A Model for Calculating Interconnection Costs in Telecommunications PPIAF Paul Noumba Um Laurent Gille Lucile Simon THE WORLD BANK Christophe Rudelle A Model for Calculating Interconnection Costs in Telecommunications A Model for Calculating Interconnection Costs in Telecommunications Guidebook prepared by Paul Noumba Um, Laurent Gille, Lucile Simon, and Christophe Rudelle THE WORLD BANK PPIAF Copyright © 2004 This guidebook is a publication of the Public­Private Infrastructure Advisory International Bank for Reconstruction and Facility (PPIAF). PPIAF is a multidonor technical assistance facility, aimed at Development helping developing countries improve the quality of their infrastructure THE WORLD BANK through private sector involvement. For more information on the facility, see 1818 H Street, NW Washington, DC 20433, USA the website at www.ppiaf.org. The findings, interpretations, and conclusions expressed in this guide are Telephone 202-473-1000 entirely those of the authors and should not, in any manner, be attributed to Internet www.worldbank.org the PPIAF, theWorld Bank or its affiliated organizations, or to members of its E-mail feedback@worldbank.org Board of Executive Directors or the countries they represent. Neither PPIAF All rights reserved nor theWorld Bank guarantees the accuracy of the data included in this pub- Printed in the United States of America lication or accepts responsibility for any consequence of their use. The First printing: December 2003 boundaries,colors,denominations,and other information shown on any map 1 2 3 4 07 06 05 04 in this report do not imply, on the part of PPIAF or the World Bank Group, any judgment on the legal status of any territory, or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. The World Bank holds copyright on behalf of both the PPIAF and itself. Dissemination of this work is encouraged and, in such instances, the World Bank will grant permission promptly,and will not require a fee when reproduction is for noncommercial purposes. Any request to photocopy portions of this publication should be addressed to: Copyright Clearance Center, Inc., 222 Rosewood Drive, Dan- vers, MA 01923, U.S.A., telephone 978-750-8400, fax 978-750-4470, or through the Internet at www.copyright.com. In regard to questions concerning this publication, or information on ordering more copies,please refer to the PPIAF website,or contact PPIAF by email at the address below. PPIAF c/o The World Bank 1818 H Street Washington, DC 20433 Fax: 202-522-7466 www.ppiaf.org Email: info@ppiaf.org ISBN 0-8213-5671-2 Library of Congress Cataloging-in-Publication Data Modèle de détermination des tarifs d'interconnexion. English. A model for calculating interconnection costs in telecommunications / edited by ... [Paul Noumba Um]. p. cm. Includes bibliographical references. ISBN 0-8213-5671-2 1.Telecommunications--Rates--Africa,Sub-Saharan--Mathematical models. 2.Telecommunications--Africa, Sub-Saharan--Costs-- Mathematical models. I. Noumba Um, Paul. II.Title. HE8467.M6313 2003 384.6'4--dc22 2003067270 Contents Foreword viii Acronyms and Abbreviations ix About the Authors x 1. Introduction 1 Preliminary Results Provided by the Cost Model 2 The Limits of the Model 4 2. Guidebook Overview 6 Cost Modeling Principles 6 User Guide 8 The Model 9 3. Cost Modeling Principles 11 Regulation of Interconnection 11 Essential Facilities 11 Cost Determination Methodologies 14 Economic Criteria for Assessing Costs 15 The LRIC Concept 18 Advantages and Drawbacks of LRIC 21 Definition of Interconnection Services 22 Definition of the Increment 23 Application to African Telecommunications Networks 25 The Average Size of an African Telecommunications Network 25 Nodes and Links 26 Transit 27 4. Modeling Principles The Modeling Principles 29 A Long-Run Approach 29 A Forward-Looking Approach 29 An Efficient Approach 29 An Economic Approach, Not an Accounting Approach 30 A Bottom-Up Approach 30 The Working Units 30 The Model Structure 30 Applications of the Cost Model 32 v Contents 5. User Guide 34 Filling in the Assumptions 35 Demand Assumptions 35 Main Technical Assumptions 38 Main Assumptions on Routing Factors 41 Main Cost Allocation Assumptions 42 Assumptions for the Mobile Network 45 The Results Sheets 47 The Results 47 Printing Reports 49 Testing the Sensitivity of the Results 49 Managing the Model 51 6. Operations of the Cost Model 52 Basic Principles 52 Logic of the Intermediary Spreadsheets 53 Capacities (Capa Sheet) 53 Transmission Capacities (Capa ElTr Sheet) 53 Infrastructure Capacities (Capa Infra Sheet) 54 Costs 57 The Mobile Network Calculations 58 Capacity of the Mobile Network (CapaMob Sheet) 58 Cost of the Mobile Network (Costs Mob Sheet) 58 Total Costs of the Mobile Network (Tot Mob Sheet) 59 Appendixes 60 Appendix 1: Economic Cost Approach 60 Appendix 2: Capital Cost Approach 63 Appendix 3: Radio Concentrator Solutions 66 Bibliography 67 Tables 1.1 A Sample of Interconnection Rates 1 1.2 Overview of the Telecommunications Sectors in Burkina Faso, Côte d'Ivoire, Cameroon, and Zambia 3 1.3 A Sample of Interconnection Rates Calculated by the Cost Model 3 2.1 Typology of Nodes and Links in a Telecommunications Network 7 3.1 Correlation Matrix between Retail Services and Elementary Services 23 3.2 Matrix of Inputs Needed to Provide Retail Services 24 3.3 Snapshot of Telecommunications Development in Africa in 1999 25 3.4 Matrix of Transmission Links 27 A3.1 Number of Subscribers Depending on the Traffic per Subscriber (in mE) 66 Figures 1.1 Comparison between Interconnection Prices: Côte d'Ivoire 4 1.2 Comparison between Interconnection Prices: Cameroon 4 2.1 Cost Model Architecture 10 3.1 The Costs of a Multiproduct Firm: An Example for Five Products 14 3.2 Transition from Historical Accounting Costs to Economic Costs (LRIC) 19 3.3 Cost Structure 20 3.4 Entry Points of Interconnection Service 23 4.1 Cost Model Architecture 32 6.1 Cost Model Architecture 52 A3.1 Example of the Architecture of an IRT (TRT-Lucent Network) 66 vi Contents Boxes 3.1 Different Definitions Selected for Cross-Subsidization 17 5.1 Erlang 39 6.1 Sizing of the SDH Network (example of RCU-LS routes) 54 vii Foreword The tremendous economic and social impact Furthermore, it demonstrates the World Bank resulting from pervasive, as well as effective, use of Group commitment to assist governments in develop- information and communication technologies has ing the proper enabling environments and effective convinced more and more African governments to regulatory institutions, and hence, contribute effec- embark on the restructuring and progressive liberaliza- tively in mitigating the regulatory risk as perceived by tion of their telecommunications sectors. potential investors. Helping newly established regula- In a liberalized market, a procompetitive regulatory tory agencies develop their capacity has been under- framework is a factor in establishing a level playing scored as a key pillar of the new World Bank Group field for a favorable environment that encourages Information and Communication Technologies (ICT) greater participation by the private sector. In that strategy (www.worldbank.org/ict). respect, establishing a sound interconnection frame- We sincerely hope that this guidebook and its cost work that ensures equal treatment, nondiscrimination model will be used by telecom regulators and opera- among market players, and cost-based tariffs is assumed tors to settle interconnection disputes. However, we to be the main engine for developing competition in would urge regulators to customize and expand the the sector. cost model to better match their internal needs. Determining interconnection tariffs is a complex The guidebook consists of six chapters. Chapter 1 and extremely sensitive task. In Africa, the absence of introduces the guidebook. Chapter 2 provides an accurate cost information has rendered the situation all overview of the guidebook. Chapter 3 reviews cost the more complex. In fact, some telecommunications modeling principles for calculating interconnection regulators resolve interconnection disputes on the rates, while chapter 4 illustrates how these principles grounds of available tariff benchmarking, although are integrated in the cost model. Chapter 5 is a user these tariffs may not always be relevant. guide and chapter 6 reviews cost modules composing This guidebook provides a sound methodology to the model and illustrates how interconnection costs are help regulators and telecommunications operators computed. adopt a tariff regime and deal with interconnection disputes on the basis of a rigorous cost model. Mohsen A. Khalil viii Acronyms and Abbreviations ABC Activity-based accounting MSC Mobile switching center ADM Add-drop multiplexer MUX Multiplexer BHCA Business hour call attempts NRA National regulatory authorities BHE Business hour Erlangs PDH Plesiochronous digital hierarchy BSC Base station controller PPIAF Public­Private Infrastructure BTS Base terminal station Advisory Facility CS Central station RCU Remote concentrator unit ECPR Efficient component pricing rule RIO Reference interconnection offer EU European Union SDH Synchronous digital hierarchy FDC Fully distributed costs TDMA Time division multiplexing access GDP Gross domestic product TELRIC Total element LRIC GICT Global Information & TRX Transmitter/receiver Communication Technologies TS Terminal station IS International switch TSLRIC Total service LRIC IT Information technology TSW Transit switch LRAIC Long run average incremental costs USO Universal service obligation LRIC Long run incremental cost WACC Weighted average cost of capital LS Local switch WBI World Bank Institute ix About the Authors Paul Noumba Um is lead infrastructure specialist, Télécommunications (ENST) Paris, 46 Rue Barrault, World Bank Institute (WBI),The World Bank Group. Paris 75013 France. Previously, he was head of the reg- Previously, he was regional coordinator for Africa in ulation group in BIPE SA. the policy division of the Communications and Infor- mation Technology department. Lucile Simon and Christophe Rudelle are consultants with BIPA SA ­ L'Atrium, 6 place Abel Gance, Laurent Gille is head of the department of economics Boulogne Billancourt, 92652 France. and social science at École Nationale Supérieure des x 1 Introduction Since the past decade, several Sub-Saharan African a fair and competitive basis. Despite the clarity and governments, through technical assistance provided by soundness of the legislative provisions in that respect the World Bank and other donors, have undertaken to (cost oriented, nondiscriminatory, fair, and transpar- reform their telecommunications sectors, by imple- ent), the number of interconnection disputes has menting market liberalization policies, privatizing the increased, and long-lasting interconnection disputes incumbent public operator, and creating autonomous have discredited the reputation and credibility of new and independent regulatory bodies.The core objective regulatory regimes. of these reforms is to significantly improve access, and One has to admit, however, that deriving optimal affordability, to telecommunications services on the interconnection rates from the principles codified in basis of the assumption that a more friendly and pre- national laws is tricky in countries constrained by dictable business environment will attract more private scarce resources (human resources with relevant expe- investment. However, the provision of interconnection rience, management and information systems that are services, on fair and efficient terms, has rapidly unable to provide accurate and comprehensive data, emerged as a main bottleneck. and so forth). As illustrated in table 1.1, interconnec- In fact, new legislation and regulations enacted in tion rates are mostly derived from international bench- Sub-Saharan Africa recognize the interconnection marking. Although the international benchmarking rights ascribed to all telecommunications service approach may be a satisfactory starting point, it is not providers and network operators. In addition, these always relevant, since it may not take into account spe- regulations also request the incumbent fixed operator cific country parameters that affect the industry cost to supply interconnection services to new entrants on structure. Table 1.1 A Sample of Interconnection Rates Euro cents/ minute Benin Burkina Faso Burundi Cameroon Côte d'Ivoire Mali Mauritania Togo Local 4.5 3.5 3.8 4.0 3.8 3.8 2.6 6.1 Simple transit 4.6 27.8 3.8 8.6 19.8 16.4 7.6 6.1 Double transit 19.8 27.8 3.8 19.8 19.8 16.4 7.6 6.1 Transit 1.8 1.2 3.8 2.3 Mobile termination 19.8 13.4 3.8 22.1 23.7 30.0 7.6 9.9 Source: Data collected from participants of the Access Pricing Workshop held in Ouagadoudou, Burkina Faso, in March 2002. 1 A Model for Calculating Interconnection Costs in Telecommunications At the beginning of the sector reform in Africa, in The guidebook was prepared by a team led by Paul the mid-1990s, it was believed that incumbent fixed Noumba Um (World Bank), including Laurent Gille, operators would hold their dominant position in the Lucile Simon, and Christophe Rudelle from BIPE SA long run. However, the recent explosive development (France).The team made use of comments and guid- of cellular business substantially contradicted that pre- ance provided by Antonio Estache and David Satola diction. In most countries, mobile operators connect (World Bank). Daniel Benitez (IDEI, University of more subscribers than the fixed service.In a short time, Toulouse) reviewed the cost model and provided sug- cellular operators have achieved significant market gestions for its improvement. We gratefully acknowl- power and are becoming dominant players themselves. edge valuable comments provided by participants to Therefore, regulation of interconnection and access the Interconnection Day, which was organized by the pricing should not only focus on calls terminating in GICT Department on April 10, 2003. the fixed network, but also should consider calls origi- We equally wish to express our gratitude for the nating or terminating in the mobile networks. valuable support offered, throughout this project, by TheWorld Bank Global Information & Communi- Michele Rajaobelina, Lizmara Kirchner, and Lucy cation Technologies (GICT) Department believes that Cueille (World Bank). This exercise could not have developing a more accurate and robust methodology to been completed without the meaningful collaboration assess interconnection rates, based on long-term incre- and support provided by telecommunications opera- mental costs, will support efforts by borrower countries tors and regulators in the following countries: Burkina to implement best practice regulation, and would sig- Faso, Côte d'Ivoire, Cameroon, Senegal, and Zambia. nificantly improve the reputation and credibility of KC Translations Services LLC revised the English ver- national regulatory authorities (NRAs). Our hope is to sion of the guidebook. see this generic cost model customized and expanded by NRAs and be used to solve pervasive interconnec- Preliminary Results Provided by the Cost tion disputes, which have flourished across Africa. Model The guidebook includes a CD-ROM that contains the bottom-up cost model.The cost model was devel- From July 2002 to February 2003, the World Bank oped by BIPE SA1 under a Public­Private Infrastructure GICT Department commissioned BIPE to conduct field Advisory Facility (PPIAF) grant managed by the World visits to Burkina Faso, Cameroon, Côte d'Ivoire, and Bank. It builds, to some extent, on a model initially Zambia. During these visits, workshops were conducted developed by Europe Economics,2 at the request of the to train regulator and telecom operator staff on how to European Commission. However, the proposed model use the cost model.This section provides an overview of takes into account the specific features characterizing the preliminary findings derived from these visits. telecommunications development in Africa (embryonic The four countries have implemented telecommu- size of the network, predominance of microwave tech- nications reforms and established sound regulatory nology for transmission links and limited roll out of frameworks enabling competition in specified market fiber-optic cables, rollout of expensive time division segments such as the mobile market. Although these multiple access [TDMA] systems to connect rural local- countries are not homogeneous, they do, however, ities, and limited regulatory capacity), and calculates reflect a series of commonalities in terms of regulatory routing factors,in the light of parameters entered by the framework, market structure, and overall telecommu- user. It generates interconnection rates for fixed-to- nications development. Interconnection disputes and fixed, fixed-to-mobile, and mobile-to-fixed calls. Other the urgent need for regulators to ensure their effective value-added services are not captured in the attached and timely settlement were identified as among the version but could be easily added by users. primary concerns of private mobile operators.The fol- The cost model is available at www.world lowing table provides an overview of the telecommu- bank.org/cit and the Europe Economics model3 is nications sector situation in these countries. available at www://europa.eu.int/ISPO/infosoc/ With the exception of Cameroon, and Côte telecompolicy/en/Study-en.htm. d'Ivoire, fixed incumbent operators have been granted 2 Introduction Table 1.2 Overview of Telecommunications Sectors in Burkina Faso, Côte d'Ivoire, Cameroon, and Zambia 2001 Data Burkina Faso* Côte d'Ivoire Cameroon Zambia Subscribers to the fixed network 60,000 300,000 105,000 105,000 Subscribers to the mobile network 111,145 730,000 510,000 150,000 Number of mobile operators 3 3 2 3 * 2002 data. Source: BIPE 2003. mobile licenses. In general, the dominant position of 2001 by ATCI was appealed by the fixed incumbent to fixed incumbent operators is seriously challenged by CTCI, the sector's appeal court for disputes between the explosive growth observed in the mobile market operators and ATCI. In November 2002, CTCI pub- segment. lished its decision on interconnection rates. Neither In the four countries, interconnection rates were the ATCI arbitration nor the CTCI appeal decision finally decided by regulators in an attempt to settle was based on sound economic analysis. lengthy interconnection negotiations. In Zambia, The preliminary results provided by the cost model interconnection disputes were subsequently brought to during these field visits are summarized below. In the law courts by mobile operators,and are still not set- Zambia, the fixed incumbent was unable to provide tled. In Côte d'Ivoire, following the submission of the traffic information required to run a cost simulation.In revised interconnection rates in 2000 by the fixed Côte d'Ivoire, arbitration was decided by CTCI, and incumbent, mobile operators filed complains to the resulting rates remain higher than the ones recom- local regulator (ATCI). The arbitration published in mended after the workshop. In Burkina Faso, the regu- Table 1.3 A Sample of Interconnection Rates Calculated by the Cost Model Euro Cents Burkina Faso Côte d'Ivoire Cameroon Zambia Fixed Network Current Rates Local 3.1 9.8 4.0 NA Simple transit 14.9 9.8 8.5 5 Double transit 14.9 19.8 19.8 NA Cost Model Rates (with TDMA systems included) Local 1.0 3.3 1.7 NA Simple transit 1.8 4.4 9.6 NA Double transit 2.4 5.2 12.7 NA Transit 0.5 0.8 3.1 NA International transit 2.5 1.0 6.4 NA Mobile Network Current Rates Local 9.5 15.2 22.1 5.0 Cost Model Rates Celtel Telecel Telmob Orange Telecel Orange MTN Celtel Telecel Originating 13 9 8 5.7 7.7 27.2 24.4 14.3 11 Termination 13 9 8 9.8 9.4 27.2 24.4 14.3 11 NA, not applicable. Notes:The interconnection prices for collection and termination are not theoretically symmetrical in so far as this traffic does not call on the same network elements in a symmetrical way.This lack of symmetry can be taken into account by means of routing factors. However, in the cases of Cameroon and Zambia, we worked with default routing factors that do not discriminate between collection and termination traffic, hence the equality of the results found.We indicated to the operators and regulators that they should refine the calculation of the default routing factors in order to better take into account the real use of the various network elements at traffic collection and termination levels. In reality, the notion of transit corresponds to the difference between double transit and single transit, in a case where a third party operator wants to route traffic on the fixed-line operator's network from one transit zone to another transit zone, while collecting and terminating the traffic itself. For further details, see the model's user guide. Source: Authors' own calculations. 3 A Model for Calculating Interconnection Costs in Telecommunications lator (ARTEL) enacted an interim regulation on inter- The Limits of the Model connection rates following a complaint filed by one of the mobile operators. During the workshops, the dis- Although the proposed cost model builds on African cussions and debates were less emotional and were telecommunication network specificities, it can be more focused on specific issues related to the relevance applied in non-African environments provided appro- of the model assumptions or parameters. priate adaptations are made.The model is easy to use, Although it is too early to derive any definitive and requests information that regulators or operators conclusion from this experience, the empirical results can easily find.However,there remain some limitations. provided by the cost model are quite robust. In Côte First,the model is designed for"small"networks that do d'Ivoire, the interconnection rates generated by the not yet implement complex transit functions or routing model were below the rates ratified by the regulator for algorithms. Second, the model does not seek full opti- single, double transit, and call termination on the mization when rebuilding the transmission network. mobile networks. For these services, interconnection is More specifically, the model does not optimize the charged above costs, and interconnection providers are nodes, as the current network topology is kept extracting monopoly rents. Conversely, local intercon- unchanged ("scorched node" or Brownfield approach). nection seems to be cost oriented. Additional actions Third, it does not provide a detailed modeling of the from the local regulator are therefore needed. In con- cables and duct networks, though it discriminates trast, the simulation in Cameroon provided mixed among different types of geography (urban, suburban, results.The incumbent's network is not optimized, and rural) or nature of the trenches (wrapped, ducted, costs incurred to provide interconnection services to buried). Such a module could be developed by each competitors are therefore abnormally high. Finally, the regulator,although node location optimization may not results show important differences between mobile be critical at the current stage of telecommunications termination costs in Côte d'Ivoire and Zambia on the network development in developing countries. one hand,and Cameroon on the other hand.These dif- In conclusion, the proposed cost model provides ferences can be explained by substantial investments accurate cost proxy estimates when applied to net- made by mobile operators in Cameroon to bypass the works with less than 1.5 million main lines and when incumbent,which did not have excess capacity to meet the incumbent's network topology (number and loca- the demands of the mobile operators. tion of nodes) is optimized or close. As shown in Figure 1.1 Comparison between Interconnection Prices: Figure 1.2 Comparison between Interconnection Prices: Côte d'Ivoire Cameroon Euro cents Euro cents 25 30 25 20 20 15 15 10 10 5 5 0 0 Local Single Double Collection Termination Local Single Double Collection Termination transit transit transit transit Current prices Model s results - fixed Current prices Model s results - fixed Model s results - mobile 1 Model s results - mobile 2 Model s results - mobile 1 Model s results - mobile 2 4 Introduction Cameroon, the results obtained when the incumbent's network is underoptimized may be misleading. In a subsequent version of the cost model, an "earth node" module will be added to generate interconnection costs when node location and links are optimized. Therefore, regulators will be able to calculate a bounded interval of interconnection rates. The lower bound will correspond to a configuration of a fully optimized network (nodes and links), while the upper bound will correspond to a partially optimized net- work (only links are optimized). Notes 1. BIPE SA can be contacted at: L'Atrium, 6, WC2A 1QU, Tel.: (+44) (0) 20 7831 4717, nomics. This study resulted in the produc- place Abel Gance, F92652 Boulogne Billan- Fax: (+44) (0) 20 7831 4515. tion of a model spreadsheet in MS-Excel court Cedex,Tel.: 33 (0)1 46 94 45 22, 3.See "Final Report on the Study of an format [EN, 4 Mb] (with a voluminous user Fax: 33 (0)1 46 94 45 99, Adaptable Model Capable of Calculating the guide),which is described in the main report E-mail:Accueil@bipe.fr, http://www.bipe.fr. Forward-Looking, Long-Run Incremental [EN, 440 kb] and an executive summary 2.Europe Economics Research Ltd.(Europe Costs of Interconnection Services for EU [EN, 65 kb] (both available here as Adobe Economics) can be contacted at: Chancery Member States" (April 2000), prepared for Acrobat *.pdf files). House, 53-64 Chancery Lane, London the European Commission by Europe Eco- 5 2 Guidebook Overview Interconnection negotiations and settlements are by an efficient African network operator using the best among the main regulatory issues to reckon with in available technologies. Considering the tremendous regard to the development of competition in the development of mobile communications in Africa, the telecommunications sector in African countries. model also calculates the interconnection cost for Although the majority of legislation includes provi- fixed-to-mobile calls, and conversely. sions ensuring the interconnection rights for new In general, the model requires the entry of substan- entrants and sets cost orientation principles for deter- tial amount of information characterizing the net- mining the interconnection rates, only a few regulators works that are interconnecting (topology, architecture are equipped to effectively implement these regula- principles and rules, traffic matrix and patterns, costs tions in practice. Regulators do not have relevant cost elements, and so forth).To palliate the frequent gaps in information that would allow for effective arbitration the information systems, the model proposes default of interconnection disputes. Moreover, regulators are values that the user could consider if needed. This not equipped to assess the cost orientation of intercon- applies mainly to the routing factors, which are rarely nection rates proposed by fixed incumbents. Hence, documented in most countries. they are ill equipped to ratify reference interconnec- The overall objective is to provide regulators and tion offers. operators with a decision tool that can enhance common Furthermore, it is difficult to replicate cost models understanding and cooperation in dealing with a sensitive developed for more advanced economies in Africa. subject.The next sections briefly present the cost mod- African networks are small in size and quite spread out. eling principles and the user guide and review the cost They rely on specific architectures and technologies to model structure. reconcile their small market size in volume to the scat- tered habitat. Transmission links are mostly over Cost Modeling Principles microwave technology, and the roll out of fiber-optic systems remains limited to urban centers. Nobody could envisage dynamic competition in the It is in this context that the World Bank contracted telecommunications market without the actual BIPE to develop a cost model that captures specificities enforcement of the interconnection regulations.With- prevailing in Africa and that could be easily replicable. out interconnection, competing operators will be The model belongs to the bottom-up, long run incre- obliged to duplicate onerous infrastructure, and con- mental cost (LRIC) models family, and determines sumers would have to subscribe to connections to dif- interconnection rates by calculating the costs incurred ferent operators'networks.Conversely,with an effective 6 Guidebook Overview interconnection regime,a seamless communication sys- accounting and involve allocating historical costs to tem is more likely to develop, enabling consumers to different services according to criteria prescribed by contract the service with whichever supplier, and be the regulators. Others estimate costs by reconstituting ensured of receiving all incoming calls, from wherever the networks on the basis of currently available tech- they originate. Nonetheless, the interconnection nologies. It is generally admitted that the latter are the regime has to be implemented with market liberaliza- most appropriate for estimating interconnection ser- tion in view;and this calls for free negotiation and con- vices costs. LRIC methodology estimates the costs tracting. Hence, interconnection agreements have to be incurred, while offering a subset of services.The costs freely negotiated in line with regulations. That key considered are those that would be avoided if these principle--as it is applied--opens the door to abuse services were not offered. and strategic behavior from the operator enjoying a For estimating the cost of the interconnection ser- dominant market position. Fixed incumbent operators vices, the selected increment comprises network ele- are likely to enjoy a strategic comparative size advantage ments belonging to the core network, that is, those (number of subscribers) when the market is opened up shared among all the network's users, excluding net- to competition,which can distort or hamper the devel- work elements dedicated to end users. opment of competition. Safeguards are, therefore, The proposed cost model takes into account the needed to protect new entrants from these anticompet- specific nature of African networks and has several fea- itive behaviors; hence, regulators are mandated by law tures: to ratify reference interconnection offers submitted by · A low number of main lines spread out, however, incumbents. It is also required that they ratify agree- over large territories. ments negotiated by the parties. In so doing, regulators · Traffic concentrated over a small number of net- must not only make sure that the agreements are work nodes. entirely in compliance with existing regulations but · Transit function almost nonexistent,and low capac- they must also check their consistency with the guiding ity of the transmission network,which relies merely principles, such as interconnection rates cost orienta- on microwave technology. tion. In checking compliance with the cost-orientation · Predominance of rural concentration systems that principle, regulators need to access relevant and accu- use TDMA-type radio systems. rate cost information to motivate their decisions.How- · The presence of domestic satellite networks. ever, assessing the cost orientation also requires an These specific features are taken into account in the ability to assess the way costs for interconnection ser- model, and rural radio concentrators are integrated in vices are formed and distributed. the increment. This is a complex and cumbersome task. Indeed, a The cost model takes into account six types of telecommunications operator manages a complex nodes and five types of links between these nodes, as business.It rolls out networks using different technolo- summarized in the following table: gies and investment layers spread out over time. It offers a portfolio of services, which are interdependent and call on the same productive resources. Further- Table 2.1 Typology of Nodes and Links in a more, some of the services are sold on a retail basis (in Telecommunications Network the final market) while others, such as interconnection Nodes IS TS LS RCU CS TS services, are sold to other operators forming a whole- IS sale range of products. In such a context, assessing or TS to IS TS-LS TS-LS identifying costs incurred by each category of service LS TS-LS RCU-LS (Local RCU link) requires the implementation of a sophisticated cost allo- CS CS-TS cation management and information system. TS Economists have proposed various cost allocation IS, international switch;TS, terminal station; LS, local switch; RCU, remote methodologies. Some are based on the operator's concentrator unit; CS, central station. 7 A Model for Calculating Interconnection Costs in Telecommunications The cost model assumes the existence of two tran- The parameters describing functionality shared by sit levels: international switch (IS) and domestic transit, several networks are filled in by default values.The user which are often not implemented in practice.The tran- can modify these values whenever needed.Two situa- sit functions are often performed by the local switches tions are possible: (LSs), which are also used to connect subscribers · Default values are not calculated values and are directly or indirectly through remote concentrator filled in light green boxes. In this case, the user can units (RCUs). Finally, TDMA technology radio con- modify these values. centrators, characterized by central stations (CSs) and · The default values are calculated from the user's terminal stations (TSs), are widely used to connect inputs and are filled in table format on the right of remote rural localities to the fixed telephone network. blank light blue cells. In case the user wants to pro- The cost modeling assumes that costs are: ceed with default values, then he/she should avoid · Long-term--meaning that all the cost compo- filling in alternative values. nents are variable. Once all of the required information is entered or · Forward looking--implying that the model con- validated, the user can view the results presented in siders current costs and not historic costs. subsequent tables. · Efficient--this implies that the model takes into · The first table gives the cost per minute of various account the best available technology.This is done network elements (nodes and links). without modifying the network topology. In this · The second table gives interconnection services case, the cost model retains the scorched node costs as follows: approach. ­ The first line provides the average interconnec- · Economic--not accounting costs. For instance, tion costs resulting from selected routing factors. the model converts investment costs into constant ­ The second line supplies the interconnection equivalent annuities. costs with or without factoring inTDMA system · Bottom-up--this implies rebuilding the network costs. These costs are compared with European according to the previous principles. best practice rates provided in euros. · Computed per minute. ­ Finally, the table translates average costs obtained The cost model is Excel based and includes 21 in tariffs according to the pricing structure in spreadsheets,7 of which can be accessed via an interface force (depending on time or day). provided in the Menu sheet. For fixed networks, the · Similar simplified tables are provided for cost of ter- model calculates interconnection costs for local, single minating or originating calls from mobile networks. transit, double transit, and international transit calls. For The user can refine certain assumptions and con- mobile networks, the model calculates interconnection duct a sensitivity test.This can be done for six prese- costs for both terminating and originating calls. lected variables: · Traffic level at peak time (as a percentage of the User Guide overall traffic). · Total length of trenches: variation by percentage. The user guide is intended to facilitate the use of the · Proportion of staff dedicated to the core network model by staff from regulatory agencies and telecom- functions (maintenance, operation, and so forth). munications operators. · Employee average annual cost. After selecting a working language (English or · Markup to the equipment capital cost incurred by French), the user is requested to provide generic the operator (this is a proxy reflecting exogenous parameters defining the network's configuration,archi- factors that increase capital cost). tecture, and topology. · The proportion of debt in the total capital structure. The user enters in the blue boxes parameters char- The user can print a report including the simula- acterizing the network's configuration in terms of size, tion results. Similarly, the user can save the sensitivity architecture, topology. test modifications. 8 Guidebook Overview The Model preventive and corrective maintenance,energy con- sumed). The model is described in greater detail in the third · Staff costs related to operation and maintenance section of this guidebook. The overall logic of the activities. model is simple. For small but spread-out networks, staff costs can · The model begins with a nomenclature of network hardly be appraised as a percentage of the investment elements (nodes and links). costs.This is, indeed, a difficult task.The model calcu- · Each service uses these elements in different pro- lates the number of staff needed to efficiently operate portions.The routing factors represent the average the network and derives salary costs accordingly. The number of times a given element is used by the ser- staff cost is then distributed among different network vice considered.The model then calculates the total elements according to parameters specified by the user. load supported for each network element. Similarly, the investment costs are calculated by · The model calculates the size of network elements considering the estimated volume of traffic to be han- (transmission elements) within the framework of dled by each network element at peak hour.The size of the selected topology. each network element is accordingly derived using · The model adds up all corresponding network ele- engineering procedures. For each network element, ment costs and calculates the per-minute cost for the model calculates the investment cost incurred and each network element. generates an investment annuity. Operating and non- · Finally, the model calculates the interconnection attributable costs are then allocated to it to obtain a costs on the basis of the routing factors. global cost per minute for each element. For each ser- The network is considered to be made up of net- vice,relevant costs per minute for the network element work elements, namely, nodes and links. These ele- are added up to obtain the interconnection cost.The ments convey interconnection traffic throughout the latter value is adjusted using the gradient of retail prices network.This implies the mobilization of network ele- to derive the interconnection rate. ments according to the complexity of the interconnec- The model calculates the interconnection costs tion service requested. over two stages. Furthermore, the model calculates the investment · First, the model determines the size of the switching cost and also the cost of operation and maintenance elements. Switching elements are considered to be incurred by these network elements. These costs are the nodes of the network. Their investment costs distributed over four components. depend on the switching system BHE (business hour Erlangs transformed into 2 megabytes per second Investment Operating [Mbps]) and the number of subscribers connected. costs costs · Second, the size of the links connecting network's Attributable costs nodes is calculated. The model differentiates the Common costs infrastructure and transmission layers: ­ The transmission layer enables the user to design Common costs are expressed as a percentage of the the electronic transmission equipment by choos- attributable costs.The percentage is usually decided by ing and sizing the capacity of the most appropri- the regulator and is to some extent arbitrary. The ate technology to be rolled out (synchronous attributable costs are the costs directly caused by the digital hierarchy [SDH] rings, plesiochronous interconnection service and would have been avoided digital hierarchy [PDH] technology). by the provider.They consist of operating and invest- ­ The infrastructure layer enables the determina- ment costs.The operating costs are composed of two tion of the links' substratum or the physical ele- terms: ments that will support the transmission link: · Maintenance and operating cost related to the net- trenches for fiber-optic cables, microwave towers work element (spare parts, equipment section of and masts, and satellite earth stations. 9 A Model for Calculating Interconnection Costs in Telecommunications The trenches are broken down by geo-type (urban, · One sheet for calculating the peak load traffic and suburban, and rural) corresponding to various burying determining the network's elements size. techniques (wrapped trench, ducted, fully buried).The · Two sheets for determining the transmission and microwaves are characterized by the nature of their infrastructure size. mass (light, medium, or heavy). · Three sheets to calculate the network element costs The links are sized by the traffic carried at peak (switching, transmission, and infrastructure). hour expressed in megabit per second (Mbit/s). The · One sheet for summing up the total interconnec- transmission link size is determined, to enable normal tion costs (including the shared costs) and calculat- flow of traffic expected from switching centers and ing the interconnection costs per minute and per leased lines connected to them. Certain infrastructure element. elements are shared by different categories of traffic; · One sheet to present the results. this is particularly applicable to SDH rings in the access Figure 2.1 below summarizes the cost model archi- network. The model enables such costs to be shared. tecture. For example, SDH rings in the access network will be In conclusion, this cost model is a tool provided to used to carry core network traffic flow and the access regulators and operators to help them: network traffic. · Develop a better understanding of the interconnec- The model comprises 21 sheets organized as fol- tion costing and economics. lows: · Determine economic-oriented interconnection · One menu sheet. rates for terminating and departing traffic from, or · Twelve sheets forming the core of the model, as to, fixed and mobile networks. described below. · Ratify crucial assumptions pertaining to traffic · Four sheets for the mobile networks. demand, network optimization, and cost allocation · One sheet to conduct the sensitivity analysis. between final and intermediary services. · Three specific management sheets (two sheets for · Report on the benchmarking exercise, for the rat- publishing the fixed-line and mobile reports, and ification of the above mentioned assumptions. one sheet to manage the two languages and the Instead of benchmarking interconnection rates, the regula- default values). tor should instead focus on benchmarking key cost driv- The 12 sheets forming the core of the model are: ers. · Four sheets for assumptions and parameters · Better identify indicators to be monitored, on a enabling the configuration of the network. regular basis, by the regulator. Figure 2.1 Cost Model Architecture Routing Unit Hypotheses Demand Network Factors Costs Demand Tech FactRout UCosts Traf c Traf c Capa Transmission Infrastructures Sizing Capa Eltr Capa Intra Cost Switching Transmission Infrastructures Costs Sw Costs Tr Costs Infra Totals Results Result tot results 10 3 Cost Modeling Principles Regulation of Interconnection tion demands presented by new entrants reasonable? If so, can they be met by the incumbent in a reasonable Since the past decade,the majority of African countries time frame? Second, the interconnection agreements have adopted legislation opening up their telecommu- are assumed to be freely negotiated, as with any other nications market, or certain segments of it, to competi- commercial contracts, although they cannot become tion. The mobile communications segment was effective unless approved by the regulator. Are parties opened up to competition, with two to five mobile likely to reach a fair agreement? If so, how much time operators authorized to commercialize their services. should be given to them to reach such agreements? Is In regard to fixed-line networks, the scope of compe- the regulator well equipped to review and ratify agree- tition remained limited as temporary exclusivity peri- ments resulting from private negotiations? Whenever ods were accordingly granted to incumbent fixed negotiations fail, is the regulator equipped to effectively operators before or after their privatization. arbitrate ensuing disputes? What should the procedure In parallel, independent or autonomous regulatory be,and what should the appeal procedure be?There are bodies were established in the majority of African laws stipulating that interconnection tariffs must be cost countries with the mandate to establish a level and oriented to ensure productive efficiency. In practice, competitive playing field. The right to interconnect newly created regulators are ill equipped to perform was included in various national laws,and national reg- that role. Field experience also shows that most inter- ulators were authorized to effectively enforce its connection disputes are related to rates,highlighting the implementation. As expected, regulating interconnec- overall importance of the economic dimension of tion has been one of their main activities.To sum up, interconnection pricing.The remainder of this guide- the key cross-cutting questions are: (a)Which firms are book treats that dimension. subject to interconnection rules? (b) What network components are subject to the rules? (c) On what Essential Facilities terms can these specific components be shared with Normally, at an early stage of competition develop- competitors? (d)What is the most appropriate term for ment, the incumbent will enjoy a "Stackelberg first interconnection contracts? (e) How should the net- move advantage," and will accordingly be considered work owners be compensated for interconnection to, to be a dominant player.In such a situation,because the and re-use of, their embedded systems? incumbent controls access to inputs that new entrants Implementing interconnection regulations raises need in order to compete, and can prevent or hamper multifaceted issues. First, there are technical-related competition development in downstream markets, considerations that supersede the effective implementa- there is a need for specific regulations. As a result, the tion of interconnection agreements. Are interconnec- access to these facilities must be regulated.The essential 11 A Model for Calculating Interconnection Costs in Telecommunications facilities concept refers to those facilities owned or drastically limits the scope of competition. Indeed, in controlled by one of the players,and whose duplication the absence of an interconnection regulatory frame- is too costly and uneconomical for new entrants. work, new entrants would have no other alternative Essential facilities should therefore be mutual or shared than to roll out their own local loop network in order resources, and the need to ensure equal access is to reach existing telephone consumers. Consequently, straightforward in certain infrastructure sectors as we consumers would need to subscribe to as many service illustrate here. providers as possible, and the overall social value of the Let us assume that there are two airlines in country telephone system would be suboptimal. A.Airline X was the only provider of public air trans- port service before the government decided to liberal- Pricing of Essential Facilities ize the market. In the aftermath of air transport Although interconnection is a mandatory obligation liberalization, the government licensed a new airlineY. embodied in national regulations, concluding and Prior to the liberalization, airline X was operating 50 managing interconnection agreements are generally routes from the main airport built and owned by the left to concerned parties, as are other commercial company. To compete with X, airline Y needs to agreements.1 Consequently, the obligations mandated develop its airport or negotiate access to airline X's air- by legislation and regulations are considerably weak- port.What would be the outcome of the competition? ened when the parties' objectives differ significantly. It It is likely that airline Y's capacity to compete with X is, therefore, important that regulations provide sound will be seriously undermined unless it gains fair access safeguard measures protecting the most vulnerable to the airport facility controlled by X. In conclusion, players from anticompetitive practices by the incum- the new legislation that enacted the liberalization of bent. During the early stage of competition, the the airline market must also ensure fair access to the incumbent's survival does not depend on fair access to airport facilities controlled by X. Then, the question competitors' networks.The survival of new competi- becomes how to set the terms for providing and regu- tors, however, is highly dependent on the terms of lating this access. access to the incumbent's network and customers. Obviously, operations and airport ownership During, the interconnection agreement negotiations, should be unbundled to allow fair competition incumbents will likely charge high prices for access to between airlines. Consequently, airport infrastructures their networks, while entrants will seek cheaper prices. are usually owned by the state or its representative Incumbents will also try to delay the provision of while the management of the facility is often delegated interconnection services, as long as possible, to com- to a commercial entity. Unbundling ownership and petitors.In other words,the regulator's role is crucial in operation of the airport facility hence allows for the deciding fair access terms. establishment of a level playing field for airline service Specific regulations are needed, and are justified, provision. Of course, further regulations are needed to whenever one of the market players enjoys a dominant ensure that all licensed airlines have equal access to air position. However, in an extremely rare situation in transport services. In the absence of such a regulatory which there is no dominant player, and assuming that framework, each new airline will have to run a private all new entrants have sound and mutual interests in airport, or be extremely dependent on facilities owned accessing each other's network at competitive terms, and managed by other competing airlines, hence limit- enforcing specific interconnection regulations may be ing the development of competition. needless. Nonetheless, the most common situation is Similarly to the airline industry example discussed when one of the players dominates some market seg- above, some components of the fixed telecommunica- ments, but not necessarily all of them. In that case, the tion network owned and operated by the incumbent dominant player finds no incentive in facilitating the can be considered as being essential facilities, at least conclusion of fair interconnection agreements. during the early development stage of competition.Any It is essential to spell out dominance criteria, as failure to provide new entrants access to these facilities clearly as possible, and to outline the obligations of the 12 Cost Modeling Principles dominant player.The new European regulatory pack- · Ratify the terms and conditions of negotiated age (2002) specifies that "an undertaking shall be interconnection agreements. deemed to have significant power if, either individually · Effectively arbitrate interconnection disputes. or jointly with others, it enjoys a position of economic strength affording it the power to behave, to an appre- Cost Orientation ciable extent, independently of competitors, customers Nondiscrimination and transparency are not the only and ultimately consumers. Where an undertaking has obligations imposed on dominant operators. Intercon- significant market power on a specific market, it may nection rates are also required to be cost oriented. also be deemed to have significant market power on a There are two ways for competition to blossom in the closely related market,where the links between the two telecommunications industry. One way is to support markets are such as allow the market power held in one infrastructure-based competition. According to that market to be leveraged into the other market, thereby way,policies and regulations are developed with incen- strengthening the market power of the undertaking."2 tives to promote investments in infrastructure. The In practice, controlling a market share of greater rationale in this strategy is that without affordable than 25 percent is often considered as being in a domi- infrastructure, competition in service provision will nance situation.3 In general, it is recommended to refer remain limited. It is therefore important to support the to any relevant jurisprudence that national competition development of alternative infrastructure providers; commissions or authorities would have established in this will make the best possible use of technological that area.As a result, fixed telecommunications incum- innovations and will lower entry costs to service bent operators would be rightly considered as being providers.The second way is to support service-based powerful or dominant operators.Whenever an operator competition.According to that way, policies and regu- is dominant or powerful, it is more likely to implement lations should be developed with incentives to pro- tariff discrimination, hence distorting competition. It is mote investments in service provision that maximize therefore essential to ensure that interconnection ser- the load factor of existing infrastructure. vices are provided on a nondiscriminatory basis. Inter- Regulation of input prices is justified whenever connection services provided to new entrants should be there is a risk of substantial anticompetitive practices. identical, in terms of quality, technical conditions, and Input prices can be priced at excessively high levels or rates to similar services that an incumbent would be can be priced at overly low levels. In both cases, com- providing to its own affiliates.4 petition is limited or constrained,either downstream or Regulating interconnection agreements implies upstream. However,"the imposition of a price control that the regulator be appropriately staffed and by national regulatory authorities should not have a equipped with a variety of regulatory tools that make negative effect on long-term competition,nor discour- it possible to: age investment in different infrastructures.The national · Effectively enforce the accounting separation prin- regulatory authorities should take into account the ciple or the unbundling of regulated activities from investments made by operators providing these inputs, the competitive ones. factoring in the risks incurred accordingly."6 · Ensure that interconnection rates are nondiscrimi- In practice, the cost orientation principle does not mean natory.5 selling at marginal cost.It is about determining the average cost · Publish a detailed reference interconnection offer incurred by an efficient operator using the best available tech- (RIO), including the description of relevant offers nology. Consequently, the above-mentioned average broken down into network elements as demanded cost incorporates possible economies of scale and scope by the market, and complemented by the corre- achieved by the operator providing the input.The over- sponding modalities, conditions, and prices. all objective is, therefore, to ensure that input prices · Ratify interconnection reference offers submitted reflect the industry productive efficiency frontier. by dominant operators according detailed proce- To ensure the orientation toward costs, there are dures. two main approaches: 13 A Model for Calculating Interconnection Costs in Telecommunications · The first approach relies on benchmarking input duced by the firm. Consequently, these expenses are prices in similar or comparable environments.This tied to the production of a specific service or product approach contains a serious shortfall. In general, and should not exist if that production is stopped. economic conditions differ from one country to Direct costs can be fixed or variable. another, and differences cannot always be explained Joint costs are generated by a family of services or by market factors. Some of the differences could products (for example, buildings costs for a telephone just be related to the geography or other specific firm). From an economic viewpoint, joint costs are socioeconomic conditions. Consequently, this costs incurred in fixed proportions7 every time a ser- approach only provides rough estimates of possible vice or a product belonging to the same family is pro- cost frontiers. Regulators should refrain from rely- duced by the firm. For example, a telephone company ing excessively on benchmarking to set intercon- incurs joint costs whenever it conveys a local, interur- nection rates. ban, or international call. · The alternative approach is analytical and reviews Common costs are shared by all the services or the cost structure of the regulated operator against products of the company (for example, the fixed costs the one provided by an efficient operator. Because of acquiring licenses). Common costs include the the operator provides a broad range of services, it is remainder of the costs that are not directly attributable important to be able to differentiate costs according or joint, and which are incurred by the firm. to their relevance to the specific input considered. In conclusion, summing up joint and common To ensure the cost orientation of interconnection costs boils down to the total shared costs incurred by rates submitted by dominant operators, the regulator the firm. These costs can be attributed to services or must build and enhance its knowledge with respect to the products manufactured by the firm using more or less industry cost frontier and cost drivers.The next section fur- arbitrary criteria. However, whenever shared costs can ther develops the cost allocation methodologies and be attributed in a nonarbitrary way, reflecting the highlights their relevance and limitations. causality factor, they are referred to as indirectly attribut- able costs. Conversely, whenever the attribution can Cost Determination Methodologies only be arbitrary, it is referred to as nonattributable costs. For multiproduct firms, determining the cost incurred to produce a specific product or service is a delicate and complex exercise. With the limitations that apply, this Figure 3.1 The Costs of a Multiproduct Firm: section presents the different methodologies that could An Example for Five Products be used to assess a multiproduct firm cost structure. Product Product Product Product Product The Generic Costs Borne by a Multiproduct Firm A B C D E Figure 3.1 illustrates the complexity of a multiproduct firm cost structure. It is assumed that the firm manu- Variable costs Variable costs factures five products (A, B, C, D, and E).The overall Directly attrributable costs total cost incurred by the firm is also known.The issue, therefore, is to determine the total cost incurred in producing each product. How should the unit cost per Fixed costs Fixed costs product be determined? To begin with, it is important to define the following concepts: Direct costs or directly attributable costs are expenses that are incurred when producing a specific Joint costs Joint costs service or a series of services or products. In other terms, direct costs attributed to product A will cease to exist if product A is no longer manufactured or pro- Common costs 14 Cost Modeling Principles Within directly attributable costs,it is important to dif- Average cost is the unit cost obtained by dividing ferentiate fixed from variable costs as follows: the total cost by the number of units produced. The Fixed costs represent the proportion of the firm's average cost function will decrease as production expenses that does not depend on, or vary with, the increases up to a threshold beyond which it then activity of the firm. Fixed costs include production increases (at least in the short term) with the output. capacity costs and other preinvestment expenses The average total cost corresponds to the sum of the incurred when preparing the launch of the firm's variable average costs and the average fixed costs. activities. In the event there is a major variation of the Marginal cost is the total cost variation resulting firm's activities, the fixed costs component will also from a variation of the firm's production.Marginal cost vary as a result of the capacity adjustment. However, is defined by economists as the incremental cost result- these adjustments are not necessarily below specified ing from the production of an additional unit (or cost thresholds. From an economic viewpoint, fixed costs of the last unit produced).A more formal definition is are assumed to be independent of the volume of pro- given by the first derivative of the total cost function duction and are borne by the firm even if it is not relative to the produced quantity. operating. Whenever the firm's activities are shut Average and marginal costs are basic concepts in down, some of the fixed costs incurred by the firm economics, and the definitions given above hold for become sunk costs. Sunk costs are considered to be monoproduct firms. In summary, it is important to nonrecoverable after the firm's activities cease. recall that marginal cost represents the theoretical bot- Variable costs are closely related to the level and tom cost that a firm has to recover in the short run. the development of the firm's production and market- These definitions are slightly modified for a multi- ing operations.When some operations are halted, the product firm. In fact, the total costs for a firm produc- corresponding variable expenses disappear. Con- ing several goods depend on the quantities and the versely, when operations develop, variable costs also proportions of goods produced. It is therefore impor- move in the same direction.Variable costs include raw tant to differentiate two polar situations: (a) propor- material costs, labor costs, other intermediary input tions of produced goods do not change; (b) costs, as well as variable marketing costs (delivery proportions of produced goods do change. The con- expenses, brokerage, commissions, allowances). Vari- cepts of radial and incremental cost are then used to able costs are not strictly proportional to the develop- refine marginal and average costs. ment of the activity because of the evolution Average radial cost. Whenever the family of characterizing production factors or technology inno- goods produced by the firm remains unchanged dur- vation. For instance, if the raw material costs vary pro- ing the production cycle, it is more appropriate to use portionally to production, that is not the case with the concept of the average radial cost instead of average salary costs. cost (that is, with a constant proportion of products). The sum of the fixed costs, the variable costs, Average incremental cost. Whenever there is a the joint costs, and the common costs gives the change in the composition or proportion within the total production cost or global cost. The global family of goods produced by a firm, it is appropriate to cost or total production cost is directly related to the use the concept of average incremental cost.The aver- production volume (total cost increases with produc- age incremental cost is defined as the average cost asso- tion increase). However, in the presence of scale ciated with a product or a group of products among economies, the unit cost drops as production increases. those manufactured by the firm. The average incre- Whenever scope economies are present, it is econom- mental cost for a product group usually decreases with ically more efficient to have only one firm serving the the increase in the number of product groups (scope market than to have several competing firms. economies). Two fundamental cost concepts yield from the total In theory,the marginal,radial,and incremental costs cost definition recalled above: the average cost and the refer solely to the variable component of the cost func- marginal cost. tion. Pricing at marginal cost does not enable the firm 15 A Model for Calculating Interconnection Costs in Telecommunications to recover the fixed cost. This situation occurs in an (1 + i)n before adding them. Above, i represents the industry in which scale and scope economies prevail. cost of capital, that is, the cost applying to resources One way to solve this problem is to use long-term borrowed from the financial market or resources pro- incremental average costs, as all costs are then variables. vided by the shareholders in terms of equity. However, joint and common costs still have to be We also refer to the following: financed. Other important cost concepts exist, but we · Total discounted cost, which is given by [the do not discuss them here.Instead,we refer the reader to initial investment9] ­ [the resale value in the dis- more advanced economics textbooks.8 counted terminal year] + [discounted operating The cost typology,discussed above,has been simpli- costs]. fied on purpose to illustrate issues and problems per- · Economic cost or average discounted cost, taining to cost-allocation methodologies when dealing which is the constant annuity equivalent to the dis- with a multiproduct firm. Conversely, the presentation counted current cost. does not reflect the refined cost allocation criteria · Discounted marginal cost for year n, in the found in most recent economic literature. absence of a resale value, is equal to the operating costs. If the resale value is not 0, it is the sum of the Economic Criteria for Assessing Costs operating costs and gap between the discounted Taking into account the economic dimension of the resale value in year n ­ 1 and resale value in year n. various costs categories, as discussed above, entails a better understanding and definition of cost allocation TRENDS ON COST DYNAMIC EVOLUTION. Historical criteria. cost is the cost value entered in the firm's books (pur- The difference between fixed and variable costs is chasing cost or production cost). This cost obviously time related. Generally, fixed costs are long term, in the cannot represent the real cost of the asset at the end of sense that they reflect expenses incurred by the firm to several years for a series of reasons, including wear, develop capacity and meet its production objectives. obsolescence, depreciation of the currency, and aging. Conversely, variable costs are directly related to the Forward-looking, long-term cost: Contrary to day-to-day operations of the firm;hence,they are short historical cost, forward-looking cost relies on the best term. However, in the long run, even fixed costs are available technologies,assuming that the firm's produc- variables. It is therefore possible to reconcile cost tion cycle is globally at optimum. accounting analysis to the economic analysis. As a general rule, for a company producing several outputs,provided that it is less expensive to jointly pro- NOTIONS OF ECONOMIC COST. The notion of eco- duce all the outputs than to produce them separately, nomic cost involves bringing a series of costs spread the total cost for producing one output is lower in a out over time back to one base year.Adding up all of joint production structure than in a stand-alone pro- these costs does not make it possible to measure their duction. Otherwise, it would be in the company's economic importance.As such, if we have an income interest to produce these different outputs separately. amount Q in year 0, it would be expected that if this Joint and common costs,therefore,illustrate economies amount were invested according to market conditions of scope, which the industry production structure (interest rate i), it would yield an income Q such entails.However,allocating these costs among products that Q = Q × (1 + i) in year 1, and Q = Q × (1 + remains a difficult and complex task. i)2 in year 2, and Q = Q × (1 + i)n in year n, etc. There are several methodologies in regard to allo- Conversely, if one intends to spend an income cation of joint and common costs. However, none of amount D in year n, that implies he or she should, these methodologies provides a satisfactory solution. today, have an income equivalent to D/(1 + i)n today. Depending on the methodology used, the resulting A sound measurement of a series of expenses with dif- allocation is biased by the arbitrariness embodied in ferent time occurrence requires discounting the each criterion. In general, we look for a cost distribu- expenses (that is,dividing the expenses in year n by the tion that avoids cross-subsidization among services or 16 Cost Modeling Principles products to the greatest extent possible.As discussed in based accounting (ABC) systems. Furthermore, the the box below, the concept of cross-subsidization also allocation of joint and common costs is done using has several definitions, depending on the context. arbitrary distribution keys. There are clear-cut differences in the way the allo- There are two families of LRIC models: (a) top- cation of joint and common costs is handled by various down LRIC and (b) bottom-up LRIC. methodologies. Four main cost allocation methodolo- Top-down LRIC is essentially an ABC methodol- gies are available. Some are cost allocation methods in ogy.Top-down LRIC models derive incremental costs the strict sense of the term, while others are pricing by summing up the costs that can be directly attributed methods from which allocation principles are derived. to the service and adding a markup that covers a pro- The four methodologies are: portion of joint and common costs. The markup is 1. The fully distributed costs (FDC) methodology. determined with the help of arbitrary distribution 2. The efficient component pricing rule (ECPR) keys. Therefore, the resulting cost is more backward methodology. oriented than forward looking. 3. The Ramsey-Boiteux and Laffont-Tirole method- Bottom-up LRIC is a constructivist methodology ology, which rely on demand-price elasticity. for determining forward-looking service cost. The 4. Long-term incremental costs (LRIC) methodol- methodology involves simulating the cost incurred by ogy. an efficiently operated network, using best available None of the above methodologies is fully satisfac- technology, to provide the service. Bottom-up LRIC tory. All of them build on sound economic rationale, models are highly recommended for regulatory deci- and can be criticized or defended depending on one's sions, though highly criticized for their lack of realism. point of view.The cost model proposed herein builds The LRIC methodology takes into account fixed costs on the long-term incremental costs methodology, caused by the provision of interconnection services, which will be developed later. but does not take into account the common costs, The FDC methodology records the expenses which do not vary proportionately with the provision incurred by a firm and allocates them to respective of interconnection services. In practice, the implemen- products based on the causality principle. In that tation of LRIC models requires a systematic assessment methodology, a cost breakdown procedure is used that of demand and cost evolutions, as well as the interde- groups costs by nature, function, and, according to an pendencies that may result.A systematic analysis of net- intertwined nomenclature, hierarchy. This approach work element load factors is also critical to accurately relies heavily on the availability of reliable accounting size the resources needed to convey the incremental information, which is usually generated by activity- traffic. In other words, the LRIC methodology makes Box 3.1 Different Definitions Selected for Cross-Subsidization10 · The public policy view:From a public policy perspective,cross-subsidization occurs in a regulated industry when the regulated firm uses revenues from one market to keep operations in another market that is financially not viable.The cross-subsidy is considered anticompetitive if there are cash flows from noncompetitive to competitive markets.The cross-subsidy is considered a universal ser- vice obligation (USO) if the cash flow (1) goes anticlockwise; (2) occurs only because regulatory rules create it; and (3) would not occur if the government policy were absent, and/if the markets were competitive. · The cost allocation view: In more general usage, if a service's prices do not make a reasonable contribution to overhead costs, it could be argued that the service is not carrying a fair share of the overheads, and is, therefore, being subsidized. · The Baumol-Faulhaber view: Baumol and Faulhaber have taken the view that cross-subsidization occurs when prices for a ser- vice do not cover the service's incremental cost and the company still earns a normal profit (that is, zero economic profit) overall. This implies a maximum price of stand-alone cost. · A more comprehensive economic view:More recent economic studies have shown that cross-subsidization occurs when prices for a service are higher than would be charged by the next most efficient competitor, and the company still earns a normal profit. 17 A Model for Calculating Interconnection Costs in Telecommunications it possible to base the cost determination on forward- these costs to the generic LRIC, we obtain a total ser- looking costs and not on historical costs.The next sec- vice (TS) or total element (TE) LRIC. TSLRIC or tion provides a more extensive description of the TELRIC involves allocating pertinent12 joint and LRIC methodology. common costs to A. Finally,it is worth outlining how the cost allocation The LRIC Concept is implemented in practice.Two options are generally The LRIC methodology, also sometimes referred to as considered, although either of them can also lead to an the long run average incremental costs (LRAIC) alternative: methodology, estimates additional costs incurred in 1. Historical costs form a first option. The method producing a service, relative to the costs already therefore involves evaluating the costs on the basis incurred by producing a portfolio of other services. of their accounting values,13 possibly adjusted to The incremental costs of a service or element A some- take inflation into account. how represent the cost savings, which result from not 2. The forward-looking or current costs are the costs producing or not implementing A. In other words, the that would be incurred if the production system costs incurred to produce A over and above the portfo- were rebuilt on the date of calculation. lio of existing products are considered as the incre- mental cost. The first approach is often described as a top-down The long-run concept involves taking the costs approach and the second as a bottom-up approach. incurred in a long-term perspective. In the long run, This is even more essential if we introduce a consider- production fixed costs can be considered as being vari- ation of technical progress in cost evaluation. In fact, able costs.The long-run incremental costs of a service the production of A on the date t by an incumbent may or element A therefore represent all of the costs that not require specific new investment, but will result could be avoided if A were not produced or imple- from investments already made in previous production mented. Hence, the incremental costs include all the cycles. As a result, working with outdated technology costs directly attributable to A, whether these are vari- (the network architecture unchanged) or with a better able in stricto sensu (depending on the level of traffic at architecture, as available on date t, are options that can a given capacity) or fixed (making up the capacity). be considered. In retaining the historical production However,A can also make use of elements,services, system, we end up using historical costs or current or functions needed jointly with other services or ele- costs to value the incremental cost of producing A. ments. The incremental costs (even long run, Conversely, in retaining the most efficient current pro- strictly speaking) only take into account a portion duction system, we end up valuing the incremental of these costs in the case of joint costs (pro rata cost of producing A with current costs. For practical their incidence), and not common costs.11 Incre- reasons, an "average historical" architecture14 is gener- mental costs are important in the sense that they reflect ally selected.This point will be discussed in more detail the company's decision for producing A. Usually, in later. deciding to produce A, the firm expects that revenues generated by A should exceed the incremental cost We then assume that: incurred and provide an earning income to capital 1. The operator is an efficient operator that minimizes above the cost of capital. costs for a given production volume. Nonetheless, incremental costs, as we have just 2. The costs are current costs. strictly defined, are difficult to use in pricing access to As we have seen, this accounting and historical a service or network element, to the extent that they method involves breaking down the company's costs only cover part of the costs. Given that A also uses and allocating them among its different products.While other network's resources and is partly responsible for assigning directly attributable costs to respective prod- joint costs incurred by the firm, it is necessary to allo- ucts is a straightforward issue, allocating joint and com- cate to A a portion of these joint costs. By adding all mon costs is complex and sensitive.The difference between 18 Cost Modeling Principles the FDC and theTSLRIC method, using historic accounting to retail services and are not shared with the intercon- costs and technologies, therefore resides in the nonincorporation nection services. Inversely, interconnection services of nonrelevant common costs.Beginning with a total cost generate specific costs, which should not be included estimate obtained with an FDC approach,we can move in the increment costs (for example, co-location costs to LRIC, step by step, by removing layers of cost ineffi- for new entrant's equipment, costs for links between ciencies as shown in figure 3.2. the incumbent's network and new entrants' networks, Effective regulation of interconnection costs, there- cost of modifying the information technology [IT] sys- fore, introduces the following concepts: the specific tems, specific billing cost, interconnection manage- nature of costs and the pertinence of costs. ment service cost, and so forth).These costs should be The increment generally relates to a group of ser- passed on to interconnection customers only.To enable vices using the same production infrastructure. The the provider to recover the costs incurred, a two-part costs incurred from the provision of a specific service tariff scheme is usually applied. of the family should be determined within that frame- The concept of pertinence affects the handling of work. For example, a telephone company supplies joint and common costs. Joint and common costs can retail and wholesale services. The cost incurred from only be applied to the increment on the condition that providing retail services should be derived from the they are linked to it, either directly or indirectly. incremental costs incurred from providing wholesale Revealing a causal link takes for granted that an in- and retail telephony services. However, the costs for depth technical review of the cost structure is per- marketing retail telephone services are costs that apply formed. Common costs include research and Figure 3.2 Transition from Historical Accounting Costs to Economic Costs (LRIC) Fully Non- Economic Efficient TELRIC Current TELRIC Pertinent LRIC allocated pertinent lifetime supplier top-down costs FL joint and FL historical common adjustment adjustment historical adjustment bottom-up common costs costs costs FDC adjustment adjustment LRIC Adjustment Forward Looking Total Increment Note: FL, forward looking. 19 A Model for Calculating Interconnection Costs in Telecommunications development costs, costs associated with headquarters than the one that could be derived with an FDC expenses and the operator's operational structure, costs assessment!15 for staff members who are no longer in their positions · Depending on assumptions made, the TELRIC or are on leave, costs related to developing the brand method can lead to a relatively wide range of esti- name reputation or marketing, and costs associated mates. It is important to recall that directly attribut- with unused buildings.Among these costs, the regula- able costs result from the service or elements tor has to assess and decide which ones are pertinent to segmentation assumed in the model.Therefore, the be accounted for.Only pertinent common costs can be magnitude of joint and common pertinent costs allocated proportionally to interconnection services. depends on the segmentation refinement. Further- Figure 3.3 illustrates the distinction between these dif- more, allocating joint and common costs remains a ferent categories of costs. major hurdle,even though the use of more accurate In general,there is a risk in overestimating common distribution criteria can induce smaller cross-subsi- fixed costs. For instance, the former monopoly could dization between services or products. argue high common costs to "squeeze" out its com- Selecting specific TELRIC methods leads to some petitors in downstream retail markets by reporting to kind of arbitrage exercise, which is similar to position- interconnection some of the costs derived from com- ing of a cost cursor within bounded values. The cost petitive activities. In consequence, it is necessary to value (obviously high), which is supposed to be favor- ensure that the common costs attributed to the incre- able to incumbents,will be at one extreme;at the other ment are sound. In any case, the costs attributed should extreme will be the cost value (obviously low) that reflect the costs borne by the most efficient operators. favors new entrants. Two other considerations are In practice, a certain number of questions are raised advanced in regard to positioning this cost cursor: in regard to the theoretical framework discussed.These · "Theoretical" considerations limit the admissible questions are synthesized in the two points below: spread of cost values, by suggesting that the spread · Depending on the network services or elements be constrained by the implementation of a system considered, the TELRIC method can prove to be eliminating cross-subsidization, and that resulting less favorable to new entrants than the FDC estimates be below the separate cost of providing method. If a network section A is subject to sub- the same services. stantial depreciation, and if current costs for recon- · "Political" considerations related to the appropriate structing A are quite similar,or even higher than the economic signal are provided to market players and historical costs, aTELRIC can lead to a higher cost enable them to arbitrate between investing and pur- chasing the inputs from the incumbent (for exam- ple, the play or pay principle). Two concerns are, Figure 3.3 Cost Structure however, raised.The first concern is to ensure that input prices are not below the costs incurred by the Specific costs incumbent. The second concern is to ensure that Variable costs input prices allow for efficient allocation of resources. Nonspecific costs An excessively low rate for input can lead to cross- Fixed costs subsidization toward this service and send undesirable Increment signals to the market.An interconnection price below cost (in case of an efficient operator) could also endan- + Pertinent joint costs ger alternative infrastructure development, as that Joint costs Nonpertinent joint costs should make new investments less attractive and prof- itable. Conversely, an interconnection price above cost + Pertinent common costs (in case of an efficient operator) is also likely to bias the Nonpertinent common costs market by switching demand to infrastructures that are 20 Cost Modeling Principles actually less efficient, and does not enhance productive must, therefore, be equal to long-run incremental costs efficiency in the sector. More specifically, the incum- in order to maximize economic efficiency. LRIC pric- bent does not have the incentive to shift to a more effi- ing schemes are forward looking and provide better cient production process. incentive for static cost efficiency. In a dynamic frame- It is, therefore, up to the regulator, using cost mod- work, the impact of LRIC methodology in determin- els to determine the"correct"price level for this input. ing interconnection rates remains inconclusive (Laffont In so doing, the regulator collects and handles various and Tirole 2001). information and has to fix a certain number of the Tight access pricing regulation prevents the incum- model's parameters.These parameters and assumptions bent, controlling essential facilities, from extracting all are improved on, with use and time. the monopoly rents related to its dominant position Unfortunately,regulators and operators do not have over these resources. Although these regulations a good knowledge of the industry cost structure. encourage the efficient utilization of the resources, Therefore, theTELRIC bottom-up method represents they can also discourage further investments, subse- a decision-making tool that induces improvements on quently causing significant social welfare losses.This is procedures and processes implemented by regulators or the most probable risk that could be implied by inef- operators to collect and retrieve cost information.This fective access pricing regulation. also improves the quality of investment decisions for The LRIC methodology has been criticized by firms, as it naturally enhances the transparency of the several authors. Salinger (1998) and Laffont and Tirole interconnection services market. (2001) argued that LRIC regulation provides the reg- The following section reviews questions raised by ulators with a key tool to manage industry entry. It is, the LRIC method for determining interconnection therefore, crucial to ensure that, whenever a mistake is cost.The model deals with the issue of economic pric- made, it is made in favor of overinvestment rather than ing of interconnection services and does not cover underinvestment. issues related to co-location or management services In a long-run framework, firms can reassign their that are associated with interconnection. inputs according to input price and output.Whenever an output is unprofitable, firms can freely dispose cor- Advantages and Drawbacks of LRIC responding assets by selling them or by dispatching The following subsection summarizes the main advan- them to other activities. However, in the telecommu- tages and drawbacks of the LRIC methodology.With nications industry, most assets cannot be freely dis- these comments in mind, we expect the user to be able posed, because exiting an activity is costly.As shown by to outline the limitations of the LRIC methodology Hausmann (1996), when the regulatory framework and notice how its implementation can help in sorting allows entrants to exit and divest some of their assets, out and, probably, in clarifying a complex and sensitive and does not provide similar options to the incumbent, subject. entrants are less likely to invest in building their own One of the most effective features attached to infrastructure. LRIC pricing schemes is the sharing of the productiv- Furthermore, the value of most telecommunica- ity gains that the various market players could derive. tions operators' assets with current available technolo- In so doing, implementing LRIC pricing schemes gies is lower than their corresponding book value. impedes excessive profits by the interconnection ser- Consequently, entrants have no incentive to build their vice provider.The relevance of the LRIC methodol- own networks and will prefer using the incumbent's ogy therefore depends on the efficiency concept. In so infrastructure. LRIC can, thus, introduce inappropriate doing, interconnection rates are derived from a bench- incentives for entry, because of probable cross-subsi- mark provided by an efficient operator. Cost models dization from the incumbents to entrants.16 are, indeed, developed to simulate, with some accuracy, It is also worth pointing out that the determination the cost frontier that could prevail in a specific eco- of long-run incremental costs remains after all discre- nomic and market environment. Interconnection rates tionary. Salinger (1998) suggests that the use of LRIC 21 A Model for Calculating Interconnection Costs in Telecommunications is theoretically sound, but its implementation in prac- Traditionally, three possibilities are considered in tice is rather complex and could undermine the prof- regard to the network's nodes that handle the traffic itability of the incumbent's investment if poorly collection or termination: executed. Similarly, Valleti (2001) argues that access · When the interconnection is handled at a local charges based purely on LRIC are an appropriate switch, the service is termed local. benchmark when retail-level distortions are eliminated · When the interconnection is handled at a transit or dealt with effectively by other regulatory instru- switch, with the latter serving only a limited or ments. Consequently, when the incumbent's retail specified transit area, the service is termed single prices are not cost oriented because of universal service transit. obligations or delays in rebalancing retail rates, there is · When the interconnection is handled at a transit a need to add a uniform markup to the LRIC esti- switch,with the latter serving or providing access to mates, although doing that does not reflect any sound the overall network's nodes, the service is termed economic analysis. double transit. Another problem is related to the pertinence of the In principle, whenever a network's topology and essential facilities concept in the telecommunications architecture are compared, the collection and termina- industry. Celani, Petrecolla, and Ruzzier (2002) tion costs are identical.The equality is not sound,but is pointed out the following problems that are not prop- established in principle. In fact, the lack of reliable data erly addressed.What telecommunications market seg- on possible routing alternatives does not allow for fur- ments could qualify as essential facilities? What assets ther distinction of the cost of handling a terminating or can be considered as being essential facilities?This calls a collected call. for a sound methodology that enables disaggregation In other words, five interconnection services are of the access market segment in order to identify those generally considered: eligible for the essential facility concept. Obligations could then be limited to these market segments. Origin­ In developing countries, the key market players are Destination mostly multinationals. Regulators should therefore Interface Point Correspondent Network closely monitor how transfer pricing schemes are Local switch Local Transit switch Single implemented between affiliates and mother compa- in the area transit nies. Similarly, an appropriate database on cost infor- Transit switch in a Double Domestic mation should be developed to limit the scope of different area transit transit International International opportunistic behaviors. switch transit Definition of Interconnection Services Interconnection services are offered to operators Domestic transit is generally calculated as the dif- (fixed-line or mobile network operators) in order to ference between double transit and single transit.Inter- collect or terminate their traffic from or to other com- national transit is a surcharge applicable to peting operators. In general, four major categories of international calls. traffic are distinguished, as illustrated below: The following diagram shows the entry points, depending on the nature of the interconnection ser- vice. International service as assumed in the model is routed via an international transit switch providing an Origin­Destination Recipient of Interconnection Supplier of Interconnection entry point in the transit area covered. If the entry Traffic collection Traffic origination/termination point is located in a different area, the cost of domestic Traffic termination Traffic origination transit has to be added. 22 Cost Modeling Principles Figure 3.4 Entry Points of Interconnection Service tions, and (c) value-added services, for shared services offered by intelligent functions of the network.Trans- Domestic port services can be broken down in various elemen- transit tary services, according to traffic collection and Double transit Transit delivery points. A correlation can be established area International between these elementary services, and the services marketed by the operator. Such a correlation is illus- trated in the table below: International Table 3.1 Correlation Matrix between Retail Services Single and Elementary Services transit Transit area Local Subscrip- Long Entry point area tion Local Distance Supple- Intercon- Fixed local Exit point Connec- Communi- Communi- mentary nection Local area tion cation cation Service Service Retail sales service X X X X Added-value Definition of the Increment service X A telecommunications operator generally offers a wide Transport range of services, whether to its subscribers or to other service Local X X X operators.These services have been traditionally classi- Single fied according to their commercial nature. For transit X X instance, one usually differentiates local from long dis- Double transit X X tance and from international services.Added-value ser- Access vices are also differentiated from plain old telephony service X X services. However, another service classification approach is based on the use of network elements fur- ther described below. In general, there are three major As shown in the matrix, there is a correlation categories of network elements: between the operator's elementary services (repre- · Elements that are dedicated to end users (which are sented by horizontal lines in the first column) and ser- used solely by a subscriber). These elements are vices sold to end users (represented by columns in the generally found in the local loop network or access first line). The matrix provides, for each service, the network and include a subscriber's connection line cost components incurred for its provision. Of course, to the switch; the matrix is not exhaustive. A column representing · Elements that are shared among users (dedicated to leased lines could also be added to it. For instance, the users on a dynamic basis upon request).These net- provision of interconnection services involves trans- work elements are allocated temporarily to a sub- port services (local, simple, double transit). Conversely, scriber and are generally found in the core the provision of a local communication involves in telephony network. addition to transport services, the retail sale service. · Elements that are shared by users but are used for Furthermore,an operator may wish to implement a the provision of complementary or supplementary retail pricing strategy that is not necessarily cost ori- services (telephone card services or other services ented, at least, as an entry strategy to develop its busi- offered by intelligent network features). ness line. In such situations, the operator will likely opt The three services discussed above can be labeled for a pricing strategy featuring high connection fees as: (a) access services, for services dedicated to each and low rental charges. Sector regulation should pre- user, (b) transport services, for services shared among a serve the pricing freedom as long as it does not intro- network's users and offered by the core network func- duce a squeeze and distort competition (that is, as long 23 A Model for Calculating Interconnection Costs in Telecommunications as it does not eliminate new entrants from the market). mented and its maintenance do not have to be It is therefore important to define and ensure the con- adjusted to bear the flow of this additional traffic. sistency between the pricing of the inputs needed to · In contrast, transport services are affected by the provide retail services and the pricing of the retail ser- traffic resulting from interconnection services. vices. The table below provides an illustration of the As a result, the cost of access services must be cov- linkages to help process the consistency review. ered by retail service revenues. Unless these services The retail sales activity includes all the functions rates are rebalanced, there are risks for anticompetitive associated with providing the service and managing practices and cream skimming. In other words, market the customer relationship.The retail sales activity pur- deregulation imposes a certain cost orientation for chases all the horizontal services required to provide dominant operators' tariffs, and, consequently, implies a the final service that the end user consumes. In so price restructuring that eliminates the largest existing doing, it implicitly transforms respective cost elements pricing averaging.17 used into pricing components. In checking the cost The core network is used to provide a broad range structure of regulated services,the regulator must pay spe- of services. Apart from interconnection services, it cific attention to the squeeze that could materialize as a result facilitates the provision of retail services, including the of undue cross-subsidization of competitively provided services provision of other services such as leased lines. Leased by noncompetitively provided ones. The regulator should lines are either aimed at final customers (companies for also ensure that interconnection services, which are their corporate networks), or at other telecommunica- wholesale services, are priced below corresponding tions that operators can use to roll their own networks. retail services. The operator can also use leased lines for its own inter- As discussed earlier,the incremental cost methodol- nal consumption (to operate another network such as ogy distinguishes directly attributable from nonattrib- the telex network,packet data transmission network,or utable costs.It involves ascertaining,whether or not the even in many countries to carry the radio and televi- production cost of "horizontal" service is increased, sion programs). whenever a "vertical" service is added to the basket of It is also important to note that the core network other services produced. usually shares resources with other networks managed Considering the example of interconnection ser- by the operator, or with the access network whose vices, it is important to underscore the following three ducts are frequently shared in urban areas. It is impor- conclusions: tant to report to the core network portions of the joint · The provision of interconnection services does not costs incurred by other networks and, in particular, the modify the retail sales service,as the latter only con- access network. cerns sales to final subscribers. In conclusion, the increment to be considered for · The provision of interconnection services does not the calculation of interconnection service costs com- modify the access service, as the capacity imple- prises the core network and excludes the access net- work's elements that are dedicated to users. Table 3.2 Matrix of Inputs Needed to Provide Retail Services Retail Subscription Local Long Distance Supplementary Interconnection Sales Connection Communication Communication Service Service Retail sales activity X X X X Added-value service X Transport service Local X X Single transit X X Double transit X X Access service X 24 Cost Modeling Principles Application to African Telecommunications double the size of an average European country. Its Networks population density is four times less, implying higher rollout and maintenance costs for telecommunications African telecommunications networks are specific infrastructure than in densely populated countries. because of their early development stage. Excluding Furthermore, its average income per capita is 70 times South Africa, we can define a representative African lower than in EU countries,and the fixed teledensity is country (Afriland) with respect to its telecommunica- 100 times lower. The average consumption of tele- tions common specificities.To do that, we aggregated phone services is estimated at slightly more than 20 data for a sample of 40 countries and were able to minutes per inhabitant per year. On average, only 20 derive relevant common features, which are embodied percent to 25 percent of the population has access to a in this cost model.18 telephone within walking distance. In terms of infrastructure,the coverage of the fixed- The Average Size of an African Telecommunications line network remains rather limited and connects Network about 75,000 lines. The number of subscribers con- Afriland,as an African representative country,is charac- nected to the mobile networks exceeds the subscriber terized as follows: base of the fixed network and is approaching 100,000. The mobile segment is served by two or three opera- Table 3.3 Snapshot of Telecommunications tors offering essentially prepaid services (95 percent of Development in Africa in 1999 the consumers of mobile are prepaid). Furthermore, approximately 80 percent of the Low Income, stock of subscribers and 80 percent of the traffic is Low European concentrated in the capital city. On the whole, the Teledensity Union Average outreach of the network is limited to the most Situation Africa (40 (15 EU in 1999 countries) countries) Afriland Country Ratio important urban centers. Extending the coverage to Area 20,927,000 3,191,000 525,000 215,600 rural communities remains a major challenge for the km2 km2 km2 km2 0.41 years to come, as rural areas are also those with the Popula- 600 376 15 25 tion million million million million 1.67 lowest income and population density.As a matter of Popula- fact, 80 percent of the stock of telephones serves only lation 20 percent of the population. Consequently, if the Density (inhabi- teledensity is about 2 percent in the main urban cen- ants/km2) 29 117 29 117 4 ters of Afriland, it is only about 0.125 percent in the GDP/ rest of the country, though the overall national tele- capita 300 a 21,000 a 300 a 21,000 a 70 TotalGDP density is 0.5 percent. Afriland is characterized by a (billions of very low level of telephone service demand, which is euros) 180 7,900 4.5 525 117 a direct consequence of the limited income per Stock of main capita. lines 3,140,000 200,000,000 75,000 13,300,000 180 The implications in terms of network architecture, Tele- topology, and costs are: density (fixed) 0.5% 53.2% 0.5% 53.2% 100 · The transit network is almost nonexistent. Usually, Density the transit functions are implemented in local of lines/ km2 0.15 61.8 0.14 61.8 440 exchanges or switches and are located principally in the capital and in two or three main cities. Source: BIPE 2001. · The transmission network has an unusual topology. When SDH rings are set up, they are used to con- In terms of size,Afriland is larger though less pop- nect telephone switches and their RCUs within the ulated than an average European country. Afriland is same urban center or region. 25 A Model for Calculating Interconnection Costs in Telecommunications · The networks still reflect choices of technology ment's universal access policy, whether or not to made in the 1990s. The number of small digital include the cost related to TDMA systems. switches operated is higher than what could be achieved by using modern technology.With the best Nodes and Links technology available today, most of these networks A thorough description of the core network is needed could be equipped with only a couple (one or two) before its modeling is attempted. This description of digital switches.The scope of inefficiencies result- involves determining the types of nodes and the nature ing from this technology legacy is rather great. of the links between the nodes. In telecommunica- · To reach remote rural areas, subscriber concentra- tions, a node is generally characterized by a switching tion systems (TDMAs) are installed.These systems function, while the links form the transmission net- serve from 8 to 256 subscribers and use time divi- work. In developing the proposed cost model, we have sion multiplexing technology, known as TDMA. assumed that an average African network has five types TDMA cost per line is particularly high.19 of nodes and five types of links. · Fiber-optic systems are not yet massively rolled out. These nodes are: In many cases, the demand does not justify the roll- · Subscriber RCUs, which do not have independent out of broadband transmission systems; however, switching command capacity,except in the event of more and more operators are beginning to replace a breakdown of the connection link with the host their microwave transmission links with fiber-optic exchange. In the model, the small rural exchanges cables. that are still in service and electromechanical · Finally, in some countries, domestic satellite net- exchanges have been considered as RCUs. The works are operated as a means of extending the rationale is that digital RCUs are more likely to coverage to isolated areas (rain forest, islands, replace these old technologies. desert). · LSs to which RCUs are connected. Some of these The proposed model seeks to account for all these exchanges have a transit capacity. specific features so that the cost calculation can be as · "Pure" transit switches (TSWs),20 which are rare in close as possible to the reality from which should be the African context, but are integrated into the derived the industry cost frontier. As an illustration, model for the sake of completeness. TDMA systems have been included in the increment, · ISs or international transit exchange. These though one could argue against inclusion. We have exchanges are rarely taken into account in the mod- assumed,however,that telephony circuits implemented eling carried out in industrial countries because the through these systems are not dedicated to a user, but international transit node is not differentiated from are shared among all users of the network, including national transit21.This is not yet the case in Africa those connected to the RCU.As a result, we consider where the international segment is not yet unbun- that the traffic generated by servicing the interconnec- dled from the incumbent's monopoly. tion demand (termination, origination) contributes to · Finally,TDMA systems lead to two types of nodes: the sizing of these systems. ­ CSs, which are generally coupled with associated Nevertheless, in costing the interconnection ser- switches to manage the signaling and the interface vice, the model provides two options: with the transmission systems component; · A cost "with TDMA systems" included in the ­ TSs, which are the base stations installed in rural increment. communities to provide the last mile connection · A cost "without TDMA systems" included in the to subscribers. increment. Similarly, there are five types of transmission links: The results obtained when including TDMA are · RCU-LS links are the transmission links between very different from the ones obtained when it is not RCUs and their host exchange. included.We suggest that the regulator decides, based · LS-LS links are the transmission links between two on national specificity and according to the govern- different LSs. In some cases, these links could also 26 Cost Modeling Principles involve LS links to TSs, to the extent that the tran- Transit sit function is implemented. Because of the small size of African networks, local · Specific links between transit centers (TS-TS). exchanges (LSs) generally provide transit function. · Links to ISs (between LS and IS). In certain cases, Whenever the network becomes larger in size, in the IS function can also be implemented in an LS. terms of coverage, local areas are generally grouped · Internal links within TDMA systems, or the links into transit areas. The single transit feature involves between the central stations and terminal stations. collecting or delivering traffic in the transit area. As a result, we obtain the matrix representation Conversely, double transit involves collecting and ter- provided below: minating traffic beyond the initiating transit area. When a network does not have a transit switch, the Table 3.4 Matrix of Transmission Links distinction between local and single transit becomes irrelevant. Nodes IS TSW LS RCU CS TS For the purpose of the exercise conducted here, we IS will assume the hierarchy between interconnection TSW To IS TSW-TSW TSW-LS LS LS-LS RCU-LS (Local services (intra-LS, single transit, and double transit) is RCU link) established along the incumbent's local rates zones or CS CS-TS derived from the administrative or government terri- TS torial organization (municipalities, counties, districts, provinces, regions). It should be noted that local links (LS-CS, RCU, The first option links the interconnection services and the central station in a TDMA system) are not hierarchy to the retail pricing structure of the incum- often taken into account because of the frequent co- bent,while the second option,although more arbitrary, location of these nodes. does not. Notes 1. General interest obligations can be imposed, providers, whereas it is itself in competition 10. Mark A. Jamison, available at for example,access to emergency numbers. with these in the same market downstream http://bear.cba.ufl.edu/centers/purc/ 2.Article 13 of the draft"Directive on a Com- ("Directive on Access to Electronic Com- primary/jamison/Pricing.pdf. mon Regulatory Framework for Electronic munications Networks and Associated In- 11. Joint costs refer to costs incurred by two Communications Networks and Services." stallations, as well as Their Interconnection." or several products in the same production 2001. Com 380, European Commission. 2001. Com 369, European Commission). process, in a constant proportion.We talk of 3. This measure supposes a prior definition 6."Directive on Access to Electronic Com- common costs when the costs are incurred of the relevant market, that is, the one in munications Networks and Associated In- by several products and remain unchanged which power has to be measured. In general stallations, as well as Their Interconnection." regardless of the relative proportion of these it is a market that brings together products 2001. Com 369, European Commission. products (the salaries for operators' head- among which significant substitutions and 7. Strict economic definition. quarters functions),that is,when a product is complementarities are present or possible 8. For example, notions of opportunity cost. offered, the second product is produced by from the customer point of view. (In our context, a cost that enables the party the same production without a supplemen- 4. Definition from the draft "Directive on selling the resource to a third-party "whole- tary cost. Access to Electronic Communications Net- saler" to obtain remuneration equivalent to 12. That is, which demonstrate a causality works and Associated Installations, as well as that which it would have obtained by selling relationship. Their Interconnection." 2001. Com 369, it in the end market.This cost is therefore the 13. These are also referred to as embedded European Commission. sale price less the cost of the retail price. It is costs. 5.The regulator can impose an obligation on equivalent for the supplier to sell on the in- 14. By taking over the topology of the his- a vertically integrated company to make its termediary market or on the final market. In torical network, that is, the same intercon- internal wholesale and transfer prices trans- some ways, it is an access cost for resellers.) nection equipment locations inside the net- parent in cases where the market analysis re- 9. This is assumed to be undertaken here in work (switching, concentration, distribution veals that the operator concerned supplies year 1 (otherwise,it would be necessary to take and so forth).This is,thus,the scorched node facilities that are essential to other service the discounted sequence of investment costs). option, which involves retaining the net- 27 A Model for Calculating Interconnection Costs in Telecommunications work's real hierarchy and the current traffic When incumbents make investment deci- vices and a possible temporary "access management rules. Imagining an optimum sions, technology that the regulator may deficit" (until the price restructuring is network would lead to a certain number of consider to be best may not be available. achieved) must compensate for the pricing criticisms concerning its feasibility, and its 17. Within this context, for operators who equalization in force. possible operational capacity, on the impact have not undertaken the required price re- 18. This roughly means African countries of this virtual architecture on other prices. structuring for whatever reason, there may whose gross domestic product (GDP) per 15. This explains that in several countries, be a continued need to subsidize some loss- inhabitant is lower than EUR 1,000. the FDC method was considered as the most making services from profitable services and, 19. These radio concentrators are presented favorable to new entrants in terms of access on an almost general basis,a subsidy from in- in appendix 3. to the local loop, where civil engineering ternational and long distance calls, to the ac- 20. Also called a tandem switch. costs are dominant,although less favorable in cess and/or local communication service. In 21.This has to do with market liberalization, terms of interconnection. this case, two phenomena have to be distin- which has now enabled operators to estab- 16. LRIC discriminates between incum- guished: the interconnection service must lish interconnection point of presence across bents and entrants in favor of the latter. measure the value of these cost-oriented ser- the borders. 28 4 Modeling Principles The model reconstructs a network, as would be done to invest or purchase services (pay or play) should be by an efficient operator using a forward-looking LRIC made in the light of the prevailing economic situa- methodology. The reconstruction of the network is tion. done in line with the realities of the Sub-Saharan African environment, as discussed earlier. Before dis- An Efficient Approach cussing the modeling in detail, let us recall the follow- The cost modeling must reflect decisions made by an ing principles. efficient operator, producing the increment services at the best cost while taking into account the technolo- The Modeling Principles gies available. As a result, the model simulates a net- work that, at a given production level, minimizes the A Long-Run Approach total cost by using best available technologies. The LRIC method adopts a long-run approach. The This requirement raises a certain number of ques- reconstruction of all network elements is assumed at tions with respect to the network architecture. As we year 1 and includes the increment. have already stated, an incumbent inherits a network topology, which is largely determined by successive A Forward-Looking Approach generations of technologies.Therefore, in dealing with Selecting a forward-looking approach means consider- the efficiency problem, there are two possible ing both the best technologies available and their cur- approaches: rent costs. From a pragmatic viewpoint, this comes · Modeling a network providing the expected ser- down to considering the digital technologies available vices after optimizing its topology and architecture on the shelves today. Consequently, when modeling, and eliminating the technology legacy.This option we replace old technology with modern "equivalent" is called "scorched earth." technology, which is more efficient and cost-effective. · Keeping the existing network's topology (the loca- For the transmission system, forward looking tion of network nodes), while reconstructing the implies selecting SDH systems over fiber-optic cables. nodes and links with the best available technolo- Although fiber optic is considered to be the most flex- gies.The result is, in a way, a topologically identical ible, efficient, and economic technology in regard to network.This option is called "scorched node." It is the bandwidth unit cost, it is not yet rolled out system- the one selected in the proposed cost model. atically in Africa. If the resulting topology is clearly nonoptimal as it As regards the costs to be taken into account, these would be recommended to ensure production effi- have to be estimated at the current acquisition price, ciency, then the regulator can prescribe a better net- and not at book value.As such, the decision whether work configuration. 29 A Model for Calculating Interconnection Costs in Telecommunications Beyond the choice of the best technologies with an services. In some cases, top-down models are also existing network structure,the question of efficiency as implemented with current costs. regards network operation also arises.Digital technolo- · Bottom-up cost models imply the development of gies, in general, and, more precisely, current network engineering and economic models in order to cal- monitoring and management systems, make it possible culate the costs of network elements used to pro- to downsize factors of production, buildings space, vide particular services, assuming an efficient labor, and so forth. Regulators and operators must operator that uses best available technologies. agree on the optimum level of efficiency. LRIC mod- els should not, under any circumstance, take into In principle, both methods should lead to the same account excess staffing from previous management result. In fact, this can only happen if the same assump- because market deregulation should be a strong incen- tions are made for operation efficiency and depreciation. tive for the incumbent to improve its efficiency. The proposed cost model belongs to the second category. An Economic Approach, Not an Accounting Approach The Working Units To obtain cost annuity, it is necessary to transform the The model has to determine unit interconnection long-run incremental costs incurred, including those costs. It is, therefore, necessary to select a metric sys- resulting from investments, into an annuity. An tem.Traditionally,the traffic flow through the core net- accounting approach would have led to considering work is measured in minutes.The cumulated duration yearly depreciation installments based on accounting of calls, measured in minutes, is considered to be the lifetime of respective equipment, calculated according important parameter in determining costs. However, to fiscal criteria (linear, tapering, or accelerated depre- some interconnection prices and some cost models ciation).The cost annuity is determined here using an also take into account the number of calls. economic approach. Economic-oriented costs are In practice, network elements also handle the fol- assumed to be the effective tool for regulating industry lowing noninvoiced traffic: (a) the call set-up time, (b) entry and investments. It is assumed that entry or the closing time, and (c) the time spent in handling investment decisions are made on the grounds of prof- unsuccessful calls.The sizing of some network elements its or reasonable return on investment expected from therefore depends on the number of calls conveyed. the activity. In the proposed cost model, investment This is particularly relevant for switching elements that costs are converted into annual economic costs, as have to handle call attempts.1 The number of calls described in appendix 1. essentially continues to determine the cost incurred by To proceed further with the discussion, we now the temporal occupation of noninvoiced network ele- need to introduce the cost of capital concept. As ments (waiting time and unanswered calls).The cost per described in appendix 2, the cost of capital represents minute is, by convention, the representative unit cost. the cost incurred by sponsors or promoters in mobiliz- ing financial resources.The cost of capital concept fac- The Model Structure tors in technology progress, economic, and country-specific risks. The logical structure of the model is relatively simple: · The model begins by proposing a nomenclature of A Bottom-Up Approach network elements (nodes and links). Two main alternative approaches can be used to esti- · Each service uses these elements in different pro- mate the long-run average incremental costs: the top- portions. On the basis of routing factors, the model down model and the bottom-up model. These two calculates the total load supported in traffic minutes approaches can be summarized as follows: for each element. · Top-down cost models rely on the accounting data · The model sizes network elements specifically for and allocate costs to different services on the basis the transmission system, within the confines of the of the causality correlation between the costs and topology the user chooses. 30 Modeling Principles · The model then aggregates network elements costs. volume of traffic handled.These costs per minute are · The model finally calculates the interconnection then added up to reflect the network elements costs, depending on what network elements are involved in handling a specific interconnection service used to supply the service. request.They are then adjusted by the hourly gradient The network is made of network elements. Any applied to retail services to obtain the final intercon- communication or any interconnection service (which nection rate. is a special form of communication) uses, on average, x To sum up, the model reconstructs the costs of the times each of these elements (with x varying from 0 to network at two levels: a few units at a maximum). The x factors are called · At the level of nodes (switching elements), invest- routing factors. This allows for the calculation of the ment costs are structured in fixed and variable com- traffic flow per network element.The model calculates ponents according to BHE (business hour Erlangs the investment cost incurred to satisfy the traffic transformed into 2 Mbps) and to the number of demand and derives respective operating costs. subscribers, respectively. · At the level of links, calculations are done at two Investment Operating sublevels: costs costs ­ A transmission level at which electronic trans- Attributable costs mission equipment (mainly on SDH rings) is Common costs designed and sized; ­ An infrastructure level at which the link substra- The operating costs consist of two terms: tum is designed. Generally, the infrastructures · A cost subcomponent, proportional to investment comprise three types of technologies: trenches, costs and reflecting the cost of maintenance and microwaves, and satellite links. direct operation of network element (spare parts, The trenches are broken down by geo-type (urban, equipment section of preventive and corrective suburban, and rural) corresponding to different bury- maintenance, energy consumed). ing techniques (wrapped trenches,ducted,buried).The · A cost subcomponent for the staff allocated to microwaves are characterized by the nature of their operations. masts (light, medium, or heavy). For small and spread-out networks, the staff cost The links are sized by the traffic at peak times can hardly be evaluated as a percentage of investment expressed in capacity (Mbps) and are sized to handle costs. For the sake of accuracy, the user is requested to the switched telephone network traffic,as well as leased state a reasonable number of staff required to run the lines bandwidth.Assumptions regarding the sharing of network.The model then uses unit labor cost input to certain elements of basic infrastructure with other net- calculate the total of operating costs and divide it works (for example, access networks) subsequently among the number of network elements. allow for the sharing of costs,which are not fully borne Obviously, investment costs are subject to detailed by the core network. calculations. Based on an estimate of the volume of To sum up,the model computation modules are pro- traffic to be handled by each network element, and the vided in the following 21 sheets structured as follows: unit cost for each element, the model determines the · A menu sheet for the user navigation. total investment cost. To achieve this, the amount of · Twelve sheets forming the core piece of the model traffic assigned to each network element at peak time is described afterwards. used for its sizing. Using the unit investment cost per · Four sheets specific to mobile networks (whose network element, the model calculates the total invest- results appear on the general results sheet). ment cost per element. The operating costs are allo- · One sheet on the sensitivity of the model to some cated to the investment unit costs to obtain a global parameters. cost per element. A unit cost per minute handled is · Three specific management sheets (two sheets for then obtained by dividing the global total by the total the publication of fixed and mobile reports and a 31 A Model for Calculating Interconnection Costs in Telecommunications sheet to manage the two languages and the default tion from the operators and sets up procedures for values). the ratification of the information supplied (for The 12 sheets in the general model break down as example, by means of accounting audits, establish- follows: ing ABC, reviewing investment invoices). · Four sheets to gather the assumptions. · Focusing the benchmarking on a limited amount of · One sheet to calculate the traffic and sizing of the sensitive information used as inputs for the decision network elements. process. As a result, the regulator shifts away from · Two other sheets to size the transmission and infra- standard benchmarking of tariffs observed between structures. countries to benchmarking key input or parame- · Three sheets to calculate costs (switching, transmis- ters. sion, and infrastructures). · Providing the opportunity to the regulator to · One sheet to recapitulate the total costs (including improve on its information and knowledge about the common costs) and to calculate the unit costs network architecture and topology. With a better per minute and element. understanding of the network structure, the regula- · One sheet to present results. tor can refine procedures implemented to collect The core of the model is synthesized in the follow- information from regulated operators. In fact, data ing diagram: collection must be executed according to formal and recurrent procedures. In conclusion, it is important that regulators ensure Figure 4.1 Cost Model Architecture the confidentiality of the information collected from operators. Despite most national regulations that give regulators inquiry powers, additional attention is Routing Unit Hypotheses Demand Network Factors Costs needed to ensure and preserve the confidentiality of Demand Tech FactRout UCosts the collected information.In general,newly established Traf c Traf c regulators do not yet have sound and effective safe- Capa guard measures ensuring the confidentiality of infor- Transmission Infrastructures Sizing mation collected from operators. Unless such measures Capa Eltr Capa Intra are implemented, it is likely that regulated operators Cost Switching Transmission Infrastructures Costs Sw Costs Tr Costs Infra will remain reluctant to fully disclose their strategic Totals Results information. Result tot results Like other costing methodologies, LRIC models also have consequential limitations, which are related to the wide scope of assumptions that foster their Applications of the Cost Model development. A system of check and control handled in collaboration with the industry can limit the arbi- In practice, the cost modeling exercise enables regula- trariness of the most sensitive assumptions.The follow- tors and operators to improve their knowledge of the ing are likely to be among the most sensitive model industry cost structure and its efficiency frontier. Of parameters that should be reviewed: course, the challenge is managing the information · The selected network topology (number and loca- asymmetry that characterizes the regulator and the tion of nodes). regulated firm relationships. The following are the · The percentage of traffic at peak hour and other three main advantages that cost models can provide: demand data critical to sizing the network ele- · Providing a catalytic effect on cost information col- ments. lection and retrieval by both the regulator and the · The traffic growth as predicted by the industry. regulated firm. To run the model, the regulator · An estimate of the staff number that could be con- draws up a comprehensive list of required informa- sidered "efficient." 32 Modeling Principles · Major cost drivers specific to the country and how they impact the investment costs in comparison with international best practice. · The proportion of joint and common costs. · The problematic decision to factor in TDMA con- centrator systems in the increment leading to the calculation of interconnection cost. A contradictory discussion on the model assump- tions would most likely improve on the accuracy of the range of cost estimates provided. This accuracy will improve as procedures prevailing to data collection also improve. It will also improve the soundness of the argument against the theoretical criticism of LRIC models. Note 1. As a matter of fact, for a long time the number of call attempts during peak time (business hour call attempts [BHCA]) was used to specify the overall processing capac- ity of switching exchanges. In view of the enhanced processing power of available processors and the resulting decreasing cost of memory, the number of calls processed at busy hours is no longer a relevant criterion to effectively differentiate switching systems. Vendors now agree to consider the total du- ration of calls (measured in Erlangs) and the number of subscribers accessing the node as the most important engineering parameters for switching exchange design. 33 5 User Guide The cost model includes a friendly interface. It was Once the language is selected, the developed as a Microsoft 2000 Excel file that contains Command/Menu sheet is shown: several folders. The model depends on several macro commands written inVisual Basic. In other words, the program cost model is less than 1 mega-octet in size; hence, it is easily portable. After launching the program, the user can choose whether or not to activate the macros links.Activating the macros is essential for the Menu sheet functions. This sheet includes three sets of commands: · On the left is a column allowing access to the Assumptions menu. · In the center is a column providing access to the Options menu. After the macros are activated, the user is asked to · On the right is a column allowing the user to dis- select the language.There are two options: French or play or print out results and test their sensitivity. English.The model is then presented with captions in IMPORTANT: Save your working file in a the selected language. different name. 34 User Guide Filling in the Assumptions It is not advisable to attempt modification of the model program orVisual Basic macros,unless one has a The assumptions are entered on five sheets, which are good knowledge of computer programming. The accessible via the Menu buttons in the left column of entire range of sheets and all the formulas are accessi- the Menu sheet: ble.The user can follow a step-by-step overview of the · Demand assumptions. model's computation process. However, modification1 · Network assumptions. outside the sheets that are accessible via the Menu · Routing factors. sheet (and the dark green boxes on these sheets) may · Cost elements assumptions. alter the model's performance. · Specific assumptions for the mobile network. The color code used to differentiate cells is as fol- Demand Assumptions lows: The Demand sheet is accessed via the Demand · Blue cells: exogenous assumptions to be filled in by assumptions button on the menu (see below). the user. These assumptions characterize the net- work and are essential for the modeling. Traffic Data · Light green cells: default information that can be The incumbent's traffic has to be broken down into 11 modified.These assumptions can be shared by dif- traffic categories.The total traffic in minutes has to be ferent networks. placed in the first column. In the second column, the · Light blue cells:default values that can be modified. average call duration in minutes (with at least 1 or 2 These default values are computed by the user and decimals) is provided by type of calls.The total number are, generally, presented in a table on the right. of calls is then calculated in the third column. Finally, However, to run the model with provided default the predicted growth rate applied to the traffic volume values, the user should avoid modifying the filled (in minutes) that is used to measure the network is matrix that contains default routing factors. entered in the fourth column. · Dark green cells: result cells, which under no cir- The traffic growth rate should remain moderate, to cumstances should be changed. avoid major oversizing of the network.The cost calcu- Margin for growth (%) Existing traffic over Number Average length Number of Number incumbent's network of minutes of calls successful calls of minutes Local telephony 0 Internet calls 0 Long distance 0 International * incoming 0 * outgoing 0 Calls to mobiles * domestic 0 Calls from mobiles * domestic 0 * international 0 Interconnection * local 0 * long distance 0 Others (Kiosk. switched X25 etc.) 0 Total 0 0 35 A Model for Calculating Interconnection Costs in Telecommunications lation is done on a yearly basis. As a result, the model Retail rate Rates considers the predicted increment of the traffic for the gradient (ratio) Traffic (%) 100 = peak Gradient year under review. Peak 100 The following input is needed for the calculation: Off-peak Weekend Other 1 Traffic statistic Successful calls Unsuccessful calls Other 2 Average 0 Average call set-up time in seconds (time to answer) 15 30 Percentage of successful calls 75% Traffic in busiest hour of the year (as a percentage Subscribers of the total) 0.00040 Data on subscribers and their connection mode are requested in the second part of this sheet.First,the net- First of all, the average answering time for success- work topology is filled as follows: ful and unsuccessful calls is taken by default at 15 and 30 seconds.The percentage of successful calls is taken at Node 75 percent.These values can be modified. information RCU LS TSW IS CS TS Total The percentage of the total traffic at the peak hour Number of nodes (business hour) is used to size the network's capacity. (total) 0 This value is estimated by default at 0.04 percent Nodes linked by (= 0.00040), but can be modified. It is considered that satellite 0 there are 335 peak days in a year (11/12 of the year), Other with on average 7.5 peak hours per day.This gives in all nodes not linked 2,513 peak hours per year.The total peak hours only through represents 1/2513 of the total annual traffic,hence 0.04 a SDH percent. The user should note that this variable is ring 0 Nodes on extremely sensitive and exerts a direct impact on the SDH rings 0 0 0 0 0 sizing of the network. The "gradient," which is related to retail services pricing schemes, should be calculated.The model pro- The first line of this table contains the total number poses five pricing levels: a peak rate, a reduced rate for of nodes in each category.We will count: off-peak hours, a reduced rate for weekend rate, and · RCUs: the total number of remote concentrator two other reduced rates to fit with more refined pric- units included in the network. ing schemes, if needed. · LSs: the total number of local exchanges equipped In the first column, the breakdown of the total traf- with their own processing and command units. LSs fic is entered as a percentage for each pricing level dis- may also perform domestic or international transit cussed above. In the second column, the level for each functions. pricing level is entered in relative terms, with the peak · TSWs: transit exchanges, dedicated exclusively to hour rate being the reference. For example, if the peak handling domestic transit. hour rate is 60 cents and the reduced rate is 40 cents, · ISs: international exchanges, dedicated to interna- we will enter 40/60 = 0.667 in the second box of the tional transit. second column of the table. · CSs: central stations for TDMA systems, used to The gradient is calculated in column 3 and will be extend the network coverage to rural communities.2 used to calculate interconnection rates in these differ- · TSs: terminal stations in TDMA systems, used to ent pricing ranges. connect about 40 subscribers. 36 User Guide The second line of the table deducts the nodes The table also includes information on the stock accounted for in the first line,and which are connected of subscribers (connected capacity), and differenti- through the domestic satellite system. In general, only ates those connected to RCUs from those connected RCUs or LSs are connected by satellite. The other to LSs: nodes of the network are assumed to be connected to · The first column contains the total number of net- each other through SDH rings. However, these rings work subscribers. are not always rolled out for several reasons.The third · The second column contains those connected to line of the table deducts the nodes already accounted nodes linked by satellite. for, and which are not connected by SDH rings.This · The third column contains those connected to represents an arbitrage in terms of efficiency that regu- nodes that are neither linked by satellite or SDH lators should assess and decide upon.These nodes will rings. be connected by microwave, at a higher cost than that · The fourth column contains the total number of of equivalent SDH rings.The last line then calculates subscribers connected by TDMA systems. the number of nodes interconnected by SDH rings. The following table shows data relating to the sub- Leased lines Number Margin for scribers' connection: growth (%) Analogue leased lines (64 kbits/s equivalents) Digital leased lines Form of subscriber connection (64 kbits/s equivalents) Percentage (%) o.w. of installed Installed TDMA Number of local tariff zones lines: capacity systems Number of LS with ­ Remote Concen- transit capabilities trator Units and Number of TS interconnected rural switches to mobile networks ­ Local switches Ratio od traffic from TDMA Total 0 subscriber to non o.w. % on local TDMA subscriber switch with transit Share of direct long distance capabilities calls between LS o.w. % on local switch % of direct routes between without transit LS with transit capabilities capabilities Percentage of o.w. o.w. non- o.w. First of all, the user enters the number of 64 kbps subscriber lines Used linked by linked by TDMA connected to: capacity satellite SDH systems equivalent analogue or digital leased lines.The user also ­ Remote Concen- provides the predicted growth rate to be taken into trator Units and account for network sizing. It should be remembered rural switches ­ Local switches that all the links using the core network must be taken Total 0 0 0 into account: · Telegraph and telex lines. · Leased lines provided to private or public users The first two lines should provide the installed (banks, transporters, civil service departments, capacity of the RCUs and LSs.This total appears on police, army). the following line, beside which the total capacity of · Leased lines provided to third-party operators the TDMA systems is indicated. The two following (mobile networks). lines enable the user to enter the proportion of sub- · Leased lines used by the operator to set up data scribers' lines connected to RCUs and LSs. It also transmission networks (X25, Frame Relay, IP net- allows for access to the proportion connected to LSs works). with transit features. Doing that improves the accuracy · Leased lines eventually provided to carry TV and of the routing factor estimates. radio programs. 37 A Model for Calculating Interconnection Costs in Telecommunications This is followed by additional information on refin- Main Assumptions forTransmission Links ing the default calculation of the routing factors: In transmission, a link is the media over which infor- · Number of local pricing zones. mation is conveyed between nodes.The information is · Number of LSs with a transit function. conveyed through different routes that may exist over · Number of exchanges (LS-TS) where there is an transmission links connecting the nodes.The number interconnection with mobile networks. of routes that can be created between nodes only · Ratio of the average traffic for a TDMA subscriber depends on the network security constraints. relative to a non-TDMA subscriber (generally The first step is to determine the number of 2 lower than 1). Mbps channels that are needed to carry the traffic flow. · Proportion of the long distance calls between LSs Consequently, the user is asked to provide information through direct routes. characterizing the transmission routes. · Proportion of direct routes between LSs with a transit function. RCU- LS-LS- LS TS to IS CS-TS Main Technical Assumptions Utilization level of This second ("Tech") sheet containing complementary transmission elements (%) 60% 70% 80% 50% assumptions concerning the network is accessed via Erlangs per circuit 0.5 0.6 0.7 0.6 the Network assumptions button on the menu. This Average length of sheet enables the user to enter important assumptions transmission routes across all geo-types on the network structure. in meters Main Assumptions on Switching Exchanges · In the first line, the load of the transmission ele- We start with a small number of assumptions whereby ments is provided as a proportion of the total capac- default values are proposed: ity available. · In the second line, information on traffic expressed in Erlangs is provided. Co-location of switches The information provided above is used to com- Percentage of tandem pute the size of the transmission network and the switches co-located with number of ports for each switching exchange. local switches 50% Maximum number of The average length of transmission routes across all nodes on a ring the geo-types is then filled.The average route length is Maximum number of the means of length routes between two nodes. nodes on a ring 16 The following table portrays the transmission net- Utilization level of switching nodes (%) RCU LS TS IS work structure for SDH systems. On the right, band- Used capacity (Erlangs) 95% 95% 95% 80% width capacity used for each link has to be provided. The first assumption relates to the percentage of Distribution of traffic links transit switches co-located with LSs so that the host Mix of STM systems buildings are not counted twice. If the network does in a fully SDH not include transit switches,the assumption is not used. transmission network The second assumption refers to the maximum RCU-LS LS-LS-TS to IS number of nodes connected by an SDH ring. The STM 1 default value is 16 nodes. STM 4 STM 16 STM 64 Total = 100% 0% 0% 0% Note:These may be re-allocated to meet required capacity. 38 User Guide The second indicator shows the average distance Box 5.1 Erlang between regenerators on SDH rings.The default num- An Erlang is a telecommunications traffic measurement metric. ber is equal to 64 km.The load factor of leased lines is For example,if a group of users makes 30 calls in 1 hour and each taken at 100 percent. call has an average duration of 5 minutes, the total traffic flow is The diversity indicators make it possible to take expressed as follows: into account the specificities of the real topology and their implications, with respect to an optimized topol- Minutes of traffic = number of calls x duration = 30 x 5 = 150 ogy retained by the model. Traffic in Erlang = 150 / 60 = 2.5 Main Assumptions on Infrastructures Agner Krarup Erlang was born in Linborg, Denmark, in 1878. He This part of theTech sheet captures all the assumptions was a pioneer of telephone traffic analysis. He proposed a formula for calculating the proportion of callers served by an exchange needed to determine the size of the infrastructures. who are compelled to wait for their turn. In 1909, he published his The first table captures the total length of infra- first result, titled The Theory of Probabilities and Telephone Conversa- structures per type of technology: tions. His work earned him worldwide recognition, as well as that of the British General Post Office, which endorsed his formula. He worked for 20 years at the CopenhagenTelephone Company until RCU- LS-LS- Length (meters) LS TS to IS Total his death in 1929. During the 1940s, the Erlang became the Total length of trench accepted unit for measuring telephone traffic. by transmission link (meters) 0 Total length of micro- wave by transmission Routes are distributed according to their band- link (SDH closure) 0 width needs. That makes it possible to size the elec- Total length of micro- wave by transmission tronic transmission elements.These indications can be link (non-SDH) 0 adjusted to the maximum number of authorized nodes Total (meters) 0 0 0 per route. These calculations are carried out in the Capa ElTr sheet. The following are additional default values used in The model distinguishes the total length (in the computation of the network size: meters) by type of: · Trenches. Average length between microwave · Microwave networks closing SDH rings. towers (meters) 40,000 · Microwave networks supporting non-SDH links. Distance between regenerators (in meters) 64,000 Utilization level of leased lines 100% The cost distribution between LS-LS-TS and TS- Diversity for STM multiplex and LTES TS is provided: (percentage) 15% Other equipment LS-LS TS-TS Transmission cross connects per IS 2 Breakdown of transmission and Diversity for regenerators 2 infrastructure costs between single Number of repeater stations of TDMA systems and double transit 75% 25% Utilization level of satellite transponders (%) The first indicator shows the average distance Trench specificities are provided: (a) wrapped between installed microwave towers (in meters). The trenches in urban environment, (b) ducts in suburban default value is equal to 40 km.This length depends on environment, (c) buried trenches in rural environment. the relief, stretches of water, and so forth. The user is asked to fill in the distribution that is relevant: 39 A Model for Calculating Interconnection Costs in Telecommunications Mbit/s. For non-SDH routes, the user is asked to fill Proportion of total length of trench in in information on the number of routes per link type each geo-type (%): RCU-LS LS-LS-TS to IS equipped with microwave, and the distribution over ­ urban (duct) 34, 8, and 2 Mbit/s microwave. ­ suburban ­ rural (buried) Cost elements for microwave sites are discussed 0% 0% 0% next: Then, some complementary indicators related to the cable network are required: % of shared masts on different routes Backbone network Trench sharing TDMA systems Percentage of trenches that are shared Percentage of microwave by geo-type: routes served by RCU- LS-LS- - urban different masts: LS TS to IS CS-TS - suburban ­ Light - rural 0% ­ Medium Percentage of shared trench ­ Heavy attributable to conveyance 25% Total 0% 0% 0% 0% Cables Cables per duct 1 · How many routes share the same microwave sites? · The percentage of trenches (wrapped/ducts) What is the distribution between the core and equally shared with other networks, including the access networks, or the TDMA network? access network, the cable network, or a private · How many links share the same antennae masts? transmission network.In general,fully buried cables ­ Light mast (lower than 40 meters) are dedicated to a single type of network. ­ Medium mast (between 40 and 60 meters) · The percentage of trenches not equally shared with ­ Heavy mast (higher than 60 meters) other networks. In general, in such circumstances, Other costs shared between the core and the access the core network only occupies about 25 percent of networks are discussed: the trench's capacity.This is given as a default value. · The number of cables per duct (generally one). This is followed by assumptions concerning Site costs Access Core Other microwaves. Percentage of RCU sites attributed to service 50% 50% 0% Percentage of LS sites attributed to service 50% 50% 0% Microwave Percentage of TDMA TS sites RCU- LS-LS- attributed to service 30% 70% 0% Closure of SDH rings LS TSW to IS Number of routes Percentage of site costs allocated to transmission (as RCU- LS-LS- opposed to switching) 25% Non-SDH routes LS TSW to IS Percentage of TDMA costs Number of routes allocated to transmission (as Breakdown per capacity opposed to switching) 15% 34 Mbps 8 Mbps 2 Mbps 0% 0% 0% The following costs are attributed to the core net- work by default: · 50 percent of the costs for the RCU and LS site. How many SDH rings are closed by microwaves? · 70 percent of the costs of the site for terminal sta- Microwave links considered for closure operate at 155 tions in TDMA systems. 40 User Guide The following are attributed by default to the IMPORTANT: As soon as a blank table is switching node: filled in by the user (even a single box), it will be · 25 percent of site costs (and therefore 75 percent to taken into account by the model.Any table that transmission). substitutes for the default table must therefore · 15 percent ofTDMA system costs (and therefore 85 be filled in completely. Some boxes can be left percent to transmission). empty, but all of the boxes that do not have a null value must be filled in.The user can specify Main Assumptions on Operation a table without having to specify them all. Finally, this Tech sheet provides the number and the The first table relates to the nodes: distribution of staff between the core network, access, and other activities. Routing factors RCU LS TS IS CS TS The user is asked to provide an estimate of the Local telephone number of staff considered optimal by an efficient Internet calls operator.The total obtained is then distributed over the Long distance switching, transmission, and infrastructure activities: International ­ incoming ­ outgoing Calls to mobiles Personnel ­ domestic (core network) Number Calls from mobiles Switching ­ domestic Transmission ­ international Infrastructure Interconnection Breakdown RCU LS TS IS CS TS ­ local Switching 25% 25% 0% 5% 15% 30% ­ long distance RCU- LS-LS- to Others (Kiosk, LS TS IS CS-TS switches X25 etc.) Transmission 15% 50% 5% 30% Infrastructure 15% 50% 5% 30% Thus, the first line should indicate how many RCUs, LSs, and TSs are crossed on average by a local telephone call and, subsequently, for all the types of Main Assumptions on Routing Factors calls according to their characteristics. The RoutFact sheet is accessed by the Routing factors The second table requires the same information for button on the menu. This sheet includes 12 routing the different types of links, while adding leased lines: factor matrixes. A table showing the default routing factors calcu- RCU- LS-LS- lated on the basis of the information supplied in the Routing factors LS TS to IS CS-TS two previous sheets is placed on the right. A similar Local Internet calls blank table is presented on the left,allowing the user to Long distance enter his/her own computed routing factors if the International ­ incoming default values calculated are unsuitable. ­ outgoing The first three tables are the routing factors for the Calls to mobiles fixed-line network, while the latter three are for the ­ domestic Calls from mobiles mobile network. In each case, the first two tables indi- ­ domestic cate how many times on average each type of call uses ­ international each network element.The first information is for the Interconnection ­ local "nodes" elements and the second is for the "links" ele- ­ long distance ments.The latter indicates how many times each type Others (payphone, of interconnection service uses these same elements ISDN, etc.) Leased lines (nodes and links) on average. 41 A Model for Calculating Interconnection Costs in Telecommunications Finally, the third table gathers the same information Local for each interconnection service: Model currency currency versus Euro Euro Name Inter- Exchange Local Single Double national rate versus level transit transit Transit transit Euro 1,000 Switching Customs RCU and transit LS duties TS Market IS surcharge Total CS Insurance TS and freight 100% Transmission Model RCU-LS currency Euro Local currency LS-LS-TS Annual cost of TS-TS an employee to IS CS-TS Options for cost annuity: 1 1 annuity including price trend 2 This third table covers both the nodes and links. annuity without price trend Main Cost Allocation Assumptions The Ucosts sheet is accessed via the Costs button on 1. The first is the level of the customs and transit duty the menu. applied to imported telecommunications equip- This sheet includes four major series of assump- ment. In general, customs duties can be very high, tions: though, temporarily, exemptions are often granted 1. General assumptions on costs. to operators as part of incentive schemes imple- 2. Assumptions in regard to common costs,and apply- mented in order to develop foreign direct invest- ing to attributable costs calculated by the model. ment.In the proposed cost model,it is assumed that 3. Assumptions in regard to the unit costs for each all imported equipment is subject to custom duties. network element for the fixed-line network. 2. The second is a market surcharge reflecting price dis- 4. Complementary assumptions for the mobile net- tortions related to the small size of African work. economies.This surcharge applies to imported equip- ment.Equipment prices inAfrica are generally higher General Assumptions than international prices. Because of the tiny size of These assumptions are important. First, they relate to African telecommunications markets, orders placed the currency in which the results will be presented and by operators are small in volume. Consequently, the costs calculated. The name of this currency is African telecommunications operators' negotiation sought in the first line and then its exchange rate to power, vis-à-vis major vendors, is limited. euros. 3. The third is related to transport and insurance costs The model can be run with local currency or in that are significantly higher for landlocked countries. euros or U.S. dollars.An exchange rate is provided for These extra charges considerably increase the cost the conversion into local currency.The proposed cost of imported equipment. The model provides the model uses the euro as the reference currency, but this opportunity to factor in this consideration when reference can be easily changed. Consequently, default reviewing the network's costs.Salary and labor costs are unit costs are expressed in euros. in local currency.These costs are obtained by dividing The model then requires three elements of supple- the staff expenditure incurred by the operator by the mentary information on the investment cost: number of employees. 42 User Guide Finally, the model requires that the user clearly ferent. Equity has a risk premium calculated on the show how cost annuities are computed. For further basis of the average return on investment observed in details on different methodologies, the reader is the local stock market and a sector-specific factor (beta referred to appendix 1. Two options are provided: (a) coefficient). Meanwhile, debt has a specific risk pre- option 1 is to factor in the equipment price trend in mium factor. the computation; (b) option 2 does not integrate any price trend for the equipment. In general, the user will Common Cost Assumptions select an option that takes into account the price The second part of the sheet deals with the common development of each network's element. costs ratio.There is no strict standard in this area.The The following are the key elements of the financial applicable common cost ratio does not necessarily have calculations: to be the ratio derived from the operator's books.The 1. Appendix 2 reviews the cost of capital calculation. ratio should refer to those observed from efficient Information required to compute the cost of capi- operators. This is an area for which the regulator tal is provided below. should conduct effective benchmarking. 2. The user introduces the margin of working capital The following are the default values provided by needed to efficiently operate the network. This is the cost model: provided as a percentage of the total attributable and common over investment and operating costs. Common costs attributed to in % RCU LS TS IS CS TS Investments 5% 5% 5% 5% 5% 5% Level of gearing [D/(D+E)] 35% Operation 10% 10% 10% 10% 10% 10% Risk free return 8% Average rate of return of the overall market 15% Common costs RCU- LS- TS- to Risk premium on stock (beta coefficient) 0.80 attributed to in % LS LS-TS TS IS CS-TS Risk premium on debt (spread) 2% Income tax rate 35% Investments 5% 5% 5% 5% 5% Operation 10% 10% 10% 10% 10% Cost of capital (pre-tax) (%) 10.0% Working capital surcharge (%) 0.0% Common investment costs include, for example, The cost of capital resulting from the assumptions the vehicle fleet, other investment at the headquarters, described above is then presented in a dark green cell. and so forth. These costs are distributed among the In other words, the operator offering interconnec- firm's activities. tion services has two options to finance its investment: debt and equity.The gearing ratio captures the leverage Unit Cost Assumptions exposure of the operator. Debt and equity are usually The following unit cost assumptions are considered for priced at different rates as their respective risks are dif- each network element: Operating User Default costs as a Equipment input value Installation Evolution Scrap value percentage of Price local Equipment costs Asset Price of as a % of equipment local currency Price (% of capita life trend Capital* equipment capital Equipment X currency FAB Euros FAB costs) (years) (%) price capital cost cost (%) Equipment X Fixed cost of equipment -- 147,000 10% 11 ­8% 0.92 1% 3.0% Site cost of switch -- 200,000 10% 38 ­1% 0.99 11% 5.0% Cost per line -- 10 10% 12 ­8% 0.92 1% 2.0% Cost per trunk -- 1,500 10% 12 ­8% 0.92 1% 2.0% 43 A Model for Calculating Interconnection Costs in Telecommunications 1. 1st column (column C in the model): contains the 8. 8th column (J): residual value of the equipment at result of the unit cost calculation.The user should the end of its lifetime. The residual value is avoid changing this information.This value derives deducted from the capital cost used in calculating either from input placed in the third column the constant economic annuity. It is primarily sig- (expressed in euros) or from the value entered by nificant for buildings and sites for which the land the user in the second column.The following cal- keeps a high residual value.This column contains a culation is done : default value that can be modified by the user. a. On the basis of the default value expressed in 9. 9th column (K): operating cost incurred by the euros (third column), the exchange rates and the equipment as a percentage of the capital cost.This three rates stipulated for imported equipment cost is the operation-maintenance cost incurred (customs, market surcharge, and transport) are directly by the equipment, and does not include applied; staff or labor costs. It comprises spare parts, repair b. On the basis of a value entered by the user (entry costs, and expenditure for equipment consumables in column 2, column D, and takes precedence (energy).This column contains a default value that over the default value) in local currency, and two can be modified by the user. of the three rates stipulated in reference to The number of unit cost items amounts to 68 and imported equipment (customs and transport, but is grouped below as follows: no market surcharge) are applied. · Eighteen headings for switching. 2. 2nd column (D): data entered by the user as a unit · Twenty-four headings for transmission (SDH elec- cost that is more suitable than the default cost. Data tronics, microwaves, towers). in national currency takes precedence over the · Nine headings for infrastructures (cables and ducts). default cost. · Four items for other costs (including one related to 3. 3rd column (E): default value of the unit cost mobile networks). expressed in euros. · Thirteen cost headings specific to mobile networks. 4. 4th column (F): equipment installation cost (as a Some of the headings deserve the following specific percentage of the capital cost). This cost includes comments: possible engineering costs (survey, planning, · Locally supplied equipment or local construction design), costs for monitoring and possibly inspect- (buildings, civil engineering) are not subject to ing the manufacturing process, the installation costs imported expenses (custom duties and taxes). per se, the costs for system testing, and costs for · Cost heading related to Domsat repeaters; leased training (on site or abroad).This column contains a lines for mobile networks belong to operating cost. default value, which can be modified by the user. · For some investment, the model does not identify 5. 5th column (G): an economic lifetime and not an specific cost items.These investments are assessed as accounting lifetime. This duration is implicitly a percentage of investment cost implied by a basket adjusted if cost annuity is done according to option of items. The model selects four baskets for these 1 (price trend).This lifetime is determined by the complementary investments: equipment's operating longevity,availability of spare ­ Switching investments: this item includes every- parts, and so forth.This column contains a default thing needed for the operation of basic services, value that can be modified by the user. excepting value-added services, and does not 6. 6th column (H):price development trend.This is the show price entered in the equipment. This trend for the long-term development of equipment applies to: SS7 signaling, signal transfer points cost (+ or ­ x% per year). The column contains a (generally included with the switches), synchro- default value that can be modified by the user. nization, centralized network management sys- 7. 7th column (I): contains the cost of capital and of tems and, possibly, training equipment. the price development.This value is used to com- ­ Transmission investments: this includes every- pute cost annuity when option 1 is selected. thing related to centralized management, alarm 44 User Guide checking, and testing and installation equipment network costs.These are entered in the Mobile sheet, that is not included in material supplies already which the user can access via the Mobile assumptions accounted for. button on the Command (Menu) sheet. ­ Infrastructure investments: testing, layout, instal- The sheet is organized into three major sections: lation equipment. 1. Demand assumptions. ­ Mobile investments: items equivalent to the 2. Subscriber and network assumptions. mobile networks that are not included in the 3. Operating cost assumptions. unit prices for equipment. In general, the cost of network elements is divided Demand Assumptions into two categories of cost item: The user can enter aggregate traffic elements carried 1. The fixed cost of equipment (processing unit of a by the mobile network, similarly to what was done for switch, racks, management bays). These costs vary the fixed-line network: with capacity (subscribers, BHE, number of fibers). 2. The cost of the site where the equipment is installed. Margin for growth (%) Existing Number Average Number Number The site costs deserve clarification. In general, the traffic over of length of success- of site includes the following cost elements: mobile network minutes of calls ful calls minutes · Land acquisition cost. Outgoing · Land development and building costs (fencing, Internal 0 to mobile access road). network 0 · Specific costs for fitting out buildings (technical to fixed network 0 Incoming floors, air conditioning, protection against lightning from mobile and fire, security system). network 0 · Energy cost. from fixed network 0 A telecommunications building often contains: Other (CRM...) · Switching and transmission equipment. Other (CRM...) 0 · Equipment for the core network or for the access Total 0 0 network, and even for other activities (for example, accommodation for public telephone boxes, a sales branch). Demand and network elements for mobile are, of Consequently, site costs should be shared out course, different: among the various activities. The model handles the allocation of these costs in the Tech sheet, which has MSC BSC BTS already been reviewed. The cost for sites (like civil % of internal calls using the same... 80% 60% 20% engineering work) is considered to be a locally pro- Successful Unsuccess- Traffic Statistics calls ful calls vided service. Hence, it is not subject to the three rates Average call set-up time in specified in the model (customs, market cost overrun, seconds (time to answer) 15 30 transport). The same applies to costs of masts and Percentage of successful calls 77% Traffic in busiest hour of the year microwave sites. (as a percentage of the total) 0.0005 Retail tariff gradient Traffic (%) Tariffs Gradient (ratio) 100 = peak Assumptions for the Mobile Network Peak 100 Several assumptions in the preceding sheets are used to off-peak calculate interconnection costs for mobile networks. Weekend other 1 This, in particular, is the case with the Cost sheet other 2 reviewed earlier. Nonetheless, specific demand and Average 0 technical elements are required to determine mobile 45 A Model for Calculating Interconnection Costs in Telecommunications In addition, the model indicates the proportion of The average number of calls transmitted by trans- internal calls that terminate in the same mobile switch- mitter/receiver (TRX; this is the radio equipment that ing center (MSC), base station controller (BSC), and manages transmission at base station level) is between base terminal station (BTS) area. four and eight.The number ofTRXs per base stations, and the number of sectors by the BTSs and the num- Assumptions for the Mobile Network ber of BTSs, are computed and entered by the user. This section contains information that specifically These values are divided on the right by the number of characterizes the mobile network and is required for BTSs to ensure that the average values obtained are cost determination: actually within the technical ranges accepted. 1. Capacity installed on MSCs and the number of 5. Data on the transmission network of the mobile subscribers: operator are entered when these links can be leased from the incumbent fixed-line operators or from Number any other licensed operator or are rolled out by the Installed capacity (MSC) mobile operator (owned microwave or fiber-optic Subscribers (post-paid) links). Subscribers (prepaid) Transmission 2. Equipment utilization levels: network MSC- MSC- BSC- Utilization level of Microwave MSC BSC BTS Total switching nodes (%) MSC BSC-BTS Number of routes 0 Used capacity Total length (m) 0 Erlangs 95% 80% Break- Break- MSC-MSC MSC-BSC BSC-BTS down down Breakdown of total of total Utilization level per capacity routes length of transmission 155 Mbps elements (%) 90% 70% 60% 34 Mbps Erlangs per circuit 0.5 0.6 0.5 8 Mbps 2 Mbps 3. Node information: The user enters the type of link, the number of routes, length of routes, and distribution by capacity. Node information MSC BSC BTS Number of nodes (total) 6. The breakdown of the number of masts is entered, BSC co-located with MSC thus: Transmission MSC- MSC- BSC- MSC- MSC- BSC- electronics MSC BSC BTS MSC BSC BTS 4. Data on BTSs: 155 Mbps 0 0 0 34 Mbps 0 0 0 8 Mbps 0 0 0 BTS data Total per BTS 2 Mbps 0 0 0 Number of 0 0 communications per TRX 4 < x < 8 Total number Total number of TRX of masts (Full duplex channel) 6 < x < 16 Percentage of different masts: ­ on roof Number of sectors ­ Light (cells) 1 < x < 6 ­ Medium Number of sites < 1 ­ Heavy 46 User Guide The sizing of electronics equipment per link is The Results Sheets done automatically.If the breakdown appearing in cells F76:H79 (model) does not match the actual situation, After entering all the inputs and assumptions, the user the user can enter the relevant data in cells C76:E79. can display, print, or test the sensitivity of the model's The total number of physical masts should be entered results.These buttons activate four sheets: in cell C81. 1. A Results sheet is displayed. The model then requests a breakdown of masts in 2. The user can print the model's results by selecting the following categories:(a) masts installed on roofs,(b) the fixed-line or the mobile network simulation. masts requiring light masts (<40 meters), (c) medium- Each report includes the results and assumptions. sized masts (between 40 and 60 meters), and (d) heavy However, the intermediary computations are not masts (>60 meters). printed, though these can be displayed in different spreadsheets. 7. Leased lines used by the mobile network: 3. The user can test the results' sensitivity to predeter- mined parameter variations. Leased lines MSC- MSC- BSC- (2 Mbps equiv.) MSC BSC BTS The Results Urban The simulation results for fixed-line and mobile oper- ­ Number Nonurban ator are presented in Print sheets. Each simulation ­ Number includes: ­ Total length (km) · Costs per minute for each network element. The model requires differentiating urban and · Interconnection services costs obtained from net- nonurban leased lines. The user is requested to enter work elements costs as follows: network elements the number of leased lines per type and per length. costs that are respectively multiplied by corre- sponding routing factors to derive interconnection Assumptions on Operating Costs service costs. The Mobile sheet includes data needed to calculate the For the fixed-line network,the cost model provides labor costs and their breakdown: the option to add rural TDMA systems costs to the interconnection service costs.The model provides two interconnection cost estimates. The first estimate Personnel (network operation) Number includes TDMA systems costs, while the second one Switching does not. Depending on universal access policies, the Transmission regulator will decide whether to account for TDMA Breakdown MSC BSC BTS Switching 25% 25% 50% systems costs. MSC-MSC MSC-BSC BSC-BTS Transmission 5% 15% 80% Cost per Switching minute The second part of the Cost sheet includes the RCU assumptions required for common costs allocation. LS TS Data provided by default can be modified by the user. IS CS TS Common costs Transmission attributed to in % MSC BSC BTS RCU-LS Investments 5% 5% 5% LS-LS-TS Operation 10% 10% 10% TS-TS Common costs to IS attributed to in % MSC-MSC MSC-BSC BSC-BTS CS-TS Investments 5% 5% 5% Operation 10% 10% 10% 47 A Model for Calculating Interconnection Costs in Telecommunications Local Single Double Transit 0 level transit transit Transit international Interconnection rates (with TDMA systems) Interconnection rates (without TDMA systems) Africa "Best current practice" peak rates (Model Currency) Europe "Best current practice" peak rates (Model Currency) Local Single Double Transit Euros cents level transit transit Transit international Interconnection rates (with TDMA systems) Africa "Best current practice" peak rates Europe "Best current practic" peak rates (Euros cents) 0.5 ­ 0.9 0.8 ­ 1.5 1.5 ­ 1.8 0.3 ­ 0.7 The costs are expressed in the currency selected in For the mobile networks, the results provide the the unit cost sheet.They are converted into euros for cost per minute for network elements and the cost of comparison purposes. These costs are average costs.They interconnection services. A distinction is made are then converted into costs according to time-of-day between call termination and origination, whenever slots with the gradients specified in the demand justified by the routing factors. assumptions: Switching Cost per minute Time of day MSC charging BSC-BTS (Model BTS currency) Transmission Transit MSC-MSC With TDMA Local Single Double inter- MSC-BSC systems level transit transit Transit national BSC-BTS Peak Off-peak 0 Originating Terminating Weekend Interconnection rates Other 1 Euro cents Other 2 Interconnection rates Time of day charging Originating Terminating (Model currency) Peak Off-peak Weekend Other 1 Other 2 48 User Guide Printing Reports The printer and printing parameters can then be Printing can be launched from the Menu sheet: specified. · Results and assumptions for the fixed-line network. The same procedure applies to printing the mobile · Results and assumptions for the mobile network. networks report. When launching the printer, the user is asked to provide a name for the printed simulation.When pro- Testing the Sensitivity of the Results vided, the name is stated at the top of the printed Once an initial determination of the interconnection report for identification purposes.The statement is not costs is obtained, the sensitivity of the results can be compulsory. tested. The user can perform the sensitivity analysis through the Sensitivities button on the Menu sheet. A snapshot of the sheet is provided below: The document is printed using a print preview function that enables the user to change the format and adapt it to the available printer, if necessary. Printing generally requires eight pages. To test the model sensitivity, captions for six prede- termined parameters are provided.The user can enter input as described below: · Traffic at peak hours (as a percentage of the total traffic) (cell C31 of the Demand sheet). · Total length of the trenches (sum of cells C48 to E48 of the Tech sheet). · Total staff (sum of cells C105 to C107 in the Tech The printer selection box is called up by clicking sheet). on the Print button on the upper bar: · Average annual cost of an employee (cell E16 of the Ucosts sheet). · Market surcharge ratio (cell C12 of the Ucosts sheet). · Level of gearing [D/(D + E)] (cell C21 of the Ucosts sheet). The user can modify these variables, either by entering a relative variation expressed in percentage (total length of trenches, staff, annual cost of an employee) or by adding to or subtracting from the original value, which is already a percentage (peak hour traffic, market surcharge, level of gearing). 49 A Model for Calculating Interconnection Costs in Telecommunications The modification can be made directly by entering a value in the cell assigned for new value; the user can increase or decrease the variable value by utilizing the vertical "cursors." The simulation is run by pressing the Test button when the new values are entered. The original values and the new sensitivity param- eter values appear in columns E and F of an Excel sheet.The new results can be compared with the old ones, as shown in the following table: A : Interconnection rates B : Interconnection rates with TDMA cross-subsidization Before change Single Double International Local level transit transit Transit transit A Old values B After change Single Double International Local level transit transit Transit transit A New values B Change Single Double International Local level transit transit Transit transit A Differences B The user can iterate several sensitivity tests before If the user decides to keep original values, the confirming and saving the resulting values.This is done model's assumptions are not modified. Conversely, if by pressing the Set sensitivities button, which opens a the user decides otherwise,the model's assumptions are dialogue box that requests the user to either validate modified accordingly.To be able to keep both original the modifications made or to cancel them. and new values, the user must save respective files in different names. 50 User Guide Managing the Model 2. After providing all the input, the user can display · Users are advised to save completed simulations the results. with suitable names.The user can clear the model's 3. Test the sensitivity of these results to variations of assumptions or restore the default values of the main or key parameters. model. 4. Print out the reports, including the simulation results, along with the underlying assumptions. IMPORTANT: Whenever a logical problem (such as division by zero) arises, or there is any inconsistency among assumptions, the model will not compute the interconnection costs. For example, sizing the capacity of SDH rings assumes that all input provided by the user is consistent with the traffic that will be conveyed. In this particular case, the user has to ratify the assumptions step by step. Doing this, he or she can either modify the capacity of the SDH sys- tems (cells C26 to E31 of the Tech sheet), or To run a simulation with the model, the user must: review the calculations done by the Capa EITr 1. Enter the different assumptions or parameters sheet. required. If needed, the user can also modify the default values provided. Notes 1. This includes the formulas as well as the layout (addition or suppression of lines or columns, and so forth). 2. Thus, a TDMA system connected to an exchange that included a capacity of 512 subscribers would count as two CSs. 51 6 Operations of the Cost Model This section presents the proposed model in greater different types of nodes and five different types of links. detail. These elements are extensively reviewed in chapter 2 The model was developed using Excel 2000 and of this guidebook. Visual Basic.All intermediary calculation sheets are in In terms of sizing, LS-LS and TS-TS links are bun- Excel 2000. dled under the same LS-LS-TS caption to take into The model's synopsis structure is shown below: account the fact that subscriber exchanges also offer transit features in Africa.As discussed in chapter 2, few Figure 6.1 Cost Model Architecture African networks have dedicated transit exchanges. However, when calculating the interconnection costs, Routing Unit the model splits these costs into two categories: LS-LS Hypotheses Demand Network Factors Costs Demand Tech FactRout UCosts links and TS-TS links. To size the various elements, the model uses total Traf c Traf c Capa and the peak hour traffic information.This is done by the Capa sheet.The network size is calculated for: Transmission Infrastructures Sizing Capa Eltr Capa Intra · Transmission electronics equipment (Capa ElTr sheet). Cost Switching Transmission Infrastructures Costs Sw Costs Tr Costs Infra · Infrastructures (Capa Infra sheet). Totals Results Result After the sizing is done, the model calculates the tot results network costs for each element: · The nodes comprise the subject of a cost break- down on the Costs Sw sheet. Assumption sheets are described in the user guide. · The link costs are broken down on two sheets: This section further spells out how the intermediary ­ The transmission costs are calculated on the computations are implemented by the model. Costs Tr sheet. ­ The infrastructure costs are calculated on the Basic Principles Costs Infra sheet. Total traffic and costs (including the common costs) The model builds on the breakdown of the core net- are consolidated before the model determines the unit work, as discussed in previous chapters, into network costs (Tot sheet).The Result sheet then presents unit elements. Eleven network elements are identified: six interconnection costs according to routing factors. 52 Operations of the Cost Model Logic of the Intermediary Spreadsheets a. The capacity of transmission elements is calculated. b. An adjustment factor is then applied to oversize the This section describes in detail the intermediary calcu- capacity.The adjustment factor is taken equal to 1, lations. by default. It can be adjusted in line 10, if excep- tional factors that apply to certain types of links Capacities (Capa Sheet) have to be taken into account. This sheet calculates the total traffic and the peak hour c. The model then calculates the proportion of the traffic using the information on volume of traffic in link's capacity not handled by SDH links from sub- minutes, as entered by the user: scriber connections, which are not connected to a. The volume of traffic is adjusted to factor in non- SDH links. In the case of international transit, the billed traffic, which is costly (time needed to set model assumes that links to the international transit up the successful calls. Column C: number of calls exchange are point-to-point SDH. If this is not the * average time to answer successful calls in sec- case, the user should, herein, indicate the propor- onds /60). tion of these links that do not use SDH technology b. The resulting value is adjusted with the time con- (cell E11). sumed to convey unsuccessful calls (column D: col- d. The capacity (in Mbit/s) that has to be served in umn C + number of unsuccessful calls * average SDH, in Domsat, and in non-SDH terrestrial lines time to respond to unsuccessful calls/60). is then derived (lines 12 to 14). c. This is then adjusted with the predicted traffic e. The model then considers the number of nodes and growth rate, as indicated in the Demand sheet. the positioning of the nodes with respect to links d. As a result, the adjusted volume of traffic (invoiced (lines 17 and 18), as well as the mix of SDH systems and noninvoiced with a growth rate) is broken for the various categories of links (lines 27 to 30). down by network element (four nodes and three Three types of links are served by SDH technolo- links, excluding the radio concentrator equipping gies: RCU-LS links, LS-LS-TS links, and links to the base stations of TDMA systems), based on routing IS. For each, the model calculates the number of rings factors provided in the RoutFact sheet. needed.The algorithm builds on a former one devel- e. The model provides the volume of traffic in min- oped by Europe Economics,and takes into account the utes supported by each network element. maximum number of nodes per loop as entered in the f. The total traffic obtained is converted into Erlangs Tech sheet (cell C10). (BHE) and used to size nodes and links.This infor- The simulation carried out assumes that RCUs are mation,depending on the charge factor of each ele- linked to LSs through SDH systems.LS-LS and LS-TS ment (percentage of occupation), is adjusted to links are implemented through an upper-level SDH ensure traffic fluidity. ring.The ring's capacity is sized according to the flow g. In the case of links, these BHEs are transformed of traffic between its nodes.It is assumed that the min- into Mbit/s. imal number of nodes to justify a ring is three. The h. The same calculation is applied to leased lines the traffic capacity that can be handled per ring is deter- number of which is converted into Mbit/s and mined according to the following factors (STM 1, 4, adjusted with the predicted demand growth rate. 16, and 64): i. The core network's capacity is adjusted according · The number of rings constituting the core net- to resources required by the public switched teleph- work. ony network and leased lines traffic. · The average number of nodes per ring. · The number of physical routes per ring (number of Transmission Capacities (Capa ElTr Sheet) nodes + 1). This sheet breaks down the transmission equipment, · The number of ADMs and termination multiplex- which includes add-drop multiplexers (ADMs) and ers (MUXs) (2 * number of rings + total number of termination equipment for SDH rings (STM 1, 4, 16, nodes). and 64). · The number of regenerators. 53 A Model for Calculating Interconnection Costs in Telecommunications Links to ISs are assumed to be SDH point to point · Trenches and ducts. and are, therefore, much easier to size. The following · Radio links. box summarizes the algorithm used by Europe Eco- The trenches are, themselves, differentiated in three nomics' model to size transmission links. subcategories, according to their location: · Urban areas: wrapped ducts. Infrastructure Capacities (Capa Infra Sheet) · Suburban areas: ducts. This sheet sizes the infrastructures needed to roll out · Rural areas: buried cables. the core network transmission links. In general, there In the proposed model, aerial cables are not used are two types of infrastructures: for conveying traffic in the core network. Box 6.1 Sizing of the SDH Network (example of RCU-LS routes) The transmission equipment required is evaluated by type of link. · Capacity required in Mbit/s = total capacity required in Mbit/s (adjusted) taken in the assumption section of the same sheet. · Distribution of the required capacity (Mbit/s). STM1 = required capacity * total STM1 system mix (the two data items are taken in this sheet) (A) STM4 = idem (B) STM16 = idem (C) STM64 = idem (D) · Number of nodes = number of RCUs taken in the same sheet, assumptions section. · Total number of nodes per STM1 capacity (STM1 share in the systems total by the number of nodes). capacityrequired in STM1(A) capacity required for all STM ( A - B - C - D ) * number of RCUs · Same formula for the STM4, STM16. and STM64 systems. · Same formula for the STM4, STM16. and STM64 systems. STM1: number of nodes per capacity for STM1 capacity required for STM1 maximum number of nodes on a ring ("technical" sheet) maximum capacity of a piece of SDH equipment IFthen take the first term; else take the second. ; Column 2: the preceding result * by the maximum capacity of a piece of SDH equipment (same sheet in assumptions). The same procedure is applied to STM4, STM16, and STM64 systems. · Number of extra nodes needed to have at least 3 nodes per ring: logical function. number of RCUs3 * per capacity ; then marked "check" ; IFelse rings several cases are possible represented by logical functions: 3 * (total number of STM1 rings ­ total number of nodes(RCU)) 0; IFthen take this figure + IF 3 * the same thing for STM4 ; + IF 3 * the same thing for STM16 ; ( ) ( ) + IF 3 * the same thing for STM64 ; else 0 ( ) (Continued on page 55.) 54 Operations of the Cost Model Box 6.1 Sizing of the SDH Network (example of RCU-LS routes) (continued) · If we obtain "check" for the preceding calculation, i.e., if the number of RCUs is under 3 * the sum of the rings per capacity we then recalculate the total number of rings per STM1, STM4, STM16, and STM 64 capacity using an IF logical function. · We recalculate the total number of nodes per capacity (STM1, STM4, STM16, and STM64) adjusted by 1 as a logical function again as a function of the calculation of the number of extra nodes needed to obtain at least 3 nodes per ring. For example, for STM1 (the same formula is applied for the other capacities): C39(cell - number) = "check"; then take a figure in the recalculated previous table i.e., IFnumber of nodes * number of3 STM1 rings / total number of rings of all capacities; * total number of rings per capacity total number of nodes (RCU) ; else take a logical function IFthen take 3 * the total number of rings per capacity; else take the total number of nodes · Reallocation of nodes on rings by capacity. Column 1: for example for STM1 systems (the same formula is applied for other capacities): C39 = "check"; then take the number of nodes above; IFelse IF number of nodes ­ 3 * number of rings 0; then take this number ; else take 0 (( ) ) Column 2: C39 = "check"; then take 0; IFelse take (column1 STM1/ column 1) * the number of extra nodes needed to obtain 3 nodes per ring · The total number of nodes by adjusted capacity 2: For example, for STM1 (the same formula is applied for other capacities): C39 = "check";then take the adjusted rounded number of nodes 1; else IF the adjusted number of nodes 1 = total number of rings, take the rounded total adjusted number of nodes 1; IFelse take the total adjusted number of nodes 1­ preceding reallocated nodes column 2 · The average number of nodes per ring. For example, for STM1 (the same formula is applied for STM4, STM16, and STM64): total adjusted number of nodes by capacity 2 = 0 then 0 ; C39 SI = "check"; then IFelse the total adjusted number of nodes by capacity 2 ; adjusted total number of rings else IFtotal number of nodes by adjusted capacity 2 = 0; then 0; else adjusted total number of nodes by capacity 2 total number of rings · Average number of physical routes per STM1 = IF(average number of nodes per ring = 0; then 0; else take the average number of nodes per ring +1) · Average number of physical routes per STM4 ring = same formula as STM1. · Average number of physical routes per STM16 ring = same formula as STM1. · Average number of physical routes per STM64 ring = same formula as STM1. (Continued on page 56.) 55 A Model for Calculating Interconnection Costs in Telecommunications Box 6.1 Sizing of the SDH Network (example of RCU-LS routes) (continued) · Total number of physical routes: C39 = "check"; then sommeprod (average number of physical routes * total number of rings by adjusted capacity for b98 IF= "check"; else sommeprod (average number of routes * total number of rings by capacity) The "sommeprod" function gives the sum of products of corresponding elements for several matrices : sommeprod matrix1;matrix2;matrix3;... ( ) · Number of termination systems = total number of physical routes (above) *2. · Number of multiplexers (Gateway MUX and ADM) by capacity: For example, for STM1: IF C39 = " check"; then 2 * the total number of rings adjusted to take account of B98 = "check"; else 2 * the total number of rings ( ) + the number of nodes by adjusted capacity 2 · Number of regenerators: RCU - LS average distance ("technical" sheet) roundup average distance between regenerators ("technical" sheet) , 0 -1 * diversity of regenerators ("technical" sheet) * total number of physical routes The rounded function up (z,0) rounds the result of the ratio z upwards to the next integer. Trenches (all geo-types considered [Tech line 22]) multiplied by The first table recapitulates the length of trenches by the number of physical routes taken from the Capa ElTr geo-type (lines 7 to 10). Then the breakdown of sheet (route between two ADMs).The total obtained is trenches is implemented as follows: then multiplied by four, assuming that each route is 1. Nonshared trenches (with other networks) are served by four fibers in general. By dividing the length identified and corresponding costs allocated to the of the fibers by the length of the cables, the average core network (lines 15 to 18). number of fibers per cable is obtained. And, from this 2. Shared trenches are identified (lines 20 to 23). latter information,the model determines the minimum 3. The proportion of shared trenches costs that are fiber capacity of cable (6, 12, 24, 36, 48, and 96 fibers). attributable to the general network is then derived. 4. The model adds directly attributable costs (non- Microwave Network shared trenches) and indirectly attributable costs Two cases are envisaged: (proportion of shared trench costs) to obtain the · A microwave link is used to close an SDH ring.This total cost of trenches allocated to the core network. is usually justified when it is too costly to lay down The length of cables is calculated in terms of the a cable (stretches of water, mountains, crossing for- number of cables declared per duct (1 by default, cell eign countries). Tech C70).The length of fibers required is calculated.It · The transmission network has not been upgraded is equal to the average length of the transmission routes to SDH technology. 56 Operations of the Cost Model The closure of SDH rings is done with 155 Mbit/s Switching Costs (Costs Sw Sheet) systems. The non-SDH links are only recommended This sheet calculates the costs for the six types of node for narrow band systems (bandwidth inferior to 155 as shown below. Five cost headings are retained. Note Mbit/s).The non-SDH links are therefore dedicated to that these headings are not relevant for all the nodes. connect areas or localities for which SDH systems are not viable.Three capacity levels are retained: 34, 8, and RCU LS TS ISI CS TS 2 Mbit/s. Fixed costs of For each type of link and technology, the model the switch calculates: Site cost of the switch a. The number of radio systems from the number of Cost per line routes. Cost per trunk b. The length of the networks and the average dis- Other related costs tance of leaps. c. The number of pylons and the breakdown of these pylons by size (light, medium, and heavy). · There are no subscriber line costs for transit d. The necessary antenna equipment. switches (TS and IS). · There are no site costs for central stations of TDMA Costs systems (co-localized with the switches or the RCUs). The costs are calculated on three distinct sheets: one · The cost of subscriber cards is included in cost per for the switching costs (Costs Sw), one for the trans- line. mission costs (Costs Tr), and one for the infrastructure · There is no cost for trunk in radio concentrator costs (Costs Infra). equipment. At the head of each sheet, the model recapitulates · The number of equipment is derived from the the size and number of equipment planned, and infor- Demand and Capa sheets (for two Mbit/s ports). mation needed to allocate cost of shared equipment. The investment costs per equipment type are calcu- Transmission Costs (CostsTr Sheet) lated as follows: The transmission cost calculation is made as follows: Investment = volume required * unit cost * RCU- LS- to (1 + installation cost (in percent)) * LS LS TS-TS ITS CS-TS STM 1 (1 ­ residual value/(1 + life span) ^ (capital cost)) STM 4 STM 16 An investment annuity is calculated as follows STM 64 Regenerators (discount rate with or without incidence of the Digital cross equipment price trend): connects Line termination systems STM 1 Annual repayment = investment / phi (life span, Line termination discount rate) where phi is the annuity function systems STM 4 shown in appendix 1. Line termination systems STM 16 Line termination The final step is to calculate operating costs attrib- systems STM 64 utable to each network element. TDM relay stations Domsat central These sheets are organized as follows: station Domsat local station Switching sites Other related costs 57 A Model for Calculating Interconnection Costs in Telecommunications The cost headings, mentioned above, do not apply Total Costs (Tot Sheet) to all transmission links. For example, in the Domsat TheTot sheet consolidates all the data calculated in the network, the central station is accounted for at the previous sections and calculates cost annuities for net- transit level, while local station costs are accounted for work elements used to provide interconnection ser- at the level of local exchanges. For operating costs, the vices (see lines 34 to 40 of the Ucosts sheet). In doing repeaters and staff costs are added. so, one obtains for the switching and the transmission elements, respectively, the total traffic these elements Infrastructure Costs (Costs Infra Sheet) carry, and therefore derives their costs. Dividing the The following costs items are accounted for: total cost of elements of a network by the total traffic · Costs of cables. gives the unit cost per minute. · Costs of civil engineering. · Costs of microwave systems. The Mobile Network Calculations The following table summarizes how these ele- ments are taken into account: The mobile network is modeled with three types of node: MSC-MSC, MSC-BSC, and BSC-BTS. Three RCU- LS- TS- To sheets are devoted to the intermediary calculations of Cable : LS LS TS IS the interconnection costs for originating and terminat- Cable 6 fibers Meters ing services on the mobile network. Cable 12 fibers Meters Cable 24 fibers Meters Cable 36 fibers Meters Capacity of the Mobile Network (Capa Mob Sheet) Cable 48 fibers Meters The capacity of the mobile network is determined Cable 96 fibers Meters according two factors: the total traffic and the peak Trench Duct (meters) - hour traffic.The sizing exercise is conducted for each transport alone Urban network element. Suburban Rural Duct (meters) - Cost of the Mobile Network (Costs Mob Sheet) shared Urban The number of equipment required and the cost com- Suburban Rural putations are carried out on the same sheet.The Cost Wireless sheet is similar in structure to the one reviewed earlier Radio RCU- LS- TS- To CS- for the fixed network (number of equipment, invest- system LS LS TS IS TS ments, annualized costs, operating costs). 155 Mbps Number 34 Mbps Number 8 Mbps Number 2 Mbps Number Antenna equipment Wireless sites with environment Total number of pylons by dimension ­ Light Number ­ Medium Number ­ Heavy Number Other related costs The calculations are similar to those of the other Cost sheets. 58 Operations of the Cost Model Switching MCS BSC BTS Equipment Number Sites Number Subscriber lines (variable) Number TRX Number Primary digital block (2 Mbit/s port) Number of ports Microwave MCS-MCS MCS-BSC BSC-BTS Radio system 155 Mbps Number 34 Mbps Number 8 Mbps Number 2 Mbps Number Antenna equipment Microwave sites with environment Total number of pylons by dimension ­ rooftop Number ­ Light Number ­ Medium Number ­ Heavy Number Leased links MCS-MCS MCS-BSC BSC-BTS Urban Number Interurban Number Interurban Km MCS BSC BTS Personnel Number MCS-MCS MCS-MCS BSC-BTS Personnel Number The calculation is thus based on: · Exchanges equipment functions, features, number of equipment, number of subscribers, and number of installed TRXs. · Microwave systems allowing connections of remote equipment to the exchanges. · Capacity of leased lines. · The number of staff operating the network. Total Costs of the Mobile Network (Tot Mob Sheet) The structure of the sheet is identical to that for the fixed network, apart from the fact that no system of realignment is implemented. 59 Appendixes Appendix 1: Economic Cost Approach counted sum of the original investment and of the operating-maintenance costs less the discounted resale Calculation without Taking into Account Technical Progress value of the equipment at N. Given the expression The total discounted cost of using equipment for N retained for fn = f0, we have: years in the absence of technical progress is written: N CN = I0 + (1+i) f0 - VN n N (1+ i)N CN = I0 + (1+i) fn - VN 1 N ( n 1 VN n=1 (1 + i)N = I0 + f0 - 1 1+ i)n (1+ i)N where: = I0 + f0Ni - VN · N is the economic lifetime of the investment; n the (1 + i)N current year. · I0 is the cost of the investment paid in year 0, or the It is then necessary to consider the average annual discounted sum in year 0 of the investment costs if economic cost of implementing this investment, which these are spread over several years. corresponds to the LRAIC.To do so the annual install- · fn is the operating-maintenance costs of year n, n = ment equivalent to the total discounted cost is considered 1, 2,..., N; these costs are generally increasing over (that is to say,the sum which,paid annually from year 1 time,but generally the assumption is made that they to year N, would enable it to be repaid). are constant and equal at f0. · Vn is the resale value in year N (with conventionally CN CN I0 VN 1 V0 = I0). XN = = = + f0 - N We write i (phi function) function1: 1 Ni Ni (1+ i)N Ni N (1+i) n 1 N 1 VN Ni = (1+i 1 = (1+ i)N - 1 = (I0 - ) + f0 )n Ni (1+ i)N 1 i(1 + i)N As defined above, the total discounted cost of the The total annual cost is the sum of the: implementation of the original investment I0 over the · total investment less the discounted residual value, period N is equal to CN that is equivalent to: the dis- the whole discounted by phi (N,i) and 60 Appendix 1: Economic Cost Approach · the annual operating costs According to this definition, the residual value of The sum XN spent every year during N years is new equipment on date n = 0 is its purchase price. Its equivalent to the cost CN. It is equivalent to spend CN residual value at date N* is nil, because the user will be straightaway or to spread expenditure XN over N indifferent to losing it. By assumption, the equipment years. would be replaced with identical equipment over a theoretically indefinite period. The maximum price Calculation Taking into Account Technical Progress one is ready to pay for old equipment is determined by When the technical progress is considered, there is a comparing the costs obtained, either by procuring need to introduce a factor capturing how the equip- equipment of identical age or by buying new equip- ment is renewed over time (every N years). ment straight away. More precisely, the usage value Un is such that the discounted costs corresponding to the two possible solutions are equal: DN = CN + CN + CN + ... + CN + ... (1+ i)N (1+ i)2N (1+ i)kN · Payment of the price Un in year n, operating the equipment in question from year n + 1 to year N, This is the sum of a geometric series and can be resale in N initially planned, renewal in identical written as: fashion over an indefinite period as from year N. · Purchase in year n of new equipment, operating during N years, then renewal in identical fashion DN = CN = (1 + i)N CN over an indefinite period. 1- 1 (1 )N + i - 1 (1 + i)N The usage value Un is then the maximum that the company is ready to pay to pursue the operation with identical equipment, at the same economic cost If this expression is related to the constant annual price as that obtained with the equipment in question. installment equivalent to the cost CN, we have: N N Un DN = XN + fk VN XN * i (1 )n - = + i k=n+1 (1 + i)k (1 + i)N k=n+1(1 + i)k The value of the equipment during its lifetime is In replacing the economic cost price by the sum of frequently used in economic calculations.This is justi- the operating costs fk and of the economic depreciation fied by the need for a company to assess the valuation ak (an = XN ­ fn), we have: of the stock of equipment on a given date.What max- N imum price would a company be willing to pay for Un = ak + VN identical equipment (same age, same characteristics) if (1 )n + i k=n+1 (1 + i)k (1 + i)N this equipment was lacking? In other terms,what is the opportunity cost that would be incurred by the loss of this equipment or the extra cost of anticipating its renewal by N* ­ n years where N* is its economic life- By subtracting the two consecutive years, we have: time and n its age (usage value or residual value).This lat- ter definition amounts to creating the fiction of a Un -1 otherwise, -1 - Un = an perfect secondhand market. In fact, in such a market, (1 )n + i (1 + i)n (1 + i)n no businessperson would agree to buy used equipment more expensively than the net cost he/she would suf- an = (1 + i) · Un ­1 ­ Un fer if forced to buy new equipment prematurely, or sell equipment cheaper than the net cost it would entail to where (1 + i)Un is the discounted value in year n of -1 replace it. the usage value Un-1. 61 A Model for Calculating Interconnection Costs in Telecommunications The economic depreciation is thus interpreted as a i by a factor g, which is the decrease in prices over the loss of discounted usage value.2 The economic depre- period in question. ciation is thus an annual cost representing the use of the capital. More practically,this economic depre- Recurrent and Nonrecurrent Section of Debt Recovery ciation can be obtained by subtracting the amount of Once an annual economic cost C for a given service is operating costs at each time n from the economic cost determined,it can be recovered over a lifetime T of the price. productive resource. Assuming Rt, the recurrent Let us make simplifying assumptions.Suppose VN = amount recovered every year, and a nonrecurrent form 0 and technical progress over time. If the cost of the R1 recoverable in year 1. investment decreases regularly at a rate g, then, The principle that must prevail is that the dis- counted sum of the Rt's increased with the nonrecur- In = I0 /(1 + g)n rent part R1 be equal to the total C to be recovered, that is: By h, we mean the composite rate defined by (1 + T h) = (1 + i) * (1 + g). C = R1 + Rt Moreover, we retain operating costs constant over ( t=1 1+ i)t time, fn = f0. In the first case, the economic depreciation is equal In general, one accepts recovery of the specific to the constant annual installment equivalent to the interconnection costs (costs of co-localization, of con- investment costs. Technical progress comes down to nections between operators, and so forth) in the non- increasing the rate of discount i by a sum g, that is to recurrent part and the costs of network use (call say, taking it as equal to h. origination and termination) in the recurrent part. We then have: The model given here does not deal with specific costs. XN = I0 + f0 N j = I0 + f0 Nh Ni Nh since j = i (growth of operating costs nil). Thus, taking into account technical progress comes down to increasing the rate of remuneration of capital Notes 1. Geometric sum of reason 1/(1 + i) by putting a = 1/(1 + i),we obtain 1 ­ a = ia and: annuity by means of which the initial invest- If Sn = a + a2 + a3 + ... + an, then we have ment could be repaid if it were borrowed at the following equations: (1 + i)n - 1 = 1(1 - 1 a rate equal to the discount rate.The nomi- ) Sn ­ Sn i nal loss of usage value thus corresponds to ­1= an and a + aSn Sn = ­1 = Sn and that i(1 + i)n (1 + i)n gives: the share of repayment in capital of this an- Sn = a 1 - an 2. If a more financial interpretation is given, nuity. The usage value thus corresponds to 1 - a the economic depreciation represents the the capital not yet repaid. 62 Appendix 2: Capital Cost Approach Appendix 2: Capital Cost Approach invested upon, and on the financial market in demand. Activities where competition is liveliest usually entail a The cost of the operators' capital must reflect the higher risk. The cost of borrowing rd also varies opportunity cost of the funds invested in the compo- between activities and between companies but to a nents of the network and the other connected assets. lesser extent than the cost of equity capital re for a Traditionally it reflects the following elements: given financial market. As far as the structure of the · The average (weighted) cost of the indebtedness for capital (E and D), is concerned, it should also reflect the various means of financing available to each the balance sheet of each main activity.When there is operator. only one main balance sheet for several activities, it is · The cost of the equity capital, measured by the acceptable to assume that these activities share the same return the shareholders require, to invest in the net- capital structure. In this light, it can usually be assumed work--taking into account the risks tied to this that the cost of borrowing rd is the same for all activi- investment. ties, unless their results are judiciously different. · The value of the borrowed capital and the equity The WACC must be applied to a capital value for capital. the components of the network and the other related This information can then be used to determine assets, in order to determine the return to be attained, the weighted average cost of capital (WACC) accord- with the help of interconnection fees.While it is rela- ing to the following formula: tively easy to determine the value of borrowed capital and equity capital for an operator as a whole, it is not WACC = re × E/(D + E) + rd × D/(D + E) easy to determine these values for each of the opera- tor's activities.This is because decisions on financing by where re is the cost of the equity capital, rd is the cost of borrowing are to a large extent company decisions. borrowing,E is the total value of the equity capital,and They are determined by various factors, such as the D is the total value of the interest-producing debt. historical loan facilities and tax management consider- The calculation of the WACC for a given operator ations. It follows that the indebtedness of the company considered globally would be relatively direct,provided is liable not to correspond exactly to the financial possible arguments on the exact calculation and the needs of its various activities. value of the input data of the WACC formulas are set To fix prices, the regulators and the operators are aside. Nevertheless, it may be that the regulators must interested in the average capital employed during a establish whether the application of the global capital given period, rather than the capital employed at a cost represented by the WACC is appropriate for the given moment, for example, at the end of the financial regulated activities of the operators; when such is the year.This is justified by the fact that a "snapshot" of the case, the global WACC could serve to determine the situation, at a given time, will not likely represent the interconnection fees. average level of capital committed by the operator.To Otherwise, the regulators can take into account the be precise, the balance of working capital at a given fact that various risk premiums are normally applicable time cannot represent the average need for liquid assets to different activities, which could translate into differ- over a lengthy period.The separate operators' account- ences on the level of the cost of equity capital re, even ing must, therefore, indicate the average capital if the financial structure is the same. In that case, there engaged, and not an end-of-year balance. could be a different WACC for each branch of activity or each activity broken down (mobile telecommunica- Accounting Values and Market Values tions, cable television, or international services). The great increase in the valuation of telecommunica- The financial economy, and the actual behavior, of tions operators in 1999­2000 raised the question of the investors teache that the cost of equity capital re is equal choice of an accounting value for E (equity capital to the cost of borrowing without risk, to which is entered in the balance sheet) or of a market value added a risk premium that depends on the activity (stock market capitalization). It is generally accepted 63 A Model for Calculating Interconnection Costs in Telecommunications that the cost of capital must reflect the minimum what is called the market trend, linking the expected remuneration expected by the fund providers as a profitability to the level of risk.The return of the shares whole (shareholders and creditors). From this point of is then expressed as follows: view, the E value should reflect the stock market capi- talization as long as this is not the result of a speculative re = rf + × (r m - rf ) bubble. In fact, in this latter case, the financial markets are inefficient.Therefore,regulators and operators must where: agree on a sort of target financial structure where the rf is the risk-free return. capitalization is coherent with the expectation of rm is the average return expected on the market (for future profits. example, the return represented by a market reference This assessment is important. In fact, the indebted- index of the type of Dow Jones, S&P, CAC) (level of ness lever (D/E relation) can be very different depend- market return). ing on whether an accounting value of E or a stock is a weighting coefficient of the market differen- market value is considered; however, as re is greater tial (equity beta). than rd, this choice could exert a considerable impact on the WACC. It is generally accepted that interconnection repre- sents less risk than the fixed telephony activity, which Effect of Profits Tax itself represents less risk than mobile telephony. As in In the WACC formula, re represents the cost required general a rate of return corresponding to the global on the share capital and rd the cost of the financial activity is retained (for the reasons recalled at the start debt; this is situated before tax, that is to say that rd = rd of this appendix), the risk premium tends to decrease (1 ­ ), where rd is the actuarial rate of the financial when the interconnection activity increases. debt and q the rate of profits tax. If rf and rm are given by the financial market con- The use of WACC in determining tariffs is not in cerned, the regulator must assess the operator's beta common use and leads to adapting it. As it is defined coefficient, and adapt it, to take into account the share above, theWACC is net of tax. However, the tariff of a in its interconnection activity. , therefore, measures an service is fixed before tax. It is therefore necessary to elasticity of the sensitivity of the operator's security to increase theWACC by the rate of profits tax and to use variations in the market index:if the profitability of the the corrected WACC: market varies by 1 point, the profitability of the secu- rity varies by b points.In the African context,telecom- munications operators are generally considered as less WACC* = re E D (1+ ) × D + E+ rd* × D + E risky values than the average value of the market: is therefore generally lower than 1. where rd* represents the actuarial rate of the financial As far as rd* is concerned, it is generally calculated debt. as the risk-free return to which a premium (spread or debt premium) specific to the operator concerned is Appreciation of re and rd added.This return can be assessed either on the basis of re and rd are calculated on the basis of a reference rate the market values (risk-free return plus premium) or called rate without risk, corresponding in general to on the basis of contractual values presented by the the return on long-term (10-year) risk-free bonds operator (annual average weighted cost of the opera- (state bonds).To this rate a risk premium must be added tor's indebtedness) corresponding to the activity involved and to the type Thus, the WACC taken into account in the model of finance considered. is expressed as follows: The risk premium may be assessed on the past fluc- tuations of the security, or better on the fluctuations of WACC* = rf + ×(rm - rf) × E +(rf + sp)× D similar securities (sector assessment) so as to apprehend (1+ ) D + E D + E 64 Appendix 2: Capital Cost Approach To calculate it, therefore, we need: · D/(D + E), share of the debt on the total financial structure (which enables us to deduce E/(D + E); this ratio is sometimes called the level of gearing or leverage. · rf, risk-free return of the financial market consid- ered. · rm,average return of the financial market considered (expected growth of the reference stock market index). · , risk note of the security. · sp, risk premium of the operator. · , rate of profits tax. These six values are required to calculate the WACC. By default, values are proposed. For more details on the application of these concepts, refer to Alexander and others (1999). 65 A Model for Calculating Interconnection Costs in Telecommunications Appendix 3: Radio Concentrator Solutions Figure A3.1 Example of the Architecture of an IRT (TRT-Lucent Network) Today, there are only a small number of TDMA tech- nology radio concentrator manufacturers. In fact, the purchase of LucentTRT by SRT,the giving up of such solutions by the majors such as Alcatel or Siemens, means that SRT and NEC and a few small constructors from low-density countries (Australia) remain the last suppliers of these technologies, which are now chal- lenged by other solutions (fixed global system mobile, satellite, and so forth). Radio concentrators are particularly suitable for low-density rural zones with wide gaps between each village.They enable subscribers far from the exchange (up to 1,600 km) to be connected, and can tolerate Table A3.1 Number of Subscribers Depending on constraints specific to rural areas (for example, electric the Traffic per Subscriber (in mE) power provided by solar panels). A leap between sta- Traffic per subscriber Number of subscribers 50 1024 tions can be up to 50 km. 60 850 These systems can operate in the 500 MHz, 1.5 70 700 GHz, 2.5 GHz, and 3.5 GHz bands.The last leap can 80 650 90 600 be by wire or else wireless thanks to a wireless local 100 520 loop termination.These systems offer basic telephone 110 500 and public phone services, group 3 fax, data transmis- 120 470 130 450 sion services as well as basic Integrated system digital 140 420 network (2B+D) services. They use TDMA distribu- 150 400 tion, which enables the available spectrum to be opti- Note: It is assumed that 30 percent of calls are local; otherwise, the system loses between 5 percent and 15 percent of its subscribers. mized.The nodal and remote stations connected only use one frequency pair.They often useYagi antennas. The capacity of these systems can go up to 4,096 subscribers and traffic of 188 E. It should be noted that these calculations are carried out for a loss rate of 1 percent. In fact, the traffic capacity is greater because local calls do not use resources.Two subscribers con- nected to the same station can communicate without occupying a channel. The central station (CS) connects to the splitter of the automatic exchange of the public network.The CS is linked by microwave to the remote stations.The radio subassembly may be deported and linked by cable or radio at 2 Mbit/s. The repeater station (RS) serves as a relay between sites not seen by the central station. The repeater station can provide up to 120 circuits at 64 kbit/s or 240 circuits at 32 kbit/s. The terminal station (TS; SDE on the diagram) connects the subscribers.The terminal station can con- nect cabled subscribers or those with a DECT solution. 66 Bibliography The word processed describes informally re- ------.1998."The Incentive for Non-Price ------. 2001. Competition in Telecommunica- produced works that may not be commonly Discrimination by an Input Monopolist." tions. Munich Lectures, The MIT Press, available through libraries. International Journal of Industrial Organiza- 2nd ed. tion 16:271­84. Laffont, J. J., P. Rey, and J.Tirole. 1998."Net- Alexander, I.,A. Estache,A. Oliveri. 1999."A Economides,N.,and L.J.White.1995."Access work Competition: I. Overview and FewThingsTransport Regulators Should and Interconnection Pricing: How Effi- Non-discriminatory Pricing;II.Discrim- Know about Risk and the Cost of Capi- cient Is the `Efficient Component Pricing inatory Pricing." RAND Journal of Eco- tal."TheWorld Bank Institute. Processed. Rule'?" Antitrust Bulletin 40:557­79. nomics 29:1­56. Armstrong, M. 1998. "Network Intercon- Hausman, J. 1996. "Reply Affidavit of Prof. Salinger, M. 1998. "Regulating Prices to nection in Telecommunications." Eco- Jerry Hausman, FCC CC Docket No. Equal Forward Looking Costs: Cost- nomic Journal 108:545­64. 96-98" and "Valuation and the Effect of Based Prices or Price-Based Costs?" Jour- Armstrong, M., C. Doyle, and J. Vickers. Regulation on New Services in nal of Regulatory Economics 14:149­64. 1996. "The Access Pricing Problem: A Telecommunications." Brookings Papers Sidak, J. G., and D. Spulber. 1997. "The Synthesis." Journal of Industrial Economics on Economic Activity: Microeconomics. Tragedy of the Telecommons: Govern- 44:131­50. Jamison, M. 1998. "Regulatory Techniques ment Pricing of Unbundled Network El- Armstrong,M.,and J.Vickers.1998."TheAc- for Addressing Interconnection, Access, ements under the Telecommunications cess Pricing Problem with Deregulation: and Cross-Subsidy in Telecommunica- Act of 1996." Columbia Law Review A Note." Journal of Industrial Economics tions." In Margaret Arblaster and Mark A. 71:1081­161. 46:115­21. Jamison, eds., Infrastructure Regulation and Spulber, D., and J. G. Sidak. 1997."Network Benitez, D., A. Estache, M. Kennett, and C. Market Reform: Principles and Practice.Aus- Access Pricing and Deregulation." Indus- Ruzzier. 2000."Are Cost Models Useful tralian Competition and Consumer try and Corporate Change 6:757­82. for Telecoms Regulators in Developing Commission and the Public Utility Re- Valletti,T.1998."The Practice ofAccess Pric- Countries?" Policy Research Working search. ing: Telecoms in the UK." The World Paper 2384.TheWorld Bank Institute. ------. 1999."Does Practice Follow Princi- Bank, EDIRP. Processed. BIPE. 2000."What Telecom Regulation for ple? Applying Real Options PrinciplesTo ------.2001."TheTheory of Access Pricing Low-Income African Countries?" Study Proxy Costs in U.S. Telecommunica- and Its Linkage with Investment Incen- for the European Commission. Available tions." In James Alleman and Eli Noam, tives." In Martin Cave, Papers on Access at http://europa.eu.int/ISPO/intcoop/ eds., Real Options: The New Investment Pricing Investment and Entry in i_acp.html. Theory and its Implications for Telecommuni- Telecommunications. Warwick Business BIPE. 2003. Field Units Reports: cations Economics. Boston: Kluwer Acade- School, Research paper series. Cameroon, Côte d'Ivoire, Zambia, Burk- mic Publishers. Valletti,T.,andA.Estache.1998."TheTheory ina Faso. Laffont,J.J.1998."Translating Principles into of Access Pricing:An Overview for Infra- Celani, M. O. D. Petrecolla, and C. A. Practice." EDI Regulatory Reform Dis- structure Regulators." The World Bank Ruzzier. 2002. "Desagregación de redes cussion Paper. World Bank, Washington, Institute.Available at http://www.world- en telecomunicaciones. Una visión desde D.C. http://www.worldbank.org/wbi/ bank.org/wbi/regulation/wp.htm. la política de defensa de la competencia." regulation/wp.htm. Wildman, S. S. 1997."Interconnection Pric- Processed. Laffont, J. J., and J. Tirole. 1996. "Creating ing, Stranded Costs, and the Optimal Economides, N. 1996. "The Economics of Competition through Interconnection: Regulatory Contract." Industrial and Cor- Networks." International Journal of Indus- Theory and Practice." Journal of Regula- porate Change 6:741­55. trial Organization 14:673­99. tory Economics 10:227­56. 67 T he liberalization of the telecommunications markets in Sub-Saharan Africa led to increased competition on the provision and pricing of communication services. But, due to the lack of appropriate regulatory tools, newly established regulators are poorly equipped to arbitrate increasing interconnection disputes between competing operators. This guidebook and its associated CD-ROM, including the cost model, were prepared to provide Sub-Saharan Africa regulators and operators with a sound regulatory tool allowing the determination of accurate interconnection costs, thus facilitating the settlement of lengthy and costly interconnection disputes between fixed and mobile operators. The cost model belongs to the family of "Bottom-Up" models, which calculate interconnection cost incurred by an efficient operator using the Long Run Incremental Cost (LRIC) methodology. The proposed cost model takes into account most features characterizing the development stage of telecommunications networks in Sub-Saharan Africa (small size of fixed network, importance of rural telephony, excessive reliance on microwave technology, explosive demand for mobile service, and weak regulatory capacity). THE WORLD BANK PPIAF 1818 H Street, NW PPIAF Program Management Unit Washington, DC 20433 c/o The World Bank U.S.A. 1818 H Street, NW, MSN I9-907 Telephone: 202.473.1000 Washington, DC 20433 Fax: 202.477.6391 U.S.A. Web Site: www.worldbank.org Telephone: 202.458.5588 E-mail: feedback@worldbank.org Fax: 202.522.7466 Web Site: www.ppiaf.org E-mail: info@ppiaf.org TMxHSKIMBy356715zv,:&:':=:" ISBN 0-8213-5671-2