66236 D I R E C T I O N S I N D E V E LO P M E N T Infrastructure Airport Economics in Latin America and the Caribbean Benchmarking, Regulation, and Pricing Tomás Serebrisky Airport Economics in Latin America and the Caribbean Airport Economics in Latin America and the Caribbean Benchmarking, Regulation, and Pricing Tomás Serebrisky © 2012 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org 1 2 3 4 14 13 12 11 This volume is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dis- semination of its knowledge, this work may be reproduced, in whole or in part, for noncom- mercial purposes as long as full attribution to the work is given. For permission to reproduce any part of this work for commercial purposes, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN (paper): 978-0-8213-8977-5 ISBN (electronic): 978-0-8213-8933-1 DOI:10.1596/978-0-8213-8977-5 Library of Congress Cataloging-in-Publication Data has been applied for. Cover photo provided by the author (courtesy of ADC&HAS). Cover design by Debra Naylor. Contents Preface xiii Acknowledgments xvii About the Author xix Abbreviations and Acronyms xxi Overview 1 The Air Transport Sector 2 Investment in the LAC Airport Sector 4 Summary of This Report 6 Conclusions 16 Note 19 References 20 Chapter 1 Recent Evolution of the Air Transport Sector 21 Latin America and the Caribbean Overview 26 Notes 34 References 34 Chapter 2 Investment in the Airport Sector 35 Private Project Financing in the Airport Sector Worldwide 42 v vi Contents Private Investments in the Airport Sector in Developing Countries 44 Private Investment in the Airport Sector in Latin America and the Caribbean 49 Conclusions 52 Notes 53 References 54 Chapter 3 Efficiency Estimation 55 Partial Performance Indicators in LAC Airports: Cross-Comparison for 2005 60 Partial Performance Indicators: Time Series 84 Measuring Technical Efficiency of Airports in LAC Countries 111 Conclusion 130 Notes 132 References 135 Chapter 4 Institutional Design and Governance of Airport Regulators in Latin America 137 Literature Review 139 Methodology and Data Sources 141 Regulatory Governance 143 Economic Regulation 157 Conclusions 161 References 163 Chapter 5 Benchmarking of Aeronautical Charges at Latin American Airports 165 Overview 165 Methodology 167 Conclusion 195 Notes 197 References 198 Appendix A Survey of Airport Performance for Operators 199 Appendix B Governance of Airport Regulators Survey 209 Appendix C Technical Efficiency Calculation 237 Contents vii Appendix D Data Sources 245 Air Transport Research Society (ATRS) 245 Airports Council International (ACI) 246 Private Participation in Infrastructure Database (PPI) 246 Dealogic ProjectWare Database 247 Asociación Latinoamericana de Transporte Aéreo (ALTA) 248 Airport Charges 249 Figures 1 Air Transport Sector Demand and World GDP, 1980–2008 2 2 Domestic and International Passenger Share, 2008 3 3 Partial Performance Indicator: Passengers per Employee, 2005 8 4 Evolution of Turnaround Costs for an Airbus A320, 1995–2009 15 5 Structure of Turnaround Costs for an Airbus A320 17 1.1 Growth Rates in the Air Transport Sector and Global GDP, 1980–2008 22 1.2 Passenger Traffic Growth, by Region, 2007 and 2008 23 1.3 Domestic and International Passenger Share, 2008 24 1.4 Volume of Cargo Moved, by Region, 2008 25 1.5 Aircraft Movements, by Region, 2008 26 1.6 GDP Growth and Passenger Growth in LAC, 1995–2008 27 2.1 Project Financing in the Airport Sector by Number of Projects, Total Project Amount, and Region, 1996–2008 43 2.2 Share of Project Financing in the Airport Sector, by Region, 1996–2008 44 2.3 Private Investment Commitments to Infrastructure Projects in Developing Countries, by Sector, 1990–2008 47 2.4 Investment Commitments to Transport Projects with Private Participation in Developing Countries, by Subsector, 1990–2008 48 2.5 Total Investment Commitments to Airport Projects with Private Participation in Developing Countries, by Region, 1991–2008 49 viii Contents 2.6 Investment Commitments to Airport Projects with Private Participation in Developing Countries, by Type of Project, 1991–2007 50 2.7 Investment Commitments to Transport Projects with Private Participation in Latin America and the Caribbean, by Subsector, 1990–2008 51 2.8 Investment Commitments to Airport Projects with Private Participation in Latin America and the Caribbean Countries, by Type of Investment, 1993–2008 52 3.1 Passengers per Aircraft Movement, 2005 63 3.2 Cargo per Aircraft Movement, 2005 65 3.3 Passengers per Employee, 2005 67 3.4 Aircraft Movements per Runway, 2005 68 3.5 Labor Costs as a Share of Operating Costs, 2005 70 3.6 Labor Cost per Passenger, 2005 71 3.7 Operating Costs per Passenger, 2005 73 3.8 Operating Costs per Aircraft Movement, 2005 75 3.9 Total Revenue per Passenger, 2005 76 3.10 Aeronautical Revenue Share, 2005 78 3.11 Aeronautical Revenue per Aircraft Movement, 2005 79 3.12 Passengers per Boarding Bridge, 2005 81 3.13 Passengers per Square Meter of Terminal Area, 2005 82 3.14 Evolution of the U.S. Dollar–Euro Exchange Rate, 1999–2009 89 3.15 Passengers per Employee 91 3.16 Labor Costs per Passenger 95 3.17 Operating Costs per Passenger 98 3.18 Total Revenue per Passenger 102 3.19 Total Revenue per Employee 105 3.20 Passengers per Boarding Bridge 108 3.21 DEA-CRS and DEA-VRS Frontiers 113 3.22 Malmquist Index of Total Factor Productivity Change 125 4.1 Decision-Making Autonomy 146 4.2 Appointment Authorities 147 4.3 Budget Composition 149 4.4 Procedure to Remove Decision Makers 150 4.5 Reasons Directors Leave Positions 151 4.6 Bureaucratic Quality 152 4.7 Bureaucratic Quality by Type 153 Contents ix 4.8 Transparency in Airport Regulators 155 4.9 Transparency by Type 156 4.10 Dimensions of Accountability in Airport Regulators 157 4.11 Dimensions of Accountability in IRAs and Non-IRAs 158 5.1 Landing Fees for an Airbus A320, Daylight Operation 170 5.2 Landing Fees for an Airbus A320, Night Operation 171 5.3 Changes in Landing Fees for an Airbus A320, Daylight Operation 172 5.4 Landing Fees Percentage Change for an Airbus A320, Daylight Operation 174 5.5 Parking Charge for an Airbus A320, for 2 Hours 175 5.6 Changes in Parking Charges for an Airbus A320, for 2 Hours 176 5.7 Landing Fees and Parking Charge for an Airbus A320, for 2 Hours, 2009 178 5.8 Landing Fees and Parking Charge for an Airbus A320, for 2 Hours, 1995–2009 179 5.9 Boarding Bridge Charges for an Airbus A320, for 2 Hours, 2009 180 5.10 Boarding Bridge Charges for an Airbus A320, for 2 Hours, 1995–2009 182 5.11 Passenger Charges per Passenger (Charges Levied by the Airport) 183 5.12 Charges and Taxes Levied on Passengers, per Passenger 184 5.13 Changes in Passenger Charges per Passenger (Charges Levied by the Airport) 187 5.14 Turnaround Costs for an Airbus A320 (2 Hours, Daylight Operation) 188 5.15 Changes in Turnaround Costs for an Airbus A320 (2 Hours, Daylight Operation) 190 5.16 Turnaround Costs for a Boeing 767-300 (2 Hours, Daylight Operation) 191 5.17 Changes in Turnaround Costs for a Boeing 767-300 (2 Hours, Daylight Operation) 192 5.18 Turnaround Costs Levied on Airlines for an Airbus A320 (2 Hours, Daylight Operation) 193 5.19 Changes in Turnaround Costs Levied on Airlines for a Boeing 767–300 (2 Hours, Daylight Operation) 194 5.20 Turnaround Costs Levied on Passengers, for an Airbus A320 196 x Contents Tables 1 LAC Region’s Share of the Air Transport Sector, 2008 4 2 Private Investment Commitments to the Airport Sector in the LAC Region, 1993–2008 5 3 Criteria for Determining Regulatory Agency Governance Ratings 12 1.1 Latin America and the Caribbean Snapshot of the Airport Sector, 2008 29 1.2 Global and LAC Airports Ranking: Passengers, Cargo, and Aircraft Movements, 2008 30 1.3 LAC Airport Ranking (Top 10) by Cargo, 2008 33 1.4 LAC Airport Ranking (Top 10) by Aircraft Movements, 2008 33 2.1 Latin American and Caribbean Airports by Type of PSP Arrangement 36 2.2 Total Project Financing in the Airport Sector by Income Level, Region, and Country, 1996–2008 45 3.1 Partial Performance Indicators Commonly Used in the Airport Sector 57 3.2 Latin American and Caribbean Airports Sampled 61 3.3 Summary of Airport Partial Performance Indicators—Top and Bottom Performers, 2005 85 3.4 Descriptive Statistics by World Region, 2005–06 115 3.5 Average Technical Efficiency Scores and Scale Efficiency by Region, 2005–06 115 3.6 Average Technical Efficiency Scores for LAC Airports, 2005–06 117 3.7 Peer Analysis, DEA VRS, 2005 119 3.8 Potential Explanatory Factors of Technical Inefficiency, 2005–06 121 3.9 Truncated Regression—Marginal Effects 123 3.10 Descriptive Statistics by Period 126 3.11 Average Annual Total Factor Productivity by Airport and Subperiod 127 3.12 Average Total Factor Productivity by Airport Categories 129 3.13 Malmquist Total Factor Productivity Index Decomposition—Averages by Period 130 4.1 Aspects of Governance of Airport Regulators 142 4.2 Mapping of Regulator and Legal Configuration 144 Contents xi 4.3 Answers to Selected Questions on Economic Regulation in the Airport Sector 160 5.1 Airport Sample Used for the Aeronautical Tariff Benchmarking Analysis 166 5.2 Key Parameters of the Aircraft Used in the Analysis 168 5.3 Passenger Charges and Taxes per Departing Passenger 185 C.1 Results for the Technical Efficiency Scores for All Airports Other Than Latin American Airports 237 C.2 LAC Airports Total Factor Productivity Change 241 C.3 Average Technical Efficiency Scores and Scale Efficiency by Region (2005–06 average) 243 Preface Expanding and enhancing the provision of air transport infrastructure has become an increasingly important policy issue on the development agenda of both high-income and developing countries. The growth of air transport demand, along with the associated need to have efficient airport infrastructure to support it, has prompted the need to evaluate the effects of ownership schemes and regulation on airport performance. Traditionally, air transport infrastructure was exclusively under govern- ment ownership and management in the Latin America and Caribbean (LAC) region. Starting in the late 1990s, private capital flows began to play an increasingly important role through the financing of air transport sector infrastructure and the management of airport operations. The intro- duction of private sector participation responded to myriad policy objec- tives, including bringing innovation and efficiency to the management of airports and boosting resources to finance the growing demand for airport infrastructure expansions and maintenance. In this context, governments have undertaken important institutional and regulatory reforms, which in several countries have resulted in the separation of planning and policy formulation functions from the day-to-day operation of airports through the establishment of independent regulatory agencies. As a global pioneer in the introduction of private sector participation in air transport infrastructure, the LAC region serves as an informative xiii xiv Preface context through which to investigate the evolution of performance in the airport sector and answer a series of pertinent policy questions: Are LAC airports technically efficient? How has efficiency evolved in the past decade? Are privately run airports more efficient than state-operated air- ports? How do independent regulators compare with government agen- cies in accountability, transparency, and autonomy? How have the level and structure of airport tariffs changed in recent years? Purpose of the Report This report presents the findings of a first-ever, comprehensive study of how LAC region airports have evolved during a notable period of transi- tion in airport ownership. It is an unbiased, positive analysis of what hap- pened, rather than a normative analysis of what should be done to reform and to attract private sector participation to the airport sector. It takes the first step to respond to the need for more conclusive information about the influence of airport ownership on economic performance. The report centers on the study of three dimensions of performance: productive effi- ciency, institutional setup for the governance of the sector, and financing. Structure of the Report This multifaceted report uses a range of advanced quantitative and qualitative methods to assess the relationship between airport ownership and performance in the LAC region. After a comprehensive overview, chapters 1 and 2 provide the necessary background for the air transport sector and the evolution of private sector participation and investment in airport infrastructure. In chapter 3, questionnaires submitted to airport operators and regulators led to the creation of the unique data sets, which were first used to compare performance across 14 partial performance indicators, and next used to develop aggregate measures of efficiency necessary for the benchmarking exercise. In chapter 4, a qualitative study of the relationship between type of regulating agency (independent or government-led) and transparency, accountability, and bureaucracy pro- vides insight into how recent reforms have also affected the quality of regulatory governance. Chapter 5 provides an in-depth analysis of the evolution of tariff structures in the region as compared to a sample of international airports. Although this report considers Latin America and the Caribbean as its focal region, the questions raised, and the analytical tools employed to Preface xv respond to those questions, may be applied to other regions. In the future, researchers seeking to evaluate the productive performance of airports can use this study as a guide to anticipate potential challenges as well as to develop successful strategies to overcome them. Several important topics were not included in this report but should be the focus of future research. In particular, the evolution of the quality of services in airports deserves greater attention, as airports are increasingly becoming business centers and key gateways for trade competitiveness. The other main topic that requires detailed practical research is climate change and its relation- ship with the airport sector. Acknowledgments This study was produced by a task team led by Tomás Serebrisky, of the Sustainable Development Department in the Latin America and the Caribbean Region of the World Bank. Members of the core team were Sebastián López Azumendi, Matías Herrera Dappe, Raquel Fernandez, and Juan Matías Ortner. The early preparatory stages of the report benefited from inputs and advice provided by Raúl Medina Caballero (Ministry of Transport, Spain). The study was conceived by a group that included Tomás Serebrisky, Luis Andrés, and José Luis Irigoyen of the World Bank. Several individuals contributed to the preparation of the report, including Sebastián López Azumendi (analysis of governance of airport regulators), Sergio Perelman (calculation of aggregate measures of techni- cal efficiency), and Andy Ricover (benchmarking of airport tariffs). Diana Cubas, Gwyneth Fries, and Sivan Tamir edited the report and provided suggestions on improving its organization. The report benefited extensively from discussions and feedback pro- vided by Jean François Arvis, Raúl Medina Caballero, Baher El-Hefnawy, Antonio Estache, Shomik Raj Mehndiratta, Aurelio Menéndez, Charles Schlumberger, and Jordan Schwartz. xvii xviii Acknowledgments The author would like to express his gratitude to all individuals responding to the questionnaires. Regulators and airport operators spent valuable time completing the questionnaires and addressing in detail the follow-up clarifications. Financial support for the preparation of this report was provided by the Public-Private Infrastructure Advisory Facility (PPIAF). About the Author Tomás Serebrisky received a Ph.D. in economics from the University of Chicago in 2000. From 2000 to 2002 Mr. Serebrisky worked in Argentina as the Chief Economist of the Antitrust Commission and as a Professor in Universidad Torcuato Di Tella. In 2002 he joined the World Bank and is currently working as Senior Infrastructure Economist in the Latin American Region. His areas of expertise are the economics of infrastruc- ture investments, public-private partnerships, logistics, economic regula- tion, and antitrust. Mr. Serebrisky has published extensively in refereed journals, including: Journal of International Economics, Transport Reviews, Journal of Maritime Policy and Management, Telecommunications Policy, Journal of Air Transport Management and World Competition. xix Abbreviations and Acronyms ACI Airports Council International AIP Aeronautical Information Publication ALTA Asociación Latinoamericana de Transporte Aéreo (Latin America and the Caribbean Air Transport Association) ANAC Agencia Nacional de Aviação Civil (National Civil Aviation Agency of Brazil) ATI air transport infrastructure ATM air traffic movement ATRS Air Transport Research Society BOT build, operate, and transfer BROT build, rehabilitate, operate, and transfer CAA Civil Aviation Authority, Panama CPI Consumer Price Index CRS constant returns to scale DEA data envelopment analysis DINACIA Dirección Nacional de Aviación Civil e Infraestructura Aeronaútica (Uruguay) IATA International Air Transport Association ICAO International Civil Aviation Organization INFRAERO Empresa Brasileira de Infra-Estrutura Aeroportuaria (Brazilian Airport Administrator) xxi xxii Abbreviations and Acronyms IRA independent regulatory agency IRR internal rate of return LAC Latin America and the Caribbean MTOW maximum takeoff weight OECD Organisation for Economic Co-operation and Development PPI Private Participation in Infrastructure (World Bank database) PPIAF Public-Private Infrastructure Advisory Facility PSP private sector participation RFI Regulatory Framework Index RLT rehabilitate, lease or rent, and transfer ROT rehabilitate, operate, and transfer SFA Stochastic Frontier Analysis TC technical change TE technical efficiency TEC technical efficiency change TFP total factor productivity TFPC total factor productivity change VRS variable returns to scale WLU workload unit Abbreviations and Acronyms xxiii Airport Codes AEP Aeroparque Jorge Newbery, Buenos Aires, Argentina ASU Silvio Pettirossi International, Asunción, Paraguay ATL Hartsfield-Jackson Atlanta International, United States BAQ Ernesto Cortissoz International, Barranquilla, Colombia BOG El Dorado International, Bogotá, Colombia BSB Presidente Juscelino Kubitschek International, Brasilia, Brazil CCS Simón Bolivar International, Caracas, República Bolivariana de Venezuela CDG Charles de Gaulle International, Paris, France CGH Congonhas International, São Paulo, Brazil CLO Alfonso Bonilla Aragón International, Cali, Colombia CUN Cancún International, Cancún, Mexico EZE Ministro Pistarini International, Buenos Aires, Argentina FRA Frankfurt am Main International, Frankfurt, Germany FTE El Calafate Airport, Argentina GDL Miguel Hidalgo y Costilla International, Guadalajara, Mexico GIG Antonio Carlos Jobim International (Galeão), Rio de Janeiro, Brazil GRU Governador André Franco Montoro International, Guarulhos, São Paulo, Brazil GUA La Aurora International, Guatemala City, Guatemala GYE José Joaquín de Olmedo International, Guayaquil, Ecuador ICN Seoul Incheon International, Republic of Korea JFK John F. Kennedy Airport, New York, United States KIN Norman Manley International, Kingston, Jamaica LAX Los Angeles International, Los Angeles, United States LHR Heathrow International, London, United Kingdom LIM Jorge Chávez International, Lima, Peru LPZ El Alto International, La Paz, Bolivia MAD Barajas International, Madrid, Spain MAO Brigadeiro Eduardo Gomes International, Manaus, Brazil MDE José María Córdova International, Medellín, Colombia MEM Memphis International, United States MEX Benito Juárez International, Mexico City, Mexico MIA Miami International, Miami, United States xxiv Abbreviations and Acronyms MFM Macau International, Macao SAR, China MGA Augusto C. Sandino International, Managua, Nicaragua MTY General Mariano Escobedo International, Monterrey, Mexico MVD General Cesareo Berisso International, Carrasco, Montevideo, Uruguay NAS Lynden Pindling International, Nassau, The Bahamas POS Piarco International, Port of Spain, Trinidad and Tobago PTY Tocumen International, Panama City, Panama SAL Comalapa International, San Salvador, El Salvador SCL Comodoro Arturo Merino Benítez International, Santiago de Chile, Chile SDF Louisville International, United States SDQ Las Américas International, Santo Domingo, Dominican Republic SJO Juan Santamaría International, San José, Costa Rica SNA John Wayne Airport, Santa Ana, United States TGU Toncontín International, Tegucigalpa, Honduras UIO Mariscal Sucre International, Quito, Ecuador VCP Viracopos-Campinas International, São Paulo, Brazil VVI Viru Viru International, Santa Cruz, Bolivia XMN Xiamen Gaoqi International, China Overview As core components of the air transport sector, airports play a key role in catalyzing social and economic development at the regional, national, and global levels. As a dynamic service industry with multiple inputs and out- puts, the airport sector facilitates domestic and international trade (by providing access to markets); creates employment opportunities related to both aeronautical and nonaeronautical activities; and enhances communi- cation and integration between people, countries, and cultures through tourism, business activities, and merchandise trade. Airports operate in different environments (large cities, remote areas) and have users with varying needs (business and leisure travelers), thus making efficiency assessments very challenging. Multiple stakeholders, including airlines, regulatory agencies, ground-handling companies, and many others, have varied interests and objectives that further complicate an evaluation of airport performance. This overview includes developed countries, such as Japan and Australia, in the World Bank regional designation of East Asia and Pacific. 1 2 Airport Economics in Latin America and the Caribbean The Air Transport Sector The air transport sector is uniquely volatile (figure 1). Over time, its fluc- tuations have followed those of the global economy, though they have been more intense. Heavily dependent on business activity, trade flows, and tourism, the sector has experienced long periods of continued growth alternated with brief crisis periods of negative growth. This amplifying effect has meant that global crises, such as the 1979 oil crisis; the Gulf War in 1990; the terrorist attacks of September 11, 2001; and the 2008 global financial crisis had a profoundly negative impact on the air transport sector as compared to other sectors of the economy. Among relevant stakeholders in the air transport sector, air- lines are particularly sensitive to severe global downturns. The progres- sive liberalization of different aviation markets, notably in the European Union and the United States in the late 1990s and 1970s, respectively, led to an overall increase in competition and to narrower operating mar- gins, which further increased the particular vulnerability of airlines. Airports themselves, with facilities that can often be classified as natural monopolies, are less sensitive to these effects. The air transport sector (in terms of passenger and cargo demand) is dominated by Europe and North America (Canada and the United States), Figure 1 Air Transport Sector Demand and World GDP, 1980–2008 15 13 11 9 7 percent 5 3 1 –1 –3 –5 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 passengers GDP per capita Source: World Bank estimation based on data from Airports Council International (ACI), International Air Transport Association (IATA), and International Monetary Fund (IMF) data. Overview 3 which together account for more than 60 percent of the market (figure 2). Airports handled 4.874 billion arriving and departing passengers in 2008, of which approximately 2 billion were international and 2.8 billion were domestic. Of these, North America (Canada and the United States) repre- sented 48 percent of domestic traffic, while Europe represented more than half of global international traffic. The share of passengers, and especially Figure 2 Domestic and International Passenger Share, 2008 a. Domestic passengers (% of global share by region) Africa 2% Asia-Pacific 26% North America 48% Europe 15% Latin America and Middle East the Caribbean 1% 8% b. International passengers (% of global share by region) North Africa America 5% 10% Middle East 5% Latin America and Asia-Pacific the Caribbean 21% 6% Europe 53% Source: Author’s estimation based on ACI data. 4 Airport Economics in Latin America and the Caribbean of cargo in North America and Europe, has fallen slightly (5 percent) since 2000, with the East Asia and Pacific region picking up most of the gains, primarily because of the significant increase of air traffic demand in China. In 2008, the East Asia and Pacific region accounted for 38 percent of the cargo market (measured in volume), while North America was second with 33 percent. Latin America and the Caribbean (LAC) accounts for a small share of the air transport sector worldwide. Based on 2008 figures, the region only accounted for 7 percent of total passengers, 5 percent of cargo, and 8 per- cent of aircraft movements (table 1). Airports are relatively small when ranked on a global scale. LAC has a total of just 4 airports among the top 100 airports worldwide and 14 airports among the top 200. Aeropuerto Internacional Benito Juárez in Mexico City, ranked 43rd globally, is the most important airport in the region in terms of passenger traffic, han- dling a total of about 26.2 million passengers in 2008 (approximately three times less than the number handled by first-ranked Hartsfield- Jackson Airport in Atlanta). As for cargo, the entire LAC region handled a total of 4.6 million metric tons in 2008, only 1 million metric tons more than the amount of cargo traffic handled by the global leader, Hong Kong International Airport (3.6 million metric tons) and three times as much as Miami, North America’s cargo hub (1.5 million metric tons). Investment in the LAC Airport Sector For much of the 20th century, commercial and business pressures were weak within the airport sector since airports around the world were not only owned and managed by governments, but also seen solely as public Table 1 LAC Region’s Share of the Air Transport Sector, 2008 Domestic passengers 224,531,098 Share of global domestic passengers 8% International passengers 113,850,200 Share of global international passengers 6% Total passengers 338,381,298 Share of global total passengers 7% Growth rate of total passengers (2007–08) 8% Cargo (metric tons) 4,589,092 Share of global cargo 5% Growth rate of cargo (2007–08) 4% Share of global aircraft movements 8% Growth rate of aircraft movements (2007–08) 0% Source: Author’s estimation based on Airports Council International data. Overview 5 utilities and strategic assets for national defense purposes. During the late 1980s and early 1990s, however, there was a slow shift toward a view of airports as more commercially oriented enterprises. Consequently, several countries introduced private sector participation (PSP) into the operation of airports. According to the ProjectWare database, US$64 billion in private investment went to a total of 110 air transport infrastructure projects between 1996 and 2008. Australia; Hong Kong SAR, China; and Turkey led the globe over the studied time period, representing 57 percent of total project financing. Australia has clearly been leading, with total proj- ect financing of US$19,326 million, followed by Hong Kong SAR, China, with US$11,050 million and Turkey with US$6,188 million. From 1993 to 2008, the private sector invested more than US$9.5 billion in the LAC region’s airports. Argentina, Colombia, and Mexico together represented almost 80 percent of total investments in the LAC region (table 2). Compared to other regions, LAC was a pioneer in introducing PSP in the airport sector, though the intensity of the process has decreased dra- matically in recent years. According to the World Bank’s Private Participation in Infrastructure (PPI) Database, which is perhaps the most complete public source of information on private investment in infra- structure, within just the developing world, the LAC region accounted for 30 percent of total investment commitments in the airport sector between 1991 and 2008. However, the relative share of total private investment in Table 2 Private Investment Commitments to the Airport Sector in the LAC Region, 1993–2008 Investments Country (US$ millions) Share of total (%) Mexico 3,223.9 33.9 Argentina 2,375.4 25.0 Colombia 1,224.3 12.9 Ecuador 665.0 7.0 Peru 430.0 4.5 Dominican Republic 350.0 3.7 Chile 345.0 3.6 Uruguay 195.0 2.0 Jamaica 175.0 1.8 Costa Rica 161.0 1.7 Venezuela, RB 134.0 1.4 Honduras 120.0 1.3 Bolivia 116.6 1.2 Source: Private Participation in Infrastructure (PPI) Database. 6 Airport Economics in Latin America and the Caribbean LAC fell from 70 percent of commitments in the late 1990s to only 12 percent between 2000 and 2008. The PPI database reports that invest- ments in the airport sector in LAC peaked in 2006 with a total of US$2,346 million, but fell to US$746 million in 2007 and US$231 mil- lion in 2008. This reduction could be the result of the successful upgrade of airport infrastructure, or it could also be that the region lost its attrac- tiveness or that individual country governments have decided not to open the sector for new or more private investment. For example, as this report was being written, Brazil had yet to decide whether to open its airport sector to PSP. Summary of This Report This report presents the findings of a first-ever, comprehensive study of how LAC region airports have evolved during a notable period of transi- tion in airport ownership. It is an unbiased, positive analysis of what hap- pened, rather than a normative analysis of what should be done to reform the airport sector or to attract and structure PSP. It takes the first step to respond to the need for more conclusive information about the influence of airport ownership on economic performance and the measurable side of operational performance. The report is centered on the study of three dimensions of performance: productive efficiency, institutional set up for the governance of the sector, and financing. The analytical weight is divided into three chapters. In chapter 3, a benchmarking exercise provides a thorough analysis of the technical per- formance of LAC region airports. Chapter 4 compares the performance of independent regulatory agencies and government regulatory agencies as it relates to transparency, accountability, and the quality of their bureaucracies. Chapter 5 investigates the growth and change of airport tariff levels within the LAC region. Efficiency Performance: A Benchmarking Approach The use of benchmarking to measure performance in the transport sec- tor and airport subsector, more specifically, is relatively new. Increased PSP in the 1990s led to a call for a more thorough evaluation of airport performance, both (a) to negotiate the terms and conditions of private involvement and (b) to track the improvements or lack thereof resulting from such involvement. As a result of this process, in the late 1990s, benchmarking began to be accepted as an important management tool within the airport industry. However, current papers using advanced Overview 7 efficiency techniques neglect Latin American airports, focusing instead on those of Asia, Europe, and North America. This report is a first attempt to bring this kind of advanced analysis to the LAC region, and includes four separate but complementary sections: (a) an investigation of techni- cal efficiency using partial performance indicators, (b) the positioning of LAC airports on a global efficiency frontier, (c) an analysis of the relation- ship between airport performance and selected socioeconomic factors and unique airport characteristics, and (d) an assessment of the evolution of the airports’ productivity in the LAC region from 1995 to 2007. The first part of chapter 3 investigates airport efficiency through par- tial performance indicators, which are widely used not only in the airport sector but also in other infrastructure sectors, such as water and electricity and telecommunications.1 First, 14 partial performance indicators from 2005 put the LAC region in a global perspective through a comparison of mean levels for the East Asia and Pacific region, Europe, and North America. Second, an analysis of how these indicators changed over the period from 1997 to 2005 provides some insight into how the advent of PSP affected the technical efficiency of the region’s airports. Responses to an original questionnaire from a representative sample of LAC airports that covers more than 80 percent of passengers and aircraft movements and 70 percent of air cargo allowed for a global comparison with partial performance data collected by the Air Transports Research Society for its periodic reports calculating airport technical efficiency in Asia, Europe, and North America. Figure 3 shows results for one partial performance indicator, passengers per employee, which is taken as an example for this overview. For this particular indicator, Comodoro Arturo Merino Benítez International in Santiago, Chile, and Congonhas International Airport in São Paulo, Brazil, are the top performers. Partial performance indicators in the airport sector should be interpreted with extreme care. In a multi- input and multi-output service industry, like airports, they do not allow for a conclusive identification of performance. For example, a high num- ber of passengers per employee could represent either high efficiency or low quality of service. This particular stage in the analysis revealed a great deal of variation in the performance of LAC region airports. However, Congonhas International Airport (CGH) in São Paulo, Brazil; Cancún International Airport (CUN) in Mexico; and Comodoro Arturo Merino Benítez International (SCL) in Santiago, Chile, were the airports that most fre- quently appeared among the top three performers in the 14 partial per- formance indicators calculated. 8 Airport Economics in Latin America and the Caribbean Figure 3 Partial Performance Indicator: Passengers per Employee, 2005 SCL CGH CLO AEP CUN MTY GDL BSB SJO MEX LIM EZE GRU GYE FTE SDQ BAQ GIG SAL MAO PTY VCP mean LAC mean NA mean EU mean AP 0 10 20 30 40 50 60 70 80 90 # of passengers, thousands airports publicly operated airports privately operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. Overview 9 The second section of chapter 3 conducts an analysis of efficiency using aggregate measures and econometric techniques to compute a global efficiency frontier for the airport sector and to identify the position of Latin American airports relative to the best practice worldwide. The Data Envelope Analysis (DEA) method used for this stage positions LAC region airports around the frontier relative to best-performing peers of the same scale. Results relating to technical efficiency in global perspec- tive reinforced the findings of the initial analysis of partial performance indicators. Privately operated airports were positioned closer to the fron- tier than were their publicly operated counterparts, though this effect was not significant across all the different specifications tested. Two final tests round out the benchmarking exercise in chapter 3. First, a truncated regression was performed to investigate the relationship between socioeconomic factors and airport performance, using the aggre- gate technical efficiency measures from the previous section. Results suggest that variation in technical efficiency is largely the result of factors exogenous to airport management. The models identified hub airports and population size as the main drivers of technical efficiency in the air- port sector. Hub airports are, on average, 10 to 15 percent more efficient than other airports. Airports located in areas with more than 5 million inhabitants are 17 to 20 percent more efficient than airports that serve less populated areas. The only variable within the control of airport man- agement that appeared to drive technical efficiency was the proportion of revenue acquired through sources other than aeronautical tariffs. Those airports that rely on sources other than aeronautical tariffs tend to be more efficient. This relationship could not be used to make further conclusions on the relationship between type of ownership and airport efficiency, because both public and private airports surveyed varied con- siderably in terms of the proportion of total revenue acquired from aero- nautical tariffs. Finally, a Malmquist quantity index of total factor productivity change shows how airport productivity has changed across three sequences: 1995 to 1999, 2000 to 2003, and 2003 to 2007. From 2003 to 2007, strong average annual productivity growth (3.9 percent) of the airport sector reflected the strong economic growth of the region as a whole. Larger airports tend to register faster productivity growth. Both publicly and privately operated airports performed similarly over the three time peri- ods, with publicly operated airports performing slightly better over the whole period. The Malmquist index requires panel data for each unit sampled. Because this panel data was largely unavailable for the region’s 10 Airport Economics in Latin America and the Caribbean airports, the first and second time series have very small sample sizes and consequently produce results that are largely skewed by outliers. For example, Argentina’s financial crisis precipitated the airport sector’s aver- age annual productivity change of −18.1 percent over the 2000 to 2003 period, which pulled down the index’s reported regional average of −1.2 percent over the same period. Overall, thorough data collection and extensive quantitative analysis in chapter 3 suggests that, when multiple factors are considered, LAC air- ports are not radically better or worse performers than those of Asia, Europe, or North America. Within the LAC region, results were not sig- nificant enough to declare a definitive relationship between ownership (public or private) and technical efficiency. Technical efficiency appears to be driven largely by factors outside of the control of airport manage- ment, though high levels of nonaeronautical revenues (that is, revenues accruing from commercial sources rather than airport tariffs) appear to have a positive relationship with technical efficiency. With this study being an initial attempt to perform benchmarking analysis on LAC region airports, inconclusive results are to be expected. Data limitations hampered the scope of the analysis of chapter 3 and influenced decisions on the types of models used and analysis per- formed, which, in some cases, led to less forceful results (these limita- tions are diligently described within chapter 3). More frequent data collection, combined with a common methodology, will considerably improve the usefulness of the LAC experience as a resource for the study of PSP and technical efficiency in the airport sector. A regional body of airport regulators or an air transport specialized institution, such as Airports Council International (ACI) or the International Civil Aviation Organization (ICAO), would be best poised to design this methodology. Given the wide variety of private participation schemes used by Latin American countries, further research should consider individual airports on a case-by-case basis. In addition, future research should also assess financial efficiency as well as the impact of PSP on the quality of services delivered. Institutional Design and the Governance of Airport Regulators Changes to the structure of economic regulation of LAC region airports accompanied the increased role of private investment in airport infra- structure. Chapter 4 addresses the realities and challenges of airport regulators from a public sector governance perspective and analyzes insti- tutional design, comparing both independent regulatory agencies (IRAs) Overview 11 and government agencies (non-IRAs). It focuses on only those aspects of governance that are directly related to economic regulation. The ultimate objective of this governance analysis is to identify under which arrange- ment regulatory governance can be enhanced. In Latin America, the introduction of PSP in the airport sector was often accompanied by the creation of IRAs to enforce concession con- tracts and quality of service. In cases where the bulk of airport services remained state owned, the role of regulator was placed in the hands of government departments, with limited independence from sector author- ities. Brazil represents an interesting case, in which an independent regu- lator was created but only regulates one state-owned enterprise. In chapter 4, qualitative comparative analysis is used to describe the design and practices of airport regulatory agencies. Survey responses from 13 LAC region airport regulators (4 independent and 9 government agencies) provided information on four main aspects of the governance of airport regulators: (a) the autonomy of the decision-making process, (b) the transparency of policies implemented by airport regulators, (c) their accountability to stakeholders, and (d) the quality of bureau- cracy (table 3). Regulatory agencies were assigned values between 0 and 1 for each of the four main aspects of governance according to predeter- mined criteria. Regardless of the existence of private sector provision of airport ser- vices, an institutional design associated with an IRA appears to provide a better channel for good regulatory governance than a government depart- ment. Both regional and international experiences show the importance of a government body that is highly specialized and has consumers as the focus of its policies. At the same time, a regulatory agency is not capable on its own to introduce institutional quality into an airport system where policies are ill designed. However, even in an adverse context, chapter 4 shows that regulatory agencies enable an adequate representation of stakeholders and act as a filter against discretional decisions. A clear advantage of making regulations in regulatory agencies rather than in government departments is related to measures aimed at enhanc- ing the transparency of regulation. The division of transparency into different dimensions within the report allowed for the identification of several advantages in IRAs versus government departments. Consultations are the most notable of these advantages. The consumer orientation of regulatory agencies versus government departments, whether in the con- text of state-owned companies or private providers, is a powerful factor in bringing stakeholders’ opinions into the decision-making process. 12 Table 3 Criteria for Determining Regulatory Agency Governance Ratings Autonomy of decision making Transparency Accountability Quality of bureaucracy • Regulatory powers • Civic engagement • Appeals of agency’s decisions • Structure of staff positions (tariffs, quality of service, in rule making • Effects of consultations within the agency and so forth) • Consultations • Evaluation of agency’s • Educational levels • Status of agency • Publication of agency’s performance of agency’s staff Characteristics • Procedures to appoint decisions • Accountability instrument • Publication of vacancies or remove board members • E-government • Performance instrument • Budget sources • Registry of board meetings and decisions Source: Authors’ elaboration. Overview 13 Technical expertise is another aspect where IRAs show advantages. The measure of bureaucratic quality found higher bureaucratic quality levels in independent commissions than in government departments, on average. These results are reflected not only in the educational levels of the staff but also in the way vacancies are posted and filled. The most controversial aspect of the governance of IRAs is autonomy. The measure of autonomy found, on average, more guarantees of autonomy in IRAs than in non-IRAs. A worrisome outcome of the surveys’ analysis was the serious defi- ciency of economic regulation in the airport sector in the LAC region. On the one hand, very few of the agencies in charge of enforcing regulations have in place the necessary information systems (regulatory accounting manuals, economic and financial models) necessary to perform their tasks correctly. On the other hand, even when agencies claim to have the adequate information systems in place, the vast majority are not using them to estimate the weighted average cost of capital, which is an essen- tial variable for a regulator. In addition, the regulatory frameworks do not seem to provide appropriate incentives for regulators to properly carry out a frequent oversight of the quality of services provided by operators. Despite the overall advantage of the IRA as a model for good regulatory governance, conclusions should not be interpreted as a “one model fits all� approach. Rather, they should be used to identify those mechanisms that better guarantee open and sound decision making in the regulation of air- port services. The comparison between IRAs and non-IRAs as alternative institutional arrangements to regulate airports allowed the disaggregation of governance into different dimensions and the identification of advan- tages and disadvantages in both models. It is up to policy makers to priori- tize those aspects that better fit their institutional and policy frameworks. Financing Performance: Evolution and Benchmarking of Aeronautical Charges at Latin American Airports Given the size of the demand for air transport services and the significant minimum investments necessary to have adequate airport services, most airports in LAC can be considered natural monopolies. Accordingly, the economic theory indicates that tariffs should be carefully regulated. Aeronautical tariffs are, indeed, heavily regulated in Latin America and the Caribbean. However, survey responses illustrate the poor record of the LAC region’s airport regulators and ministerial departments when it comes to the use of regional tariff benchmarking tools, indicating that decisions about tariff levels and structure are often poorly informed. In 14 Airport Economics in Latin America and the Caribbean some cases, either airport regulators lack the technical capacity to per- form this kind of analysis, or structural inefficiencies prevent or deter qualified individuals from doing so. The tariff benchmarking analysis presented in this report constitutes an important first step in fostering dialogue on these issues and in set- ting the basis for a more robust tariff benchmarking exercise at the regional level, a task that should be led by sector regulators. Survey responses from 26 airports in 20 LAC countries provide the basis for the identification of changes in tariff structures and levels in three dif- ferent years: 1995, 2003, and 2009. The selection of years responds to the objective of identifying whether changes in tariff structures and levels were the direct outcome of the introduction of private sector participation in the management of airports. Since most airport conces- sions in the region took place before 2002, 2003 was selected to discern whether changes in tariff levels and structure corresponded with the introduction of PSP in the airport sector. The year 2009 was included to present the most recent tariffs available at the time this report was written, while 1995 was chosen because PSP had not yet come to occupy a prominent role in the LAC region. Within this overview, regulated tariffs are understood as the total turn- around costs faced by an aircraft, including landing fees (and night sur- charges for lighting), aircraft parking, use of boarding bridges, and passenger charges (passenger facility charges, security). The aircrafts selected for comparison, the Airbus A320 and the Boeing 767 are consis- tent with the type of fleets most commonly found in the LAC region in 2009. To provide an international reference to the benchmarking analysis, the following airports were included in the sample: New York (JFK), Los Angeles (LAX), Miami (MIA), Madrid (MAD), Paris (CDG), London (LHR), and Frankfurt (FRA). These European and North American air- ports concentrate most of the Latin America and Caribbean–based air- lines’ international flights outside of the LAC region. The following preview of results from chapter 5 shows how, in most cases, total turnaround costs for most LAC region airports have increased in recent years (see figure 4). Turnaround costs, as defined in this report, for an Airbus A320 increased by 34 percent in real terms at most LAC airports between 1995 and 2009. Very similar increases apply to a Boeing 767. For both types of aircraft used in this report, current total turn- around costs in LAC region airports are, on average, at a comparable or higher level than those in European and U.S. airports that are most fre- quently served by Latin American and Caribbean airlines. Figure 4 Evolution of Turnaround Costs for an Airbus A320, 1995–2009 NAS CCS UIO GIG GRU CLO BOG LIM EZE MEX MGA GUA SJO SCL CUN LPD MVD TGU SDQ ASU VVI MTY PTY SAL GDL KIN 0 1 2 3 4 5 6 US$, thousands (constant 2008) 1995 2003 2009 Source: World Bank elaboration based on information from IATA (1995, 2003, and 2009), Aeronautical Information Publication (AIP) Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infraestructura Aeronaútica (DINACIA—National Authority of Civil Aviation and Aeronautical Infrastructure), Uruguay. Note: Calculated turnaround costs assume a load factor of 71 percent; a daylight operation includes landing, parking (initial 2 hours), boarding bridge, passenger facility charge, and security. Figure assumes a 71 percent load factor. For a list of airport codes and the airports they represent, see page xxiii. 16 Airport Economics in Latin America and the Caribbean The increase in turnaround costs in real terms between 1995 and 2009 for the Airbus A320 and Boeing 767 has been accompanied by changes in the tariff structure. Fees paid by airlines decreased between 1995 and 2009, while those levied on passengers increased. In fact, charges applied to passengers, which currently account for over 85 percent of total aero- nautical charges, increased in real terms by 44 percent between 1995 and 2009. The current tariff structure in LAC airports is similar to that pre- vailing in the sample of European and U.S. airports, with a slightly higher percentage of the share devoted to passenger charges as opposed to air- line charges in the LAC region (figure 5). The tariff benchmarking analysis carried out in this report does not allow for definitive conclusions on the relationship between changes in aeronautical charges and the introduction of private sector participa- tion. The increase in aeronautical charges observed between 1995 and 2009 was shared by both publicly and privately operated airports. Further research through a case-specific approach should be conducted (a) to assess whether the introduction of private sector participation has led to an increase in aeronautical charges and (b) to link changes in aeronautical charges to the changes in the level and quality of air- port services. The study of airport tariffs is followed by a bibliography of sources used in the creation of this report, as well as appendixes that include the surveys submitted to airport operators to measure performance and to airport regulators to gather information on their governance. Conclusions The air transport sector in the LAC region faces the same basic problem as the other transport subsectors (roads, ports, rail, and urban transport): the lack of objective data to construct a reasonable baseline to assess its economic performance. Using that well-known initial diagnostic, this report presents a comprehensive assessment of the evolution of airport performance, investments, tariffs, and governance institutions. The assess- ment is the result of extensive research to compile the very limited pub- lic information available, complemented with questionnaires developed exclusively for this report. In summary, the main findings of the report are as follows: • In the LAC region, pioneering the introduction of PSP in the operation and expansion of airport infrastructure has led to total investments in Overview 17 Figure 5 Structure of Turnaround Costs for an Airbus A320 NAS CCS UIO GIG GRU CLO BOG LIM EZE MEX MGA GUA SJO SCL LAC sample average CUN LPD MVD TGU SDQ ASU VVI MTY PTY SAL GDL KIN CDG FRA LAX MAD JFK MIA 0 1 2 3 4 5 6 US$, thousands (constant 2009) paid by airlines paid by passengers Source: World Bank elaboration based on information from IATA 2009, AIP Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: Calculated turnaround costs assume a load factor of 71 percent; a daylight operation includes landing, parking (initial 2 hours), boarding bridge, passenger facility charge, and security. Figure assumes a 71 percent load factor. For a list of airport codes and the airports they represent, see page xxiii. 18 Airport Economics in Latin America and the Caribbean excess of US$10 billion since 1995. Increased investment has not been confined solely to large, privately operated airports. Demonstration effects may have led publicly operated airports to emulate the success- ful example of private counterparts through the pursuit of increased investment. • From 1995 to 2007, LAC region airports have become increasingly productive, though they remain on average consistently less efficient than those of Asia, Europe, and the United States. Even though the smaller size of LAC airports prohibits them from exploiting economies of scale, the alignment of management to international best practices improved their productive performance in global comparisons. • From 1995 to 2009, both publicly and privately operated airports saw an increase in aeronautical charges of more than 30 percent in real terms. The structure of aeronautical tariffs also changed toward higher tariffs for passengers and lower tariffs for airlines. Among possible explanations are a decision to set tariffs following a cost-recovery prin- ciple; less reliance on public sector subsidies; a need to cover higher costs associated with better quality of services; and the need to com- pensate private operators and more commercially oriented, corpora- tized public airport operators. • Airport economic regulation in the LAC region is weak. Independent agencies and government departments do not meet the international best practice criteria for transparency and accountability. Lack of tech- nical capacity, inadequate funding, and the incorrect or insufficient use of regulatory instruments are all likely causes. Several key questions regarding the quality of airport services remain unanswered. Did an increase in PSP affect the evolution and improve- ment of airport service quality? How much? Were improvements cost- effective? Who paid? Some anecdotal evidence indicates that quality improved mainly owing to the expansion of related air and land infra- structure. A proper impact evaluation of airport investments, including micro- and macroeconomic effects, is overdue but requires data on qual- ity that are currently unavailable. To improve the productive performance of LAC region airports, this report recommends, first and foremost, the enhancement of the capacity Overview 19 of airport regulators to measure the impact of public policies. Higher- quality regulation will call for consistent data collection and analysis, allowing for the generation of a robust and well-grounded benchmark of airport performance that highlights best performers. Better analysis will make it possible to determine whether policies (introduction of PSP, expansion of capacity, changes in the level of tariffs) achieve the desired objectives. A strong foundation of information will increase the quality of decision making, thereby reducing the unpredictability of regulatory decisions and consequently the cost of capital. Ultimately, stronger air- port regulation will further enhance the positive image of PSP in the LAC region’s airports and encourage sustained investment. National efforts to strengthen airport regulation will be most effective if supported by the knowledge and experience of established institutions, such as Airports Council International (ACI) and the International Civil Aviation Organization (ICAO). Each of the analytical chapters (chapters 3, 4, and 5) suggests addi- tional next steps to enrich future studies. In addition to the analysis per- formed in chapter 3, further research into technical efficiency should collect and explore information on the quality of service provided, as this is a major determinant of airports’ costs and a key input for strengthening programs aimed at increasing competitiveness and growth (through tour- ism, industry, and clusters of development or high-value-added air cargo trade). Chapter 4 emphasizes continued investigation into regulatory governance on a case-by-case basis. Chapter 5 recommends the system- atic incorporation of regional tariff benchmarking exercises into the regu- lar operations of regulatory agencies, in addition to further research into the due diligence performed, the actual process for setting aeronautical tariffs in Latin America, and the incentives they provide for infrastructure investments. The overall purpose of this report is to enhance the understanding of airport performance in the LAC region. It is expected that the findings of the report will motivate further analytical work to provide a menu of policy options aimed at increasing the contribution of the airport sector to economic growth. Note 1. See Andrés et al. (2008) for a survey of the recent literature and an applica- tion of partial performance indicators in the electricity, water distribution, and fixed telecommunications sectors. 20 Airport Economics in Latin America and the Caribbean References ACI (Airports Council International). 2009. “World Airport Traffic Report 2009.� ACI, Geneva, Switzerland. Andrés, L. A., J. L. Guasch, T. Haven, and V. Foster. 2008. The Impact of Private Sector Participation in Infrastructure: Lights, Shadows, and the Road Ahead. Washington, DC: World Bank. IATA (International Air Transport Association). 1995. Airport and Air Navigation Charges Manual. Montreal: IATA. ———. 2003. Airport and Air Navigation Charges Manual. Montreal: IATA. ———. 2009. Airport and Air Navigation Charges Manual. Montreal: IATA. Private Participation in Infrastructure (PPI) Database. World Bank, Washington, DC. http://ppi.worldbank.org/. CHAPTER 1 Recent Evolution of the Air Transport Sector The evolution of the air transport sector has been closely linked with the fluctuations of the global economy. Air transport demand, which is heav- ily dependent on business activity, trade flows, and tourism, has experi- enced long periods of continued growth alternated with brief crisis periods of negative growth (figure 1.1). Air traffic fluctuations are more intense than changes in the gross domestic product (GDP). In fact, air transport traffic, measured as pas- senger-kilometers (km), has a high income elasticity of demand of about 2.1 This amplifying effect has meant that, in times of crisis (such as those associated with the second oil crisis in 1979, the Gulf War in 1990, the terrorist attacks of September 11, 2001, or the global financial crisis of 2008), the impact on the sector has been much more negative than on other segments of the economy. This feature has especially affected the airlines because the progressive liberalization of the most important aviation markets (most notably the liberalization process initiated by the United States and the European Union in the late 1970s and 1990s, respectively) resulted in an overall increase in competition and in the This chapter includes developed countries, such as Japan and Australia, in the World Bank regional designation of East Asia and Pacific. 21 22 Airport Economics in Latin America and the Caribbean Figure 1.1 Growth Rates in the Air Transport Sector and Global GDP, 1980–2008 15 13 11 9 7 percent 5 3 1 –1 –3 –5 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 passengers GDP per capita Source: World Bank estimation based on Airports Council International (ACI), International Air Transport Associa- tion (IATA), and International Monetary Fund (IMF) data. narrowing of their operating margins, which increased their vulnerabil- ity in periods of crisis. In the airport sector, where many of its facilities are natural monopolies and consequently are regulated, these effects have not been so evident. Evidence of the impact of the economic slowdown that began in late 2008 confirms the strong relationship between the level of economic activity and air transport passenger demand. According to traffic statis- tics released by the International Air Transport Association (IATA), international passenger traffic fell by 3.5 percent in 2009 relative to 2008 (IATA 2009). The significant passenger traffic growth observed between 2007 and 2008 has been heterogeneous across regions. Figure 1.2 demonstrates that all regions experienced high rates of growth in passenger demand in 2007 but the rate of growth has since decreased sharply across regions. In 2008, the last year for which annual data across regions were available (at the time this report was written), the Middle East experienced the greatest increase in passenger traffic (5.8 percent), followed by Africa (4.9 per- cent), and Latin America and the Caribbean (2.1 percent). Europe and the East Asia and Pacific region both grew by 1.2 percent. North America, on the other hand, was the only region with a negative growth rate, at −3.1 percent. Recent Evolution of the Air Transport Sector 23 Figure 1.2 Passenger Traffic Growth, by Region, 2007 and 2008 14 12 10 8 % growth 6 4 2 0 –2 –4 a c pe st be nd ica ifi ric Ea ro rib a a er an ac Af e Eu Am -P Ca ic dl ia e er id rth As Am M No tin th La 2007 2008 Source: World Bank estimation based on ACI data. In absolute numbers, the airport sector handled 4.874 billion arriving and departing passengers in 2008, as compared to 4.869 billion in 2007 and 4.5 billion in 2006, of which approximately 2.0 billion were interna- tional and 2.8 billion were domestic. As shown in figure 1.3, North America (the United States and Canada) by itself represented 48 percent of domestic traffic, with 1.3 billion domestic passengers, and Europe rep- resented more than half of global international traffic, with approximately 1.1 billion international passengers. The results for global air cargo traffic for 2008 show that traffic slowed down from the previous year by 3.7 percent, with domestic freight declining more severely than international freight, at −5.4 per- cent versus 2.4 percent. Such a deceleration could be attributed in part to increases in fuel prices, which diverted traffic to other transport alter- natives such as maritime, road, and rail. More recently, passenger and cargo traffic have been considerably affected by the global economic crisis that caused a major drop in international trade volumes; world- wide demand for air cargo capacity began to dwindle in December 2008. The latest data from IATA indicate that compared to 2008, air freight fell by 10.1 percent in 2009, representing the largest decline the 24 Airport Economics in Latin America and the Caribbean Figure 1.3 Domestic and International Passenger Share, 2008 a. Domestic passengers (% of global share by region) Africa 2% Asia-Pacific 26% North America 48% Europe 15% Latin America and Middle East the Caribbean 1% 8% b. International passengers (% of global share by region) North America Africa 10% 5% Middle East 5% Latin America and Asia-Pacific the Caribbean 21% 6% Europe 53% Source: World Bank estimation based on ACI data. Recent Evolution of the Air Transport Sector 25 industry has seen in the postwar period. This fall has been driven pri- marily by reductions experienced in Africa, Europe, and North America. These regions experienced year-on-year output declines of significant proportions: 11.2 percent, 16.1 percent, and 10.6 percent, respectively (IATA 2009). Disaggregating total cargo by region, figure 1.4 shows that North America and the East Asia and Pacific region contributed the greatest share (33 percent and 34 percent, respectively) to the industry’s 86 mil- lion cargo tons handled in 2008, followed by Europe (20 percent), the Middle East (5 percent), Latin America and the Caribbean (5 percent), and Africa (3 percent). The total aircraft movements handled by airports in 2008 was 77 million, a decrease of 2.1 percent compared to 2007. This figure includes cargo, military, general aviation, and passenger aircraft move- ments and translates into 87.3 passengers per movement. Ranking of airports by number of aircraft movements shows that 9 out of the top 10 airports are located in the United States, with the exception of Figure 1.4 Volume of Cargo Moved, by Region, 2008 percent Africa 3% Asia-Pacific 34% North America 33% Middle East 5% Europe Latin America and 20% the Caribbean 5% Source: World Bank estimation based on ACI data. 26 Airport Economics in Latin America and the Caribbean Figure 1.5 Aircraft Movements, by Region, 2008 percent Africa 4% Asia-Pacific 14% North America 33% Europe 28% Middle East Latin America 1% and the Caribbean 8% Source: World Bank estimation based on ACI data. Charles de Gaulle airport in Paris, France. On the other hand, as indi- cated by figure 1.5, the regions with the lowest share of aircraft move- ments are Latin America and the Caribbean, along with the Middle East and Africa. Together, they comprise only 13 percent of global aircraft movements. Latin America and the Caribbean Overview The objective of this report is to gain a better understanding of the airport sector in Latin America and the Caribbean (LAC) through an analysis of the evolution of airport and air industry performance. Consequently, it is important to present a general framework of recent regional trends, thus expanding on the previous global analysis. The LAC region has experienced great fluctuations in GDP growth, with particularly sharp declines from 1997 to 1999 and 2000 to 2001 (figure 1.6). Periods of high growth rates, on the other hand, took place between 1996 and 1997 and between 2004 and 2008. Specifically, in 2007, GDP grew at 5.6 percent, as commodity exporters benefited from Recent Evolution of the Air Transport Sector 27 Figure 1.6 GDP Growth and Passenger Growth in LAC, 1995–2008 12 10 8 % annual change 6 4 2 0 –2 –4 –6 95 97 99 01 03 05 07 08 19 19 19 20 20 20 20 20 GDP passenger growth rate Source: GDP data obtained from World Bank Open Data; available at http://data.worldbank.org. Passenger data obtained from ACI. Note: GDP in constant U.S. dollars. GDP growth rates calculated as the weighted average of the following coun- tries: Antigua and Barbuda, Argentina, The Bahamas, Barbados, Belize, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominica, Dominican Republic, Ecuador, El Salvador, Grenada, Guatemala, Guyana, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Suriname, Trinidad and Tobago, Uruguay, and República Bolivariana de Venezuela. record prices and rapid growth in global demand. In 2008, however, growth eased by 1 percent, due mainly to spillovers from the slowdown in worldwide activity and to decreased demand for commodity exports arising from the global economic crisis. Overall, the region’s airport sector, measured by changes in passenger traffic, has followed the economic cycle, and given the high elasticity of demand with respect to GDP, changes in passenger growth rates fluctu- ated more than GDP growth rates. The LAC region accounts for a small share of the air transport sector worldwide. Even though its total GDP is approximately 30 percent of the U.S. GDP, the size of the air transport sector is one-fifth that of the U. S. sector. Clearly, it has significant room for growth, which will depend primarily on economic growth, but also on a wide combination of vari- ables, including availability and quality of infrastructure (airports, access to airports), an efficient air traffic control system, adequate investment climate, and tourism development, among others. 28 Airport Economics in Latin America and the Caribbean In 2008, the LAC region handled approximately 338 million passen- gers, 4.6 million metric tons of cargo, and 6.4 million aircraft movements. Globally, this translates into 7 percent of passenger traffic, 5 percent of cargo traffic, and 8 percent of aircraft movements (table 1.1). Table 1.2 provides a ranking of airports in the LAC region in a global context, as measured by passenger numbers and organized according to the 10 largest airports worldwide, followed by all LAC airports included in the sample used for this report. Mexico City Airport, ranked 43rd glob- ally, is the most important airport in the region in terms of passenger traffic, handling a total of about 26.2 million passengers in 2008, approx- imately one-third the number handled in Atlanta, which ranked first worldwide. Furthermore, it should be noted that LAC countries have only four airports among the top 100 airports worldwide and 14 airports in the top 200.2 As for cargo, the entire LAC region handled a total of 4.6 million met- ric tons in 2008, only 1 million metric tons more than the amount of cargo traffic handled by the Hong Kong International Airport, the global leader with 3.6 million cargo metric tons, and three times as much as Miami, the North American cargo hub that handled 1.5 million metric tons. Within the region, the top 10 cargo airports account for approxi- mately 59 percent of the region’s cargo volume (see table 1.3). Among those, Brazil boasts four airports (Guarulhos, Manaus, Viracopos, and Galeão); Mexico two (Mexico City and Guadalajara); Chile one (Santiago de Chile); Colombia one (Bogotá); Peru one (Lima); and Argentina one (Ezeiza, Buenos Aires). On the aircraft movements level, table 1.4 outlines the 10 top-perform- ing LAC airports, out of which six (Bogotá, São Paulo GRU, Brasilia, Rio de Janeiro, Cancún, and Santiago de Chile) experienced positive growth between 2007 and 2008, with the Brasilia airport taking the lead. Several stylized facts can be drawn from the available data: (a) consid- ering passengers as the unit of measurement, airports in LAC, on average, are smaller than those in North America, Europe, and the East Asia and Pacific region; (b) airports in LAC, on average, have fewer aircraft move- ments than airports in North America, Europe, and the East Asia and Pacific region; (c) the most significant difference in output size between the average airport in LAC and that of the other regions is cargo; and (d) airports in LAC tend to rely heavily on international passengers rela- tive to airports in North America and the East Asia and Pacific region. Also, it is important to note that there is great heterogeneity among LAC airports with respect to how they rank in terms of passengers, aircraft Table 1.1 Latin America and the Caribbean Snapshot of the Airport Sector, 2008 Share of Share of Share of Growth rate Growth Share of Growth rate global global global of total Share of rate of global of aircraft Domestic domestic International international Total total passengers Cargo global cargo Aircraft aircraft movements passengers passengers passengers passengers passengers passengers (2007–08) (metric tons) cargo (2007–08) movements movements (2007–08) 224,531,098 8% 113,850,200 6% 338,381,298 7% 8% 4,589,092 5% 4% 6,403,629 8% 0% Source: Airports Council International (ACI) 2009. 29 30 Table 1.2 Global and LAC Airports Ranking: Passengers, Cargo, and Aircraft Movements, 2008 Global % change Cargo % change Aircraft % change rank Airport Passengers 2007–08 (metric tons) 2007–08 movements 2007–08 World 1 Atlanta, United States (ATL) 90,039,280 0.7 655,277 −9.0 978,824 −1.6 2 Chicago, United States (ORD) 69,353,876 −9.0 1,332,123 −13.1 881,566 −4.9 3 London, United Kingdom (LHR) 67,056,379 −1.5 1,486,260 6.5 478,518 −0.6 4 Tokyo, Japan (HND) 66,754,829 −0.2 852,444 −0.1 339,614 2.4 5 Paris, France (CDG) 60,874,681 1.6 2,280,050 −0.8 559,806 1.3 6 Los Angeles, United States (LAX) 59,497,539 −4.7 1,629,525 −11.9 622,506 −8.6 7 Dallas, United States (DFW) 57,093,187 −4.5 660,036 −8.7 656,310 −2.0 8 Beijing, China (PEK) 55,937,289 4.4 1,365,768 14.5 431,670 8.0 9 Frankfurt, Germany (FRA) 53,467,450 −1.3 2,111,031 −2.7 485,783 −1.4 10 Denver, United States (DEN) 51,245,334 2.8 250,994 −6.1 619,503 0.9 Latin America and the Caribbean 43 Mexico City, Mexico (MEX) 26,210,217 1.3 382,417 −7.0 366,561 −3.1 62 São Paulo, Brazil (GRU) 20,990,662 7.3 470,404 −3.7 194,186 3.3 96 São Paulo, Brazil (CGH) 13,661,227 10.4 32,521 −6.8 186,356 −9.3 99 Bogotá, Colombia (BOG) 13,456,330 4.9 547,928 −1.8 248,642 7.0 105 Cancún, Mexico (CUN) 12,786,423 11.3 16,496 −6.2 121,397 6.4 120 Brasilia, Brazil (BSB) 10,892,330 −6.2 56,619 −18.2 141,477 11.5 122 Rio de Janeiro, Brazil (GIG) 10,695,992 −0.8 114,581 −1.2 130,595 8.9 144 Santiago de Chile, Chile (SCL) 9,017,718 7.4 298,457 −1.1 101,103 7.0 156 Lima, Peru (LIM) 8,285,688 10.4 239,112 6.1 98,734 6.3 165 Buenos Aires, Argentina (EZE) 8,012,794 7.0 205,506 0.3 71,037 0.7 172 Guadalajara, Mexico (GDL) 7,393,500 −5.0 113,340 −8.8 152,353 −7.2 182 Monterrey, Mexico (MTY) 6,749,240 −1.7 40,979 −1.0 110,150 −5.7 207 Buenos Aires, Argentina (AEP) 5,687,221 0.4 14,690 4.3 85,793 5.5 242 Panama City, Panama (PTY) 4,549,170 19.6 86,588 5.0 80,694 8.4 306 San José, Costa Rica (SJO) 3,238,602 6.8 78,850 −1.0 77,114 2.6 307 Guayaquil, Ecuador (GYE) 3,236,768 8.0 66,936 −8.9 74,205 4.1 334 Santo Domingo, Dominican Republic (SDQ) 2,719,899 −2.0 54,500 −5.4 41,454 6.0 (continued next page) 31 32 Table 1.2 (continued) Global % change Cargo % change Aircraft % change rank Airport Passengers 2007–08 (metric tons) 2007–08 movements 2007–08 339 Nassau, Bahamas, The (NAS) 2,665,000 0.8 NA −1.2 NA NA 344 Piarco, Trinidad and Tobago (POS) 2,566,200 7.0 31,535 −3.9 65,401 −1.4 356 Cali, Colombia (CLO) 2,418,644 −0.7 41,354 −1.2 55,502 0.7 360 Medellín, Colombia (MDE) 2,367,555 1.4 99,078 −20.4 46,470 1.1 365 Guatemala City, Guatemala (GUA) 2,109,086 5.7 58,834 −15.3 102,519 11.2 384 Manaus, Brazil (MAO) 1,957,050 −13.1 130,723 −23.2 44,925 1.4 430 San Salvador, El Salva- dor (SAL) 1,570,012 −1.7 28,162 −4.3 33,922 −4.7 483 Campinas, Brazil (VCP) 1,260,112 4.5 223,023 −2.8 32,399 10.9 493 Barranquilla, Colombia (BAQ) 1,207,084 4.3 33,023 6.1 37,168 7.7 708 El Calafate, Argentina (FTE) 494,722 14.1 120 7.3 6,355 20.9 Source: Author’s estimation based on ACI 2009 and the World Bank Benchmarking LAC Airports Database. Note: Global rank is determined by total number of passengers. Rankings for Nassau (NAS), Guatemala City (GUA), and Santo Domingo (SDQ) correspond to 2007 data. Accordingly, the percentage change corresponds to the change between 2006 and 2007. Recent Evolution of the Air Transport Sector 33 Table 1.3 LAC Airport Ranking (Top 10) by Cargo, 2008 Cargo Percentage change Rank LAC Airport (metric tons) (2007–08) 1 Bogotá, Colombia (BOG) 547,928 −1.8 2 São Paulo, Brazil (GRU) 470,404 −3.7 3 Mexico City, Mexico (MEX) 382,417 −7.0 4 Santiago de Chile, Chile (SCL) 298,457 −1.1 5 Lima, Peru (LIM) 239,112 6.1 6 Campinas, Brazil (VCP) 223,023 −2.8 7 Buenos Aires, Argentina (EZE) 205,506 0.3 8 Manaus, Brazil (MAO) 130,723 −23.2 9 Rio de Janeiro, Brazil (GIG) 114,581 −1.2 10 Guadalajara, Mexico (GDL) 113,340 −8.8 Source: Author’s estimation based on ACI 2009 and the World Bank Benchmarking LAC Airports Database. Table 1.4 LAC Airport Ranking (Top 10) by Aircraft Movements, 2008 Aircraft Percentage change Rank LAC Airport movements (2007–08) 1 Mexico City, Mexico (MEX) 366,561 −3.1 2 Bogotá, Colombia (BOG) 248,642 7.0 3 São Paulo, Brazil (GRU) 194,186 3.3 4 São Paulo, Brazil (CGH) 186,356 −9.3 5 Guadalajara, Mexico (GDL) 152,353 −7.2 6 Brasilia, Brazil (BSB) 141,477 11.5 7 Rio de Janeiro, Brazil (GIG) 130,595 8.9 8 Cancún, Mexico (CUN) 121,397 6.4 9 Monterrey, Mexico (MTY) 110,150 −5.7 10 Santiago de Chile, Chile (SCL) 101,103 7.0 Source: Author’s estimation based on ACI 2009 and World Bank Benchmarking LAC Airports Database. movements, and cargo. For example, airports such as Guarulhos International in São Paulo (GRU) and Mexico City’s Benito Juárez International Airport (MEX) exhibit a similar scale of rankings across the three outputs (passengers, aircraft movements, and cargo). However, other airports rank differently for different outputs. Cancún International Airport (CUN), for instance, ranks high in terms of passengers, average in terms of aircraft movements, and low in terms of cargo. Another example is Viracopos-Campinas International (VCP), which ranks low in terms of passengers and aircraft movements but is the sixth highest in terms of cargo, with about 223,000 metric tons in 2008. In summary, the LAC region accounts for a small share of the air trans- port sector worldwide. It accounts for only 7 percent of total passengers, 34 Airport Economics in Latin America and the Caribbean 5 percent of cargo, and 8 percent of aircraft movements. Airports are relatively small when ranked on a global scale.3 The LAC region has four airports among the top 100 airports worldwide and only 14 among the top 200. Notes 1. Doganis 2006. An income elasticity of demand of 2 implies that when income (GDP) grows by 1 percent, demand for air travel grows by 2 percent. 2. The airport that serves the city of Caracas in the República Bolivariana de Venezuela occupies position 148 and handled 8.9 million passengers in 2008. This airport was not included in the table because it was not possible to obtain a response to the questionnaire submitted to the operator. Similarly, the Luis E. Magalhaes Airport, serving the city of Salvador in Bahía, Brazil, occupies position 186, but it was not included in this report. 3. The average airport in LAC served almost 5.8 million passengers in 2005, whereas the average airports in North America, Europe, and the East Asia and Pacific regions served 21.2, 17.8, and 16.5 million passengers, respectively. References ACI (Airports Council International). 2009. “World Airport Traffic Report 2009.� ACI, Geneva, Switzerland. Doganis, R. 2006. The Airline Business. London: Routledge. IATA (International Air Transport Association). 2009. “Air Transport Market Analysis.� IATA, Montreal, Quebec. World Bank Benchmarking LAC Airports Database. World Bank Open Data (database). World Bank, Washington, DC. http://data .worldbank.org/. CHAPTER 2 Investment in the Airport Sector Several Latin American and Caribbean (LAC) countries embarked upon a structural reform process in the 1990s. This process included, as a major component, the deregulation and privatization of several infrastructure services. In this context, the airport sector experienced a transformation that resulted in the introduction of private sector participation (PSP) in most LAC countries. A wide variation of PSP schemes was adopted. While Argentina opted to concession its airport network to a single operator, Chile adopted a case-by-case strategy and Mexico concessioned its airports by groups. Peru used a mix of single and group concessions, while Colombia and Costa Rica opted for the single concession scheme. The most important economy in the region, Brazil, continues to operate the largest airports through a state-owned corporatized enterprise. However, in 2008 the federal government launched a consultation pro- cess to introduce private participation in the airport sector. Table 2.1 shows the countries that, as of 2008, have introduced PSP in the manage- ment of airports and details the type of contractual arrangement chosen to incorporate the private sector. This chapter includes developed countries, such as Japan and Australia, in the World Bank regional designation of East Asia and Pacific. 35 36 Table 2.1 Latin American and Caribbean Airports by Type of PSP Arrangement Financial Type of PSP Subtype of PSP Contract period Total investment Country Project name closure year arrangement arrangement (years) (US$ millions) Argentina Islas Malvinas 1996 Concession Rehabilitate, 30 1996: 8; 2007: 6 International operate, and Airport transfer Argentina Airport 1998 Concession Rehabilitate, lease 30 1998: 1,581; System or rent, and 2007: 698 transfer El Calafate Airport 2000 Concession Build, rehabilitate, 25 2000: 25 2007: 15 Terminal operate, and transfer Neuquen Airport 2001 Concession Build, rehabilitate, 20 42 operate, and transfer Bolivia Bolivia Airports 1996 Concession Rehabilitate, lease 25 100 Concession or rent, and transfer Bolivian Airports 2000 Divestiture Full n.a. 17 Fuel Terminals Chile Diego Aracena 1995 Concession Build, rehabilitate, 12 8 Airport operate, and transfer El Tepual Airport 1996 Concession Build, rehabilitate, 12 6 operate, and transfer El Loa Airport 1997 Concession Build, rehabilitate, 12 4 operate, and transfer La Florida Airport 1997 Concession Build, rehabilitate, 15 4 operate, and transfer Santiago 1997 Concession Build, rehabilitate, 15 1997: 220; 2004: 22 International operate, and Airport transfer Carriel Sur Airport 1999 Concession Build, rehabilitate, 16 32 operate, and transfer Cerro Moreno 1999 Concession Rehabilitate, 10 10 Airport operate, and transfer Carlos Ibanez Del 2000 Concession Build, rehabilitate, 9 10 Campo Airport operate, and transfer Colombia El Dorado 1995 Greenfield project Build, operate, 20 145 International and transfer Airport Runway El Dorado 2006 Concession Build, rehabilitate, 20 650 International operate, and Airport transfer Rafael Nunez 1996 Management and Lease contract 15 22 International lease contract Airport 37 (continued next page) 38 Table 2.1 (continued) Financial Type of PSP Subtype of PSP Contract period Total investment Country Project name closure year arrangement arrangement (years) (US$ millions) Ernesto Cortissoz 1997 Management and Lease contract 15 9 International lease contract Airport Cali Alfonso Bonilla 2000 Concession Build, rehabilitate, 20 178 Airport operate, and transfer San Andres and 2007 Concession Rehabilitate, 20 20 Providencia operate, and Airports transfer Costa Rica San Jose 2000 Concession Build, rehabilitate, 20 161 International operate, and Airport transfer Dominican Republic Dominican 2000 Concession Build, rehabilitate, 20 265 Republic Airport operate, and Network transfer La Romana 2000 Greenfield project Merchant n.a. 55 International Airport Licey al Medio 2000 Greenfield project Merchant n.a. 30 Airport Ecuador Mariscal Sucre 2002 Management and Management n.a. 0 Airport lease contract contract New Quito Airport 2005 Greenfield project Build, operate, and 35 585 transfer Guayaquil 2004 Concession Build, rehabilitate, 15 80 International operate, and Airport transfer Honduras Honduras Airport 2000 Concession Build, rehabilitate, 20 120 Network operate, and transfer Jamaica Sangster 2003 Concession Build, rehabilitate, 30 175 International operate, and Airport transfer Mexico Southeast 1998 Concession Build, rehabilitate, 50 1998: 120; 2000: 394; Airports Group operate, and 2001: 28; 2002: 19; transfer 2003: 7; 2004: 32; 2005: 61 Pacific Airports 1999 Concession Build, rehabilitate, 50 1999: 264; 2000: 57; Group operate, and 2001: 26; 2002: 52; transfer 2003: 29; 2004: 64; 2005: 73; 2006: 1,000 Northern Central 2000 Concession Build, rehabilitate, 50 2000: 230; 2005: 203; Airports Group operate, and 2006: 376 transfer Puebla Airport 2000 Concession Rehabilitate, n.a  80 operate, and transfer Toluca Airport 2006 Concession Build, rehabilitate, 50 100 operate, and transfer 39 (continued next page) Table 2.1 (continued) Financial Type of PSP Subtype of PSP Contract period Total investment 40 Country Project name closure year arrangement arrangement (years) (US$ millions) Nuevo Laredo 2007 Greenfield project Build, operate, 20 7 Cargo Terminal and transfer Peru Jorge Chavez 1998 Greenfield project Build, operate, 30 8 Airport Cargo and transfer Terminal Jorge Chavez 2001 Concession Build, rehabilitate, 30 2001: 110; 2005: 92 Airport operate, and transfer Regional Airport 2006 Concession Rehabilitate, 25 220 Network Group I operate, and transfer Uruguay Laguna del Sauce 1993 Concession Build, rehabilitate, 26 31 Airport operate, and transfer Punta del Este 1996 Concession Build, operate, own 20 30 Airport Carrasco 2003 Concession Build, rehabilitate, 20 164 International operate, and Airport transfer Venezuela, RB Margarita General 1994 Concession Rehabilitate, 20 1994: 100; 2004: 34 Santiago Marino lease or rent, International and transfer Airport Source: Authors’ compilation based on the World Bank’s Private Participation in Infrastructure (PPI) Database and ProjectWare. Note: The projects listed for each country correspond to those listed in the PPI database. The column for total investment reports investment commitments. When new investment commitments are reported, the year (in italics) and amount are included. Otherwise, the amount reported corresponds to the financial closure year. n.a. = not available. Investment in the Airport Sector 41 It is important to highlight that the need to attract new investment financing sources to improve the quality of services has been the state- ment most commonly used by governments in the LAC region to justify introducing PSP in airport infrastructure. The LAC region, with its diver- sity in PSP schemes and more than 10 years of experience with the pri- vate management of airports, is able to provide valuable insights into the nature of investments in the sector. An analysis of the evolution of invest- ments in the airport sector in the LAC region, therefore, is useful in answering questions such as the following: Did the investment commit- ments that were announced when the contractual agreement was signed with the private airport operators eventually materialize? Were invest- ments allocated to address the most urgent infrastructure needs? Were there savings in construction costs brought about by the private conces- sionaires? Did airport regulators satisfactorily supervise the compliance of investment commitments made by airport operators? Questions along these lines should also be answered by state-owned airport operator com- panies to allow a comparison between the performance of public and private airport operators. Data requests on investment were a central part of the surveys distrib- uted to airport operators and regulators in LAC during the preparation phase of this report. Approximately half of the airports provided detailed responses regarding airport investment commitments, but only a few regulators reported on the compliance of investment commitments by airport operators. In addition to the incomplete nature of the investment data, comparability is difficult whenever investment information is gath- ered from different operators and countries. Investment reporting is not homogeneous because (a) tax laws allow for different depreciation meth- odologies, (b) regulatory accounting methods differ with respect to the types of investments that can be considered operation and maintenance or capital costs, and (c) investments in airports can be made in aeronauti- cal activities and nonaeronautical activities, with each definition being different among airports. Given the lack of a complete set of responses and the difficulties in producing homogeneous estimates, this report does not answer several of the questions raised in previous paragraphs. The only possible way to answer them is through an in-depth case-specific analysis of each airport and airport operator, a task that is pending for the LAC region. Given the aforementioned limitations on the data gathered through this study’s survey methodology, this report relies on specialized databases 42 Airport Economics in Latin America and the Caribbean to track the evolution of private investment in LAC airports, comparing it to private investment in airports in other regions as well as to that in other infrastructure sectors. Two data sources are considered: the Private Participation in Infrastructure (PPI) Database, a joint initiative of the World Bank and the Public-Private Infrastructure Advisory Facility (PPIAF), and ProjectWare, a database produced by a private firm, Dealogic. Both databases collect airport investment information, but whereas the PPI database tracks private investment commitments exclusively for developing countries as classified by the World Bank, ProjectWare tracks project financing for both developing and developed nations. It should be noted, however, that ProjectWare is not as complete as the PPI database, since some cases of private financing are not recorded.1 Overall, the PPI and ProjectWare databases present partial investment information. Their major limitation when analyzing investments in the airport sector is that none of them report public investment, and thus they underestimate total investments. For instance, neither database reg- isters airport investments in Brazil, the largest economy in LAC, where Infraero, the country’s state-owned airport operator, channels investments through operating resources or through transfers made by the federal government. A similar problem is found for the biggest air transport mar- ket in the world, the United States, where investments in airports are done through federal funds, by issuance of bonds with municipal or state guarantees, and by airlines. Private Project Financing in the Airport Sector Worldwide The ProjectWare database, which covers financing in the airport sector from 1996 to 2008, reported a total of 110 projects worldwide amounting to US$64 billion during this period. Figure 2.1 details the historical investment in airports worldwide: 2003 experienced the larg- est volume and largest annual increase for airport financing, measured by the number of projects across all regions. More recently, however, the number of airport projects receiving financing as reported by the database has decreased, from 14 in 2007 to 9 in 2008. With respect to project financing in value terms, the greatest financing amount took place in 1996, with 77 percent of the total amount attributed to the East Asia and Pacific region alone as a result of significant investment commitments of approximately US$10 billion for the Hong Kong SAR, China, airport. Investment in the Airport Sector 43 Figure 2.1 Project Financing in the Airport Sector by Number of Projects, Total Project Amount, and Region, 1996–2008 16 18 14 16 12 14 12 10 US$ billions # projects 10 8 8 6 6 4 4 2 2 0 0 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 20 20 20 20 20 20 20 20 20 East Asia and Pacific Europe and Central Asia Latin America and the Caribbean North America South Asia projects (right axis) Source: Author’s estimation based on ProjectWare data. When analyzing regional contributions to private airport project financing, it becomes evident that, historically, the two regions receiving the largest financing share are East Asia and Pacific, and Europe and Central Asia. Traditionally, both regions accounted for approximately 90 percent of total project financing in the airport sector. These have been followed by Latin America and the Caribbean, South Asia, and finally North America (see figure 2.2). If countries with project financing are divided into income-level cate- gories, it would be reasonable to assume that countries with higher levels of economic development would have higher levels of financing, as those are the countries that handle greater amounts of passenger and cargo traf- fic and therefore require greater investments to maintain or expand their airport capacity. Moreover, these countries tend to provide better condi- tions to attract large quantities of private financing. The evidence sup- ports this hypothesis, as high-income countries received most of the private financing, accounting for 81 percent of the total, while middle- income countries accounted for 19 percent for the period 1996–2008. 44 Airport Economics in Latin America and the Caribbean Figure 2.2 Share of Project Financing in the Airport Sector, by Region, 1996–2008 North America 2% South Asia Latin America and 4% the Caribbean 5% East Asia and Europe and Pacific Central Asia 50% 39% Source: Author’s estimation based on ProjectWare data. These results are reason for concern, as developing nations within the middle-income category have received fewer resources for investing in their airport sectors. Given that, in recent years, airports in developing countries have experienced very high rates of growth of passengers and cargo volumes, it is important to realize that they also require significant investments to upgrade their facilities and broaden their operations in response to such growth in demand. The economic crises that began in 2008 has reduced the pressure on the available infrastructure as demand fell, but if the relative growth rates return to the levels observed prior to the crisis, the remark made about investment needs in fast-growing devel- oping regions will hold true. Table 2.2 disaggregates project financing to the income level, region, and country between 1996 and 2008. Across countries, Australia has clearly been leading, with total project financing of US$19,326 million, followed by Hong Kong SAR, China, with US$11,050 million, and Turkey with US$6,188 million. Combined, these three countries repre- sented 57 percent of total financing for airport projects worldwide. Private Investments in the Airport Sector in Developing Countries Investment commitments to infrastructure projects across sectors (energy, telecommunications, transport, and water and sewerage) in developing countries with private participation have been increasing on Investment in the Airport Sector 45 Table 2.2 Total Project Financing in the Airport Sector by Income Level, Region, and Country, 1996–2008 Income level, region, and country US$ millions High Income: OECD East Asia and Pacific Australia 19,326 Japan 1,302 Korea, Rep. 127 New Zealand 115 Europe and Central Asia Belgium 1,544 Denmark 1,369 Germany 924 Greece 2,700 Hungary 2,660 Italy 4,220 Spain 155 United Kingdom 4,068 North America United States 1,275 High Income: Non-OECD East Asia and Pacific Hong Kong SAR, China 11,050 Europe and Central Asia Cyprus 783 Netherlands Antilles 55 Latin America and the Caribbean Bahamas, The 170 Upper Middle Income Europe and Central Asia Turkey 6,188 Latin America and the Caribbean Chile 463 Costa Rica 161 Jamaica 145 Mexico 509 Panama 70 Uruguay 31 Lower Middle Income East Asia and Pacific Philippines 629 Europe and Central Asia Albania 65 Armenia 30 (continued next page) 46 Airport Economics in Latin America and the Caribbean Table 2.2 (continued) Income level, region, and country US$ millions South Asia India 2,678 Latin America and the Caribbean Colombia 795 Dominican Republic 265 Peru 378 Source: Author’s estimation based on the ProjectWare database. Note: OECD = Organisation for Economic Co-operation and Development. average over the years (figure 2.3). However, there was a reduction in private investment between 1999 and 2004 and in 2008 due to the financial crises.2 More specifically for the transport sector, roads have been at the forefront of private investment in developing countries every year since 1990, except for 1999 when they were led by railways (figure 2.4). Regional contributions to investment commitments in airport projects were heterogeneous between 1991 and 2008. Overall, the LAC region accounts for 30 percent of total commitments (figure 2.5). If this time period is divided in two: from 1991 to 2000 and 2001 to 2008, the LAC region would account for 70 percent of commitments between 1991 and 2000 and only 12 percent between 2001 and 2008. This fact shows that LAC was a pioneer in introducing PSP in the airport sector compared to other regions and that the intensity of the process has recently decreased dramatically, either because most airports have already received the nec- essary private investment to upgrade airport infrastructure, the region lost its attractiveness, or governments decided not to open the sector for new or more private investment. An important dimension to consider when analyzing private sector participation in infrastructure is the extent of participation of the private sector. Generally this is summarized by the type of contractual agreement and type of project. The PPI database divides investment commitments into four subtypes of private participation in infrastructure: management and lease contracts, concessions, greenfield projects, and divestitures. Concessions, in turn, include three categories: (a) rehabilitate, operate, and transfer (ROT); (b) rehabilitate, lease or rent, and transfer (RLT); and (c) build, rehabilitate, operate, and transfer (BROT). Greenfield projects, on the other hand, include a variety of different types of categories, Figure 2.3 Private Investment Commitments to Infrastructure Projects in Developing Countries, by Sector, 1990–2008 180 160 140 US$ billions (2008) 120 100 80 60 40 20 0 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 energy telecommunications transport water and sewerage Source: Private Participation in Infrastructure (PPI) Database. 47 48 Figure 2.4 Investment Commitments to Transport Projects with Private Participation in Developing Countries, by Subsector, 1990–2008 40 140 35 120 30 US$ billions (2008) 100 new projects 25 80 20 60 15 40 10 5 20 0 0 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 airports railways roads seaports new projects (right axis) Source: PPI database. Investment in the Airport Sector 49 Figure 2.5 Total Investment Commitments to Airport Projects with Private Participation in Developing Countries, by Region, 1991–2008 Sub-Saharan Africa 2% East Asia and South Asia Pacific 15% 15% Middle East and North Africa 7% Europe and Central Asia 31% Latin America and the Caribbean 30% Source: PPI database. among which the most prevalent in the airport sector is build, operate, and transfer (BOT). As can be seen in figure 2.6, from 1993 to 2008, the predominant type of private participation in the airport sector in terms of investment has been BROTs, followed by management and lease con- tracts, BOTs, and ROTs. Private Investment in the Airport Sector in Latin America and the Caribbean Focusing on LAC, the two databases under consideration present signifi- cant differences. According to the PPI database, consistent with global trends, the airport sector in LAC represents a small fraction of projects and investment amounts in infrastructure in the region compared to roads (figures 2.7 and 2.8). Investments in the airport sector in LAC peaked in 2006 with a total of US$2,346 million but fell to US$746 million in 2007 and US$231 million in 2008. The observed fall in the investment commitments does not imply a significant reduction in the actual investments directed to the airport sector because, as figure 2.8 50 Figure 2.6 Investment Commitments to Airport Projects with Private Participation in Developing Countries, by Type of Project, 1991–2007 9,000 8,000 7,000 US$ millions (2007) 6,000 5,000 4,000 3,000 2,000 1,000 0 91 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 BOT BROT divestiture other greenfield projects RLT ROT management and lease contract Source: PPI database. Figure 2.7 Investment Commitments to Transport Projects with Private Participation in Latin America and the Caribbean, by Subsector, 1990–2008 20,000 60 50 15,000 US$ millions (2008) 40 new projects 10,000 30 20 5,000 10 0 0 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 20 20 20 20 20 airports railways roads seaports new projects (right axis) Source: PPI database. 51 52 Airport Economics in Latin America and the Caribbean Figure 2.8 Investment Commitments to Airport Projects with Private Participation in Latin America and the Caribbean Countries, by Type of Investment, 1993–2008 2,500 2,000 US$ billions 1,500 1,000 500 0 93 95 97 99 01 03 05 07 08 19 19 19 19 20 20 20 20 20 payments to the government investments in physical assets Source: PPI database. shows, almost US$1,500 million reported in 2006 were payments made to governments in the form of concession rights. As shown in the PPI database, from 1993 to 2007, private participa- tion in the airport sector in the region has been led by Mexico, followed by Argentina and Colombia (see table 2 in the overview). Together, these countries account for 71 percent of total investments in the region. In Mexico, all projects have been concessions, except for the most recent greenfield project of the Nuevo Laredo Cargo Terminal. In Argentina, on the other hand, most of the investments have been allocated to the reha- bilitation of 33 airports under the AA2000 concession, while Colombia shows a mixture of greenfield projects, management and lease contracts, and concessions. Conclusions Among all developing regions, LAC pioneered the introduction of PSP in the airport sector. Until 2001, according to the PPI database, LAC accounted for 70 percent of private sector investment in airports among developing regions. This region has also implemented the widest variety of types of public-private partnerships in the airport sector. In order to gain a deeper understanding of the role and impact of such private participa- tion in the sector, one must be able to effectively track investments in a standardized and reliable manner. Despite the availability of comprehen- sive databases that aim at tracking private investments in infrastructure Investment in the Airport Sector 53 sectors, the information is insufficient. The major problem is that these databases do not track investments in airports operated by public compa- nies. Moreover, there are significant differences in the investment infor- mation reported by the databases. The ProjectWare database shows very different total amounts and sequencing (the date the investment is regis- tered) of investment projects: signed airport projects for LAC amounted to US$2,986 million from 1996 to 2007, which represents 27 percent of the investment amount reported by the PPI database. There is a clear need for further analysis, which should be made on a case-by-case basis and implemented with a common methodology to measure investments to be able to aggregate and make cross-country comparisons. Such analysis would be important if the endeavor to com- plete the information vacuum described is carried out by a regional body of airport regulators or by institutions specializing in air transport, such as the Airports Council International (ACI) or the International Civil Aviation Organization (ICAO). Notes 1. The PPI data set represents the more comprehensive and detailed attempt to quantify investments in LAC countries. For instance ProjectWare does not include private financing in the Argentine airport sector, which accounted for 25 percent of total private sector investment in the LAC airport sector from 1993 to 2007, according to the PPI database. These databases differ in the types of data collected. PPI records total investment commitments entered into by the project entity at the beginning of the project (at contract signature or financial closure). ProjectWare, in contrast, presents total project amounts and their breakdown by financing sources, including shares in loans, bonds, and equity, from which investment then is identified. In the ProjectWare database, project amounts reflect investments in infrastructure in the form of construc- tion, expansion, and refurbishment of physical assets as well as in the financing of acquisitions and the refinancing of existing debt. Any given project can consist of one or a combination of any of the above. ProjectWare presents details on projects in five categories: preapproval, in tender, in finance, signed, and cancelled projects, with the purpose of making the PPI and the ProjectWare comparable; therefore, in this report, investment data included in the category “signed� in ProjectWare were considered as investment. 2. This section relies exclusively on information from the PPI database. Information contained in this database is based on contractual arrangements with and without investments in which private parties have assumed operating risks. Projects included in the database are not limited to those that are entirely 54 Airport Economics in Latin America and the Caribbean privately owned, financed, or operated, but rather some of them have some degree of public participation. To be included in the database, projects are required to have at least 25 percent participation from the private sector except for divestitures, which are required to have at least 5 percent of equity owned by private parties. Investment amounts reflect investments that were committed to at the time of contract signature or financial closure, not the investments that have actually been executed. They represent the total of private and public investments for a particular project. The database does not provide information about the difference between investment commitments and actual investment. Accordingly, the PPI database provides an upper bound of total private investment in infrastructure (in other words, actual investment is at most equal to the one reported by the database). References Private Participation in Infrastructure (PPI) Database. Public-Private Infrastructure Advisory Facility (PPIAF) and World Bank, Washington, DC. http://ppi .worldbank.org/. ProjectWare (database). Dealogic. http://www.dealogic.com. CHAPTER 3 Efficiency Estimation Commercial firms constantly search for ways to improve their opera- tional performance. In identifying potential areas for improvement, the first step a company should take is to assess its performance. There are two general approaches to achieve this. At the firm level, an analysis can be conducted based on the evolution of the firm’s performance through time. Alternatively, the same firm’s performance can be evaluated through a comparison with that of other firms in the same industry, namely, through a benchmarking exercise. This section of the report intends to appraise the performance of airports in the Latin America and Caribbean (LAC) region using this second approach. The use of benchmarking in the infrastructure sector is a relatively recent practice. Having gained significant ground with the creation of economic regulators, the first attempt to use benchmarking for regulatory purposes dates back to the late 1990s, when an initial methodology was developed by the Office of Water Services (OFWAT) in the United Kingdom. This chapter includes developed countries, such as Japan and Australia, in the World Bank regional designation of East Asia and Pacific. 55 56 Airport Economics in Latin America and the Caribbean In the transport sector, and airport subsector more specifically, bench- marking has made limited strides. This can be partly attributed to the fact that for a long period of time, commercial and business pressures were weak within the airport sector since airports around the world were not only owned and managed by governments, but also seen solely as public utilities and strategic assets for national defense purposes. During the late 1980s and early 1990s, however, there was a slow shift toward a view of airports as more commercially oriented enterprises. Consequently, this led several countries to introduce private sector participation in the operation of airports, which in turn called for a more thorough evaluation of airport performance, both to be able to initially negotiate the terms and conditions of private involvement and to track the improvements or lack thereof resulting from such involvement. As a result of this process, in the late 1990s, benchmarking began to be accepted as an important management tool within the airport industry. It was not until this time that the academic literature embarked upon the study of airport bench- marking through advanced efficiency estimation techniques, when the first papers were published in academic journals.1 Most of the academic literature focuses on developed countries, with the exception of several papers written by University of British Columbia researchers, which use data from several Asian countries, most of them in China. For Latin American countries, a very limited quantity of papers using advanced efficiency estimation techniques were written.2 As highlighted by the specialized literature, benchmarking, as a tool to assess performance, is a multifaceted task that must be viewed within the context of a firm’s complexity. A firm can be seen as a collection of dif- ferent processes needed to bring its products or services to the market. Therefore, even though it would be ideal to evaluate a firm based on a single measure of performance, this is not feasible, since a firm’s perfor- mance depends on the performance of each one of these processes and their interaction. Thus, a firm’s performance needs to be measured from several angles, with the three main ones being economic, financial, and operational (see table 3.1 for several examples of performance indica- tors). First, the economic angle takes into account factors such as the mix of inputs used, technology to transform inputs into outputs, and the firm’s productive scale. Second, the financial perspective addresses the mix of financial resources and profitability indicators. Finally, the opera- tional perspective refers to the quality of the products or services pro- vided and can be divided into two categories: the unobserved perception of quality, which looks at clients’ satisfaction, and the measured quality, Efficiency Estimation 57 Table 3.1 Partial Performance Indicators Commonly Used in the Airport Sector Key performance Perspective indicators Metric Financial Revenue diversification Aeronautical revenue as a percentage of total revenue Depreciation impact Depreciation costs as a percentage of total revenue ROCE Return on capital employed Operating profits Operating margin as a percentage of total revenue Economic: ATM staff productivity Aircraft movements per employee Productivity Pax staff productivity Passenger throughput per employee ATM capital productivity Aircraft movements per capital employed Pax capital productivity Passenger throughput per capital employed Economic: Unit ATM service cost Total revenue per air traffic Cost-effectiveness movement Unit pax service cost Total revenue per passenger Unit staff employment cost Total staff costs per passenger Unit operating cost Total operating costs per passenger Quality of service Stand availability Stand availability per landing Runway capacity availability Average throughput capacity vs. maximum capacity Passenger satisfaction Aggregated output of passenger satisfaction surveys Baggage system availability Aggregated serviceable hours of system vs. desired hours Source: Author’s compilation based on IATA. which is based on the firm’s measurement of the product or service qual- ity. The focus of this report is on economic performance and the measur- able side of operational performance.3 In some cases, infrastructure utilities have a single output, such as an electricity generator, which solely produces energy. In others, firms have several outputs, but the main production technologies and inputs are clearly different for each output. For example, a water utility provides drinkable water and sewage treatment, but the production technologies and inputs to provide each service are different. In the airport sector, on the other hand, there are three different outputs (passengers, aircraft movements, and cargo), but the production technologies and inputs are 58 Airport Economics in Latin America and the Caribbean shared among all of them. The multi-input characteristic of airports’ pro- duction function might explain why benchmarking performance has developed faster in the energy and water sectors. This chapter introduces the benchmarking of airports’ performance through partial performance indicators. The calculation of partial perfor- mance indicators is the simplest and most intuitive way to compare air- ports’ performance. Most of the airport regulatory agencies, airport operators, and industry organizations rely on the information obtained from partial performance indicators to adopt sector-specific policies, including airport tariffs, taxes, and investments. Due to data availability, the first section of this chapter presents several partial performance indi- cators for 2005. With the intention of conducting a more in-depth analy- sis, the second section of this chapter develops time series of some partial indicators for selected airports for the period 1995–2006. Finally, follow- ing the latest developments in the literature, the third section of the chapter conducts an analysis of efficiency using aggregate measures and econometric techniques to compute a globally efficient frontier for the airport sector and identify the position of the LAC countries’ airports relative to the best practice worldwide. As mentioned earlier, partial performance measures are widely used not only in the airport sector but also in other infrastructure sectors such as water and electricity and telecommunications.4 The main advantage of these measures is that they are simple to calculate and easy to under- stand, a feature shared by the traditional accounting ratios widely used for financial performance. At the same time, partial performance measures have severe limita- tions. As partial performance indicators ignore the interaction between inputs and outputs produced, they can provide a distorted picture of performance. For example, good performance translated into a high num- ber of passengers per check-in desk may reflect underperformance in another partial measure, such as waiting time per passenger spent in line to check in. In addition, partial measures do not reflect differences in fac- tor prices nor take into account possible substitution of inputs. For instance, if labor relative to capital is cheaper in city A than in city B, partial performance measures can signal that the city A airport is using too much labor even though the airports in cities A and B are both using an efficient mix of labor and capital given the input prices in their respec- tive markets. Another problem with partial indicators is that they do not account for differences in economic frameworks. For example, indicators such as Efficiency Estimation 59 aircraft movements per employee could be misleading if some countries have more rigid labor markets than others. This is most likely to be rele- vant when analyzing the time series of this type of indicator, since labor market rigidities in some countries might make it impossible for airports to cut their workforce during a recession. Furthermore, partial indicators can be problematic because they do not take into account airports’ differences with respect to scales and demands (infrastructure and personnel endowment are different for international and domestic passenger airports). Finally, partial measures do not take into account differences in operating environments between firms and are unable to handle multiple outputs, thus ignoring the multi- output characteristics of airports. As previously stated, there are many reasons behind why the use of benchmarking for airports is harder than for other industries. First, appro- priate outputs have to be defined, and also have to reflect the quality dimension of the services airports provide. Second, even if outputs are relatively homogeneous, data adjustments have to be made to take into account differences in the operational environment and the legal frame- work under which each airport operates. Finally, airports are faced with lump-sum investments, and different airport investment cycles may dis- tort efficiency comparisons if those investments are not properly taken into account. To overcome the shortcomings of partial performance indicators, aggregate measures and estimation techniques were developed by the academic literature in the last few years, and their use is becoming increasingly popular. The problem of sophisticated productivity measure- ment methods is their inherent complexity, which poses a problem for their use by airport regulators, especially in some developing countries where regulators lack the technical expertise and where the quality of data is inadequate. Aggregate measures include stochastic and nonsto- chastic methods, as well as parametric and nonparametric ones. A nonsto- chastic and nonparametric method is that of price index numbers such as the Tornqvist total factor productivity (TFP). This method requires the aggregation of all outputs into a weighted output index and all inputs into a weighted input index. It is a price index number because the prices of inputs and outputs are used as weights. Another nonstochastic technique is Data Envelopment Analysis (DEA), which compares a weighted out- put index relative to a weighted input index. The key difference between DEA and price index number TFP is that the weights in DEA are not predetermined but instead the result of linear programming. Hence, the 60 Airport Economics in Latin America and the Caribbean data requirements in DEA are less demanding than in price index num- bers. On the other hand, an example of a stochastic method is Stochastic Frontier Analysis (SFA), which is also a parametric method. It estimates a production frontier by allowing decomposition of the model residuals into a random component and an error component that represents the actual level of inefficiency. The third section of this chapter explains and uses DEA techniques to benchmark the performance of LAC airports.5 Partial Performance Indicators in LAC Airports: Cross-Comparison for 2005 The information used to compute partial performance indicators was obtained from the responses to the questionnaire developed for this report, a copy of which is included in appendix A. Table 3.2 lists all LAC airports that submitted a response to the questionnaire. To calculate the cross-comparison of partial performance indicators, this report used data for 2005. Two reasons can be cited for the selection of 2005 as the com- parison year: (a) it is the year for which the dataset is the most complete; and (b) 2005 is the year used in the latest available version of the Airport Benchmarking Report by the Air Transport Research Society (ATRS 2008) at the time this report was written. The ATRS report, which covers airports from North America (Canada and the United States), Europe, and the East Asia and Pacific region, along with the Airport Performance Indicators report by Jacobs Consultancy (2007),6 are the only two peri- odic reports that calculate partial performance indicators for airports. The ATRS report provides the regional mean for a series of partial perfor- mance indicators, thus providing the opportunity for comparison with the means for North America, Europe, and East Asia and Pacific. The two reports are the most widely known sources for benchmarking analysis in the airport sector. However, neither of them have data for Latin American airports. No source was found that systematically collects data and esti- mates partial performance indicators for Latin America and the Caribbean. Thus, the partial performance indicators presented in this chapter and the subsequent calculation of technical efficiency are the first attempts to conduct an overall assessment of airport efficiency for the main airports of the LAC region. The sample assembled for this chapter is representative of the air transport sector in the LAC region. It accounts for more than 80 percent of passengers and aircraft movements and for 70 percent of air cargo. As such, the database has a similar representativeness compared to the sample constructed by the ATRS and Jacobs Consultancy. Efficiency Estimation 61 Table 3.2 Latin American and Caribbean Airports Sampled Location Airport name IATA code Buenos Aires, Argentina Aeroparque Jorge Newbery AEP Buenos Aires, Argentina Aeropuerto Internacional Ministro Pistarini EZE El Calafate, Argentina Aeropuerto Internacional El Calafate FTE Nassau, The Bahamas Lynden Pindling International Airport NAS São Paulo, Brazil Aeroporto Congonhas CGH São Paulo, Brazil Aeroporto Internacional de Viracopos- Campinas VCP São Paulo, Brazil Aeroporto Internacional de Guarulhos Governador Andre Franco Montoro GRU Brasilia, Brazil Aeroporto Internacional Presidente Juscelino Kubitschek BSB Manaus, Brazil Aeroporto Internacional Eduardo Gomes MAO Rio de Janeiro, Brazil Aeroporto Internacional Antonio Carlos Jobim/Galeao GIG Santiago de Chile, Chile Aeropuerto Internacional Comodoro Arturo Merino Benítez SCL Bogotá, Colombia Aeropuerto Internacional El Dorado BOG Cali, Colombia Aeropuerto Alfonso Bonilla Aragón CLO Barranquilla, Colombia Aeropuerto Internacional Ernesto Cortissoz BAQ Medellín, Colombia Aeropuerto Internacional José María Córdova MDE San José, Costa Rica Aeropuerto Internacional Juan Santamaría SJO Guayaquil, Ecuador Aeropuerto Internacional José Joaquín de Olmedo GYE San Salvador, El Salvador Aeropuerto Internacional Comalapa SAL Guatemala City, Guatemala Aeropuerto Internacional La Aurora GUA Guadalajara, Mexico Aeropuerto Internacional Miguel Hidalgo y Costilla GDL Monterrey, Mexico Aeropuerto Internacional General Mariano Escobedo MTY Mexico City, Mexico Aeropuerto Internacional Benito Juárez MEX Cancún, Mexico Aeropuerto Internacional de Cancún CUN Panama City, Panama Aeropuerto Internacional de Tocumen PTY Lima, Peru Aeropuerto Internacional Jorge Chávez LIM Santo Domingo, Dominican Aeropuerto Internacional de Las Américas Republic SDQ Port of Spain, Trinidad Piarco International Airport and Tobago POS Source: Author’s compilation. The rest of this section presents several partial performance indica- tors. The indicators are those most commonly found in assessments of the air transport industry. The results of the calculations of all partial performance indicators are illustrated in the figures and described through examples from airports in the sample. An assessment of practical 62 Airport Economics in Latin America and the Caribbean limitations and caveats to consider when analyzing partial perfor- mance indicators is also included for most of the indicators presented. In all figures, those airports that were operated by the government or by a government-owned enterprise in 2005 are in light gray, while those that were operated by a private enterprise are in dark gray. Regional means appear in black. Each airport is identified by its code (three letters). Passengers per Aircraft Movement The average number of passengers per aircraft movement in 2005 for LAC was approximately 59, which is quite similar to the average for North America, but smaller than for Europe and the East Asia and Pacific region (figure 3.1). There are airports such as Buenos Aires (EZE), Cancún (CUN), or Santiago de Chile (SCL) where the average flight in 2005 carried about 100 passengers. At the other end of the scale are air- ports such as Nassau, The Bahamas (NAS), or Guatemala City (GUA), where the average flight in 2005 carried only 20 passengers. Ideally, this indicator should be calculated differentiating passengers and air traffic movements (ATMs) by destination (domestic and inter- national). Due to data limitations, it was not possible to disaggregate this indicator. However, it is safe to argue that differences in the number of passengers per aircraft movement cannot be fully explained by the percentage of international passengers served by the airports. For exam- ple, the airports in Buenos Aires (EZE), Cancún (CUN), Nassau (NAS), and San Salvador (SAL) are four of the five airports with the largest percentage of international passengers, but they show very different values for the ratio of passengers per aircraft movement. The mix of airplane sizes may explain such differences, and that mix in turn depends on the geographical location of the airport that determines the distance from the most popular origin and destination markets served. As a general rule, the closer the distance is between the airport and the markets it serves, the smaller the aircraft the airport will handle. However, this hypothesis cannot be tested, since several of the airports in the sample did not provide information about the size of the air- planes they handle. It can be concluded, therefore, that the indicator of passengers per aircraft movement is not necessarily an accurate indicator of an airport’s performance, as it is influenced by the airport’s geographical location; the type of passengers served; the structure of the network designed and served by airlines—for instance, SCL, GRU (São Paulo, Brazil), and EZE Efficiency Estimation 63 Figure 3.1 Passengers per Aircraft Movement, 2005 EZE CUN SCL GRU FTE GIG LIM SDQ MEX AEP CGH MTY PTY BSB CLO SAL BOG GDL SJO POS GYE MAO BAQ VCP GUA NAS mean LAC mean NA mean EU mean AP 0 20 40 60 80 100 120 140 # of passengers airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. 64 Airport Economics in Latin America and the Caribbean are terminal airports with a high percentage of long-haul flights served by large aircrafts; and runway and apron characteristics. Cargo per Aircraft Movement The airport in Santiago de Chile (SCL) not only ranks in third place in terms of passengers per aircraft movement in 2005, but also ranks high regarding cargo per aircraft movement, with almost 4 tons per aircraft movement (figure 3.2). Santiago de Chile’s airport is surpassed only by Viracopos-Campinas International (VCP) in São Paulo, with the average flight in 2005 carrying almost 6 tons of cargo and only 27 passengers. VCP is the only major airport dedicated to cargo in the LAC region— although the share of passengers is increasing due to the expansion of the catchment area of São Paulo and to the capacity constraints faced by the two other São Paulo airports, GRU and Congonhas (CGH)—with an important presence of industry consolidators (FEDEX, among others). An average flight in LAC in 2005 carried only 1.5 tons of cargo, which is more than what an average flight in Europe and North America carried for that same year (1.3 tons), but it is 3 tons less than the cargo in an average flight in East Asia and Pacific. Those airports included within the sample that are located in tourist areas within LAC, such as in Cancún (CUN) and El Calafate (FTE), handle the lowest cargo per aircraft. In a good example of the limitations of partial performance indicators, figure 3.2 seems to suggest that the LAC air cargo market is larger than that of North America or Europe. The ratio of cargo-to-aircraft move- ments appears smaller for North American and European markets, not because there is less cargo, but because there are more aircraft move- ments. In order to handle large volumes of cargo, North America, Europe, and the East Asia and Pacific region rely more on dedicated freight flights than on using space in regular passenger commercial flights to carry cargo. Whereas cargo per aircraft movement might be a fitting measurement for Latin America, tons of cargo per cargo-dedicated aircraft movement would provide a better idea of the size of North American and European air cargo markets relative to Latin America.7 Passengers per Employee Probably the most popular partial performance indicator in airports is the ratio of passengers per employee.8 If we assume that the number of pas- sengers is the only output of an airport and labor is its only input, then we could conclude that in 2005 the average airports in North America and East Asia and Pacific were more efficient than the average airport in Efficiency Estimation 65 Figure 3.2 Cargo per Aircraft Movement, 2005 VCP SCL MAO EZE GRU BOG LIM PTY SDQ SJO MEX BAQ CLO GYE GIG GDL MGGT SAL BSB POS MTY AEP CGH CUN FTE mean LAC mean NA mean EU mean AP 0 1 2 3 4 5 6 7 tons airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. 66 Airport Economics in Latin America and the Caribbean Latin America, which, with a little over 22,500 passengers per employee, was more efficient than the average airport in Europe (figure 3.3). This is certainly a surprising result. Given the relatively lower cost of labor in Latin America and the smaller scale of airports compared to Europe, one would have expected a lower ratio of passengers per employee in Latin America. As figure 3.3 shows, the airport in Santiago de Chile (SCL) and Congonhas International (CGH) in São Paulo are the most efficient in terms of number of passengers per employee. Also, the airports in Santiago de Chile (SCL) and Buenos Aires (EZE) served almost the same number of passengers in 2005, but the number of employees in EZE is over four times the number of employees in SCL. However, since this is a partial measure, which is strongly influenced by the degree of outsourc- ing in each airport, one should be careful when drawing conclusions regarding economic and operational efficiency.9 Aircraft Movements per Runway In 2005, the average airport in the LAC region had the lowest number of aircraft movements per runway when compared with North America, Europe, and the East Asia and Pacific region. If we consider aircraft movements as the only output and available runways as the only input, then the average airport in LAC is less efficient than the average air- ports in North America, Europe, and East Asia and Pacific. This finding indicates that in the average airport in LAC there was more excess capacity or less congestion than in airports in the other regions. The results of figure 3.4, as with all partial performance indicators, should be interpreted with caution. A simple look at figure 3.4 indicates that some airports in LAC are underutilized. Although this may be true for airports at the lower end of the scale, at least one caveat should be mentioned. The case may be that a given airport constructed a new runway in 2005 or in recent years to accommodate future demand growth. Since airport runways are a typical example of lump-sum investments, this partial performance indicator may not be providing accurate information. On the other hand, it is possible that despite available capacity, an airport is not allowed to accommodate additional aircraft movements due to restrictions imposed by air traffic control agencies that have low endowment of technological equipment and human capital resources. As a matter of fact, this is very common in the LAC region. The airport in Mexico City (MEX) and Congonhas International (CGH) in São Paulo have the most congested runways in Latin America, Efficiency Estimation 67 Figure 3.3 Passengers per Employee, 2005 SCL CGH CLO AEP CUN MTY GDL BSB SJO MEX LIM EZE GRU GYE FTE SDQ BAQ GIG SAL MAO PTY VCP mean LAC mean NA mean EU mean AP 0 10 20 30 40 50 60 70 80 90 # of passengers, thousands airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. 68 Airport Economics in Latin America and the Caribbean Figure 3.4 Aircraft Movements per Runway, 2005 MEX CGH BOG CUN GRU BSB AEP GDL LIM SJO SCL GYE POS GIG MTY CLO MAO SDQ EZE VCP PTY SAL FTE mean LAC mean NA mean EU mean AP 0 20 40 60 80 100 120 140 160 180 # of aircraft movements, thousands airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. Efficiency Estimation 69 with 166,500 and 141,314 aircraft movements per runway, respectively. According to the sample assembled for this report, the airport with the least congested runway is El Calafate (FTE) in Argentina, with almost 2,000 aircraft movements in 2005. Labor Costs as a Share of Operating Costs As expected, labor costs represent a smaller share of operating costs in LAC than in North America, Europe, or East Asia and Pacific (figure 3.5). According to the data reported by the airport operators and further calcu- lations of total labor costs produced for this report, labor costs represent, on average, only 16 percent of the costs incurred in the operation of an airport in Latin America in 2005. A striking finding is the large difference between the Colombian airports in Barranquilla (BAQ) and Cali (CLO), with labor costs representing almost 40 percent of operating costs in BAQ and only 5 percent in CLO. One possible explanation could be that the outsourcing of labor-intensive jobs is more prevalent in CLO than in BAQ. However, this is just speculation, since we do not have information on the degree of outsourcing in each airport. The airport comparison of this indi- cator can suffer from an important accounting bias. It is common that concessioned airports include as operating costs their annual payments to the government for the right to operate the airport. This would tend to increase operating costs of privately operated airports when compared to public airports that do not have this expense, thus reducing the ratio of labor costs over operating costs. Given that the two effects, outsourcing and the reporting of annual payments as operating costs, move in the same direction, we should expect privately operated airports to show a lower value for this indicator, providing a bias that cannot identify whether pri- vately managed airports have lower labor costs due to higher average labor productivity. Labor Cost per Passenger The smallest average labor cost per passenger in 2005 was found in LAC, where the average airport spent US$1 per passenger in labor- related costs. In Europe, the average airport spent almost US$6 per passenger in labor-related costs. As figure 3.6 shows, the two extremes in LAC were represented by the Tocumen International Airport (PTY) in Panama, which spent US$3.20 per passenger in labor-related costs, while Congonhas International (CGH) in São Paulo, Brazil, spent only US$0.14 per passenger. 70 Airport Economics in Latin America and the Caribbean Figure 3.5 Labor Costs as a Share of Operating Costs, 2005 BAQ GDL BSB VCP MTY GIG MAO MEX SJO CGH GRU SCL CUN EZE GYE AEP CLO mean LAC mean NA mean EU mean AP 0 5 10 15 20 25 30 35 40 45 50 percent airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. Efficiency Estimation 71 Figure 3.6 Labor Cost per Passenger, 2005 PTY MAO GIG BAQ SDQ FTE MEX GRU GYE GDL EZE MTY SJO CUN BSB AEP CLO SCL CGH mean LAC mean NA mean EU mean AP 0 1 2 3 4 5 6 7 US$ per passenger airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. 72 Airport Economics in Latin America and the Caribbean Obtaining definitive conclusions from this indicator is a difficult task as its construction suffers from the same problems as the indica- tor of labor costs as a share of operating costs. The labor practices adopted by each airport, reflected in how much labor is outsourced, can make the calculation of this indicator meaningless for perfor- mance benchmarking. Operating Costs per Passenger When analyzing operating costs per passenger for 2005, we observe that even though the average airport in LAC is the least costly to operate, with US$5.56 per passenger, it is quite similar to those in North America and East Asia and Pacific (figure 3.7). On the other end of the scale is Europe, with an average of US$14.23 of operating costs per passenger in 2005. The available data do not allow us to conduct a cross-regional comparison of the determinants of airports’ operating costs in each region. The fact that the average airport in LAC shows a similar value of this indicator to that in North America and East Asia and Pacific is somewhat worrisome, given that labor costs as a share of operating costs are much lower in LAC (see figure 3.5). This would indicate that airport operators might be including cost items as operating costs that operators in other regions do not include (possible candidates are payments to the government and higher rates of depreciation). More research consisting of in-depth case studies should be carried out to understand the differences in input costs across regions. The Brazilian airports are an interesting case regarding operating costs per passenger. The top two airports in this category are Eduardo Gomes International (MAO) in Manaus and Galeão International in Rio de Janeiro (GIG), while the lowest and third-lowest airports are Congonhas International in São Paulo (CGH) and Juscelino Kubitschek in Brasilia (BSB). MAO is one of the smallest airports in terms of number of pas- sengers in the sample, while CGH, BSB, and GIG are the second-, fourth-, and sixth-largest airports in that same category, respectively. The most striking difference is between BSB and GIG. The airport in Brasilia (BSB) had approximately US$19 million in operating costs and served 9.4 mil- lion passengers, while the airport in Rio de Janeiro (GIG) had approxi- mately US$91 million in operating costs and served 8.6 million passengers. With the exception of Guayaquil (GYE), in Ecuador, and Buenos Aires (EZE), in Argentina, privately operated airports seem to have lower operating costs per passenger than those operated by public companies. Efficiency Estimation 73 Figure 3.7 Operating Costs per Passenger, 2005 MAO GIG GYE EZE SAL GRU MEX CLO CUN SJO BAQ AEP MTY GDL SCL BSB LIM CGH mean LAC mean NA mean EU mean AP 0 2 4 6 8 10 12 14 16 US$ per passenger airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. 74 Airport Economics in Latin America and the Caribbean Operating Costs per Aircraft Movement The average airports in Europe and East Asia and Pacific spent signifi- cantly more U.S. dollars per aircraft movement than those in North America and LAC (figure 3.8). An interesting finding is that the aver- age airport in North America has a lower operating cost per aircraft movement than the average airport in LAC. The most likely explana- tion for this result is the difference in scale, as airports in North America have, on average, three times as many aircraft movements per year than LAC airports. Moreover, the network structure of the air transport market in North America, which is based on the hub-and- spoke system, generates a more intensive use of aircraft than the LAC system, leading to increased numbers of aircraft movements. The ratio of operating costs to aircraft movements for all the LAC airports, with the exception of Viracopos-Campinas (VCP), was less than US$1,000 in 2005. As in the case of operating costs per passenger, Congonhas (CGH) in São Paulo and Juscelino Kubitschek (BSB) in Brasilia are the lowest and third-lowest airports, in terms of U.S. dollars spent on operating costs per aircraft movement in 2005. Viracopos-Campinas (VCP), with US$1,174 of operating costs per aircraft movement, was the costliest airport to run per aircraft movement, while Galeão International (GIG) with US$881, ranked second in this category. Total Revenue per Passenger The average airport in LAC had almost US$10 of revenue per passenger in 2005, which is similar to the revenue per passenger in the average North American airport (figure 3.9). However, this number is small com- pared to the revenue per passenger generated in the average European airport, which was US$23 in 2005. Airports generate revenues from two distinct categories: (a) aeronauti- cal services and (b) nonaeronautical services. Aeronautical revenue refers to the income directly related to the aviation activities at an airport, including landing fees, passenger and terminal charges, and in some cases ground-handling charges. Traditionally, aeronautical revenues are the pri- mary source of income for airports. However, increasingly, more airports are actively seeking other sources of nonaeronautical revenues, including car parking, retail shops and concessions, and real estate leasing, among others. Airport regulators tend to look closely at the aeronautical services, as these are services considered, for most airports in the world, a natural monopoly. Accordingly, from a policy perspective, aeronautical revenue Efficiency Estimation 75 Figure 3.8 Operating Costs per Aircraft Movement, 2005 VCP GIG EZE GRU MAO CUN MEX SAL GYE CLO AEP MTY SCL SJO BAQ GDL BSB LIM CGH mean LAC mean NA mean EU mean AP 0 200 400 600 800 1,000 1,200 1,400 US$ per aircraft movement airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. 76 Airport Economics in Latin America and the Caribbean Figure 3.9 Total Revenue per Passenger, 2005 MAO SJO CUN SAL GRU GDL GYE POS GIG MEX MTY EZE BAQ SCL CLO LIM AEP BSB CGH mean LAC mean NA mean EU mean AUNZ mean AS 0 5 10 15 20 25 US$ per passenger airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. Efficiency Estimation 77 and its determinants (that is, tariff structures and levels) tend to be stud- ied in more detail by regulators and airlines. The worldwide trend in commercial airports has been an increase in nonaeronautical revenues. Most regulatory regimes generate incentives for airport operators to increase revenues from commercial activities. Aeronautical Revenue Share In the average airports in North America and East Asia and Pacific, aero- nautical revenue amounts to 50 percent of total revenue. In LAC, the average airport draws 56 percent of its revenue from aeronautical sources, which is 4 percent more than the average airport in Europe (figure 3.10). The Colombian airports in Cali (CLO) and Barranquilla (BAQ), as well as the airports in Lima (LIM) and Santo Domingo (SDQ), drew more than 80 percent of their total revenue from aeronautical sources. Santiago de Chile’s airport (SCL) and Viracopos-Campinas (VCP) are on the other end of the ranking for LAC, with only around 15 percent of their revenue in 2005 coming from aeronautical fees and charges. The low value for Santiago de Chile (SCL) is explained by the type of conces- sion, which exclusively covers the terminals, and consequently the opera- tor does not receive landing fees, a major aeronautical revenue factor. Chapter 5 provides a comparison of airport tariffs and their evolution. Aeronautical Revenue per Aircraft Movement The average airport in LAC trailed those in Europe, Asia, and Australia and New Zealand10 in terms of aeronautical revenue per aircraft move- ment in 2005 (figure 3.11). The average airport in LAC generated US$339 in aeronautical revenue per aircraft movement, while those in the other three regions generated US$972, US$1,168, and US$687, respectively. An unexpected finding is that the average airport in LAC earned US$41 more than the average airport in North America during 2005. The top four LAC airports earned more than US$500 per average flight in 2005, with the airport in Cancún (CUN) earning more than US$1,000, while the six lowest-earning airports received no more than US$200 from similar sources. The Mexican airports tend to earn more aeronautical revenues per aircraft movement than the Brazilian airports. Ideally, according to the International Civil Aviation Organization (ICAO) guidelines, aeronautical revenues should cover the cost of pro- viding aeronautical services. Aeronautical revenues include several tariffs, some charged to aircraft and others to passengers. The most common tariffs are landing, parking, gate use, and passenger. There are different 78 Airport Economics in Latin America and the Caribbean Figure 3.10 Aeronautical Revenue Share, 2005 CLO BAQ LIM GDL MEX MTY GYE CUN SAL CGH POS EZE BSB GIG GRU AEP SJO MAO SCL VCP mean LAC mean NA mean EU mean AP 0 10 20 30 40 50 60 70 80 90 percent airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. Efficiency Estimation 79 Figure 3.11 Aeronautical Revenue per Aircraft Movement, 2005 CUN MEX EZE GRU SAL MTY GIG GDL GYE LIM VCP CLO POS BAQ SJO MAO AEP SCL CGH BSB mean LAC mean NA mean EU mean AUNZ mean AP 0 200 400 600 800 1,000 1,200 1,400 US$ per aircraft movement airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU: European Union; AUNZ = Australia and New Zealand; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. 80 Airport Economics in Latin America and the Caribbean tariff regimes that are used to finance airports’ costs, with the most com- mon being single till, dual till, compensatory, and residual schemes. It is seldom the case that aeronautical tariffs are set to cover the cost of pro- viding aeronautical services. Instead, tariff structure and levels are set to accomplish other objectives, including providing enough revenue to cross-subsidize loss-making airports, subsidizing domestic passengers, and generating revenue for the government (taxes included in tariffs). Chapter 5 presents a detailed benchmarking of aeronautical tariffs for 26 airports in Latin America. This comparison provides more detailed information than figure 3.11 about the costs airlines and passengers incur when using a specific airport. The benchmarking shows that there is a significant heterogeneity in the tariff structure across airports in the LAC region, indicating that when setting tariffs, regulatory authorities have very different objectives (the observed dispersion in aeronautical tariffs among airports cannot be explained exclusively by different cost functions). Passengers per Boarding Bridge and Passengers per Square Meter of Terminal Area Two interesting partial measures of efficiency are passengers per boarding bridge (figure 3.12) and passengers per square meter of terminal area (figure 3.13). The number of boarding bridges and the area of the termi- nal are both proxies for capital inputs, while the number of passengers that fly through an airport is one of the outputs of an airport. Hence, these two measures can be read as output per capital input. At the same time, these measures give an idea of the quality of the service pro- vided by the airport, as a large number of passengers per boarding bridge or per square meter of terminal area tell us whether the service provided by the airport could be improved. In the first indicator, if we observe a large number of passengers per boarding bridge, we could infer that the airport relies heavily on remote aircraft parking and bus transportation, which generates discomfort, since passengers need to walk to and from planes unprotected from weather conditions.11 In the second indicator, a large number of passengers per square meter of terminal area could indi- cate that the terminal is too crowded, meaning passengers cannot move around at ease. On the other hand, if these ratios are very small, we could conclude that the physical facilities are underused or even infer that the capacity was inadequately planned or, in the other extreme, that the air- port is a “white elephant.� The problem when assessing these types of quality variables is that in the case of a low ratio of number of passengers per boarding bridge or per Efficiency Estimation 81 Figure 3.12 Passengers per Boarding Bridge, 2005 CGH AEP CUN LIM GYE GDL MEX BSB GRU MTY EZE SDQ BOG SCL CLO MAO GIG PTY FTE POS MDE SAL mean LAC mean NA mean EU mean AP 0 0.5 1.0 1.5 2.0 2.5 millions airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. 82 Airport Economics in Latin America and the Caribbean Figure 3.13 Passengers per Square Meter of Terminal Area, 2005 CGH MTY AEP CUN BSB GYE EZE GRU LIM MEX SCL FTE SAL CLO SDQ MAO MDE VCP GIG mean LAC mean NA mean EU mean AP 0 50 100 150 200 250 300 350 # of passengers airports privately operated airports publicly operated Source: Author’s calculation. Note: LAC = Latin America and the Caribbean; NA = North America; EU = European Union; AP = Asia-Pacific. For a list of airport codes and the airports they represent, see page xxiii. Efficiency Estimation 83 square meter of terminal area, for example, a growth of the indicator would reflect a more efficient use of the installed capacity (available capital). Once a certain threshold is reached, however, the high intensity of use of the installed capacity would translate into congestion, possible delays in flights and discomfort to passengers, sending a clear signal that more investment is needed. As is the case with runways and terminals, boarding bridges are lump sum investments that generate an optimal cycle of underutilization followed by operation at full capacity. The average airport in LAC has a small number of boarding bridges relative to the number of passengers (and more shuttle transportation) when compared with those in North America, Europe, and East Asia and Pacific. In North America and Europe, the average ratio of number of passengers to boarding bridges in 2005 was less than 300,000. In that same year, the average airport in East Asia and Pacific had 360,000 passengers per boarding bridge. In Latin America there were 560,000 passengers per boarding bridge in the average airport. The data seem to suggest that LAC airports need more investments in boarding bridges. Eleven airports in LAC had a ratio of more than 500,000 passengers per boarding bridge during 2005. The worst case is Congonhas International (CGH), where there were more than 2 million passengers per boarding bridge. In the case of passengers per square meter of terminal area, we see that the average airport in Latin America is quite similar to that in East Asia and Pacific, but this ratio is lower than the average airports in Europe and North America. Again, the outlier by far is Congonhas International (CGH), where there were 333 passengers per square meter of terminal area. On the other end of the ranking we find another Brazilian airport, Galeão International (GIG) in Rio de Janeiro, with 22 passengers per square meter. This indicator suggests that there are no significant capacity constraints in LAC airports, or in other words, that there have not been overinvestments in the airport sector. However, the limitation of this indicator is that it does not consider specific daily or seasonal capacity constraints. It is usually the case that airports in LAC have a high concen- tration of flights during a few hours, especially those that serve interna- tional passengers. Anecdotal evidence in LAC indicates that several airports suffer from significant concentrations of passengers at some point during the day, resulting in high usage intensity of available capacity, and accordingly, low level of service quality. The problem is that airports are an example of lump-sum investments that should not be planned based only on peak demand periods. Furthermore, in order to regulate congestion 84 Airport Economics in Latin America and the Caribbean and spread the use of facilities throughout the day, economic incentives to airlines and passengers through differentiated tariffs should be put in place. The partial performance indicators calculated in this section of the report for LAC airports do not allow us to construct a unique perfor- mance ranking of airports. However, the indicators showed weak evi- dence that a couple of airports, Congonhas (CGH) in Brazil and Santiago de Chile (SCL), are more efficient than the other airports in the region, as they show up more times in the top three performers (table 3.3). By no means was it possible, from the calculated indicators, to conclude whether in 2005, airports operated by the private sector were more effi- cient than those operated by a state-owned company. These results are not surprising due to the problems associated with the use of partial per- formance indicators in multi-input, multi-output services (explained in detail in the introduction of this chapter). The next section, by studying the evolution through time of partial performance indicators, tests whether it is possible to identify airports that significantly improved their performance and whether privately operated airports improved their performance more than public airports. Partial Performance Indicators: Time Series Studying the evolution of partial performance through time gives the opportunity to answer more questions than the calculation of partial performance indicators for one point in time. By tracking the evolution in time of partial performance indicators, short-run macro- (appreciation of exchange rate, economic recession) and microshocks (construction of a new runway, strike in the major airline in a given airport) do not distort comparisons. In addition, and more interestingly for this report, time series data could give valuable information to test the hypothesis that the introduction of private sector participation (PSP) in airports brought about improvements in performance. The questionnaires prepared for this report and distributed to the main airport operators in Latin America (see appendix A) covered the period from 1995 to 2007. The purpose of asking for data starting in 1995 was to analyze the evolution of airports’ performance in the last decade. In addition, the data would provide the opportunity to make a before and after analysis of the introduction of PSP, given that private sector participation in the operation of the major airports in the region began in the late 1990s. Table 3.3 Summary of Airport Partial Performance Indicators—Top and Bottom Performers, 2005 Partial performance indicator Top 3 performers Bottom 3 performers Passengers per aircraft movement • Buenos Aires, Argentina (EZE) • São Paulo, Brazil (VCP) • Cancún, Mexico (CUN) • Guatemala City, Guatemala (GUA) • Santiago, Chile (SCL) • Nassau, Bahamas, The (NAS) Cargo per aircraft movement (tons) • São Paulo, Brazil (VCP) • São Paulo, Brazil (CGH) • Santiago, Chile (SCL) • Cancún, Mexico (CUN) • Manaus, Brazil (MAO) • El Calafate, Argentina (FTE) Cargo per dedicated aircraft movement (tons) • Santiago, Chile (SCL) • Panama City, Panama (PTY) • Lima, Peru (LIM) • Santo Domingo, Dominican Republic (SDQ) • Buenos Aires, Argentina (EZE) • Monterrey, Mexico (MTY) Passengers per employee • Santiago, Chile (SCL) • Manaus, Brazil (MAO) • São Paulo, Brazil (CGH) • Panama City, Panama (PTY) • Cali, Colombia (CLO) • São Paulo, Brazil (VCP) Aircraft movements per runway • Mexico City, Mexico (MEX) • Panama City, Panama (PTY) • São Paulo, Brazil (CGH) • San Salvador, El Salvador (SAL) • Bogotá, Colombia (BOG) • El Calafate, Argentina (FTE) Labor costs as a share of operating costs • Guayaquil, Ecuador (GYE) • Barranquilla, Colombia (BAQ) • Buenos Aires, Argentina (AEP) • Guadalajara, Mexico (GDL) • Cali, Colombia (CLO) • Brasilia, Brazil (BSB) Labor cost per passenger (US$) • Cali, Colombia (CLO) • Panama City, Panama (PTY) • Santiago, Chile (SCL) • Manaus, Brazil (MAO) • São Paulo, Brazil (CGH) • Rio de Janeiro, Brazil (GIG) Operating cost per passenger (US$) • Brasilia, Brasil (BSB) • Manaus, Brazil (MAO) • Lima, Peru (LIM) • Rio de Janeiro, Brazil (GIG) • São Paulo, Brazil (CGH) • Guayaquil, Ecuador (GYE) (continued next page) 85 86 Table 3.3 (continued) Partial performance indicator Top 3 performers Bottom 3 performers Operating cost per aircraft movement (US$) • São Paulo, Brazil (VCP) • Brasilia, Brasil (BSB) • Rio de Janeiro, Brazil (GIG) • Lima, Peru (LIM) • Buenos Aires, Argentina (EZE) • São Paulo, Brazil (CGH) Total revenue per passenger (US$) • Manaus, Brazil (MAO) • Buenos Aires, Argentina (AEP) • San José, Costa Rica (SJO) • Brasilia, Brazil (BSB) • Cancún, Mexico (CUN) • São Paulo, Brazil (CGH) Aeronautical revenue share • Cali, Colombia (CLO) • Manaus, Brazil (MAO) • Barranquilla, Colombia (BAQ) • Santiago, Chile (SCL) • Lima, Peru (LIM) • São Paulo, Brazil (VCP) Aeronautical revenue per aircraft movement (US$) • Cancún, Mexico (CUN) • Santiago, Chile (SCL) • Mexico City, Mexico (MEX) • São Paulo, Brazil (CGH) • Buenos Aires, Argentina (EZE) • Brasilia, Brazil (BSB) Passengers per boarding bridge • São Paulo, Brazil (CGH) • Port of Spain, Trinidad and Tobago (POS) • Buenos Aires, Argentina (AEP) • Medellín, Colombia (MDE) • Cancún, Mexico (CUN) • San Salvador, El Salvador (SAL) Passengers per square meter of terminal area • São Paulo, Brazil (CGH) • Medellín, Colombia (MDE) • Monterrey, Mexico (MTY) • São Paulo, Brazil (VCP) • Buenos Aires, Argentina (AEP) • Rio de Janeiro, Brazil (GIG) Source: Author’s compilation. Note: Top performers for the indicators: labor cost per passenger, operating cost per passenger, and operating cost per aircraft movement are those airports for which the indicators show the lowest value. As noted in the text, the highest value of the indicators’ aeronautical revenue share and aeronautical revenue per aircraft movement should not be directly interpreted as synonymous with top performance. Efficiency Estimation 87 As was explained in the introduction of this chapter, airport perfor- mance in developing countries has seldom been the subject of in-depth research. This has not been the case for other infrastructure sectors in developing countries, in particular in Latin America. This report ben- efited from a recent research project by Andrés, Guasch, and Lopez (2008), who conducted a thorough evaluation of the impact of PSP on electricity distribution, fixed-line telecommunications, and water and sewerage in Latin America by comparing the evolution of selected indicators before and after the introduction of private sector participa- tion in the management of utilities.12 These authors identify three distinct periods: (a) the pretransition or preprivatization period, refer- ring to the three years before the transition period; (b) the transition period, starting two years before the privatization or concession was awarded and ending one year after award; and (c) the posttransition or postprivatization period, referring to the four years after the transi- tion. The results indicate that changes in management and ownership generated significant improvements in labor productivity, efficiency, and product/service quality in the three infrastructure sectors ana- lyzed. However, changes are not very remarkable in the posttransition period, suggesting that most of the efficiency gains took place during the transition period. Unfortunately, very few of the private airport operators that responded to the questionnaire provided data for the years before the change in ownership. The reason cited was that the data are not available and that they could not share the data used for the preparation of the ownership changeover bidding documents. This was an expected although not a desirable outcome of the preparation process of this report. It is usually the case that private operators have limited access to information about the firm before the change of control. Moreover, once private sector participation is introduced, the former state operator, and in some cases the ministerial department that supervised the state operator, are dis- mantled and human resources and institutional memory lost as a result. Thus, constructing reliable time series of key performance data is a daunt- ing and often impossible task. For those airports under private operation included in our dataset, the available data start in the year the concession was awarded. Hence, it is not possible to undertake a before and after comparison of performance indicators. Given these data constraints, the focus of this section is to identify patterns and major changes, if any, in the evolution of performance indicators in the posttransition or post- privatization period. 88 Airport Economics in Latin America and the Caribbean Before proceeding to show the results of the calculation of time series partial performance indicators, it is important to make a note about the use of currencies, exchange rates, and inflation. To ensure consistency and comparability between airports from several countries, measures of income and cost used need to be expressed in the same currency. All partial per- formance measures that are expressed in monetary terms are affected by the exchange rate, which can bias the comparison of these measures among airports located in different countries. For example, the analysis of partial performance measures elaborated for this report for the year 2005 show that the average European airport ranked at the bottom on any of the cost-related partial measures. This result could indicate that a differ- ence in real costs exists between Europe and other regions or it could be the result of a temporary appreciation of the euro during 2005. As figure 3.14 illustrates, that was not the case, since the euro actually suf- fered a temporary depreciation in 2005. However, if the analysis had been done for the year 2008, it is likely that the cost difference between the average European airport and the average airports in other regions would have been bigger, given the higher value of the exchange rate (U.S. dollar per euro) in 2008. Also, an analysis based on one point in time (in this case only one year, 2005) might be influenced by the fact that different countries might be at different stages of the economic cycle. For instance, if a country is just coming out of a recession that negatively affected passenger volumes, this will likely bias some partial performance indicators, such as passengers per square meter or passengers per boarding bridge. When analyzing the evolution of partial performance measures through time, it is necessary not only to express income and cost mea- sures in the same currency, but also to remove the effect of inflation. If costs are not expressed in real prices, a generalized increase in prices could cause an increase in some partial performance measures, such as operating costs per employee or labor costs per passenger, wrongly signal- ing a decrease in productivity. In addition, by expressing all income and cost measures in U.S. dollars of a given year, the effect of fluctuations in the exchange rate is removed. For these reasons, the income and cost measures used in the analysis were first expressed in local 2005 prices and then in U.S. dollars using the average exchange rate for 2005.13 For presentation purposes, only selected partial performance indica- tors and a representative sample of airports that responded to the ques- tionnaire were used in the time series analysis.14 The database assembled enables the classification of LAC airports into three groups based on the Figure 3.14 Evolution of the U.S. Dollar–Euro Exchange Rate, 1999–2009 1.6 1.5 1.4 1.3 dollars per euro 1.2 1.1 1.0 0.9 0.8 Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 89 Source: Author’s calculation. 90 Airport Economics in Latin America and the Caribbean number of passengers handled in 2005. The first group, with airports serving more than 15 million passengers, is the smallest and includes only three airports: Mexico City’s Benito Juárez International Airport (MEX), Congonhas International (CGH), and Guarulhos International (GRU), both located in São Paulo. The second group consists of 10 air- ports that served between 5 million and 10 million passengers in 2005, but performance indicators are shown for five airports: Cancún International Airport (CUN), El Dorado International Airport (BOG) in Bogotá, Comodoro Arturo Merino Benítez International Airport (SCL) in Santiago de Chile, Ministro Pistarini International Airport (EZE) in Buenos Aires, and Jorge Chávez International Airport (LIM) in Lima. The third group comprises 14 airports that served less than 5 million passengers in 2005, and results are shown for four airports: Juan Santamaria International Airport (SJO) in San José, José Joaquín de Olmedo International Airport (GYE) in Guayaquil, Tocumen International Airport (PTY) in Panama City, and Piarco International Airport (POS) in Port of Spain. In the following figures the airports operated by the government or a government-owned enterprise are in light gray, while those under conces- sion are in dark gray,. All graphs also include the average of the partial performance measure for the airports in LAC for the year 2005. Each airport is identified by its code (three letters). Passengers per Employee Figure 3.15 shows that, with the exception of a few cases, there is not a clear increase in the number of passengers per employee. The airports in Santiago de Chile (SCL), San José (SJO), Panama City (PTY), and São Paulo (CGH and GRU) had an increase in productive efficiency mea- sured by the indicator passengers per employee. Simple observation of these figures does not make it possible to draw any conclusion about a trend in the evolution of this indicator, as some privately operated air- ports improved while others did not. Moreover, a similar pattern is observed for publicly operated airports. The case of Ministro Pistarini Airport (EZE) in Buenos Aires is interesting and illustrates important caveats to the analysis of perfor- mance in the airport sector. EZE experienced a significant drop in passengers per employee between 2000 and 2002 (a drop on the order of 50,000 passengers per employee to 16,000). The change was mainly due to a 34 percent decrease in the number of passengers, which was caused by a significant economic crisis after Argentina devalued its Efficiency Estimation 91 Figure 3.15 Passengers per Employee a. Group 1 1999 2000 2001 2002 MEX 2003 2004 2005 2006 2007 1995 1996 1997 1998 1999 2000 CGH 2001 2002 2003 2004 2005 2006 2007 1995 1996 1997 1998 1999 2000 GRU 2001 2002 2003 2004 2005 2006 2007 mean LAC 2005 0 10 20 30 40 50 60 70 80 90 # of passengers, thousands airports privately operated airports publicly operated (continued next page) 92 Airport Economics in Latin America and the Caribbean Figure 3.15 (continued) b. Group 2 1999 2000 2001 2002 CUN 2003 2004 2005 2006 2007 1999 2000 2001 2002 SCL 2003 2004 2005 2006 2007 2000 2001 2002 2003 EZE 2004 2005 2006 2007 2001 2002 2003 LIM 2004 2005 2006 2007 mean LAC 2005 0 10 20 30 40 50 60 70 80 90 100 # of passengers, thousands airports privately operated airports publicly operated (continued next page) Efficiency Estimation 93 Figure 3.15 (continued) c. Group 3 2001 2002 2003 SJO 2004 2005 2006 2007 2005 GYE 2006 2007 2003 2004 PTY 2005 2006 2007 mean LAC 2005 0 5 10 15 20 25 # of passengers, thousands airports privately operated airports publicly operated Source: Author’s calculation. Note: For a list of airport codes and the airports they represent, see page xxiii. 94 Airport Economics in Latin America and the Caribbean currency and defaulted on its external debt. Because airports face a derived demand from the demand for air transport services, it is diffi- cult to adjust production inputs during recessions or negative demand shocks. Accordingly, when benchmarking airports, caution needs to be exercised, as airports could be facing very different economic environ- ments. It could be the case that some of the most commonly cited efficiency performance indicators can deteriorate for reasons out of the airport operator’s control. This is one of the main reasons why airport regulators have not fully incorporated benchmarking techniques to set efficiency gain variables in tariff regimes (for a comprehensive justifi- cation and discussion of the use of benchmarking techniques to set tariffs, see CAA 2000). Labor Costs per Passenger Figure 3.16 depicts labor costs per passenger for the three groups of airports. Most of the airports in the three groups have reduced the amount spent per passenger in labor-related expenses during the period analyzed. In 1996, Guarulhos International (GRU) spent US$2.98 per passenger in labor-related expenses, while in 2006 it spent only US$1.15 per passenger. In Mexico City’s airport (MEX), the ratio of labor costs per passenger, measured in 2005 U.S. dollars, shifted from US$1.33 in 2000 to US$0.93 in 2006. In Cancún’s airport (CUN) this ratio decreased from US$0.74 in 1999 to US$0.44 in 2004, but then increased to US$0.84 in 2006. In the case of Lima’s airport (LIM), data are available for three years (2001–03). During this period the amount spent per passenger in labor-related expenses increased almost 40 percent. Operating Costs per Passenger When analyzing a broader measure of cost efficiency, such as operating costs per passenger, only a few airports present a clear trend during the period analyzed (figure 3.17). Only Guarulhos International (GRU) and the airport in Guayaquil (GYE) experienced a steady decrease in the ratio of operating costs to number of passengers. The airport in San José (SJO) also experienced a decrease in this ratio for all but one of the years analyzed, while for Congonhas International (CGH), the ratio of operat- ing costs to number of passengers decreased for most of the years until 2005, but started increasing in 2006. At the Ministro Pistarini (EZE) airport in Argentina, the operating costs per passenger, measured in 2005 U.S. dollars, ranged from US$7.51 to US$32.35 during the eight-year Efficiency Estimation 95 Figure 3.16 Labor Costs per Passenger a. Group 1 1999 2000 2001 2002 MEX 2003 2004 2005 2006 1996 1997 1998 1999 2000 CGH 2001 2002 2003 2004 2005 2006 1996 1997 1998 1999 2000 GRU 2001 2002 2003 2004 2005 2006 mean LAC 2005 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 US$ (2005) airports privately operated airports publicly operated (continued next page) 96 Airport Economics in Latin America and the Caribbean Figure 3.16 (continued) b. Group 2 1999 2000 2001 2002 CUN 2003 2004 2005 2006 1999 2000 2001 2002 SCL 2003 2004 2005 2006 2000 2001 2002 EZE 2003 2004 2005 2006 2001 LIM 2002 2003 mean LAC 2005 0 0.2 0.4 0.6 0.8 1.0 1.2 US$ (2005) airports privately operated airports publicly operated (continued next page) Efficiency Estimation 97 Figure 3.16 (continued) c. Group 3 2001 2002 2003 SJO 2004 2005 2006 2005 GYE 2006 mean LAC 2005 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 US$ (2005) airports privately operated airports publicly operated Source: Author’s calculation. Note: For a list of airport codes and the airports they represent, see page xxiii. period starting in 2000. In the case of Santiago de Chile’s (SCL) airport, this same partial performance measure ranged from US$2.00 to US$3.30 between 1999 and 2007. The evolution of the two cost efficiency measures, labor costs per passenger and operating costs per passenger, does not enable us to draw definite conclusions. It appears the airports experienced an overall reduction in both cost indicators between 1996 and 2007. However, there are wide differences across airports. No conclusion can be drawn 98 Airport Economics in Latin America and the Caribbean Figure 3.17 Operating Costs per Passenger a. Group 1 1999 2000 2001 2002 MEX 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 CGH 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 GRU 2003 2004 2005 2006 2007 mean LAC 2005 0 2 4 6 8 10 12 14 US$ (2005) airports privately operated airports publicly operated (continued next page) Efficiency Estimation 99 Figure 3.17 (continued) b. Group 2 1999 2000 2001 2002 CUN 2003 2004 2005 2006 2007 1999 2000 2001 2002 SCL 2003 2004 2005 2006 2007 2000 2001 2002 2003 EZE 2004 2005 2006 2007 2001 2002 2003 LIM 2004 2005 2006 2007 mean LAC 2005 0 5 10 15 20 25 30 35 US$ (2005) airports privately operated airports publicly operated (continued next page) 100 Airport Economics in Latin America and the Caribbean Figure 3.17 (continued) c. Group 3 2001 2002 2003 SJO 2004 2005 2006 2007 2005 GYE 2006 2007 mean LAC 2005 0 1 2 3 4 5 6 7 8 9 10 US$ (2005) airports privately operated airports publicly operated Source: Author’s calculation. Note: For a list of airport codes and the airports they represent, see page xxiii. when airports are divided by type of operator (public or private) or even by size. Revenue Indicators Revenue indicators are important for efficiency analysis, as they provide a signal on how well an airport is generating revenues with the same or Efficiency Estimation 101 even fewer inputs than its competitors (figure 3.18). Considering the airports included in our dataset, only one airport (Cancún) shows a steady growth trend of the ratio of total revenues to passengers. The evi- dence for the other airports is mixed, with some airports showing impor- tant gains and others showing a significant reduction. As with the previous indicators, it is not possible to draw conclusions on the behavior of the ratio of total revenues per passenger across size of airport or type of operator (public or private). Revenue per Employee When analyzing the total revenue generated by the average employee in each airport (figure 3.19), a contrast can be observed between the air- ports in Santiago de Chile (SCL) and in Buenos Aires (EZE). The total revenue per employee in SCL was increasing from 1999 to 2007, with only small decreases in 2002 and 2006. In contrast, the total revenue per employee in EZE significantly decreased from 2000 to 2007, and par- ticularly between 2000 and 2003. Congonhas International (CGH) and Guarulhos International (GRU) both in Brazil have also experienced a decrease in the total revenue per employee, but these have been more moderate than the decrease in EZE. In CGH and GRU, the total revenue per employee decreased 12 percent and 27 percent, respectively, between 2000 and 2007, while in EZE it decreased 67 percent during the same period. In Cancún (CUN) the total revenue per employee had an erratic behavior, while in Mexico City’s airport (MEX) as well as in Lima’s (LIM) it has remained stable. According to the information gathered in the database, the fluctuation of this indicator is explained mostly by changes in revenues rather than in the quantity of employees. That is expected, as employment is diffi- cult to adjust. Even though it is not possible to provide a robust conclu- sion regarding the relationship between ownership and the level and trend of this indicator, it seems that private operators perform better, both in the improvement through time of this indicator and the absolute level (that is, revenue per employee is higher in airports operated by the private sector). Quality Indicators There are several other partial performance indicators that can be stud- ied. Ideally, revenue and cost indicators should be complemented and expanded with quality indicators. However, there are very few airports and virtually no regulator that releases data on the evolution of quality variables. There are, nonetheless, certain measures that give a proxy of the 102 Airport Economics in Latin America and the Caribbean Figure 3.18 Total Revenue per Passenger a. Group 1 2001 2002 2003 2004 MEX 2005 2006 2007 1998 1999 2000 2001 2002 CGH 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 GRU 2003 2004 2005 2006 2007 mean LAC 2005 0 5 10 15 20 25 US$ (2005) airports privately operated airports publicly operated (continued next page) Efficiency Estimation 103 Figure 3.18 (continued) b. Group 2 1999 2000 2001 2002 CUN 2003 2004 2005 2006 2007 1999 2000 2001 2002 SCL 2003 2004 2005 2006 2007 2000 2001 2002 2003 EZE 2004 2005 2006 2007 2001 2002 2003 LIM 2004 2005 2006 2007 mean LAC 2005 0 5 10 15 20 25 30 35 US$ (2005) airports privately operated airports publicly operated (continued next page) 104 Airport Economics in Latin America and the Caribbean Figure 3.18 (continued) c. Group 3 2001 2002 2003 SJO 2004 2005 2006 2007 2005 GYE 2006 2007 mean LAC 2005 0 5 10 15 20 25 US$ (2005) airports privately operated airports publicly operated Source: Author’s calculation. Note: For a list of airport codes and the airports they represent, see page xxiii. quality of service provided by the airport, such as the previously described passengers per boarding bridge and passengers per square meter of termi- nal area. Figure 3.20 shows the evolution of passengers per boarding bridge. Noteworthy are the large drops in the number of passengers per boarding Efficiency Estimation 105 bridge in Mexico City (MEX), Cancún (CAN), Santiago de Chile (SCL), Bogotá (BOG), Guayaquil (GYE), and Panama City (PTY). The decrease in the number of passengers per boarding bridge observed in Mexico City’s airport (MEX) in 2001 and 2007 are the consequence of an addi- tion of eight and twenty-three new boarding bridges, respectively. On the Figure 3.19 Total Revenue per Employee a. Group 1 2001 2002 2003 MEX 2004 2005 2006 2007 1998 1999 2000 2001 2002 CGH 2003 2004 2005 2006 2007 1998 1999 2000 2001 GRU 2002 2003 2004 2005 2006 2007 mean LAC 2005 0 50 100 150 200 250 300 US$, thousands (2005) airports privately operated airports publicly operated (continued next page) 106 Airport Economics in Latin America and the Caribbean Figure 3.19 (continued) b. Group 2 1999 2000 2001 2002 CUN 2003 2004 2005 2006 2007 1999 2000 2001 2002 SCL 2003 2004 2005 2006 2007 2000 2001 2002 2003 EZE 2004 2005 2006 2007 2001 2002 2003 LIM 2004 2005 2006 2007 mean LAC 2005 0 200 400 600 800 1,000 1,200 US$, thousands (2005) airports privately operated airports publicly operated (continued next page) Efficiency Estimation 107 Figure 3.19 (continued) c. Group 3 2001 2002 2003 SJO 2004 2005 2006 2007 2005 GYE 2006 2007 mean LAC 2005 0 50 100 150 200 250 300 350 400 450 US$, thousands (2005) airports privately operated airports publicly operated Source: Author’s calculation. Note: For a list of airport codes and the airports they represent, see page xxiii. other hand, Cancún’s airport added 11 new boarding bridges in 2007, while Santiago de Chile’s airport added 6 boarding bridges in 1999 and 7 in 2001. This was also the case in Bogotá’s airport in 2001 and Guayaquil’s and Panama City’s airports in 2006. The calculation of the time evolution of partial performance indicators carried out in this section provided more information to assess airports’ 108 Airport Economics in Latin America and the Caribbean performance. To facilitate the presentation, airports were divided in groups using size as the grouping criterion. As was the case with the com- parison of partial performance indicators for the year 2005, the time series evolution of partial performance indicators does not allow the Figure 3.20 Passengers per Boarding Bridge a. Group 1 1999 2000 2001 2002 MEX 2003 2004 2005 2006 2007 2004 2005 CGH 2006 2007 1995 1996 1997 1998 1999 2000 GRU 2001 2002 2003 2004 2005 2006 2007 mean LAC 2005 0 500 1,000 1,500 2,000 2,500 # of passengers, thousands airports privately operated airports publicly operated (continued next page) Efficiency Estimation 109 Figure 3.20 (continued) b. Group 2 1999 2000 2001 2002 CUN 2003 2004 2005 2006 2007 1995 1996 1997 1998 1999 2000 BOG 2001 2002 2003 2004 2005 2006 2007 1998 1999 2000 2001 2002 SCL 2003 2004 2005 2006 2007 2000 2001 2002 2003 EZE 2004 2005 2006 2007 2005 LIM 2006 2007 mean LAC 2005 0 200 400 600 800 1,000 1,200 1,400 1,600 # of passengers, thousands airports privately operated airports publicly operated (continued next page) 110 Airport Economics in Latin America and the Caribbean Figure 3.20 (continued) c. Group 3 2005 GYE 2006 2007 2003 2004 PTY 2005 2006 2007 2001 2002 2003 POS 2004 2005 2006 mean LAC 2005 0 100 200 300 400 500 600 700 800 900 # of passengers, thousands airports privately operated airports publicly operated Source: Author’s calculation. Note: For a list of airport codes and the airports they represent, see page xxiii. Efficiency Estimation 111 generation of a single measure of efficiency to rank the performance of airports. To that end and following the recent evolution in the specialized literature, the next section calculates aggregate measures of efficiency and compares its evolution for LAC airports. Measuring Technical Efficiency of Airports in LAC Countries In order to overcome the shortcomings of partial performance indicators, this section computes aggregate productivity measures. By taking advan- tage of a database collected by the Air Transport Research Society for the main airports in Asia, Europe, and North America and the use of the responses received to the questionnaire sent to Latin American airports, this section presents the computation of a worldwide benchmark for Latin American airports. The content of this section is quite technical, as it reflects the latest developments in the specialized literature on the computation of effi- ciency frontiers. To the extent possible, the intuition behind the results is laid out, and an introduction based on graphical representations of theo- retical concepts precedes the results.15 The DEA approach was chosen among the different productivity mea- sures most frequently used to compute technical efficiency scores. The selection of DEA is explained by the availability of data and their quality. Ideally, further research should be conducted with other approaches to test whether the results are consistent across estimation techniques. This section is divided into three parts. The first computes a DEA frontier for commercial airports around the world using data for the years 2005 and 2006. This estimation allows the identification of the performance of LAC airports relative to the best practices in the sector. For each airport in the LAC region, it was possible to assess if it is on the frontier that is defined by the most efficient airports in the sample. If it is not, then the set of airports (referred to as peers) with similar productive characteristics that make up the frontier for each airport is identified. The second part of this section attempts to identify factors that drive the differences in observed efficiency in the airport sector. In order to do this a truncated regression model is estimated, using the efficiency scores obtained from the DEA efficiency frontier estimation as depen- dent variables. Several variables that attempt to capture the institu- tional framework and socioeconomic environment in which the airport 112 Airport Economics in Latin America and the Caribbean operates, as well as specific characteristics of each airport (share of aeronautical revenues, hub airport, among others) are included as inde- pendent, or explanatory, variables. Finally, the third part measures the total factor productivity change (TFPC) for LAC airports over the period 1995–2007. The methodology used, which is explained in detail below, consists of the computation of a Malmquist quantity index of TFPC based on the nonparametric DEA approach. Computing a Worldwide Airport Efficiency Frontier In this subsection, technical efficiency scores for 148 airports world- wide are computed and a DEA activity frontier is built. The data com- prises the years 2005 and 2006 from 22 LAC airports, 23 airports from East Asia and Pacific, 40 from Europe, and 63 from Canada and the United States. DEA is a deterministic nonparametric approach used to build a bench- mark, the best practice frontier, based on available information. One of the main advantages of this approach is that it takes into account the multi-output and multi-input dimensionality of production, which is a characteristic of the production function of airports. Another advantage is that computations are based exclusively on measures of physical out- puts and inputs, without the need to use prices, which are very difficult to collect and compare, particularly at the international level. Two models are estimated under the competing assumptions: constant returns to scale (CRS) and variable returns to scale (VRS).16 This allows us to compute scale efficiencies and to identify for each airport the returns-to-scale region—increasing, constant, or decreasing—in which it operates. The calculations of this report assume that airports have as a production target the maximization of outputs for a given input combi- nation; therefore, an output-oriented framework is used. Figure 3.21 illustrates CRS and VRS frontiers in a simple one-output (y) one-input (x) setting. The points P, Q, R, S, and T illustrate the observed quantities of input used and output produced by different pro- duction units (in our case airports). Two frontiers, best practice convex envelope in the DEA terminology, are computed assuming constant and variable returns to scale (CRS and VRS). R is the only point at which a production unit is technically efficient under both CRS and VRS. In other words, the production unit at R operates at the optimum constant returns scale. S and T are efficient under the VRS assumption, with S in the region of increasing returns to scale and T in the region of decreasing Efficiency Estimation 113 Figure 3.21 DEA-CRS and DEA-VRS Frontiers CRS frontier VRS frontier y PC T PV R output P Q S 0 A x input Source: Coelli et al. 2003. returns to scale. Finally, P and Q are technically inefficient. They would be able to produce more output units using the same input quantities. For instance, production unit P uses quantity A of input x to produce quan- tity AP of output y. The vector PP measures the distance to the best C practice frontier. It can be decomposed into two parts: the distance PPV corresponds to the pure technical inefficiency, while the distance PV PC denotes technical inefficiency due to the scale of operation. As can be seen from figure 3.21, production unit P is compared with firms R and T (its peers), which form the piecewise linear combination benchmark to which unit P is compared. Similarly, the peers for production unit Q are R and S. Finally, note that under the CRS hypothesis, production unit R is the benchmark for all the other production units. The literature that estimates aggregate airport efficiency measures is very recent but has grown quickly in the decade thanks to the avail- ability of comparable data, mostly in developed countries. Pestana Barros and Dieke (2008) present an overview of this literature, showing that most studies use the DEA approach, which takes into account the multi-output and multi-input nature of the business. However, there are considerable differences across studies in defining inputs and out- puts. On the outputs side, the more complete and often-used model specification includes three output dimensions: passenger, freight, and aircraft movements. On the inputs side there is lesser consensus in the literature, mainly due to data availability problems. In any case, most studies include a bundle of variables representing labor and capital 114 Airport Economics in Latin America and the Caribbean inputs. The most commonly used variables are the number of employ- ees, as proxy for labor input, and the number or size of runways, termi- nal size, and number of boarding bridges, as proxies for capital. When comparable accounting data are available, inputs are represented by operating costs and capital stock. In addition to physical variables, the data available for the preparation of this report also include information on airport revenues and operating expenditures and in some cases valuation of capital investments. However, given the lack of homogeneity in the definition of these variables and the fact that they were gathered from different sources, they were excluded from the analysis. Consequently, the use of physical input quantities remained as the only possible choice for the calculation of the efficiency frontier. This may be an inferior solution but one that seems to have fewer potential measurement biases. In summary, the availability and comparability of data at an interna- tional level allow the specification of the airport business as a three- output and three-input production function: Outputs Passengers Tons of freight Aircraft movements Inputs Employees Runways Boarding bridges The data, corresponding to the years 2005 and 2006, are well balanced for the 22 LAC airports but unbalanced for the other regions of the world, particularly for European airports.17 For this reason, the data were pooled to carry out the benchmark study. In other words, a single DEA frontier was computed for the period 2005–06. Table 3.4 presents descriptive statistics for outputs and inputs by region. LAC airports are on average smaller than those from other regions in terms of the three outputs: passengers, tons of freight, and aircraft movements. Despite these differences in the scale of production, on aver- age, LAC airports employ nearly as much staff as Canadian and U.S. air- ports. At the same time, in terms of capital investments, the number of runways and boarding bridges is several times lower in LAC airports than in Canadian and U.S. airports. Table 3.5 presents the technical efficiency (TE) results for the airports in the four regions, which were calculated by performing DEA computa- tions using the Data Envelopment Analysis (Computer) Program (DEAP; Efficiency Estimation 115 Table 3.4 Descriptive Statistics by World Region, 2005–06 Outputs (× 1,000) Inputs Tons of Aircraft Boarding Statistics Passengers freight movements Employees Runways bridges LAC (22 airports, 44 observations) Mean 6,430.6 117.2 96.1 424.0 1.5 11.3 STD 6,033.6 119.0 82.9 412.0 0.5 9.7 Min. 181.0 0.2 1.9 20.0 1.0 0.0 Max. 24,727.0 470.9 356.0 1,568.0 2.0 38.0 East Asia and Pacific (23 airports, 39 observations) Mean 18,776.7 836.0 148.2 1,044.0 1.7 52.3 STD 12,432.4 970.7 82.7 1,107.3 0.6 35.5 Min. 1,293.3 10.3 10.5 137.0 1.0 0.0 Max. 45,100.0 3,600.0 286.5 4,873.0 3.0 143.0 Europe (40 airports, 66 observations) Mean 19,305.0 318.3 211.8 2,029.4 2.3 67.9 STD 15,728.4 515.7 127.8 2,982.6 1.0 58.3 Min. 1,218.9 3.6 29.8 298.0 1.0 6.0 Max. 67,915.0 2,131.0 533.0 17,528.0 6.0 264.0 Canada and the United States (63 airports, 125 observations) Mean 21,318.4 406.5 310.9 549.9 3.4 69.9 STD 17,976.6 641.8 196.7 480.7 1.2 42.6 Min. 2,657.1 3.6 60.5 119.0 1.0 14.0 Max. 85,907.4 3,713.4 980.4 3,000.0 7.0 178.0 Source: Author’s compilation using World Bank Airports LAC Benchmarking Database and ATRS 2008. Note: STD = standard deviation; min = minimum value, max = maximum value. Table 3.5 Average Technical Efficiency Scores and Scale Efficiency by Region, 2005–06 Returns to scale diagnosis Technical efficiency (% of observations) World region CRS VRS SE IRS CRS DRS Latin America 0.532 0.690 0.801 70.5 9.1 20.5 East Asia and Pacific 0.670 0.771 0.869 84.6 12.8 2.6 Europe 0.490 0.530 0.927 43.9 6.1 50.0 Canada and the United States 0.540 0.616 0.875 23.2 8.0 68.8 All 0.545 0.629 0.875 44.5 8.4 47.1 Source: Author’s estimation. Note: CRS = constant returns to scale; DRS = decreasing returns to scale; IRS = increasing returns to scale; SE = scale efficiency; VRS = variable returns to scale. 116 Airport Economics in Latin America and the Caribbean Coelli 1996). The average TE score of airports in all regions is 0.545 under the constant returns to scale (CRS) assumption. This means that, on average, the airports included in the sample could almost double their outputs (passengers, tons of freight, and aircraft movements) with the same quantity of inputs they currently use. However, part of the distance to the best practice CRS frontier is explained by the scale of operation. Under the variable returns to scale (VRS) assumption, the average TE is 0.629 and the average scale effi- ciency is 0.875.18 Table 3.5 also shows the distribution of airports in each region according to the type of production scale (increasing, con- stant, or decreasing). The last three columns of table 3.5 report the percentage of airports corresponding to this classification. Grouping all regions, 44.5 percent, 8.4 percent, and 47.1 percent of the airports in our dataset operate under increasing, constant, and decreasing returns to scale, respectively. LAC airports appear to be the ones that suffer the most from a subop- timal scale operation. Scale inefficiency is close to 20 percent (scale effi- ciency [SE] = 0.801), mainly concentrated in the increasing returns to scale area (70.5 percent of observations). This means that on average, airports in LAC could improve their efficiency 20 percent if they were to increase their scale of operation to the optimal scale. Contrary to that finding, nearly 70 percent of Canadian and U.S. airports operate in the decreasing returns to scale region. The results of returns to scale diagnosis coincided with the intuition: airports in LAC are smaller, and given that the production technology of airports is characterized by large fixed investments (runways, terminals), it is logical to expect that smaller air- ports are still in the increasing returns to scale zone of the production function. It should be noted that airports identified as operating at the optimal scale (CRS) in our database handle between 20 and 30 million passengers each year, a result that exceeds previous estimates.19 The rel- evant policy question is whether airports can influence the scale of operations. The answer depends on many factors, including the availabil- ity of land to build new facilities, existence of competition, congestion of existing facilities, and the possibility of changing airport tariffs. However, airports have strong limitations in the extent of influence on the demand they face. It is a well-accepted fact that airports face a derived demand, and consequently, they cannot significantly alter the outputs when they change inputs. Notable exceptions are airports that suffer strong conges- tion, but the argument is valid only for airport expansion (increase in inputs). When airports face a scenario of output contraction caused by a Efficiency Estimation 117 macroeconomic crisis (for instance, the financial crises that began in 2008), there is not much they can do to adjust the scale of operation because inputs remain constant (runways, terminals20) when output falls due to factors completely out of their control. Table 3.6 presents detailed results for LAC airports.21 Only two airports in the region are technically efficient under both CRS and VRS: Congonhas (CGH) and Viracopos (VCP), both in São Paulo. However, it is important to highlight that VCP is a special case: it is an efficient unit in DEA by default, which occurs when a production unit has no peers to which it can be compared. VCP is an airport that during our sample period can be characterized as a dedicated freight airport, as it has virtually no passenger Table 3.6 Average Technical Efficiency Scores for LAC Airports, 2005–06 Scale Country Airport CRS VRS efficiency Argentina AEP 0.612 0.998 0.614 EZE 0.414 0.417 0.993 FTE 0.115 1.000 0.115 Brazil BSB 0.498 0.536 0.931 CGH 1.000 1.000 1.000 GIG 0.318 0.320 0.994 GRU 0.677 0.678 0.998 MAO 0.377 0.692 0.544 VCP 1.000 1.000 1.000 Chile SCL 0.786 1.000 0.786 Colombia BAQ 0.329 0.524 0.628 CLO 0.496 0.734 0.676 Costa Rica SJO 0.594 0.983 0.605 Dominican Republic SDQ 0.260 0.372 0.699 Ecuador GYE 0.472 0.646 0.739 El Salvador SAL 0.114 0.127 0.900 Mexico CUN 0.860 1.000 0.860 GDL 0.643 0.649 0.991 MEX 0.961 0.963 0.998 MTY 0.403 0.410 0.982 Panama PTY 0.164 0.178 0.926 Peru LIM 0.621 0.961 0.646 All 0.532 0.69 0.801 Source: Author’s estimation. Note: CRS = constant returns to scale; TE = technical efficiency; VRS = variable returns to scale. For a list of airport codes and the airports they represent, see page xxiii. 118 Airport Economics in Latin America and the Caribbean movement and no boarding bridges. Other results of table 3.6 can be sum- marized as follows: (a) TE scores for LAC airports show notable variations, from airports on the frontier (with a value of 1) to airports that have TE scores close to 0; (b) when assuming CRS, only two airports, CGH and VCP, are on the frontier; and (c) when VRS is assumed and, consequently, scale efficiency is isolated, the TE of LAC airports improves. Out of 22 airports, 6 are on the frontier. The later subsection on sources of technical efficiency tries to identify the variables that explain the observed differ- ences in TE scores across airports. As previously mentioned in this subsection, the DEA approach allows the identification of peers for each airport, which are the set of efficient airports that make up the relevant frontier for a given airport. Table 3.7 presents the peers for LAC airports in 2005 under the DEA VRS model. It should be noted that, by construction, technically effi- cient airports do not have other airports as peers. Technically inefficient airports have, on the contrary, a benchmark composed by other units. Given the three-output and three-input dimensionality of the produc- tion setting, the maximum number of peers is six, but an airport can have fewer than six peers. It is important to remark that some LAC airports are peers for other airports. They serve as peers not only for other airports in the LAC region but also for other airports around the world. This is the case mainly of CGH (Congonhas, São Paulo), which is a peer for 28 observations (2005 and 2006 airport observations taken together). Other airports playing the same role of peers are AEP (Aeroparque Jorge Newbery, Buenos Aires), SCL (Comodoro Arturo Merino Benítez, Santiago de Chile), CUN (Cancún), and, to a lesser extent, FTE (Calafate, Argentina) and SJO (San José, Costa Rica). An interesting result is that all LAC airports in our sample, with the exception of MAO (Manaus, Brazil), have at least one LAC airport as a peer. Eight airports from outside the LAC region act as peers for LAC airports: XMN (Xiamen), ICN (Seoul), SDF (Louisville), LAX (Los Angeles), MEM (Memphis), SNA (Santa Ana, California), ATL (Atlanta), and MFM (Macao SAR, China).22 For illustration purposes, consider in more detail one observation: BSB airport (Juscelino Kubitschek, Brasilia). For this airport, the computed DEA TE score was 0.552, which corresponds to a 45 percent output inef- ficiency diagnosis. The airports identified as peers for BSB are CGH (Congonhas, São Paulo) and three U.S. airports: LAX (Los Angeles), MEM (Memphis), and SNA (Santa Ana, California). If BSB is compared with CGH, its only LAC peer, and one looks at some of their main Efficiency Estimation 119 Table 3.7 Peer Analysis, DEA VRS, 2005 As peer Peers TE VRS for other Country Airport 2005 airports 1 2 3 4 5 Argentina AEP 1.000 9 AEP EZE 0.404 0 CGH (CGH) (XMN) (ICN) (SDF) FTE 1.000 7 FTE Brazil BSB 0.552 0 CGH LAX MEM SNA CGH 1.000 28 CGH GIG 0.316 0 (CGH) (XMN) (ICN) ATL GRU 0.680 0 (CGH) (XMN) (ICN) ATL MAO 0.680 0 SJO (XMN) MFM SNA VCP 1.000 0 VCP Chile SCL 1.000 10 SCL Colombia BAQ 0.507 0 FTE SJO (XMN) SNA CLO 0.747 0 (FTE) SCL SNA Costa Rica SJO 1.000 6 SJO Dominican SDQ 0.386 0 AEP (LIM) (SCL) SNA (XMN) Republic Ecuador GYE 0.814 0 (FTE) SJO (XMN) SNA El Salvador SAL 0.131 0 (CGH) LAX MEM SNA Mexico CUN 1.000 11 CUN GDL 0.615 0 CGH FTE (XMN) (SDF) MEX 0.947 0 CGH ICN (XMN) ATL SNA MTY 0.424 0 CGH (FTE) (ATL) MEM SNA Panama PTY 0.188 0 CGH ICN (XMN) (SDF) SNA Peru LIM 0.922 0 AEP (LIM) (SCL) (XMN) SNA Source: Author’s estimation. Note: DEA = Data Envelopment Analysis; VRS = variable returns to scale. Underlined peers are LAC airports. Ob- servations in parentheses are 2006 observations. Other airports: ICN (Seoul, Republic of Korea); MFM (Macao SAR, China); XMN (Xiamen, China); ATL (Atlanta, Georgia); SDF (Louisville, Kentucky); MEM (Memphis, Tennessee); LAX (Los Angeles, California); SNA (Santa Ana, California). For a list of airport codes and the airports they represent, see page xxiii. output-input features (for the year 2005), DEA results are confirmed. On the output side, BSB handles 9.4 million passengers per year, against the 17.1 million passengers of CGH. Similarly, BSB had 171,600 aircraft movements in 2005, against 282,600 aircraft movements in CGH. Finally, on the input side, BSB had 365 employees and 13 boarding bridges, while CGH had 225 employees and eight boarding bridges. 120 Airport Economics in Latin America and the Caribbean Identifying Sources of Technical Efficiency The previous subsection presented the estimation of technical effi- ciency for 148 airports in the world and showed the results for the LAC airports included in the sample. With the TE scores in hand, the logi- cally subsequent question is: What are the variables that explain the observed differences in technical efficiency across airports? The previous subsection showed that a fraction of the variation in TE scores can be explained by the scale of operation. However, even when a VRS model, which isolates the scale component of technical inefficiency, is used, significant differences in technical efficiency exist. A potential factor behind the observed differences in efficiency is qual- ity. It is likely that, other things being equal, airports operating with a large staff and/or a large number of boarding bridges provide better ser- vice quality to passengers. Unfortunately, survey data on users’ satisfac- tion are not yet available at an international scale, so we were not able to include quality indicators in our analysis. It is possible to divide the measurable potential drivers of efficiency into two groups that can be distinguished by the degree of control that each airport has over these variables. Among the exogenous (out of air- ports’ control) drivers, the institutional setting or the demographic and socioeconomic environment in which airports operate can be included. Within the group of variables for which airports have a higher degree of control (endogenous), the percentage of passengers in transit (an attri- bute of hub airports), and the importance of nonaeronautical activities (duty-free shops, parking, local transportation, and so forth) can be included. This subsection tests the effects of some of these potential factors using available information from different sources. To that end, a trun- cated regression model is estimated using the airport TE scores of the previous subsection as dependent variables and the exogenous and endogenous drivers as explanatory variables. The choice of a truncated model is dictated by the nature of the TE measure (which is by definition truncated at 1.0) and by the use of this model in the most recent aca- demic literature (Simar and Wilson 2007).23 Table 3.8 presents average values by region for the candidate variables to account for observed differences in technical efficiency. Starting with the institutional setting, table 3.8 shows that, on average, LAC airports operate under a more liberalized framework. Indeed, more than half of LAC airports (54.5 percent) in the sample operate as private concessions, and 31.8 percent are regulated by an independent regulatory agency. In Efficiency Estimation 121 Table 3.8 Potential Explanatory Factors of Technical Inefficiency, 2005–06 Canada Latin and United Explanatory factors America Asia Europe States Institutional framework Private airport (%) 54.5 25.6 37.9 0 Independent regulatory agency (%) 31.8 10.3 16.7 0 Socioeconomic environment GDP per capita (US$) 5,442 17,397 32,598 42,219 Tourism expenditures per capita (US$) 69 532 943 393 Population concentration Population in the area (1,000s) 7,719 6,709 3,200 3,984 Population > 5 million (% of observations) 45.5 48.7 22.7 34.4 Airport characteristics Hub airport (%) 9.1 17.9 40.9 27.2 Passengers connecting (% of passengers) 7.9 9.5 32.8 23.4 Aeronautical revenues (% of total revenue) 56.9 53.8 51.6 49.2 Source: Author’s estimation. Note: The value of 5 million corresponds to the mean of the population of the cities where airports are located. contrast, only 25.6 percent of Asian airports and 37.9 percent of European airports are under private management, while 10.3 percent and 16.7 percent of Asian and European airports, respectively, are regu- lated by an independent regulatory agency. Finally, all airports in Canada and the United States are operated by state-owned enterprises, and regu- latory agencies in those two countries still depend directly on a political authority (a ministry or a government agency). Another potential factor that could have a role in the explanation of airport performance is the socioeconomic environment in which air- ports operate. This effect is incorporated with two indicators: GDP per capita (measured in nominal U.S. dollars) and tourism expenditures (also measured in nominal U.S. dollars). However, it is worth stressing that these variables are only available at the country level and do not necessarily correspond to the area of influence of the airports.24 The demographic environment is represented by the concentration of population in the area served by the airport. On average, LAC air- ports appear to serve very large urban agglomerations (45.5 percent of airports), like their Asian counterparts (48.7 percent). Compared to European and North American airports, which are on average located in cities with 3 million to 4 million inhabitants, LAC airports are on average located in cities with 8 million people. In the regression analysis, this information will be incorporated with a binary (dummy) 122 Airport Economics in Latin America and the Caribbean variable that takes a value of 1 for airports located in cities with more than 5 million people and of 0 otherwise. Finally, a set of variables that represent characteristics that are particu- lar to each airport is introduced. One of them is their specialization as a hub, represented by the percentage of connecting passengers. LAC air- ports have the lowest percentage of connecting passengers, with 7.9 per- cent (and also have the lowest percentage of hubs—9.1 percent), followed by Asian airports. The highest percentage is observed among European airports, where nearly one-third of passengers are connecting. Another variable that is particular to each airport is the share of aeronautical rev- enues in total revenues. Table 3.8 shows that aeronautical revenues are on average more important for LAC airports (where they represent almost 60 percent of total revenues) than for airports in any other region. Table 3.9 reports the results, in the form of marginal effects, of estima- tions for alternative truncated regression models. The first two columns show the estimates of two models with VRS TE scores as dependent variables, with and without dummies, for each world region. The third column presents the estimates of a model with CRS TE scores as depen- dent variables, without regional dummies. The Likelihood Ratio Tests (LT) indicate that in all three cases, the explanatory variables included in the model, taken all together, have a statistically significant effect on the dependent variable. Even though the results are sensitive to the returns to scale assumption, overall, the sign and magnitude of marginal effects are comparable for VRS and CRS assumptions. The main difference con- cerns the statistical significance of most explanatory variables, which tend to be nonsignificant when CRS is assumed. Several results are worth highlighting. First, it should be noted that there are two variables that appear as the main drivers of technical efficiency in the airport sector. On the one hand hub airports are, on average, 10 percent to 15 percent more efficient than other airports. On the other hand, the population size in the area served by the airport also seems to matter: airports located in areas with more than 5 million inhabitants are 17 percent to 20 percent more efficient than airports that serve less-populated areas. Second, the results show that the institutional variables (whether the airport is private or public and whether it is regulated by an independent regulatory agency) are associated with positive marginal effects. However, these variables are not statistically significant, with the exception of the dummy for private airports under the VRS assumption. According to these results, privately operated airports tend to be more efficient, with a Efficiency Estimation 123 Table 3.9 Truncated Regression—Marginal Effects VRS TE with VRS TE without CRS TE without regional dummies regional dummies regional dummies Explanatory Marginal Marginal Marginal factors effect (std) effect (std) effect (std) Institutional framework Private airport (dummy) 0.064 (0.036)* 0.082 (0.035)** 0.068 (0.041) Regulation authority (dummy) 0.048 (0.048) 0.041 (0.050) 0.083 (0.059) Socioeconomic environment GDP per capita 0.006 (0.002)*** 0.001 (0.001) 0.001 (0.001) Tourism expenditures per capita −0.033 (0.049) −0.005 (0.033) −0.045 (0.036) Population concentration Population > 5 million (dummy) 0.169 (0.025)*** 0.201 (0.027)*** 0.173 (0.031)*** Airport characteristics Hub airport (dummy) 0.122 (0.028)*** 0.099 (0.031)*** 0.153 (0.031)*** Aeronautical revenues −0.150 (0.081)* −0.183 (0.085)** −0.134 (0.102) Control variables (dummies) Asia 0.059 (0.047) — — — — Europe −0.200 (0.059)*** — — — — Canada and the United States −0.201 (0.069)*** — — — — Year 2006 −0.023 (0.023) −0.107 (0.024) −0.210 (0.274) LR test Chi2 (11) 110.3*** Chi2 (8) 80.8*** Chi2 (8) 56.7*** Observations 251 251 251 Source: Author’s estimation. Note: LR = likelihood ratio; CRS = constant returns to scale; VRS = variable returns to scale; — = not applicable. Significance level: * = 10 percent, ** = 5 percent, *** = 1 percent. TE score that is on average 6 percent to 8 percent points higher than publicly operated airports. Another important feature that distinguishes airports is the impor- tance of aeronautical activities in their operation. As expected, the impor- tance of these activities, summarized by the share of aeronautical revenues in the total airport revenue, plays a negative effect on efficiency (although this effect is statistically significant only when we use VRS TE scores as the dependent variable). In other words, airports in which nonaeronautical (that is, commercial) activities are more important tend to be more efficient. The estimated marginal effect indicates that, on 124 Airport Economics in Latin America and the Caribbean average and holding the other variables constant, a 10 percent increase in the share of aeronautical revenues produces a loss in technical efficiency of nearly 2 percent. GDP per capita seems to have a positive effect on airport efficiency. However, this estimate is only significant in the VRS model (with regional dummies). In this case, when GDP per capita increases by US$10,000, the technical efficiency of airports is expected to increase 6 percent. Finally, tourism expenditures are not significant in the three specifications. Measuring Productivity Change of LAC Airports The objective of this subsection is to assess how airport productivity evolved in Latin America and the Caribbean. This exercise tracks the evolution of productive efficiency among LAC airports. It is possible, then, to identify those airports that experienced the largest efficiency gains and that can be categorized as best performers. To that end, total factor productivity change (TFPC) for LAC airports over the period 1995 to 2007 is computed. The period covered was determined by the data compiled through the questionnaires distributed for the elaboration of this report. The computation relies on the same three-output three-input model specification used in the international benchmark study presented above, and the methodology consists of the computation of a Malmquist quantity index of TFPC based on the nonparametric DEA approach (Färe et al. 1994). Figure 3.22 illustrates the computation of the Malmquist index in a simple one-output (y) and one-input (x) setting. Points Mt and Mt+1 cor- respond to consecutive observations of production unit M at period t and t+1, respectively. Based on available information for the sector, two DEA frontiers are computed, one for period t and another one for period t+1, under the assumption of constant returns to scale. The technical efficiency (TE) of unit M in period t corresponds to the ratio AMt/AB, an output-oriented distance function in the produc- tion activity terminology. In period t+1, the TE of unit M is given by the distance function DMt+1/DF. Proceeding in the same way, and using the same information, it is possible to compute two auxiliary distance functions. One measures the distance separating Mt from the frontier in period t+1, given by AMt/AC, and the other measures the distance separating Mt+1 from the frontier computed in period t, given by DMt+1/DE. Efficiency Estimation 125 Figure 3.22 Malmquist Index of Total Factor Productivity Change y period t+1 frontier F period t frontier output Mt+1 C E B Mt 0 A D x input Source: Färe et al. 1994. Färe et al. (1994) show how to compute a Malmquist index of TFPC, based on the four distance functions introduced above. In terms of figure 3.22, total factor productivity change of firm M from period t to period t+1 is computed as follows: 0.5 ⎛ DMt +1 AMt DMt +1 AMt ⎞ TFPCM = ⎜ × ⎟ � DE AB DF AC ⎠ (1) After some simple algebraic manipulations, this formula can be restated as: 0.5 ⎛ DMt +1 AMt ⎞ ⎛ AC DF ⎞ TFPCM = ⎜ ⎟×⎜ × ⎟ � DF AB ⎠ � AB DE ⎠ (2) From figure 3.22 it can be easily verified that the first term in brackets in the right side of equation 2 corresponds to the productivity improve- ment of unit M, from period t to period t+1, in terms of technical effi- ciency. This term is known as technical efficiency change (TEC). The second term in brackets, known as technical change (TC), measures the frontier shift between period t and period t+1. It corresponds to the fron- tier shift computed as the geometric mean of the change in technology between the two periods. 126 Airport Economics in Latin America and the Caribbean The Malmquist index of TFPC presents two advantages with respect to traditional index numbers. On the one hand, prices are not needed in order to calculate this index. On the other hand, the index can be decom- posed into a measure of technical progress (TC) of the activity level taken as a whole, and another measure (TEC) that captures how each unit is catching up with respect to the technological frontier. Its main disadvan- tage compared with traditional index numbers is that it cannot be com- puted separately for each unit. Its computation relies on the estimation of sequential frontiers. And for this purpose, panel data must be available for representative units operating in the sector. This section of chapter 3 relies on a panel composed exclusively of LAC airports for the period 1995–2007. Unfortunately, the international panel including airports around the world is only available for the years 2005 and 2006, and consequently the computation of TFPC could not be done and used as a benchmark for LAC airports. Table 3.10 presents descriptive statistics for the three subperiods in which the sample is decomposed: 1995–99, 2000–03, and 2004–07. For each of these three subperiods, the number of airports in the sam- ple varies noticeably, from 7 to 22. As a consequence, the benchmark Table 3.10 Descriptive Statistics by Period Outputs (× 1,000) Inputs Tons of Aircraft Boarding Statistics Passengers freight movements Employees Runways bridges 1995–99 (7 airports, 26 observations) Mean 5,039.7 145.4 119.1 723.5 1.5 9.7 STD 4,586.1 125.7 80.9 690.0 0.5 10.1 Min. 250.6 21.4 30.5 77.0 1.0 0.0 Max. 14,705.1 409.2 293.8 2,056.0 2.0 38.0 2000–03 (17 airports, 60 observations) Mean 6,136.6 132.8 112.4 429.3 1.5 11.8 STD 5,314.4 124.2 88.0 465.8 0.5 10.9 Min. 654.8 10.4 29.5 56.0 1.0 0.0 Max. 21,694.0 418.9 334.5 1,940.0 2.0 38.0 2004–07 (22 airports, 85 observations) Mean 6,579.4 121.1 99.1 433.9 1.5 12.0 STD 5,992.7 120.6 83.6 421.9 0.5 10.7 Min. 157.9 0.0 1.9 20.0 1.0 0.0 Max. 25,882.0 470.9 379.0 1,598.0 2.0 56.0 Source: Author’s estimation based on World Bank Airports LAC Benchmarking Database and ATRS 2008. Note: STD = standard deviation; min = minimum value; max = maximum value. Efficiency Estimation 127 used for TFPC computations varies as well, and the results should be interpreted carefully, mainly for the TFPC decomposition into TEC and TC.25 Table 3.11 presents the estimation of TFPC by subperiod and by LAC airport.26,27 Average productivity growth oscillated over the three subpe- riods. Between 1995 and 1999 airports in the region posted an average annual productivity growth of −2.7 percent. However, it should be noted that this growth corresponds to the average scores of Brazilian airports and the airport in Barranquilla, Colombia, the only airports for which data are available for this period. The results are driven by the strong negative growth of the airport in Barranquilla. Table 3.11 Average Annual Total Factor Productivity by Airport and Subperiod percentage Country Airport 1995−99 1999−2003 2003−07 Argentina AEP — −7.0 −3.0 EZE — −18.9 4.0 FTE — — 22.9 Brazil BSB 10.0 5.4 2.9 CGH 13.8 2.6 −4.0 GIG 7.4 −5.5 16.3 GRU 3.5 −0.9 2.7 MAO −2.3 0.3 6.8 VCP 0.9 −7.6 −0.8 Chile SCL — 1.3 2.0 Colombia BAQ −23.0 −8.4 1.5 CLO — −6.2 −5.1 Costa Rica SJO — 22.1 0.0 Dominican Republic SDQ — — −3.7 Ecuador GYE — — 8.1 El Salvador SAL — 2.7 1.4 Mexico CUN — 6.6 −0.3 GDL — −6.1 9.5 MEX — 1.1 4.9 MTY — 5.8 4.7 Panama PTY — — 7.4 Peru LIM — — 9.7 All −2.7 −1.2 3.9 Source: Author’s estimation. Note: For a list of airport codes and the airports they represent, see page xxiii. 128 Airport Economics in Latin America and the Caribbean Productivity growth during the intermediate subperiod (1999– 2003) was negative (−1.2 percent per year on average) and was driven mainly by some airports that experienced dramatic losses in productiv- ity, like EZE (Ministro Pistarini, Buenos Aires) which showed a loss in productivity of −18.9 percent per year over this period as a direct con- sequence of the economic and financial crisis Argentina suffered dur- ing 2001/02. Conversely, positive rates of growth appear to be the norm (with only some exceptions) during the last subperiod (2003 to 2007). The average TFPC rate was 3.9 percent during this period, with many air- ports experiencing annual productivity growth rates close to, or even higher than, 10 percent. Different complementary explanations could be driving the high rates of TFPC of this period. But, as negative eco- nomic shocks are a likely explanation of the reduction in productivity between 1999 and 2003, the strong economic growth enjoyed by LAC economies in the period 2003–07 is a strong driver of improvements in airports’ TFPC. One of the main questions that motivated the elaboration of this report was whether privately operated airports in LAC, a region that has experimented with a wide variety of private sector participation schemes for the operation of airports, have higher productivity levels. Table 3.12 presents the evolution of airport productivity by type of ownership and size. To avoid reaching biased conclusions caused by dif- ferences in airport size, airports are weighted using the workload unit (WLU) measure. The results reported in table 3.12 show that the largest airports are the ones that registered faster productivity growth. In particular, those air- ports that handle between 7.5 million and 10 million passengers per year posted an average annual growth rate of 5.4 percent for the whole period, and an even higher growth of 7.0 percent during the last subperiod. Interestingly, the category made up by the three biggest airports in the region (CGH, GRU, and MEX, which handle more than 10 million pas- sengers per year) grew faster during the first subperiod but then grew at a rather low rate over the two last subperiods. Public airports appear to have performed better on average over the whole period compared to private airports (annual productivity changes of 2.9 percent and 0.7 percent, respectively). Nevertheless, if the analy- sis focuses on their evolution over the last two subperiods, for which the available information is more complete, both groups behaved quite similarly, registering negative productivity growth during the period Efficiency Estimation 129 Table 3.12 Average Total Factor Productivity by Airport Categories annual percent change Airport categories 1995−99 1999−2003 2003−07 ALL Nonweighted Size (million passengers) < 5.0 −5.6 −1.8 3.5 0.4 5.0 to 7.5 − −4.3 3.7 0.5 7.5 to 10.0 8.9 1.7 7.0 5.4 > 10.0 8.5 0.9 1.8 3.4 Private vs. public Private −23.0 −1.6 3.4 0.7 Public 5.3 −0.8 4.5 2.9 All 2.7 −1.2 3.9 1.9 Weighted Private vs. public Private −23.2 −0.5 2.7 1.3 Public 6.1 0.2 4.4 3.2 All 5.5 0.0 3.7 2.6 Source: Author’s estimation. Note: Weighted using workload units. Airport categories are composed as following: Public: BSB, CGH, GIG, GRU, MAO, and VCP (Brazil); SAL (El Salvador); MEX (Mexico); and PTY (Panama). Private airports: AEP, EZE, FTE (Argentina); SCL (Chile), BAQ and CLO (Colombia); SJO (Costa Rica); GYE (Ecuador); CUN, GDL, and MTY (Mexico); and LIM (Peru). Airport size: less than 5.0 million passengers: BAQ, CLO, FTE, GYE, MAO, PTY, SAL, SDQ, and SJO; 5.0–7.5 million passengers: AEP, EZE, GDL, LIM, MTY, and SCL; 7.5–10.0 million passengers: BSB, CUN, and GIG; more than 10.0 million passengers: CGH, GRU, and MEX. 1999–2003 and positive growth between 2003 and 2007 (although with a slightly more favorable profile for public airports). These results are confirmed when TFPC averages are weighted using WLU as the weight variable. Finally, table 3.13 presents the decomposition of the Malmquist TFPC index into its two main components, TEC and TC. The table presents both nonweighted and weighted (by WLU) averages. Note that weighted averages give a better approximation of the average productivity growth for the airport activity in the region. Since larger airports performed bet- ter than smaller ones, the weighted average TFPC is higher than the nonweighted average (2.6 percent compared to 2.2 percent). Despite these differences, in general, the results are very similar when nonweighted and weighted averages are used. Both averages show that 130 Airport Economics in Latin America and the Caribbean Table 3.13 Malmquist Total Factor Productivity Index Decomposition—Averages by Period annual percent change Nonweighted Weighted Period TEC TC TFPC TEC TC TFPC By period 1995−99 −1.5 6.0 4.4 −0.6 6.1 5.5 1999−2003 2.3 −3.3 −1.2 1.1 −1.1 0.0 2003−07 6.4 −2.4 3.9 4.3 −0.6 3.7 All 1995−2007 3.6 −1.4 2.2 2.4 0.2 2.6 Source: Author’s estimation. the airport industry in the LAC region did not experience any improve- ment in productivity due to technical change over the period. In other words, there was no significant change in the production frontier of the industry between 1995 and 2007, as the estimated TC index is near zero or even negative, except for the first subperiod. In fact, the table shows that the main source of TFPC corresponds to improvements in TEC, particularly during the last subperiod. This result has to be interpreted in terms of a catching-up process. Most LAC airports grew during the sample period (1995–2007), mainly by better allocating inputs in a framework of well-known technologies and production processes. This process allowed them to position themselves closer to the activity frontier than they were at the beginning of the period. Conclusion This chapter presented a detailed analysis of LAC airports’ performance. It starts by a comparison of the most frequently used partial performance indicators. More than 20 airports in LAC are compared using 2005 data. The graphs that illustrate the partial performance indicators include the average values observed in other regions of the world (Asia, Canada and the United States, and Europe) and the accompanying text explain the main findings and caveats necessary to consider when linking the results of partial performance indicators and coming up with a conclusion about the performance ranking of airports. To enrich the analysis of partial per- formance indicators, the evolution in time of these indicators was carried out whenever data were available. Airports were divided by size to facili- tate the graphical presentation. Efficiency Estimation 131 Overall it is difficult and not always correct to obtain clear-cut conclu- sions about airport performance solely by considering partial performance indicators. The text presents several examples of why reaching conclu- sions about performance just by looking at selected indicators could give misleading conclusions. Still, an effort to rank airports was produced and reflected in table 3.3, which identifies best and worst airport performers in LAC. Relying on some of the most advanced techniques currently in use by specialists in the measurement of productivity, this report presents aggre- gate measures of efficiency. The literature review carried out for the preparation of this report did not identify any publicly available attempt to measure productive efficiency for a representative sample of LAC airports using aggregate productivity measurement techniques. Thus, this report presents the first comprehensive calculation of technical efficiency of airports in the Latin America and Caribbean region. The results indicate that technical efficiency for LAC airports show notable variations: from airports on the frontier (with a value of 1) to airports that have technical efficiency scores close to 0. In the best-case scenario, when variable returns to scale are assumed, out of the 22 LAC airports in the sample, 6 are on the frontier. The results obtained when using aggregate measures of efficiency tend to coincide with comparisons using partial performance indicators. On average, LAC airports are less efficient than Asian and North American airports when constant returns to scale are used, but they are more efficient than European airports. However, when boarding bridges are excluded and not considered proxy for capital investments, LAC air- ports are on average significantly less efficient than in the other regions included in the study. Using information for more than 148 airports worldwide, several factors that explain the observed differences in airport efficiency were identified using regression analysis. As expected, the regression analysis shows that airports that serve as hubs tend to be more efficient. Moreover, airports that are located in cities with more than 5 million inhabitants are also more efficient than airports located in smaller cities. The level of income (GDP per capita) also seems to positively influence productive efficiency. Airports that rely more on revenue sources other than aeronautical tariffs also tend to be more efficient, a finding consistent with the recent literature (ATRS 2008). Finally, airports that are privately operated tend to stand closer to the efficient frontier than their publicly operated counterparts, although this effect is not significant across all the different specifications tested. 132 Airport Economics in Latin America and the Caribbean When analyzing in more detail how LAC airports’ productivity evolved between 1995 and 2007, the calculations indicate that productivity growth has been driven mainly by improvements in technical efficiency and not by pure technical change. This finding implies that the efficient production frontier of the sector did not experience any major shift between 1995 and 2007, but many airports were able to raise their effi- ciency level and become more productive, a process by which they were able to come closer to the efficient frontier. Probably the most unex- pected result is that privately operated airports in LAC have not outper- formed publicly operated airports. Given the wide variety of private participation schemes used by LAC countries, this result should lead to more detailed, case-by-case research to assess the effects of private par- ticipation on airport performance. In addition, future research should also assess financial efficiency as well as the impact of private participation on the quality of service delivered. Notes 1. Gillen and Lall (1997), Parker (1999), and Murillo Melchor (1999) are among the first published papers that measure performance of airports using aggregate measures of efficiency. 2. Flor and de la Torre (2008) use Data Envelopment Analysis (DEA) methods to analyze efficiency and total factor productivity of airports in Peru. Similarly, Fernandes and Pacheco (2002) also employ DEA methods to compute a production frontier using data for Brazilian airports. Gómez-Lobo and González (2008) use DEA to compare the airport of Santiago de Chile with airports in developed countries. The literature review conducted for this report did not identify other papers that use efficiency estimation techniques applied to airports in Latin America. 3. This report would have benefited from an analysis of perception of quality, but the lack of public information regarding passengers’ and airlines’ experi- ence during the consumption of airport services did not allow us to pursue such analysis. 4. See Andrés et al. (2008) for a survey of the recent literature and an applica- tion of partial performance indicators in the electricity, water distribution, and fixed telecommunications sectors. 5. The book A Primer on Efficiency Measurement for Utilities and Transport Regulators by Coelli et. al. (2003) provides an excellent introduction to aggre- gate productivity measurement methods. 6. The Airport Performance Indicators by Jacobs Consultancy was formerly known as TRL Airport Performance Indicators. Efficiency Estimation 133 7. Though information is available for cargo per cargo-dedicated aircraft move- ments for most of the LAC region, regional averages for North America, Europe, and Asia Pacific were not available, nor were data for Brazil’s air- ports, including Viracopos-Campinas International (VCP). Without this information, a graphical representation of this indicator would not have been sufficient to represent the region or provide a basis for global com- parison. 8. The questionnaire developed for this report asked the operator to provide information about staff directly employed by the airport operator and total employees in the airport. Most operators provided data for the for- mer category of staff but provided virtually none for the latter. The same problem with staff information was reported by ATRS for airports in Europe, North America, and Asia. It is difficult for airport operators to have information about staff employed by companies in charge of out- sourced services. 9. Even though our data set does not contain labor (staff) data before and after concessions took place, anecdotal evidence suggests that when operation and management are transferred to the private sector, airports tend to increase outsourcing through service contracts, mainly for security and cleaning. It is interesting to note that from figure 3.3, it is not possible to assert that pri- vately managed airports in LAC show higher values for the ratio of passengers per employee. 10. For this partial performance indicator, the ATRS report split the region East Asia and Pacific in two: Asia and Australia–New Zealand. 11. The use of buses instead of boarding bridges is particularly uncomfortable for handicapped passengers. 12. Most of the literature uses before and after time comparisons to evaluate the impact of privatization, even though the ideal strategy would be to compare utilities under private operation with publicly operated utilities sharing simi- lar characteristics. The reason for most researchers’ selection of the before and after methodology lies in the difficulty of identifying comparable firms (firms with identical characteristics) operating under different ownership regimes. Even a comparison using before and after scenarios for a given firm is a difficult exercise to carry out due to lack of available and reliable data. 13. That is, nominal prices were adjusted by local inflation and expressed in 2005 constant prices and then converted to U.S. dollars using the average value of the exchange rate in 2005. 14. Time series of the partial performance indicators for those airports not pre- sented in this section are available upon request. The selection of airports was somewhat arbitrary but was made looking for a balance among size, owner- ship, and country coverage. 134 Airport Economics in Latin America and the Caribbean 15. The nonexpert readers interested in efficiency estimation are encouraged to read A Primer on Efficiency Measurement for Utilities and Transport Regulators, Coelli et al. (2003). 16. Constant returns to scale implies that when all production inputs are increased by 10 percent, the output increases by 10 percent. When DEA is used and CRS technology is used, it is assumed that all airports operate under constant returns to scale. Variable returns to scale, on the other hand, calculates techni- cal efficiency and isolates the scale component (that is, it allows identification of whether an airport is inefficient because it operates at a scale other than constant returns to scale). 17. A balanced sample in this context means we have observations for years 2005 and 2006 for all airports. 18. By construction, TE under VRS multiplied by scale efficiency (SE) equals TE under CRS. Table C.3 in appendix C replicates table 3.5 but adds the results of computing average TE scores using a model with two inputs (leaving in runways and staff and removing boarding bridges). Investment in boarding bridges shows a significant underinvestment in LAC (569,000 passengers per boarding bridge, compared with 359,000, 284,000, and 305,000 in Asia, Europe, and North America, respectively), and given that DEA cannot mea- sure quality of service, it tends to reward airports that underinvest in capital. When removing boarding bridges from the calculation, the average TE score for LAC airports falls significantly relative to the average in other regions. 19. Doganis (1992) found that airports experience significant increasing returns to scale up to 1 million passengers and that unit costs continue to decline up to 3 million passengers, but that they level off thereafter. Our estimates indi- cate that constant returns to scale are reached at a much higher volume of passengers. 20. It could be argued that airports can close a runway or terminal, but this is usually not the optimal strategy. It is better for airports to maintain assets in proper condition rather than abandon them and then invest in reha- bilitation. 21. Table C.1 in appendix C presents detailed results for the calculation of TE scores for all airports other than those in Latin America. 22. Peer airports are the equivalent of points R, T, and S in figure 3.21. 23. We estimate truncated regressions using the “truncreg� procedure of STATA 9.0. 24. Given that our data set contained a lot of airports in the United States, and given the availability of data for these airports, we used GDP per capita of the state in which each airport is located instead of GDP per capita for the coun- try as a whole. Efficiency Estimation 135 25. The only criterion used to split the data was to obtain three subperiods with an equivalent number of years. The sample covers a large range of airport sizes. Measuring size by the number of passengers per year, the sample ranges from 158,000 to 25.8 million passengers. Zero values are reported for some variables. On the output side, this is the case for freight transportation for at least one airport. On the input side, at least one airport was still not equipped with boarding bridges in 2007. 26. In order to avoid potential biases due to the presence of an unbalanced panel, Malmquist index computations were performed separately for each two-year sequential period using in each case a balanced panel of airports. 27. The TFPC index values reported in tables 3.11–3.13 exclude 14 observations (out of a total of 154). The excluded observations correspond to airports that introduced major changes in their capital stock in a particular year. These changes, given by increases in either the number of runways or boarding bridges, are reported in a given year and thus represent a significant discrete change in the inputs of production (the moment where the investment is ready to use). Given that these types of investments are lumpy by nature, they tend to have a big negative impact on the measures of productivity change (for instance, when one runway is added in year X, it is expected that the quantity of aircraft movements per runway will go down significantly in year X). Appendix C reports the results for all airports and years. In table C.2 values in bold indicate the year of changes in the capital stock of either the number of runways or boarding bridges. In most cases the TFPC index cor- responding to these observations are, as expected, highly negative. References Andrés, L., J. L. Guasch, and A. S. Lopez. 2008. “Regulatory Governance and Sector Performance: Methodology and Evaluation for Electricity Distribution in Latin America.� Policy Research Working Paper 4494, World Bank, Washington, DC. ATRS (Air Transport Research Society). 2008. “2008 Airport Benchmarking Report.� ATRS, Vancouver, Canada. CAA (Civil Aviation Authority). 2000. “The Use of Benchmarking in the Airport Reviews.� Background paper, CAA, London. Coelli, T. 1996. “A Guide to DEAP: A Data Envelopment Analysis (Computer) Program.� Department of Econometrics, University of New England, Armidale, Australia. Coelli, T., A. Estache, S. Perelman, and L. Trujillo. 2003. A Primer on Efficiency Measurement for Utilities and Transport Regulators. Washington, DC: World Bank Institute. 136 Airport Economics in Latin America and the Caribbean Doganis, R. 1992. The Airport Business. New York: Routledge. Färe, R., S. Groskskopf, and C. Lovell. 1994. Production Frontiers. Cambridge, U.K.: Cambridge University Press. Fernandes, E., and R. R. Pacheco. 2002. “Efficient Use of Airport Capacity.� Transportation Research Part A: A Policy and Practice 36: 225–38. Flor, L., and B. de la Torre. 2008. “Medición no paramétrica de eficiencia y pro- ductividad total de los factores: El caso de los aeropuertos regionales de Perú.� Revista de Regulacion en Infraestructura de Transporte 1 (1): 99–114. Gillen, D., and A. Lall. 1997. “Developing Measures of Airport Productivity and Performance: An Application of Data Envelopment Analysis.� Transportation Research 33 (4): 261–73. Gómez-Lobo, A., and A. González. 2008. “The Use of Airport Charges for Funding General Expenditures: The Case of Chile.� Journal of Air Transport Management 14: 308–14. Jacobs Consultancy. 2007. “Airport Performance Indicators 2007.� London. Murillo Melchor, C. 1999. “An Analysis of Technical Efficiency and Productive Change in Spanish Airports Using Malmquist Index.� International Journal of Transport Economics 26: 271–92. Parker, D. 1999. “The Performance of BAA before and after Privatization.� Journal of Transport Economics and Policy 33 (2): 133–46. Pestana Barros, C., and P. U. C. Dieke. 2007. “Performance Evaluation of Italian Airports: A Data Envelopment Analysis.� Journal of Air Transport Management 13: 184–91. Simar, L., and P. W. Wilson. 2007. “Estimation and Inference in Two-Stage, Semi- Parametric Models of Production Processes.� Journal of Econometrics 136 (1): 31–64. CHAPTER 4 Institutional Design and Governance of Airport Regulators in Latin America A wave of structural reform, market liberalization, and privatization swept across Latin America and the Caribbean (LAC) during the 1990s. The airport sector was not spared. By mid-2000, several LAC countries had begun introducing private sector participation in the management of airport services. Yet ownership change was not uniform; different modes and arrangements were adopted. While Argentina opted to con- cession its airport network to a single operator, Chile adopted a case-by- case strategy, and Mexico concessioned its airports by groups (Lipovich 2008). Colombia, Costa Rica, and Peru are among other countries that embarked on reforming their airport sector. Countries that introduced private sector participation in the airport sector had faced the most challenging aspect of the privatization model in Latin America: how to design and implement effective and efficient economic regulation. The debate about the necessary conditions to implement sound regulatory decisions included not only the content of regulatory policies (for instance, tariff methodologies) but also the insti- tutional design of the government authority (as independent commis- sions or government departments). This chapter focuses primarily on the latter. It addresses the realities and challenges of airport regulators from a public sector governance perspective and analyzes the institutional design 137 138 Airport Economics in Latin America and the Caribbean of regulators in terms of their autonomy from authorities formulating policies, the transparency of their procedures, and the quality of their bureaucracy. The analysis does not cover areas related to sector planning, safety, security, licensing of airlines and pilots, and other areas that require a sound regulation; it concentrates only on governance aspects directly related to economic regulation. International practices exemplify two main typologies of airport regulators. The first approach may be characterized by the presence of an independent regulator as the main decision maker in the sector. In the second approach, using primarily competition law, the policy frame- work dispenses with any kind of direct regulation, relying heavily on consultations between an airport and its users. Legal provisions enable competition authorities to control anticompetitive behavior, including the possibility of imposing tighter regulation and price controls when consultations do not prove satisfactory. A third approach, a blend of the two mentioned above, exists in Australia. The Australian Competition and Consumer Commission has broad responsibility for administering competition policy as well as regulation in all sectors with essential facilities. Latin American countries demonstrate a governance design that matches more closely with the first typology of regulators, the indepen- dent regulator. Most countries that concessioned airport services had cre- ated regulatory agencies as their preferred institutional arrangement to enforce concession contracts and the quality of services. In cases where the bulk of airport services remained as state owned, the role of regulator was placed in the hands of government departments with limited indepen- dence from sector authorities. The region also demonstrated the presence of independent regulators in the context of state-owned enterprises. This is the case of Brazil, where the national airport administrator (INFRAERO) is regulated by the National Civil Aviation Agency (ANAC). Independent regulatory agencies in the airport sector, as in other infra- structure industries, were given the highest levels of administrative and legal independence and subject to accountability before the congress. Their decision-making authority was placed within a board of directors, which would be composed of technical and nonpolitical members. The agencies were also given significant regulatory competencies to determine tariffs and minimum requirements for quality of service. In countries where airports remained publicly owned, regulatory func- tions were kept in the hands of nonindependent agencies (usually under Institutional Design and Governance of Airport Regulators in Latin America 139 the name of administraciones aeronauticas). These institutions, sometimes having a separate status from the government, possess overall policy implementation responsibilities, although decisions are made by policy formulators such as the line ministry. In this chapter, institutional attributes of independent regulatory agen- cies (IRAs) are compared with nonindependent regulatory agencies or government departments (non-IRAs) in the airport sector. The goal is to identify under which arrangement regulatory governance can be enhanced. Moreover, the multidimensional approach of this chapter allows the dis- entanglement of different aspects of regulatory governance to test their individual strengths in both IRAs and non-IRAs. Literature Review The literature on independent regulators in Latin America has mainly focused on the electricity and water sectors, where the highest number of IRAs have been set up. This section reviews this literature as it contrib- uted to the development of the methodology used in this study because no previous paper has conducted a cross-country assessment of the gov- ernance structure of airport regulators in Latin America. Literature on IRAs in the LAC region has adopted two main approaches. The first approach has been quantitative, establishing correlations between different indexes of agencies’ autonomy, transparency, and accountability, and sector performance indicators, such as coverage, quality, and labor productivity. The second approach has been qualitative, making use of institutional mapping and benchmarking techniques to assess the presence of several institutional attributes in regulatory agencies. Studies addressing IRAs in a more comprehensive manner are limited. Andrés et al. (2008), an example of the first approach, explore the cor- relations between different measures of the governance of regulatory agen- cies and sector indicators (company level) in the electricity sector of the LAC region. Through principal component analysis (PCA), they develop different indexes of agencies’ governance, establishing links between utilities’ performance and the existence of a regulatory agency, the expe- rience of the regulatory agency (given by the years of the regulators since establishment), and the governance levels of regulatory agencies. (An aggregated index of governance in regulatory agencies produced weight- ing of several dimensions of governance, including autonomy, transpar- ency, accountability, and administrative capacities.) They find a positive 140 Airport Economics in Latin America and the Caribbean and significant correlation between the three measures of existence, experience, and governance levels and sector performance. Along the same lines as Andrés et al. 2008, Gutierrez (2003) estimates the impact of regulation on telecom outcomes in 22 Latin American and Caribbean countries. Gutierrez measures regulation through a regulatory governance index composed of different formal characteristics of regula- tory agencies as well as patterns of the regulatory framework. Gutierrez finds a positive correlation in sector performance when associating the aggregated index or its separate components. Estache and Rossi (2008) find that the introduction of regulatory agencies in electricity distribution in developing countries is associated with more efficient firms and with higher social welfare. The second approach could be defined as qualitative. It makes use of rather descriptive and normative types of analysis, focusing on certain attributes of independent regulatory agencies such as their autonomy, the transparency of their procedures, and their accountability to both institu- tional and noninstitutional actors. A common research design has been to construct different indexes to benchmark IRAs in infrastructure sectors. Gutierrez (2003) develops a Regulatory Framework Index (RFI) to assess the evolution of regulatory governance in the telecommunications sector during the period 1980–2001 in 25 LAC countries. The index, an aggre- gated measure of formal legal and institutional attributes of IRAs in telecommunications, ranks agencies in the region on their performance in each component. Andrés et al. (2007) benchmark IRAs of LAC countries in electricity distribution. Through different indexes that combine formal and informal attributes of the governance of regulatory agencies, the study compares 19 national electricity regulators of the region. An aggregated index of regulatory governance (Electricity Regulatory Governance Index) and 16 indexes of formal and informal aspects of regulatory agencies related to their autonomy, transparency, and accountability are developed. The paper makes an interesting distinction between different dimensions of each of the variables. For instance, the variable autonomy is analyzed in terms of its political, regulatory, and managerial dimensions. Results show several shortcomings in the implementation of the independent regula- tory model in electricity, with autonomy as the variable with the lowest value among all indexes that aim at measuring regulatory governance. Correa et al. (2006) provide a detailed analysis of Brazilian regulatory agencies in different infrastructure sectors (six federal and 15 state regu- latory agencies in electricity, natural gas, water and sanitation, ground Institutional Design and Governance of Airport Regulators in Latin America 141 transportation, petroleum, railroads, and telecommunications). Agencies’ governance is measured through three indexes. The first index, Regulatory Governance Index, is the baseline indicator and represents the most com- prehensive data of all the indexes. The second index, the Parsimonious Index, captures those variables of the survey that are less subjective. The third index, the Facto Index, is related to actual practices of regulatory agencies. The report finds that the independence and accountability attri- butes are more developed than regulatory means and instruments (par- ticularly qualified personnel and regulatory tools) and decision-making procedures (particularly with respect to those mechanisms that can guar- antee consistency of decisions and reduce arbitrariness). It also finds a clear partition between federal and state regulatory agencies, with the former achieving higher results in the autonomy, decision-making, and decision tools components of the Regulatory Governance Index. Despite the significant developments in understanding the role of IRAs in the performance of utilities, the research on the political and governance aspects of regulation in the LAC region remains limited. The main gap in the existing literature is the low explanatory power. Few, if any, attempts have been made to address issues of causality, sequencing, and complex interaction effects that contribute to a better explanation of IRAs in policy making. Methodology and Data Sources This chapter combines both qualitative and quantitative techniques, with emphasis on the former. Qualitative comparative analysis is used to describe the design and practices of airport regulatory agencies. The frame- work of analysis focuses on four main aspects of the governance of airport regulators: the autonomy of the decision-making process, the transparency of the design, the accountability of the design, and the quality of the bureaucracy (table 4.1). The analysis includes both IRAs and non-IRAs. Our analysis focuses on the institutional design of airport regulators, omitting indicators related to actual effectiveness. Thus, the reader should be aware that when, for instance, autonomy is measured, a degree of factual independence is not automatically attributed to the agency. Institutional design refers to the inputs or characteristics of IRAs and government departments that would allow them to be more autonomous and accountable. Nevertheless, even when the institutional design incor- porates the best possible attributes, that does not guarantee either effec- tive autonomy or accountability. 142 Table 4.1 Aspects of Governance of Airport Regulators Autonomy of Quality of Aspects decision making Transparency Accountability bureaucracy - Regulatory powers - Civic engagement in rule - Appeals of - Structure of staff (e.g., tariffs, quality making agency’s decisions positions within the of service) - Consultations - Effects of consultations agency - Status of agency - Publication of agency’s - Evaluation of agency’s - Educational levels of - Procedures to appoint/ decisions performance agency’s staff Components remove board - E-government - Accountability - Publication of vacancies members - Registry of board instrument - Budget sources meetings and decisions - Performance instrument - Publication of job vacancies Source: Author’s elaboration. Institutional Design and Governance of Airport Regulators in Latin America 143 A study that addresses effectiveness has to rely on in-depth case stud- ies. At this stage, considering the gap in the literature on the subject, the analysis of this chapter is focused on identifying typologies and institu- tional design patterns at the regional level. The ultimate goal is then to compare regulators in terms of their institutional design. The hope is that the findings of this chapter will lead to further research aimed at estab- lishing correlations between the performance of the sector and the design and practices of regulators. Data were collected through a survey (a copy can be found in appen- dix B) submitted to 24 airport regulators in the LAC region. Both IRAs and non-IRAs were included. The survey was the result of a thorough consultation process and literature review; its framework builds on simi- lar surveys carried out by Andrés et al. (2007). Final respondents include 13 regulators, four of them IRAs and nine nonindependent regulators. Questions cover each of the dimensions of governance summarized in table 4.1 as well as several questions on economic regulation. Measures of autonomy, transparency, accountability, and bureaucratic quality in airport regulators were created by assigning values, between 0 (worse) and 1 (best), to different indicators. These measures are aimed only at providing the reader with a quantitative approximation of the governance structure of airport regulators; they are not indexes or tools of benchmarking. The methodology section of this chapter describes each indicator and the criteria used in assigning values. Table 4.2 maps each regulator, country, and legal configuration (IRA or non-IRA). Regulatory Governance Autonomy of Decision Making The discussion around the autonomy of independent agencies has absorbed most of the space dedicated by policy analysts to the subject of regulatory authorities. Although this debate has been especially present in developing countries, it is also a current subject in developed govern- ments (OECD 2002). Independent commissions were the result of an agreement between President Franklin D. Roosevelt and both the judiciary and the Congress of the United States in 1930. The occurrence of the market crisis at that time convinced President Roosevelt of the need for stronger regulation in the economy. His response, creating public bodies in charge of establish- ing these standards, resulted in opposition from the other two branches of government, which were reticent to give the executive branch full 144 Airport Economics in Latin America and the Caribbean Table 4.2 Mapping of Regulator and Legal Configuration Regulator Country IRA/Non-IRA Organismo Regulador del Sistema Argentina IRA Nacional de Aeropuertos Department of Civil Aviation Bahamas, The Non-IRA Superintendencia de Transportes Bolivia IRA Agência Nacional de Aviação Civil Brazil IRA Dirección de Aeropuertos, Chile Non-IRA Ministerio de Obras Públicas Unidad Administrativa Especial de Colombia Non-IRA Aeronáutica Civil Dirección General de Aviación Civil Costa Rica Non-IRA Comisión Aeroportuaria Dominican Non-IRA Republic Dirección General de Aviación Civil Ecuador Non-IRA Autoridad de Aviación Civil El Salvador Non-IRA Dirección General de Aviación Civil Guatemala Non-IRA Autoridad Aeronáutica Civil Panama Non-IRA Organismo Supervisor de la Peru IRA Inversión en Infraestructura de Transporte de Uso Público Source: Author’s compilation. Note: IRA = independent regulatory agency. powers to intervene in the economy. The agreed–upon framework involved creating commissions, in which part of their public administra- tion structure would be subject to the accountability of the Congress. Their main decision-making body would be a board of directors com- posed of members appointed by the president with the agreement of Congress. This was precisely the origin of the term independent, that is, commissions that would be independent from the executive branch and subject to the accountability of Congress. These agencies were created to regulate intrastate trade, communica- tions, energy, and transport. Nonetheless, in the 1980s under the Reagan administration, and in a different policy context, they were used to imple- ment a vast economic deregulation process. The independence and administrative configuration of these agen- cies have been discussed since their creation. Nowadays, the prevailing opinion is that independent commissions have an institutional design that is only explained historically but not functionally. Moreover, differ- ent specialists in the matter have suggested the transformation of IRAs into single decision-making bodies, with their members subject to the same stability of the board of directors and their adjudication powers Institutional Design and Governance of Airport Regulators in Latin America 145 assigned to a special tribunal within the same agency (Verkuil 1988). Moreover, in the United States, the introduction of the regulatory review process by the Office of Information and Regulatory Affairs has added one more reason to believe that the idea of absolute indepen- dence of IRAs is an illusion. The discussion around the autonomy of regulatory agencies in Latin America has become abstract in some countries and increasingly relevant in others. It has become abstract in countries where regulatory agencies were affected by political discretion, having little influence in policy mak- ing. Yet it has become relevant in countries that have made significant progress in their regulatory institutions. In the latter, the emergence of professionalized and more influential agencies has generated policy debates around the equilibrium between policy formulators and regula- tory agencies, the oversight of regulation, and the mechanisms to guaran- tee autonomy (that is, the discussions in Brazil around the accountability of regulatory agencies to sector ministries). This study defines autonomy in terms of four main patterns of decision makers: (a) the composition and appointment of the authorities, (b) budget independence, (c) the procedure to appoint and remove the main decision makers, and (d) the reasons by which decision makers can be removed. Our aggregate measure of autonomy finds IRAs in a better position than government departments. This is not an indication of actual levels of autonomy but of the inputs agencies would need to perform in an envi- ronment where decisions can be made with reasonable levels of transpar- ency and independence. Figure 4.1 shows a clear advantage of IRAs versus non-IRAs in this regard. Nevertheless, the disaggregation of our measure of autonomy in different variables shows advantages and disad- vantages for both IRAs and non-IRAs. While an independent regulator provides more guarantees in terms of the meritocracy with which author- ities are appointed and removed, not yet clear are the advantages of IRAs vis-à-vis non-IRAs in terms of their regulatory powers and the stability of their decision-making authorities. IRAs have more independent budget- ary sources than government departments, although non-IRAs have sig- nificant contributions from inspection taxes (a parallel form of the regulatory tax that IRAs charge to private operators). Composition and appointment of decision-making authorities. A critical aspect of the autonomy of any government body, with or without inde- pendent status, is the way the top authorities are appointed. In fact, the creation of IRAs in LAC countries was intended to cut the cycle of 146 Airport Economics in Latin America and the Caribbean Figure 4.1 Decision-Making Autonomy a. Autonomy of decision making in IRAs 1 0.8 index (0–1) 0.6 0.4 0.2 0 Peru Brazil Bolivia Argentina b. Autonomy of decision making in non-IRAs 1 0.8 index (0–1) 0.6 0.4 0.2 0 a r e r a ica ic ile do o bi al Th bl ad Ch aR em m ua pu , v as lo st al Ec at Re m Co S Co Gu ha El n ica Ba in m Do Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). political appointees and discretional decision making in the infrastructure sectors. The initial configuration of independent regulators included the appointment of a board of directors with sufficient proficiency in eco- nomic regulation and a reasonable level of political independence. Unfortunately, this trend was reversed in the majority of the cases. Independent regulators are in 50 percent of the cases appointed by the executive with different levels of intervention of the parliament. The involvement of the parliament is generally seen as positive, especially because it allows the participation of other stakeholders (especially the opposition parties) in the selection of directors to the board. The Institutional Design and Governance of Airport Regulators in Latin America 147 involvement of the executive (both through the president and the line minister) explains 70 percent of appointments in non-IRAs. Interestingly, in 30 percent of the cases, special agencies and trade associations linked to the airport sector also appoint representatives to the board. This last aspect of non-IRAs could be considered a positive development, consid- ering the nontransparent norm of leaving the appointment of decision makers entirely to elected officials (figure 4.2). Figure 4.2 Appointment Authorities a. Appointment authorities in IRAs president and ministers president 25% 25% president and congress 50% b. Appointment authorities in non-IRAs involvement of technical bodies and line minister trade associations 29% 29% president 42% Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). 148 Airport Economics in Latin America and the Caribbean Budget independence. The composition of the budget is a critical aspect of the regulator’s independence, perhaps its most salient characteristic (figure 4.3). Inspired by international best practices (Latifulhayat 2008), regulatory agencies of the region established a regulatory tax that would be charged to service providers. This source of funding would prevent the agency from being fully dependent on government support. In the case of airport regulators, the majority of IRAs finance their bud- get with a tax or fee charged to service providers. The only exception is ANAC in Brazil, which receives 30 percent of funds from the central gov- ernment. In accordance with the literature, we assume that agencies with budget autonomy have more freedom and flexibility to design programs and monitor operators. In fact, our measure of autonomy gives higher values to agencies whose budget is integrated with taxes or fees charged either to service providers (majority of IRAs) or passengers (some govern- ment departments). Non-IRAs enjoy a combination of sources. Although in the majority of cases they receive government support, they also integrate their budgets with different taxes charged to passengers and airlines. Non-IRAs with autonomous funding present an alternative scheme to an institutional design of an independent regulator. So far, the literature in the LAC region has not yet addressed the benefits of these arrangements in the context of regulatory agencies. Procedure to appoint and remove the main decision makers. Similar in relevance to the appointment of decision-making authorities is the way they can be removed from office (figure 4.4). In the case of IRAs, the legal statute requires a justified cause to proceed with the removal. In the cases of government departments, public servants may be removed under the sole discretion of the line minister or the president. Reasons decision makers leave their positions. A complementary aspect related to the procedure to dismiss directors is the actual reasons they leave their positions (figure 4.5). Agencies were given four options: end of mandate, voluntary leave, external pressure, or retirement. Our mea- sure of autonomy gives higher values to the first option. We would expect that a director that ends her or his mandate is the most desirable situation for the independence of regulators. We could also assume the same for a director that leaves the agency based on her or his own will. Nonetheless, voluntary leave may also reflect a disagreement with policy formulators (line minister or president) or undue pressures from the same actors. Institutional Design and Governance of Airport Regulators in Latin America 149 Figure 4.3 Budget Composition a. Budget composition in IRAs 100 80 60 percent 40 20 0 Argentina Bolivia Brazil Peru government budget registration tax b. Budget composition in non-IRAs 100 80 60 percent 40 20 0 Colombia Ecuador Guatemala Panama Bahamas, The government budget inspection tax Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). 150 Airport Economics in Latin America and the Caribbean Figure 4.4 Procedure to Remove Decision Makers non-IRAs, 33% IRAs, 67% Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). Note: Directors can only be removed under a bad performance cause. End of mandate explains about 30 percent of the reasons why regula- tors leave both agencies and government departments. This percentage is higher in non-IRAs (35 percent) than in IRAs (29 percent). The main difference between IRAs and non-IRAs with regard to why decision makers leave is seen in voluntary leave. Directors of IRAs leave office voluntarily in 57 percent of the cases, while in government depart- ments this number is 29 percent. On the other hand, dismissal explains around 40 percent of the cases in government departments, while it is only 14 percent in the case of IRAs. The previous numbers show contradictory results. The evidence shows that the likelihood of directors leaving voluntarily in a regulatory agency is higher than that for a civil servant in a government department. While this is consistent with the flexibility of a private law regime in regulatory agencies (they are hired under private law in most of the cases), it is not a positive sign for the stability of regulatory policies. It can even show that in practice, influence by the executive over directors is high, and they react by leaving the IRA. Criteria to appoint authorities. In the context of highly volatile political environments and undue influence, meritocracy emerges as a critical fac- tor in making regulatory decisions that are sound and transparent. The survey asked regulators to identify the criteria under which top decision makers are appointed. Institutional Design and Governance of Airport Regulators in Latin America 151 Figure 4.5 Reasons Directors Leave Positions a. Reasons directors leave positions in IRAS dismissal 14% end of mandate voluntary leave 29% 57% b. Reasons directors leave positions in non-IRAs dismissal voluntary leave 36% 29% end of mandate 35% Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). Our sample shows that in the cases of non-IRAs, requirements to be appointed as directors are soft. In the cases of IRAs, statutes require tech- nical expertise in the appointment of decision-maker authorities. In only 25 percent of the cases, IRAs do not require any criteria for appointment. In the case of government departments, this number reaches 55 percent. Regulatory autonomy. Another aspect included in the autonomy dimen- sion of governance is the power of the agency vis-à-vis the government, the airport operator, and other institutions to set tariffs, quality of service 152 Airport Economics in Latin America and the Caribbean standards, and other regulatory competencies. Surprisingly, results are similar for both IRAs and government departments in the LAC region’s airport sector. Quality of Bureaucracy It is usually argued that one of the contributions of regulatory agencies to policy making is technical rationality (Thatcher 2007). An agency com- posed of directors appointed under meritocratic criteria and well-paid officials would constitute relevant factors to insulate it from politics and improve decision making. In this section we focus on the bureaucracy of airport regulators. We define bureaucratic quality in airport regulators in terms of three main aspects: (a) educational levels of the regulator’s staff, (b) the flexibility and powers of the agency to decide its own human resources policies, and (c) the publication of the agency’s vacancies. Our definition of bureau- cracy excludes directors to the board. Agency staff was defined in terms of three main categories: managers, technical workers, and administrative employees. In general, responses for educational levels were low com- pared to other questions. In our measure of bureaucratic quality, IRAs present better scores than government departments (figures 4.6 and 4.7). This is reflected not only in the educational levels of the staff but also in the way vacancies are advertised and promoted. Figure 4.6 Bureaucratic Quality 1 0.8 index (0–1) 0.6 0.4 0.2 0 graduate degrees college degrees among graduate degrees publication of among managers administratiave staff among technical staff vacancies IRA non-IRA Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). Institutional Design and Governance of Airport Regulators in Latin America 153 Figure 4.7 Bureaucratic Quality by Type a. Bureaucratic quality in IRAs 1 0.8 index (0–1) 0.6 0.4 0.2 0 Peru Bolivia Argentina b. Bureaucratic quality in non-IRAs 1 0.8 index (0–1) 0.6 0.4 0.2 0 Ecuador Bahamas, The Dominican Guatemala Republic Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). At the level of managers, IRAs present better results than government departments. On average, the majority of managers in IRAs have graduate degrees. Technical employees present a different landscape. Government departments show better results for graduate education than IRAs. Nevertheless, technical employees with college degrees have a higher incidence in IRAs than in government departments. IRAs also present better results among administrative employees. When measured by those with a college degree, IRAs show, on average, a higher percentage of employees with this background. 154 Airport Economics in Latin America and the Caribbean Another important aspect of a high-quality bureaucracy is the way staff of the agency is selected. Regulators were given four options: (1) no publication, (2) publication on the agency’s website, (3) publication in a newspaper, and (4) both options 2 and 3. We assigned higher values to those agencies that use both websites and newspapers to publish job vacancies. In this measure, IRAs also show, on average, more transparent human resource policies than government departments. Transparency of Decision Making The establishment of IRAs in the regulation of the infrastructure sector has been considered as a way of opening the regulatory process to affected parties. Attached to decision making in regulation was the development of different instances of consultations with both providers and consumers. Arrangements to promote and advocate consumers’ rights include users’ councils within the agency, consumer organiza- tions, and consultations. The a priori expectation is that these tools are more likely to be used in the context of an IRA than in a government department. In this section we compare practices of transparency in regulatory agencies and government departments. Our measure of transparency focuses on five main aspects: public consultations, legal effect of consulta- tions, publication of the agency’s decisions, publication of vacancies, and registration of board meetings. Overall, IRAs offer a better framework for more transparent regulatory policies than government departments (figures 4.8 and 4.9). On average, IRAs achieve better results than government departments in most of the dimensions. The regulatory agency model seems to provide a more suitable space for the involvement of consumers and other stake- holders in rule making and consultations. Consultations seem to have, according to the responses obtained, a larger influence in IRAs than in government departments. ANAC in Brazil is the agency with the largest number of consultations. Since its establishment in 2006, the agency has conducted 12 consulta- tion procedures that focused on tariffs, licenses, investment, safety, and consumers’ rights. IRAs in Argentina and Peru, to a lesser extent, also perform consultations. In the case of Peru, consultations have focused mainly on tariff regulation. Several IRAs of the region have established consultative committees as advisory bodies to the board of directors. OSITRAN, the transport Institutional Design and Governance of Airport Regulators in Latin America 155 Figure 4.8 Transparency in Airport Regulators 1 0.8 0.6 index (0–1) 0.4 0.2 0 ta n in tin rd tio of io of nc of ee oa ta s cis n ca n fs gs ns ns s ul ct tio ie de atio va atio rif m fb ns effe ta o ic ic ul rd bl bl ns co pu pu co re co IRAs non-IRAs Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). infrastructure investment regulator in Peru, created a consumer council to deal with demands from different sector stakeholders. ANAC in Brazil established a consultative committee that provides advice to the board of directors. A frequent criticism of these councils is the low levels of involvement of final consumers and the majority presence of service providers. For instance, there is only one association of passen- gers involved in the consumer council of OSITRAN, the majority of representation coming from airlines and other trade associations. Accountability The balance between independence and accountability is one of the most critical issues in the governance of independent agencies (OECD 2005). Politicians have traditionally questioned the independence of agencies headed by nonelected officials. From their perspective, regulatory agen- cies are part of the public administration and, as such, they should be held accountable to the government. Policy responses, in terms of the 156 Airport Economics in Latin America and the Caribbean Figure 4.9 Transparency by Type a. Transparency in IRAs 1 0.8 index (0–1) 0.6 0.4 0.2 0 Brazil Peru Argentina Bolivia b. Transparency in non-IRAs 1 0.8 index (0–1) 0.6 0.4 0.2 0 Costa Rica Ecuador Guatemala Chile Colombia Bahamas, Dominican The Republic Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). accountability of IRAs, range from those that prefer providing agencies with a significant playing field (Australia and the United Kingdom) to those imposing different controls and standards (New Zealand and the United States) (figures 4.10 and 4.11). Our definition of accountability includes both its internal and external dimensions. Measures related to internal accountability are represented by the agency’s staff evaluations. Measures of external accountability include public consultations and the instrument the agency uses to report its performance to external stakeholders. The definition also includes judicial accountability, or the review of the agency’s decision by the courts. Institutional Design and Governance of Airport Regulators in Latin America 157 Figure 4.10 Dimensions of Accountability in Airport Regulators 1 0.8 0.6 index (0–1) 0.4 0.2 0 effects of public evaluation of accountability publication of consultations personnel instrument accountability instrument IRA non-IRA Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). Our measure of accountability shows better results for non-IRAs than for independent regulators. Results may be consistent with the nature of government departments. Contrary to IRAs, government departments are subject to more public sector controls; hence, the monitoring of their decisions may have stronger accountability measures than regulatory agencies. In fact, as previously mentioned, the introduction of IRAs as completely independent entities has made governments question if inde- pendent commissions are the best institutional response to regulating infrastructure sectors. With no exception, all IRAs and non-IRAs prepare an annual report of their performance. In some cases, such as Chile and Colombia, directors must give a presentation on sectoral issues before the congress when asked to do so. The agency’s website is the preferred way to publish annual reports. Economic Regulation The main concern of all economic agents in the airport sector is how tariffs are set. Tariff regulation in airports should be concerned exclusively with those services that have characteristics of natural monopoly and thus 158 Airport Economics in Latin America and the Caribbean Figure 4.11 Dimensions of Accountability in IRAs and Non-IRAs a. Accountability in IRAs 1 0.8 index (0–1) 0.6 0.4 0.2 0 Peru Argentina b. Accountability in non-IRAs 0.8 0.6 index (0–1) 0.4 0.2 0 Ecuador Panama Chile Guatemala Bahamas, The Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). where price regulation is necessary. These services are usually referred to as aeronautical services and include runway use, aircraft parking, air traf- fic control, meteorological services, and passenger handling. There are different approaches to setting regulated tariffs, but they can be grouped under two main methodologies: single till and dual till. Under the first of these methodologies, regulated operational costs related to aeronautical Institutional Design and Governance of Airport Regulators in Latin America 159 services are recovered through revenues generated by aeronautical and nonaeronautical (that is, commercial) activities, with profits from com- mercial activities used to help maintain low levels of regulated aeronauti- cal tariffs. On the other hand, under a dual till model, operational costs related to aeronautical services are exclusively recovered through reve- nues obtained from charging for the provision of aeronautical services (Starkie and Yarrow 2000). With the risk of oversimplifying the differ- ences between the two approaches, the single till method allows the existence of cross-subsidies between airports’ aeronautical and commer- cial activities. Following the example set by the United Kingdom, most airports in Latin America rely (explicitly in a few cases and implicitly in most) on the single till approach. Table 4.3 shows that six countries reported set- ting tariffs following a single till model. However, the answers provided by regulatory agencies regarding this issue are contradictory. For instance in Argentina, Brazil, El Salvador, and Guatemala, regulatory agencies claim that their tariff-setting mechanism responds to the single till model. However, in a separate question, these four regulatory agencies claim that the costs associated with the provision of aeronautical services are fully recovered through aeronautical tariffs. To be able to set tariffs, regulatory agencies should develop and main- tain a comprehensive economic and financial model. If an agency does not have an economic and financial model, it will have only a partial understanding of the performance of the regulated airports. To feed an economic and financial model, regulators develop regulatory accounting manuals specifying the information that operators need to submit to regulatory agencies. Within our sample, only 8 of the 14 regulatory agen- cies that answered the questionnaire stated that they employ an eco- nomic and financial model. However, it is not easy to assess if the economic and financial models are being used in practice to regulate airport operators. Other responses indicate that of the eight agencies that rely on an economic and financial model, only four use a regulatory accounting manual. The regulator needs to make sure that the operator reaches an eco- nomic and financial equilibrium. In other words, it needs to make sure the operator’s internal rate of return (IRR) is equal to the weighted aver- age cost of capital (WACC). If the WACC is higher than the IRR, the operator will exit the market. Only 5 of the agencies that responded to the questionnaire estimate the WACC faced by the operators (and only 3 of those 5 perform these estimations on a regular basis). Finally, only 5 160 Table 4.3 Answers to Selected Questions on Economic Regulation in the Airport Sector Are aeronautical Does Was agency services’ costs Does agency Does agency agency created before recovered Are commercial use an conduct regulate introduction Single till, through tariffs subject economic and Manual of Does agency economic and private of private dual till, or aeronautical to any financial regulatory estimate cost financial Country operators? management? hybrid? tariffs? regulation? model? accounting? of capital? audits? Argentina Y Y Single till Y Y Y Y N Y Bahamas, The Y Y Single till N Y Y Y Y Y Bolivia Y Y Single till N N N N Y N Brazil Y N Single till Y N Y N N N Chile Y Y Dual till Y Y N N Y N Colombia Y Y Single till N N Y Y N Y Costa Rica Y Y Single till/ Y Y Y N N Y hybrid Dominican Y Y Dual till Y Y Y N N N Republic Ecuador Y Y Hybrid Y N N N n.a. N El Salvador n.a. n.a. Single till Y N n.a. N N N Guatemala N – Single till Y Y N N N N Panama N – Dual till Y N Y N Y N Peru Y Y Dual Y N Y Y Y Y Till/hybrid Source: Authors, based on responses to the Governance of Airport Regulators survey (see appendix B). Note: n.a. = answers were missing, incomplete, or unclear. a. Ecuador’s regulatory agency estimates the cost of capital for only those airports that are still operated by the state (nonconcessioned airports). Institutional Design and Governance of Airport Regulators in Latin America 161 out of the 14 agencies frequently conduct economic and financial audits of the airport operators they regulate. Airport regulators are well aware of the trade-off between tariffs and quality of service. When tariffs are regulated (and in particular, when they are subject to a binding price cap), airport operators will naturally tend to lower the quality of service in order to reduce their costs and thus obtain higher profit margins. Against this backdrop, it becomes essential for regulators to introduce the right incentives for operators to increase the quality of the service they provide. Strikingly, only three agencies in our sample (Brazil, Costa Rica, and Peru) responded positively to the question of whether the regulatory framework includes tools to monitor the evolution of quality of service and to design proper economic incen- tives for improvement. The results from our survey suggest that there are serious deficiencies regarding economic regulation in the airport sector in the LAC region. On the one hand, very few of the agencies in charge of enforcing regula- tions have in place the necessary information systems (regulatory accounting manuals, economic and financial models) necessary to per- form their tasks correctly. Even when agencies declared that they have adequate information systems in place, they are not using them to esti- mate the WACC, which is an essential variable for a regulator. In addition, the regulatory frameworks do not seem to provide appropriate incentives for regulators to properly carry out frequent oversight of the quality of services provided by operators. Conclusions Regardless of the existence or nonexistence of a private sector provision, an institutional design identified with an independent regulatory agency appears to provide a better space for good regulatory governance than a government department. Both regional and international experiences show the importance of a government agency that is highly specialized and makes consumers the focus of their policies. Nonetheless, a regula- tory agency is not capable, on its own, to introduce institutional quality into an airport system where policies are ill-designed. But it may, even in an adverse context, enable an adequate representation of stakeholders and a filter to discretional decisions. The division of transparency into different dimensions allowed the identification of several advantages in IRAs versus government depart- ments. Consultations are the most notable of these advantages. The 162 Airport Economics in Latin America and the Caribbean consumer orientation of regulatory agencies versus government depart- ments, whether in the context of state-owned companies or private providers, is a powerful factor to bring stakeholders’ opinions into the decision-making process. Technical expertise is another aspect where IRAs show advantages. Our measure of bureaucratic quality found, on average, higher bureau- cratic quality levels in independent commissions than in government departments. These results are reflected not only in the educational levels of the staff but also in the way vacancies are posted and filled. The most controversial aspect of the governance of IRAs is auton- omy. Our measure of autonomy found, on average, more guarantees of autonomy in IRAs than in non-IRAs. Nevertheless, non-IRAs also show similar regulatory powers and lower turnover rates in their policy makers. Regional experiences provide interesting findings in support of our arguments. The cases of Brazil and Peru are, perhaps, the most illustra- tive. The introduction of regulatory agencies has, in both countries, contributed to more transparent and accountable decision making. These cases are interesting as they present two situations of regulatory agencies in contexts of private sector (Peru) and state (Brazil) provision of the service. A worrisome outcome of the surveys’ analysis was the serious defi- ciency of economic regulation in the airport sector in the LAC region. On the one hand, very few of the agencies in charge of enforcing regu- lations have in place the information systems (regulatory accounting manuals, economic and financial models) necessary to perform their tasks correctly. On the other hand, even when agencies claim to have adequate information systems in place, the vast majority are not using them to estimate the weighted average cost of capital, which is an essential variable for a regulator. In addition, the regulatory frameworks do not seem to provide appropriate incentives for regulators to properly carry out frequent oversight of the quality of services provided by operators. Despite the overall advantage of IRAs for good regulatory governance, conclusions should not be interpreted in a “one model fits all� approach. Rather, they should be used for the purpose of identifying those mecha- nisms that better guarantee open and sound decision making in the regu- lation of airport services. The comparison between IRAs and non-IRAs in airports allowed the disaggregation of governance in different dimensions and the selection of advantages and disadvantages in both models. It is up Institutional Design and Governance of Airport Regulators in Latin America 163 to policy makers to prioritize those aspects that better fit in their institu- tional and policy frameworks. References Andrés, L., J. L. Guasch, M. Diop, and A. S. Lopez. 2007. “Assessing the Governance of Electricity Regulatory Agencies in Latin America and the Caribbean: A Benchmarking Analysis.� Policy Research Working Paper 4380, World Bank, Washington, DC. Andrés, L. A., J. L. Guasch, T. Haven, and V. Foster. 2008. The Impact of Private Sector Participation in Infrastructure: Lights, Shadows and the Road Ahead. Washington, DC: World Bank. Correa, P., C. Pereira, B. Mueller, and M. Melo. 2006. “Regulatory Governance in Infrastructure Industries: Assessment and Measurement of Brazilian Regulators.� World Bank/Public-Private Infrastructure Advisory Facility (PPIAF), Washington, DC. Estache, A., and M. Rossi. 2008. “Regulatory Agencies: Impact on Firm Performance and Social Welfare.� Policy Research Working Paper 4509, World Bank, Washington, DC. Gutierrez, L. H. 2003. “Regulatory Governance in the Latin American Telecommunications Sector.� Utilities Policies 11 (4): 225–40. Latifulhayat, A. 2008. “The Independent Regulatory Body: A New Regulatory Institution in the Privatised Telecommunications Industry.� International Journal of Technology Transfer and Commercialisation 7 (1): 15–33. doi: 10.1504/IJTTC.2008.018800. Levi-Faur, D., and J. Jordana. 2004. “The Rise of the Regulatory State in Latin America: A Study of the Diffusion of Regulatory Reforms Across Countries and Sectors.� Centre on Regulation and Competition, Institute for Development Policy and Management, University of Manchester, U.K. Lipovich, G. A. 2008. “The Privatization of Argentine Airports.� Journal of Air Transport Management 14: 8–15. OECD (Organisation for Economic Co-operation and Development). 2002. “Regulatory Policies in OECD Countries: From Interventionism to Regulatory Governance.� OECD, Paris. ———. 2005. Designing Independent and Accountable Regulatory Authorities for High Quality Regulation. Proceedings of an Expert Meeting in London, United Kingdom, January 10–11. Starkie, D., and G. Yarrow. 2000. “The Single Till Approach to the Price Regulation of Airports.� Civil Aviation Authority, London, U.K. http://www.caa.co.uk/ docs/5/ergdocs/starkieyarrow.pdf. 164 Airport Economics in Latin America and the Caribbean Thatcher, M. 2007. “Regulatory Agencies, the State and Markets: A Franco-British Comparison.� Working Paper RSCAS 2007/17, European University Institute, Florence School of Regulation, Florence, Italy. Verkuil, P. 1988. “The Purpose and Limits of Independent Agencies.� Faculty Publications Paper 1029, College of William and Mary Law School, Williamsburg, VA. CHAPTER 5 Benchmarking of Aeronautical Charges at Latin American Airports Overview A study of tariffs was not envisaged at the time this report was conceived. However, the study of performance indicators and the feedback received from airport regulators, airlines, and other stakeholders consulted during the preparation of this report prompted us to calculate the evolution of airport tariffs and generate a regional benchmark that constitutes the only one publicly available. The tariff benchmarking exercise includes 26 airports within 20 Latin American and Caribbean (LAC) countries (see table 5.1) and for three years: 1995, 2003, and 2009. The selection of years responds to the dual objective of including ample data and measuring the changes in tariff structures and levels as a result of introducing private sector participation in airport infrastructure management. Since most airport concessions in the region took place before 2002, the year 2003 was selected to test the assumption that price increases took place after airports were conces- sioned to the private sector.1 The year 2009 was included to present the latest available information on tariffs at the time this report was written, while 1995 was chosen because in that year no private sector participa- tion policy discussions were held in LAC. 165 166 Airport Economics in Latin America and the Caribbean Table 5.1 Airport Sample Used for the Aeronautical Tariff Benchmarking Analysis Country Airport name 1 Argentina Buenos Aires – Ministro Pistarini International 2 Bahamas, The Nassau – Lynden Pindling International Airport 3 Bolivia La Paz – El Alto International 4 Santa Cruz – Viru Viru International 5 Brazil Rio de Janeiro – Galeão International 6 São Paulo – Guarulhos International 7 Chile Santiago – Comodoro Arturo Merino Benitez International 8 Colombia Bogotá – El Dorado International 9 Cali – Alfonso Bonilla Aragón International 10 Costa Rica San José – Juan Santamaría International 11 Dominican Republic Santo Domingo – Las Américas International 12 Ecuador Quito – Mariscal Sucre International 13 El Salvador San Salvador – Camalapa International Airport 14 Guatemala Guatemala City – La Aurora International 15 Honduras Tegucigalpa – Toncontín International 16 Jamaica Kingston – Norman Manley International 17 Mexico Cancún International 18 Guadalajara – Miguel Hidalgo y Costilla International 19 Mexico City – Benito Juárez International 20 Monterrey – General Mariano Escobedo International 21 Nicaragua Managua – Augusto C. Sandino International 22 Panama Panama – Tocumen International 23 Paraguay Asunción – Silvio Pettirossi International 24 Peru Lima – Jorge Chávez International 25 Uruguay Montevideo – Carrasco International 26 Venezuela, RB Caracas – Simón Bolivar International 27 France Paris – Charles de Gaulle International 28 Germany Frankfurt – am Main International 29 Spain Madrid – Barajas International 30 United Kingdom London – Heathrow International 31 United States Los Angeles – Los Angeles International 32 New York – John F. Kennedy Airport 33 Miami – Miami International The preparatory work for this chapter included extensive research to obtain cross-country comparisons of tariffs, including an exploration of airports’ and regulators’ web pages and specialized publications, as well as consultations with airport and airline international organizations. Research showed that international benchmarking of aeronautical tariffs of LAC airports is not publicly available, but several sources contain Benchmarking of Aeronautical Charges at Latin American Airports 167 limited information for a price. However, in most cases, these sources offer incomplete data for LAC airports, and there is limited tariff data prior to 2005 for a large sample of LAC airports. Even private consulting firms engaged in international benchmarking exercises have scarce infor- mation on LAC airports. Surprisingly, regulators and ministerial departments reported during the preparatory work of this study that they do not carry out frequent benchmarking studies. This is particularly worrisome given that regional tariff benchmarking studies should be a basic instrument for regulators, especially when they need to make informed decisions about tariff changes as part of ordinary tariff review processes and when contract renegotiation with a private operator is required. To provide an international reference to the benchmarking analysis, the following airports were included in the sample: New York (JFK), Los Angeles (LAX), Miami (MIA), Madrid (MAD), Paris (CDG), London (LHR), and Frankfurt (FRA). These European and North American air- ports concentrate most of their Latin American international flights out- side of the LAC region. The tariff benchmarking presented in this section focuses solely on the aircraft-passenger tariff dimension of the charges that aircrafts (airlines) and passengers pay according to established norms and regulations. No attempt is made to analyze the tariff structure between the components (landing fees, aircraft parking, and use of boarding bridges, among other tariffs) and the economic incentives embedded in them. This topic, how- ever, merits further research as it is very relevant for economic regulation of airports and planning of infrastructure investments. Methodology The tariff benchmarking exercise includes an analysis of regulated tariffs that are part of the total turnaround costs established by airports on any given flight. The analysis covers the following regulated charges: landing fees (and night surcharge for lighting), aircraft parking, use of boarding bridges, and passenger charges (passenger facility charges, security, and federal taxes). To make sure the benchmarking analysis is a true cross-comparison, specific assumptions were made regarding the type of aircraft, time spent on the ground, number of passengers on board (or percentage of aircraft seats occupied), and other variables. 168 Airport Economics in Latin America and the Caribbean The aircraft selected for comparison are consistent with the types of fleets most commonly found in the LAC region in 2009. The Airbus A320 is the most popular aircraft in LAC, serving short- and medium- haul routes. The Boeing 767-300 is the most widely used aircraft for medium- and long-haul destinations. The parameters and assumptions used are summarized in table 5.2. While charges were estimated and compared for both types of aircraft, detailed calculations of aeronautical tariffs are presented only for the Airbus A320 to avoid producing a very long and repetitive chapter. Summary graphs, including total turnaround costs, are presented for the Boeing 767-300. Information to carry out the benchmarking exercise was obtained from two main sources: the International Air Transport Association’s (IATA) Airport and Air Navigation Charges Manual and countries’ Aeronautical Information Publications (AIP 1993, 2003, 2009). The for- mer contains detailed information on regulated charges for about 300 airports worldwide, which is updated on a regular basis.2 For this study, the information provided by IATA was crosschecked with tariffs pub- lished in the AIPs of the countries included in the sample. In cases of discrepancies between information provided by IATA and the AIPs, the data used correspond to the AIP data sets. Tariffs are measured in real terms (in 2009 U.S. dollars). Nominal prices were converted to real prices according to the United States Consumer Price Index (CPI), as reported by the International Monetary Fund’s (IMF) World Economic Outlook database. The U.S. CPI was selected instead of each country’s CPI because the price adjustment clauses of airport privatization contracts in the region generally use the U.S. CPI as the benchmark. One important assumption in the analysis was the type of flight: all flights considered for this analysis correspond to international flights. Some countries discriminate airport tariffs according to the origin and destination of flights. In some countries, domestic flights enjoy lower Table 5.2 Key Parameters of the Aircraft Used in the Analysis Parameter Airbus A320 Boeing 767-300 Fuselage Narrow-body Wide-body Range Short, medium Medium, long Maximum take-off weight (MTOW) 77 tons 187 tons Seating capacity 150 seats 269 seats Assumed load factor 71% 71% Source: Author’s estimation with data from Airbus S.A.S. and Boeing Commercial Airplanes. Benchmarking of Aeronautical Charges at Latin American Airports 169 regulated charges despite the fact that the aircraft is the same and demands the same infrastructure service as a flight with an international destination. Thus, the “international� price was used at airports that set different rates for international and domestic operations. Landing Charge and Lighting Surcharge Landing fees at every airport in the sample are based on the aircraft maximum takeoff weight (MTOW), which is a standard manufacturer feature of the aircraft. In 2009, rates in LAC ranged from US$0.4 to US$13.30 per ton. However, calculation methods differ greatly from airport to airport. Some airports charge per ton of MTOW, some include a fixed-amount component, and some employ weight bands to establish the unit rate per ton. Given the variety of approaches and to ensure a true cross-comparison, landing fees were compared for the cost of land- ing an aircraft of the same characteristics (for an Airbus A320 and Boeing 767–300, with parameters listed in table 5.2). Using this method, the calculation of the landing fees included every aspect of the tariff struc- ture, such as fixed amounts, minimums, and weight bands. In other words, it compares how much the same aircraft pays at each one of the different airports. Some airports, especially those in Europe, levy a noise charge that var- ies according to the aircraft noise category (which is usually defined by the airport). Since these charges were created either to avoid (through the operation of quieter aircraft) or to penalize the generation of noise during landing and takeoff, they were included in the landing fees calculation. Half the airports in the sample also feature a lighting surcharge (or night surcharge) for operations taking place during the night hours. The following graphs show the landing fees (total charge) for an Airbus A320 during day hours (figure 5.1) and during night hours (figure 5.2). According to the sample, average landing charges for an Airbus A320 on a daylight operation is US$298 in LAC, and US$591 for the European and U.S. airports in our sample. Quito has the highest landing fees in LAC, at US$781, while Santo Domingo has the lowest, with total landing charges of US$54. For landing fees during night operations, Quito and Santo Domingo are also the most and least expensive airports in LAC, as landing fees total US$1,014 and US$54, respectively, for an Airbus A320. The average charge for an Airbus A320 on a night operation in LAC is US$330 and rises to US$663 in Europe and the United States. Landing fees have decreased in real terms between 1995 and 2009 at most airports in Latin America and the Caribbean (figure 5.3). Figure 5.1 Landing Fees for an Airbus A320, Daylight Operation UIO VVI LPD CCS LIM GIG GRU MVD GIG KIN BOG SCL LAC sample average EZE MGA SAL MEX TGU PTY CUN GDL NAS MTY SJO ASU GUA SDQ JFK CDG MAD LAX FRA MIA 0 200 400 600 800 1,000 1,200 US$ (2009) Source: Author’s elaboration based on information from IATA (2009), Aeronautical Information Publication (AIP) Co- lombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infraestructura Aeronaútica (DINACIA) Uruguay. Note: Buenos Aires assumes that payments are made on time; San Jose includes airside infrastructure charge; Santo Domingo assumes landing during off-peak hours; Guatemala includes ramp fees; Mexico assumes tariff A (off-peak); Los Angeles assumes the airline is a signatory carrier; New York assumes landing during off-peak hours. For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 171 Figure 5.2 Landing Fees for an Airbus A320, Night Operation UIO CCS VVI LIM LPD MVD GK GRU SCL CLO EZE BOG KIN LAC sample average MGA TGU SAL MEX PTY CUN GDL NAS SJO MTY ASU GUA SDQ CDG JFK MAD LAX FRA MIA 0 200 400 600 800 1,000 1,200 US$ (2009) Source: Author’s elaboration based on information from IATA (2009), AIP Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: See note for figure 5.1. 172 Airport Economics in Latin America and the Caribbean Figure 5.3 Changes in Landing Fees for an Airbus A320, Daylight Operation UIO VVI LPD CCS LIM GIG GRU MVD VLO KIN BOG SCL EZE MGA SAL MEX TGU PTY CUN GDL NAS MTY SJO ASU GUA SDQ 0 100 200 300 400 500 600 700 800 900 1995 2003 2009 Source: Author’s elaboration based on information from IATA (1995, 2003, and 2009), AIP Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 173 In the sample, nominal landing fees increased on average by 4.7 per- cent between 19953 and 2003 and remained constant between 2003 and 2009. However, if the inflationary effect is incorporated, landing fees in real terms declined in the majority of airports in the region from 1995 to 2009. As seen in figure 5.3, landing fees measured in constant dollars decreased at 10 out of 13 airports between 1995 and 2003 and at 11 of those airports between 1995 and 2009. When only the 2003–09 period is considered, there was a reduction in real landing fees at 24 of the 26 airports in the sample (figure 5.4). On average, landing fees fell by 14 percent in real terms between 1995 and 2003, and by 10 percent between 2003 and 2009. San José is clearly the airport with the most significant price escalation (167 per- cent between 2003 and 2009), although, as significant as the increase seems in relative terms, it did not have a dramatic impact in absolute terms, as San José features one of the cheapest landing fees in the region. If San José was excluded from the sample, landing fees on average would have fallen by 17 percent in real terms between 2003 and 2009 in the LAC region. Aircraft Parking Charges Parking fees are time and weight based and, as with landing fees, airports employ different methods of charging airlines, varying significantly from airport to airport. Most of the airports include a grace period (free time) after landing, which ranges from one hour to up to six hours from the moment of engine shutdown (“chocks-in�). Figure 5.5 compares parking charges for a two-hour period for an Airbus A320. Zero values indicate that the two-hour period is within the free time allowance included in the landing charge, and hence the airline is not levied with any additional charges for parking. Only nine airports in LAC charge parking for two-hour periods. Fees for an Airbus A320 range from US$373 in Cancún to US$15 in Managua, with an average of US$145. In Europe and the United States, four out of seven airports included in our sample charge for parking for two-hour periods, and the average price is US$72. Parking fees have declined in real terms between 1995 and 2009, as seen in the figure 5.6. Airports where parking fees were raised in real terms between 1995 and 2003 experienced reductions between 2003 and 2009. The net effect is that at five of the six airports where prices for 1995 were available, 2009 parking fees were lower in real terms than those of 1995. 174 Airport Economics in Latin America and the Caribbean Figure 5.4 Landing Fees Percentage Change for an Airbus A320, Daylight Operation SJO UIO KIN LIM CLO VVI SAL CCS GIG GRU SCL PTY ASU MUD MGA GUA BOG LPD NAS GDL SDQ CUN MTY EZE MEX TGU –100 –50 0 50 100 150 200 % of US$ (2009) 2003 vs. 1995 2009 vs. 2003 Source: Author’s elaboration based on information from IATA (1995, 2003, and 2009), AIP Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 175 Figure 5.5 Parking Charge for an Airbus A320, for 2 Hours CUN GDL MEX MTY LAC sample average GUA LIM EZE SJO MGA ASU BOG CLO CCS KIN LPO MVD NAS PTY UIO GIG SAL UVI SCL SDQ GRU TGU LHR FRA CDG JFK LAX MAD MIA 0 50 100 150 200 250 300 350 400 US$ (2009) Source: Author’s elaboration based on information from IATA (2009), AIP Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: Buenos Aires assumes parking in operating apron; São Paulo and Rio de Janeiro assume parking in operat- ing apron; Mexico assumes tariff A (off-peak); London assumes off-peak period. For a list of airport codes and the airports they represent, see page xxiii. Figure 5.6 Changes in Parking Charges for an Airbus A320, for 2 Hours CUN GDL MEX MTY GUA LIM EZE SJO MGA ASU NAS LPD VVI GIG GRU SCL BOG CALI SDQ UIU SAL TGU KIN PTY MVD CCS 0 50 100 150 200 250 300 350 400 450 US$ (2009) 1995 2003 2009 Source: Author’s elaboration based on information from IATA (1995, 2003, and 2009), AIP Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 177 Landing and Parking Charges Less than half of the 26 airports in this sample separate landing from parking charges. In the rest, the fee paid for landing includes limited free time on the ground. Therefore, in order to make an accurate comparison of landing fees, fees have to be aggregated with the cost for parking. Figure 5.7 shows the consolidated cost of landing and parking for the same sample of airports. Figure 5.7 illustrates that when aggregating landing and parking fees into one measurement, the ranking of airports changes significantly. For example, Quito is now the most expensive airport in the sample, followed by Cancún, although Cancún was among the least expensive when the comparison considered only landing fees costs. It should be highlighted that airports in Central America and the Caribbean tended to congregate at the lower end of the graph, indicating that they generally are among the cheapest in the LAC sample when landing and parking fees are ana- lyzed jointly. More generally, under this scenario, landing and parking charges declined in real terms between 1995 and 2009 (figure 5.8). For the airports where 1995 prices were available, average landing and parking fees dropped from US$511 in 1995 to US$408 in 2003 and US$386 in 2009. If all the airports in the sample are considered, the average charge in real terms decreased from US$398 in 2003 to US$348 in 2009. Boarding Bridge Charges The charge for the use of a boarding bridge also differs among airports in the sample: some charge a fixed amount per usage (connection fee), while others consider the aircraft’s MTOW and the time it stays connected to the gate. Figure 5.9 presents the calculation of boarding bridge charges for an Airbus A320, relying on the assumptions outlined in table 5.2. As seen in figure 5.9, there is less variation for boarding bridge charges than for landing fees, as charges in the majority of the airports consider only the time the aircraft is connected with the gate. Airports like San José, Santo Domingo, Asunción, Montevideo, and Quito are exceptions, as they contemplate the type of aircraft when charging for the use of a boarding bridge. An Airbus A320 is charged for a two-hour boarding bridge usage between US$200 in Buenos Aires and US$25 in Santo Domingo, with an average of US$89 in the 22 LAC airports that charge for this service.4 There is a significant variance in this charge among the European and U.S. airports included in our sample. Some airports, such as London 178 Airport Economics in Latin America and the Caribbean Figure 5.7 Landing Fees and Parking Charge for an Airbus A320, for 2 Hours, 2009 UIO CUN GDL VVI LPD LIM CCS GIG GRU MEX MVD CLO MTY LAC sample average KIN BOG SCL EZE MGA SAL SJO TGU PTY NAS GUA ASU SDQ JFK CDG MAD LAX FRA MIA 0 200 400 600 800 1,000 1,200 US$ (2009) Source: Author’s elaboration based on information from IATA (2009), AIP Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: Buenos Aires assumes on-time payments and parking in operating apron; São Paulo and Rio de Janeiro assume parking in operating apron; San José includes airside infrastructure charge; Santo Domingo assumes landing during off-peak hours; Guatemala includes ramp fees; Mexico assumes tariff A (off-peak); Los Angeles assumes the airline is a signatory carrier; New York assumes landing during off-peak hours. For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 179 Figure 5.8 Landing Fees and Parking Charge for an Airbus A320, for 2 Hours, 1995–2009 UIO CUN GDL VVI LPD LIM CCS GIG GRU MEX MVD CLO MTY KIN BOG SCL EZE MGA SAL SJO TGU PTY NAS GUA ASU SDQ 0 100 200 300 400 500 600 700 800 900 US$ (2009) 1995 2003 2009 Source: Author’s elaboration based on information from IATA (2009), AIP Colombia, AIP Costa Rica, El Salvador Airport, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: For a list of airport codes and the airports they represent, see page xxiii. 180 Airport Economics in Latin America and the Caribbean Figure 5.9 Boarding Bridge Charges for an Airbus A320, for 2 Hours, 2009 EZE CCS LIM MEX BOG CLO KIN UIO LAC sample average GDL CUN PTY LPD VVI SCL NAS GUA TGU MVD SJO ASU MTY SDQ MGA GIG SAL GRU MAD CDG FRA MIA LHR LAX JFK 0 50 100 150 200 250 300 350 400 US$ (2009) Source: Author’s elaboration based on information from IATA (2009), AIP Colombia, AIP Costa Rica, Aerodom, Dominican Republic, El Salvador Airport, AIP Guatemala, AIP Honduras, AIP Nicaragua, Panama CAA, Direccion Nacional de Aviación Civil (DINAC) Paraguay, Montevideo Airport. Note: Mexico assumes tariff A (off-peak); Madrid assumes normal tariff. For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 181 LHR, Los Angeles, and New York’s JFK, do not charge for the use of a boarding bridge, while others, such as Madrid and Paris CDG, charge more than US$300. It is difficult to draw a general conclusion regarding the evolution of boarding bridge charges given the lack of information for 1995 prices at many airports (figure 5.10). Considering all the airports in the sample where prices were available in 2003 and 2009, the average boarding bridge charge in LAC increased by less than 10 percent, from US$80 in 2003 to US$86 in 2009 (there were increases in real terms at 8 airports, and reductions at 10). Passenger Charges Charges levied on passengers—regardless of whether they are collected directly from passengers or through airline tickets—are referred to as “passenger charges� for the purpose of this analysis. Passenger charges include the passenger facility charge (also commonly referred to as the “boarding fee�), security fees, and other taxes. Some passenger charges are imposed by the national government (such as tourist taxes), and the airport may not collect nor receive those funds. They could be included by the airlines in the ticket price or collected from passengers upon check-in at airport counters or through commercial banks located at the airport. Although these country-specific taxes levied on passengers are not part of the airfare, they do represent an integrated cost of the journey for the passenger. Consequently, depending on the price elas- ticity of demand, taxes could have a substantial impact on the decision to travel. Two different evaluations were carried out. The first (figure 5.11) includes only those charges that are levied by the airport, while the sec- ond (figure 5.12) contains all charges and taxes levied on the passengers, including airport-related services and federal taxes.5 All passenger charges presented in this section of the study pertain to departing international passengers. For airports charging separately for arriving and departing pas- sengers, both charges were considered as if collected from departing pas- sengers. Charges other than the passenger facility charge and security fees were labeled as federal taxes, since ultimately they serve the same pur- pose. The federal taxes concept is summarized in table 5.3 and can include tourist taxes, taxes levied within tickets, customs and immigration fees, among others. The passenger facility charge ranges from US$44.10 in Nassau to US$7.60 in Kingston, with a LAC sample average of US$27.70. Including Figure 5.10 Boarding Bridge Charges for an Airbus A320, for 2 Hours, 1995–2009 EZE CCS LIM MEX BOG CLO KIN UIO GDC CUN PTY LPD VVI SLL NAS GUA TGU MVD SJO ASU MTY SDQ GIG GRU SAL MGA 0 50 100 150 200 250 US$ (2009) 1995 2003 2009 Source: Author’s elaboration based on information from IATA (2009), AIP Colombia, AIP Costa Rica, Aerodom, Do- minican Republic, El Salvador Airport, AIP Guatemala, AIP Honduras, AIP Nicaragua, Panama CAA, Direccion Nacional de Aviación Civil (DINAC) Paraguay, Montevideo Airport. Note: Mexico assumes tariff A (off-peak); Madrid assumes normal tariff. For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 183 Figure 5.11 Passenger Charges per Passenger (Charges Levied by the Airport) NAS CCS UIO GIG GRU BOG CLO GUA MGA SJO EZE LIM SCL MEX LAC sample average SDQ TGU ASU MVD LPD CUN MTY VVI PTY SAL GDL KIN CDG FRA LHR LAX MIA MAD JFK 0 10 20 30 40 50 60 US$ (2009) passenger facility charge security Source: Author’s elaboration based on information from IATA (2009), AIP Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama CAA, DINACIA Uruguay. Note: Mexico assumes tariff A (off-peak); Madrid assumes normal tariff. For a list of airport codes and the airports they represent, see page xxiii. Figure 5.12 Charges and Taxes Levied on Passengers, per Passenger EZE NAS MEX UIO CCS CUN MGA MTY BOG CLO GIG GRU LAC sample average GDL GUA SJO KIN LIM SCL SDQ TGU ASU MVD LPD VVI PTY SAL LHR LAX MIA JFK CDG FRA MAD 0 20 40 60 80 100 120 US$ (2009) passenger facility charge security federal taxes Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Mexico assumes tariff A (off-peak); Madrid assumes normal tariff. For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 185 Table 5.3 Passenger Charges and Taxes per Departing Passenger U.S. dollars Passenger facility Security Federal taxes Airport charge fee Charge Remarks Buenos Aires EZE 29.0 2.5 10.0 + 40.0 Customs and immigration + ticket tax (5% of fare, assumes US$800) Nassau 44.1 7.0 n.a. La Paz 24.0 n.a. n.a. Santa Cruz 20.0 n.a. n.a. Rio de Janeiro 36.0 n.a. n.a. São Paulo 36.0 n.a. n.a. Santiago 30.0 n.a. n.a. Bogotá 33.0 n.a. 5.0 Tourist tax Cali 33.0 n.a. 5.0 Tourist tax San José 29.3 2.4 n.a. Santo Domingo 27.5 n.a. n.a. Quito 35.8 3.0 5.0 Tourist tax San Salvador 19.9 n.a. n.a. Guatemala 30.0 2.5 n.a. Tegucigalpa 26.4 n.a. n.a. Kingston 7.6 1.4 11.6 + 10.0 Air passenger tax + tourist tax Cancún 23.2 0.2 18.1 Tourist tax Guadalajara 15.0 0.2 18.1 Tourist tax Mexico City 29.6 0.2 18.1 Tourist tax Monterrey 21.1 0.2 18.1 Tourist tax Managua 32.0 n.a. 8.0 Tourist tax Panama 20.0 n.a. n.a. Asunción 25.0 n.a. n.a. Lima 30.3 n.a. n.a. Montevideo 25.0 n.a. n.a. Caracas 42.9 n.a. n.a. Paris CDG 26.9 14.6 9.9 + 5.6 Civil aviation tax + solidarity tax Frankfurt 27.7 12.3 n.a. Madrid 10.2 2.0 n.a. London LHR 34.7 n.a. 65.5 Air transportation tax Los Angeles 12.9 5.0 17.5 + 30.8 Immigration, customs + air transport tax Miami 10.9 5.0 17.5 + 30.8 Immigration, customs + air transport tax New York JFK 4.5 5.0 17.5 + 30.8 Immigration, customs + air transport tax Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: n.a. = not applicable. 186 Airport Economics in Latin America and the Caribbean security charges, which are levied at 10 LAC airports (in some countries the service is provided by the government and paid for with general taxes), total passenger charges levied by the airport vary from US$51.10 to US$9.00 in the LAC region. In the European and U.S. airports included as benchmarks, the passenger charge average is US$24.40, and the maxi- mum and minimum charges are US$41.50 in Paris CDG and US$9.50 in New York JFK, respectively. When considering all charges and taxes levied on passengers (figure 5.12), Buenos Aires is the most expensive airport to depart from, with US$81.50 per passenger.6 Travel taxes can also become a significant part of the ticket price in Europe and the United States, as is the case, in our sample, of London, Los Angeles, Miami, and New York. Table 5.3 details all passenger charges and taxes. When measured in real terms, passenger charges have clearly increased since 1995, as seen in figure 5.13. Out of 14 airports for which 1995 information was available, 11 raised passenger charges in real terms. The average passenger facility and security charge at those 14 airports rose from US$20 in 1995 to US$29 in 2009. Total Turnaround Cost As some charges are levied on the aircraft and others on the passengers, the appropriate method to compare aeronautical charges as a whole (and their evolution in time) is by calculating the cost of a turnaround.7 The cost of a turnaround is an agglomeration of all the above-mentioned charges levied by the airport, including landing fees, parking fees, board- ing bridge charges, passenger facility charge (or boarding fees), and secu- rity charges. Charges levied by other entities, such as tourist and travel taxes, are excluded from the analysis, as they are subject to great variation between countries. The calculation was performed for an Airbus A320 and for a Boeing 767-300 with a 71 percent load factor, for a two-hour turnaround.8 Figure 5.14 presents turnaround costs for an Airbus A320. Total aeronautical charges (paid by airlines and passengers) for a two- hour turnaround for an Airbus A320 range from US$5,603 in Nassau to US$1,378 in Kingston. In the sample, average aeronautical charges in LAC for an Airbus A320 with a 71 percent load factor on a two-hour turnaround is US$3,433, whereas for the selected airports in Europe and the United States the average charge is US$3,233. Figure 5.13 Changes in Passenger Charges per Passenger (Charges Levied by the Airport) NAS CCS UIO GIG GRU BOG CLO GUA MGA EZE LIM SCL MEX SJU SDQ TGU ASU MVD LPD CUN MTY VVI PTY SAL GDL KIN 0 10 20 30 40 50 60 US$ (2008) 1995 2003 2009 Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Mexico assumes tariff A (off-peak); Madrid assumes normal tariff. For a list of airport codes and the airports they represent, see page xxiii. Figure 5.14 Turnaround Costs for an Airbus A320 (2 Hours, Daylight Operation) NAS CCS UIO GIG GRU CLO BOG LIM EZE MEX MGA GUA SJO SCL LAC sample average CUN LPD MVD TGU SDQ ASU VVI MTY PTY SAL GDL KIN CDG FRA LAX MAD JFK MIA 0 1 2 3 4 5 6 US$, thousands (2009) paid by airlines paid by passengers Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Includes landing, parking, boarding bridge, passenger facility charge, and security. Assumes a 71 percent load factor. For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 189 Total aeronautical charges, as defined in this report, increased in real terms at most airports between 1995 and 2009, as seen in figure 5.15. Considering the 14 airports for which 1995 information was available, the average total turnaround cost increased in real terms from US$2,617 in 1995 to US$3,241 in 2003, and then to US$3,516 in 2009. This rep- resents a 20 percent increase between 1995 and 2003, an 8 percent increase between 2003 and 2009, and a 26 percent increase between 1995 and 2009. Considering all the airports in the sample, the average turnaround cost increased 9.8 percent between 2003 and 2009. The ranking of airport turnaround costs is virtually unchanged when the analysis is done for a Boeing 767-300 (figure 5.16 and figure 5.17). The evolution of turnaround costs between 1995 and 2009 for a Boeing 767-300 is the same as for the A320. The only significant dif- ference in total turnaround costs is the absolute magnitude of costs. Turnaround costs are, as expected, much higher for a Boeing 767-300, as it carries many more passengers and the fee structure relies more heavily on charges to passengers. Furthermore, in cases where charges are defined by MTOW, Boeing 767-300 aeronautical charges are natu- rally higher. Although total charges for both types of aircraft increased in real terms from 1995 to 2009, it is important to note that the structure of charges changed during this period. Fees paid by airlines decreased between 1995 and 2009, while fees levied on passengers increased. This result is independent of the type of aircraft. Figure 5.18 and figure 5.19 show charges levied on airlines for the Airbus A320 and the Boeing 767- 300. Figure 5.18 shows that charges paid by airlines for an A320 have decreased in real terms at most airports. For the airports where 1995 charges were available, the average aeronautical charges paid by airlines decreased from US$535 in 1995 to US$472 in 2003 and to US$462 in 2009. When considering all the airports in the sample, average charges dropped from US$454 in 2003 to US$424 in 2009. The reductions in charges measured in real terms are mainly caused by the effects of infla- tion, as nominal prices either remained constant or grew slightly. In 2009, charges levied on passengers account for, on average, over 85 percent of the total aeronautical charges. Considering only those airports where 1995 prices were available, average aeronautical charges paid by airlines for a Boeing 767-300 declined from US$1,353 in 1995 to US$1,129 in 2003 and to US$1,089 in 2009. If the 26 airports in the sample are included, charges in real terms decreased from US$1,028 in 2003 to US$935 in 2009. Figure 5.15 Changes in Turnaround Costs for an Airbus A320 (2 Hours, Daylight Operation) NAS CCS UIO GIG GRU CLO BOG LIM EZE MEX MGA GUA SJO SCL CUN LPD MVD TGU SDQ ASU VVI MTY PTY SAL GDL KIN 0 1 2 3 4 5 6 US$, thousands (2009) 1995 2003 2009 Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Includes landing, parking, boarding bridge, passenger facility charge, and security. Assumes a 71 percent load factor. For a list of airport codes and the airports they represent, see page xxiii. Figure 5.16 Turnaround Costs for a Boeing 767-300 (2 Hours, Daylight Operation) NAS UIO CCS GIG GRU EZE BOG CLU LIM MEX MGA SCL SJU GUA LAC sample average CUN LPD TGU MUD SDQ ASU VVI MTY PTY SAL GDL KIN CDG FRA MAD LAX JFK MIA 0 2 4 6 8 10 12 US$, thousands (2009) paid by airlines paid by passengers Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Includes landing, parking, boarding bridge, passenger facility charge, and security. Assumes a 71 percent load factor. For a list of airport codes and the airports they represent, see page xxiii. Figure 5.17 Changes in Turnaround Costs for a Boeing 767-300 (2 Hours, Daylight Operation) NAS UIO CCS GIG GRU EZE BOG CLO LIM MEX MGA SCL SJO GUA CUN LPD TGU MVD SDQ ASU VVI MTY PTY SAL GDL KIN 0 2 4 6 8 10 12 US$, thousands (2009) 1995 2003 2009 Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Includes landing, parking, boarding bridge, passenger facility charge, and security. Assumes a 71 percent load factor. For a list of airport codes and the airports they represent, see page xxiii. Figure 5.18 Turnaround Costs Levied on Airlines for an Airbus A320 (2 Hours, Daylight Operation) UIO LIM CCS CUN GDL VVI MEX LPD EZE CLO MVD GIG GRU KIN BOG SCL MTY PTY MGA TGU NAS SJU SAL GUA ASU SDQ 0 100 200 300 400 500 600 700 800 900 1,000 US$ (2009) 1995 2003 2009 Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Includes landing, parking, and boarding bridges. For a list of airport codes and the airports they represent, see page xxiii. Figure 5.19 Changes in Turnaround Costs Levied on Airlines for a Boeing 767–300 (2 Hours, Daylight Operation) UIO LIM CUN GDL EZE CCS MEX VVI LPD GIG GRU BOG CLO SCL KIN MTY MGA MVD SJO PTY SAL TGU NAS GUA ASU SDQ 0 500 1,000 1,500 2,000 2,500 US$ (2009) 1995 2003 2009 Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Includes landing, parking, and boarding bridges. For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 195 With respect to charges levied on passengers, on the other hand, these have been raised in real terms at more than half of the airports in the sample. Average charges levied on passengers increased from US$2,120 in 1995 to US$2,767 in 2003 and US$3,057 in 2009 for those airports where 1995 information was available. If the 26 airports in the sample are considered, average charges grew from US$2,672 in 2003 to US$3,010 in 2009 (figure 5.20). Conclusion According to the sample of airports gathered for this report, total aero- nautical charges in LAC increased between 1995 and 2009 (34 percent for those with 1995 data and 9.8 percent between 2003 and 2009 for all airports). The tariff benchmarking analysis does not permit us to reach a conclusion about the relationship between changes in aeronautical charges and the introduction of private sector participation. A simple visual analysis indicates that the increase in aeronautical charges observed between 1995 and 2009 is shared by both publicly and privately operated airports. Further research through a case-specific approach should be conducted to assess whether the introduction of private sector participa- tion has led to an increase in aeronautical charges and to link changes in aeronautical charges to the changes in the level and quality of airport services. An important conclusion regarding aeronautical charges is that for both types of aircrafts, the Airbus A320 and the Boeing 767-300, total turnaround costs in LAC in 2009 are, on average, at a comparable or higher level than those in European and U.S. airports that are most fre- quently served by Latin American airlines. Several questions, which merit further research, need to be answered in order to fully understand why this is the case. For example, are aeronautical tariffs set on a cost-recovery basis? Do aeronautical tariffs reflect an adequate due diligence process? How are they modified through time? Do aeronautical tariffs provide the right incentives for infrastructure investments? Finally, the results indicate that the structure of aeronautical charges has changed in the last decade. The importance of charges applied to pas- sengers is increasing relative to those levied on airlines. Passenger charges currently account for over 85 percent of total aeronautical charges. On the other hand, charges levied on airlines (such as landing fees, aircraft parking, and boarding bridges) either remained constant in nominal terms or grew at a slower pace than inflation, demonstrating that charges in real Figure 5.20 Turnaround Costs Levied on Passengers, for an Airbus A320 NAS CCS UIO GIG GRU BOG CLO GUA MGA SJO EZE LIM SCL MEX SDQ TGU ASU MVD LPD CUN MTY VVI PTY SAL GDL KIN 0 1 2 3 4 5 6 US$, thousands (2009) 1995 2003 2009 Source: Author’s elaboration based on information from International Air Transport Association (IATA) Airport and Air Navigation Charges Manual (2009), Aeronautical Information Publication (AIP) Colombia, AIP El Salvador, AIP Honduras, AIP Nicaragua, Panama Civil Aviation Authority (CAA), Dirección Nacional de Aviación Civil e Infrae- structura Aeronaútica (DINACIA) Uruguay. Note: Includes passenger facility charge and security. Assumes a 71 percent load factor. For a list of airport codes and the airports they represent, see page xxiii. Benchmarking of Aeronautical Charges at Latin American Airports 197 terms are lower today than they were in 1995. The current tariff structure in LAC airports is similar to that prevailing in the sample of European and U.S. airports, with a slightly higher percentage of the share of pas- senger charges versus airline charges in LAC. Two main explanations can be provided to account for the changes in the tariff structure. The first, based on a political economy argument, is that airlines as a group have a higher negotiating power through their trade associations, such as IATA or the Latin American Airline Association (ALTA), whereas individual travelers have neither the resources nor the organization to fight tariff increases. This is not to say that airlines show resistance only toward price increases in aircraft-based charges and not toward charges paid by passengers. Clearly, it is in the airlines’ best inter- est to ensure that passengers assume the lowest possible travel cost. As ticket prices increase, in turn, demand is reduced, affecting the airlines’ bottom lines. Another potential explanation is that the current tariff structure better reflects relative demand elasticities. If this is the case, then relative charges were modified by regulatory agencies because pas- sengers have lower demand elasticity (for the use of a given airport) than airlines. While further research needs to be conducted in order to provide a more in-depth analysis of the evolution of tariffs, the present work repre- sents an important first step in fostering dialogue on these issues and in laying down the basis for a more robust tariff benchmarking exercise. Notes 1. This assumption has significant weaknesses. Concessions took place in differ- ent years throughout LAC. To correctly test the hypothesis that tariffs increase after the introduction of the private sector, a detailed study of the evolution of tariffs in each airport in each country should be conducted. We opted to include, arbitrarily, the year 2003, as it is the year with the most comprehensive information on tariffs available. 2. The frequency depends on the type of airport, but for the major airports, information on charges is collected on a semiannual basis. 3. Aeronautical charges for 1995 could be obtained for only a set of 13 airports. 4. It is interesting to note that INFRAERO, the Brazilian company operating the airports in Brazil, does not have a separate charge for boarding bridges. The use of the bridges is included in the parking charge, which is the same for remote stands as for contact positions. According to collected information, the reason for pricing both services equally is that although boarding bridges are generally more expensive to purchase (cost of capital) and operate (increased 198 Airport Economics in Latin America and the Caribbean level of service following a principle of the airline’s “willingness to pay�) than remote positions, the latter have to cover the extra cost of shuttle services for passengers (following a principle of operating cost recovery). 5. Visa costs are not considered for the purpose of this study. 6. Airfares for international travel in Buenos Aires (EZE) are charged with a tourist tax calculated as 5 percent of the airfare (called “DNT� or Dirección Nacional de Turismo). For the purpose of this analysis, an assumption of a US$800 airfare was employed; aggregated fees, taxes, and duties total US$40 on top of the ticket cost. 7. Turnaround refers to all activities involved in handling an aircraft between its arrival and its departure (typically known as “from chocks-in to chocks-out�). In this chapter, turnaround costs do not include any operational costs for the airline other than airport charges. 8. Average load factor for Latin America in 2009 was reported to be 71 percent, according to Air Transport World, June 2009. References Air Transport World Media Group. 2009. Air Transport World. June. http:// atwonline.com. IATA (International Air Transport Association). 1995. Airport and Air Navigation Charges Manual. Montreal, Quebec: IATA. ———. 2003. Airport and Air Navigation Charges Manual. Montreal, Quebec: IATA. ———. 2009. Airport and Air Navigation Charges Manual. Montreal, Quebec: IATA. APPENDIX A Survey of Airport Performance for Operators QUESTIONNAIRE ON AIRPORT SECTOR PERFORMANCE IN LATIN AMERICA AND THE CARIBBEAN Please fill in the questions to the best of your knowledge. We realize that it may not be possible for you to find all the information we are asking for. For those years for which data are available, we would appreciate it if you could provide as many details as possible. Finally, please indicate if there is any specific information that you would prefer for us to keep confidential. Airport Information and Point of Contact: Country: Airport name: Code (IATA): Name of the point of contact: Phone number: 199 200 Airport Economics in Latin America and the Caribbean Fax: E-mail: General Information: 1. What is the airport’s form of ownership (Public/Management Contract/Concession/Private)? 2. If a concession, management contract, or privatization, what year did the transition take place? 3. If a concession or management contract, what is its duration? 4. What is the name of the airport’s operator? 5. Please list the names of the major shareholders/companies that operate the airport. Specify the percentage of shares and voting rights of each. 6. What is the airport’s fiscal year? (e.g., April to March) If your data do not relate to a calendar year, please include data in the box for which the year ends, e.g., data for April 1998 to March 1999 would cor- respond to the 1999 box. OUTPUT VARIABLES 1. Passenger Data: 1.1–1.5 How many passengers of each type were handled by the airport? (Numbers in thousands) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 1.1 International passengers 1.2 Domestic passengers 1.3 Scheduled passengers 1.4 Non- scheduled passengers (i.e., charter passengers) 1.5 Transfer passengers Survey of Airport Performance for Operators 201 1.6 What was the total number of passengers handled by the airport? (Numbers in thousands) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 1.6 Total passengers 2. Cargo Data: 2.1–2.3 What were the total tons of cargo (freight and mail) handled by the airport? (Thousand metric tons) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2.1 Total cargo handled (loaded and unloaded) INTERNATIONAL 2.2 Total cargo handled (loaded and unloaded) DOMESTIC 2.3 Total cargo handled (loaded and unloaded) 3. Intermediate Output Data: 3.1–3.5 What were the total aircraft movements registered by the airport (ATMs)? (Numbers in thousands) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 3.1 Passenger aircraft INTERNATIONAL 3.2 Passenger aircraft DOMÉSTIC 3.3 Total passenger aircraft 3.4 Cargo-only aircraft 3.5 General aviation and other aircraft 202 Airport Economics in Latin America and the Caribbean 3.6 (3.6.1–3.6.6) What was the aircraft mix? (Percentage of total aircraft movements related to each of the following type of aircrafts) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 3.6.1 Largest wide- body aircraft (e.g., 747/777/ A340/A330) 3.6.2 Other large wide-body aircraft (e.g., DC10/DC11/ L1011) 3.6.3 Medium wide-body aircraft (e.g., 757/767/ A300/A310) 3.6.4 Narrow-body aircraft (e.g., 727/737/ A320/DC9/ MD80/MD90) 3.6.5 Commuter/ Turboprops 3.6.6 Other/ General aviation 4. Traffic Peaking: 4.1 What were total passenger numbers in the busiest month of the year? 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 4.1 Peak month passenger traffic 4.2 What were total passenger numbers during the busiest hour of the year? 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 4.2 Peak hour passenger traffic 4.3 What were total ATMs (air traffic movements) during the busiest hour of the year? 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 4.3 Peak hour ATMs Survey of Airport Performance for Operators 203 4.4 During the busiest month or hour of passenger numbers, where can the capacity restriction be found? In the arrival or departure of passengers? Arrival Departure 5. Financial Data: Costs and Revenues In all cases, we are specifically interested in data referring to the airport indicated on page 1 only. If, however, your airport is part of a larger group and there are no airport-level data available, please provide any corporate figures at the group level. In addition, please provide details on any assump- tions that have been made to allocate costs across airports when answering the questions. Currency used for financial data in this section: Units used for financial data in this section (e.g., thousands or millions): Assumptions: (Please make any clarification you deem appropriate. Indicate the types of services that are accounted for in the costs and revenues.) COSTS: 5.1 What were total operating costs? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.1. Total operating costs Please provide any additional information with respect to the definition of operating costs: 5.2 Are there any corporate costs associated with your airport that are not included in the total operating costs provided in 5.1 above? Yes No If yes, what is the estimated level of corporate costs associated with your airport that are not included in the total operating costs given in 5.1 above? 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.2 Corporate costs (not included in operating costs) 204 Airport Economics in Latin America and the Caribbean Please provide details with respect to the estimation method used or any further details on these corporate costs: 5.3 What was the total capital expenditure? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.3 Total capital expenditure 5.4 What were the total depreciation costs? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.4 Total depreciation costs 5.5 What was the total operation and capital expenditure? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.5 Total operation and capital expenditure REVENUES: 5.6 What was the aeronautical revenue? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.6 Aeronautical revenue 5.7 What was the nonaeronautical revenue? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.7 Nonaeronau- tical revenue 5.8 Was there any other kind of revenue? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.8 Other revenue Please specify what kind of revenue the data refer to: Survey of Airport Performance for Operators 205 5.9 Has the airport received any operational subsidy? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.9 Operational subsidy Please specify types of subsidies and sources (local, federal governments): 5.10 What was the airport’s total revenue? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.10 Total revenue 5.11 What were the airport’s earnings before interest, taxes, depreciation, and amortization (EBITDA)? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.11 EBITDA 5.12–5.16 Please provide the following information regarding assets and liabilities of your airport: 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 5.12 Fixed assets 5.13 Current assets 5.14 Capital + reserves 5.15 Current liabilities 5.16 Long-term liabilities 6. Capital Assets and Capacity Utilization: 6.1–6.11 What were the total: 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 6.1 Number of runways (number) 6.2 Runway capacity (movements per hour) (continued next page) 206 Airport Economics in Latin America and the Caribbean 6.1–6.11 What were the total: (continued) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 6.3 Passenger terminal capacity (passengers per hour) 6.4 Terminal size (square meters) 6.5 Terminal space used for retail activities (square meters) 6.6 Number of contact gates (boarding bridges) (number) 6.7 Aircraft parking stands (remote + bridges) (number) 6.8 Cargo terminal capacity (square meters) 6.9 Number of check-in desks (number) 6.10 Number of baggage claim units (number) 6.11 Number of seats provided by airport defined by airside seating after security check (number) Survey of Airport Performance for Operators 207 7. Quality Data: 7.1–7.3 What were the following? 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 7.1 Average baggage delivery time (minutes) 7.2 Average check-in waiting time (minutes) 7.3 Average security waiting time (minutes) Please provide any clarification or explanation with respect to the measure- ment of the data contained in boxes 7.1 to 7.3: 8. Employee Composition: 8.1 How many employees were contracted directly by the airport management company? 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 8.1 Employees contracted directly by the airport management company (number) 8.2 How many employees worked in the airport? (total, includes those who are contracted directly by the operator plus those employed in services outsourced) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 8.2 Total employees in airport (number) 9. Total Aeronautical Fees for an Airbus A320 Aircraft: Characteristics of an Airbus A320 Aircraft: Aircraft type: A320-200 Maximum takeoff weight: 73.5 metric tons 208 Airport Economics in Latin America and the Caribbean Passengers: 120 (load factor: 73.8%, typical seating: 162) Type of flight: Regular, international (LAC country/LAC country) Turnaround time: 2 hours (at peak hour) Please do not include discounts (i.e., transfer passenger discounts) when reporting data. 9.1–9.9 What were the fees for each of the following? (Local nominal currency) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 9.1 Landing 9.2 Passengers 9.3 Security 9.4 Parking 9.5 Contact gate (aerobridge) 9.6 Terminal navigation 9.7 Noise/ pollution 9.8 Other (please specify below) 9.9 Total 9.9 Please specify what other aeronautical fees, if any, are charged to the operation of an Airbus A320 aircraft: EXTRA VARIABLES 10. Fee Structure: Please describe the fee structure for each of the following. Of particular importance is the change of each fee structure over time. For landing, please specify if the tariff changes by weight of aircraft, time of day, or other variables. For passenger, please differentiate between domestic and international tariffs. APPENDIX B Governance of Airport Regulators Survey QUESTIONNAIRE ON THE GOVERNANCE OF REGULATORY AGENCIES IN THE AIRPORT SECTOR IN LATIN AMERICA AND THE CARIBBEAN Regulatory agency Country Name of the person in charge of answering the questionnaire Position in the agency Telephone number E-mail The present questionnaire is divided into three main sections. The first section is composed of general questions related to the regulation of the airport sector in your country. The second contains questions that intend to identify different aspects related to the governance of regulatory 209 210 Airport Economics in Latin America and the Caribbean agencies in the airport sector. Finally, the third section contains questions that ask your opinion on the institutional scheme adopted to regulate airports in your country. The questionnaire contains a glossary of those terms that could lead to confusion. In the present questionnaire, the word agency is used indistinctly to refer to both independent regulatory bodies and civil aviation administrations (without independent regulator characteristics). In accordance with the objectives of this research project led by the World Bank, it is very important that you answer the present question- naire based on your experience and objectives as regulatory agency of the sector. A better understanding of this subject will allow us to conduct a more comprehensive analysis of the airport sector in Latin America and the Caribbean. When answering the questionnaire, you will note that in several cases, the questions are not numbered consecutively. The reason for this kind of nontraditional numbering is explained by the need to facilitate the com- parison of the answers received by agencies in the airport sector to those received by other infrastructure sectors that are answering similar ques- tionnaires. I. General Questions: 1. List the major airport operators in your country: OPERATOR MANAGED AIRPORTS MAJOR SHAREHOLDERS 2. Who has decision-making competencies over the following aspects? Government Not Agency operator Airport Other applicable Tariff structure (aeronautical) Tariffs (commercial services) Tariff modifications (aeronautical) (continued next page) Governance of Airport Regulators Survey 211 Government Not Agency operator Airport Other applicable Tariff modifications (commercial) Quality of service User complaints Investment plans (ex ante approval) Investment plans (ex post, fulfillment) Slot allocation Anticompetitive practices Merger and acquisition reviews Authorization of ground handling providers Technical/security standards 3. Since 1995 until today, have there been significant changes in the man- agement and regulation of airports (concessions, legislative changes, changes in airport operators, among others)? Please describe briefly. 4. When did the agency begin to operate? Month/Year: 5. Describe the main functions that the agency performs according to the specified mandate within the legal instrument that created it. 6. Is the agency sectoral or multisectoral? If it is multisectoral, specify the sectors that are regulated by the agency. 7. If your agency regulates private airport operators, was the agency cre- ated before the introduction of private management? Yes No 212 Airport Economics in Latin America and the Caribbean If the answer is yes, what was your agency’s role in the process through which private management was introduced? (mark all that apply): The agency: Issued nonbinding opinions Was actively involved in the design of the concession Developed the economic-financial model Developed the technical specifications Defined the concession’s initial aeronautical tariffs Participated in the selection of the concessionaire If you wish to explain the agency’s role in the process of introduction of private sector management in the airport sector in more detail, please do so below: 8. Tariff regulation: a) What is the option that better describes the regime or modality of changes in the aeronautical tariffs of the airports under your jurisdic- tion? Mark (only one) The tariffs are freely set by the airports without any intervention from a state entity. The tariffs are freely set by the airports, but the state reserves the right to revise them when necessary. The level and change in tariffs is negotiated between the operator and the agency without a process that has been established in norms or manuals. Tariffs are modified through a formal petition from the operator, following a formal administrative process that requires approval from your agency or from a ministry. Tariffs are fixed for a predetermined period of time (for example, every five years) and are revised through a formal process that has been established in a concession contract or in the regulatory agency’s procedures. Other mechanisms. Please explain the methodology and procedure of tariff setting and review in your country. Please specify authorities involved, timeline, and mechanisms. Governance of Airport Regulators Survey 213 9. Describe the current tariff structure in the airports under your juris- diction (for example, specify if different tariffs apply to international passengers, if landing tariffs vary by hour or day and weight of the aircraft, etc): Passengers: Landing: Security: Others: 10. Have there been significant modifications since 1995 in the tariff structure mentioned in question 9? 11. Are the tariffs corresponding to nonaeronautical (commercial) ser- vices subject to any kind of regulation? Yes No If yes, please describe the tariff structure and the modification mech- anisms of these tariffs. 12. Are the costs associated with the provision of aeronautical services recovered through the aeronautical tariffs? Yes No 13. The current mechanisms for tariff setting respond better to: Mark (only one) Single till (The operational costs related to aeronautical services are recovered through revenues generated by the charging of aeronautical services and through revenues obtained from non-aeronautical [commercial] activities.) Dual till (The operational costs related to aeronautical services are exclusively recovered through revenues generated by the charging of aeronautical services.) Hybrid If hybrid, please explain: 14. Does the agency use an economic-financial model as a basis to calculate changes in tariffs? Yes No If yes, describe if this model was developed by the agency or by consultants. 214 Airport Economics in Latin America and the Caribbean 15. Does the agency rely on a manual of regulatory accounting devel- oped for the airport sector in your country to request information from the operator(s) (e.g., detailed information on costs differenti- ated by type of services)? Yes No If yes, describe if this model was developed by the agency or by consultants. 16. Does the agency estimate the capital costs incurred by the operators in the airport sector? How often is this estimation performed? Is it performed by internal staff or by external consultants? Explain 17. Does the regulatory framework allow operators to grant airport tariff discounts to airlines (e.g., by amount of flights or types of airplanes)? Yes No Explain 18. Does the regulatory framework establish minimum levels of service quality in the airports? Yes No If yes, who sets the service quality levels? 19. What economic incentives do airport operators have to increase the quality of their services? 19B. Does the legal framework allow the airport operator to charge dif- ferent aeronautical tariffs as a function of the quality of service provided (e.g., if a low-cost airline wants to receive a lower quality of service, can the airport operator charge lower tariffs)? Yes No Explain: 19C. Does the agency conduct economic-financial audits of the airport operators? If so, with what frequency? Yes No Frequency: Governance of Airport Regulators Survey 215 19D. Is the ground-handling service liberalized in your country? Yes No Is there any regulation for this type of service? If there is a limit to the number of ground-handling service pro- viders, please indicate the total number allowed, the current number, and their names (and shareholders if the information is available). 19E. Slots allocation policies in your country: What entity grants slots? In what airports are slots allocated? Do airlines have to pay for them? Can they buy them and sell them? 19F. What is the agency’s role in mergers and acquisitions and in anti- trust cases? Mark (only one) Final decision made by the aviation agency Final decision made by another agency with previous mandatory consultation with the aviation regulatory agency. Final decision made by another agency. The final decision maker is not obliged either to request or to consider the authority’s opinion. Other If other, please explain: 19G. Does your agency regulate the provision of services related to ter- minal (air side) navigation? Yes No Explain how this service is financed. 19H. Does your agency regulate the provision of services related to air- port security? Yes No Explain how this service is financed. 216 Airport Economics in Latin America and the Caribbean 19I. Does your country have a master plan for airport investments? Yes No Please indicate who has the authority to develop this investment plan and what airports are included in it. 19J. Specify, contingent upon data availability, the capital investments (runways, terminals, airport systems, among others) that the airport system has undergone since 1997 (in local nominal currency or U.S. dollars. Please specify currency used). 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Investments Airports that received investments 19K. If the airports that your agency regulates are concessioned: Does the agency approve investment plans? Yes No Does the agency determine, before authorizing the investment, if it corresponds to capital or maintenance? Yes No 19L. Conflict resolution: describe if the regulator performs a mediating role in conflicts between users (airlines) and the airport operator. 19M. Did the airport system in your country run a surplus in 2006? Yes No 19N. Considering the 10 airports with the highest annual passenger volumes, indicate which have run surpluses or deficits. Names of the airports with surpluses: Name of the airports with deficits: How are losses from deficitary airports covered? Contributions from the Treasury or Ministry Cross-subsidies from airports with financial surplus Governance of Airport Regulators Survey 217 Debt issued by airport Other If other, please specify II. Agency’s Governance Variables: 1. Autonomy: 20. Through what legal instrument was the agency created? Please indicate the norm’s number and year. Law Number/Year Decree Number/Year Ministerial Resolution Number/Year Other (please indicate) 21. What is the agency’s legal status? It is a separate and autonomous entity from the sectoral minister. It is a separate, but non-autonomous entity from the sectoral minister. There is no agency as regulation is conducted by a ministry. Other (please indicate) 22. Can the agency be intervened? If so, please indicate what authority has the power to intervene. SEE DEFINITION IN GLOSSARY Yes Authority No 23. Has the agency ever been intervened? Yes No Number of interventions and dates: 218 Airport Economics in Latin America and the Caribbean 25. What institution is competent in the economic regulation of the air- ports in your country? Agency only Agency and another independent agency Agency and Parliament Agency and government Agency has only consultative competencies 28. Is the agency’s independence explicitly established? If so, please indicate what legal instrument establishes this independence. Yes No Law/decree where the agency’s independence is established: Clause/article of the law/decree where this independence is estab- lished (provide its text): 29. Have there been any major changes during the past five years in the responsibilities of the regulatory agency? Yes No Yes, responsibilities have decreased Please specify Yes, responsibilities have increased Please specify 30. What are the agency’s competencies? (Mark all that apply) Consultative/advisory Oversight Contract/license approval Tariff approval Normative creation Other If other, please indicate Governance of Airport Regulators Survey 219 32. How do you evaluate the degree of interference by the sectoral min- ister (e.g., Transport or Public Works) in the decisions adopted by the agency? Very high High Low Very low 33. If there is conflict over the application/interpretation of a norm, what is the administrative authority in charge of making the final deci- sion? Explain 34. What is the mechanism for the selection of the agency’s directors? The minister appoints the members of the board/director The president appoints the members of the board/director The president appoints the members of the board/director with authorization from Congress Other (please specify) 35. Assign a value between 1 (WORST) and 5 (BEST) to the following aspects of the board member’s selection process: Transparency Merit-based Insulation from political influence 36. What are the necessary requirements to be designated as a director? College degree Experience in airport regulation Political independence There are no requirements Other (please specify) 220 Airport Economics in Latin America and the Caribbean 36A. What are the previous positions and educational levels of the agency’s current directors? Director 1 Public sector Private sector Previous position and organization Education Director 2 Public sector Private sector Previous position and organization Education Director 3 Public sector Private sector Previous position and organization Education Director 4 Public sector Private sector Previous position and organization Education Director 5 Public sector Private sector Previous position and organization Education 37. What is the duration of a director’s mandate? Fixed mandate Number of years Undefined mandate 38. Is it renewable? For how long? No Yes, for an additional period Yes, for more than one period 39. If mandate is fixed, how many directors have not completed their mandates? Less than five Please indicate the number of directors. Governance of Airport Regulators Survey 221 More than five Please indicate the number of directors. 39A. What authority is responsible for removing the agency’s directors? Parliament (or Congress) President Minister Other Please describe the dismissal procedure. 39B. Should the removal of the agency’s director be carried out accord- ing to specific causes? Yes No If justification is necessary, please specify the required causes for dismissal. 40. Has the mechanism for the dismissal of directors ever been used? Yes If so, how many times? No There are no mechanisms for the dismissal of directors 41. Select the reasons for which directors leave their positions: Yes No Removal External pressure Retirement Voluntary leave End of mandate Other If others, please specify 222 Airport Economics in Latin America and the Caribbean 42. Does the agency have the power to establish its administrative/orga- nizational structure (e.g., creation of new departments/units/divi- sions in the organizational framework, management assessment mechanisms, appointments, etc.)? Yes No If no, please specify who is the responsible authority. 43. Identify the labor regime that regulates the following situations: Private law Civil service law Directors of the board Managers Technical employees Rest of the staff Other(s) 45. Is the agency free to make its own personnel decisions (e.g., hire, promote, discipline)? Yes No If no, please identify the authority with the power to make those decisions. 46. What are the sources of the agency’s budget? Identify the percentage of each. Percentage Government budget % Fines % Donations % Tariffs % Specify type of tariff and percentage % Governance of Airport Regulators Survey 223 Tariff % Tariff % Tariff % Other(s) % 47. Does the agency have financial autonomy to determine its own expenses? Yes No If no, please identify the authority with the power to assume this role. 48. What has been the evolution of the agency’s budget over the past three years? (Local nominal currency or U.S. dollars. Please specify currency.) 2005 2006 2007 2. Accountability: 51. To whom is the agency accountable? SEE DEFINITION IN GLOSSARY. Congress Government Both 51A. Is the agency’s performance evaluated? SEE DEFINITION IN GLOSSARY. Yes No 51B. What are the main areas examined in the agency’s performance evaluation and who performs this evaluation? Administrative Efficiency (delays in addressing a complaint, transparency in appointments, other institutional quality measures) Evaluating Authority 224 Airport Economics in Latin America and the Caribbean Economic Efficiency (impact of the agency’s decisions on the market) Evaluating Authority Budgetary Performance Evaluating Authority Please describe the areas that are evaluated in further detail 54. Can the regulating agency’s decisions be appealed? Yes No 55. By whom are the appeals considered? Please identify the court/ tribunal. General law courts (excluding the Supreme Court of Justice/ Supreme Tribunal) Name of the tribunal Tribunals established specially to treat regulatory aspects Name of the tribunal Ministry/government Name of the tribunal Special administrative tribunal to deal with regulatory matters Name of the tribunal Combination of the above Name of the tribunal Other(s) (please specify) 3. Transparency: 59. Does the agency publish the methodology/data and other tools used in the application of it regulatory decisions in economic matters (e.g., the calculation of price caps)? Yes No If yes, please specify how the data is published (i.e. through the agency’s web site, bulletins, etc.). Governance of Airport Regulators Survey 225 60. How are the agency’s procedures for the elaboration of rules and the due process regulated? The agency has its own procedures The agency is subject to the same administrative procedures as those of the rest of the public sector There are no procedures for the elaboration of rules Please describe the procedures. 61. Does the airport sector legislation establish the participation of the main economic agents and of civil society (businesses, users, etc.) in the agency’s rule-making process? Yes No Please describe the procedure/mechanism through which the vari- ous actors participate in the agency’s rule-making process. 63. Does the agency perform public consultations when changes in tar- iffs are undertaken? SEE DEFINITION IN GLOSSARY. Yes No Who is invited to participate in the consultations? Airlines Yes No Passengers Yes No Nongovernmental organizations (NGOs) Yes No Others______ 64. If yes, how are the public consultations regulated? Informally Formally If formally, please specify what legal instrument regulates the public consultations. 65. What are the matters that need to be consulted with the economic agents (airlines)? Changes in tariffs Approvals of investment plans Variables that affect the quality of service Others 226 Airport Economics in Latin America and the Caribbean 65A. What is the legal effect of the agency’s consultations? The consultation’s outcome is binding for the agency The consultation’s outcome is NOT binding for the agency The outcome does not bind the agency, but the agency must justify why it made a different decision 66. How frequently and how many consultations are performed by the agency? Every two months Every six months Annually Other How many public consultations? 66A. Please list and describe the main public consultations/hearings performed to date. Please list them in order of importance. Hearing/Consultation (name) Date Outcome Other comments 1. 2. 3. 4. 5. Other hearings/consultations: 67. Is the agency obliged to publish its decisions? Yes No Please specify how the agency’s decisions are published. 70. Does the agency have a collective or individual decision-making structure? Collective Individual 71. Are there quarantine rules for the directors? SEE DEFINITION IN GLOSSARY. Yes No Governance of Airport Regulators Survey 227 72. If so, for how long? 4. Regulatory, Management, and Institutional Tools: 74. Is benchmarking used by the agency? Yes No 75. If the agency uses benchmarking, what is the methodology used? Partial indicators Total factor productivity Data evolving analysis Statistical techniques Process comparison Customer service comparison Model engineer corporation Combination of these If a combination, please specify what methods are included. 77. How would you rank the agency’s effectiveness in the enforcement of its decisions in matters of economic regulation? Very high High Medium Low Very low Comments 78. Has the agency developed its own structure of posts and salaries? Yes No Please, briefly describe the agency’s staff grades and salary scales. 228 Airport Economics in Latin America and the Caribbean 79. How many employees does the agency have? Please specify the number of technical and administrative staff under each range. Technical staff Administrative staff (Area of economic regulation) Less than 20 Between 21 and 50 Between 51 and 100 More than 100 79A. Please specify, using percentages, the current educational levels in the agency (elementary, middle school, high school, college, graduate level). Elementary Middle High College Graduate School School School Managers Percentages % % % % % Technical staff Percentages % % % % % Administrative staff Percentages % % % % % 79AA. Does the agency hire external consultants to carry studies/work on economic regulation? Yes No Please specify how many individual consultants and firms were contracted between 2005 and 2007 and the tasks they per- formed. 79B. How does the agency evaluate its staff? SEE DEFINITION IN GLOSSARY. There is a periodic evaluation according to preestablished assess- ment mechanisms (e.g., performance indicators) There is an ad hoc, discretionary evaluation, in a nonsystematic or regular way Governance of Airport Regulators Survey 229 The agency does not evaluate staff performance Please describe the evaluation mechanisms. 80. Does the agency publish its job openings and if so, where? Yes No Newspaper Agency’s website Both Other(s) 81. Does the agency use performance-based payments for its employ- ees? Yes No If so, briefly describe the payment system. 82. From the positions listed below, please select those whose hiring requires public examinations. Director1 Manager Technical staff Administrative assistants Rest of the staff Public examinations are not required 82A. How would you describe the salary levels in the agency? Similar to those of businesses in the sector Below the market level in the sector but above the public sector level Similar to those of the public sector 83. How would you rate the training the agency’s employees receive? Excellent Very good 230 Airport Economics in Latin America and the Caribbean Good Bad Very bad There is no training available 83A. In what areas does the agency provide training to its employ- ees? Leadership Briefly describe the kind of training. Sector regulation (tariff regime, investment evaluation, regulatory law, regulatory accounting, etc.) Briefly describe the kind of training. Financial and auditing Briefly describe the kind of training. 83B. What is the budget share that is annually devoted to employees’ training and development? % 84. What is the agency’s reporting instrument? SEE DEFINITION IN GLOSSARY. Annual report Agency governing authorities’ hearings before the Parliament Both There are no reporting instruments Other 85. Are consumers’ rights and obligations legislated in regulatory or nonregulatory legal instruments? Regulatory instruments Nonregulatory instruments There is no regulation Governance of Airport Regulators Survey 231 Please identify the legal instruments that regulate consumers’ rights and obligations. 86. Does the agency evaluate customers’ (i.e., users) satisfaction with the quality of the service provided? Yes No Please describe the evaluation procedure. 87. Does the agency prepare an annual report? SEE DEFINITION IN GLOSSARY. Yes No 88. If yes, is the report published? Yes No Please specify through which medium (website, printed publication, etc.). 91. Does the agency have a website? Yes No 91A. What type of information does the website contain? Airport legislation Content (periods and conditions) of concession and/or service provision contracts Public releases of the agency’s decisions/resolutions Annual performance report/accountability report Addressing of customers’ complaints Name and résumé of the board’s directors Sector indicators Please specify Other Please specify 232 Airport Economics in Latin America and the Caribbean 92. Please answer the following questions related to users’ claims: A. Number of complaints received (per year) Year 2000 2001 2002 2003 2004 2005 2006 2007 2008 Passengers Airlines B. Reasons for complaints (percentage) Quality of service Increases in taxes/tariffs Airport installations Others (specify) % % % % C. What is the average time to resolve a user complaint? D. What are the legal steps to solve a complaint? 1. The airports’ regulatory agency has final decision authority 2. Although the airports’ agency intervenes in the process, the final decision is made by another administrative authority 3. The affected party can go to the courts without the need for a final administrative decision 4. The affected party can only go to the courts after a final administrative decision is made 93. Does the agency apply regulatory quality standards to its regulations (i.e., cost-benefit analysis, alternatives to regulation, administrative simplification, regulatory impact analysis)? Yes No If so, please identify these standards and describe each of them: Cost-benefit analysis of regulations Description: Alternatives to regulation Description: Regulatory impact analysis Description: Governance of Airport Regulators Survey 233 Administrative simplification Description: User participation in the development of regulations Description: 94. Are the Board’s meetings recorded? Yes No 95. Does the agency publish its audited accounts? Yes No The agency’s accounts are not audited If yes, identify the media through which the audited accounts are published (website, print publications, etc.). 96. Does the agency have norms of ethics? Yes, it has its own norms/codes of ethics Yes, it applies the norms/codes of ethics of the public administration No 97. Have these norms been used during the past five years? Yes, it resulted in the dismissal of one of the agency’s officials/employees Yes, it resulted in a minor punishment of one of the agency’s officials/employees No If yes, please specify the employee’s position and the type of sanction applied. III. AGENCY’S POINT OF VIEW: In this section, please answer the following questions according to your opinion and point of view. Please feel free to expand upon each question as you may deem necessary. 234 Airport Economics in Latin America and the Caribbean A) Do you agree with the institutional framework that has been estab- lished to regulate the airport sector in your country? What changes, if any, would you make? We would appreciate if you could refer to the entity in charge of regulation, the regulatory framework (tariff regulation, private sector participation, granting of licenses, permits, etc.), the role of operators, the government, as well as any other refer- ence that you consider relevant. B) Are you satisfied with your agency’s performance? We would appreci- ate if you could frame your answer in the context of the four themes of our analysis: autonomy; transparency; accountability; regulatory, management/institutional tools and capacities. C) Autonomy: Transparency: Accountability: Tools and capacities: D) Please include any other thoughts you consider relevant for a better understanding of the dynamics and functioning of the institutional framework of the airport activity in your country. GLOSSARY OF DEFINITIONS: Question 22. Agency intervention: By “agency intervention.� we mean the abil- ity to suspend the agency’s authorities by the Executive branch or the Legislature, to overcome an extraordinary situation affecting the normal functioning of the body. In these cases, the intervening institution designates a power controller to act on their behalf during the transitory period. Question 51. Agency accountability: In this question we try to identify the authority before which the agency must be accountable by complying with their duties. Generally, the authority is the same that created the body. In the Common Law, the independent administrative agencies are accountable for their perform- ance to the Parliament. Accountability is understood here in a broad sense, not being limited exclusively to the budget. Governance of Airport Regulators Survey 235 Question 51A. Evaluation of the agency management: This question is com- plementing the previous one and inquires to the agency about the existence of procedures to assess the management area. In other words, we are interested in identifying mechanisms by which the entity’s performance or management is assessed. We differentiate the assessment of the agency’s management in three main areas: administrative efficiency, economic efficiency, and budget perform- ance. Both questions 51 and 51A are related, but do not necessarily address the same issue. While it may be the case that the same authority to which the agency is accountable can also be in charge of evaluating the performance, it can also be the case that the agency is assessed by a completely different body. It could be the case, for example, that the agency is subjected, and accountable, to the Parlia- ment (because this institution determines the budget and appointments) and the evaluation of their performance (taking into account the previously identified issues) is done by an entity other than the Parliament. Question 63. Public consultations: Procedure by which the agency makes available to the public at large particular issues for discussion and consideration. Unlike the decision-making procedures, in this kind of public consultation (e.g., a public hearing), the agency publicly releases a rule or decision that has already been drafted or that is in its final request of definition. It is worth clarifying that each system has its regulatory and institutional peculi- arities and that this difference (between participation in the development of standards and public hearings) may not be as clear, in some cases being confus- ing. In such cases, please make the clarifications that you may consider applicable to the case. Question 71. Quarantine rules: Prohibitions by which the directors of the entity cannot serve in a private provider within the same sector after the end of their mandate at the agency. This ban is for a fixed term. It tries to prevent the perpetration of acts of collusion and abuse of influence in the industry during their mandate as directors. Question 79B. Assessment of the agency staff: In this question we are inter- ested in identifying the mechanisms used by the agency to evaluate the perform- ance of their employees and officials. The options are three. The first one is related to the evaluation of the staff of the company on a regular basis and according to preestablished performance indicators. The second option refers to the evalua- tion of the staff of the company in a sporadic and incidental way, not obeying a 236 Airport Economics in Latin America and the Caribbean constant and regular practice of the company. The third option is the absence of any personnel evaluation. Question 84. Agency’s reporting mechanism: In this question, we would like to know how the agency is made accountable (question 51). As we stated in the question 51A, this question can also be linked with the assessment of the man- agement of the agency. The options can be a report or annual management report, the appearance of the agency directors before the Parliament, or any other mechanism you may have established for accountability purposes. Question 87. Annual report by management: Report or reports containing, in detail, the actions that took place during the year. This document is of the utmost importance as it is, in some cases, the unique instrument of accountabil- ity of the agency for the users and the rest of society. Ideally, this report should contain the goals and objectives that were proposed at the beginning of the year and the rate of success fulfilling them. Also, it should contain an account of the obstacles and challenges faced by the agency in the implementation of its policies and regulatory decisions Note 1. In this question, the term director refers to chiefs of units/departments/divisions and excludes the agency’s governing authorities (Board Members/Directors). APPENDIX C Technical Efficiency Calculation Table C.1 Results for the Technical Efficiency Scores for All Airports Other Than Latin American Airports Scale Airport IATA code CRS VRS efficiency Auckland, New Zealand AKL 0.648 0.879 0.737 Bangkok, Thailand BKK 0.935 0.951 0.983 Brisbane, Australia BNE 0.655 0.718 0.912 Guangzhou, China CAN 0.651 0.665 0.979 Jakarta, Indonesia CGK 0.854 0.867 0.985 Christchurch, New Zealand CHC 0.357 0.371 0.964 Chiang Mai, Thailand CNX 0.245 0.329 0.745 Haikou, China HAK 0.366 0.421 0.870 Hat Yai, Thailand HDY 0.134 0.208 0.645 Hong Kong SAR, China HKG 1.000 1.000 1.000 Phuket, Thailand HKT 0.393 0.528 0.743 Seoul, Republic of Korea ICN 0.962 0.962 1.000 Osaka, Japan KIX 0.743 1.000 0.743 Kuala Lumpur, Malaysia KUL 0.652 0.657 0.992 Macao SAR, China MFM 0.465 0.844 0.555 Tokyo, Japan NRT 0.860 0.876 0.982 Penang, Malaysia PEN 0.386 0.898 0.430 Shanghai, China PVG 0.909 0.931 0.976 (continued next page) 237 238 Airport Economics in Latin America and the Caribbean Table C.1 (continued) Scale Airport IATA code CRS VRS efficiency Seoul, Republic of Korea SEL 0.618 0.619 0.999 Changi, Singapore SIN 0.927 0.934 0.993 Sydney, Australia SYD 0.828 0.837 0.991 Shenzhen, China SZX 0.721 0.949 0.760 Xiamen, China XMN 1.000 1.000 1.000 Amsterdam, Netherlands AMS 0.637 0.856 0.745 Stockholm, Sweden ARN 0.410 0.453 0.905 Athens, Greece ATH 0.464 0.471 0.987 Barcelona, Spain BCN 0.731 0.763 0.959 Birmingham, United Kingdom BHX 0.365 0.367 0.994 Brussels, Belgium BRU 0.370 0.419 0.881 Bratislava, Slovak Republic BTS 0.102 0.105 0.971 Budapest, Hungary BUD 0.331 0.332 0.996 Paris, France CDG 0.826 0.922 0.896 Cologne, Germany CGN 0.299 0.388 0.771 Rome, Italy CIA 0.488 0.612 0.797 Copenhagen, Denmark CPH 0.369 0.385 0.960 Dublin, Ireland DUB 0.457 0.468 0.976 Düsseldorf, Germany DUS 0.324 0.341 0.951 Edinburgh, Scotland EDI 0.693 0.799 0.868 Rome, Italy FCO 0.517 0.585 0.885 Frankfurt, Germany FRA 0.759 0.846 0.897 Geneva, Switzerland GVA 0.418 0.443 0.953 Hamburg, Germany HAM 0.384 0.386 0.993 Helsinki, Finland HEL 0.326 0.409 0.796 Istanbul, Turkey IST 0.611 0.716 0.853 London, United Kingdom LGW 0.995 1.000 0.995 London, United Kingdom LHR 0.998 0.999 0.999 Lisbon, Portugal LIS 0.527 0.538 0.980 Ljubljana, Slovenia LJU 0.198 0.207 0.958 Madrid, Spain MAD 0.969 0.995 0.974 Manchester, United Kingdom MAN 0.488 0.498 0.981 Valletta, Malta MLA 0.144 0.144 1.000 Munich, Germany MUC 0.678 0.743 0.913 Paris, France ORY 0.556 0.570 0.974 Oslo, Norway OSL 0.526 0.539 0.975 Prague, Czech Republic PRG 0.319 0.397 0.807 Riga, Latvia RIX 0.206 0.227 0.909 Sofia, Bulgaria SOF 0.186 0.205 0.910 London, United Kingdom STN 0.911 0.970 0.940 Tallinn, Estonia TLL 0.163 0.180 0.902 Berlin, Germany TXL 0.372 0.377 0.986 (continued next page) Technical Efficiency Calculation 239 Table C.1 (continued) Scale Airport IATA code CRS VRS efficiency Vienna, Austria VIE 0.507 0.508 0.997 Warsaw, Poland WAW 0.349 0.355 0.986 Zurich, Switzerland ZRH 0.434 0.497 0.874 Albuquerque, United States ABQ 0.351 0.420 0.851 Albany, United States ALB 0.195 0.243 0.827 Atlanta, United States ATL 1.000 1.000 1.000 Austin, United States AUS 0.381 0.453 0.867 Nashville, United States BNA 0.344 0.362 0.958 Boston, United States BOS 0.401 0.572 0.703 Baltimore, United States BWI 0.402 0.484 0.835 Cleveland, United States CLE 0.318 0.397 0.800 Charlotte, United States CLT 0.793 0.835 0.949 Cincinnati, United States CVG 0.550 0.626 0.877 Washington, D.C., United States DCA 0.495 0.617 0.803 Denver, United States DEN 0.599 0.898 0.667 Dallas, United States DFW 0.596 0.840 0.711 Detroit, United States DTW 0.419 0.572 0.736 Newark, United States EWR 0.892 0.939 0.951 Ft. Lauderdale, United States FLL 0.559 0.584 0.956 Honolulu, United States HNL 0.458 0.653 0.702 Washington, D.C., United States IAD 0.510 0.553 0.935 Houston, United States IAH 0.575 0.702 0.814 Indianapolis, United States IND 0.581 0.705 0.823 Jacksonville, United States JAX 0.291 0.306 0.953 New York, United States JFK 0.973 0.973 1.000 Las Vegas, United States LAS 0.721 0.855 0.845 Los Angeles, United States LAX 0.956 1.000 0.956 New York, United States LGA 1.000 1.000 1.000 Kansas City, United States MCI 0.255 0.298 0.855 Orlando, United States MCO 0.574 0.651 0.881 Chicago, United States MDW 0.690 0.697 0.990 Memphis, United States MEM 0.996 0.999 0.998 Miami, United States MIA 0.505 0.675 0.747 Milwaukee, United States MKE 0.560 0.562 0.998 Minneapolis, United States MSP 0.556 0.617 0.906 New Orleans, United States MSY 0.245 0.247 0.993 Oakland, United States OAK 0.839 0.849 0.988 Ontario, United States ONT 0.442 0.464 0.956 Chicago, United States ORD 0.768 1.000 0.768 West Palm Beach, United States PBI 0.468 0.485 0.964 Portland, United States PDX 0.457 0.520 0.881 Philadelphia, United States PHL 0.510 0.678 0.751 (continued next page) 240 Airport Economics in Latin America and the Caribbean Table C.1 (continued) Scale Airport IATA code CRS VRS efficiency Phoenix, United States PHX 0.698 0.718 0.973 Pittsburgh, United States PIT 0.262 0.437 0.606 Raleigh, United States RDU 0.425 0.481 0.892 Richmond, United States RIC 0.306 0.314 0.978 Reno, United States RNO 0.257 0.299 0.886 San Diego, United States SAN 0.826 1.000 0.826 San Antonio, United States SAT 0.376 0.492 0.814 Louisville, United States SDF 0.970 0.971 0.999 Seattle, United States SEA 0.743 0.768 0.967 San Francisco, United States SFO 0.585 0.677 0.865 San Jose, United States SJC 0.433 0.459 0.945 Salt Lake City, United States SLC 0.439 0.677 0.649 Sacramento, United States SMF 0.402 0.441 0.912 Costa Mesa, United States SNA 0.893 1.000 0.893 St. Louis, United States STL 0.288 0.435 0.662 Tampa, United States TPA 0.451 0.506 0.893 Edmonton, Canada YEG 0.332 0.335 0.990 Halifax, Canada YHZ 0.289 0.303 0.955 Ottawa, Canada YOW 0.297 0.317 0.935 Montréal, Canada YUL 0.311 0.418 0.743 Vancouver, Canada YVR 0.510 0.634 0.804 Winnipeg, Canada YWG 0.502 0.518 0.970 Calgary, Canada YYC 0.732 0.745 0.983 Toronto, Canada YYZ 0.371 0.484 0.765 Source: Author’s estimation. Table C.2 LAC Airports Total Factor Productivity Change annual % Costa Argentina Brazil Chile Colombia Rica Year AEP EZE FTE BSB CGH GIG GRU MAO VCP SCL BAQ CLO SJO 1995–1996 - - - 9.9 8.4 4.1 11.7 −21.5 −4.4 - - - 1996–1997 - - - 23.6 20.3 9.3 5.3 −3.0 9.0 - - - 1997–1998 - - - 9.6 17.7 8.8 0.2 15.2 8.5 - −32.1 - 1998–1999 - - - −1.5 9.1 −22.7 −2.5 4.0 −8.4 - −12.7 - 1999–2000 - - - 8.2 12.9 5.6 1.2 5.9 20.0 11.8 9.2 - 2000–2001 −40.1 −24.8 - −1.5 13.1 −1.6 −5.7 −7.6 −1.7 −9.7 −27.8 - 2001–2002 −15.8 −41.9 - 10.0 3.7 −8.2 −1.4 6.2 −22.7 −10.4 −0.9 −23.7 57.0 2002–2003 2.8 22.2 - −4.4 −16.3 −16.5 2.3 −2.5 −20.0 3.7 −9.7 15.3 −5.0 2003–2004 5.9 20.0 60.3 12.1 −84.2 6.5 2.6 9.6 0.6 2.8 −2.8 −17.2 1.2 2004–2005 −2.5 −9.0 14.6 −39.0 9.0 43.7 9.4 2.0 −6.6 5.7 3.0 −5.2 −0.6 2005–2006 −9.3 5.2 3.9 −11.0 −15.4 2.3 −6.7 3.0 −18.6 −2.2 4.5 −2.6 −4.1 2006–2007 −5.5 1.9 19.7 9.2 −26.5 16.9 6.0 12.8 26.5 −15.2 1.3 5.9 3.6 241 (continued next page) 242 Table C.2 (continued) Dominican Ecuador El Salvador Mexico Panama Peru Republic Year GYE SAL CUN GDL MEX MTY PTY LIM SDQ 1995–1996 - - - - - - - - - 1996–1997 - - - - - - - - - 1997–1998 - - - - - - - - - 1998–1999 - - - - - - - - - 1999–2000 - - 18.5 - 0.8 −1.2 - - - 2000–2001 - - −3.3 - −9.9 −5.9 - - - 2001–2002 - 7.4 1.9 - 0.4 18.0 - - - 2002–2003 - −1.8 10.4 −6.1 2.2 14.1 - - - 2003–2004 - 12.4 10.2 5.1 6.0 1.7 9.0 - - 2004–2005 - −0.6 −8.2 −13.9 3.3 3.8 6.8 - −9.1 2005–2006 −28.2 −0.9 −2.1 12.3 5.5 −0.5 −20.3 9.6 −10.5 2006–2007 8.1 −4.7 −1.7 11.3 −6.9 14.2 6.3 9.8 2.0 Source: Author’s estimation. Note: Values in bold indicate the year of changes in capital stock, either in the number of runways or in the number of boarding bridges. Technical Efficiency Calculation 243 Table C.3 Average Technical Efficiency Scores and Scale Efficiency by Region (2005–06 average) Returns to scale diagnosis Technical efficiency (% of observations) World region CRS VRS Scale IRS CRS DRS Model with 3 outputs and 3 inputs (runways, staff, and boarding bridges) Latin America 0.532 0.690 0.801 70.5 9.1 20.5 Asia 0.670 0.771 0.869 84.6 12.8 2.6 Europe 0.490 0.530 0.927 43.9 6.1 50.0 Canada and United States 0.540 0.616 0.875 23.2 8.0 68.8 All 0.545 0.629 0.875 44.5 8.4 47.1 Model with 3 outputs and 2 inputs (runways and staff) Latin America 0.283 0.399 0.796 63.6 6.8 29.5 Asia 0.477 0.528 0.901 38.5 2.6 59.0 Europe 0.454 0.512 0.886 47.0 7.6 45.5 Canada and United States 0.443 0.491 0.911 36.8 5.6 57.6 All 0.425 0.487 0.885 43.8 5.8 50.4 Source: Author’s estimation. APPENDIX D Data Sources Air Transport Research Society (ATRS) The Air Transport Research Society is a nonprofit organization that gath- ers individuals from various sources to exchange research ideas and results on issues of air transportation. Specific sources include air trans- port researchers from established institutions all over the world, senior policy makers from various government organizations and think tanks, and experts from the aviation industry, ranging from airports, airlines, aerospace manufacturers, and aviation consulting services. One of ATRS’s most important outputs is its Annual Global Airport Benchmarking Report, which comprehensively assesses airport perfor- mance based on productivity, efficiency, and unit cost competitiveness data. It provides over 30 performance measures identifying effects of the operating environment of the airport, business diversification efforts, out- sourcing, and service quality. Airports are benchmarked among peer air- ports within geographical boundaries, which currently span three regions: North America, Europe, and Asia Pacific and Oceania. For this particular report, we used ATRS’s 2007 Airport Benchmarking Report, which uses 2005 data for its analysis. The results from the bench- marking report provided a basis of comparison for the performance of the Latin American and Caribbean (LAC) airports included in our sample. 245 246 Airport Economics in Latin America and the Caribbean Comparisons were made using 2005 data as reported by the airport operators surveyed for our study. Website: http://www.atrsworld.org Airports Council International (ACI) ACI is an international association of the world’s commercial service airports, which represents the interests of airport operators at interna- tional forums; develops standards and recommended practices in the areas of safety, security, and environmental initiatives; and fosters coop- eration with partners throughout the air transport industry. It includes 597 members operating over 1,679 airports in 177 countries and territo- ries. Regular members represent over 96 percent of the world’s passenger traffic and are owners or operators, other than airlines, of one or more civil airports with commercial air services. An important part of ACI’s mission is to provide members with indus- try knowledge, advice, and assistance. In achieving this, it produces a wide range of publications that address global airport policies, standards and guidelines, industry statistics, operational surveys, analytical reports, briefs, and position papers. For this particular report, we made use of ACI’s 2007 World Airport Traffic Report, which provides airport- and country-specific passenger and cargo traffic results, in addition to aircraft movement statistics. Website: http://www.airports.org Private Participation in Infrastructure Database (PPI) The Private Participation in Infrastructure Database is a joint product of the World Bank’s Infrastructure Economics and Finance Department and the Public-Private Infrastructure Advisory Facility (PPIAF). It pro- vides information on private participation in infrastructure projects in low- and middle-income countries and regions as classified by the World Bank. The projects are grouped into four sectors with some monopoly or oligopoly characteristics: energy, telecommunications, transport, and water– sewerage. More competitive sectors, such as airlines and gas production, are not included. Currently, the database contains data on more than 4,100 infrastruc- ture projects dating from 1984 to 2007. Projects must meet the following three criteria: (a) Private parties must have at least a 25 percent partici- pation in the project contract, except for divestitures, which are included Data Sources 247 with at least 5 percent of equity owned by private parties. (b) Projects must directly or indirectly serve the public; captive facilities (such as cogeneration power plants and private telecommunications networks) are excluded unless a significant share of output (20 percent) is sold to serve the public under a contract with a utility. (c) Projects must have reached financial closure after 1983 (database coverage currently extends to 2007). With over 30 fields per project record, the database details the project’s country, financial closure year, infrastructure services provided, type of private participation, technology, capacity, project loca- tion, contract duration, private sponsors, investment commitments (in the form of physical assets and payments to government), and develop- ment bank support. For purposes of this report, the PPI database was used to produce an overview of investment levels in the airport sector of developing coun- tries at both the global and LAC-specific levels. It should be noted that the analysis derived from the use of this database presents only a partial picture of investments in the airport sector for three reasons: First, given that the database is compiled through publicly available information, some projects, particularly those involving local and small-scale operators, tend to be omitted because they are usually not reported by major news sources, databases, government websites, or other sources used by the PPI projects database. Second, the database does not record important public projects such as the network of Brazilian airports operated by the state- owned company, INFRAERO. Third, with few exceptions, the investment amounts in the database represent the total investment commitments entered into by the project entity at the beginning of the project (at con- tract signature or financial closure), not the planned or executed annual investments. In addition to contributing to the overview of investment flows in the airport sector, the PPI database also proved useful in providing a general picture of the degree of private participation in the airport sectors in specific LAC countries. The latter was included within each of the case studies prepared for this benchmarking project. Website: http://ppi.worldbank.org/ Dealogic ProjectWare Database Dealogic ProjectWare is a database containing information on project and trade finance transactions since 1994 in both developing and developed nations. Collected directly from the banks and organizations involved in 248 Airport Economics in Latin America and the Caribbean the deals, the data include financial and nonfinancial information that covers projects from preapproval to signing. For this study, we made use of the Dealogic ProjectWare database to complement the overview of investment levels in the region produced through the use of the PPI database. Different from the PPI database, which records total investment commitments entered into by the project entity at the beginning of the project at contract signature or financial closure, ProjectWare presents total project amounts and their breakdown by financing sources, including shares in loans, bonds, and equity. Project amounts in the ProjectWare database reflect investments in infrastructure in the form of the construction, expansion, and refurbishment of physical assets as well as in the financing of acquisitions and the refinancing of existing debt. Any given project can consist of one or a combination of any of the above. Given that ProjectWare presents details on projects in five categories—preapproval, in tender, in finance, signed, and cancelled projects—it is important to mention that with the purpose of making the PPI and the ProjectWare data as comparable as possible, our analysis uses only those projects that have achieved financial closure and whose status was reported as “signed.� Website: http://www.dealogic.com Asociación Latinoamericana de Transporte Aéreo (ALTA) ALTA is a private, nonprofit organization composed of Latin American commercial airlines whose objective is to combine and coordinate its members’ efforts to facilitate the development of air transport in the Latin American and Caribbean region. As part of its objective to establish appropriate systems of informa- tion to be used by its members in order to promote safe and efficient air transport services in the LAC region, ALTA produces a yearly capacity analysis that is a comprehensive compendium of air transport statistics. More precisely, the ALTA analysis contains valuable information, includ- ing a ranking of the most important airports and city pairs in the region in terms of volume and growth. The 2008 analysis, more specifically, includes information on 496 airports and 1,918 city pairs throughout LAC and compares April 2008 figures with April 2007, as well as the average annual growth rates between 2000 and 2008. It identifies the top and fastest-growing airports (in terms of international, domestic, and total flights) and city pairs (in terms of available seats) across the region. Data Sources 249 For the case studies produced for this report, the 2008 LAC Capacity Analysis data served as a tool to generate a picture of flight composi- tion and growth at the regional and country level as well as at selected airports. Website: http://www.alta.aero Airport Charges The Airport Charges database is a source of published airport charges worldwide. It contains over 2,000 charges documents covering airports on every continent and allows for comparison between regions, countries, and airports. Information contained within the database includes up-to- date fuel prices for every airport and historic evolution of charges dating back to 2005. Website: http://www.airportcharges.com Historically, airports in Latin America were exclusively owned and managed by the government. Since the late 1990s, several Latin American countries have implemented innovative public-private partnerships that transferred the financing and management of airports to the private sector. Airport Economics in Latin America and the Caribbean presents a comprehensive study of how the airports in the region have evolved during this notable period of transition. The book presents a positive and unbiased analysis of events, rather than a normative analysis of what should be done to reform the airport sector or to attract private participation. It focuses on three dimensions of performance: productive efficiency, the governance of the sector, and financing of airport infrastructure. Airport Economics in Latin America and the Caribbean responds to the need for more conclusive information about the influence of airport ownership on economic perform- ance. Using rigorous analytical tools, this book evaluates the introduction of private sector participation in the Latin American airport sector and answers several policy questions, including: Are Latin American airports technically efficient? How has efficiency evolved in the last decade? Are privately run airports more efficient than state-operated airports? How do independent regulators compare with government agencies in accountability, transparency, and autonomy? How has the level and structure of airport tariffs changed in recent years? This book will be of interest to air transport practitioners, transport regulators, decision makers in transport ministries, airport investors, and academics. ISBN 978-0-8213-8977-5 SKU 18977