatfield Model), of the National Association of Regulatory UtilitiesCommissioners (the Gabel/Kennet Model) and of a mixed industry group (theBenchmark Cost Model and its derivatives). Most of these latter models are so-called proxy models that, without modifications, can be applied to all regions ofthe U.S. Only the data input changes from locality to locality. Proxy modelstend to depend on a smaller number of data than firm-specific models for whichall necessary data are available to the ILECs (but are proprietary). They can beused to check the accuracy and dependability of firm-specific models. In theUK, British Telecom and an industry group (that included BT) developedcompeting models (BT a top-down model and the industry a bottom-up model)that were reconciled by the Office of Telecommunications with the help of aconsulting firm (NERA). In all these processes, a fairly strong consensus acrossmodels and jurisdictions emerged about the long-run incremental costs of somenetwork elements, switching and transport in particular. In contrast, the costmeasurement for local loops and for operating costs remains contentious.

Measuring costs has been an ongoing and controversial issue. Costmeasurement problems arise for a number of reasons.


85

• There can be conflict about the appropriate economic cost category tobe measured. The candidates are marginal costs, incremental costs,stand-alone costs (and fully allocated costs).

• Once agreement has been reached on the cost category, a method ofmeasurement has to be chosen. At stake are the proper mix ofbookkeeping, engineering and econometric methods as well as bottom-up versus top-down and firm- or location-specific versus general(proxy) model.

• Last, the data inputs for the models have to be collected. These can beactual purchase and consumption data, which are often proprietary andtherefore unavailable, or they can be publicly available data fromstatistics and the trade press.

Cost measurement based on actual data appears to be naturally superior tomeasurement based on estimated or average data. For that reason actual datashould be made available to the regulator. An incumbent TO cannot complainthat data used by outsiders are incorrect but at the same time withhold thecorrect data. But the stated superiority of actual data may not hold at all, for thefollowing reasons:

• Future and efficient costs are being measured, and the actual data mayreflect cost inefficiencies. These could include waste, old technologies,excess capacities and more. For example the loop costs in the formerEast Germany may be exaggerated because not all households areconnected yet or because capacity leap-frogging has occurred.

• The way data are presented may distort cost measurement. Forexample, Burton, Kaserman and Mayo (1997) argue that regulatedmonopolists in the U.S. have an incentive to overstate common costsrelative to directly attributable costs.

The use of actual cost data in regulation (and long-term contracts) is knownto provide weak incentive effects. In contrast, because proxy models are builton (price and quantity) data that are not firm specific and location specific, theycan function as benchmarks with strong incentive effects. Thus, it is not clearthat, for pricing purposes, firm-specific models are superior to proxy models.For the purpose of determining universal service costs, the FCC is now tryingto reconcile various cost models.


86

The FCC, in its Local Competition Order, developed a cost methodologythat closely resembles the one suggested above in Section 2.2.3. The FCCderived proxy cost data from cost models used in state regulatory proceedingsthat had used similar methodologies. The prescription of these numbers and themethodology for state regulatory proceedings has, however, (until now)successfully been challenged before the courts. Thus, at this point, there doesnot exist an official U.S. costing methodology, although most states haveadopted the FCC methods, though not necessarily the FCC’s numbers.

No good models currently exist for operating costs and for the costing of theretail stage. In these cases, bookkeeping costs have been used at least as thestarting point. This is not so bad because, in this case, most of the costs areincurred simultaneously with their expenses so that depreciation and valuationissues hardly arise. However, they may still be inefficient. What could, inprinciple be done is the estimation of the frontier cost function for these itemsfrom the standardized bookkeeping data of all ILECs.

4. U.S. SOLUTIONS TO GERMAN PROBLEMS?

4.1. Cross-country Learning?Learning from other countries’ experiences is always difficult, especially if

countries differ substantially. Many past differences between thetelecommunications sectors of the U.S. and Germany have shrunk or vanishedrecently. No longer is Deutsche Telekom a fully state-owned monopoly.Rather, it is a regulated, partially privatized firm facing competition in all itsmarkets from 1998 onwards. However, the U.S. and Germany continue to differin at least four areas.

First, Germany lacks some of the regulatory problems encountered by theU.S. division of labor between federal and state regulation, although one canargue that similar problems arise for Germany relative to the EuropeanCommission. As a result of the U.S. federal system, U.S. telecommunicationsregulation has a complicated and sometimes inefficient two-tier structure andvaries substantially between the fifty states of the union. As is only too obviousright now, the two-tier U.S. structure leads to disputes about the correctregulatory level to apply, and these disputes retard regulatory reform efforts.Thus, for example, not so much can be learned from the experience with theU.S. Telecom Act because much of it is in the hands of the courts and not yetimplemented.60

Second, as a result of the 1984 AT&T divestiture that followed the 1982consent decree (known as the Modified Final Judgment or MFJ), most of theU.S. telecommunications markets have been vertically separated into local and


87

long-distance providers. While reintegration occurs at this time, verticalseparation is still dominant among ILECs.

Third, the U.S. has the much larger land area that, on average, is as thinlypopulated as the least populated areas in Germany (the 5 percent of Germansliving in the least populated areas still experience twice the population densityof the forty-eight contiguous U.S. states.). Population density is a maindeterminant of telephone line density which, in turn, is highly correlated withloop costs, transport costs and, to a lesser extent, switching costs. Nevertheless,the U.S. has several more densely populated states and population centers thatare quite comparable to Germany. So, careful cost inferences can be drawn. Amajor implication of the different population densities (and the more unequalU.S. income and wealth distribution) is that the universal service problem in theU.S. is and has for a long time been much more pronounced than in Germany.

Fourth, the U.S. has much more experience than Germany with regulationof private firms and with competition in the telecommunications sector. Thislast difference is, however, the main reason why Germany may have somethingto learn from the U.S. At the same time, some of the consequences of a longhistory of regulation do not apply to Germany. One can argue that the longhistory of regulation has resulted in some property rights in the status quo ofregulation and that therefore some practices required to facilitate competitioncould represent regulatory takings in the U.S. (Sidak and Spulber, 1996). Asidefrom the fact that this view is highly controversial,61 it would not apply toGermany because of its lack of regulatory history. This does not, however,totally preclude takings claims in Germany that could be based on a violation ofthe German Telecommunications Act of 1996 (hereafter German TelecomAct), which itself was fully known at privatization of Deutsche Telekom.

4.2. German Problems4.2.1. Interconnection/Bottlenecks

History sometimes repeats itself. Worldwide, incumbent DOs tend to dragout interconnection negotiations with new rivals by demanding highinterconnection charges and by raising technical feasibility issues. This hashappened, for example, in the UK between BT and Mercury in 1984/85 and inthe U.S. between ILECs and new entrants in the early 1990s. It has mostrecently happened between Deutsche Telekom and various entrants. Byrefusing to interconnect, the incumbent DO can prevent a competitive entrantfrom having its subscribers complete calls to subscribers on the DO’s network.This hurts the entrant much more than the DO because only few calls need to becompleted in the opposite direction. This asymmetry of interests provides theincumbent DO with incentives to drag out interconnection negotiations,


88

provide poor interconnection services and charge high interconnection prices.Thus, it appeared to be clear that the incumbent DO can only win by delaybecause the DO has little to gain from quick interconnection while the entrant’svery existence depends on it. However, in all three of the above cases theincumbents’ approach backfired and they missed out on the opportunity tosettle at mutually acceptable negotiated conditions. Instead, regulators werecalled in. In the UK this resulted in speedy interconnection at rates that(initially) favored Mercury. Similarly, in the U.S. interconnection conditionsand rates as stipulated by the U.S. Telecom Act did away with advantages thatILECs probably could have maintained in voluntary negotiations.62 NowDeutsche Telekom finds itself in a similar position. The UK and U.S.experience should be sufficiently powerful to prove (a) that variousinterconnection arrangements (including unbundling) are technically andeconomically feasible and (b) that cost-based interconnection charges do notcripple incumbent DOs. Because competition in Germany is just starting, thereare likely to be more bottlenecks than in the U.S. and UK where duplicatefacilities often already exist.

The current bottlenecks in German telecommunications appear to be thefollowing:

• network access to single-line subscribers (local loops)• local switches• telephone numbers• access to network intelligence and to certain databases.

Will they go away? Have bottlenecks gone away in the past? Long-distancecompetition may look like a good example. Long-distance networks today arenot considered bottlenecks. However, in the U.S., there were no bottleneckseven twenty years ago. There are likely to be no bottlenecks in Germany aswell.63 So, the absence of bottlenecks in the long-distance parts of networkscannot be used as an argument that all bottlenecks will eventually disappear.Any reduction in bottlenecks will depend on a number of developments, such asthe density of networks, growth of markets (geographic and otherwise),available capacities, and new technologies with less sunkness.

The U.S. experience with interconnection of bottleneck services andnetwork components started with MCI’s entry in long-distance services thatrequired access to local networks. Then came the interconnection ofinformation service providers in the so-called ONA (open networkarchitecture) proceedings that started to open the network componentsthemselves to firms other than incumbent local exchange carriers. Although


89

ONA was not a very successful undertaking, it lay the groundwork for physicalaccess to local networks in the Expanded Interconnection proceedings thatpromoted physical and virtual collocation of other firm’s interconnectionequipment at ILEC network nodes.64 At the time of these proceedings, in 1993,Ameritech offered to provide unbundled network components to competingcarriers (in exchange for the permission to offer long-distance services) andRochester Telephone offered to split itself into a wholesale provider and a retailservice provider. All these developments are the predecessors on which the U.S.Telecom Act builds. The act establishes the interconnection requirementsmentioned above. Since the German Telecom Act has fairly similarrequirements, the U.S. experience would be highly relevant. The U.S.provisions have, however, not all been fulfilled yet, as expressed by the fact thatthe FCC and Justice Department have so far rejected all RBOC applications forpermission to enter long-distance services in their territories on the grounds thatlocal competition has not yet been established. Nevertheless, experiencedemonstrates the feasibility of providing resale of local services and ofproviding a number of unbundled network elements, local loops in particular.By spring 1997, Ameritech Michigan, for example, had provided 70,000unbundled loops to competitors. At the same time, unbundled switching had notbeen provided at all. Problems have appeared with the provision of supportservices, such as repair and maintenance, that go along with unbundledelements. Also, the speed of transfer of loops has been an issue. Thus, theexperience is that most elements can be unbundled but perfect and seamlessunbundling (under which customers do not detect that they continue to beserved by the ILEC) remains elusive. So, while the U.S. experience withbottleneck services is positive, it is worth observing whether the price forperfection is not too high.

4.2.2. Incumbency Burdens (Altlasten)Incumbents often claim that they face burdens relative to entrants. The most

common of these claimed burdens are sunk in the sense that they are notinfluenced by current decisions. This could hold for ongoing costs of assets thatwere purchased in the past at high historic costs and have not been depreciatedsufficiently to reflect their current market value. Deutsche Telekom, forexample, claims that its DM45 billion investment in the former East Germaninfrastructure is such a burden.65 It also faces pension costs for civil servantsformerly employed by its predecessors. (For the next few years these costs arepredetermined payments per annum which Deutsche Telekom has to pay into agovernment fund.) Incumbents also may face future expenses (or reducedrevenues, see below) imposed on them by past or current regulatory restrictions


90

but which depend on current decisions of the firm. For Deutsche Telekom themain future expenses are the costs of reducing its workforce.

There appear to be three reasons why incumbency burdens might giveincumbents an excuse to charge higher prices or to receive subsidies.

The first is that fairness may require compensating the incumbent forburdens that other firms (entrants) are not facing. At first blush, this seems to beentirely reasonable. However, two other fairness considerations counter it:

• Usually, the incumbent enjoyed past privileges that others were cutoff from (through entry prohibitions). These privileges enriched theincumbent. However, as the downside, the incumbent may haveinvested or made commitments beyond the time when the privilegeswould expire. The question therefore is if the incumbent could havetrusted that privileges would persist. This can only be a valid excuse ifthe incumbent made irreversible decisions with the assurance that therewould be no competition. In the U.S. this is a much-debated issuebecause competition had cast its shadows for a long time. In Germany,the question is if such expectation could have existed at the time ofDeutsche Telekom’s partial privatization in late 1996. Until that pointin time all decisions regarding Deutsche Telekom and its regulationwere internal to the federal government so that no outside owners hadcommitted funds that would have been adversely affected byregulatory decisions. Expectations at the time of privatization andthereafter would be relevant, though. Such expectations could be basedmainly on the German Telecom Act and, to some extent, on DeutscheTelekom’s prospectus for the sale of shares (which partially includesthe government’s interpretation of the Act).

• The stronger counter-argument against accepting incumbencyburdens under the fairness argument, however, is that incumbentsenjoy present incumbency advantages in the form of unmatchedname recognition, know-how, infrastructure, etc., for which, in turn,the entrants would have to be compensated. These advantages havebeen accumulated under monopoly in part because entrants have beenkept out. Such advantages usually by far outweigh the incumbencyburdens (in the case of Deutsche Telekom probably by a 10:1 margin).

A fairness (or equity) reason for providing relief from a burden is rooted insome popular consensus that may allow for some sacrifice of efficiency. Wehave not found sufficient reason for such a consensus. However, there may


91

actually exist efficiency reasons for granting such relief. The argument herewould be that, without the relief, the DO (or entrants) would make decisions thatwould hurt either static or dynamic efficiency.66 The second reason is thus oneof static efficiency. It relates to two cases:

• The first case involves historic decisions by the incumbent that lead tofuture outlays that entrants do not face because they start with a cleanslate. This holds, for example, for Deutsche Telekom’s payments intoa fund for pensions and other expenses for the civil service status ofsome of its employees. However, these are adjustment costs that ariseon account of technical, market and organizational change rather thanthrough competition per se.

• The second case relates to assets etc. that become obsolete becausecompetition reduces or otherwise changes the demand for theincumbent’s products. Clearly, under competition the incumbent DOloses market share. In principle, this argument holds for any dominantfirm in any market if competition arrives. The main question from anefficiency perspective is whether the total costs in the market areincreasing because competition is rendering some of the incumbent’sassets obsolete. This may occur precisely if the incumbent charges toohigh interconnection prices to competitors who therefore build theirown facilities. Thus, to prevent this inefficiency, interconnectioncharges have to be set such that duplicative investment is avoided if itincreases total costs.

The third reason is one of dynamic efficiency. It says that the incumbenthas had a regulatory contract that, for example, required a large workforce in thepast or disallowed market-based depreciation of assets. If the firm receives nofull compensation for these incumbency burdens it cannot or will not investefficiently in the future (nor will others that may one day be in a similarposition). This is actually an argument about the lack of commitment by aregulator that may lead to long-run inefficiency. Here are again two cases:

• The first is that, because of the financial hardships (reduction in cashflow) caused by incumbency burdens, the incumbent may be unable toraise the funds necessary for investments. Thus, even an efficientincumbent may be precluded from efficient investment.


92

• The second case relates to the reputation effect from breach ofregulatory commitment. Even if the incumbent DO generates enoughcash flow, new investments may appear to be unprofitable because ofthe danger that the regulator will in the future “expropriate” the assetsby not allowing the firm to recover its investment. This argument canspill over to other industries and countries where sub-optimalinvestments would be observed. This argument is unlikely to hold fora firm like Deutsche Telekom that has just been privatized and issubject to a new law and a new regulator, provided the GermanTelecom Act is applied correctly and public promises made by thegovernment before privatization are kept. Then, no regulatory contractwould be breached or reputation of regulatory commitment negativelyaffected. The reputation argument may hold for an evolutionarysituation, like in the U.S., but even here the question is to what extentthe change could have been reasonably anticipated at the time ofinvestments or at the time of entering into employment contracts.

Again, on the efficiency side we found little reason in favor of grantingrelief for incumbency burdens. The case is weakened further by the fact that therelief itself would consist in price increases that would hurt competition andefficiency.

In the U.S., incumbency burdens are usually invoked with respect todepreciation provisions (stranded investments) and services provided in low-density areas. ILECs have generally dealt with excessive workforces in a morequiet way. Most of them are now slim, in part due to incentive regulation, in partdue to the threat of competition. ILECs also have generally adjusted thebookkeeping valuation of their assets to current market values. By far thelargest issue is that of low-density service areas; this is currently attacked underthe universal service provisions discussed below.

In conclusion, the argument of incumbency burdens has only a weak basisin equity and efficiency. This holds for Deutsche Telekom in particular becauseof strong continuing incumbency advantages and because no regulatorycontract would be breached by denying incumbency burdens.

4.2.3. Access DeficitsThe access deficit issue refers to incumbency burdens on the pricing side.

In particular, charges for retail subscriber access (connection fees and monthlyfees) may be underpriced relative to usage fees. As in the case of cost burdens,questions of equity and efficiency arise. Contrary to the case of cost burdens,however, price burdens are more easily resolvable, provided the regulator


93

allows the incumbent DO to rebalance its rate structure. This is something thatU.S. ILECs may have a harder time with than Deutsche Telekom. U.S. localrates continue to be regulated by state PUCs that are under political pressure tokeep residential local rates down. My interpretation is that Congress, throughthe pricing provisions of the U.S. Telecom Act on the unbundled networkelements and wholesale discounts, wanted to force states to allow rebalancingof local tariffs between and within states. However, that has not yet happened.

The German Ministry of Posts and Telecommunications recently decidednot to allow access deficit contributions as part of Deutsche Telekom’sinterconnection charges. The regulator expressly reiterated that the firm mayrebalance its price structure, and Deutsche Telekom had previously publiclyacknowledged that it does not feel constrained by the regulator (DeutscheTelekom, 1996, p.26). Thus, Deutsche Telekom must have some preference forits current rate structure. I have argued elsewhere (Vogelsang, 1996) thatDeutsche Telekom’s rate structure does not appear to be cross-subsidized. One-time and monthly subscriber payments are at least as high as in the UK and inthose U.S. states that have comparable population densities. However, for thesake of argument, let us assume that Deutsche Telekom had priced consumeraccess below its cost. Would that be a sign of a regulatory constraint? Notnecessarily, since a monopoly firm may well maximize its profits (in a two-partpricing arrangement) by pricing access below cost, as shown, for example, bySchmalensee (1981) in a constant-returns-to-scale framework. This can happeneven without income effects and network externalities, simply because lowerend-user subscription charges lead to more subscribers which, in turn, increasedemand for usage.

Thus, an incumbent monopolist or DO could have a long-run interest in across-subsidized rate structure. However, with the advent of competition and(coincidentally) with full telephone penetration, the incumbent’s interest in lowaccess/high usage prices is likely to fade. At the same time, rebalancing mayentail a rate shock that could prove unpopular and reduce Deutsche Telekom’sgoodwill. Thus, Deutsche Telekom definitely prefers high interconnectioncharges to higher fixed subscriber tariffs.

4.2.4. Cross-Subsidies and Universal ServiceAlthough a pricing policy that is financed within the industry can affect

“universal service” only if it has some focus on marginal consumers, virtuallyall past universal service policies have involved such broad-based cross-subsidies that almost every user was both a payer and a receiver of the subsidies.This is usually inefficient but can be politically desirable, especially underincomplete and asymmetric information about the consequences of price


94

structures. If a universal service policy is maintained in the long run it willrequire some form of regulation that may, at the same time, impede deregulationof competitive aspects of the industry. Is a universal service policy thereforereally required in the long run? Universal service policies are so popularbecause telecommunications has a broad user base that includes almost the totalpopulation. Universal service is a catchword similar to the “security of energysupply” that can be used by politicians to channel subsidies and reducecompetition. In the U.S., this has worked for a long time during which implicitand explicit subsidies to rural areas have increased substantially.

Regulation-induced cross-subsidies are a major tool of regulatory interestgroup policy. They totally interfere with competition. Such cross-subsidieshave been substantially reduced in the UK in a twelve-year process that startedwith the first price-cap regime. In the U.S., rate rebalancing progresses atvarious speeds, depending on state public utility commissions. As a rule, raterebalancing is more advanced the more progressive a commission is inimplementing competition. Forerunners in rebalancing are Illinois andCalifornia. Major rebalancing can be expected from the ongoing U.S. universalservice reform and is part of the recent access charge reforms. It will also likelybe forced upon state regulators by the implementation of the competition rulesof the U.S. Telecom Act. Can such rebalancing be expected in Germany? Notlikely. As argued below, the cost of universal service in Germany is much lowerthan in the U.S. where vast rural areas are subsidized. Deutsche Telekom hasnot threatened rebalancing for the near future.67

Regulation-induced cross subsidies tend to vanish only slowly and mayrequire a conscientious policy of dealing with large rent transfers and remainingequity issues in the form of an explicit universal service policy. In that sense,independent of its long run economic justification, a universal service policymay be necessary to overcome resistance (of DOs and affected user groups)against competition, free entry and deregulation of retail (and possiblywholesale) rates. Opposition of the DO against incremental-cost-basedinterconnection prices can be overcome much more easily if funds from auniversal service pool can be used to support high-cost or low-revenueendusers. In fact, universal service policy may help to make the introduction ofcompetition a Pareto improvement.

In the U.S., universal service has a long tradition, although most of theuniversal service policy has in the past been implicit, in the form of broad cross-subsidies, rather than explicit (lifeline and link-up programs). Germany’swaiting lines for telephone services that persisted well into the 1970s wouldhave been politically unacceptable in the U.S. The U.S. Telecom Act has justwidened the scope of universal service to include enhanced telecommunica-


95

tions services for health care providers, schools and libraries. The universalservice provisions were required to get approval to the law by representativesfrom rural states. The law also allows for the definition of universal service tobe adapted by the FCC to new circumstances. Given that the U.S. has vast low-density and remote areas, universal service is soon officially going to becomea multibillion-dollar industry. Abolishing universal service or reducing it to awelfare program is likely to be extremely difficult here. Countries, such as theUK and Germany, that start out universal service with a more limited approachand have denser and more homogeneous populations are more likely to be ableto deregulate universal service should it become obsolete or dwindle.

In contrast to the U.S., calculations of universal service costs in the UKindicate that, from a financing perspective, universal service could be anonexistent problem.68 That does not mean that specific user groups in remoteareas or with low income would not be harmed without access to lowsubscription rates. The empirical question is if the market without such policieswould not provide those rates. To what extent would, without universal serviceobligations (and/or subsidies), a dominant or competitive telecommunicationsoperator deaverage geographic rates or charge high rates to low users? This isan empirical question. Evidence from other competitive service markets (e.g.,airlines) suggests that price differentiation often works in favor of users withlow willingness to pay who are content with lower quality service. Thus,complete cutoff of users may be an exceptional problem precisely becausecompetition generally results in increased and even excess capacities. Theremaining universal service issue could then be dealt with as an annex to other,welfare-type programs that are user-specific and require no interference in thetelecommunications industry. In the U.S., foodstamps would be the analogy.They certainly do not make the food retail sector a regulated industry.

While the German Telecom Act provides for a universal service policy, thisis not triggered automatically. Rather, invoking universal service policy firstrequires a threat by the incumbent DO(s) to abandon service in an area. Holdingback on universal service policy makes sense in a country with no specificuniversal service policy tradition, high telephone penetration, fairly smallincome disparities and fairly small disparities in population density. At thesame time, the costs of universal service are likely to be as small in Germany asin the UK. If that is true, then the real economic costs of a universal servicepolicy are its administrative burdens and its potential interference withcompetition.69 The real danger of universal service policies is that they preventcompetition from fully effecting prices. That is why universal service policieshave to be restricted to a small set of deserving subscribers. If there is anythingto be learnt from the U.S. it is not to emulate their universal service policy.


96

4.2.5. CollusionThe current contentious relationships between incumbent DOs and entrants

do not raise the imminent specter of collusion. The telecommunications sectoris probably subject to most of the same collusion problems as otherconcentrated industries. On top of those, however, telecommunications maybecome particularly prone to collusion for reasons similar to those that favorintegration of telecommunications companies. In addition to competing witheach other telecommunications firms have to collaborate with each other in theform of interconnection, provision of unbundled network elements and in theresale of services. All these collaborations occur, in principle, between allcompeting firms in the industry and on an ongoing basis. The collaboration isalso sufficiently complex that it requires close relationships. Such collaborationbetween multiproduct firms in the form of reciprocal dealings has, in the past,been viewed as one of the reasons for reduced competition in some othersectors, such as the oil industry. We have little experience in this respect withthe telecommunications industry. Due to the antagonistic political process thatjump-starts competition, initially the problem of collusion will be minor.However, over time it may well increase significantly. This is further enhancedby merger activities among new competitors in local networks. Suchconsolidation has occurred in both countries but may be more of a problem inGermany than in the U.S., since in Germany the number of interactingcompetitors is likely to remain smaller. Collusion then need not occur in explicitforms but will be more subtle and therefore less subject to antitrust scrutiny. Thequestion then arises if intermediation in collaborations between competitors byregulators could help reduce the incentive to collude.

4.2.6. Determination of CostsOne might think that, if at all, measurement of its network costs is Deutsche

Telekom’s problem. However, if interconnection charges are to be based oncosts, cost measurement becomes a general problem for the telecommunica-tions sector.

In terms of costing, Germany could, in principle, learn from the U.S.experience about cost models. I have alluded to some of these lessons in anearlier study, which resulted in a substantial dispute.70 I do not want to repeatthose arguments and numerical conclusions here. However, there are meta-conclusions for interconnection costing and pricing suggested by this episode.The first is that costing itself is such a controversial issue that the incumbent DOand the entrants are unlikely to agree on the types of costs to be applied and oncost figures corresponding to each cost type. The confidentiality with whichDeutsche Telekom treats its cost data exacerbates this problem. Thus, it is


97

important to find a way to bridge the gaps. What the U.S. (and UK) experiencesuggests is that the regulator should function as a mediator for cost modelingefforts by the industry. Since Deutsche Telekom by law has to reveal cost datato the regulator but not to its competitors, the regulator is the only institution thatcan question Deutsche Telekom’s cost data. That does not, however, precludeothers, such as the entrants and researchers, from developing alternative costmodels (and their own cost data). It also does not preclude the regulator fromsharing the cost modeling efforts with the industry and experts. Something thatdefinitely has to be avoided is that Deutsche Telekom uses its proprietary dataand a proprietary cost model that cannot be critiqued by outsiders. Rather, themodel should be transparent and available so that it can be analyzed and can befed with cost data supplied by others. In order to agree on such a model, theregulator should establish a group of experts representing the licensedtelecommunications operators, the regulatory body and independent experts(such as the WIK).

5. CONCLUSIONS

The lessons from the U.S. on interconnection policy are a mixed bag.Interconnection and equal access were at the heart of successful competition inthe U.S. long-distance markets. After establishment of equal access, thereduction in interconnection charges was responsible for most of the pricereductions enjoyed by long-distance customers. Using interconnection policyto open local telecommunications markets to competition appears to besubstantially more difficult because it is here where bottlenecks have to be madeaccessible to competitors. This is a contentious and complex task.

Paramount among the competitive provisions of the U.S. Telecom Act arethe duties of the ILECs to provide unbundled network elements and to providewholesale rebates to resellers of all their services to end-users. As a result, newcompetitors can, in principle, start offering local services anywhere, withoutnecessarily having to invest in their own networks. In addition, they have a rightto interconnection with the ILECs, to nondiscriminatory access to rights-of-way, and to number portability. At the same time, most of the largest ILECs willreceive the long-awaited permission to offer long-distance services in theirregions only if they can demonstrate that local competition is happening there.

Although the new competitors and the ILECs are free in principle to reachvoluntary agreements about their mutual cooperation, the regulators are alwaysthe arbitrators of last resort. As a result, in the short run, regulation actually hasbeen increasing. In addition, traditional cross subsidies will soon have to bereplaced by an explicit universal service policy, and a reform of interexchange


98

carrier access charges is being implemented. In principle, the arbitragepossibility of long-distance carriers to substitute ILEC access charges withunbundled network elements may force ILECs to lower long-distance accesscharges even more quickly. Similar forces may be at work with regard to ILECend-user tariffs. However, as long as local competition is not well established,ILECs are not forced to rebalance their retail prices. On the contrary, statepublic utilities commissions, by setting maximum prices for basic services,often prevent ILECs from rebalancing.

At this point in time, U.S. policy on local competition has been frustrated bytechnical difficulties and by legal battles before courts and regulatorycommissions.

The U.S. experience shows that:

• Resale of local services is important for entrants but hard to provide insuch a way that quality equal to the incumbent’s service can be guaranteed.• Unbundled network elements vary in importance, the most importantbeing unbundled local loops. Again, quality is important, in particular,when it comes to the quality of repair and maintenance. Instead of opting forelusive perfection price discounts (like for unequal access) may be in order.These could decline as quality improves over time.• The U.S. carries some historic ballast, for example, fairly high long-distance access charges. It is now recognized that interconnection chargesfor the same service should not be differentiated by use.• The U.S. overemphasis on universal service is motivated largely bydistributional and political concerns that do not hold in a similar way inGermany.• The U.S. and the UK have experience with cost measurement that wouldbe useful for Germany.

REFERENCES

Albach, Horst, Günter Knieps, John Panzar und Bell Communications Research,Costing und Pricing in wettbewerblichen Telekommunikationsmärkten, Study forDeutsche Telekom, January 1997.

Areeda, Philip, “Essential Facilities: An Epithet in Need of Limiting Principles,”Antitrust Law Review 58, 1990, pp.841-869.

Armstrong, Mark, Chris Doyle and John Vickers, “The Access Pricing Problem: ASynthesis”, Journal of Industrial Economics 44, 1996, pp.131-150.


99

Brunekreeft, Gert, “Local versus global price cap: A comparison of foreclosureincentives,” Discussion paper #36, Institut für Verkehrswissenschaft undRegionalpolitik, Albert-Ludwigs-Universität Freiburg/Breisgau, June 1997.

Arnbak, Jens, Bridger M. Mitchell, Werner Neu, Karl-Heinz Neumann, and IngoVogelsang, “Network Interconnection in the Domain of ONP,” Final WIK/EACReport of Study for the European Commission, Brussels, November 1994.

AT&T (1996), “Reply Comments of AT&T Corp.,” CC Docket 96-98, May 30.

Baumol, William J., “Some Subtle Issues in Railroad Regulation,” Journal of TransportEconomics 10, 1983, pp.1-2.

Baumol, William J., and J.Gregory Sidak (1994), Toward Competition in LocalTelephony. Cambridge, Mass.: MIT Press and American Enterprise Institute Press.

Brock, Gerald W., “Local Competition Policy Maneuvers,” in G.L. Rosston and D.Waterman (eds.), Interconnection and the Internet, Selected Paper from the 1996Telecommunications Policy Research Conference, Mahwah, NJ: LawrenceErlbaum Associates, 1997, pp.1-14.

Burton, Mark L., David L. Kaserman and John W. Mayo, “The Economics of CommonCosts: Lessons from (and for) Regulatory Policy,” Paper Presented at AEIConference “Pricing and Costing a Competitive Local TelecommunicationsNetwork,” Washington, November 4, 1997.

Deutsche Telekom AG Bonn, “Unvollständiger Verkaufsprospekt,” 3. Oktober 1996.

“Deutsche Telekom mit Gewinnsprung und deutlichem Umsatzplus im 1. Halbjahr1997,” www.dtag.de/aktuell/presse/archiv/091897.htm.

FCC, Implementation of the Local Competition Provisions in the TelecommunicationsAct of 1996, First Report and Order, CC Docket 96-98, FCC 96-325, issued August8, 1996.

Gabel, David, and Richard Gabel, “The Application of Cost Data in theTelecommunications Industry,” Paper Presented at the 25th TelecommunicationsPolicy Research Conference, Alexandria, Virginia, September 27-29, 1997.

Hylton, Keith H., “Economic Rents and Essential Facilities,” Brigham YoungUniversity Law Review, 1991, pp.1243-1291.


100

Johnson, Leland L., “Telephone Assistance Programs for Low-Income Households: APreliminary Assessment.” Technical Report R-3603 NSF/MF, RAND Corpora-tion, Santa Monica, February 1988.

Katz, Michael L., “Economic Efficiency, Public Policy, and the Pricing of NetworkInterconnection Under the Telecommunications Act of 1996,” in G.L. Rosston andD. Waterman (eds.), Interconnection and the Internet, Selected Paper from the1996 Telecommunications Policy Research Conference, Mahwah, NJ: LawrenceErlbaum Associates, 1997, pp.15-32.

Knieps, Günter, “Phasing out Sector-Specific Regulation in CompetitiveTelecommunications,” KYKLOS, Vol.50, 1997, Fasc.3, pp.325-339.

Kress, Carl B., “The 1996 Telekommunikationsgesetz and the Telecommunications Actof 1996: Toward More Competitive Markets in Telecommunications in Germanyand the United States,” Federal Communications Law Journal 49(3), April 1997,pp.551-619.

Laffont, Jean-Jacques, and Jean Tirole, A Theory of Incentives in Procurement andRegulation, Cambridge, MA: MIT Press, 1993.

Laffont, Jean-Jacques, and Jean Tirole, “Access Pricing and Competition,” EuropeanEconomic Review 38, 1994, pp. 1673-1710.

Laffont, Jean-Jacques, and Jean Tirole, “Creating Competition Through Interconnec-tion: Theory and Practice,” Journal of Regulatory Economics 10, 1996, pp. 227-256.

MacAvoy, Paul W., The Failure of Antitrust and Regulation to Establish Competitionin Long-Distance Telephone Services, MIT Press and AEI Press, 1996.

Masmoudi, Hautam, and Francois Prothais, “Access Charges: An Example ofApplication of the Fully Efficient Rule - Mobile Access to the Fixed Network,”mimeo, April 1994.

Mitchell, Bridger M., “Incremental Costs of Telephone Access and Use,” Report R-3909-ICTF, RAND Corporation, Santa Monica, July 1990.

Mitchell, Bridger M., Werner Neu, Karl-Heinz Neumann, and Ingo Vogelsang, “TheRegulation of Pricing for Interconnection Services,” in Gerald Brock (ed.), Towarda Competitive Telecommunication Industry - Selected Papers from the 1994Telecommunications Policy Research Conference, Mahwah, NJ: LawrenceErlbaum, 1995, pp. 95-118.


101

Mitchell, Bridger M., and Ingo Vogelsang, “Markup Pricing for Interconnection: AConceptual Framework,” in David Gabel (ed.), Opening Networks to Competition:The Regulation and Pricing of Access, Boston: Kluwer Academic Publishers, 1998.

Oftel, “Universal Telecommunication Services: Proposed Arrangement for UniversalService in the UK from 1997,” Consultative Document, London, 1997.

Perl, Lewis, and Jonathan Falk, “The Use of Econometric Analysis in EstimatingMarginal Cost,” in Telecommunications Costing in a Dynamic Environment,Proceedings of the Bellcore – Bell Canada Conference on TelecommunicationsCosting, Held 5-7 April 1989 in San Diego, pp. 825-846.

Schmalensee, Richard, “Monopolistic Two-Part Tariff Arrangements,” Bell Journal ofEconomics 12, 1981, pp.445-466.

Shin, Richard T., and John S. Ying, “Unnatural Monopolies in Local Telephone,”RAND Journal of Economics 23, 1992, pp.171-.

Sidak, J. Gregory, and Daniel F. Spulber, “Deregulatory Takings and Breach of theRegulatory Contract,” New York University Law Review 71, 1996, pp. 851-999.

Vogelsang, Ingo, “Kosten des Ortsnetzes,” Studie des VTM, Mai 1996.

Vogelsang, Ingo, “Die DTAG-Studie ‘Costing und Pricing in wettbewerblichenTelekommunikationsmärkten’: Ansätze, aber keine Antworten,” mimeo, BostonUniversity, August 15, 1997.

Vogelsang, Ingo, and Bridger M. Mitchell, Telecommunications Competition: The Last10 Miles, MIT Press and AEI Press, 1997.

Williamson, Oliver E., “Deregulatory Takings and Breach of the Regulatory Contract:Some Precautions,” New York University Law Review 71, 1996, pp. 1007-1020.

Willig, Robert D., “The Theory of Network Access Pricing,” in Harry Trebing (ed.),Issues in Public Utility Regulation, East Lansing: Michigan State University, 1979,pp. 109-152.


102

INFRASTRUCTURE COMPETITIONAND LOCAL-LOOP UNBUNDLING

Martin Cave & Peter Crowther

1. INFRASTRUCTURE VERSUS SERVICE COMPETITION

The debate about the effectiveness of infrastructure and servicecompetition has occupied the telecommunications industry and its regulators inEurope for some time. Proponents of infrastructure competition argue that it isnecessary to replicate networks over a substantial area in order to get the benefitof full cost reduction and innovation made available by competitive markets.They argue that the telecommunications industry, in which most providers relyfor most of the time upon the network provided by the historic operator, willlimit customer choice and condemn the sector to perpetual and stiflingregulation. When taxed with the argument that duplication of infrastructureimposes avoidable costs on the network, because it prevents the fullexploitation of economies of scale and scope, they argue that in practice secondnetworks either benefit from an economy of scope—being provided, forexample, in conjunction with cable television or other utilities such aselectricity transmission or distribution—or that they rely upon a newtechnology such as wireless, which imposes lower capital demands.Proponents of infrastructure competition also invoke, with good reason, thespur to innovation which infrastructure competition generates. This relates notonly to new methods of delivery as discussed above, but also to newtechnological capacity, such as wider band width.

Infrastructure competition also provides a stimulus to efficiency, by tworoutes. First and most obviously, the incumbent is in head-to-head competitionwith entrants for customers and must match their offerings in price and quality.Secondly, data from entrants provide the regulator with a crucial check on theincumbent’s reported cost data.

Unsurprisingly, proponents of service competition attach considerableweight to the avoidable cost of network duplication and argue, particularly inrelation to broadband delivery systems, that it is more efficient to have a singlebroadband pipe passing all customers’ homes.

Within Europe, the United Kingdom in particular has favored infrastructurecompetition, through a variety of measures sustained over more than a decadedesigned to create incentives to build alternative infrastructure. The first ofthese was the duopoly policy, from 1984 to 1991, which confined new entry toa single operator, Mercury. In return for this advantage, Mercury had to commititself to network roll-out targets. This approach is now generally regarded as


103

conservative and ineffective. Subsequently, entrants received a benefit in theform of relief from access deficit contributions for a limited period.

It is impossible to evaluate the effects of this second measure, which wereprobably small. In any event, the objective of infrastructure competition in theUK has in large measure been achieved, to the extent that there are a largenumber of providers of network services for long-distance conveyance, agrowing number of infrastructure-based providers of international telecommu-nications services, and a growing cable sector which is licensed to pass 65percent of UK homes and already passing two-thirds of them. However,significant infrastructure building has taken place in a number of othercountries which have not relied upon the explicit entry promoting measuresadopted by the UK company.

The European Commission has attempted to adopt a neutral position overthe question of infrastructure or service competition, favoring neither. Definingsuch a regime in terms of regulatory policy creates, however, considerabledifficulties.

This can be illustrated with particular clarity in relation to local loopunbundling. This involves the unbundling of local switching and access throughan arrangement in which a competitor leases the copper wire connecting asubscriber’s premises to the switch to the boundary of the access network (theremote concentrator or local switch) and “hardwires” the copper wire into aswitch of its own, located on the incumbent’s premises (co-location) orelsewhere. Local-loop unbundling is thus a special case of the more generalphenomenon of network unbundling, in which the nature of the cost of theservice provided makes leasing the appropriate charging mechanism. Inpractice, any operator might choose voluntarily to introduce local-loopunbundling. In practice, the term normally signifies that the unbundling occursat the regulator’s behest.

It is widely observed that infrastructure competition is particularly slow totake hold in the local-loop. The reasons for this are not hard to find. First, thelocal-loop is very capital intensive, and requires in particular up-front sunkinvestment. Second, the normal pattern of prices observed in Europeantelecommunications generate high returns to long distance and internationalcalls, but low returns to the local-loop and local calls. Where this lack ofbalance in the tariff is the result of regulatory intervention, the situation issometimes corrected by access deficit contributions or the like. Where it arisesfrom commercial decisions on the part of the incumbent operator, perhapsbased upon a reluctance to introduce rapid changes in the structure of the tariff,the incentive to build a competing local-loop is substantially reduced.


104

Local-loop unbundling could, at a stroke, solve for a competitor theproblem of acquiring local-loop assets without making a significant sunkinvestment. The decision whether to build a rival local-loop, whether to leasethe incumbent’s or whether to buy the local-loop and local switching from theincumbent in a bundled form depends upon the current and expected pattern ofpricing. It would not be difficult to generate a set of prices which would makeany of the three options preferred. In particular, making an unbundled local-loop available at exceptionally low prices—both current and expected—wouldhave a major impact on the extent of infrastructure competition. We explorethis crucial point below, after a further discussion of incentives to invest ininfrastructure.

2. INCENTIVES TO INVEST IN THE LOCAL-LOOP

The standard way of appraising an investment proposal is to apply the netpresent value (NPV) approach, that is, to project the constant revenuesassociated with the investment, discount them at a rate equal to the cost ofcapital and establish whether the resulting net present value is positive ornegative. However, as Dixit and Pindyck have pointed out, “the net presentvalue rule is based on some implicit assumptions that are often overlooked.Most important it assumes that either the investment is reversible, that is it cansomehow be undone and the expenditures recovered should market conditionsturn out to be worse than anticipated, or, if the investment is irreversible, it is anow or never proposition, that is, if the firm does not undertake the investmentnow, it will not be able to.”71

Yet many investments, including investments in the local-loop have neitherof these two characteristics. An investment in the local-loop is irreversible, inthe sense that much of the expenditure is not recoverable in the event that thecompany ceases to operate, and potential investors are not faced with now ornever propositions: they can adjust the start date and speed of construction ofthe loop to any desired level.

In these circumstances, a firm with an opportunity to invest possessessomething equivalent to a call option on a financial instrument. The holders ofa call option can exercise that option at a time of their choosing, but do not haveto do so (and obviously will not do so if the market price is less than the optionprice). But the observation that the market price exceeds the option price is notenough to trigger the use of the option. The holder may choose to wait untilfurther information arrives concerning the value of the underlying security.

Similarly, a potential new investor in the local-loop can predict fairlyaccurately the cost of making an investment, which is fairly constant from one


105

period to the next. Its value, however, is changing all the time as marketprojections change. By undertaking the investment the firm gives up thepossibility of waiting for new information to arrive. As Dixit and Pindyck putit, “this lost option value is an opportunity cost that must be included as part ofthe cost of the investment. As a result, the NPV rule ‘invest when the value ofa unit of capital is at least as large as its purchase and installation costs’ must bemodified. The value of the unit must exceed the purchase and installation cost,by an amount equal to the value of keeping the investment option alive.”72

This way of viewing the investment decision, from the standpoint of so-called “real option theory” is helpful in understanding investment in the local-loop. A firm estimate of the NPV of such an investment is constantly changingas new information arrives from the market and from the regulator. Amomentarily positive NPV is not enough by itself to trigger investment.Moreover, the availability of an option to lease capital equipment (the “buy”option) is an alternative which influences the NPV of the investment decision(the “make” option).

This new way of thinking about the investment decision applies to theincumbent too. Its investments are irreversible but it can modernize thenetwork at a time of its choosing. If entrants can immediately lease themodernized network that may well chill the incumbent’s incentives to invest.

Although Dixit and Pindyck and their successors have made considerableprogress in applying real option theory both to a monopoly and a perfectlycompetitive market, employing the technique to the case of an oligolopy haspresented hitherto insurmountable difficulties, derived from a combination ofthe smaller number of players and the stagastic nature of the events that unfoldover time. Firms in such circumstances will be torn between a desire to “waitand see” and a desire the gain first mover advantages. The approach can,however, be useful in evaluating in an informal way the effects of policytowards local-loop unbundling. In general, its availability at a reasonable priceis likely to chill investment in the local-loop because it provides a method ofbuying time, in the sense that an entrant can offer its services on the marketwithout incurring the cost of an irreversible investment. This is particularlyadvantageous if customers are indifferent between having service provided tothem by an entrant employing the incumbent’s unbundled loop or employing itsown loop. This way of conceptualizing the investment decision also opens upthe possibility of the regulator manipulating the option value by specificintervention. We return to the possibility below.


106

3. IS LOCAL LOOP UNBUNDLING REQUIRED UNDER EC LAW?

Many EU member states are considering local loop unbundling, and it istherefore timely to consider how it is likely to be viewed under ECtelecommunications regulation and competition law.

EC Telecommunications RegulationEC telecommunications law does not provide a definitive answer to the

question of whether unbundled access to the local loop should be granted.Viewed from the perspective of the type of access which is being granted by theincumbent, it would appear that the nature of such access is not“interconnection” within the meaning of the Interconnection Directive.73 Thebetter view would seem to be that the local loop cannot constitute a“transmission system,” although this is not defined in the Directive. TheEuropean Commission’s own view, set out in its Recommendation onInterconnection Pricing,74 is consistent with this analysis, that access to the localloop is not governed by the rules on interconnection.

On the other hand, it may be that access to the local loop falls within themeaning of “special network access,” attention to which is given in Article 16of the amended ONP Voice Telephony Directive.75 Article 16 provides thatreasonable requests for special network access should be “dealt with,” and thisobligation is only to be limited on a case by case basis, where two conditions arefulfilled: first, that technically and commercially viable alternatives to therequested access exist; and second, that the requested access is inappropriate inrelation to the resources available to meet the request. It is difficult to see wherejustification in respect of access to the local loop might arise under the secondlimb. In the case of the first limb, clearly, “traditional” local access is atechnically viable alternative, in terms of the objective sought by entrants, i.e.,to gain access subscribers; whether this proves to be commercially viable willdepend upon the pricing mechanism adopted by the local access provider, towhich the ordinary rules of analysis apply.

Competition LawAs is well known, Article 86 EC prohibits the abuse of a dominant position

by a company. Viewed in the context of this Article, the question is whether arefusal to supply unbundled access to the local loop could be regarded asanticompetitive. Whilst space does not permit a complete analysis of thisquestion, it is helpful to offer a number of observations. First, in general terms,the European Court of Justice has on a number of occasions found refusals tosupply anticompetitive. However, in the present context, the question is


107

perhaps more appropriately: is the provision of local interconnect (i.e., accessto the local switch and the local loop etc) to be regarded as a form ofanticompetitive tying. There is no clear answer to this question, but apreliminary indication may be obtained in the light of the fact that the secondarylegislation and the competition provisions are, in principle, to be appliedconsistently. With this in mind, it should be emphasized that the secondarylegislation, which imposes stricter obligations than does competition law, stopsshort of requiring unbundled local loop access. Therefore, although there mayin principle be an argument that refusing unbundled local loop access mayconstitute an abuse, this may require some solid justification.

4. THE COST AND PRICE OF LOCAL-LOOP UNBUNDLING

I have referred above to the key role of the price charged for an unbundledlocal-loop in determining take-up and hence in influencing the rate ofconstruction of the competitive local-loop. There are two generic approacheswhich might be employed toward pricing this particular unbundled service,whether within the framework of Competition Law or as part of a regulatoryprocess. The first is based upon cost, while the second has its starting point inthe incumbent’s retail tariff.

The question of how the prices of interconnection services should be relatedto costs has been discussed exhaustively. Our discussion is based on theassumption that an appropriate starting point is average long-run incrementalcost, supplemented by a mark-up to recover common costs between theconveyance network and the access network. Because the replacement cost ofassets in the access network is likely to exceed their historic cost (because ofrelatively low rates of technical progress there), current cost estimates are likelyto be higher than historic costs. This proposition is shown in BT’s financialaccounts.

The alternative approach is to base wholesale prices on the incumbent’sretail charges for similar services. This could be done by stripping out from theretail monthly payment costs associated with the provision of the retail servicealone, on the assumption that marketing and billing costs associated with awholesale client are sufficiently small to be ignored. The most persuasiveargument for basing interconnect charges on retail prices starts from theproposition that the regulator is seeking to introduce efficient competition incircumstances where tariffs are unbalanced. This argument is encounteredmost frequently in connection with entry into profitable activities such as long-distance or international calls. If the regulator has set tariffs for these servicesat a high level in order to cross-subsidize access, then a less efficient operator


108

than the incumbent may be able to cream-skim these markets. To prevent thatoccurring, it is argued that interconnection charges should be set at a level equalto the incremental cost of the service in question augmented by the “opportunitycost,” which can be defined as the profit foregone by the incumbent as a resultof its loss of long-distance or international business to its competitor. Providedthat the overall level of profits earned by the incumbent is controlled, and ifother conditions apply, it can be shown that this so-called efficient componentpricing rule (ECPR) will ensure that entry only occurs where the entrant is moreefficient in providing the service than the incumbent.76

In relation to local-loop unbundling, the equivalent argument could bemade that the service should be offered to entrants at the incumbent’s retailtariff, reduced by the incremental cost to the incumbent of providing the retailservice associated with the provision of a line. In such circumstances, entrantswill only come into the market if they can perform the retail functions moreefficiently than the incumbent.

If this argument is to be successfully applied, then two conditions must befulfilled. Firstly, it must be the case that the structure of retail prices isdetermined by the regulator, in pursuit of public policy objectives, rather thanby the operator, on commercial considerations. (Otherwise the interconnectionpricing regime effectively causes competitors to conform with the incumbent’spricing strategy.) Second, the principle must be applied to all interconnectionservices, both those used to provide profit-making and those used to provideloss-making services. In particular, if it applies only to loss-making services,and if the unbundled local-loop is loss-making at existing tariffs, then entrantsare able to purchase the unbundled local-loop at a subsidized price, and can useit as a basis to attack profitable markets, without having to contribute to the costof the local-loop subsidy. In other words, if retail prices are to be used as thestarting point for charging for any interconnection service, they should be thestarting point for charging for all such services, including cases where anentrant by-passes the incumbent’s facilities.

In summary, price is obviously a key element in determining the impact ofany policy of local-loop unbundling. The application of now quite widelyestablished principles of interconnection pricing implies that a long-runincremental cost basis, with mark-up, is appropriate. Where the tariff isunbalanced, a change starting from the incumbent’s retail price (without furtheradjustments) carries significant risks of encouraging inefficient entry.


109

5. REGULATORY POLICY TOWARDSLOCAL-LOOP UNBUNDLING

A telecommunications regulatory agency typically operates under sector-specific legislation which gives it different duties and usually greater powersthan apply under general Competition Law. They can, for example, use thesepowers to promote particular competitors or particular forms of competitiveentry. In the case of local-loop unbundling, the key instruments at their disposalare decisions about whether (or for how long) to mandate local-loop unbundlingand the determination of the price at which it should be made available.

The significance of the pricing issue was discussed in section 4 above. Iemphasized there that, unless the regulator is pursuing an explicit policy ofentry promotion, by tipping the playing field against the incumbent, regulatorypricing of interconnection services could reasonably be based upon the uniformapplication of one of two competing principles—long-run incremental cost(with mark-up) or charges based on the incumbent’s retail prices. The lattermay be appropriate where there is a strong regulatory interest in maintaining aparticular structure of retail prices. In the absence of such a motive, LRIC pricesare more appropriate. The key point, however, is that, if distortions tocompetition are to be avoided, the same approach should be applied to allinterconnection services, access or conveyance.

We now address the question of whether it is appropriate as a matter ofregulatory policy, rather than of competition law, to mandate local loopunbundling. This must depend on a variety of factors, including in particular theobjectives of policy pursued by particular governments. I noted above theEuropean Commission’s preference for neutrality between infrastructure andservice competition, but, in my opinion, the task of defining and enforcing suchneutrality during a period of transition to competition is in practice impossible.This leaves governments and regulators with considerable flexibility of how toproceed.

On our reading of international experience of the development ofcompetition in telecommunications, cases of significant infrastructuredevelopment have been observed, without local-loop unbundling. Thissuggests that such unbundling is not a necessary condition for competition todevelop. At the same time, experience of lineside unbundling is so restrictedthat it is impossible to say whether it is a complement or precursor to thedevelopment of infrastructure competition or whether it is a substitute for it.(Nor will the issue be resolved soon, as the time required to set up sucharrangements—let alone observe their effects—may be considerable.) In logic,a case can be made for either proposition. Local-loop unbundling might


110

provide an entrant with a relatively quick means of gaining direct access tocustomers, who then transfer to the entrants own infrastructure. Alternatively,if local loop unbundling is available at a low price, incentives to buildinfrastructure are eliminated. As in the whole of this discussion, prices are offundamental importance.

There is, however, a position intermediate between mandating and notmandating lineside interconnection - mandating it for a limited period. This isthe approach employed in Canada. It is sometimes represented as a means of“jump-starting” competition. The idea is that entrants acquire customers in thelocal-loop through leasing the incumbent’s facilities. This provides them witha sounder basis for their own investment positions in infrastructure. However,because local-loop unbundling is only mandated for a limited period, they haveto take the plunge or risk losing their business. In terms of analysis set out insection 2 above, this policy is designed to increase the NPV of the investmentwhen it occurs and to control its timing by raising the opportunity costs of delay.What is not clear, however, is the basis on which local-loop unbundling wouldbe priced in the interim period. This would require considerable judgement. Ifit were set too low, there is a risk of entry by short-run cream-skimmers whohave no intention of constructing their own local-loop. If it were set too high,the opportunity would not be taken up. A long-run incremental cost approach(with mark-up) might be most appropriate here, as it would confine entryassistance to bringing its timing forward, rather than offering an explicitsubsidy.

Questions remain, however, about whether this particular form ofintervention, which is geared at managing the transition to competition, isnecessary or desirable. Evidence from other countries suggests that it is notnecessary. The question of its desirability is tied up with the general issue ofentry promotion. In my judgement, now that entrants have establishedthemselves in a number of countries and typically are supported by incumbentsin other major markets, the burden of proof rests with those who favor explicitentry-promoting measures.

There will also be effects on the incumbent’s investment decisions. Inparticular, the incumbent may choose to defer network modernization, so thatit comes into effect only after the period of mandatory unbundling has expired.This possible or likely detriment should also be placed in the balance when thepolicy decision is made.


111

6. CONCLUSIONS

This paper has reviewed a number of questions relating to local-loopunbundling, in particular, whether a dominant incumbent should be required tounbundle its local-loop, and if so, at what price. Inevitably, these issues arenested within a broader consideration of government policy towardsinfrastructure or service competition.

At present there is insufficient evidence to judge whether economic andsocial welfare is better served by infrastructure competition or servicecompetition. However, infrastructure competition has established itself in anumber of countries without the availability of local loop unbundling.

Moreover, because of the complexity of regulatory interventions intelecommunications markets, which embrace decisions about the level andstructure of retail prices and the charges for and availability of wholesaleservices, it is impossible to define an entirely neutral approach towardsinfrastructure and service competition, and for this reason Europeangovernments have considerable and legitimate scope in deciding what policiesto adopt.

No conclusive case can be made for the proposition that local loopunbundling is mandatory under EU Competition Law, and no such case hasbeen brought by the Commission. It is impossible to divorce the question ofmandatory unbundling of the local loop from the price which is charged for it,but as with other interconnection services, consistent cases can be made eitherfor pricing in relation to long-run incremental costs (to provide efficient entrysignals) or in relation to the structure of retail prices (but only where these arefixed by regulators to meet social objectives). In order to avoid inefficiententry, a consistent approach must be taken to pricing access and conveyanceservices.

Account should also be taken of the likely effect of local loop unbundlingon discouraging the incumbent from investment in network modernization. Aregulatory body with a remit explicitly to promote entry can use local-loopunbundling available at a low price to stimulate entry. Given the pattern ofentry at present, such entry assistance is unnecessary or should in any case beshort-lived.

A temporary period of mandatory local-loop unbundling might bringcompetition forward, rather than distort its form over the long term. Such apolicy does, however, carry with it the risks of promoting inefficient entry, andof (temporarily) reducing the incumbent’s investment incentives.


112

In summary, a convincing regulatory case for local loop unbundling has notbeen made. But if it is adopted, it should be done at prices which do not distortthe competitive process—i.e., normally LRIC prices with mark-up.


113

UNIVERSAL SERVICE OBLIGATIONS: COMPARISON OFTHE UNITED STATES WITH THE EUROPEAN UNION

Barbara A. Cherry

Abstract. The United States (U.S.) and the European Union (EU) areboth pursuing joint policies of liberalization of telecommunicationsservice markets and universal service. As a matter of law, there aremany similarities in their respective policies, such as the elimination ofmonopolies and the establishment of explicit universal service fundingmechanisms. However, cultural and political differences as well asother dissimilar regulatory rules may yield differences inimplementation. There are also some substantial differences in statedpolicy. For example, the U.S. views affordability of services in terms ofuniformity of rates, whereas the EU approach is to mitigate the effectsof mandatory rate rebalancing towards costs. In addition, unlike theEU, the U.S. mandates special rates for special groups of customers,namely, certain schools, libraries and health care providers. This latterrequirement alone has resulted in creation of a $2.65 billion annualfund by the Federal Communications Commission, which is to befunded by contributions from telecommunications service providers.These important differences may likely provide telecommunicationsproviders in the EU with competitive advantages over U.S. carriers inthe international telecommunications service marketplace.

1. INTRODUCTION

Both the United States (U.S.) and the European Union (EU) have beenpursuing policies of liberalization of telecommunications service markets whileproviding for universal service. For the U.S., these policies are codified in theTelecommunications Act of 1996 (TA96) [1]. For the EU, these policies areembodied in numerous directives, such as the Voice Telephony Directive [2],the Full Competition Directive [3], and the recently adopted Interconnectionand Universal Service Directive (“IC/US Directive”) [4].

In TA96, Congress enacted legislation setting forth the regulatoryframework for incumbent providers and new entrants in a competitiveenvironment, with specific provisions on universal service. Consistent with thedual jurisdictional nature of regulation set forth in the Communications Act of1934 [5], the new law continues to provide for regulation of interstate servicesby the Federal Communications Commission (FCC) and regulation of intrastate


114

services by state regulatory commissions. In particular, section 254 of TA96provides specific policies for the provision of universal service. Some aspectsof universal service are to be implemented by the FCC, pursuant to which theFCC recently issued its Universal Service Order [6]. Other aspects are left toimplementation by the States, subject to consistency with the Act and certainpreemption powers granted to the FCC under section 253(d).

Similarly, the EU’s directives establish policies for harmonization of theregulatory environment and standards for services and technologies among itsmember states. The member states have the authority to implement suchpolicies on a national level, subject to consistency with the law of the EuropeanCommunity and the oversight authority of the European Commission (ECOM)as to implementation of numerous requirements by the national regulatoryauthorities.

There are many similarities in the policy provisions of TA96 and the FCC’sUniversal Service Order, on the one hand, and of the EU Directives and ECOMCommunications [7] [8], on the other. Their provisions contain strikinglysimilar approaches by government to ensure universal service to customers,such as the elimination of monopolies, the establishment of explicit fundingmechanisms for net costs arising from universal service obligations, and thepossibility of multiple providers bearing universal service obligations.

However, there are some important differences in policies between the U.S.and EU. These differences relate primarily to the relationship of raterebalancing to the concept of affordability, as well as the greater scope infunding to be provided for certain educational institutions and health careproviders in the U.S.

A comparison of universal service policies between the U.S. and EUprovides valuable insights as to the likelihood of success among the differentnations in achieving the goals of universal service. Given my greater familiaritywith the implementation process of universal service policy in the U.S., acomparison of policies reveals problems, which have arisen in the U.S., that Ibelieve are less likely to be encountered in the EU. However, it also revealsproblems that may arise in the EU if the experience of the U.S. is not taken intoaccount. The purpose of this paper is to describe many of the similarities anddifferences in universal service policy between the U.S. and EU, and to exploresome of the ramifications flowing from them. It is hoped that this may triggerfurther interest in sharing the successes and failures in policies among nationsand in broadening the research endeavors of policy experts.


115

2. SIMILARITIES IN UNIVERSAL SERVICE POLICIES

The similarities in universal service policies between the U.S. and the EUrelate to some basic principles embedded in a regulatory framework oftelecommunications services. These principles include the elimination ofmonopolies, a fairly narrow initial definition of universal service for generalresidential and business customers, the notion that the definition of universalservice is an evolving one, the possibility of multiple carriers bearing universalservice obligations in a given serving area, the lack of a presumption as to whichcarriers may bear universal service obligations, and the establishment ofexplicit funds to support universal service objectives.

However, these similarities in policies may not necessarily lead tosimilarities in implementation of policy nor of performance within thetelecommunications industry. For example, in the long run, cultural differencesmay lead to dissimilar evolving definitions of universal service. In the shorterrun, the interlata entry barrier to Regional Bell Operating Companies in the U.S.will likely lead to more strategic behavior and creamskimming activities bynew entrants in the U.S. compared to new entrants in EU member states.Furthermore, the combinatorial effects of other Federal and State regulatoryrules with the collection mechanisms for universal service may pose greatercompetitive non-neutrality problems in the U.S.

2.1. Elimination of MonopoliesIn both the U.S. and the EU, monopolies for the provision of

telecommunications services are required to be eliminated. Under section253(a) of TA96, no state or local government statute, regulation or legalrequirement “may prohibit or have the effect of prohibiting the ability of anyentity to provide any interstate or intrastate telecommunications service.” Tothe extent that a state or local government action violates section 253(a), undersection 254(d) the FCC shall preempt its enforcement to the extent that it isnecessary to correct the violation.

In the EU, Article 2 of the Full Competition Directive requires memberstates to withdraw all measures granting exclusive rights for the provision oftelecommunications services by January 1, 1998. However, the Directiveleaves open the possibility of additional implementation periods for fullliberalization for member states with less-developed or small networks. In thisregard, additional periods to implement full liberalization have been granted toGreece, Ireland, Luxembourg, Portugal, and Spain [9].


116

2.2. Definition of Universal ServiceYet, reliance on more competitive markets to provide telecommunications

services may impede the ability of all individuals to have access to the publicswitched telecommunications network and the use of certain basictelecommunications services. Under section 254(c)(1), Congress states that theFCC shall establish, based on recommendations from a Federal-State JointBoard, the definition of services to be supported by federal universal servicesupport mechanisms. It also provides that universal service is an evolving levelof telecommunications services that the FCC shall establish periodically, givencertain enumerated criteria.77

In its Universal Service Order, the FCC defined core services to receiveuniversal service support with respect to residential and business users. Thecore services are: single party service; voice grade access to the public switchednetwork; dual tone multifrequency signaling (DTMF) (i.e., touch tone); accessto emergency services; access to operator services; access to interexchangeservices; access to directory assistance; and toll limitation services forqualifying low-income customers (par. 56). Universal service support is also tobe provided for some amount of local usage, although at a level of usage yet tobe determined (pars. 65-70). With respect to low income individuals, theUniversal Service Order also provides for expansion of existing Lifeline andLinkup programs (pars. 346-382), the availability of toll limitation services(par. 385), prohibitions on disconnection of low service for nonpayment of tollcharges (par. 390), and restrictions on use of service deposits (par. 398). Specialneeds of disabled individuals are handled in separate orders and proceedings.

In the Voice Telephony Directive, the member states of the EU are requiredto ensure that end users can obtain upon request a connection to the fixed publictelephone network (Article 3(a)). Furthermore, national regulatory authoritiesare to ensure, among other things, that DTMF, direct dialing-in, callforwarding, and calling-line identification are provided where technicallyfeasible and economically viable (Article 9); that special tariffs may beavailable for emergency services, low usage users, or specific social groups(Article 14); that targets are set for the provision of itemized billing (Article 15);that directories of subscribers to voice telephony services are made available tousers in either printed or electronic form, with public directory informationavailable upon request (Article 16); and that public pay telephones be providedto meet the reasonable needs of users in terms of both numbers and geographicalcoverage (Article 17). Furthermore, member states are authorized to takemeasures to ensure that disruption of service is confined as narrowly asnecessary for nonpayment of bills (Article 23).


117

The recent IC/US Directive expands upon the Voice Telephony Directiveby expressly defining the fixed public telephone network to support not onlyvoice telephony but also certain facsimile communications and voice band datatransmission via modems at a rate of at least 2400 bits/second (Annex I, Part 1).The fixed public telephone network is also to provide to end users access toemergency services and the ability to originate and receive national andinternational calls, as well as the provision of operator assistance, directoryservices, public pay phones, and special terms or facilities for customers withdisabilities or special social needs (Annex I, Part 1). Furthermore, in itspreamble, the European Parliament and Council state that the concept ofuniversal service must evolve to keep pace with advances in technology, marketdevelopment, and changes in user demand (par. (7)).

From the preceding description, it is clear that, as a starting point forresidential and business end users, both the U.S. and EU are still primarilyconcerned with the availability of voice telephony; although both jurisdictionsinclude the requirement of access to certain technical functionalities—such asDTMF (touch tone services)—that make other uses of the network, such asfacsimile and modems, possible. In addition, both require access to emergencyservices as well as services related to access to information, such as operatorassistance and directory services. Special needs of certain groups, such as lowincome users and the disabled, are also addressed. However, as will bediscussed in section 3.1, the overall regulatory framework within the EUpermits a more targeted approach in setting rates for special groups.

Yet, in the U.S. a controversy yet to be resolved is the amount of local usageto be included in the definition of universal service. This issue is to bedetermined at a later time by the FCC, for federal support purposes, and by theState commissions for state support, if any. This issue highlights conflictbetween those pushing for a more expansive definition of universal service andthose desiring to keep universal service funding contained. Given that the EUhas not specified local usage as a component of universal service, this issue isless likely to arise, as national universal service funding schemes can not beused to fund costs incurred outside the scope of universal service (see [8], p. 6).

As to the future, both the U.S. and the EU contemplate the evolution of themeaning of universal service as technology, customer demands, and societalneeds change. But, at this time, neither requires the availability of moreadvanced services requiring broad bandwidth to general residential andbusiness end users. However, it is unclear whether the expansion of the meaningof universal service will follow similar paths in the U.S. and throughout the EU.The development of a truly global economy would encourage similardevelopments. But, cultural differences may prove significant impediments to


118

common social goals. An example of such a difference is discussed in section3.2 of this paper with respect to universal service to educational institutions andhealth care providers.

2.3. Imposition of Universal Service Obligations on ProvidersThere are some basic elements of imposing universal service obligations

(USOs) on providers of telecommunications network facilities and/or servicesin the U.S. and the EU that are essentially the same. In both jurisdictions, morethan one provider may bear USOs and be eligible for funding. Furthermore, inneither jurisdiction is there a presumption that the incumbent local exchangeproviders (ILECs) are to bear USOs.

In the U.S., carriers that bear USOs and are eligible for universal servicesupport are called eligible carriers. In particular, section 214(e)(1) defines aneligible carrier as one that, throughout the service area for which the designationis received, offers the services supported by federal universal service supportmechanisms (either using its own facilities or a combination of its own facilitiesand resale of another carrier’s services) and advertises the availability of andcharges for such services using media of general distribution. Statecommissions are to designate eligible carriers, satisfying this definition, for aservice area; more than one eligible carrier shall be designated for non-ruralareas and may be designated for rural areas (section 214(e)(2)). However,under section 214 there is no presumption that any particular carrier, such as theILEC, shall be an eligible carrier.

In the EU, the IC/US Directive provides that national regulatory authoritiesare to notify to the European Commission by January 31, 1997, the names ofthose telecommunications organizations which have USOs (Article 18). Morethan one organization in a member state may bear USOs (par. (8) of preamble).However, according to the European Commission [8], the regulatoryframework in the EU also provides no presumption that the incumbent(formerly monopoly) provider bears USOs (Annex D). Furthermore, similar tothe definition of an eligible carrier in the U.S., USOs are to be imposed throughconsideration of geographic areas and service categories (see [8], Annex D).

Despite these similarities in the regulatory frameworks for assigning USOs,there are major historical differences in the provision of telecommunicationsservices between the U.S. and throughout the EU which give rise to dissimilarpolitical issues. These differences in political forces may produce differentoutcomes as to bearers of USOs.

For example, in the U.S., the Bell Operating Companies may not provideinterlata interexchange services in areas where they provide local service untilthey convince the FCC that they satisfy the requirements contained in section


119

271(C) of TA96. There is heated debate among industry players as well asregulators as to whether any of the Bell Companies have met such conditions.As to applications for interlata entry that have been filed before the FCC,interexchange carriers claim that certain items of the competitive checklist insection 271(c)(2)(B) have not yet been met. On the other hand, Bell Companiesclaim that interexchange carriers have strategic incentives to slowly enter localexchange markets and to greatly exaggerate problems with interconnection andoperating system interfaces.

In essence, the controversy on interlata entry has simply shifted from thelegislative arena to the regulatory one. It appears that the FCC will not allow in-region interlata entry by any Bell Company for some time to come, and thatlocal exchange entry by interexchange carriers will likely continue to be slowto better ensure no interlata relief. So long as this is the case, the present ubiquityof ILEC facilities, particularly of Bell Companies, in their service areas willlikely require designation of ILECs as the providers bearing USOs. Newentrants will then have the ability to choose those areas in which they will seekto bear USOs and be eligible for universal service funding. These choices ofwhere to bear USOs may be motivated, in significant part, by creamskimmingopportunities provided through arbitrage of averaged prices for interconnectionand unbundled network elements and the deaveraged costs of servingcustomers within wire centers.

In member states of the EU, there is also the history of incumbency bymonopolists. However, monopolies were usually in the form ofgovernmentally-owned or controlled entities compared to privately-ownedentities in the U.S. Therefore, member states have been required to take stepsthat the U.S. has not, such as privatizing telecommunications providers andestablishing national regulatory authorities. As privatization and implementa-tion of national policies through national regulatory authorities occur, theforces contributing to the development of local competition may be greater atleast in some member states of the EU. First, there is no EU edict, comparableto the interlata restriction in the U.S., barring entry by incumbent providers toa lucrative telecommunications market and inducing the associated strategicbehavior of industry members. Given the U.S. experience, it would be wise formember states to also refrain from erecting any such artificial entry barriers (tothe extent that it would even be permissible under European Community law).Second, local telephony competition by cable companies is already greater insome of the member states, such as the UK. This is due not only to somedifferences in regulation of cable service, but to the fact that many cablesystems in the EU are being built at a later time than in the U.S. and specificallywith the intent to provide both cable and telephony services. Third, by


120

mandating rate rebalancing towards costs, as discussed in section 3.1 of thispaper, the EU provides greater protections against creamskimming by newentrants. Therefore, throughout the EU, there may be: (1) less incentive forstrategic behavior by new entrants both to burden ILECs with USOs and yet toaccept USOs when access to universal service funding providescreamskimming advantages; and (2) a greater number of local telephonyproviders able to bear USOs in certain geographic areas.

2.4. Explicit Funding for Universal Service ObligationsBoth TA96 and the EU Directives provide for the establishment of explicit

mechanisms to support universal service objectives, with funding based oncontributions from telecommunications providers. The scope of entitiesrequired to make contributions is also similar. Furthermore, both jurisdictionsare concerned with nondiscriminatory and equitable assessment of contributionburdens among providers.

Under TA96, universal service principles established by Congress include“specific, predictable and sufficient Federal and State mechanisms to preserveand advance universal service” (section 254(b)(5)). Such mechanisms are to befunded by contributions from telecommunications carriers—of interstateservices for Federal mechanisms and of intrastate services for Statemechanisms—on an equitable and nondiscriminatory basis (sections 254(d)and (e)).

In similar fashion, the Full Competition Directive (Article 4c) and the IC/US Directive (Article 5) authorize member states to establish national schemesto share the net cost of the provision of USOs among organizations operatingpublic telecommunications networks and/or publicly available voice telephonyservices. Contributions may be required as a part of an independent universalservice fund or as unbundled, supplementary charges added to interconnectioncharges. Furthermore, in setting the amount of contributions, member states areto take into account the principles of transparency, nondiscrimination andproportionality.

2.4.1. Providers Which Pay ContributionsIn applying section 254(d), the FCC decided in its Universal Service Order

that the definition of a carrier required to make contributions to a Federal fundbe construed as broadly as possible, including wireless and wireline providers(pars. 779-791). However, certain providers are to be excluded, such as Internetand enhanced service providers, because they are currently considered to not betelecommunications carriers under federal law (par. 788), and private networkoperators which do not lease or sell excess capacity (par. 794).


121

Likewise, in its Communication on assessment criteria for nationalschemes for costing and financing universal service [8], the ECOM states thatorganizations not liable to make contributions include private networkoperators offering corporate networking or closed user group services,providers of value-added data services, and providers of enhanced voicetelephony services (guideline 15). In addition, Internet access providers can notbe required to make contributions because voice telephony service via theInternet is not currently viewed as a publicly available voice telephony service(footnote 19). Yet, since universal service can be provided by wireline orwireless technology to a fixed location ([8], section 2.1; [4], Annex I, Part I), itappears that both wireline and wireless providers would also be required toprovide contributions.

In the U.S., a number of parties have claimed that exclusion of Internet andenhanced service providers from making universal service contributions is notcompetitively neutral, and that their exclusion will be more problematic ascompeting services, such as Internet telephony, become more widespread. Theseverity of the competitive advantage given to excluded entities will depend, ofcourse, on the degree to which excluded entities’ offerings becomesubstitutable services. The same structural problem is embedded in theframework of the EU. To the extent that excluded entities’ services becomesubstitutes for those of included entities, a cost advantage will be conferred onexcluded entities. In this way, a source of economic inefficiency will beintroduced, and it will be more difficult to determine whether the diffusion ofexcluded entities’ services is due to innovation, changes in customer needs, orsimply an artificial price advantage.

2.4.2. Differential Abilities to Pass Through Costs of ContributionsBut, a greater problem for both the U.S. and the member states of the EU is

posed by the likelihood that the various providers required to makecontributions will have differential abilities to pass through the cost of suchcontributions to customers. Such differential abilities pose not only competitiveneutrality problems, but also potentially threaten the financial viability ofincumbent providers in certain geographic areas and thereby the continuedavailability of universal service in such areas.

In its Universal Service Order, the FCC requires contributions to beassessed against end user revenues of telecommunications services (par. 844).This revenue base was chosen in order to avoid double tax problems, as wouldoccur with an assessment against gross revenues, and distortions in theincentives of new entrants to build their own facilities or purchase services forresale rather than buy access from incumbent providers (pars. 843-850).


122

However, the coexistence of other Federal and State regulatory requirementsreintroduce the distortionary effects meant to be avoided by choice of a revenuebase of end-user revenues.

First, the Universal Service Order itself prohibits ILECs from adjustingtheir rates for unbundled network elements to account for the universal servicecontributions (par. 851). Furthermore, ILECs are permitted to pass throughtheir contributions to the Federal fund only onto interstate—which areprimarily access—services. As a result, competitors will be induced to purchaseILEC unbundled elements, build their own facilities, or purchase services byresale, rather than purchase access services from ILECs. In this way, inactuality it will be more difficult for ILECs than new entrants to passcontributions through to customers rather than their shareholders.

Second, many States have price caps or rate freezes which have the effectof prohibiting the pass through of ILECs’, but not new entrants’, contributionsto customers. These rate restrictions also create a higher burden on theshareholders of ILECs compared to the shareholders of their competitors tofinance the universal service contributions. To ensure that providers have thesame ability to pass through their contributions in rates to customers, the FCCshould have provided—although it declined to do so (par. 853)—an explicitmechanism to pass through the contributions, such as an end user surcharge.

In the EU, the ECOM has also applied the principles of nondiscriminationand proportionality to require that member states implement national schemesfor universal service support through a collection mechanism that prevents“double contributions,” as would occur with contributions based on grossrevenues ([8], guideline 17.4). A number of indicators for determining theapportionment of contributions among providers might be used—such asrevenues before tax, call minutes, number of subscribers, and overall profit—aslong as criteria which unduly distort patterns of market entry, activity orinvestment by market players are avoided (guideline 17.3).

As illustrated by the above scenario in the U.S., member states should alsoavoid a regulatory framework which would yield differential abilities amongproviders to pass through contributions in rates and the associated differentialburdens of shareholders of such providers to finance the contributions. Suchdifferential impacts may arise from the choice of criteria for apportionmentamong providers, or from the combinatorial effect of multiple, even seeminglyunrelated, regulatory requirements. It deserves reiteration that the potentialeffects of such differential impacts create not only inequitable anddiscriminatory effects among providers and associated distortions to economicefficiency, but may also threaten the long term availability of universal servicein some geographical areas.


123

3. DIFFERENCES IN UNIVERSAL SERVICE POLICIES

Amid numerous similarities in universal service policies, the U.S. and EUhave some important differences. A few will be discussed here. First, the U.S.and EU have differing concepts of how rate rebalancing relates to affordabilityof universal service to end users. The U.S. views affordability as requiringuniformity of rates among classes of customers, with particular emphasis oncomparable rates between rural and urban areas. On the other hand, the EUviews affordability as mitigating the impacts of mandated rate rebalancingtowards costs.

Second, the U.S. requires special rates, as a matter of law, for some specialcategories of customers, consisting of schools, libraries and health careproviders. This has resulted in a substantial funding burden of $2.65 billion peryear which is being placed on domestic telecommunications carriers. In the EU,there is no comparable universal service burden. Both of these differences inuniversal service policy will likely lead to competitive advantages for carriersin the EU as compared to those in the U.S.

3.1. Affordability and Rate RebalancingPerhaps the most important difference in universal service policy between

the U.S. and the EU is reflected in their respective approaches to the relationshipof rate rebalancing to the affordability of universal service to end users. This isevident from a comparison of provisions of TA96 with those of EU Directives.

In its Full Competition Directive, the EU specifically directs the memberstates to allow telecommunications organizations to rebalance their ratestowards costs (Article 4c):

Member states shall allow their telecommunications organizationsto re-balance tariffs taking account of specific market conditions and ofthe need to ensure the affordability of a universal service, and, inparticular, member states shall allow them to adapt current rates whichare not in line with costs and which increase the burden of universalservice provision, in order to achieve tariffs based on real costs. Wheresuch rebalancing cannot be completed before January 1, 1998 themember states concerned shall report to the Commission on the futurephasing out of the remaining tariff imbalances. This shall include adetailed timetable for implementation.

Although member states may create an explicit fund to recover the net cost ofUSOs according to this Directive, such a fund may not be used to recover access


124

deficit contributions attributable to unbalanced national tariff structures nor anyother costs associated with activities outside the scope of universal service ([8],p. 6). To address such other costs, including any concerns of bypass orcreamskimming by new entrants because incumbent providers may not haveyet completed rebalancing of tariffs, member states must establish mechanismsseparate from national schemes for funding universal service ([8], Annexes B& D). This overall approach was reaffirmed by the EU in its IC/US Directive,which provides the principles for calculating the net costs of USOs (Article 5)and states that such a calculation “should not hinder the on-going process oftariff rebalancing” (par. 8).

In recognition that such rate rebalancing may make certain telephoneservice less affordable in the short term, member states are permitted under theFull Competition Directive, consistent with the principle of subsidiarity, toadopt special provisions to soften the impact of rebalancing (par. 21). Thisincludes the ability of the member states to maintain uniformity of nationalprices or to allow deaveraging, providing that the overall affordability of theuniversal service is not called into question ([8], Annex D). This aspect ofpolicy was left unchanged by the new IC/US Directive. Thus, other than therequirement to rate rebalance, and other requirements related to interconnectioncharges, pricing decisions are left to the member states.

This approach is in stark contrast to that in the United States. In TA96, thereis no mandate, as in the EU, that the FCC or the States rebalance end-user ratestowards costs.78 On the contrary, rather than leave the issue to the FCC or theStates, Congress mandates several forms of rate averaging in the name ofuniversal service.

For example, comparability of rates is generally required between rural andurban rates for all telecommunications and information services in section254(b)(3):

Consumers in all regions of the Nation, including low-incomeconsumers and those in rural, insular and high cost areas, should haveaccess to telecommunications and information services, includinginterexchange services and advanced telecommunications andinformation services, that are reasonably comparable to those servicesprovided in urban areas and that are available at rates that arereasonably comparable to rates charged for similar services inurban areas.

A similar requirement is imposed specifically on providers of interexchangeservices in section 254(g):


125

[T]he Commission [FCC] shall adopt rules to require that the ratescharged by providers of interexchange telecommunications services tosubscribers in rural and high cost areas shall be no higher than the ratescharged by each such provider to its subscribers in urban areas.

In addition, section 254(g) requires the FCC to adopt rules to “require thata provider of interstate interexchange telecommunications services shallprovide such services to its subscribers in each State at rates no higher than therates charged to its subscribers in any other State.” Thus, the same rates mustalso be provided for interstate interexchange services across states.

These requirements, for comparability (in essence, averaging) of rates toconsumers with no reference to costs, are an integral part of Congress’ vision ofuniversal service. Implementation of the universal service provisions of section254 by the FCC and the states must necessarily comply with these requirements.Thus, universal service mechanisms—including contributions by carriers tofund federal and state universal service funds authorized under subsections (d)and (f), respectively—necessarily include the costs of not being able todeaverage or rebalance rates towards costs among end users. Such inclusion, aspreviously discussed, is prohibited by the member states of the EU.

These differing approaches highlight a fundamental difference in theconcept of affordability between the U.S. and the EU. In the U.S., affordabilityis viewed, in large part, by uniformity of rates among classes of customers. Inthe EU, affordability is viewed as mitigating the effects of rate rebalancingthrough targeted and special tariff schemes. Although rate averaging may be anapproach used by a member state, universal service funds may not be used formaintenance of such a tariff structure.

In the long run, the EU’s concept of affordability may better position itsmember states, as compared to the U.S., to participate in a global economy. Thisis because, as competitive pressures grow in the telecommunicationsmarketplace, rate structures in EU countries will be more cost-based due to raterebalancing and less encumbered by artificial pricing restrictions as a matter oflaw. Policy makers in the U.S. should take note of the rate rebalancing policy ofthe EU, its compatibility with achieving universal service, and the internationalcompetitive advantage it is likely to confer.

3.2. Affordability for Special Categories of UsersArguably, the main focus of affordability discussed in section 3.1 relates to

general concerns with residential and business consumers. In the U.S., unlikethe EU, notions of affordability are also incorporated in legislative mandates forspecial telecommunications service prices to special categories of end users. In


126

particular, as part of universal service, Congress specifies pricing requirementsfor certain educational institutions and libraries as well as health care providers.

Section 254(h)(1)(B) of TA96 requires telecommunications carriers toprovide services within the definition of universal service—which, undersection 254(c)(3), may include additional services to those for other end users—to elementary schools, secondary schools and libraries at rates less than theamounts charged for similar services to other parties. In its recently issueduniversal service rules, the FCC has defined such services to include allcommercially available telecommunications services (section 54.502). Carriersare to receive reimbursement for providing such discounts from universalservice support mechanisms (section 254(h)(1)(B)(ii)).

Section 254(h)(1)(A) pertains to services for health care providers andprovides in relevant part:

A telecommunications carrier shall, upon receiving a bona fide request,provide telecommunications services which are necessary for theprovision of health care services in a state, including instructionrelating to such services, to any public or nonprofit health care providerthat serves persons who reside in rural areas in that state at rates that arereasonably comparable to rates charged for similar services in urbanareas in that state.

This is yet another instance in which Congress mandated comparable ratesbetween rural and urban areas. In this section, Congress further specifies that anamount, equal to the difference in rates provided to health care providers forrural areas and the rates for similar services provided to other customers incomparable rural areas in that state, should be treated as a part of universalservice obligations and mechanisms.

In implementing these provisions, the FCC has ordered support foreducational institutions and libraries in an amount of $2.25 billion per year ([6],par. 425). For provision of service to health care providers, an annual cap of$400 million has been set (par. 608). Thus, for these two groups alone, federaluniversal service funding of $2.65 billion per year is required.

The EU has provided no comparable mandate for universal service tospecific categories of consumers, such as educational institutions, libraries orhealth care providers. Instead, any special tariff schemes for targeted groups isleft to the policy making of member states. Furthermore, as discussed earlier,any special schemes that are beyond the scope of universal service as defined inthe EU Directives may not be funded by member states through the explicituniversal service funding mechanisms that were initially authorized to be


127

established under the Full Competition Directive. Rather, member states willhave to develop some other means of implementing such a policy, such as bydirect funding by national governments. In this way, the EU is avoidingimposition of the enormous price tag to support special service rates toeducational institutions, libraries and health care providers, that has beencreated in the U.S.

The U.S. policy of funding these special categories of users—as withfailure to rate rebalance towards costs—is also likely to confer someinternational competitive disadvantage to U.S. telecommunications carriers.Since U.S. carriers must pay contributions to support services at special rates toschools, libraries and health care providers, such contributions constitute anincreased cost of doing business. These increased costs will pose a competitivedisadvantage to U.S. carriers in the market for international services, because,under the Universal Service Order, the international revenues of only domesticcarriers will be included in the revenue base against which contributions areassessed (par. 779). Thus, with competition, U.S. carriers will not be able tosimply pass through these costs to international customers.

4. CONCLUSION

The U.S. and EU are both pursuing policies of liberalization oftelecommunications service markets and universal service. These policies arereflected in various legislative measures—in the Telecommunications Act of1996 in the U.S. and in several Directives of the EU—with implementation bydesignated regulatory agencies. There are significant similarities as well asdifferences in the policies of these respective jurisdictions.

Similarities include elimination of monopolies; initial definitions ofuniversal service for general residential and business customers based on voicegrade telephony and certain capabilities for facsimile and modems; the abilityfor multiple providers to bear universal service obligations; and no presumptionthat universal service obligations are placed on incumbent providers. Despitethese similarities, there may be differences in implementation. For example,cultural differences may lead to differing, evolving definitions of universalservice for customers. In addition, political differences reflected in policy, suchas the interlata entry barrier placed on Bell Operating Companies, may lead togreater strategic behavior and creamskimming activities by new entrants in theU.S. In the long run, a higher incidence of creamskimming activities in the U.S.may lead to greater difficulties for incumbent providers in the U.S. to meet theiruniversal service obligations in some geographic areas.


128

Similarities in policies between the U.S. and EU also include theestablishment of explicit funding mechanisms to support universal service, withcontributions to be paid by telecommunications service providers and to beimposed in an equitable and nondiscriminatory manner. Both the U.S. and EU,however, will face challenges in actually assessing such contributions in acompetitively neutral manner. This is due to the exclusion of providers ofsubstitutable services, such as Internet and enhanced service providers, frommaking contributions, and the differential abilities among providers to passthrough the costs of contributions to customers.

There are also some important differences in universal service policies. Thedifferences discussed in this paper relate to concepts of affordability. Forresidential and business customers in general, the U.S. policy of affordability isbased, in large part, on uniformity of rates among classes of customers.Particular emphasis is placed on comparability of rates between rural and urbanareas, as well as uniformity of interexchange rates across the states. By contrast,the EU mandates rate rebalancing toward costs for end users, and viewsaffordability as the need to mitigate the effects of this rate rebalancing. Eachmember state is provided the authority to determine the degree of uniformity ofrates among classes of customers, although failure to rate rebalance is not to befunded from universal service funds. Thus, the EU essentially creates apresumption of rate rebalancing towards costs, with deviations permitted foraffordability purposes, whereas the U.S. specifically precludes such raterebalancing in favor of uniformity of rates among various groups of customers.

The U.S. concept of affordable rates is also reflected in special statutoryprovisions under TA96 with regard to rates to be provided to special categoriesof customers. More specifically, discounted rates are to be provided to certainschools and libraries, and urban rates are to be made available for health careproviders serving rural areas. The burden of these requirements has been set inthe amount of $2.65 billion per year by the FCC. The EU Directives have nocomparable requirements.

These differences in approaches to affordability may lead to competitiveadvantages in the international marketplace for telecommunications providersfrom the EU as compared to providers from the U.S. This is due to the greaterflexibility afforded to member states of the EU to require cost-based rates byproviders, as well as the greater financial burden to be borne by U.S. providersfor provision of special rates to schools, libraries and health care providers.

Overall, the differences in the regulatory frameworks for providinguniversal service in a more competitive telecommunications environment willlikely confer competitive advantages for telecommunications providers in theEU, as well as greater success for achieving the underlying universal service


129

goals. The probability of such advantages materializing increase if the memberstates avoid the mistakes made in the U.S., such as the creation of doubletaxation problems with respect to imposition of universal service contributionson providers. In order to mitigate the advantages naturally occurring in the EU’sregulatory framework, policy makers in the U.S. should consider adoptingsome elements of the EU approach, such as the EU’s view of the relationshipbetween rate rebalancing and affordability of rates.

REFERENCES

[1] Telecommunications Act of 1996, Pub. L. No. 104-104, 47 U.S.C. Sec. 151 et seq.(West Supp. 1997).[2] Directive 95/62/EC, (March 12, 1995).[3] Directive 96/19/EC, (March 13, 1996).[4] Directive 97/ /EC, (June 11, 1997).[5] Communications Act of 1934, 47 U.S.C. sec. 151 et seq. (West 1991 & Supp. 1995).[6] In the Matter of Federal-State Joint Board on Universal Service, FCC Report andOrder, CC Docket No. 96-45 (May 8, 1997).[7] European Commission Communication, Universal Service for Telecommunicationsin the Perspective of a Fully Liberalised Environment, COM(96) 73 (March 13, 1996).[8] European Commission Communication, Assessment Criteria for National Schemesfor the Costing and Financing of Universal Service in Telecommunications andGuidelines for the Member States on Operation of Such Schemes, COM(96) 608(November 11, 1996).


130

RECENT DEVELOPMENTS IN THE REGULATION OF INTERNA-TIONAL TELECOMMUNICATIONS

Dr. Andrea Huber

INTRODUCTION

The year 1998 is filled with expectations regarding changes in theregulation of telecommunications worldwide. From the effects of liberalizationin the European Union (EU) to implementation of the World TradeOrganization’s (WTO) Basic Telecom Agreement to the need for reform of theinternational accounting rate system, regulatory bodies the world over arecalled upon to meet the challenges of the fast-changing commercialenvironment.

In most member states of the European Union, 1998 marks theliberalization of voice telephony and, thus, the abolition of the last remainingmonopoly in these countries’ telecommunications markets.79 A large part of theEU Commission’s efforts is devoted to the creation of a level playing fieldbetween incumbent operators and new entrants in the Member states.

In the United States, the implementation of the 1996 TelecommunicationsAct, which sought to introduce competition for local telephone service, is stillin progress, especially with regard to the Bell Operating Companies’ (BOCs)entry into the long distance market and the corresponding opening of their localtelephone markets to competition. In fact, the first weeks of 1998 have alreadyseen significant activity in this area, especially with regard to the application ofsection 271 of the Telecommunications Act, that is, the provision stating therequirements for BOC access to the long-distance market.

On a larger scale, the WTO Basic Telecom Agreement that was signed onFebruary 15, 1997, became effective on February 5, 1998. It provides formarket access for foreign telecommunications carriers and for adherence to abasic regulatory framework by the signatory states. Thus, the Agreementenables operators to enter the markets of the signatory states with some degreeof certainty regarding the regulatory environment.

Finally, the system of international settlement rates is in the process ofbeing reviewed, not only by the Federal Communications Commission (FCC)which issued an order containing benchmarks for international settlement rateslast year,80 but also by the International Telecommunications Union (ITU). TheITU World Telecommunications Policy Forum (WTPF) in March 1998addressed developments in world trade in telecommunications and the reformof the international settlement rate system. The 3-point opinion adopted by theWTPF is expected to improve the relationship between the ITU and the WTO,


131

support developing countries in their transition to competitive markets, andaccelerate the pace of accounting rate reform.81

One of the most important issues in the context of the liberalization oftelecommunications markets worldwide is network access and interconnection.The introduction of competition can only be successful if it is guaranteed, on theone hand, that new entrants do not have to duplicate existing facilities but cangain access to existing networks. On the other hand, to avoid inefficient marketentry and subsidization of new entrants by incumbent network operators, theoperators of existing networks must be able to recover the costs forinterconnection from the new entrants.

Implementing a fair and efficient interconnection regime is one of the mostsignificant challenges in markets currently opening to competition. Thus, notsurprisingly, it is one of the focal points of the EU Commission and nationalregulators throughout the member states. At the same time, the EUCommission is discussing the relation between international settlementagreements and cross-border interconnection. In the United States,controversial questions on the implementation of the Telecommunications Act,including the introduction of competition in local markets and network accessfor competing local exchange carriers (LECs) will likely be decided by the U.S.Supreme Court.

In discussing the issues surrounding network access for new entrants, it ishelpful to keep in mind the international regulatory environment intelecommunications and to look at the approaches chosen in different countries.To provide a framework for the contributions from individual countries thisarticle will now describe and analyze the recent developments in the regulationof international telecommunications that were already briefly mentionedabove.

This article examines the most recent of these developments in the area ofinternational telecommunications regulation. Part of the analysis is devoted tointernational agreements and initiatives such as the WTO Basic TelecomAgreement and the current attempts at a reform of the international settlementrate system. Also emphasized, because of their significant impact around theworld, are the developments in the United States and the European Union. Afocal point of this analysis is on the use of asymmetric regulation in theliberalization of telecommunications markets.


132

DISCUSSION

A. International Agreements1. WTO Basic Telecom Agreement

The WTO Basic Telecom Agreement was signed by sixty-nine states onFebruary 15, 1997, in Geneva, Switzerland. The Agreement, which could havethe widest impact of any other development for some time to come, followed alengthy period of negotiations within the so-called Group on BasicTelecommunications. Originally, the Agreement had been scheduled forsignature in April 1996. However, at that time the United States argued thatthere was no “critical mass” of countries that were willing to sign theAgreement. Therefore, it was argued, signature of the Agreement would leadto negative consequences for those countries that would open their markets. Inthe months following April 1996 more countries decided to join thenegotiations and sixty-nine of them finally signed the Agreement in early1997.82

The Agreement required implementation by the signatory countries byNovember 30, 1997. Its entry into force was scheduled for January 1, 1998.However, in November 1997 it became clear that not all signatory countrieswould be able to implement the Agreement in time to meet the deadline. Forsome countries this failure was caused only by administrative or bureaucraticimpediments. However, the United States contended that the lack ofimplementation especially in signatory states from Latin America wouldseriously impede the success of the WTO Basic Telecom Agreement.Therefore, after further discussions the signatory states agreed in January 1998to extend the deadline for ratification until July 31, 1998. The Agreement wasto become effective by February 5, 1998, for those states that had alreadyratified it. For the twelve signatory states that had not ratified the Agreement byJanuary 31, 1998, the Agreement will become effective with the date ofratification.83

a) ContentsWhat is usually referred to as the “WTO Basic Telecom Agreement” takes

the form of the Fourth Protocol to the General Agreement on Trade in Services.To this protocol are attached the schedules of commitments and lists of MostFavored Nation (MFN) exemptions undertaken by the signatory countries.Each schedule contains that country’s commitments with regard to marketaccess and national treatment, as well as any additional commitments thecountry wishes to make. Further, the schedules contain descriptions of theservices for which commitments are made.


133

An additional commitment that a majority of the signatory countriesentered into is the Reference Paper on regulatory principles. Under theReference Paper, countries are obliged to prevent major suppliers fromengaging in anti-competitive practices. The Paper particularly proscribescross-subsidization, use of information obtained from competitors, and thewithholding of necessary technical and commercially relevant informationfrom other service suppliers. Consequently, the signatory countries mustintroduce or maintain appropriate competitive safeguards.

Second, the Reference Paper requires major suppliers to ensureinterconnection at any feasible point of the network. In most WTO membercountries the term “major supplier” will apply to the monopolist or formermonopolist incumbent operator. Further, interconnection must be provided (1)under non-discriminatory terms and conditions; (2) at transparent, reasonable,and cost-oriented rates; (3) sufficiently unbundled, and (4) upon request, atpoints in addition to the network termination points. The Reference Paperprovides for the public availability of interconnection procedures andarrangements by major suppliers, that is, a reference interconnection offer mustbe available to any party interested in entering into an interconnectionagreement.

Finally, the Reference Paper requires that a dispute settlement procedureadministered by an independent regulatory body shall be made available tosuppliers requesting interconnection. There are also guidelines with regard touniversal service obligations, public availability of licensing criteria,independence of regulators, and the allocation of scarce resources.

An important consequence of the adoption of the WTO Basic TelecomAgreement is the recourse to the WTO Dispute Settlement Body which itprovides. Under Article III(3) of the WTO Agreement and the annexedUnderstanding on Rules and Procedures Governing the Settlement of Disputes,each WTO Member has the right to request a dispute settlement procedure toaddress alleged violations of the WTO Agreement, including the WTO BasicTelecom Agreement. If a WTO Member brings a case before a DisputeSettlement Body, the panel will examine the case, hear the parties and thencome to a decision within six months.

b) AssessmentThe WTO Basic Telecom Agreement is a comprehensive market opening

undertaking that, for the first time, offers enforceable mechanisms to enter aforeign telecommunications market.84 The Reference Paper provides basicregulatory tools on an international level that will allow telecommunicationsoperators to assess the chances and risks of entering a foreign market. To


134

achieve this goal of making market entry easier for foreign carriers, theReference Paper relies mostly on asymmetric regulation and the existence of anational regulatory authority.

Under the Reference Paper, major suppliers must provide interconnection,competitive safeguards against the abuse of a dominant market position arerequested, and universal service obligations must be imposed in a competitivelyneutral way. Consequently, those WTO members that do not have aninterconnection regime in place will have to introduce one in order to complywith the WTO Basic Telecom Agreement.

Interconnection is one of the most important prerequisites for a functioningcompetitive environment. Without effective interconnection, market entry willbecome extremely difficult to foreign carriers as the creation of a new networkis often prohibitively expensive and time consuming. Furthermore, duplicationof an existing network is not an economically reasonable allocation ofresources. Therefore, interconnection and access to an incumbent’s network isgenerally the most intensely fought over issue in the course of liberalization ofa monopolist market. This is true for those EU member states that have openedtheir voice telephony markets on January 1, 1998, as well as for local servicesin the United States. An example are the ongoing disputes between severalBOCs and the FCC about fulfillment of the competitive checklist contained insection 271 of the Telecommunications Act.

The Reference Paper now transfers the lessons from these experiences to aninternational level. By granting foreign carriers the right to interconnect with acountry’s major supplier, there should now be effective access to each signatorycountry’s telecommunications market. However, it also opens the door to thedisputes we know from earlier-liberalized markets in Europe and the UnitedStates. Interconnection cannot be provided for free, nor are the technicalparameters always undisputed. Thus, the judicial and regulatory challenge inthe next few years will be a task that would have challenged Salomon: to try tofairly balance the competing interests of incumbent, new entrants, andcustomers. A similar challenge is the Reference Paper’s provision fortransparent, non-discriminatory, and cost-oriented interconnection rates. Itmust be borne in mind, though, that the methods used for determininginterconnection costs have been hotly debated in the earlier-liberalizedmarkets. The better part of the FCC’s Interconnection Order,85 for example,deals with the comments and reply comments filed in the course of theRulemaking Proceeding. Therefore, it can be expected that these same issueswill surface in newly-liberalized countries when a foreign carrier claimsinterconnection rights pursuant to the Reference Paper.86


135

Does this mean, in the end, that a WTO Dispute Body will be required todetermine cost-orientation of interconnection rates in markets such asArgentina, Germany or the United States? Likely not. It is to be expected thatthe WTO, in such disputes, would rely heavily on the expertise of the nationalregulatory authority. In fact, this is another reason why the Reference Paperstrongly emphasizes the existence of independent national regulatoryauthorities. Consequently, without an independent national regulatoryauthority and its commitment to market opening and liberalization, theprovisions of the Reference Paper will not be sufficient to grant efficient marketaccess to foreign operators.

2. ITUFor several years, one of the ITU’s focal points has been the reform of the

International Accounting Rate System. Under the existing system carriersbilaterally negotiate accounting rates for the termination of internationaltelephone traffic. However, in a major part of the world, monopolist carriers areable to set accounting rates which substantially exceed the costs of calltermination. The resulting imbalances between liberalized and non-liberalizedcountries have resulted in increasing calls for a cost-oriented accounting rateregime over the past years. With its widespread membership of developed anddeveloping countries, the ITU has been so far the most prominent forum todiscuss options for a reform of the existing system.

Current developments in international telecommunications, mostprominently the implementation of the WTO Basic Telecom Agreement,mandate a reform of the international Accounting Rate system. Implementationof the WTO Basic Telecom Agreement will result in increasing liberalization inthe 69 signatory countries representing approximately 90 percent of the world’stelecommunications markets. The existing Accounting Rate system, on the onehand, is tailored for bilateral settlement agreements between two monopolistoperators. Increasing liberalization, on the other hand, will result incompetitive markets and the abolition of existing monopolies. Therefore, it isnecessary to adapt the accounting system for the termination of internationalcalls to this new, competitive environment.

In the ITU-T Study Group 3, experts from country and sector membersanalyze the future of accounting rate and settlement systems in increasinglycompetitive markets. The basis of the discussion within Study Group 3 is ITU-T Recommendation D.140 under which accounting rates must be cost-oriented,transparent, and non-discriminatory.87 This Recommendation has now beenendorsed by most countries represented in Study Group 3.88


136

The ITU’s goal in reforming the international accounting rate system is tointroduce cost-based, non-discriminatory, and transparent termination charges.These charges would have to be set based on the following elements: (1) theinternational circuit; (2) the international exchange; and (3) the nationalextension. However, as the Secretary-General points out in his report to theITU Policy Forum, the introduction of cost-based termination charges requiresa general “multilateral agreement on costing methodologies and on which costelements can be legitimately included in the calculation of the charges.”89

Existing studies, such as case studies commissioned for the Policy Forum, bestpractice rates, e.g., from the European Union, or benchmarks set by otherregulators, particularly by the FCC in its Benchmark Order,90 could provideguidance for the necessary transition.91

Further, in December 1997, Study Group 3 discussed possible means toencourage the transition from the existing accounting rate system to cost-oriented mechanisms.92 As a result, the Group agreed on severalrecommendations which, upon approval by the ITU’s member countries, willbecome part of a revised Recommendation D.140.93 Under the Study Group’srecommendations, national administrations or recognized operating agencies(ROA) should aim for a total accounting rate level of 1 SDR per minute by theend of 1998.94 By the end of 1999, national administrations/ROAs should havein place an appropriate costing methodology to determine their relevant costs.

Finally, as a result of the WTPF a Focus Group will be established whichhas the task to develop a proposal how to further reduce accounting rates as atransitional measure on the road to cost-oriented rates. The Focus Group willreport on the results of its work to Study Group 3 by November 6, 1998.

In the long run, the introduction of cost-based termination charges iscertainly the best method to account for the termination of international calls. Itis in line with the general development towards more transparency and cost-accounting in telecommunications displayed, for example, by the EUCommission’s activities in the field of cost-based interconnection pricing andaccounting separation. Also, accounting rates which exceed costs are notsustainable in a global competitive environment. Rather, their existenceencourages carriers to by-pass the existing system using, for example, refilingor call back methods. Thus, even more of an imbalance is created.

However, two things need to be considered when discussing theintroduction of cost-based termination charges for international calltermination. First, it must be ensured that a sufficient number of countries agreeto apply termination charges in place of the existing accounting rates.Otherwise, low termination charges between already liberalized countriesinvite countries with high accounting rates to refile their traffic and, thus, even


137

increase the payments they receive as settlements. Thus, a severe imbalancewould be created to the advantage of non-liberalized countries. In particular,developing countries with high accounting rates need to be convinced toimplement cost-based termination charges. While a certain percentage of freeriding can, most likely, not be avoided, a balance has to be struck between thegeneral interest in reforming the Accounting Rate system and the economicinterests of operators in liberalized countries.

Second, agreeing on and implementing a costing methodology fortermination charges will be, in all likelihood, an extremely complicatedprocess. Over the past two years, both the European Union and the UnitedStates have introduced a long run incremental cost methodology to determinecost-based interconnection charges. However, in the United States and in theEU member states the application of this methodology has led to disputes andextended court proceedings. The telecommunications markets in the UnitedStates and the EU member states are highly developed markets with efficientoperators. One can easily imagine how much harder it will be to introduce ancost accounting methodology that is acceptable to both developing anddeveloped countries with highly different market structures and infrastructureneeds. Therefore, the challenge and the hope is that the lessons learned fromthese early experiences can be effectively and efficiently applied so that cost-based accounting will become standard world-wide.

B. Recent Developments within the European Union and the United States1. European Union

As mentioned above, most EU member states liberalized their voicetelephony markets by January 1, 1998. The introduction of competition forvoice telephony was the last step in the liberalization of the EU’stelecommunications market that began with the adoption of the Green Paper onthe Liberalization of the Telecommunications Sector in 1987.95 In the yearssince, the EU has adopted various directives designed to open thetelecommunications markets in the member states, including the Competitionand the Full Competition Directives,96 the ONP Framework Directive,97 theInterconnection Directive,98 and the Voice Telephony Directive.99

In order to promote a functioning competitive environment, the EU has, onthe one hand, abolished market access barriers such as monopoly rights forstate-owned telecommunications operators. On the other hand, it has created aregulatory framework designed to enable new competitors to enter the formermonopoly markets. In this context, the EU Commission has, in the course of thepast year, particularly focused on interconnection and network access includingpricing and accounting separation.


138

The right of new market entrants to interconnect with the network of adominant provider is one of the cornerstones of EU telecommunicationsregulation. Various directives lay out the ground rules for this right. Forexample, under the ONP Interconnection Directive, organizations that providepublic telecommunications networks or services must negotiate interconnec-tion terms with each other. Other organizations, i.e., organizations that do notprovide public telecommunications networks or services, have the right torequest access to a dominant telecommunications provider’s network under theONP Voice Telephony Directive. Charges for both interconnection and accesshave to be cost-based and transparent. In order to ensure availability of anincumbent’s standard interconnection offer to new entrants, national regulatoryauthorities were required to ensure publication of a reference interconnectionoffer by July 1, 1997, under the Full Competition Directive. Finally, withrespect to the inevitable disputes that can be expected in the area ofinterconnection and access, the ONP Interconnection Directive provides fordispute settlement procedures for interconnection disputes.

As to cost orientation of interconnection or access charges, there is stilldiscussion about the adequate methodology within the EU. This pastSeptember, the European Commission published the first part of arecommendation on interconnection pricing and cost accounting forinterconnection pricing.100 Therein, the Commission recommends a bestcurrent practice approach for the time being. This approach examines peak rateinterconnection charges to an incumbent’s fixed public telephone networkunder three different scenarios: (1) local level interconnection, (2) metropolitanlevel interconnection, and (3) national level interconnection. Based upon theinterconnection charges in the three lowest priced member states, theCommission then recommends best current practice prices for interconnection.However, the Commission states in the Recommendation that, once thenecessary studies are concluded, it will favor a methodology based upon longrun average incremental costs (LRAIC) to determine interconnection costs.

As a next step in creating a basis for the determination of cost-basedinterconnection charges, this past November the European Commissionpublished a Working Document on cost accounting and accounting separationprinciples.101 This proposal is designed “to provide practical guidance toNational Regulatory Authorities on how the requirements on cost-accountingand accounting separation in the interconnection directive can beimplemented.”102 With reference to operators with significant market power103,the Commission suggests separating operating costs, capital employed, andrevenues into four business lines: core-network, local access-network, retail,and other activities.104 Thus, operators should be able to make transparent their


139

costs, revenues, and capital employed to regulatory authorities and competitors.In the Commission’s opinion, this will serve to give regulators and competitors“confidence that no anti-competitive cross-subsidies exist.”105

Another issue that the European Commission is currently looking at isinterconnection for carriers from EU member states in non-EU member states.This issue gains increasing significance in the EU because, under EUregulation, a foreign carrier has the right to terminate its cross-border traffic inany EU member state by means of interconnection. However, unless there isreciprocity with regard to the right to cross-border interconnection, this maycause imbalances in the processing of incoming and outgoing calls in EUmember states. Therefore, under Article 22 of the Interconnection Directive,the Commission must report to the European Parliament and the Council “onthe availability of rights to interconnect in third countries for the benefit ofCommunity organizations.”106 A draft of this report was made public by theONP Committee this past December.107 Therein, the Commission states that theexisting accounting rate system is no longer appropriate for charging cross-border calls between liberalized countries.108 Instead, the Commissionconsiders cost-oriented termination charges, pointing out that this mechanism isbest suited to a competitive market. Therefore, it is to be hoped that theCommission will closely watch the issue of accounting rates versus cross-border interconnection charges in the course of this year to determine possibleimpacts on the EU telecommunications industry and ensure a balance betweenintroduction of an EU-wide interconnection regime and application of theexisting Accounting Rate system.

2. United StatesIn the United States, several important aspects of the legal and regulatory

framework have been subjected to challenges and will be subjected to thescrutiny of the highest court in the land. The Telecommunications Act of 1996was designed to introduce competition in the local telephone markets where, sofar, the local Bell Operating Companies (BOCs) had a monopoly to providelocal telephone service. In order to compensate the BOCs for the loss of theirstatus as sole provider of local services, the Telecommunications Act allowsthem to enter into the long-distance market, provided that competition exists intheir local markets. Under section 271 of the Telecommunications Act, a BOCmay only offer long-distance services in its local services market if it hasentered into an interconnection agreement with a facilities-based competitor or,if there are no interested competitors, if it has an approved standardinterconnection agreement available.109 Consequently, interconnection has


140

become one of the dominant issues regarding the introduction of competition inlocal telephone markets in the United States.

In its Interconnection Order, the FCC decided that cost-basedinterconnection rates should be determined using a total element long runincremental cost methodology (TELRIC).110 This decision was widelycriticized, especially by local exchange carriers arguing that the FCC did nothave jurisdiction to determine an interconnection pricing methodology. Rather,this determination should be left to the states. In the course of the ensuinglawsuit, the Interconnection Order was stayed by the court of appeals for theEighth Circuit for jurisdictional reasons.111 The court followed the argumentsbrought by the local exchange carriers, which was essentially that the FCC didnot have jurisdiction over the issue of interconnection pricing in local markets.The court held that jurisdiction over this issue lies with the state regulators. Thecase is now before the U.S. Supreme Court, which is expected to issue anopinion by the end of 1998. However, at this point most state regulators areusing either TELRIC or a similar methodology when deciding interconnectionpricing disputes.

Another question that should be decided by the U.S. Supreme Court thisyear is the BOCs’ obligation to recombine unbundled network elements. Inorder to further competition in the local telephone markets, the FCC haddecided that BOCs were obliged to not only offer unbundled access tocompetitors but also to recombine the unbundled network elements accordingto a competitor’s demands.112 However, this obligation was also stayed by thecourt of appeals for the Eighth Circuit.113 Until the Supreme Court decides thisissue, BOCs are not being required to recombine unbundled network elements.

The latest development with regard to the opening of the local telephonemarkets to competition in the United States is the decision of the District Courtin Wichita Falls, Texas, declaring section 271 of the Telecommunications Actunconstitutional.114 In a lawsuit filed by SBC in 1997, the Court found thatsection 271 and, in particular, the so-called competitive checklist BOCs have tofulfill before they are allowed to provide long-distance services in their localservice areas, give an unfair advantage to those local exchange carriers that donot have to comply with section 271. With this decision, the BOCs would havebeen able to offer long-distance service within their local service areas withoutany restrictions. However, the U.S. Department of Justice and several long-distance carriers appealed the decision and obtained a stay from the districtcourt, which means that the court decision is ineffective until review by a highercourt is completed.

Although this latest decision is not in effect, it may have had some influenceon the FCC’s policy with regard to BOC entry into the long-distance market.


141

For example, the FCC recently announced that it will use a more cooperativeapproach when judging BOC application for entry into the long-distancemarket. Thus, a speedier granting of applications may be facilitated. This is achange from the previous procedure where only after lengthy reviews andseveral rejected applications a set of FCC precedences were formed whichcould serve as criteria for following applications.

Clearly, the implementation of the Telecommunications Act and thecorresponding changes in the U.S. telecommunications market are by no meanscompleted. As in Europe, the opening of formerly monopolist markets tocompetition creates friction between market players and regulators and givesrise to arguments which will, in the end, be decided by the courts.

CONCLUSION

So far, the expectations regarding the year 1998 have only been partlyfulfilled. On one hand, the liberalization of global telecommunications marketshas made a great leap with the implementation of the WTO Basic TelecomAgreement. In particular, implementation in the EU member states and theUnited States have led to improvements regarding market access for foreigntelecommunications companies. On the other hand, much remains to be donebefore a truly global and liberalized telecommunications market will becomereality. Effective implementation of the WTO Basic Telecom Agreement in allsignatory countries must be monitored and ensured in order to encourageinvestment and participation in foreign markets. Also, the liberalization oftelecommunications markets must play a significant role in negotiations foraccession to the WTO, e.g., with Russia or China. The international accountingrate system must be reformed to gradually adapt the cost of international calltermination to the cost of national call termination. In this context, incentivesand support must be provided for developing countries that will suffer from adecrease in foreign currency income as a result of accounting rate reform.Finally, in the domestic arena in both the United States and EU member states,a functioning competitive environment must be ensured—one that considersboth the concerns of incumbent operators and new entrants.


142

ENDNOTES

1. This is an amended and updated version of a paper presented at theTelecommunications Policy Research Conference, Alexandria, Virginia, USA onSeptember 27-29, 1997.2. Access covers non-traffic sensitive costs - local loop plus line-card; Networkcovers traffic sensitive costs, including the components used to build up interconnec-tion charges (referred to below as “conveyance”).3. Under the ECPR interconnection charges are given by the sum of the incremental costand the opportunity cost (foregone profit) on retail calls.4. BT has a few remaining analogue local switches, which are to be replaced with digitalequipment by early 1998.5. A discussion of the pros and cons of different mark-up regimes is beyond the scopeof this paper—for OFTEL’s views on mark-ups see OFTEL (1994, 1995a).6. See, for example, the series of reports prepared for OFTEL by NERA.7. With the assumption that there are no sunk costs in the long run, this is identical to theincremental cost.8. Copies of the bottom-up model on diskette are also available from OFTEL.9. In practice, this was found to lead to some differences between the models, asset byasset, but the differences were not systematic and were in both directions. Consequently,this was not a major source of difference.10. Point-to-point and intermediate multiplexed networks will have a different balanceof costs between electronics and fibre.11. There are three double tandem services offered by BT and two by Deutsche Telekom,differing by distance band. To compute the average double tandem charge, a simpleaverage has been taken of these charges.12. Switch port rental costs are included on a per minute basis assuming 1.8 millionminutes per annum per 2Mbit/s connection.13. Unbalanced tariffs (a type of access deficit calculation) account for about 0.8centimes per minute, and the universal service cost of geographically averaged tariffsaccounts for about 1 centime per minute.14. Access charges, however, will be paid on call origination. Consequently the price of“Local Carrier Select” is roughly double the price of “Local Terminating Access.”15. They are paid on inter-LATA toll calls, but not local calls (even though the networkcomponents used in each case may be identical).16. The FCC commented that a consideration in determining the definition of theincrements in TELRIC, was to minimize the common costs between elements—seeFCC (1996).17. Similarly, by construction, the sum of the component ceilings equals the stand-alonecost of conveyance. The component ceiling is therefore generally lower than the stand-alone cost of the component. Further details are set out in Annex C of OFTEL (1997d).


143

18. The author is indebted to G. Brunekreeft, W. Groß, M. Kunz, J. Merkt, H.-J. Weißand the participants of the research seminars at the Government Institute for EconomicResearch in Helsinki and the University of Freiburg and the participants of the AICGSConference in Bonn for helpful comments.19. For the analysis of technical regulations, the reader is referred to Knieps, G., DieAusgestaltung des zukünftigen Regulierungsrahmens für die Telekommunikation inDeutschland, Diskussionsbeiträge des Instituts für Verkehrswissenschaft undRegionalpolitik, Nr. 22, Universität Freiburg, Juli 1995; the analysis of an entry-compatible alternative to cross-subsidization, the so-called universal service fund, hasalready been provided in Blankart, Ch. B., Knieps, G., What Can We Learn FromComparative Institutional Analysis? The Case of Telecommunications, in: Kyklos, Vol.42, Fasc. 4, 1989, S. 579-598 .20. A separate critique of the best current practice interconnection charge at the locallevel is provided in part C.IV, although in local networks regulation still seems to benecessary.21. See Knieps, G., Phasing out Sector-Specific Regulation in CompetitiveTelecommunications, in: Kyklos, Vol. 50, Fasc. 3, 1997, pp. 328ff.22. Similar ONP-policies could also be observed in other network industries, e.g.,railroads and airlines (e.g., Knieps, G., Wettbewerb in Netzen? Reformpotentiale in denSektoren Eisenbahn und Luftverkehr, Tübingen, 1996).23. Similar ONP-policies could also be observed in other network industries, e.g.,railroads and airlines (e.g., Knieps, G., Wettbewerb in Netzen? Reformpotentiale in denSektoren Eisenbahn und Luftverkehr, Tübingen, 1996).24. Shankerman, M., Symmetric regulation for competitive telecommunications, in:Information Economics and Policy, Vol. 8, 1996, pp. 5f.25. Baumol, W.J., Panzar, J.C., and Willig, R.D., Contestable Markets and the Theoryof Industry Structure, New York, 1982.26. The argumentation in favor of asymmetric regulation in Germany is illustrated forexample in: Mestmäcker, E.-J., Witte, E., Gutachten zur Zuständigkeit für dieVerhaltensaufsicht nach dem dritten und vierten Teil des Referentenentwurfs für einTelekommunikationsgesetz (TKGE), Hamburg und München, 22. November 1995, p.12, and in: Picot, A., Burr, W., Regulierung der Deregulierung imTelekommunikationssektor, in: ZfbF (Schmalenbachs Zeitschrift fürbetriebswirtschaftliche Forschung), Vol. 48, Heft 2, 1996, p. 192.27. OECD: Working Party on Telecommunication and Information Services Policies—Universal service and public access in the technologically dynamic and converginginformation society. OECD, Paris, April 10-11, 1997.28. OECD: Working Party on Telecommunication and Information Services Policies—Corrigendum to “Universal service and public access in the technologically dynamicand converging information society.” OECD, Paris, September 15-16, 1997.


144

29. See Knieps, G., Phasing out Sector-Specific Regulation in CompetitiveTelecommunications, in: Kyklos, Vol. 50, Fasc. 3, 1997, pp. 325-339.30. Final Report and Order on the Local Competition Provisions of theTelecommunications Act, FCC-96-325, August 8, 1996, pp. 329ff.31. See Haring, J., Rohlfs, J.H., Efficient competition in local telecommunicationswithout excessive regulation, in: Information Economics and Policy, Vol. 9, No. 2, June1997, pp.119-131.32. See Griffin, J.M., The Process Analysis Alternative to Statistical Cost Functions: AnApplication to Petroleum Refining, in: American Economic Review, Vol. 62, 1972, p.47.33. See Gabel, D., Kennet, D.M., Estimating the Cost Structure of Local TelephoneExchange Network, Columbus, OH: National Regulatory Research Institute at OhioState University, 1991.34. See Gabel, D., Kennet, D.M., Economies of Scope in the Local Telephone ExchangeMarket, in: Journal of Regulatory Economics, Vol. 6, 1994, pp. 386 ff.35. See Griffin, J.M., Long-run production modeling with pseudo-data: electric powergeneration, in: Bell Journal of Economics, 1977, p. 125.36. As long as upgrading is an efficient strategy its costs should not be confused withphantom costs due to overvaluation of installed investment (based on differencesbetween economic and historical depreciation patterns (see next section)).37. The reader is referred to a series of studies NERA provided to OFTEL: TheMethodology to Calculate Long-Run Incremental Costs, March 1996; Reconciliationand Integration of Top Down and Bottom Up Models of Incremental Costs, June 1996;Reconciliation and Integration of Top Down and Bottom Up Models of IncrementalCosts, Final Report, December 1996.38. See also H. Albach, G. Knieps: Kosten und Preise in wettbewerblichen Ortsnetzen,Baden-Baden 1997, p. 31.39. The peak-load pricing principle is already developed in: Steiner, P.O., Peak load andefficient pricing, in: Quarterly Journal of Economics, Vol. 71, 1957, pp. 585-610.40. See the article by Willig, R.D., Pareto-Superior Nonlinear Outlay Schedule, in: BellJournal of Economics, 1978, Vol. 9, pp. 56-69.41. See e. g. Kahn, A.E., Shew, W.B., Current Issues in Telecommunications Regulation:Pricing, in: Yale Journal on Regulation, Vol. 4 (2), Spring 1987, pp. 191-256.42. See Henriet, D., Moulin, H., Traffic-based cost allocation in a network, in: RandJournal of Economics, Vol. 27, No. 2, Summer 1996, pp. 332-345.43. See OFTEL: Network Charges from 1997, London, May 1997.44. Commission of the European Communities: Commission Recommendation onInterconnection in a liberalized telecommunications market - Part 1 - InterconnectionPricing, Brussels, October 15, 1997 (Provisional text).45. For different views, see Brunekreeft (1997) and Knieps (1997) according to whomtelecommunications markets either represent bottlenecks (local networks) or are


145

contestable (the service and long-distance markets). This contrasts, for example, withMacAvoy (1996), who maintains that there is a lack of competition in the U.S. long-distance market, something that would not be sustainable under contestability.46. In New Zealand, the incumbent DO is regulated “only” by a price moratorium onbasic services.47. The ECPR is widely attributed to Willig (1979) and Baumol (1983). For an extensivediscussion, see Baumol and Sidak, 1994, and the Winter 1994 edition of The YaleJournal on Regulation and in the Fall 1995 issue of the Antitrust Bulletin.48. The relationship between the ECPR and the Ramsey approach is akin to the onebetween partial and general equilibrium analysis. However, in the case ofinterconnection the legitimacy of partial analysis is called into question becauseinteractions between the relevant markets are bound to be very strong.49. These effects do not yet include incentive effects as discussed in Laffont and Tirole(1993). The absence of incentive effects can be justified if the incentive-pricingdichotomy holds.50. We are here neglecting the possibility that network interconnection may efficientlybe priced below marginal/incremental costs because of network externalities or becauseof lack of competition in the retail market.51. See, for example, in Arnbak et al., 1994, Mitchell et al., 1995, Mitchell andVogelsang, forthcoming.52. See, for example, Perl and Falk (1989) or Shin and Ying (1992).53. Similarly, calculating stand-alone costs is harder than calculating incremental costs.In particular, stand-alone costs of various output combinations need to be calculated inorder to derive the relevant upper bound, because the DO’s revenues should not exceedstand-alone costs for any output combination.54. Laffont and Tirole (1996) equate the imputation requirement with the ECPR.However, imputation implies upper bounds for interconnection charges (or, minimalinternal transfer prices), while the ECPR declares these upper bounds to be optimal.55. Price adjustment formulas of similar kinds would even be advisable for privatelynegotiated interconnection agreements.56. For a different opinion, see MacAvoy (1996).57. The slow pace obviously suits the ILECs. However, it may also suit some others,such as cable TV companies, who want to be full network entrants and who take time todevelop their network technology and infrastructure.58. In principle, the sum of network element costs properly weighted should equal totalnetwork costs. However, elements may include setup and collocation costs that wouldnot be needed for a stand-alone network.59. See, for example Gabel and Gabel (1997) for a history of cost measurement in theU.S. telephone industry.


146

60. Brock (1997) believes that the ambiguity in the U.S. Telecom Act about the FCC’sauthority to interfere with intrastate matters was intentional, in order to get the actpassed.61. See Williamson’s (1996) comment to Sidak and Spulber. Williamson’s maincounter-argument against a construction of takings under Sidak’s and Spulber’s view ofregulation as a long-term contract is that long-term contracts are always incomplete andhave to be adapted. In regulatory contracts, in particular, issues of opportunism arise onboth sides (e.g., regulatory capture, not prudently incurred investments, entrydeterrence and other moral hazard issues). As a result, there cannot be guarantees oneither side. According to Williamson, a guarantee to compensate for strandedinvestment would lead to wasteful investment.62. However, the influence of state regulators mitigates part of the strong language of theU.S. Telecom Act, as interpreted by the FCC.63. As noted in Kress (1997), the U.S. Telecom Act concentrates on issues specific tolocal telecommunications (except for Sections 271/272) because long-distancetelecommunications had been experiencing competition for more than two decadesalready. In contrast, the German Telecom Act had to include local and long-distancenetworks. However, the U.S. Telecom Act did trigger a long-distance access chargereform for interstate access.64. For an extensive analysis of these developments see Vogelsang and Mitchell (1997).65. See Deutsche Telekom (1997).66. Under an efficiency approach we may define the presence of incumbency burdens bythe fact that, without its fault, the incumbent as a firm (or a well-defined part of the firm)is valued lower by the (financial) markets than it would cost to replace its assets (bymodern assets, age adjusted). This means that, for the firm, Tobins ‘q’ < 1.67. See Deutsche Telekom (1997). Based on the low telephone density of some formerEast German districts, Deutsche Telecom could actually try to claim universal servicesubsidies under the German Telecom Act.68. See Oftel (1997). Note that the issue of universal service costs is related to but notidentical with that of access deficits.69. See Johnson (1988) for the high administrative costs of U.S. lifeline and link-upprograms.70. See Albach et al. (1997) and Vogelsang (1996 and 1997).71. Dixit, Avinash K. and Pindyck, Robert S., Investment Under Uncertainty,Princetown University Press 1994, p. 6.72. Ibid73. Defined in Article 1(a) as “the physical and logical linking of telecommunicationsnetworks used by the same or a different organization in order to allow the users of oneorganization to communicate with the users of the same or another organization, or toaccess services provided by another organization. Services may be provided by theparties involved or other parties who have access to the network.”


147

74. Commission Recommendation on Interconnection in a Liberalized Telecommunica-tions Market. January 8, 199875. Directive 98/10/EC of the European Parliament and of the Council of 26 February1998 on the application of open network provision (ONP) to voice telephony and onuniversal service for telecommunications in a competitive environment.76. See William J. Baumol & Gregory Sidak, Toward Competition in Local Telephony,MIT Press/AEI, 1994 Ch 7.77. The criteria are the extent to which telecommunications services: (A) are essential toeducation, public health or public safety; (B) have, through the operation of marketchoices by customers, been subscribed to by a substantial majority of residentialcustomers; (C) are being deployed in public telecommunications networks bytelecommunications carriers; and (D) are consistent with the public interest,convenience and necessity.78. However, in section 252, Congress does mandate cost standards for pricing of certainelements, such as interconnection and unbundled network elements, between carriers.79. Of the fifteen member states, the following were granted deferments ranging from sixmonths to three years with regard to the liberalization of voice telephony: Greece,Ireland, Luxembourg, Portugal, Spain.80. International Settlement Rates, IB Docket 96-261, Report and Order, FCC 97-280.81. ITU World Telecommunications Policy Forum, Trade in Telecommunications,Report by the Chairman, March 18, 1998, available at http://www.itu.int/wtpf/sg_rep/chairman/english.doc.82. For a brief overview over the signatory countries and their individual commitments,see William J. Drake & Eli M. Noam, The WTO deal on basic telecommunications, 21TELECOMMUNICATIONS POL’Y 799, 803 [hereinafter Drake & Noam].83. In the following countries the Agreement was not ratified as of January 26, 1998:Argentina, Belgium, Bolivia, Brazil, Chile, Dominica, Dominican Republic, Ghana,Guatemala, Papua New Guinea, Philippines, Poland.84. Drake & Noam at 809.85. Implementation of the Local Competition Provisions in the Telecommunications Actof 1996 and Interconnection between Local Exchange Carriers and Commercial MobileRadio Service Providers, First Report and Order, 11 FCC Rcd 15499 (1996), Order onReconsideration, 11 FCC Rcd 13042 (1966), petition for review pending and partialstay granted, sub nom. Iowa Utilities Bd. v. FCC, 109 F.3d 418 (8th Cir. 1996), pricingrules vacated, 1997 WL 403401 (8th Cir. July 18, 1997) (“Interconnection Order”).86. Drake & Noam at 806.87. International Telecommunications Union, Telecommunications StandardizationSector. ITU Recommendation D.140, Accounting Rate Principles for InternationalCommunications Services (Geneva 1992, revised 1995).


148

88. See Third Draft of the Secretary-General’s Report to the Second WorldTelecommunications Policy Forum on Trade in Telecommunications, available at http://www.itu.int/wtpf/sg_rep/3_draft/002v3e2.doc (“Third Draft”) at para. 44.89. Third Draft at para. 47.90. International Settlement Rates, IB Docket 96-261, Report and Order, FCC 97-280.91. Third Draft at para. 47.92. Id.93. Id. at ¶ 45.94. Approximately $1.35.95. Towards a Dynamic European Economy: Green Paper on the Development of aCommon Market for Telecommunications Services and Equipment, COM(87)290final.96. Commission Directive 90/388, 1990 O.J. (L 192) 10; Commission Directive 96/19,1996 O.J. (L 74) 13.97. Council Directive 90/387, 1990 O.J. (L 192) 1.98. Parliament and Council Directive 97/33, 1997 O.J. (L 199) 32.99. Parliament and Council Directive 95/62, 1995 O.J. (L 321) 6 (A revision of the 1995ONP Voice Telephony Directive has been adopted on January 29, 1998, by the Counciland will have to be implemented by the Member states by mid-1998).100. Commission Recommendation on Interconnection in a liberalized telecommunica-tions market, Part 1 - Interconnection Pricing of 15 October 1997, 1997 O.J. (C 3148).101. Interconnection in a liberalized telecommunications market: Working document onCost Accounting and Accounting Separation of 6 November 1997, XIII/A/1.102. Id. at 2.103. An operator is “presumed to have significant market power when it has a share ofmore than 25 percent of a particular telecommunications market in the geographical areain a Member State within which it is authorized to operate.” Parliament and CouncilDirective 97/33, 1997 O.J. (L 199) 32, Article 4 (3).104. Interconnection in a liberalized telecommunications market: Working document onCost Accounting and Accounting Separation of 6 November 1997, XIII/A/1 at 5.105. Id. at 2.106. Parliament and Council Directive 97/33, 1997 O.J. (L 199) 32, Art. 22 para.1.107. ONP Committee Draft Report on the availability of rights to interconnect in thirdcountries for the benefit of Community telecommunications organizations of December23, 1997, ONPCOM98-01.108. Id. at 8.109. Telecommunications Act 47 U.S.C. § 271 (1996).110. Interconnection Order at paras. 690-693.111. Iowa Utilities Bd. v. FCC, 109 F.3d 418 (8th Cir. 1996).


149

112. Interconnection Order at paras. 292-294.113. Iowa Utilities Bd. v. FCC, 120 F.3d 753 (8th Cir. 1997), cert. granted by AT&TCorp. v. Iowa Utilities Bd.114. SBC v. FCC, 981 F. Supp. 996 (D.C.N.D. Texas)


150

CONFERENCE AGENDA

Telecommunications Reform in Germany: Lessons and Priorities

Wednesday, November 19, 1997

Opening DinnerSpeaker: Arne Börnsen, MdB, Chairman, Committee for Post and

Telecommunications

Thursday, November 20, 1997

Panel I: The Role of Regulatory ArrangementChair: Richard Simnett, Bellcore Applied ScienceSpeakers: Michael Riordan, Federal Communications Commission

Richard Schultz, McGill UniversityJörg Sander, German Ministry of Post and Telecommunications

Panel II: Access to the Dominant Network: Interconnection andUnbundlingChair: Marc Harold Scanlan, European CommissionSpeakers: Geoffrey Myers, OFTEL

Peter Quander, German Ministry for Post andTelecommunicationsThomas Mellewigt, o.tel.o communications GmbH & Co.Martin Schlieker, Deutsche Telekom AG

Panel III: Setting Interconnection Charges: Costing and Pricing ofCompetitive Telecommunications in GermanyChair: Martin Cave, Brunel UniversitySpeakers: Günter Knieps, University of Freiburg

Ingo Vogelsang, Boston UniversityWerner Neu, Institute for Communications Research (WIK)

Panel IV: Panel Discussion: Lessons and PrioritiesChair: Jürgen Müller, Berlin School of Economics (FHW)Speakers: Martin Cave, Brunel University

Barbara Cherry, Northwestern UniversityAndrea Huber, Deutsche Telekom AG


151

Karl-Heinz Neumann, RWE Alliance AGLuigi Prosperetti, University of MilanDavid Sevy, France Telecom


152

American Institute for Contemporary German StudiesAICGS Conference Reports

Gibowski, Wolfgang, Dieter Roth, and Hans Herbert von Arnim. WhoseParty is This? Transitions in the German Political Party System.Washington, D.C.: AICGS, November 1996.

Janes, Jackson and Charles C. Loveridge. The Future of Telecommunica-tions: A German-American Dialogue. Washington, D.C.: The Center forStrategic and International Studies and AICGS, November 1996.

Culpepper, Pepper. Skills for the 21st Century. Washington, D.C.:AICGS, January 1997.

Westin, Alan. Data Protection in a Global Society. Washington, D.C.:AICGS, May 1997.

Cowles, Maria Green. The Limits of Liberalization: RegulatoryCooperation and the New Transatlantic Agenda. Washington, D.C.:AICGS, July 1997.

Rahr, Alexander, Sherman Garnett and Zbigniew Brzezinski. The Futureof Ukraine: Challenges to German and American Foreign Policy.Washington, D.C.: AICGS, July 1997.

Rudolf, Peter and Geoffrey Kemp. The Iranian Dilemma: Challenges toGerman and American Foreign Policy. Washington, D.C.: AICGS, July1997.

Gibowski, Wolfgang, Josef Joffe and Dieter Roth. The Road toGermany’s National Elections. Washington, D.C.: AICGS, February1998.

The Parameters of Partnership: Germany, the U.S. and Turkey.Challenges for German and American Foreign Policy. Washington,D.C.: AICGS, April 1998.


153

Pavliuk, Oleksandr, Alexander Rahr and John Tedstrom. Ukraine inTransition and Western Strategy: Challenges to German and AmericanForeign Policy. Washington, D.C.: AICGS, August 1998.

MacDonald, Duncan. Protecting Privacy: The Transatlantic DebateOver Data Protection. Washington, D.C.: AICGS, September 1998.

Harris, Martha Caldwell, Kay Möller, Margot Schüller, and DavidShambaugh. Managing Conflict, Building Consensus: Germany, theUnited States and the People’s Republic of China. Washington, D.C.:AICGS, November 1998.

Telecommunications Reform in Germany: Lessons and Priorities.Washington, D.C.: AICGS, December 1998


154

American Institute for Contemporary German Studies 1400 16th Street, N.W. Suite 420 Washington, D.C. 20036-2217 Tel: 202.332.9312 Fax: 202.265.9531

ISBN 0-941441-37-7