LIGHTING BRAZILIAN CITIES: BUSINESS MODELS FOR ENERGY EFFICIENT PUBLIC STREET LIGHTING Authors: Megan Meyer Luiz Maurer Javier Freire Christophe de Gouvello 3 LIGHTING BRAZILIAN CITIES: BUSINESS MODELS FOR ENERGY EFFICIENT PUBLIC STREET LIGHTING Authors: Technical staff: Megan Meyer Pezco Consultoria Luiz Maurer Clara Ramalho (Lumina Consultoria) Javier Freire Castagnari Consultoria Ltda. Christophe de Gouvello Pezco Consultoria Washington, 1st edition, April 1, 2017 Copyright © April 2017 assistance program administered by the World Bank – assists low- and middle-income countries to increase their know- All rights reserved how and institutional capacity to achieve environmentally This document was produced by the staff of the International sustainable energy solutions for poverty reduction and economic Bank for Reconstruction and Development World Bank. The growth.  ESMAP is funded by Australia, Austria, Denmark, the findings, interpretations and conclusions expressed herein do European Commission, Finland, France, Germany, Iceland, not necessarily reflect the views of the World Bank Board of Japan, Lithuania, Luxembourg, the Netherlands, Norway, Executive Directors or the countries they represent. the Rockefeller Foundation, Sweden, Switzerland, the United Kingdom, and the World Bank. The World Bank does not guarantee the accuracy of the data herein and does not accept any liability arising from the use BRAZEEC of this data. The boundaries, colors, denominations and other The Energy Efficiency Program for Brazilian Cities (BRAZEEC) information shown on any map in this document do not imply is an initiative financed by ESMAP aimed at promoting energy any judgment on the part of the World Bank concerning the efficiency (EE) actions through the identification of feasible legal status of any territory or the endorsement or acceptance and replicable business models in Brazilian cities in the areas of such boundaries. of public lighting, public buildings, urban industries and Rights and Permissions transportation. The program includes a focus on institutional frameworks to strengthen energy efficiency management at the The material in this document is protected by copyright and municipal level. total or partial reproduction of the text without previous consent shall be understood as an infringement of applicable law. The The program´s overall objectives are to: (i) prepare business International Bank for Reconstruction and Development / World models to improve EE in the four above- mentioned areas that Bank encourages the dissemination of its work and permits the can be implemented by various Brazilian cities; (ii) enhance reproduction of parts of the same providing the source is quoted. the capacities of stakeholders (municipal and private) to enable them to implement EE business models; and (iii) create For permission to photocopy or reprint any or part of this a demonstration effect that, in the long term, will expand EE document, please send a request with full information to the initiatives to other city sectors. Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, U.S.A. Telephone: 978-750-8400, fax: 978-750-4470; Internet: www.copyright.com. Energy Sector Management Assistance Program The World Bank All queries regarding rights and licenses, including subsidiary 1818 H Street, NW rights, should be addressed to the Office of the Publisher, The Washington, DC 20433 USA World Bank, 1818 H Street, NW, Washington, DC, 20433, E.U.A., email: esmap@worldbank.org fax: 202-522-2422 e-mail: pubrights@worldbank.org. web: www.esmap.org ESMAP The financial and technical support by the Energy Sector Graphic Project and Layout Management Assistance Program (ESMAP) is gratefully Carlos Peliceli e Ione Flor carlospeliceli@gmail.com acknowledged. ESMAP – a global knowledge and technical 2 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table of Contents FOREWORD XI Acknowledgements XIII Executive Summary XV Introduction XV Approach to the Report XVI Overview of Global Trends in LED Public Street Lighting XVII The Brazilian Market: Overview, Opportunities and Challenges XVIII Overview XVIII Opportunities XIX Challenges XX Mapping Brazil’s Municipalities for Public Lighting XXIV Business Models for Public Street Lighting in Brazil XXVII Financing Mechanisms and Improving Credit Risk XXVIII Gaps in the Market, and Recommendations. XXXI Conclusions XXXIII 1 - Introduction 1 2 - Overview of Trends in Public Street Lighting using LEDs. 5 2.1 - LED Technology 5 2.2 - Economic and Financial Benefits of LEDs 6 2.3 - Other Benefits Associated with the Improvement of Street Lighting Systems with LEDs27 9 2.4 - International Examples 10 3 - Public Street Lighting Market in Brazil 13 3.1 - General Aspects 13 3.1.1 - Coverage of the public street lighting service 13 3.1.2 - Installed technology 15 3.2 - Bidding Schemes for Public Street Lighting 17 3.3 - Opportunities for LEDs in the Brazilian context 20 3.3.1 - Opportunity 1 - Energy and equipment prices moving in opposite directions 20 3.3.2 - Opportunity 2- Incentives for municipalities to invest in their lighting assets 22 3.3.3 - Opportunity 3 - Source of specific funds to pay for public lighting services 24 3.3.4 - Opportunity 4 - Alignment with the public climate policy 28 III Table of Contents 3.4 - Challenges for LEDs in the Brazilian Context 29 3.4.1 - Challenge 1 - Large scale funding required 29 3.4.2 - Challenge 2 - Large diversity of municipalities 29 3.4.3 - Challenge 3 - The macroeconomic situation 30 3.4.4 - Challenge 4 - Restricted public credit lines 31 3.4.5 - Challenge 5 - Restrictions on municipal indebtedness 33 3.4.6 - Challenge 6 - Credit risk and municipal political risk 35 3.4.7 - Challenge 7 - Lack of a clear regulatory framework for public lighting 35 4 - Mapping Brazilian Municipalities for Public Lighting 39 4.1 - Survey of Public Street Lighting 39 4.1.1 - Data collection 39 4.2 - Formation of Clusters 40 4.3 - Homogenous Groups 42 5 - Business Models for Public Street Lighting in Brazil 45 5.1 - Overview of the Primary Functions Considered in Business Models 45 5.2 - Introduction to Business Models 46 5.3 - Presentation of Business Models 47 5.4 - Summary of Business Models and Mapping of Groups of Municipalities 74 6 - Financing and Credit Enhancement Mechanisms 79 6.1 - Financing Mechanisms 79 6.2 - Risks and Mitigation Mechanisms 88 7 - Other Considerations in the Design of a LED Project 91 7.1 - Automation, Remote Management and Ancillary Services 91 7.2 - Modernization of the Public Street Lighting System in a City Infrastructure 93 8 - Gaps in the Market and Recommendations 97 9 - Conclusions 105 ANNEXES 109 ANNEX 1 – Summary of Business Models 109 ANNEX 2 - S / P Correction between Scotopic Vision (S) and Photopic Vision (P) 124 ANNEX 3 - Methodology for Generating Clusters and Grouping of Municipalities 125 ANNEX 4 - International Standards and Specifications for High Performance Luminaires 141 REFERENCES 142 IV Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning List of Figures Figure 1 – Largest public lighting LED retrofit projects 3 Figure 2 – Technologies in the National Public Lighting System in 2012, by % 4 Figure 3 – Largest public lighting LED retrofit projects 32 Figure 4 – Coverage of public lighting for % of households, by household income 36 Figure 5 – Concentration of light points in Brazil 37 Figure 6 – Percentages of municipalities familiar with LEDs in public street lighting 38 Figure 7 – Responsibility for maintenance of public street lighting (% of municipalities) 40 Figure 8 – Average electric energy tariff for public street lighting, % variation 42 Figure 9 – Average price of LEDs (US$ per 1,000) 43 Figure 10 – Most recent municipalities to transfer their public street lighting assets 45 Figure 11 – COSIP contribution to public lighting, by % of municipalities 46 Figure 12 – Sufficiency of COSIP to cover PL costs, by % of municipalities 47 Figure 13 – Basic annualized interest rate (SELIC) in % 52 Figure 14 – Exchange-rate (Brazilian R$/US dollar) 53 Figure 15 – Example of the structure of a PPP model 70 Figure 16 – Example of the structure of a PPP with municipal consortium model 75 Figure 17 – Example of structure of the municipal financing model 78 Figure 18 – Example of structure of electricity utilities program model 82 Figure 19 – Example of structure of the ESCO model 85 Figure 20 – Example of structure of centralized procurement model 88 Figure 21 – Example of structure of self-funding model 91 Figure 22 – Example of structure of luminaire transfer model 94 Figure 23 – Example of the use of World Bank guarantees in the PPP model 105 Figure 24 – Example of the structure of the PPP model 131 Figure 25 – Example of use of guarantees in the PPP model 133 Figure 26 – Example of the structure of the PPP/Municipal Consortium model 134 Figure 27 – Example of the structure of municipal financing model 135 Figure 28 – Example of structure of electricity utilities program model 137 Figure 29 – Example of structure with ESCO model 138 Figure 30 – Example of structure of centralized contracting model 140 V List of Figures Figure 31 – Example of structure of self-funding model 142 Figure 32 – Example of structure of luminaire transfer model 144 Figure 33 – Schematic diagram of cluster analysis 155 Figure 34 – Evolution of pseudo-F statistics for different amounts of groups 156 Figure 35 – Evolution of CCC statistic for different amounts of groups 157 Figure 35 – Dispersion diagram of the clusters averages of the variables GDP and consumer units per connection 158 Figure 36 – Diagram of dispersion of the averages of the clusters relative to the NCD/NCR variables 158 VI Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning List of Tables Table 1 – Database characteristics 11 Table 2 – Main characteristics of the groups 12 Table 3 – Municipal groups (statistics) 12 Table 4 – Summary of business models 13 Table 5 – Mapping business models tailored to each group 14 Table 6 - Mapping financial instruments tailored to business models 15 Table 7 – Mapping financial instruments tailored to business models 16 Table 8 – Prioritization of recommendations and indicative implementation timeframe 17 Table 9 – Summary of LED features compared with other technologies 27 Table 10 – Summary of results of the pre-feasibility study for the city of Rio de Janeiro 30 Table 11 – Characteristics of the vicinity of Brazilian homes - selected public services 35 Table 12 – Number of lamps in the National Public Lighting System, by % 38 Table 13 – Bidding regimes for energy efficiency projects in public lighting systems 41 Table 14 – Recommendations for COSIP design 48 Table 15 – Examples of COSIP design for São Paulo and Belo Horizonte 50 Table 16 – Main restrictions on credit operations imposed by the Fiscal Responsibility Law 56 Table 17 – Municipalities surveyed, by region 62 Table 18 – Database characteristics 63 Table 19 – Grouping of clusters according to key characteristics 64 Table 20 – Municipal groups in order of relevance 65 Table 21 – Functions and key stakeholders for public street lighting projects 67 Table 22 – Key stakeholders applicable to all the business models 68 Table 23 – Matrix of functions and actors in the M1 — municipal PPP model 71 Table 24 – PPP projects in Brazil (according to project stage) 73 Table 25 – Matrix of functions and actors in Model M2 — PPP consortia 75 Table 26 – Matrix of functions and actors in the M3 model — municipal financing 79 Table 27 – Matrix of the functions and actors in model M4 — electric utility companies’ programs 83 Table 28 – Matrix of functions and actors in model M5 — ESCOs 86 Table 29 – Matrix of functions and actors in model M6: municipal consortium or central procurement agent 89 Table 30 – Matrix of functions and actors in model M7 – self-financing 92 Table 31 – Matrix of functions and actors in Model 8 — transfer of HPS luminaires 95 Table 32 – Key stakeholders in each phase of the business models 97 VII List of Tables Table 33 – Business models for each group 98 Table 34 – Details of financial instruments of development banks 106 Table 35 – Financial instruments suitable for use in the different business models 109 Table 36 – Other risk mitigation mechanisms 110 Table 37 – Automation levels of public street lighting projects 114 Table 38 – Levels of modernization of the public street lighting infrastructure 116 Table 39 – Gaps, barriers, and recommendations 120 Table 40 – Prioritization of recommendations and indicative implementation timeframe 123 Table 41 – S / P correction factor between scotopic vision (S) and photopic vision (P) 146 Table 42 – Precision index by region 149 Table 43 – Database characteristics 150 Table 44 – Matrix of cross-correlations of selected variables 151 Table 45 – Relevance of variables in the formation of groups 157 Table 46 – Average variables in each group 159 Table 47 – Classification of clusters in each variable 160 Table 48 – Criteria for assigning score for each variable 161 Table 49 – Regrouping of clusters according to main characteristics 161 Table 50 – Relevance of groups of municipalities, absolute numbers 162 Table 51 – Relevance of groups of municipalities,% 162 VIII Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Abbreviations ABDI - Brazilian Industrial Development Agency ABGF - Brazilian Guarantees and Fund Management Agency AFD - French Development Agency AgeRio - Rio de Janeiro State Development Agency ANEEL - Brazil’s National Electricity Regulatory Agency BB - Bank of Brazil BDMG - Minas Gerais Development Bank IDB - Inter-American Development Bank BNDES - Brazilian National Bank for Economic and Social Development BNDES/Finem - business financing sector of BNDES BNDES/PMAT - Fund for the Modernization of Tax Administration and Basic Social Sectors BRL - Brazilian Real (Brazilian Real/R$) CAPEX - Capital expenditure CEF - Caixa Econômica Federal (Federal Mortgage Bank) CEMAB - Maciço de Baturité Intermunicipal Consortium of Energy/Public Lighting CEMIG - Minas Gerais Electric Energy Company CIDES/MG - Intermunicipal Public Consortium for the Sustainable Development of the Triângulo Mineiro and Alto Paranaíba CIGIP/AL - Alagoas Public Consortium for Electric Energy Management and Public Services CIP and/or COSIP - street lighting charge (on consumers bills) CONIAPE/PE - Inter-Municipal Consortium of Pernambuco Hinterland and Borders COP-21 - 21st Conference of the Parties CRI - Color Reproduction Index DCL - Net Current Debt (NCD) DISCO - Electricity distribution company DRC - Differentiated Contracting Regime EE - Energy Efficiency ESCO - Energy service company ESMAP - Energy Sector Management Assistance Program FGIE - Infrastructure Guarantee Fund FGTS - Brazilian Employees´ Guarantee Fund FI-FGTS - FGTS Investment Fund FIDC - Credit Rights Investment Fund FIP - Equity Investment Fund FIR - Referenced Investment Fund FPM - Municipal Participation Fund G-20 - Group of 20 GHG - Greenhouse gases IX Glossary of Acronyms and Abbreviations GWh - Gigawatt hour HPS - High pressure sodium lamps ICMS - Tax on the Circulation of Goods and Services IFC - International Finance Corporation IFGF - Firjan Fiscal Management Index MFIs - Multilateral Financial Institutions INDC - Intended Nationally Determined Contributions INMETRO - Brazilian National Institute of Metrology, Quality and Technology IPI - Industrialized Products Tax IR - Income Tax LED - Light Emitting Diode MDIC - Ministry of Development, Industry and Commerce MIGA - Multilateral Investment Guarantee Agency MME - Ministry of Mines and Energy NCD - Net Current Debt NCR - Net Current Revenue NDC - Nationally determined contributions NGO - Nongovernmental organization OPEX - Operating expenditure R & D - Research & Development PEE - Energy Efficiency Program GDP - Gross Domestic Product PMI - Expression of Interest Procedure PPIAF - Public-private infrastructure advisory facility PPP - Public-private partnership PR - Reference equity PRG - Partial risk guarantee PROCEL - National Program for the Conservation of Electric Energy PROCEL-RELUZ - National Program of Public Lighting and Traffic Signals Efficiency PROESCO - Energy Efficiency Program PSL - Public Street Lighting R20 - Regions of Climate Action RAB - Regulatory Asset Base DRC - Differentiated Contracting Regime RGR - Reversal Global Reserve SNIS - National Sanitation Information System SPE or SPV - Special Purpose Vehicle SSL - Solid-State Lighting TJLP - Long-Term Interest Rate TWh - Terawatt-hour WBG - World Bank Group X Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning FOREWORD The public street lighting sector in Brazil has reached a turning point. LED technology is now used on a large scale by several cities throughout the world and offers significant savings in energy and maintenance costs. Since late December 2014, all Brazil´s cities have had full responsibility for their street lighting systems, and thus have more incentives to reduce the rising energy costs involved. However, converting to LEDs is capital intensive, which leads municipalities to raise questions such as: (i) is this technology proven and will it attain its forecast lifecycle and produce the expected energy savings?; (ii) how do street lighting investments fit in with municipalities other social priorities?; and (iii) what are the options to structure and finance a substantial modernization project of the municipality´s street lighting system? To seek answers to these and other key questions, the World Bank Group, as part of its ‘green agenda’, examined the benefits, risks and financing models of public street lighting projects in Brazil based on LED technology. We began this task in late 2013 by providing technical support to the cities of Rio de Janeiro and Belo Horizonte to identify areas within the municipal sphere with energy efficiency potential. After identifying public street lighting as one of the most promising areas in these two cities, our study evolved into an analysis of possible financing mechanisms, concluding that public-private partnerships (PPPs) through administrative concession could deliver a number of advantages to large cities. Given Brazil´s 5,570 highly diversified municipalities in terms of needs, per capita income, technical knowledge and access to finance, the World Bank Group perceived that the PPP model, although appealing to a privileged group of large and medium-sized cities, would be a struggle for smaller cities or those with weaker credit profiles. With this in mind, we launched the study reported here with the aim of identifying business and financing models that could be applicable to a wider range of Brazil’s municipalities, and lead to all of them benefiting from the improved street lighting services and energy savings that LED technology can offer. This report by Meyer, de Gouvello, Freire and Maurer is a first effort to give practical meaning to the above questions. We hope that it will prove useful to government officials, mayors, financial institutions and equipment manufacturers, and others. At the same time, we wish to emphasize the need to define a clear strategy for Brazil based on joint, well-coordinated efforts by the different actors, agencies and government spheres to ensure support for municipalities to achieve these goals. The World Bank Group reaffirms its ongoing interest in working with these stakeholders in the search for solutions to meet the needs of Brazil’s diverse regions and municipalities. Antonio Barbalho Manager Energy and Extractive Industries Global Practice Latin America and the Caribbean XI Foreword XII Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning ACKNOWLEDGEMENTS This report was drafted by Megan Meyer, Javier Freire, Christophe de Gouvello (World Bank) and Luiz Maurer (IFC), based on analyses by Pezco Consultoria and under the leadership of Antonio Barbalho. It is the revised version of the paper distributed at the Forum on Business Models for Energy Efficient Public Lighting that took place in São Paulo on June 1, 2016, and incorporates feedback from the Forum participants. The report has benefited from substantial contributions by Peter Curley (The Climate Group), Marcel de Costa Siqueira (ELETROBRÁS), Paulo Oliveira (BNDES), Fernando Camacho (BNDES), Fernando de Paiva Pieroni (São Paulo Negócios), Maria Eduarda Berto (Estruturadora Brasileira de Projetos), Clara Ramalho (Lumina Consultoria), Castagnari Consultoria Ltda, Hector Gomez (IFC), Tomas Anker (IFC), Ivan Jaques (World Bank), Peter Mockel (IFC), and Ashok Sarkar (World Bank). The Pezco Consultoria Ltda team that helped prepare this report comprised Frederico Araújo Turolla, Luís Fernando Rigato Vasconcellos, Helcio Shiguenori Takeda, Jorge Luiz Dietrich, Márcio Fernandes Gabriell, Jorge Oliveira Pires, and André Yoshizumi Gomes. This work was carried out thanks to the support provided by the World Bank’s Energy Sector Management Assistance Program (ESMAP) under the Brazil Energy Efficient Cities Program (BRAZEEC). The authors also wish to thank the organizations and other bodies that shared their knowledge and opinions during the preparation of this study. Errors or omissions are the exclusive responsibility of the authors. XIII Acknowledgements XIV Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Executive Summary Introduction technology are emerging as a technically and commercially viable option for investments in energy efficiency in the public street lighting sector. LED C ities are among the world´s biggest consumers lamps are between 40 60% more efficient in terms of electric energy, accountable for two thirds of of energy consumption than the current lamps total electricity consumption and for over 70% used throughout Brazil, and they offer substantial of global greenhouse gases emissions. Public lighting reductions in operational and maintenance costs. It is systems contribute significantly to a city´s energy also possible to integrate efficient LED infrastructure consumption. In Brazil, for example, public street with smart monitoring systems, thereby creating the lighting accounts for more than 4% of the country’s backbone for a “smart city”. Furthermore, the higher total energy consumption. Moreover, the cost of quality of the new lighting technology can have energy for street lighting represents the second most the positive effect of reducing crime and improving expensive item of most municipalities’ budgets, communities’ perception of security. surpassed only by payroll expenditures. Thus, it is clear that energy efficiency projects in the public Although more energy efficient and cheaper over their lighting sector can play an important role in energy lifetime, LEDs are considerably more capital intensive conservation, as well as in reducing emissions and than the existing lighting technologies predominantly alleviating a burden on local authority budgets. used in Brazil. Municipalities are currently studying the possibility of attracting private capital - for example At the global level, member states of the 21st in the form of PPPs - to carry out retrofit projects, Conference of Parties (COP-21), held in November thus relieving local authorities of the burden of using 2015, submitted their proposals (Intended Nationally their limited budgets to finance the up-front cost of Determined Contributions – INDCs) for addressing LEDs. PPP projects are a realistic alternative for large climate change. Among other goals, Brazil committed and medium-sized cities with good credit standing; to increasing energy efficiency in the electricity sector however, the majority of Brazil’s municipalities are by 10% by 2030. Up to one-fifth of this target could small and/or face challenging fiscal situations. Given be reached by introducing more energy efficient the wide diversity of the country´s municipalities, technologies across the Brazilian municipal street business and financing models are needed that will lighting sector. enable a universal modernization of the Brazilian public street lighting sector. Lamps using the new LED (Light-Emitting Diode) XV 1. Executive Summary This report is focused on identifying business 2. Develop a methodology to group and financing models, which – by taking into municipalities: based on this data collected in account institutional environment and market the field, combined with other publically available characteristics – will enable sustainable retrofit socioeconomic data, the team developed a projects to be implemented in the public lighting methodology to group municipalities according sector throughout Brazil in the near-term, to the to characteristics that were presumed to be the benefit of municipal governments, communities, major drivers of business model development and the environment. for investments in energy efficient public street lighting. Six groups of cities were ultimately Approach to identified, representing all 5,570 Brazilian municipalities. the Report 3. Conduct a market review: identify the main opportunities and challenges within the existing regulatory, institutional and financial-economic The origin of this study dates back to 2013, when the framework of the public street lighting sector. World Bank provided technical support to the Rio de Janeiro and Belo Horizonte municipal government 4. Identification of business models: after authorities aimed at identifying the areas within those categorizing cities into six groups, the team municipalities with the greatest energy efficiency identified business models that could meet the potential. As a result, public lighting was identified needs and capacity of the six different groups as one of the most promising areas, and in 2014 pre- of Brazilian municipalities. Once eight business feasibility studies were carried out to identify projects models had been identified, the team identified for modernizing the public street lighting systems in which of the business models would be best suited the two cities1. to meet the needs of the six groups of cities. 5. Identification of financing mechanisms: After confirming the huge potential of this sector the next step was to identify financing and in Brazil, in early 2015, the World Bank decided to credit enhancement mechanisms that could be embark on a broader study to better understand the used to implement the business models. Ten wider context of the country´s public street lighting mechanisms are identified in this report, and sector. This report is the product of that work. The each are evaluated vis-à-vis their applicability to process to develop this report was conducted with the eight business models. the following approach: 6. Identification of other considerations, gaps 1. Conduct a detailed survey of the Brazilian and recommendations for next steps: the team market: this included field visits to (i) nine reviewed other important considerations – aside municipalities and consortia, and (ii) electric from the business and financing models - that utilities in three states (Ceará, Minas Gerais and municipalities may need to consider when designing São Paulo). Using information gathered in the their public street lighting program (e.g., scope of field, the team prepared an electronic survey and modernization). Finally, the team identified the key used this tool to gather street lighting data from barriers standing in the way of a full-scale rollout a statistically significant sample of more than of LEDs in Brazil, as well as recommendations for 300 municipalities across all the Brazilian states. how these barriers can be overcome. 1 In the course of preparing this report, the World Bank also provided support to personnel in the City of São Paulo engaged in preparing the tender for the modernization of that city´s public street lighting service. XVI Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Overview of addition to energy savings and reduced maintenance costs, LED lighting can result in a wide range of Global Trends other socioeconomic benefits, such as improved lighting quality, enhanced security and increased in LED Public local economic activity. The adoption of “smart” control systems allows greater flexibility in terms of Street Lighting delivering lighting options that are focused on the daily needs of citizens. Projects and trials of LED technology for public street Cities around the world of varying size, geographic lighting have proven to deliver energy savings of location, climate, etc., have begun implementing 40–70%, and even as much as 80% when combined LED retrofit projects, as shown in Figure 1. with “smart” management and control systems2. In Figure 1 – Largest public lighting LED retrofit projects3 LED Points British Columbia (Province), Canada New York, USA Belo Horizonte, Brazil Los Angeles, USA Houston, USA Guangdong (Province), China Manitoba, Canada Kent, UK Chennai, India Ontario (Province), Canada Birmingham, UK Buenos Aires, Argentina Madrid, Spain Guangzhou, China Jaipur, India Glasgow, UK Sheffield, UK Detroit, USA Oslo, Norway Durham, UK Portland, USA Foshan, China Missouga, Canada Las Vegas, USA Seattle, USA London, UK Boston, USA 0 40.000 80.000 120.000 160.0000 200.000 240.000 Source: The Climate Group, World Bank. 2 For example, the Climate Group, and international NGO focused on - inter alia - promoting energy efficient public street lighting, conducted a study in 2012 in 12 municipalities, showing 50–70% savings. The city of Los Angeles has seen savings of >60% in its city-wide LED street lighting program. 3 Projects at different planning/implementation stages. XVII 1. Executive Summary Further improving the attractiveness of LEDs is the Overview fact that prices are rapidly falling (by around 10% annually), as the equipment is benefiting from rapid Public street lighting in Brazil is estimated to consist technological innovation and economies of scale. of over 18 million light points, with a penetration rate of around 95.5% of households. The distribution These benefits make a LED technology revolution of installed lighting in Brazil relates directly to virtually inevitable; however, the pace of scaling up demographic concentration: lighting points are public street lighting to LEDs will depend on cities primarily concentrated in large cities and along the having the appropriate financing mechanisms in Southeast and Northeast Atlantic Coast. place and the political will to prioritize these projects. The installed lighting inventory consists mainly of The Brazilian high-pressure sodium lamps and, to a lesser degree, mercury vapor lamps. LED technology penetration is Market: Overview, very low, although several cities are currently rolling out pilot projects to deploy this technology. Figure 2 Opportunities shows a massive concentration of HPS and mercury vapor technologies in the Brazilian street lighting and Challenges system as of 2012. The proportion of LED technology is minimal by comparison. Figure 2 – Technologies in the National Public Lighting System in 2012, by % Mercury vapor 23,6% Others * 5,3% LEDs 0,1% HPS 71,0% *Multi-vapor metallic, halogen incandescent lamps, amongst others Brazil´s public street lighting market presents good present some of the more promising opportunities and opportunities for LED retrofits. However, the process is difficult challenges in the Brazilian market. not without its drawbacks. The following two sections XVIII Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Opportunities Opportunity 1: High cost of energy and decreasing costs of LED technology The current high energy costs in Brazil, combined with the declining cost of LED technology, translate directly into a very substantial economic and financial opportunity for Brazil’s municipalities. The average public street lighting energy tariff increased by 38.8% in 2015, following an already substantial increase of 10.9% during the previous year. The trend continues towards additional, albeit smaller, increases in the coming years. The high cost of electricity means that municipalities have strong incentive to invest in technologies that reduce their energy consumption. At the same time, prices of LEDs are rapidly falling. From 2007 to 2012, the cost per lumen decreased nine times, from US$32 to US$3.45 for every 1000 lumens produced. This reduction was partly due to the increased efficiency of LEDs, which almost doubled from 70 lumens per watt to 130 lumens per watt. Opportunity 2: Incentives for municipalities to invest in their assets A recent regulatory change has had a significant impact on Brazil’s public street lighting sector: in 2013, ANEEL (National Electricity Regulatory Agency) issued a resolution that all public lighting assets previously in the hands of electricity utilities were to be transferred to municipalities by the end of 2014. This change affected around 42% of Brazil’s municipalities, as the remaining 58% of municipalities already owned their street lighting assets. Now that this deadline has passed, all Brazilian municipalities own their public street lighting assets and are responsible for delivering public street lighting services. The transfer of assets is expected to initially lead to additional costs for the municipalities recently receiving these assets. However, on the positive side, the fact they now own their street lighting assets means that they now have a strong incentive to invest in energy efficient equipment, as they will benefit directly from the resulting reduction in energy costs of their investment. Another important incentive for municipalities is the potential political benefit that LED street lighting can bring. In the environment of the ongoing macro-fiscal crisis, the mayors of Brazilian cities have few other investment opportunities that can be implemented as quickly and offer immediate benefits to citizens’ quality of life, at a negative cost to municipalities over the life of the project. These incentives create a critical mass of municipalities seeking solutions to modernize their public street lighting. This also creates a large potential market for equipment manufacturers, public lighting installation firms, and O&M firms, among others. XIX 1. Executive Summary Opportunity 3: Ring-fenced funding to pay for public lighting A constitutional amendment of December 2002 allowed municipalities and the Federal District to collect for the ”Contribution to Public Lighting Costs“ (CIP or COSIP, hereafter called COSIP) through customers´ itemized bills. This ring fenced resource for covering the cost of public street lighting must be used to pay for the energy, maintenance, installation and modernization of public lighting equipment. This beneficial structure is not seen in most other municipal sectors, making investments in this sector attractive relative to other municipal infrastructure investments. A World Bank sample survey of 300 municipalities reveals that 81.6% now collect COSIP, and many of the remaining municipalities are currently preparing legislation to introduce this charge. The existence of COSIP means that a guaranteed resource exists to pay for public street lighting improvements. The funds could be used, for example, as collateral in business models that include financing (municipal loan repayments, municipal payments to PPP concessionaires, etc.). Opportunity 4: Alignment with public climate policies As mentioned in the introduction to this report, member countries participating in the COP 21— including Brazil—submitted their INDCs to the United Nations Framework Convention on Climate Change (UNFCCC). Brazil committed inter alia to make efficiency gains in its energy sector of 10% by 2030. Modernization of Brazil´s public street lighting sector can contribute significantly towards reaching this goal, since the sector currently represents 4% of the country´s total energy consumption. Moreover, a national conversion to LED lighting could produce, at a conservative estimate, energy savings of 50% in the public street lighting sector, thus translating into achieving one-fifth of its INDC by exploiting this potential if a national LED conversion is completed (i.e., national efficiency gains of 2%). Challenges modernization of the public street lighting sector in Brazilian cities still faces many challenges. Despite the substantial opportunities, These include: Challenge 1: Large scale of financing required LED public lighting projects can offer substantial economic benefits to Brazilian cities, since LEDs are more efficient and have a much longer life than existing technologies; however, they are significantly more capital-intensive. The estimated cost of converting the entire 18 million lighting points in Brazil could be as much as R$28 billion (US$8.6 billion ). XX Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning This means that a significant amount of capital will need to be mobilized for all Brazilian cities to benefit from the advantages of LED technology. Given that municipal budgets are constrained, city-wide LED projects will only likely move forward if municipalities and/or other agents can access financing at viable interest rates to cover these up-front costs. A further issue is that the scale of investment required is beyond the reach of purely public resources, meaning that the private sector will need incentives to become significantly involved in modernizing the Brazilian public street lighting network. Challenge 2: Wide diversity of municipalities The challenge presented by the scale of the projects is more than financial. The widely differing characteristics of Brazil´s 5,570 municipalities will require the development of a number of different types of projects tailored to these characteristics. All the projects will need well-designed business models that meet the needs of individual municipalities. At the same time, municipal personnel will need training to improve project management and/or oversight capacity. The heterogeneity of Brazilian municipalities in terms of size, capacities and fiscal situation complicates the task of finding solutions for large-scale implementation. Therefore, a key challenge is to develop business models to meet the varied needs of Brazilian municipalities, while simultaneously offering the necessary incentives to attract private sector participation. Challenge 3 – Macroeconomic scenario Brazil’s macroeconomic scenario over recent years has precipitated an unfavorable environment for infrastructure investments in general. Some factors are particularly relevant in the case of public lighting projects. First, borrowing costs, which were already high, have increased further. For example, the prime interest rate (SELIC) was 14.25% for most of 2015 and 2016—double that of early 2013. Second, the depreciation of the Brazilian Real in 2015 has led to higher costs of imported equipment—a significant drawback given that LED luminaires are not manufactured in Brazil. This volatility in the Brazilian economy increases the perception of risk by potential international investors. XXI 1. Executive Summary Challenge 4 – Limited options for public funding In view of the financing requirements and Brazil´s current high interest rates, public funding is important for cities (or their agents) seeking viable financing sources for public street lighting projects. At the federal level, the National Program for Efficient Public Lighting and Traffic Light Signaling (PROCEL-RELUZ), managed by ELETROBRAS, has been the most important historical source of investment for the sector. Another national program, the Energy Efficiency Program (PEE), administered by ANEEL, has also provided funding for investments in the public street lighting sector, albeit on a much smaller scale. At present, these two programs are not providing significant funding for investing in energy efficiency in the public street lighting system. Recent changes in legislation— particularly Law 13,280 of May 3, 2016, which allocates 20% of the PEE resources to PROCEL—may enable these programs to play more important roles. Nonetheless, given the amount of capital required for converting the public lighting network to LEDs, other sources of financing will still be needed. While multinational or bilateral bodies can fill part of the financing gap, it will be essential to leverage domestic and/or international private sector financing to develop the LED public street lighting market potential. Challenge 5 – Restrictions on municipal indebtedness Although the amount of financing needed by the public sector for modernizing the public street lighting network is significant, there are major funding restrictions on Brazilian municipalities arising from Brazil’s Fiscal Responsibility Law, which imposes on municipalities an indebtedness ceiling of 16% of Net Current Revenue (NCR). There are some exceptions to this law, such as the financing provided by multilateral bodies, federal credit institutions or development banks, providing these resources are used for investments in projects to improve tax administration or the management of financial and patrimonial assets. This exception is also applicable to PROCEL-RELUZ transactions for public lighting projects, presumably for the same rational that, as with the funds earmarked for modernizing the tax structure, investments in energy efficiency should improve a municipality´s fiscal situation by reducing its expenditure over time. However, given the relatively small resources available for PROCEL, this exception is unlikely to have a material impact across the national public street lighting sector. These restrictions will make it difficult for many municipalities to access financing directly, making it important to identify business models that can offer municipalities “off-balance sheet” solutions for financing LED street lighting projects. XXII Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Challenge 6 – Municipal credit risk Municipal credit risk is one of the toughest challenges, not only in Brazil, but in municipal financing projects throughout the world. Since Brazilian municipalities own their public lighting assets, they are now ultimately responsible for generating the cash flows that will be used to pay investment-related costs. This exposes investors in public street lighting projects to municipal, political and credit risks. As previously mentioned, COSIP provides a ring-fenced funding source to be used exclusively for public lighting. However, the mere existence of COSIP does not entirely rule out project risk. For example, COSIP revenues could still be undermined if they are frozen or reduced by oversight/ control agencies, or if COSIP collection is not sufficient to cover the investment costs throughout the investment cycle. According to the research data collected from 300 municipalities, 44.1% of those surveyed indicated that COSIP was considered to be sufficient to cover municipal expenditure on public lighting, while 31.3% considered that it was insufficient, and 24.6% were unable to assess whether COSIP was sufficient or not. It follows that the municipal law defining the COSIP charge must be well formulated. Recommendations regarding COSIP formulation are presented in the full report. Depending on a municipality´s profile, lenders or utility companies may demand other types of guarantees before investing in public street lighting projects, even for municipalities with a well- formulated COSIP, particularly those with lower financial capacity. Challenge 7 – Lack of a clear regulatory framework for public lighting Public lighting is a network service, as is, for example, the electricity distribution service. As a natural monopoly, these services should be regulated in terms of (i) entry and exit conditions for service providers; (ii) economic regulation; and (iii) service delivery quality. All this requires substantial regulatory capacity, a process that is currently not being led by the state or federal level. In the absence of an institutional knowledge base for regulating the public lighting service, an alternative is to implement some regulation through a “regulation by contract” approach. In this scenario, concession contracts are prepared in the form of detailed instruments covering a variety of situations that may arise during contract implementation, or issues related to contract extension; however, this process is cumbersome and prone to important, unforeseen exclusions. Moving to a more robust and efficient system would require regulatory action to address important questions for the market (e.g., operator incentives to invest at an optimal level, the best way to ring- fence COSIP revenues, defining payments in case of changes to exogenous factors such as electricity prices, etc.). XXIII 1. Executive Summary The current lack of regulatory clarity will inherently slow down market growth. Given this vacuum, capacity-building and regulatory standards will need to be developed at the municipal level to ensure that sustainable contracts are implemented under the existing regulation by contract system. The current lack of regulatory clarity will As mentioned in the introduction, the World Bank inherently slow down market growth. Given undertook a detailed survey of over 300 Brazilian this vacuum, capacity-building and regulatory municipalities with different population profiles and standards will need to be developed at the economic status in order to understand the context municipal level to ensure that sustainable in which the public lighting sector operates. Based contracts are implemented under the existing on data obtained in this research, together with regulation by contract system. statistical socioeconomic data covering all Brazilian municipalities, the World Bank team then grouped Mapping Brazil’s the different cities according to their capacities for developing different business models aimed at Municipalities for securing investment for modernizing the sector. Public Lighting The set of socioeconomic characteristics considered by the World Bank survey included levels of socioeconomic development, the size and density of the lighting network, fiscal status, the technologies in the existing street lighting inventory, and existing The identification of business models that could network coverage. enable municipalities to benefit from market opportunities and overcome the key challenges The procedure for grouping municipalities was divided is an important step in the development of the in two phases: (i) statistical cluster analysis; and Brazilian market. However, prior to identifying (ii) identification of homogeneous groups amongst appropriate business models, the characteristics of the clusters. The cluster analysis produced 18 the municipal public lighting market in Brazil needs initial clusters (homogenous groups), subsequently to be understood. reorganized/regrouped into six groups based on a qualitative evaluation of the clusters’ characteristics, Brazil has 5,570 highly heterogeneous municipalities namely project scale and fiscal management. This with different socioeconomic characteristics approach enabled multiple important aspects to be (income and development levels) and physical/ taken into consideration (size, development level, demographic features. A single universal business tax situation, lighting network density, sectorial model for public lighting projects is clearly out of the indicators) and, at the same time, identifying a question. Therefore, the first challenge is to group manageable number of municipal groups. For the municipalities on the basis of their similarities, with statistical cluster analysis, the variables presented in the ultimate goal of developing solutions tailored to Table 1 were used. each group of municipality. XXIV Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 1 - Database characteristics Variable Definition Objective/logic for use Unit Year GDP per capita GDP divided by number of inhabitants of Proxy of the municipality´s level of R$ 2012 municipality development Consumer units “Measurement of the water supply network Proxy of the level of “verticalization” of the Ratio 2013 per water supply density. A connection consists of the municipality, giving an idea of the public connection consumer unit/ units linked via a single street lighting network density. The public extension to the distribution network . 2 water supply system also basically covers the same area as the street lighting network. IFGF Index for measuring how taxes paid by Proxy of the level of the municipality´s Index 2013 community residents are managed by the fiscal management. municipal authorities. Composed of five indicators: Own Revenue, Personnel Costs, Investments, Liquidity, and Debt Cost. NCD/NCR Ratio between NCD (Net Current Debt– “To measure municipality´s level of debt. Ratio 2015 Consolidated Debt deducted from financial assets) and Net Current Revenue (NCR)5. Number of light Estimate of the number of light points in the Proxy of the size of municipalities; Number 2014 points public street lighting system 6 estimate the size of the public street lighting system – also a proxy for assessing the level of investment for a public lighting retrofit project. > 20% mercury Indicator (yes/no) to indicate whether the Efficiency level of the current network in Indica- 2014 vapor lamps municipality uses over 20% of inefficient order to assess the possibilities of making tor used in the mercury vapor lamps in its public street savings via a modernization program public lighting lighting network 8 network 7 % of the Percentage of the municipality not covered by Proxy for the amount of investment Percent- 2014 municipality street lighting. 9 needed to modernize the existing network age not covered by versus extension of the network to public street unserved areas. lighting Source: IBGE; FIRJAN; National Treasury; Ministry of Cities, World Bank Group. 4 For example, a building with ten apartments can have a single connection serving ten consumer units. Therefore for this building the consumer units per connection ratio is ten. 5 There are cases where the municipality may have a negative NCD, i.e., cash availability exceeding financial liabilities. In this case, a negative NCD NCR index indicates how much cash the municipality possesses in relation to Net Current Revenue. 6 Using current available data and estimates using a regression model when data unavailable. 7 Technology used in the network estimated from the use of a proportion of more than 20% mercury vapor lamps. 8 Using current available data and estimates using a regression model when data are unavailable. XXV 1. Executive Summary The final grouping results are presented in Tables clustering and grouping approach can be found in 2 and 3.10 More detailed information about the Annex 3. Table 2 – Main characteristics of the groups Number of municipalities Group Scale (Light points) Fiscal Management total % A 47 1 B 88 2 C 329 6 D 887 16 E 3.406 61 F 813 15 Source: World Bank Legend: Yellow = good; Orange = moderate; Red = limited Table 3 – Municipal groups (statistics) Population Light points Required Investments (R$)11 Group Total Total Total Average % Average % Average % (millions) (millions) (billions) (millions) A 59.9 29 1,274.,015 5.1 27 107,499 7.7 27 161.3 B 23.8 12 270,041 2.8 15 31,490 4.2 15 47.2 C 14.7 7 44,701 2.1 11 6,303 3.2 11 9.5 D 23.0 11 25,967 2.2 12 2,437 3.3 12 3.7 E 64.4 32 18,921 5.1 28 1,493 7.7 28 2.2 F 18.6 9 22,894 1.2 7 1,533 1.8 7 2.3 TOTAL 204.4 100% 36,704 18.4 100% 3,302 27.8 100% 5.0 Source: World Bank Group and Pezco Consultoria. Tables 2 and 3 show that the municipalities in Groups end of the spectrum, cities in Group E have a smaller A and B have a good scale of public lighting (over scale (less than 2,000 light points) and moderate fiscal 20,000 points) as well as strong fiscal management. management, and this group includes the highest These two groups represent only 3% of the country’s number of cities and populations of all the groups, municipalities, although they contain 41% of the representing 61% and 32% of the country, respectively, population and 42% of the light points. On the opposite and approximately 20% of Brazil’s light points. 9 In these variables we used data from a sample survey of Brazilian municipalities conducted for the World Bank Group by Castagnari Consultoria Ltda, given that no data was available for the entire universe of municipalities. 10 List of cities in each group is provided at . 11 Equipment prices based on World Bank Group surveys in June 2015. Prices are estimated at R$ 1,500 per point (excluding “smart” controls). Prices do not include the potential impact on the scale of procurement. Official exchange rate of May 13, 2016 at BRL 3.5/US$. XXVI Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Business Models A detailed understanding of the current market makes it possible to identify business models for Public Street to meet the requirements of these six groups of municipalities. Eight business models were proposed Lighting in Brazil that could be deployed to serve the needs of the six municipal groups, as described in Table 4. Table 4–Summary of business models Model Description M1 – Municipal PPP Municipalities tender a PPP for provision of energy efficient public street lighting using LEDs. A consortium of companies is awarded an administrative concession by the municipality to provide these services over a 5–35 year period. Management concession to retrofit the public street light- ing network and deliver efficient public lighting services. M2 – PPPs with Mu- Municipalities form a municipal consortium, which then becomes the legal body to tender a PPP nicipal Consortium 12 for energy efficient public street lighting on behalf of all municipalities within the consortium. Since consortiums are not able to borrow under Brazilian law, the consortium would create a Special Purpose Vehicle (SPV) that would be able to undertake financing on behalf of all munici- palities in the consortium. M3 – Municipal Issuance of municipal bonds or loans, enabling those municipalities that either do not want to ten- Financing der a PPP for LED street lighting or that lack the technical or financial capacity to do so, an option to finance the necessary investments for conversion to LEDs. M4 – Electric Utility Involves the utility company financing LED procurement on behalf of municipalities, leveraging Programs COSIP (and potentially other) resources from consumers´ electricity bills (i.e., on-bill financing approach). M5 – Energy Service ESCOs obtain financing in the market and make the necessary investments for municipal LED Companies (ESCOs) retrofitting. Operation and maintenance remain the responsibility of the municipality. M6 – Centralized Creation of municipal consortia to centralize LED equipment procurement and benefit from the Procurement resulting economies of scale for the cost of LEDs. Each municipality remains responsible for financing the investment individually. M7 – Self-Funding Municipalities use surplus revenues in any given year to fund investments in LED street lighting over a longer period of time, without obtaining any up-front financing. M8 – Transfer of Interim solution consisting of redeploying HPS (or mercury vapor) luminaires made redundant Luminaires after a city converts to LEDs, transferred to municipalities with limited prospects to convert their own systems to LED in the near future. Source: World Bank 12 Note that although the formation of a municipal consortium is a necessary precondition for the M2 model, aggregation (to scale up the network, reduce transaction costs, etc.) could be achieved through the use of municipal consortia in several other models. XXVII 1. Executive Summary A more detailed description of the eight business models can be found in Annex 1. business models is provided in Section 6 of the report, including a summary of their different Table 5 shows the connection between the characteristics, a structuring diagram, the groups municipal groups and the respective business of municipalities to which they can be applied, the models. It is important to emphasize that several advantages, disadvantages, risks and mitigation business models can be used in most of the factors of each model, as well as the identification municipal clusters, depending on local government of key stakeholders for each project phase of project preferences and the circumstances of each city. development. An abbreviated summary of the Table 5 - Mapping business models tailored to each group M1 M2 M3 M4 M5 M6 M7 M8 PPP Municipal Municipal Electric ESCO(s) Centralized Self- Transfer of Groups Municipal Consortium Funding Utility Procurement funding Luminaires with PPP Programs A B C D E F Source: World Bank Group, Pezco Consultoria Legend: = Suggested; = Possible. Financing be deployed. For example, models such as M1 and M5 may appeal to the private sector, while others such as Mechanisms and M3 rely on municipalities to play a preponderant role in financing the modernization and future operation Improving Credit Risk of the street lighting network. Table 6 provides an overview of the financing instruments identified as having potential to finance LED street lighting Each of the business models described above contains projects in Brazil. specificities that affect which financing options could XXVIII Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 6 - Mapping financial instruments tailored to business models Mechanism Description COSIP A tax collected by utility companies through consumers’ electricity bills and passed to the munici- pality to cover the current expenses and expansion costs of its public street lighting system. It is an “earmarked” fund and cannot be used for any other purpose. Loans from Public Banks These institutions (Banco do Brasil and the Caixa Econômica Federal) are physically present all over (BB, CEF) Brazil and play an important role in providing loans for municipal infrastructure projects (many local authorities are active clients). The CEF possesses decentralized, qualified technical teams that could assist smaller municipalities to review their public lighting modernization projects.. BNDES-Finem This is a line dedicated to energy efficiency projects, replacing the previous PROESCO line from BNDES. BNDES-Finem accepts transactions with values equal to or greater than R$5 million, cover- ing up to 70% of financeable items. Private Equity This entails private capital investment as shareholder equity in a Special Purpose Vehicle (SPV), aimed at financing a public street lighting system retrofit. Credit Enhancement Credit enhancement mechanisms can be essential for financing public lighting retrofits. Even when the Mechanisms COSIP tax is well-formulated and offers strong guarantees for lenders, there might be a perception of residual risk of insufficient COSIP resources and, thus, exposure of investors to municipal credit risk. Development Banks (do- Institutions such as BNDES, DesenvolveSP, AgeRio, BDMG, the World Bank, IFC, CAF, IDB, etc, provide mestic and multilateral) several instruments focused on financing energy efficiency projects, potentially including public street lighting projects. DesenvolveSP, AgeRio and BDMG also provide technical assistance to assist munici- pal governments with project design and structuring public-private partnerships. Sector Financing Lines Sector-specific funds such as PEE and PROCEL Reluz have played important roles in public light- ing retrofits. However, regulatory changes have reduced the amount of resources available to these financing lines. Loans from Private This is a modality where private commercial banks (domestic or foreign) provide loans directly to Banks municipal governments or to agents acting on their behalf for lighting sector retrofits. For loans to the public sector, municipalities are subject to the applicable indebtedness limits. Debentures, FIDIC, FIP, These instruments enable funds to be raised for large-scale projects in national or international capi- Green Bonds tal markets in higher volumes and at more competitive cost. FI-FGTS FGTS funding has supported several urban infrastructure projects. In view of recent institutional changes it would make sense to consider street lighting modernization projects as a part of urban infrastructure. The allocation of FI-FGTS funds requires the approval of the Fund´s Board of Trustees (Conselho Gestor). These funds are not currently available for any of the business models selected. XXIX 1. Executive Summary The applicability of these financial instruments to the eight business models is shown in Table 7. Table 7 - Mapping financial instruments tailored to business models Financial M1 - M2 - M3 - M4 - M5 - M6 - M7 - Self- M8 - Instrument Municipal PPPs for Municipal Electric ESCOS Centralized financing Transfer of / Business PPP Municipal Financing Utility Procurement Luminaires Model Consortium Program COSIP or Municipal Budget Loans from Public Banks (BB, CEF) BNDES - FINEM Private Equity Credit enhancements (e.g. WorldBank) Development Banks (domestic, multilateral) Sector Financing Lines (PEE, RELUZ, PROCEL) Loans from Private Banks Bonds, FIDC, FIP, GReen Bonds FI-FGTS Source: World Bank Legend: Full circle = more adequate; Empty circle = not applicable XXX Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning More details on financing and risk mitigation reduce the economic and social benefits that LEDs mechanisms – including mechanisms to mitigate technology can provide (energy savings, public noncredit risks (technical, operational risk, political, safety, improvement of services for community etc.) – are provided in Section 7 of the report. residents, etc.). Gaps in the Therefore, it is essential to identify gaps and seek solutions to include the highest number of Brazilian Market, and cities in this technological revolution. Section 9 of the main report lists barriers and gaps that can Recommendations. undermine the overall development of the system in six areas: (1) regulatory/legal framework; (2) public policies; (3) financing; (4) capacity-building; Most of the business models and financial and (5) technology, and provides recommendations instruments described require some type of for overcoming each of the existing barriers. institutional support to speed up implementation at the national level. Not all of the recommendations can be completed in the short term, and some are more critical to The results of our study indicate that, without the development of the LED street lighting market some type of intervention in the market, the most than others. In order to facilitate stakeholder probable scenario would be the use of the “self- uptake and implementation of the above funding” business model for over 90% of the cities, recommendations, this report categorizes the with 50% of total light points (i.e., Model M7; recommended actions in terms of their indicative Groups D, E and F). This would imply a very slow timeframe for implementation and their level of retrofit of Brazil’s public street lighting sector, and priority, as shown in Table 8 below. Table 8 – Prioritization of recommendations and indicative implementation timeframe Short-term (<1 year) Highest priority recommendations Important recommendations Identify and designate federal, state, and/or municipal leaders Create standardized project evaluation tools to be responsible for leading coordination of EE street lighting initiatives and for helping municipalities to overcome barriers Create new credit lines and/or instruments for municipalities Create guidelines for implementing or adjusting COSIP unable to attract (sufficient) private investment. Design investment funds offering new, standardized EE asset Arrange seminars for exchanging ideas between the audit class to attract a variety of investors at scale counts on the most controversial subjects affecting public lighting contracts Create instruments for mitigating municipal credit risk Introduce new databases for providing key data related to EE street lighting projects for cities and potential investors (city street lighting inventories, COSIP collection information, etc.) Standardization/certification of LED equipment XXXI 1. Executive Summary Medium-term (1–3 years) Highest priority recommendations Important recommendations Identify possible actors with regulatory knowledge to fill regu- Standardization of the legal framework cities can use to lator gap, or at least to provide regulatory advice to municipal earmark COSIP funds governments when necessary The federal government interprets and clarifies what are the Establish a uniform approach for oversight agencies to ver- legal uses of COSIP (e.g., for expenditures electricity, O&M, ex- ifying the legal provisions applied to different procurement pansion, modernization). bidding modalities. Standardize business models Federal government should design a national strategy for EE street lighting, including targets for converting the national public street lighting network to LEDs Standardize financial contracts/instruments Explore new instruments that may reduce foreign exchange risk that are affordable and available over the longer term (e.g. partial indexation in US$) Create an industrial policy for national production of LEDs (re- Create national or state technical assistance programs view import taxes, etc.) Introduce standard templates for EE public street lighting contracts Ensure efficient and effective flow of information between the municipal audit courts to promote consistency Standardize equipment guarantees provided by suppliers Increase use of other technology risk mitigants available in the market (e.g., insurance) Increase national production by creating credit lines for the national production of LEDs Long-term (>3 years) Highest priority recommendations Important recommendations Revise the legislative framework for public consortia to allow Provide electric utilities different incentives to promote EE direct financing of the public consortium´s legal entity street lighting by altering the form of compensation for ener- gy distributors, namely through decoupling sales (and MWh) and revenues Pass new legislation that allows exceptions to be made in munic- ipal debt thresholds in order to enable greater investments in EE public lighting (similar to exceptions that were provided to debt under the PROCEL program) XXXII Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning This analysis shows that there are a handful of However, it is no simple task to make this kind of high-priority activities that could be implemented energy efficiency project a reality, and no single in the short-term to help catalyze the LED recipe can be applied to all our cities. The complex street lighting market in Brazil. These include nature of the public street lighting services in (i) identifying and designating national leaders Brazilian cities presents many challenges and for leading coordination of EE street lighting opportunities that need to be faced in the quest initiatives, (ii) creating new public credit lines, (iii) for more efficient technologies. designing investment funds offering standardized EE asset class to attract private-sector investors Brazilian cities, although aware of the benefits, at scale, (iv) creating instruments for mitigating are faced with a series of economic, financial and municipal credit risk, and (v) finalizing the regulatory barriers to developing their public standardization/certification of LED equipment. street lighting projects. This is the case not Other important steps will need to be taken in only at the local level (e.g., the large amount the medium and long-term to expand the market, of investment needed while facing restrictions such as standardization of business models and under the Fiscal Responsibility Law), but also at contracts, establishment of new industrial policies the macroeconomic level, where extremely high to promote domestic LED production, and revision interest rates prevail and where depreciation of of the legislative framework for public consortia the national currency is an ever-present concern. to increase their feasibility (e.g., allow them to take on financing obligations directly). At the same time, Brazilian cities face strong incentives to investment in this sector, given the mounting costs of energy, the continued reduction Conclusions in costs of LEDs, and in most of the municipalities, the existence of specific funds available through In an increasingly urbanized world, solutions for the COSIP charge on utility bills. The 40%+ climate change and other global problems will municipalities that recently became owners of depend heavily on public policies and projects their public lighting assets also need to identify developed at the city level. By developing energy options to efficiently operate and maintain their efficiency projects in cities, which represent over street lighting networks. two-thirds of energy consumption and over 70% of emissions, it is also possible to achieve impacts With this initial survey of Brazilian municipalities, on a global scale. it has been possible to identify groups with different capacities and needs. The classification As for the public lighting situation in Brazil, huge of Brazil´s 5,570 municipalities into six groups is a potential exists for energy efficiency projects first step towards proposing solutions tailored to that can have a significant impact on energy their needs. The eight business models presented consumption and on mitigating climate change, in this report take into account the characteristics in addition to the myriad other benefits, including of the various groups of municipalities, proposing reduced operational and maintenance costs, structures and sources of financing to enable better lighting quality, increased perception of entire public street lighting systems to be scaled security, creating a smart-city platform, etc. up. Before selecting a particular business model, it is important for municipal governments to carry XXXIII 1. Executive Summary out a critical assessment of their own needs and market, and proposes a series of “next steps” capacities for designing and/or managing the to fill the gaps and overcome barriers. Most of operations inherent in each model. these recommendations require governmental intervention; therefore, it is important that Once the business model has been chosen, a national level entity take a leadership role the next step is to identify the best sources of in the sector. Recent changes at the federal finance for progressing the operation, including, level – including, for example, the creation whenever necessary, additional mechanisms of an Executive Secretariat in the Investment such as guarantees for mitigating municipal Partnerships Program (PPI) and the increased credit risk. This report presents 10 financing funding for changes to PROCEL – potentially and credit enhancement instruments, and maps provide opportunities for new federal agencies to the compatibility of these instruments with the play a more active role in promoting this agenda. business models. The World Bank Group stands ready to continue In order to speed up the implementation of energy to support the Brazilian government and other efficiency projects in public street lighting, the stakeholders to accelerate investment in energy report identifies gaps and barriers in the current efficiency in the public street lighting sector. XXXIV Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning XXXV 1. Executive Summary XXXVI Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 1 - Introduction C ities are among the world´s largest consumers in energy prices in recent years. of electric energy, accountable for two-thirds of total electricity consumption and for over The advantages of this new technology are not 70% of global greenhouse gases emissions13. Public limited to energy efficiency. It also offers significant street lighting systems contribute significantly to reductions in operating and maintenance costs, as a city´s energy consumption. In Brazil, the cost of well as important co-benefits such as reducing energy for public lighting already represents the crime and increasing the community´s perception second most expensive item of most municipalities’ of security. budgets, surpassed only by payroll expenditures. Furthermore, new regulations require all the In spite of the substantial benefits associated with municipalities to own the city’s public lighting the conversion of the installed public lighting assets, making public lighting one of the few network in Brazilian cities, major economic- sectors in which local authorities have direct financial and institutional obstacles still need to control over energy-consuming assets (contrasted be overcome. Although LEDs are more efficient, with other high energy-consuming sectors such as they are considerably more capital-intensive transport). As a result, the local authorities will than existing technologies. Although LED prices have every incentive to invest in and implement are falling at a rapid pace (estimated at 10% per lighting projects by themselves. year), the relatively high costs of the investments required remain a major challenge to converting In Brazil, the current public street lighting Brazil’s public street lighting networks. inventory primarily consists of mercury and HPS lamps, which over time will tend to be replaced In order to reap the benefits of conversion to LED, by more efficient technologies such as Light- it is necessary to design and implement business Emitting Diodes (LEDs). This new technology is models that can enable the necessary investments. already in operation in some major cities in other These business models must take into account the countries. The availability and increasing spread diversity of Brazil’s municipalities. Furthermore, of LED technology offers a unique opportunity consideration must be given to designing financial for Brazilian cities to reduce their energy solutions that can raise private sector capital consumption. This is especially important and while mitigating municipal credit and project beneficial to cities, considering the sharp increase performance risks. 13 Commonly known as GHG (Greenhouse Gases). 1 1. Introduction 2 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Given the unquestionable benefits of LEDs and indicators, installed technology, current the fact that LED prices continue to fall, the energy consumption, etc., and outlines the transition to LED technology is almost inevitable, opportunities and main challenges that need to although the pace of conversion is uncertain. This be overcome in order to scale up the Brazilian report seeks primarily to identify business and network using LED technology. financing models that, by taking into account the • Section 5 - consists of a brief review of institutional environment and characteristics of the technical task of mapping Brazilian the Brazilian market, will enable the necessary municipalities in terms of their public lighting resources to be catalyzed to advance sustainable networks, and describes the procedure for projects in the short- and medium-term, to the classifying the municipalities in homogeneous benefit of municipalities, communities, and the clusters, followed by assembling them in six environment. different groups to facilitate the development of business models. This report makes use of studies that have already been conducted by the World Bank Group • Section 6 - outlines the proposed eight specifically on the public street lighting network business models corresponding to the in Brazil. These include: pre-feasibility studies on suggested groups of municipalities. the cities of Rio de Janeiro and Belo Horizonte;14 • Section 7 - provides an overview of the a sample survey of the public lighting situation financial instruments that can be used in the in Brazilian municipalities15; field visits by World business models. Bank consultants in five areas of Brazil 16; and a cluster analysis and review of preliminary • Section 8 - considers ideas for designing business models carried out by Pezco Consultoria. a public street lighting project, focusing The study also benefited from previous World specifically on the scope for system Bank and the International Finance Corporation automation/remote management and possible (IFC) studies on other countries and regions. infrastructure improvements for the lighting sector and the city. The remainder of this report is divided into eight • Section 9 - summarizes the main gaps and sections: barriers for developing this market, and contains recommendations on how to narrow • Section 3 - reviews international trends these gaps and overcome the barriers. in public street lighting with the new LED technologies, drawing attention to their benefits • Section 10 - sets out the conclusions of the and describing international experience with report. them. This section is followed by three annexes, which • Section 4 - describes the public lighting detail technical aspects and provide more data to market in Brazil, focusing on lighting coverage support the analyses. 14 Pre-feasibility studies were carried out by World Bank Group teams, completed in August and September 2014. 15 “Study of the situation of public illumination in Brazilian municipalities”, a sample survey involving face-to-face interviews in 300 municipalities by Castagnari Consultoria Ltda, under contract to the World Bank in 2015. 16 The visits between March and April 2015 included, among others: Uberlândia-MG, Hortolândia-SP, Consortium Cides-MG and Aracoiaba-CE. 3 2. Overview of trends in public street lighting using LEDs 4 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 2 - Overview of Trends in Public Street Lighting using LEDs. T he development of solid-state lighting the public street lighting sector, accompanied by a technology17 has provided major opportunities range of benefits for the community as a whole, such to reduce energy consumption and achieve as better security and environmental quality. The better results with the use of new lighting equipment following describes some of the international trends. in residential, commercial and public environments. 2.1 - LED Technology The costs of the many applications of the new technology have been falling significantly in terms of equipment and maintenance costs. The main public street lighting technologies The focus of these developments is the public sector, currently employed in Brazil are based on metallic, which is now beginning to benefit from the new mercury and HPS lamps that use glass tubes and solid-state technology used in public street lighting internal gas. Table 9 compares LED technology with networks. There is great potential for savings across other currently available selected technologies18. Table 9 - Summary of LED features compared with other technologies Type of lamp Technical Perceived nocturnal luminous Color Lamp useful Price luminosity efficiency (lumens/watt)20 Reproduction lifespan (hours) (R$)21 efficiency19 Index (CRI) (lumens/watt) High pressure 50-150 32-95 24 15,000-24,000 $316 sodium (HPS) Metallic vapor 70-130 104-194 96 8,000-12,000 $320 Mercury vapor 35-65 38-70 17 10,000-15,000 $285 LEDs 70-160 133-304 70-90+ 40,000-90,000 $1,500 Source: Adapted from Pike Research Smart Street Lighting, published in the 3rd quarter of 2012, and from interviews with key players. 17 Solid-State Lighting (SSL). 18 There are other technologies available, such as low pressure HPS and induction, that were not included in this study for sake of simplification. 19 Photopic vision (P) is that adapted to high luminance levels such as daylight. 20 Scotopic (S) vision is that adapted to low levels of luminance, such as at night. Scotopic vision is much more sensitive to white light such as that produced by LEDs and inductive fluorescent lamps, and less sensitive to the yellowish light of HPS lamps. Therefore, the visual impact on the pupil of a lumen produced by an LED luminaire is much greater at night than the same lumen produced by an HPS. The relative impact is expressed by the S/P correction index which adjusts the photopic vision to the perception of luminosity in the pupil of the human eye. The conversion between technical (photopic) light efficiency and perceived (scotopic) light efficiency is described in more detail in Annex 2. 21 Prices based on a survey made by World Bank Group teams in June 2015. The official exchange rate of the day (05/16/2016) of R$ 3.5/US$ was used. 5 2. Overview of trends in public street lighting using LEDs As can be seen in Table 9, LED luminaires and good of performance problems, real-time monitoring of the quality high-pressure sodium lamps (HPS) are almost entire lighting system and, finally, the use of remote equivalent from the point of view of technical luminous metering. All these features have been proven to reduce efficacy. However, HPS luminaires do not produce energy and operating costs. the luminous frequencies to excite the human eye at the same level as LEDs. Thus, LED luminaires have Finally, unlike mercury vapor and HPS lamps, LED proven to be superior, after taking into account how lamps do not contain heavy metals, meaning that the light is perceived by the human eye in the dark, i.e., the risk of environmental contamination is lower, especially ‘perceived night-time luminous efficiency’22. Moreover, if their electronic components comply with the RoHS LED-based luminaires have special optics that generate standard.23 a better directed light beam pattern, producing improved luminosity on public roads, with fewer dark areas between poles. Ultimately, this means that 2.2 - Economic lower-powered luminaires are sufficient to illuminate the same area, thus increasing the potential efficiency and Financial gains with LEDs. Benefits of LEDs LED-based technologies also provide better color reproduction, instant start-up, better system LEDs can generate substantial economic and financial integration, and reduced operating costs. The Color benefits for cities that install them for public street Reproduction Index (CRI measuring how well the lighting. The benefits range from energy savings to human eye is able to distinguish colors) of LEDs is 70- reduced operational and maintenance costs due to 90% compared to the daylight index of 100%. By way the longer lifespan of LEDs compared to other current of comparison, mercury and HPS lamps have a CRI of technologies. 55% and 24%, respectively. Metal halide lamps, despite their CRI of 96%, have a considerably shorter lifecycle The Climate Group (an International NGO) was one of than LEDs.In fact, LED lamp life is at least twice that of the first organizations to conduct studies in large cities metal halide and mercury vapor lamps, and 1.5x that of with LED-based public lighting projects to evaluate the HPS lamps. This implies lower replacement costs and a performance of the new technology. The Group studied consequent reduction in operational and maintenance 12 cities around the world, including New York, Toronto, costs. The savings that can be made with LED lighting Hong Kong, London, Sydney, Adelaide and Calcutta can be substantial. and was able to demonstrate that LED technology can achieve energy savings of 50–70%, and as much Another benefit of LEDs is that they ignite virtually as 80% when combined with intelligent management instantaneously, which increases their efficiency and control systems. According to this study, adopting when the lamps are switched on and when network intelligent control systems produced greater flexibility fluctuations or voltage drops occur. LED luminaires can in terms of lighting options and improvements in the also be installed in conjunction with intelligent control quality of public lighting services for the community. systems which enable light points to be individually controlled, including dimming, increasing the current to In 2013–14, the World Bank also carried out detailed increase lighting output to compensate for the natural studies of the economic and financial viability of LED depreciation of the LEDs over time, immediate detection projects for public lighting, focusing on two Brazilian 22 The conversion between the technical / photopic luminous efficacy and perceived scotopic light efficacy is described in more detail in Annex 2. 23 Restriction of Hazardous Substances (RoHS) originated in the EU under Directive 2002/95/EC and restricts the use of six hazardous materials found in electrical and electronic products. 6 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning cities: Rio de Janeiro and Belo Horizonte. These studies two studies to other Brazilian cities and the world as revealed very substantial potential benefits from a whole, the World Bank created a model to enable investments in LED public lighting, with an internal rate cities to estimate the costs and economic/financial of return (IRR) sufficiently attractive to attract private benefits of an investment project involving the sector interest. Box 1 contains the results of the LED installation of LED public street lighting, taking into pre-feasibility study in Rio de Janeiro. account the specific characteristics of each city. This tool is available free of cost24. To disseminate the knowledge acquired from these BOX 1 - Pre-feasibility Study for the Use of LEDs in the City of Rio de Janeiro In 2014, the World Bank - in collaboration with the city of Rio de Janeiro - completed a pre-feasibility study to identify possible implementation and financial structuring options for investments in LED public street lighting in Rio de Janeiro. The study capitalized on lessons learned and work done under the Low Carbon City Development Program (LCCDP) for Rio de Janeiro, funded by ESMAP and PPIAF - two World Bank trust funds25. The WB study addresses four themes: project design, legal and institutional frameworks, financial and economic viability, and financing options for the municipality. Although the study focuses primarily on Rio de Janeiro, many of the findings could assist other cities in Brazil and around the world to assess the feasibility of implementing an energy-efficient street lighting system. The study also evaluated, in 2014, the financial and economic viability of an LED-based project (with smart-systems) in Rio de Janeiro. The rate of return and the financial and economic analyses were updated in July 2015 to take account of important changes in project assumptions (i) increased electricity prices; (ii) R$/US$ exchange rate fluctuations; and (iii) increased national production of LEDs25. The results of the studies are shown in Table 10. The financial analysis for Rio de Janeiro estimated an investment of around R$390 million over a five-year period (2017–2021) to cover the costs of LED equipment, the intelligent system and installation. The study estimated that it would be possible to save up to 57% in electricity and operational and maintenance costs (the latter involving labor to service equipment and replace burned out lamps) of R$1.2 billion over a 15-year period. The project would also save around 2.2 GWh of electricity for the city, resulting in a reduction of 655,000 tons of carbon emissions (tCO 2e). The payback period was estimated at 6.5 years, with an IRR of 27%. Moreover, the study estimated that the returns would be highly favorable, thus providing an incentive for the private sector to consider financing the project if the municipality were interested. This study showed that investing in LED street lighting is a great opportunity for Rio de Janeiro to reduce energy consumption, save on electricity and O & M expenditure, and improve public lighting provision for the inhabitants of the city. Rio de Janeiro recently published its Strategic The CityLED Tool can be found on the World Bank ESMAP (https://www.esmap.org/node/57817). 24 25 Interviews with manufacturers indicate that they are interested in increasing the level of LED manufacture in Brazil providing a minimum scale is reached to make sales worthwhile. This could be the case with the implementation of public lighting PPP projects currently underway in São Paulo and Belo Horizonte. 7 2. Overview of trends in public street lighting using LEDs Plan for 2017–2020, which includes a LED public lighting project26. Table 10 - Summary of results of the pre-feasibility study for the city of Rio de Janeiro Project analysis Unit Financial analysis July 2015 (base case) Key assumptions Total luminaires replaced in project n 318,733 Years to implement project n 5 Period of analysis years 15 Include smart system? flag Y After-tax cost per point in year 1 (technology + infrastructure) BRL 1,365 Cost of electricity at year 1 BRL/year/MWh 420 Real annual increase of electricity tariff % 0 % electricity saving to LED with project % 57% WACC (net of inflation) % 6% Exchange rate BRL/USD 3.2 Social cost of carbon dioxide (CO2e) BRL/tCO2e 96 % domestic manufacturing % 50% Estimate project results Total project investment over 15 years (CAPEX + fees) million BRL 390 Total financing savings (electricity + O&M) over 15 years million BRL 1,173 Total GHG reductions over 15 years tCO2e 655,954 Total electricity saving over 15 years GWh 2,189 Payback period (non-discounted) years 6.4 RoI % 201% NPV over 15 years (million BRL) million BRL 380 IRR % 27% Source: World Bank 26 Available at: . Some countries are implementing national programs, Brazil has not yet defined national policies in this including India and Malaysia. Both countries regard; the municipalities, now owners of the public have announced ambitious LED implementation lighting assets, are responsible for making their programs - involving fully replacing their public own decisions on whether or not to convert to LED lighting networks by 2018 and 2020, respectively31. technology. Goldman Sachs data (2015); for India the source is: . 31 11 3. Public street lighting market in Brazil 12 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 3 - Public Street Lighting Market in Brazil This section provides an overview of existing public 4.3% of the country’s electricity consumption (14.3 street lighting coverage in Brazil based on the TWh), at a cost of R$3.5 billion (excluding taxes). The following indicators: the installed lumino-technical installed lumino-technical system largely consists of systems; electricity consumption; tariff evolution; high-pressure sodium lamps and, to a lesser degree, recent changes in the institutional environment; mercury vapor lamps. LED technology penetration is specific street lighting sector public policies; and, very low. finally, the current situation in terms of technical standardization in Brazil. 3.1.1 - Coverage of the public street lighting service 3.1 - General Aspects The lighting network has a high level of coverage and Public street lighting in Brazil is estimated to penetration, with a higher number of households consist of more than 18 million light points, with a served by public lighting than by other public penetration rate of around 95.5% of households. In services such as street paving or the presence of 2015, the entire public lighting system accounted for sidewalks (Table 11). Table 11 - Characteristics of the vicinity of Brazilian homes - selected public services Lighting of public areas Street paving Existence of sidewalk Percentage of homes served in Brazil 95.5% 81.0% 68.5% Source: IBGE - Demographic Census 2010 - Urban characteristics of the vicinity of homes Notwithstanding this percentage of homes benefiting example, on the size of municipalities. Those with up from public lighting, 2.1 million homes were to 20,000 inhabitants provide public lighting services unserved in 2010. The highest coverage of homes is to an average of 94.9% of homes, while larger in the Center-West (97%), and the lowest in the North municipalities with populations of over 500,000 have (89.2%). However, it is important to note that there an average coverage of 97.1%. are significant variations in coverage depending, for 13 3. Public street lighting market in Brazil In terms of household income, households with up times the minimum wage or more have a 98.1% rate. to one quarter of a minimum wage have an average Figure 4 shows the coverage by household income coverage rate of 90.2%, while those receiving two segment32. Figure 4 - Coverage of public lighting for % of households, by household income 98.1% 97.1% 95.2% 95.5% 93.3% 92.4% 90.2% Total Up to 1/4MW 1/4 to 1/2MW 1/2 to 1MW 1 to 2 MW More than 2MW No Income Source: IBGE - Demographic Census 2010 - Urban characteristics of the vicinity of homes. Calculations and compilation by Pezco Consultoria. Income in Brazil is generally described in terms of the national minimum wage. The amount of the minimum wage amount is defined by the Federal 32 Government and updated annually (R$ 880.00 per month in 2016). 14 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning There are still challenges for public street lighting 3.1.2 - Installed technology coverage, especially with regard to coverage for lower income populations. Although this study addresses The distribution of installed lighting capacity in Brazil issues related to conversion of public lighting to relates directly to the demographic concentration more efficient technologies rather than to network represented by large cities and along the Atlantic expansion, the unserved population may possibly Coast - principally in the Southeast and Northeast, benefit from PPP projects. The savings generated as as shown in Figure 5. the result of scaling up could be used to fund further expansion of the network. Figure 5 - Concentration of light points in Brazil Source: Pezco Consultoria using Tableau Software Table 12 summarizes the national profile and its five the two available sources. regions in terms of installed technology, based on 15 3. Public street lighting market in Brazil Table 12 – Number of lamps in the National Public Lighting System, by % Eletrobras Cadastre (2012) Brazil (%) North (%) Northeast (%) South (%) Center-West (%) Southeast (%) Mercury vapor 23.6 31.3 20.7 23.9 23.0 24.4 HPS 71.1 64.5 68.6 71.4 72.2 72.5 LEDs <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 Others33 5.3 4.2 10.7 4.7 4.9 3.1 Table 12 shows a massive concentration of HPS and technologies is already fairly widespread in Brazil mercury vapor technologies in the Brazilian street (see Figure 6).. The World Bank survey of 300 lighting system. The proportion of LED technology municipalities in May 2015 shows that 84.1% of was minimal by comparison. Brazilian municipalities are aware of LED public street lighting technology. Knowledge about new solid-state lighting Figure 6 - Percentages of municipalities familiar with LEDs in public street lighting 91.2 88.0 84.1 81.1 80.2 78.7 Brazil North Northeast Center-West Southeast South Source: World Bank Group In Brazil, around 16% of municipalities are not familiar to ensure that knowledge of the new technology with the idea of LED technology for public lighting. reaches a significant number of municipalities, city In the Center-West the percentage is over 20%. These mayors and public agents. numbers indicate that there is still work to be done Multivapor metallic, halogen and incandescent lamps, among others. 33 16 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Box 3 – Technology and Population Preferences In order to gain experience of local situations, World Bank consultants visited five different areas in Brazil between March and April 2015. One of the cities surveyed was Fortaleza, the capital of the state of Ceará and fifth largest city in Brazil, with a population of 2.6 million (2014) and 183,800 light points. The public lighting service of the city was observed to be well-structured, with registered and geo- referenced points and a specific Master Plan.The lighting service structure partly reflects the fact that the city assumed responsibility for public lighting in 2001, in anticipation of the nationwide transfer of lighting assets that took place around 14 years later. Maintenance, extension and improvement of the system is outsourced to a private company. One particularly interesting point from the consultants‘ research in Fortaleza was the observation that the local population prefers ‘white’ light, with better color reproduction. Thus, although the city has 345 LED points, mainly serving roads, streets and places of touristic interest, the use of metal halide lights (54,400 points) still prevails to achieve the white light desired by residents. It is worth noting that using metal halide technology to obtain white light is inefficient compared to LED technology. Large cities such as Fortaleza could provide what their populations genuinely want (the benefits of white light) in a more economical and sustainable way if they were to invest in converting their system to LEDs. 3.2 - Bidding (except oversight) for the system to private outside operators through the award of a concession. Schemes for Public Intermediate solutions can include municipalities that outsource only certain responsibilities to private Street Lighting sector companies. According to the aforementioned World Bank Group survey, private sector participation in maintenance contracts is fairly widespread, presumably in The public street lighting system in Brazil is usually accordance with Law 8666 and/or bidding (pregão) managed by a combination of public stakeholders regimes. While around half of Brazilian cities (municipalities) and private practitioners. At one end outsource at least some of their maintenance of the spectrum municipalities may be responsible for services, there are substantial regional variations. all aspects of management, including investments in Meanwhile, administrative concession contracts system expansion, modernization, maintenance and for public street lighting services are increasingly operation. At the other extreme, the municipality may common, although these are confined to a small transfer virtually all the management responsibilities number of Brazilian cities. 17 3. Public street lighting market in Brazil Figure 7 - Responsibility for maintenance of public street lighting (% of municipalities) Own Outsourced Both 6.1 6.8 7.1 8.6 5.2 4.7 7.6 26.4 37.5 40.2 45.3 83.2 83.8 66.4 57.8 53.1 48.5 11.7 Br azil Nor th Northeast Center-West Southeast South Source: World Bank Group This section contains a brief overview of the most common instruments used by Brazilian municipalities to contract goods and/or services for public street lighting efficiency projects, summarized in Table 13.34 34 The common concession system (Law 8,987 / 1995) is not included in this report because it enables the private sector to participate in large contracts that are remunerated by consumer tariffs. Given that public lighting projects are not a specific service provided to ordinary consumers, but to municipalities, the common concession system does not apply to this sector. In the same way, the sponsored concession system is also not considered because the corresponding contract provides that part of the concessionaire´s remuneration derives from the charging of user fees. 18 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 13 - Bidding regimes for energy efficiency projects in public lighting systems Instrument Characteristics Law 8.666/1993 • Conventional public management contracts, with a one-year term, extendable by up to 48 months. • Bidding or competition required for all engineering contracts worth over R$1.5 million. • Criteria to determine the winning bidder: lowest price; best technical quality; combination of price and technical quality; highest bid or offer. • Processing time depending on complexity of goods and/or services to be tendered — from one month (goods) to six months (goods + services). Auction • Designed to facilitate the purchase of joint goods and services by municipal governments, using the (Pregão) (Law ”reverse auction“ procedure. 10.520/2001) • Lowest bid is the only selection criterion. • One-year contract, extendable for up to 60 months. • A simple process that can be concluded in less than a month after publication of the public tender notice PPP—Ad- • Concession contract in which the final contractor of the services is the public agency which remuner- ministrative ates the ”concessionaire“ monetarily and without charging user fees. Concession (Law 11.079/2004) • The selection criteria are: lowest end-user tariff; combination of lowest tariff and best technical quali- ty; lowest payment by public agency; combination of lowest payment and best technical quality. • PPP concession agreements can last 5–35 years, renewable. • Estimated timeframe for implementing a PPP project: 15–20 months. 19 3. Public street lighting market in Brazil 3.3 - Opportunities 3.3.1.1 Electricity prices for LEDs in the The opportunity for investing in Brazil´s public street lighting sector is affected by a scenario in Brazilian context which electricity and equipment prices are moving in opposite directions. While LED equipment prices tend to be increasingly cheaper as the result of The current public street lighting network presents technological innovations (with a consequent opportunities for LEDs; however, the process also reduction in the capital outlay required for converting faces challenges. to LEDs), electricity prices in Brazil continue to rise.35 It follows that projects that help reduce energy This section provides an overview of the most consumption offer obvious benefits. important opportunities and most significant challenges in the Brazilian market. The average cost of public street lighting electricity increased by 38.8% between 2014 and 2015 (from R$ 3.3.1 - Opportunity 1 - Energy 0.18 / kWh to 0.25 / kWh) after an already substantial and equipment prices moving increase (10.9%) in 2013. Figure 8 shows the annual variation of the public street lighting energy tariff. in opposite directions Figure 8 – Average electric energy tariff for public street lighting, % variation 50% 40% 30% 20% 10% 0 -10% -20% 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: ANEEL; Pezco Consultoria 35 Electricity tariffs are set by the National Electricity Energy Agency (ANEEL). This federal regulatory agency allocates tariffs per group of users according to voltage. Public lighting falls into Subgroup B4, divided into: a) B4a: the tariff applied when the public street lighting assets are municipally-owned ( i.e. the present situation of all Brazil´s municipalities). b) B4b: the tariff applied while the public lighting assets were owned by the distributors. It is estimated that this tariff was 9.6% higher than B4a, with the difference representing the remuneration for distributors´ operation and maintenance of the public lighting system. This tariff is being phased out in accordance with ANEEL Normative Resolution 414/2010. 20 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Unlike other consumer categories, where billing is 3.3.1.2 - Equipment prices done on the basis of actual, measured consumption and calculated according to a set tariff, public lighting LED prices for public lighting are rapidly falling in utilities are not required to install meters. This terms of US$ per lumen. From 2007 to 2012, the cost means that, in many cases, electricity consumption per lumen fell around nine times, from US$32 to US$ of the public lighting network is estimated on the 3.45 per 1,000 lumens. This price reduction was partly basis of the quantity and power rating of the installed due to the increased luminous efficiency of LEDs, lighting equipment36. The total power of the installed which doubled from around 70 lumens per watt to 130 equipment is multiplied by a fixed number of hours lumens per watt, as shown in Figure 9. Experts suggest of daily use to calculate the estimated price of kWh / that price reductions are in the range of 8–10% per day consumed by the public street lighting services37. year, and some studies forecast that price reductions could be as much as 16% per year38. Figure 9 - Average price of LEDs (US$ per 1,000) Price;klm Average lm/1W pk g $32 130 120 110 98 $22 $16 77 70 $13 $6.25 $3.45 2007 2008 2009 2011 2010 2012 Source: Strategies Unlimited International LED prices for recent public lighting of projects. Bloomberg studies suggest that prices tend projects in the United States, China and Mexico range to stabilize for projects involving more than 100,000 from US$250 to US$500 per light point. The costs in luminaires. This is the case, for example, of Los Angeles, India, which has a strong local LED manufacturing which achieved prices of US$300 per point on highways industry, are as low as US$150-200 per point, although and US$245 on smaller roads (2012 prices). with lifespan specifications below the Global Lighting Challenge standards. See Annex 4 for details. There is still a significant gap between the above- mentioned prices and the retail prices of LED lamps in The final price of LEDs is likely to depend on the scale Brazil. This is basically due to low local content in 36 Also called measurement by agreement (avença). 37 ANEEL determined that public lighting is to be used for 11 hours 52 minutes a day except in the case of public parks and / or tunnels, which normally have their own consumption measure. Municipalities can submit to ANEEL requests showing a deviation from the forecast standard. The city of Rio de Janeiro was able to prove to ANEEL an average daily consumption of 11 hours 32 minutes. 38 FOOTE, J .; WOODS, E. (2014).”Smart street lighting: LEDs, communications equipment and network management software for public outdoor lighting: market analysis and forecasts. ” Navigant Research. 21 3. Public street lighting market in Brazil lighting systems. This section focuses on two key incentives: (i) economic and financial incentives created by regulatory changes related to the transfer of public lighting assets to the municipalities; and (ii) political incentives presented by a combination of the relatively short time needed for implementing new public lighting compared to other infrastructure projects, and by the prospect of immediate positive impacts on community residents. 3.3.2.1 - Regulatory changes Municipalities, as owners of the assets, are responsible for regulating the public lighting market. However, for historical reasons, certain municipalities did not claim ownership and, as a result, the services remained the responsibility of the city’s electric utility. Most of these municipalities whose assets were under the responsibility of the electric utility were in the states of luminaires (typically 30%), and to taxes of around 75% São Paulo, Minas Gerais, Ceará, Pernambuco, Amapá, on imported components (i.e., double those levied on Roraima and parts of Paraná; together they represent domestically-manufactured high pressure sodium around 42% of all Brazilian municipalities. lamps). Furthermore, market prices in Brazil have not yet been impacted by the potential for lower costs Following a series of normative resolutions and public generated by large-scale projects, which are still in consultations, ANEEL Resolution 587, issued in 2013, their infancy. determined that by the end of 2014 all public lighting assets were to be transferred to the municipalities, with Given the falling prices, adoption of LED technology for cities assuming the management of all the activities public street lighting is virtually inevitable over time. related to the provision of public street lighting services However, the pace of full-scale conversion will depend previously carried out by the electric utilities. on a city’s financial status and the political will that is needed to prioritize these projects. Since the transfer, public lighting assets can no longer be included in the balance sheets of the old owners 3.3.2 - Opportunity 2- Incentives (utility companies) and cannot be calculated in the energy distribution Regulatory Asset Base (RAB), and for municipalities to invest therefore have no further impact on ANEEL-calculated in their lighting assets tariffs. Current regulatory and economic conditions in Brazil Figure 10 shows the asset transfer situation in Brazil are major incentives for municipalities to invest in the (states highlighted in yellow have recently exercised modernization (and efficiency) of their public street their rights of ownership). 22 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Figure 10 - Most recent municipalities to transfer their public street lighting assets RR AP AM PA MA CE RN PI PB PE AC AL RO TO SE BA MT GO DF MG ES MS SP RJ PR SC RS Source: ANEEL39 The public lighting assets transferred to the The transfer of the public street lighting services may municipalities generally included the following incur additional costs for municipalities, especially individual components40: a reactor; lighting arm; light in view of the widespread lack of trained and fixture; lamp (sodium, mercury, metallic vapor, LEDs); experienced personnel to manage the new assets. photoelectric relay; and, in certain cases, a pole. It is However, because the municipalities now own the important to note that not all the assets needed to assets and are obliged to pay the costs of running provide the services were transferred to municipalities. the services, they now have an economic incentive This is a source of potential economic loss for the to invest in public lighting energy efficiency with the municipalities given that prior to the transfer of assets goal of reducing the burden of electricity costs on they shared some of the distributors’ resources, such municipal budgets. as vehicle maintenance. The vertical disintegration of the public street lighting 3.3.2.2 - Political opportunities service entails new transaction costs for municipalities There is also evidence that the Brazilian population involving the transferred assets, and it raises a number greatly values the significant improvement in the of issues that cities must now confront directly, such as: perception of security that lighting modernization coordination with the energy grid operator; technical/ with LEDs can provide. In light of the current economic regulation of the street lighting service; and macro-fiscal crisis, Brazilian mayors have few other the need to invest in equipment, etc. opportunities during their four-year mandates to Available at: 50.000 > 0.6 >0 B 88 20.000–50.000 > 0.6 >0 C 329 < 20.000 = 0.6 <0 D 887 < 5.000 > 0.6 <0 E 3.406 < 2.000 > 0.4 <0 F 813 < 2.000 < 0.3 > 50 Light yellow = good; Light orange = moderate; Red= limited Source: World Bank Group; Pezco Consultoria The regrouping of the clusters in order of relevance resident population, light points and estimated is presented in Table 20 — number of municipalities, capital expenditure required for LED conversion. 63 The list provided in Annex 3 nominally identifies the municipalities contained in each group, also informing the cluster in which they were originally classified. 42 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 20 - Municipal groups in order of relevance Group Population Light points Required investments (R$) Total Total Total Average % Average % Average % (millions) (millions) (billions) (millions) A 59,9 29 1.274.015 5,1 27 107.499 7,7 27 161,3 B 23,8 12 270.041 2,8 15 31.490 4,2 15 47,2 C 14,7 7 44.701 2,1 11 6.303 3,2 11 9,5 D 23,0 11 25.967 2,2 12 2.437 3,3 12 3,7 E 64,4 32 18.921 5,1 28 1.493 7,7 28 2,2 F 18,6 9 22.894 1,2 7 1.533 1,8 7 2,3 TOTAL 204,4 100% 36.704 18,4 100% 3.302 27,8 100% 5,0 Source: Compilation and estimates by the World Bank Group and Pezco Consultoria64. Tables 19 and 20 above show that the municipalities and 12% of the light points in the country. Cities in in Groups A and B have a good scale of public street Group E have a smaller scale (less than 2,000 light lighting (over 20,000 points) as well as strong fiscal points) and moderate fiscal management. This group management. These two groups represent only 3% of includes the highest number of municipalities and the country’s municipalities, although they contain populations of all the groups, representing 61% and 41% of the population and 42% of the light points. 32% of the country, respectively, and approximately Group C has a relative good scale of lighting (typically 20% of Brazil’s light points. Finally, the cities in more than 20,000 points) and relatively good fiscal Group F are typified by their small scale public management. This group represents 6% of Brazil’s lighting (under 2,000 light points) and limited fiscal municipalities, 7% of its population and 11% of the management, representing 15% of the country´s country´s light points. Group D is characterized by its cities, 9% of its population, and 7% of all the country´s relative low scale (typically less than 5,000 points) light points. with relatively good fiscal management. It includes 16% of Brazilian municipalities, 11% of the population 64 Equipment prices based on a survey by World Bank Group in June 2015. Prices are estimated at R$1,500 per light point (excluding “smart” controls) Prices do not include the potential impact on the scale of purchases. Official exchange rate of from R$ 3.5/US$ (05/13/2016). 43 4. Mapping Brazilian municipalities for public lighting 44 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 5 - Business Models for Public Street Lighting in Brazil 5.1 - Overview of the Any business model for public street lighting modernization projects involves similar functions Primary Functions performed by different stakeholders in each model. This section presents some of the most important Considered in functions for a public street lighting project, including: generation of resources, energy supply, equipment Business Models supply, planning and bidding procedures, financing, purchase and installation of lamps, system operation and maintenance. Table 21 outlines these functions and identifies the key stakeholders involved. Table 21 – Functions and key stakeholders for public street lighting projects Function Stakeholder (s) a) Generation of resources for improving street lighting Consumers, via COSIP b) Energy supply at the light point Electric utility company c) Supplying equipment to the project LED manufacturers d) Planning and bidding Municipality or private sector e) Obtaining financing Municipality or private sector f) Purchasing lamps Municipality or private sector g) Implement installation Municipality or private sector h) Operation and maintenance Municipality or private sector Source: World Bank The first three functions and actors listed in Table 21 data in the section Models 1–8, details of these three — resource generation, power supply, and equipment functions, risks and mitigating factors are contained supply — remain virtually the same regardless of in following Table 22. the project´s business model. To avoid repetition of 45 5. Business models for public street lighting in Brazil Table 22 – Key stakeholders applicable to all the business models Functions Stakehold- Details Project risks Mitigating factors ers Generation of Consumers COSIP (included in elec- Failure to enjoy ben- To launch outreach campaigns and resources for tricity bills) paid on time or efits; dissatisfied with introduce systems to receive and payment of if necessary to the munic- the civil works. resolve complaints; to prepare with services ipality; enjoy the benefits the electric utility company a robust of better public lighting procedure for collecting COSIP. services; provide feedback about quality of the service. Energy supplied Electric utili- Provide energy at the light Energy supply short- Ensure clarity in the operational at the light point ty company point; sign an operation age; failure to fulfill agreements; increase the capacity agreement with the SPV; agreements with the of the municipality to oversee the collection and transfer municipality regarding charging process related to electricity COSIP receipts. collection of COSIP. billing and COSIP collection. Supply of equip- Manufactur- Provide equipment to the Low quality and/or Ensure performance guarantee of the ment for the ers project. high cost of equip- equipment, ideally for a minimum of project ment. 10 years (e.g., the end of the useful life of the LEDs); to provide support to the authorities and/or public banks to increase domestic manufacturing of LEDs with a view to reducing import costs of the same; reduce taxes on national production of LEDs (for limited time). Source: World Bank The remaining five functions — planning well as interviews with selected key stakeholders and bidding, financing, lamps purchasing, and an analysis of Brazilian and international lamps installation, and system operation and documentation. The eight business models are as maintenance — are specifically detailed for each follows, and are described in detail in this section: business model in the next section. • M1 — Municipal Public-Private Partnership (PPP) 5.2 - Introduction • M2 — PPPs with Municipal Consortium65 to Business Models • M3 — Municipal Financing • M4 — Electric Utility Programs Following the classification of municipalities, this study identified eight business models based • M5 — Energy Service Companies (ESCOs) on fiscal, demographic and scale variables, as 65 Note that while the formation of a municipal consortium is a necessary precondition for the M2 model, aggregation (to create scale, reduce transaction costs, etc.) could be achieved by using municipal consortia in several other models. 46 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning • M6 — Municipal Consortium or Central The PPP model is characterized by the presence Procurement Agent of a concessionaire, to which the municipality grants an administrative concession covering a • M7 — Self–Funding wide range of responsibilities.66 PPPs can be used • M8 — Transfer of Luminaires to undertake all the tasks involved in modernizing a municipality´s public street lighting network and The range of options provided in the eight business in providing efficient services to the municipality, models are meant to provide feasible solutions for to include the installation, maintenance and energy efficient public street lighting for the wide operation of the system for the duration of the variety of Brazilian municipalities. contract.67 The contract term for an administrative PPP can vary from five to 35 years. 5.3 - Presentation The concessionaire is normally a Specific of Business Models Purpose Vehicle (SPV) formed by the winning consortium, which may include, but is not limited to, an operator and an equipment manufacturer. This section contains a detailed description of The concessionaire is responsible for seeking these eight business models, focused on: financing to cover initial equipment costs, while the municipality is responsible for reimbursing the 1) A description of the model and cash flows; concessionaire via monthly payments. The selection 2) Main agents, risks and mitigating factors; of the concessionaire is by public tender, in which the main criterion of success is the minimum value 3) A brief description of the model’s advantages/ offered by pre-qualified candidates. Contracts can disadvantages; and include fixed-payment schedules and performance 4) Suggestions of which groups of municipalities clauses and can be subject to periodic reviews (from the aforementioned categories A–F) would specified in the concession agreement. be good candidates for implementing each model. If the municipality has implemented COSIP, payments to concessionaires would primarily 5) Evolution of the model in Brazil be funded by COSIP revenues from electricity At the end of the section, two tables summarize the consumers. If allowed by municipal law, the functions and main actors involved in each model COSIP revenues could be collected directly by the and map the application of the eight models to the concessionaire and transferred to a municipal fund six groups of municipalities identified in Section 5. or escrow account. The concession agreement would contain clauses governing payments from these accounts. In the event of COSIP funds not Model M1: Public-Private being sufficient to make the monthly payments (or in the absence of COSIP), the municipality may Partnership (PPP) need to call on the municipal budget to cover remaining payments due. Figure 15 illustrates Description of the model and cash flow the cash flows and the main actors involved in a sample PPP model. 66 Authorized by Law No. 11,079 of 2004, amended by Law No. 12,766, of 2012, the management concession corresponds to the contract in which the final contractor of the services is the public authority, which reimburses the concessionaire monetarily without charging tariffs or public prices to the user. The other approach permitted by the same law is the Sponsored Concession, in which part of the concessionaire remuneration comes from the collection of user fees. 67 Concession contracts could also include ancillary services, such as 4G mobile telephony, WiFi, monitoring services, etc. The apportionment of profits from these revenues must be specified in the concession agreement. 47 5. Business models for public street lighting in Brazil Figure 15 - Example of the structure of a PPP model Financier(s) Municipality Loans, debentures and/or grants $ Reembursement $ Additional Efficient street contributions lightning services SPV of EE public street Municipal Account / lighting consortium Monthly payments $ Escrow Account COSIP $ Manufacturer Operator Electricity Utility COSIP $ Financial flows Physical flows End Users Main actors, risks and mitigating factors main risks that could affect the project throughout the process, and contains recommendations for Table 23 summarizes the principal actors in each of mitigating such risks. the key stages of the PPP model. It also shows the 48 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 23 – Matrix of functions and actors in the M1 — municipal PPP model Model M1 – Municipal PPP Functions Actors Details Project risks Mitigating factors Planning and Municipality Prepare public notice and Lack of sufficient qualified Capacity-building or standard man- bidding organize bidding process for staff to prepare the public ual. contracting public lighting notice and/or assess pro- concessionaire. posals; lack of competition to ensure maximization of financial resources. Audit Court Analyze and validate docu- Delaying or suspending Higher involvement of audit court in ments/comply with require- process. preliminary consultations and at the ments published in the public initial stages of the process. tender notice. Consortia of Companies constituted to com- Lack of interested parties Municipality strike adequate balance companies pete for the concession; submit in the bidding process. between risk and return in the public project bid at a competitive bidding documents. price. Obtaining Public street Plan and implement in- Impossibility to obtain fi- Provide risk reduction mechanisms finance lighting con- vestment project; leverage nancing; financing not suf- (e.g., guarantees and insurance cessionaire resources for the investment. ficiently viable to support against technical and credit risks); the project throughout its requirements to improve quality of duration. bids (e.g. demanding proof of ability to undertake project). Lamp procure- Financiers Provide funding at competitive Lack of interest in invest- Provide data to facilitate understand- ment rates for the CAPEX. ing; high interest rates. ing of the risks involved. Implemen- Public street Conduct bidding process for Lack of capacity to cope The public tender must follow best tation and lighting con- lamps. with technical subjects; national/international practices; bid- installation cessionaire lack of adequate guaran- ding to include international competi- tees for the equipment. tion; specialist advice. Operation and Public street Manage the lamps exchange Lack of expertise about the Preparation of robust technical stud- maintenance lighting con- process and related services public lighting network, ies; capacity building. cessionaire (civil works, disposal of existing resulting in delays and/or lamps etc.). cost increases. Planning and Public street Provide efficient public street Inability to make energy Technical and performance guaran- bidding lighting con- lighting services savings and to meet the tees. cessionaire standard demanded by the municipality. Municipality Regulate and oversee public Lack of capacity- building; Regulatory and contractual capaci- street lighting services. political interference. ty-building; provision of transparent contract mechanisms to limit political interference. Source: World Bank. 49 5. Business models for public street lighting in Brazil Advantages and disadvantages One of the advantages of the PPP model is that the municipality can receive a full guarantee from the concessionaire to cover the useful lifespan of the LEDs (normally 10–15 years), since the concessionaire is responsible for purchasing, installing, operating and maintaining the lights over the entire concession period. Moreover, PPPs explicitly permit contract payments to be linked to project performance. This can result in economic and financial savings, since any consortium formed to provide equipment and services will have an interest in maximizing energy efficiency gains that will benefit both parties66. All these factors, when combined, can potentially reduce the performance risks faced by the public authority. Despite its advantages, the PPP model has some drawbacks, such as the time and transaction costs involved in preparing a PPP. These can be substantial, particularly since there are few examples of PPPs in the public lighting sector to guide municipalities and auditors (e.g., the government Audit Courts) in the process of structuring and evaluating contracts. Identification of groups to which the model would Groups A and B, and perhaps some of the larger apply cities in Group C69. Including only Groups A and B, this represents a total of 135 municipalities (2.4% of Only some of the municipal groups identified in this the total cities in Brazil). Regardless of the relatively study would be in a position to award concessions for small number, these municipalities contain 40.9% public lighting services using the PPP mechanism. of the country’s total population. Furthermore, Structuring a PPP requires the municipality to the R$1.9 billion worth of investments required initiate technical training of its personnel and make a for retrofitting the street lighting systems of these substantial investment in studies, as well as passing municipalities with LEDs amounts to 42.4% of the specific legislation (e.g., establishing an escrow investment needed for the entire country. account for COSIP). It also requires the municipality to undertake monitoring procedures and ensure Evolution of the model in Brazil compliance with special accounting requirements under Brazilian law. Several municipalities have begun to adopt the PPP model in the form of administrative concessions. On the basis of this assessment, potential candidates Table 24 presents a summary of PPP projects in the for PPPs in public street lighting include cities in Brazilian market as of October, 2016. 68 This applies to cases where the public street lighting concessionaire is responsible for paying for the electricity consumed. In cases where the municipalities continue to exercise its responsibility (e.g., the Belo Horizonte PPP), the contract could include explicit incentives (e.g. a bonus) for efficiency gains that exceed agreed values. 69 Note that at present 50 municipalities have begun organizing PPPs on an individual basis. However, smaller municipalities not belonging to Groups A or B have also been investing in the PPP solution. 50 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 24 - PPP projects in Brazil (according to project stage) Status of PPP Number Awarding Municipalities Contract signed 5 Urânia-SP; São João de Meriti-RJ; Caraguatatuba-SP; Guaratuba-SP; Belo Horizonte-MG Winner declared 4 Araguaína-TO; São José de Ribamar-MA; Marabá-PA; Mauá-SP Bidding in progress 7 Contagem-MG; Goiatuba-GO; Campo Maior-PI; Goianésia do Pará-PA; Guaíra-SP; São Cristóvão-SE; Teixeira de Freitas-BA Public consultation. 18 Almirante Tamandaré-PR; Atibaia-SP; Feira de Santana-BA; Guarapuava-PR; Lins-SP; Completed Maceió-AL; Nova Iguaçu-RJ; Aparecida de Goiânia-GO; Dois Vizinhos-PR; Dourado-SP; In- humas-GO; Patrocínio-MG; Pederneiras-SP; Porangatu-GO; São Bernardo do Campo-GO; Uberlândia-MG; Vespasiano-MG; Vitória-ES Public hearings 3 Muriaé-MG; Guanambi-BA; Penedo-AL initiated Request for Expres- 17 Bertioga-SP; Boa Vista-RR; Camaragibe-PE; Governador Valadares-MG; Içara-SC; Itu- sions of Interest iutaba-MG; Jaboatão dos Guararapes-PE; Sorriso-MT; Várzea Grande-MT; Cariacica-ES; (PMI) closed Esmeraldas-MG; Garopaba-SC; Hortolândia-SP; Maringá-PR; Niterói-RJ; Pará de Minas- MG; Rio Verde-GO Request for Expres- 15 Carolina-MA; Delmiro Gouveia-AL; Imbituba-SC; Jacundá-PA; Rolim de Moura-RO; Salto- sions of Interest SP; Salvador-BA; Açailândia-MA; Água Boa-MT; Araçatuba-SP; Araucária-PR; Distrito (PMI) initiated Federal; Gurupi-TO; Palmeira-PR; Porto Nacional-TO Modelling initiated 5 Betim-MG; Caruaru-PE; Formosa-GO; Ribeirão Preto-SP; Guarulhos-SP Publication of intent 16 Angra dos Reis-RJ; Aracaju-SE; Diadema-SP; Florianópolis-SC; Foz do Iguaçu-PR; Irace- mápolis-SP; Ponta-Grossa-PR; Santos-SP; Sete Lagoas-MG; Sombrio-SC; Sorocaba-SP; Valparaíso de Goiás-GO; Votuporanga-SP; Catanduva-SP; Palmas-TO; Porto Alegre-RS Cancelled / Sus- 95 Canela-RS; Cuiabá-MT; Uberaba-MG; Consórcio CIGIP-AL (Água Branca; Anadia; Barra de pended Santo Antônio; Barra de São Miguel; Batalha; Belém; Belo Monte; Boca da Mata; Bran- quinha; Cacimbinhas; Cajueiro; Campestre; Campo Grande; Canapi; Capela; Carneiros; Colônia Leopoldina; Craíbas; Delmiro Gouveia; Dois Riachos; Estrela de Alagoas; Feira Grande; Flexeiras; Igaci; Inhapi; Jacaré dos Homens; Jacuípe; Japaratinga; Jaramataia; Jequiá da Praia; Jundiá; Lagoa da Canoa; Major Isidoro; Mar Vermelho; Maragogi; Mara- vilha; Marechal Deodoro; Mata Grande; Matriz de Camaragibe; Minador do Negrão; Murici; Novo Lino; Olho d’Água do Casado; Olho d’Água Grande; Ouro Branco; Palestina; Pão de Açúcar; Pariconha; Paripueira; Paulo Jacinto; Pilar; Pindoba; Poço das Trincheiras; Porto Calvo; Porto de Pedras; Porto Real do Colégio; Quebrangulo; Rio Largo; Santana do Ipane- ma; Santana do Mundaú; São José da Laje; São Luís do Quitunde; São Sebastião; Senador Rui Palmeira; União dos Palmares; Viçosa); Barbacena-MG; Barueri-SP; Breu Branco-PA; Consórcio CONIAPE (Bezerros; Bom Jardim; Brejo da Madre de Deus; Casinhas; Frei Miguelinho; Jataúba; João Alfredo; Orobó; Riacho das Almas; Santa Cruz do Capibaribe; Santa Maria do Cambucá; São Caitano; São Joaquim do Monte; Surubim; Taquaritinga do Norte; Toritama; Vertente do Lério; Vertentes); Cascavel-PR; Caxias-MA; Santo André-SP; São Paulo-SP; Três Corações-MG Total 185 Source: Radar PPP (www.radarppp.com), October 3, 2016. Radar PPP updates its data daily. 51 5. Business models for public street lighting in Brazil As shown in Table 24, by October 2016, 185 A PPP contracted by a consortium has the same Brazilian municipalities initiated processes or general characteristics and modus operandi have demonstrated their intention to contract of a consortium contracted by an individual for public street lighting services under a PPP 68. municipality. These include 101 municipalities (54.6% of the total) which initiated PPP procedures individually. Due to legal restrictions currently in force, These range from large regional capitals such the legal entity of the public consortium is not as São Paulo and Belo Horizonte, and smaller permitted to receive financing directly. However, cities such as Urânia in the interior of São Paulo the concessionaire involved in a PPP bid by a state, which already has awarded the contract. consortium can obtain financing in the capital The remaining 84 municipalities (45.4%) have market to undertake the necessary investments. joined together in two public consortia — Figure 16 summarizes the main actors involved CONIAPE (representing 18 municipalities in the and the cash flow of the PPP model contracted by State of Pernambuco) and CIGIP in the State a consortium of municipalities. of Alagoas (with 66 municipalities). However, more than half of these PPPs have either been cancelled or suspended, and only five have signed contracts. This demonstrates the immediate need for increased capacity building to help cities adequately prepare PPP tenders from the outset. Model M2: Consortium of Municipalities Using PPP Description of the model and cash flow Given that it will be challenging and not economically attractive for many small- and medium-size municipalities to individually grant concessions using the PPP model, municipal consortia are a possible solution to generate the scale necessary for the implementation of a PPP.71 International experience shows that there are huge economies of scale involved in the procurement of luminaires by using this model. Once established, the consortium of municipalities is able to contract a public-private partnership and the consortium itself functions as the grantor of operate street lighting services through a PPP. 70 Compiled from the RadarPPP database and data obtained by Pezco Consultoria. 71 The public consortia concept applied to the federative entities was established by Law 11,107 of 2005. 52 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Figure 16 - Example of the structure of a PPP with municipal consortium model End User Cosip $ Financial flows Physical flows Electricity Utility Cosip $ Municipal Consortium Financier(s) Municipality 1 Municipality 2 Municipality 3 Loans, debentures, $ $ and/or grants $ R Efficient public lightning Service SPV of EE public SPV of Consortium lighting consortium of Municipalities Manufacturer Operator $ Payment $ Escrow Account Main actors, risks and mitigating factors risks that could affect the project throughout the process, and contains recommendations for Table 25 lists the main actors involved in the PPP mitigating these risks. model based on a consortium, indicates the main Table 25 - Matrix of functions and actors in Model M2 — PPP consortia M2 – Consortia for PPPs Functions Actors Details Project risks Mitigating factors Planning and Municipalities Formation of the consor- Lack of coordination between Clear governance rules in bidding tium. the municipalities in the the consortium; capaci- consortium; insufficiently ty- building or standard qualified personnel. manual. Consortia of mu- Prepare public notice and Lack of coordination between Clear governance rules in nicipalities organize bidding process the municipalities in the the consortium; capaci- for contracting public consortium; insufficiently ty- building or standard lighting concessionaire. qualified personnel. manual. Audit Court Analyze and validate Delay or suspend the process. More involvement in pre- documents/comply with liminary consultations and requirements published in at the initial stages of the the public tender notice. process. Consortia of com- Companies constituted to Lack of interested parties in Selection by the municipali- panies compete for the conces- the bidding competition. ty of adequate ratio between sion; submit project bid at risk and return in the public a competitive price. tender documents. 53 5. Business models for public street lighting in Brazil Obtaining SPV of the consor- Structure and leverage re- Complexity in the financial Provide mechanisms for finance tium sources for the investment. structuring due to the large guarantees and insurance number of participants; against technical and credit stakeholders; failure to re- risks, provide clear rules imburse the consortium can for the governance and put all the members of the treatment of default. consortium at risk. Financiers Provide funding at compet- Lack of investor interest; high Supply sufficient data itive rates. interest rates. to financeirs in order to facilitate understanding of the risks. Procurement of Public street light- Bidding for procurement of Lack of capacity to cope with The public notice to adhere lamps ing concessionaire lamps. technical subjects; lack of to best national/interna- adequate guarantees for the tional practices; bidding to equipment. include international com- petition; specialist advice Manufacturers Production/importation of In the event of local pro- Overwhelming demand for LEDs for the municipali- duction, the risks involved LED lamps can provide an ties/SPV. in financing the installation incentive for factories to be of the manufacturing plant; established in Brazil. in the event of imported LED products, the exchange rate and import taxes on the equipment. Implemen- Public street light- Manage the lamps instal- Lack of expertise about the Preparation of consistent tation and ing concessionaire lation process and related public lighting network re- technical studies; capacity- installation services (civil works, sulting in delays and/or cost building. disposal of existing lamps increases. etc.). Operation and Public street light- Provide efficient public Failure to deliver expected Technical and performance maintenance ing concessionaire street lighting services. energy savings, or failure to guarantees. comply with the standard de- manded by the municipality. Municipalities Regulate and oversee Lack of qualified staff; politi- Regulatory and contractual the public street lighting cal interference. training; provision of clear services and those of the contract mechanisms to public lighting concession- limit political interference. aire. 54 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Advantages and disadvantages for the Management of Electric Energy and Public Services (CIGIP)72, consisting of 66 municipalities The PPP consortium approach has the advantage in the state of Alagoas. However, this consortium of allowing a wider range of municipalities to has not yet offered solutions for modernizing or participate in a PPP scheme compared to the operating the public lighting network, but rather has individual PPP model. Furthermore, a potential focused to date on boosting the municipality - utility financier may be attracted to the scheme given the relationship to ensure technical support, equipment, greater spread of political and credit risks of the consulting services and training for running the various municipalities, e.g., subject to financial and street lighting network and the municipal public governance arrangements, a single municipality buildings of consortium members. Meanwhile, the default would not necessarily result in default by the state of Minas Gerais has deployed a strategy to consortium SPV. encourage municipalities to form consortia based on initiatives between the state government, the State However, if most municipalities´ profiles suggest Audit Court, CEMIG (the state electricity distributor) high political and credit risk, the PPP consortium and associations of municipalities. In Pernambuco, approach can generate risks for both financiers and the Inter-Municipal Consortium of Pernambuco consortium members. Transaction costs and risk Hinterland and Borders (CONIAPE) initiated a perception may also be high in view of the more public debate about the possibility of a PPP, but complex governance of a consortium, resulting in a this has been suspended. Tendering procedures are lack of interest by both lenders and municipalities, currently being organized in the hope of obtaining as well as potentially incurring high financing costs. comprehensive management and maintenance of the public street lighting systems of the 17 municipalities Identification of groups to which the model would in the consortium.73 apply Although national interest in the consortium Aggregation in a consortium is a more appropriate mechanism for running public services (including solution for Group C municipalities, which have low lighting networks) has increased, no consortium scale (less than 20,000 luminaires), but relatively model has yet been used to create an SPV to secure good fiscal management. 329 municipalities are in financing to implement a street lighting project. this group, and some relatively small municipalities More technical advice is essential to help interested in Group B could also participate. By including only municipalities to implement this model, particularly the Group C municipalities, the M2 model could be with regard to legal aspects involving the structuring applicable to 6% of Brazilian municipalities that of SPV financing and governance schemes. represent 7% of the total population and R$3.2 billion of investment (11% of the investment needed throughout Brazil). Model M3: Municipal Financing Evolution of the model in Brazil Description of the model and cash flow There is a trend towards forming public consortia The high level of upfront capital needed to in Brazil. The consortium to include the public implement LED technology is the main challenge street lighting sector was the Public Consortium to the conversion of the public lighting network. 72 CIGIP’s public notice was special in that it provided for investments in various sectors, including public lighting, distributed generation, emissions mitigation and control of electricity consumption in public buildings, etc. 73 Available at: . Accessed in March 2016. 55 5. Business models for public street lighting in Brazil Leaving aside the concession model, the main municipality has not implemented COSIP, or if revenue responsibility for raising financing and undertaking from COSIP is insufficient to defray loan or other the required investments will rest with the municipal payments, the municipality would need to use funds administration, normally through loans or the from its general municipal budget to finance any issuance of bonds (debentures). outstanding amounts of debt payment. In this model, the municipality would be responsible for executing or As with PPPs, municipal debt repayment could outsourcing the implementation of the project, and for be financed by COSIP revenues (if available) and overseeing the operation and maintenance services. transferred to a municipal fund (or escrow account, Figure 17 summarizes the main actors and the cash if allowed under municipal legislation). If the flow of the municipal financing model. Figure 17 – Example of structure of the municipal financing model Financier(s) Funds $ Reimbursement $ Equipment Services Manufacturer Municipality Outsourced Company Additional contributions $ COSIP$ Payment $ Payment $ Municipal Fund COSIP $ Electricity Utility COSIP $ Financial flows Physical flows End Users Main actors, risks and mitigating factors that could affect the project throughout the process and contains recommendations for mitigating them. Table 26 lists the main actors involved in the PPP model based on a consortium, indicates main risks 56 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 26 - Matrix of functions and actors in the M3 model — municipal financing M3 — Municipal Financing Functions Actors Details Project risks Mitigating factors Planning and bidding Municipality Design and prepare Lack of qualified staff. Public notice to adhere to the project; handle best national/international the bidding processes practices; competitive bidding; for equipment and specialist advice. services. Obtaining finance Municipality Structure and leverage Reaching upper limit of Provide guaranteed mech- resources for promot- indebtedness capacity; anisms against credit risks ing investment. market has credit risk including robust legislation for perception; delays caused COSIP (if it exists); Encourage by lack of sufficient fi- market interest in the munici- nancing at the beginning pality´s credit status during the of the project. first stages of the project. Financiers Provide financial Lack of investor interest; Provide data to facilitate un- resources at compet- high interest rates. derstanding of the risks. itive rates for capital expenditure (CAPEX). Procurement of lamps Municipality Organize bidding pro- Lack of capacity to cope Public notice to adhere to cess for procurement with technical subjects; best national/international of lamps. lack of adequate equip- practices; competitive bidding; ment guarantees. specialist advice. Installation Municipality Manage the process or Lack of expertise about Preparation of consistent oversee the services the public lighting system technical studies; capaci- of an outsourced resulting in delays and/ ty-building. company. or cost increases, or lack of qualified oversight staff (in the event of super- vision of an outsourced company). Operation and main- Municipality Manage the process or Lack of qualified staff. Technical and/or oversight tenance oversee the services capacity building. of an outsourced company. 57 5. Business models for public street lighting in Brazil Advantages and disadvantages Identification of groups to which the model would apply A key advantage of this business model is the low complexity involved in structuring the This business model is most applicable to project due to the smaller number of actors and the smaller municipalities in Groups B (88 the possibility of procuring equipment using municipalities) and C (329 municipalities), since the routine bidding process provided for in Law these have good fiscal management and good 8,666/1993. This can result in lower transaction scale, but perhaps without sufficient scale to justify costs compared to the PPP models. This model the transaction costs associated with structuring a therefore represents a possible way forward for PPP. This model could theoretically apply to 7% of municipalities with good fiscal management, but Brazilian municipalities, representing 19% of the without sufficient scale to justify the transaction total population and R$7.4 billion of investment costs of a PPP. (26% of the investment needed throughout the country). However, as explained in Section 4.4 (Challenges for LEDs in the Brazilian Context), municipal Evolution of the model in Brazil indebtedness capacity is restricted under the Law of Fiscal Responsibility, and many municipalities While loans contracted by a subnational entity are often unable to take on additional debt. with a financial institution to finance capital Even when municipalities have adequate fiscal investments is a traditional format in several space, they may wish to use these resources on countries, including Brazil, this model faces investments in other sectors for which it is more substantial fiscal and institutional constraints. difficult to attract private sector investment Even in Europe, where bank financing for the (e.g., schools). public sector is the dominant model, lack of access to credit for local governments acts as a brake on A further disadvantage of this model is that public investment (AFD, FCH, IPEA, 2014, p. 95). if the municipality outsources the operations Thus, credit for Brazilian municipalities generally and maintenance services, the performance involves loans from public banks supported by guarantee will usually be limited to five years - federal guarantees. the maximum period established by Law 8,666 / 1983. It follows that system operators will have In the course of the research and interviews no incentive to ensure technical performance conducted for this study, no examples were over the lifespan of LEDs (10-15 years). In the identified in Brazil of municipal financing being less likely event of the municipality assuming used to raise capital for investment in the the operations and maintenance services, there modernization with LEDs. is a risk that the municipality may not possess the human and technical capacity to manage One way to catalyze this business model would the process. Thus, this model could result in be implementation of public policies that would municipalities having to assume the majority of provide exceptions to the indebtedness limits technical and operational performance risk. governed by the Fiscal Responsibility Law for investments in municipal energy efficiency. As explained previously, debts taken on by 58 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning municipalities under the PROCEL-RELUZ program Regardless of the trend for energy utilities to were not subject to the Fiscal Responsibility Law be less involved in the public lighting sector, it indebtedness restrictions, on the basis that the is possible to envisage a scenario where they loans would reduce municipal energy costs, could play an important role in scaling up the and thus improve the fiscal situation of the roll out of LEDs throughout the network, as well municipalities in the long term. However, cities as becoming involved in a wider range of energy are unable to benefit from this exception since efficiency projects. Given the current incentive the PROCEL-RELUZ program stopped disbursing structure, changes of a legal and regulatory resources for street lighting modernization nature are likely to be necessary to make this (although PROCEL-RELUZ resources will increase happen. due to the recent changes mentioned in Section 4.4.4, albeit at a relatively small scale). The federal In some ways, this business model could be seen government could consider further exceptions to as a scaling up of the current PEE program. For assist municipal governments to obtain energy this to happen, the electric utility would be given efficiency-related resources from other sources a broader mandate to invest in energy efficiency (i.e., capital markets), with a public body such as (e.g., higher than the 0.5% of the current annual PROCEL acting as the supervisory authority to net revenue - say up to 1.0%). 74 Part of this increase approve and guarantee the submission of high would be earmarked to the public lighting sector quality projects that meet energy-saving targets. by way of loans to municipalities. However, unlike under the current PEE program Model M4: Electric Utility where utilities do not recover costs from the public Company Programs beneficiaries of the energy efficiency investments, the electric distribution utility would lend to Description of the model and cash flow municipalities at a sustainable interest rate. If needed to make the investments viable, the In the two national programs for the street lighting interest rate could be below commercial rates, sector, PROCEL-Reluz and PEE, electricity utilities the utilities could recover part of their outlay via have a key role in the financing and/or facilitation a slight increase in end-user tariffs (as is done of investments for large-scale public street under the current PEE program). lighting projects. However, these two programs currently have minimum potential for LED retrofit The profits from the program could capitalize a programs; the sole source of capitalization of revolving fund that would be used to fund new PROCEL-Reluz has been discontinued (although investments in street lighting energy efficiency the overall PROCEL program will receive new projects. The municipalities would use the loans, sources of funding from PEE), and PEE funds for together with their own municipal contribution public lighting have dwindled. Moreover, with the (funded by budgetary appropriation, COSIP or transfer of energy assets to the municipalities, another municipal source), to purchase the LED the utility companies are no longer involved in equipment for the project. Municipalities would the operational aspects of street lighting. use savings generated by the program (i.e., surplus COSIP revenue due to reduced electricity bills) to repay their loans. The municipality would Some U.S. utilities invest a higher proportion in energy efficiency (e.g. 4%). The largest investments in EE in the United States are done by gas and 74 energy concessionaires. 59 5. Business models for public street lighting in Brazil be responsible for managing the bidding process If program performance were adequate, the costs for equipment and for overseeing O & M (using of an increasingly self-sustaining program would the remaining COSIP resources or the municipal be reduced for all consumers. budget). Figure 18 - Example of structure of electricity utilities program model Revolving Fund Funds $ Repayment $ Electricity Utility End Users COSIP $ Loans $ Repayment $ COSIP $ Municipalities Municipal Account (s) $ Equipment Payments $ Services Financial flows Manufacturer Outsourced Company Physical flows Considering that this model would require further opportunities for investing in street lighting. regulatory changes, a short-term option would be for the energy utilities to earmark more PEE Main actors, risks and mitigating factors resources to the public street lighting sector. This may be possible for those with operations in Table 27 summarizes the main actors of each of areas with few low-income consumers or where the key stages of the electric utility programs these consumers already been serviced. The model, and indicates main risks that could affect new legislation eliminating the minimum 60% the project throughout the process and contains investment in low-income households could provide recommendations for mitigating them. 60 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 27 – Matrix of the functions and actors in model M4 — electric utility companies’ programs M4 — Electric Utility Programs Functions Actors Details Project risks Mitigating factors Planning and Municipality Design and prepare the Lack of capacity to prepare Hire expert advisors. bidding project; submit project a proposal or consistent proposal to electric project. utility for approval. Electric utility Design investment and Lack of qualified personnel Confirmation that the program has revolving program. to oversee the program. sufficient resources for its manage- ment ANEEL Agree to update the Failure to approve project Clearly communicate the benefits mechanism for using the concept. of the program, including climate PEE for engaging utility change targets and the possibility companies‘ financial of future self-sustainability of the interest in the sector. program Obtaining Electric utility Make financing resourc- Municipal demand for funds Limiting the criteria for municipal- finance es available for munici- exceeds supply. ities to receive resources from the palities. program (e.g. small and medium- sized municipalities). Lamp procure- Municipality Organize bidding for Lack of technically qualified Public bidding notice to adhere to ment or ESCO lamp procurement. personnel; lack of appropri- best Brazilian and international ate equipment guarantees. practices; bidding open to interna- tional competition; expert advisory. Implemen- Municipality Manage the process or Lack of knowledge of the Preparation of studies; ensure con- tation and or ESCO oversee the services public lighting system, sistent technical approach; capaci- installation provided by ESCO. resulting in delays and/or ty-building. increased costs, or lack of oversight capacity. Operation and Municipality Manage the process or Lack of capacity. Technical and/or oversight training. maintenance or ESCO oversee the services provided by ESCO. Advantages and disadvantages transaction costs and greater diversification, and possibly help reduce the cost of finance for A financing program led by electricity utilities has municipalities; and, because the model involves several advantages: borrowing costs will be lower lower costs, it could be useful for municipalities than those that municipalities could obtain in the with limited possibilities of raising funds. capital market; centralizing the funds leveraging process would result in economies of scale, lower The main drawback of the model is that ANEEL 61 5. Business models for public street lighting in Brazil would need to introduce regulatory changes increasing PROCEL’s investments in public street to provide incentives for the energy utilities lighting. This would be a first major step towards to allocate more investment for this sector. At encouraging the electric utility companies to play present, this would likely be difficult given the a more prominent role in modernizing Brazil’s recent regulatory trend towards reducing the public street lighting network. involvement of energy utilities in the public street lighting sector. Model M5: ESCO Model Identification of groups to which the model would Description of the model and cash flow apply This model involves municipal off-balance sheet This model could be applied mainly to the financing, where the investment is made by a third municipalities in Groups D and E, which represent party. In this way, the investments for modernizing more than 4,200 municipalities and around the public street lighting network are not 75% of the Brazilian population, and for which classified as municipal debt and do not affect a investment totaling R$9.5 billion would be municipality’s indebtedness limits. A company or required. In certain cases, the model could also a consortium of companies (SPV) raises funds, and be applied to groups in addition to Groups C and purchases/installs LED luminaires in exchange for F. The municipalities concerned are of relatively a fixed payment by the municipality. As in other small size, with average fiscal management. To models, the municipality would use COSIP and/ protect the utility as a lender, the municipality’s or municipal budget resources to pay the ESCO. credit risk could be mitigated with the use of an ESCOs are not normally involved in operation escrow account linked to COSIP revenues. Other and maintenance, but they would guarantee the types of credit risk mitigation could be considered technical performance of the lamps they install. for municipalities with more significant fiscal problems. The experience of PROCEL-RELUZ, ESCOs operate under one of two modalities: (i) a which explicitly required municipalities to provide share of the efficiency gains; or (ii) a fixed payment financial certification, could prove to be useful in for the investment made and for providing a this respect. product technical performance guarantee. Option (ii) has generally been preferred since it simplifies Evolution of the model in Brazil measuring and verification procedures and ensures a predictable cash flow, thus enhancing a As noted in Section 4.4.4, a recent bill of law project´s bankability. proposed the allocation of 20% of PEE resources to PROCEL (approximately R$100 million Figure 19 shows the main actors and cash flow annually), which would present an opportunity for involved in an ESCO model sample. 62 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Figure 19 – Example of structure of the ESCO model Equipment Repayment $ Manufacturer ESCO Financier Funds Payment $ Repayment $ Equipment Outsourced Services Municipalities Company Additional Payment $ COSIP $ contributions $ Municipal Account COSIP $ Electricity Utility COSIP $ Financial flows Physical flows End Users Main actors, risks and mitigating factors Table 28 summarizes the main actors in each of the key stages of the ESCO model, indicates the main risks that could affect the project throughout the process, and recommendations for mitigating these risks. 63 5. Business models for public street lighting in Brazil Table 28 – Matrix of functions and actors in model M5 — ESCOs M5 – Electric Utility Companies (ESCOs) Functions Actors Details Project risks Mitigating factors Planning and Municipality Prepare public notice Lack of sufficiently quali- Capacity building or standard manu- bidding and carry out bidding fied manpower to prepare al; competitive bids. process for contract- the public notice and/or ing ESCO (equipment, evaluate proposals; lack of installation and competition to guarantee possibly operation and maximization of municipal maintenance). resources. ESCOs Submit substantive Absence of interested parties Selection by the municipality of an project at a competitive in the bidding competition. appropriate ratio between risk and price. return in the public notice docu- ments. ESCOs Structure and leverage Lack of sufficient ESCO Increased lines of financing avail- resources for invest- resources to ensure the scale able for ESCOs (with public/non-re- ment. of investments needed for fundable grants if necessary); focus the entire project. on the ESCOs that are subsidiary companies of electric utilities; focus on small-scale projects. Obtaining finance Financiers Provide financial Lack of investor interest; Provide data to facilitate understand- resources at compet- high interest rates. ing of the risks. itive rates for capital expenditure. Lamp procure- ESCOs Promote bidding for Lack of capacity to cope with The public follows best national/ ment procurement of lamps. technical subjects; lack of international practices; bidding to adequate guarantees for the include international competition; equipment. specialist advice Implementation ESCOs Management of the Lack of knowledge of public Preparation of studies; use of consis- and installation process. street lighting network, re- tent techniques; capacity building. sulting in delays and/or cost increases. Operation and Municipality Manage the process or Lack of qualified manpower. Technical and/or oversight training. maintenance or ESCO oversee ESCO services. Advantages and disadvantages exposure to political fluctuations, etc., as would be required for the M4 model. A key advantage of this model is that, as with the M4 model, it involves an off-balance sheet financing option However, the scale and scope of this model are less for smaller municipalities. However, it does not require than that of M4 since the financial capacity of ESCOs the same degree of involvement by the regulator, is much less robust than that of an electric utility, and 64 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning because ESCOs would be more vulnerable to municipal the Regions 20 (R20) organization which has been credit risk (or less able to shield themselves from risk) instrumental in LED conversion programs in 13 Brazilian since they have no direct access to municipal resources cities. The financing concept is based on ”payment for linked to revenues from end users (e.g., COSIP, electricity the use of the asset,” using a model that involves a payments). municipal/private sector SPV for each municipality. In this scheme, the financial partner purchases the Identification of groups to which the model would apply luminaires and supplies them to the municipality at a cost that takes into account the efficiency gains resulting This model best applies to Group C, with a total of 329 from the conversion of the street lighting system. municipalities characterized by relatively small-scale and good fiscal management. These municipalities contain The ESCO model and off-balance sheet financing has 7% of the total population and require investments been considered by India, Mexico and several other amounting to R$3.2 billion for modernizing their public countries. Box 7 illustrates the financing mechanism street lighting network (11% of the investment needed used by the city of Guadalajara in Mexico. The main throughout the country). points of interest are the leasing process and the level of coordination between government agencies The evolution of the model in Brazil for developing a comprehensive program of energy efficiency in the street lighting sector. A variant of the ESCO model is being proposed by Box 7 - Guadalajara (Mexico): Lease-to-Own Model The city of Guadalajara, with a population of approximately 1.5 million, is Mexico’s fourth-largest city. It has approximately 80,000 installed luminaires, all of them of HPS, and plans to replace half of them with LEDs. Guadalajara began installing the new luminaires in 2013, benefiting from a federal energy efficiency program in the public street lighting sector. This program, mobilized by the Secretariat for Energy (SENER), the National Commission for Energy Efficiency (CONUEE), the Federal Electric utility (CFE) and the National Bank for Public Works (BANOBRAS), entails the provision of technical assistance for cities keen to develop LED projects. CONUEE also provides subsidies for municipalities to implement, verify and finance EE projects, a simulation model and a list of accredited products. Furthermore, the Secretariat is responsible for approving the technical aspects of projects as a prerequisite for obtaining Ministry of Finance authorization to employ public funds for street lighting projects. BANOBRAS also provides investment funds for projects through an energy efficiency incentives program. The Guadalajara project is being financed by EE gains, which are expected to reach 50–55% over the baseline, representing a monthly saving of US$500,000. The financial instrument used is a US$19 million 10-year leasing contract for which the municipality pays an average of US$250,000 per month. At the end of the contract, ownership of the luminaires (with an estimated lifespan of 13 years) will be transferred to the municipality. The Mexican federal government guarantees coverage of certain operational costs in the event of the Guadalajara authorities’ failing to make monthly payments. The existence of this guarantee has been a key incentive for attracting private capital to the venture. Source: Adapted from “Proven Delivery Models for LED Public Lighting.” World Bank, 2016. 65 5. Business models for public street lighting in Brazil Model M6: Municipal Consortium type mechanism, or a simpler arrangement, for the purpose of joint centralized procurement. The initial or Central Purchasing Body costs of purchasing new lamps would be borne by municipalities using COSIP resources and/or Description of the model and cash flows municipal budget funds. This model seeks to capture economies of scale Figure 20 summarizes the main actors and cash related to joint procurement of equipment (and flows of a municipal consortium/central purchasing possibly services) by multiple municipalities without body, assuming a centralized procurement process joint municipal responsibility for raising finance. involving a single contract with an equipment Although this model could involve an SPV, the supplier. main requirement is to establish a consortium- Figure 20 – Example of structure of centralized procurement model End Users COSIP $ Electricity Utility COSIP $ COSIP $ Funds $ Funds $ Financier(s) Financier(s) Municipality 1 Municipality 2 Municipality 3 Repayment $ Repayment $ COSIP and COSIP and other other sources $ sources $ Contract Equipment Payment $ and services Financial flows Physical flows Manufacturer + outsourced company Main actors, risks and mitigating factors affect the project throughout the process, and contains recommendations for mitigating these Table 29 lists the main actors of each of the key steps risks. of the municipal consortium/central purchasing body model, indicates the main risks that could 66 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 29 - Matrix of functions and actors in model M6: municipal consortium or central procurement agent M6 – Municipal Consortium or Central Procurement Agent Functions Actors Details Project risks Mitigating factors Planning and Consortia of munici- Prepare public notice Lack of sufficient qualified Capacity-building or bidding palities and organize bidding for manpower to prepare tech- standard manual; com- equipment. nical specifications. petitive bids. Consortia of munici- Promote bidding for lamp Lack of qualified manpow- Public notice follows palities procurement. er to cope with technical best national/interna- subjects; lack of adequate tional practices; bidding equipment guarantees. to include international competition; specialist advice. Obtaining Municipalities Structure and leverage Inability of smaller munici- Promotion of (public) finance resources to promote palities to secure financing credit lines for small- investment. on their own. and medium-size munici- palities. Financiers Provide financial resourc- Lack of investor interest; Provide data to facilitate es at competitive rates high interest rates. understanding of the for capital expenditure risks. (CAPEX). Lamp procure- Municipality or ESCO Manage the process or Lack of knowledge of Preparation of good ment supervise the services of the public street lighting quality technical studies; an outsourced company. network resulting in delays capacity- building. and/or cost increases; lack of qualified manpower for supervision (in the case of an outsourced company). Implementation Municipalities or con- Manage the process or Lack of qualified personnel. Technical and/or over- and installation sortia of municipalities supervise the services of sight training. an outsourced company. Advantages and disadvantages economies of scale at lower transaction costs. An important advantage of this model compared This model does not, however, resolve the difficulties to the M2 model (PPP consortia) is that it is not faced by municipalities to obtain financing, which necessary to establish an SPV and it does not require can be particularly difficult for medium-sized cities in joint financing arrangements. The individualized view of the substantial upfront costs. In such cases, responsibility to raise funds makes the model this model would need to be combined with a viable substantially less complex, and there is no risk of financing instrument. Moreover, the model would need a default by one municipality affecting the entire the technical expertise of a leading municipality to model. As a result, the municipalities benefit from coordinate the equipment-bidding process. 67 5. Business models for public street lighting in Brazil Identification of groups to which the model would Evolution of the model in Brazil apply While it is likely that the model has already been This model best applies to the municipalities in implemented for non-LED technologies in Brazilian Groups C and D. These total 1,216 municipalities, with cities, we found no specific example of centralized 21.8% of Brazil’s population and represent R$ 6.5 procurement during our surveys and interviews. billion of investment (23% of the investment needed Nonetheless, the centralized procurement modality throughout the country). This model could also be has proven to be successful in other parts of the applicable to some cities in Groups E and F (4,219 world (economies of scale, etc.), as in the case of the municipalities). province of Ontario, Canada, described in Box 8. BOX 8 - Province of Ontario: Joint Management Model for Procurement The Province of Ontario consists of 444 municipalities accounting for one third of Canada’s population. Around two-thirds of them are small with populations of under 10,000 and less than 2,500 light points. The Ontario Municipalities Association (OMA) has played a significant role in helping smaller cities modernize their public street lighting systems. The association also has a wholly-owned subsidiary company called LAS (Local Authority Services), which makes purchases of products and services – including LEDs - on behalf of member municipalities. Several factors were instrumental in AMO’s launching its LED-based public lighting program, including the high energy and operational costs of HPS lamps. The AMO/ LAS joint purchasing model basically aimed to: (i) ensure lower prices through large-scale purchasing; and (ii) reduce smaller municipalities‘ transaction costs involved in developing and managing complex specification processes associated with purchasing LED luminaires and maintenance services. As of August 2015, 127 of the smaller municipalities had participated in the LAS joint LED-purchasing program (a total of 100,000 luminaires). Many municipalities requested LAS technical assistance to select LED suppliers. RTE, the company chosen by LAS, implements and manages projects, promotes new products, and can provide financing and other services for the municipalities. Canada’s municipalities, especially in Ontario, are in good financial and credit shape due to the indebtedness limits imposed on them. Local authorities are obliged to meet strict transparency and competition requirements when conducting bids. However, many smaller municipalities suffer from the lack of experienced staff to prepare specifications and evaluate/ identify good quality suppliers and prefer to call upon AMO and LAS to bid for goods and services either centrally (which may involve the distribution of goods and services), or through straightforward centralized purchasing. 68 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning The combination of joint purchases by LAS, tax incentives, falling LED prices and the high-quality suppliers on the Canadian market has caused a rapid expansion of LEDs used in municipal street lighting in Ontario. In parallel, the number of municipalities that have adopted LED street lighting (contracted under LAS or independently acquired by larger municipalities) has grown rapidly. Joint bidding procedures that made it easier for smaller municipalities to adopt LED lighting appears to have encouraged larger municipalities to do the same. Source: Adapted from “Proven Delivery Models for LED Public Lighting.” World Bank Group, 2016. Model M7: Self-financing corresponding expenditures and investments, likely resulting in a relatively low value of annual Description of the model and cash flow investment, dependent on annual COSIP surpluses. The municipality could directly manage the O&M This model can be used by municipal governments to process, or it could be outsourced to a third party. invest in modernizing their street lighting networks without third-party intervention. Municipalities Figure 21 illustrates the main actors and cash flow undertake public lighting modernization investments involved in the self-financing model. using public lighting revenues pari passu with Figure 21 – Example of structure of self-funding model End Users COSIP $ Electricity Utility COSIP $ Municipality $ Equipment (+services) Manufacturer (+ outsourced company) Financial flows Physical flows 69 5. Business models for public street lighting in Brazil Main actors, risks and mitigating factors that could affect the project throughout the process, and contains recommendations for mitigating these Table 30 lists the key actors in each of the main stages risks. of the self-financing model, indicates the main risks Table 30 – Matrix of functions and actors in model M7 – self-financing M7 — Self-financing Functions Actors Details Project risks Mitigating factors Planning and Municipality Prepare public notice Lack of sufficiently qualified Capacity building or standard bidding and organize bidding manpower to prepare technical manual; competitive bids. for equipment. specifications. Obtaining finance Municipality Use surplus (if avail- Slow implementation of the Maximize annual investment able) from annual project, causing lost opportuni- (e.g., robust structuring of budget for the public ty for financial and energy-sav- COSIP, seeking opportunities lighting sector and/or ing benefits for the municipal- to reduce costs on O&M, etc.), from COSIP. ity. since the gains produced by the project will be greater than costs over the project lifetime. Lamp procure- Municipality Prepare bidding for Lack of qualified personnel to Public notice follows best na- ment procurement of lamps. cope with technical subjects; tional/international practices; lack of adequate equipment bidding to include international guarantees. competition; specialist advice. Implementation Municipality Manage the process Lack of knowledge of the Preparation of robust technical and installation or supervisor services public street lighting network, studies; capacity-building. of an outsourced resulting in delays and/or cost company. increases; lack of qualified manpower for supervision (in the case of an outsourced com- pany for O&M). Operation and Municipality Manage the process or Lack of qualified manpower. Technical and/or oversight maintenance supervise the services training. provided by an out- sourced company. Advantages and disadvantages municipalities to carry out a LED street lighting project, unless they have federal government The main advantage of this model is that its support in terms of capacity-building, financing, implementation does not require financial and/or financial guarantees. or institutional arrangements, thus reducing transaction costs. Moreover, the model may be the Anticipating the use of COSIP funds or municipal only way forward for smaller and fiscally fragile budgets for upfront investments in street 70 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning lighting projects may not be an option for the need for government involvement in municipal many municipalities, given that self-financing street lighting programs by providing technical requires surplus revenues to be available. Thus, assistance and concessionary funds for designing municipalities with smaller scale and a shortage and implementing projects. One solution would be of qualified staff may find it difficult to increase for public financing institutions, such as BNDES, COSIP revenues to a sufficient level for pursuing Banco do Brasil and Caixa Econômica Federal self-financing investment. A further problem is (CEF) to provide subsidized credit lines. that such municipalities lack technical expertise, and risk purchasing expensive and lower-quality equipment. Model M8: Transfer of Luminaires Description of the model and cash flow One of the main disadvantages of this model is that it involves a slow street lighting modernization This model consists of the transfer of HPS, together process given that it depends on surplus revenues with complete luminaires (housing, lens, ballast and cannot count on outside financing. All this and mounting arm), from larger municipalities reduces the prospect of improving the financial to others that still use less efficient lamps and do situation of municipalities and improving the not have the resources to buy new LED lamps. This country´s energy efficiency overall. model focuses on inter-municipal (rather than intra-municipal) transfers of equipment that can Identification of groups to which the model would be easily carried out by municipalities themselves apply without requiring purchasing or donation procedures. This model may be the only option for many municipalities in Groups D, E and F, unless some The massive conversion in the future of HPS and type of public programs are available to assist mercury vapor lamps to new LED equipment76 them. These groups represent more than 90% of will leave a significant stock of secondhand HPS Brazil’s 5,106 municipalities and require R$12.8 lamps which, after replacement, will be available billion in investment (47% of the investment needed for transfer to locations that still use other less throughout the country). The self-financing model efficient lighting technologies but that do not yet could also be applied to certain municipalities in have the means to convert to LEDs77. Group C. A coordinating structure will be needed for this Evolution of the model in Brazil task. One of the possible candidates for this could be PROCEL, which could create an auction Several large municipalities are already using the website (similar to E-bay or Alibaba, for example) self-financing model for LED pilot projects—often to facilitate buying and selling in a transparent an interim measure before implementing a larger manner. The buyers could consist of municipalities PPP project using one of the alternative business or outsourced companies acting on their behalf, models described here75. In view of the technical while the sellers would be municipalities with evolution and economic-financial benefits of LED stocks of secondhand luminaires, or private actors technology, the full-scale conversion of street who have purchased them from municipalities lighting to LED is almost inevitable. This reinforces 75 While preparing the public notice for its PPP project, the city of São Paulo used its own resources to install LED lamps on some of the city’s major viaducts. 76 In 2012 the installed network of HPS lamps amounted to 11.4 million units, far more than the 1.2 million light points currently existing in Group F municipalities. Metallic multivapor lamps totaled 201,000 and mercury vapor lamps (candidates for substitution) 3.8 million. There is also an operational public street lighting network using incandescent lamps (188,000), mixed lamps (283,000), fluorescents (160,000) and halogen lamps (10,900). 77 The city of Los Angeles prepared a tender for the sale of the retired HPS luminaires, which produced around US$ 6.5 million. 71 5. Business models for public street lighting in Brazil (or from dealers acting on their behalf) and are noting that although transferring secondhand willing to sort them, fix them if necessary, and equipment would only be a temporary measure, offer them for sale on the website. To do this, the the receiving municipalities would nevertheless products and quantities would need to be clearly gain short-term benefits while leaving the described, including their technical specifications conversion to LED lighting until their financial and residual life expectancy (with no guarantee positions improve. of lifespan). The transactions could follow the upward bidding auction model as on E-bay. Note The second issue concerns Brazil’s national that these are simply ideas to encourage debate strategy for converting to LED technology. on how to make the most efficient use of the Transferring non-state-of-the-art equipment can secondhand luminaires. represent energy efficiency gains in areas where the business models do not make LED deployment However, this model raises two important issues. worthwhile, at least in the short and medium term. The first concerns the transaction costs involved in However, a more robust public policy strategy to transferring secondhand equipment. The receiving replace the country’s entire public street lighting municipalities might also encounter problems system with LEDs would rule out the need for this in acquiring and testing the equipment, given model. the nature of the bidding processes needed, as well as problems involved in pricing and valuing Figure 22 summarizes the main actors and cash the transferred items and paying for appropriate flows of the luminaire transfer model. technical capacity-building to do this. It is worth Figure 22 - Example of structure of luminaire transfer model Municipalities A Municipalities B Municipalities C Municipalities D HPS and mercury vapor $ Technical assistance / financing Coordinating Agent Público Entity $ HPS and mercury vapor Financial flows Municipalities E Municipalities F Physical flows 72 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Main actors, risks and mitigating factors main risks that could affect the project throughout the process, and contains recommendations for Table 31 summarizes the main actors in the key mitigating these risks. stages of the luminaire transfer model, indicates the Table 31 - Matrix of functions and actors in Model 8 — transfer of HPS luminaires M8 – Transfer of HPS Luminaires Functions Actors Details Project risks Mitigating factors Planning and Coordinating To create an online plat- Complex coordination, Identify private sector bidding agent, municipal- form for the selling/buying possibly involving high actors who could provide ities or private municipalities to familiarize transaction costs; lack of coordination for logistical companies themselves with the products interest of public and/or and financial aspects of the on sale private bodies in partici- program; involvement of a pating in the process. public sector stakeholder to oversee the process. Municipality Manage the process or Technology with a lower To offer lower prices to supervise the services of an useful lifespan, or lack offset the technical risk; outsourced company. of qualified manpower capacity-building. for oversight (in the case. outsourced companies). Coordinating Organize auctions for dispos- Lack of qualified staff to Rely on private sector agent, municipal- ing of the HPS lamps. manage the process; lack expertise. ities or private of controls possibly lead- companies ing to financial risks. Obtaining Municipality Use surplus (if available) Lack of sufficient resources Robust implementation of finance from annual budget for the for purchasing LED lamps; COSIP to ensure surplus public lighting sector and/or complexity of organizing for public lighting modern- from COSIP. transactions between ization; involvement of the municipalities. private sector in carrying out the transactions. Lamp procure- Municipality Manage the process or Lack of qualified person- Technical and/or oversight ment supervise the services of an nel; high cost resulting capacity-building. outsourced company. from rapid burnout of lamps acquired in this way. Execution and Municipality Manage the exchange Technology acquired with Ensure prices of used equip- installation process or supervise the short remaining life- ment properly reflect the company outsourced to do times, or lack of capacity technical risk. Training. this work. to supervise outsourced company O&M Municipality Manage the O&M process Lack of capacity; high fail- Ensure sufficient technical or supervise the company ure rates of lamps, leading and/or supervisory skills. outsourced to do this work. to high costs. 73 5. Business models for public street lighting in Brazil Advantages and disadvantages Evolution of the model in Brazil The main advantage of this model is that it No similar model was identified as part of the offers an opportunity for smaller, lower-income research undertaken for this study. The need municipalities to improve energy efficiency and and/or attractiveness of this business model will service quality of their public lighting networks. be largely determined by whether the federal government develops a more robust public However, this model should be considered a policy strategy to support and/or encourage the temporary option, given the recommendation replacement of the country’s entire public street that the entire country should aim to convert its lighting system with LEDs. entire public street lighting network to LEDs in 5.4 - Summary of due course. Before proceeding with this option it is worth considering whether the short-term Business Models benefits can justify the possible transaction and coordination costs involved, if the secondhand and Mapping HPS lamps are to be eventually replaced by LED technology. At the same time, given the financial of Groups of limitations of smaller municipalities, this model may be their only short-term option. Identification of groups applicable to the model Municipalities Transferring secondhand lamps could benefit the 4,219 municipalities in Groups E and F where fiscal and institutional difficulties rule out the This section seeks to identify business models that, conversion to LED-based street lighting in the considering the institutional environment and short term. These groups represent 75% of all characteristics of the Brazilian market, would allow Brazilian municipalities and would require R$9.5 sufficient funds to be generated for implementing billion in investment (35% of the investment public street lighting modernization projects. Table needed throughout the country). 32 summarizes the eight models, indicating the main roles and key stakeholders involved in each key phase of the project. 74 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 32 - Key stakeholders in each phase of the business models Business M1 – Munici- M2 – Consortium M3 –Munici- M4 – Electric M5 – ESCOs M6 – Munic- M7 –Self M8 –Transfer model pal PPP for PPPs pal Financing Utilities ipal Central Financing of Luminaires Programs Procurement Agent Resource Consumers Consumers Consumers Consumers Consumers Consumers Consumers Consumers generation Energy supply Electric utility Electric utility Electric utility Electric utility Electric utility Electric utility Electric utility Electric utility to light point company company company company company company company company Supply of Manufactur- Manufacturers Manufactur- Manufactur- Manufactur- Manufactur- Manufactur- Manufactur- equipment ers ers ers ers ers ers ers Planning and Municipality Municipalities, Municipality Energy utility, ESCOs, munic- Municipal Municipality Coordination bidding or Audit Court, municipal municipalities; ipalities consortium agent, munici- consortia of consortia, Audit ANEEL palities companies Court, consortia of companies Obtaining Public lighting Public lighting Municipality Energy con- ESCOs; Municipal Municipality Municipality financing concession- concessionaire; or financiers cessionaire financiers consortia or aire; financiers financiers municipalities; financiers Lamp pro- Public lighting Public lighting Municipality Municipality ESCOs Municipal Municipality Coordination curement concession- concessionaire; or ESCO consortia agent; munici- aire; manufac- manufacturers palities turers Implemen- Public lighting Public lighting Municipality Municipality ESCOs Municipality Municipality Municipality tation and concession- concessionaire or ESCO or ESCO installation aire Operação e Public lighting Public lighting Municipality Municipality Municipality Municipal Municipality Municipality manutenção concession- concessionaire; or ESCO or ESCO consortia or aire; munici- municipalities municipalities pality. Given that, in many cases, the models apply to The municipalities themselves will be ultimately more than one group, it is important to map how responsible for choosing the business model that the eight models can be applied to the six groups of is most appropriate to their circumstances or a municipalities. Table 33 shows the interface between combination of models that could produce the best the groups and the suggested business models. solution in the short, medium and long term. 75 5. Business models for public street lighting in Brazil Table 33 - Business models for each group M1 M2 M3 M4 M5 M6 M7 M8 PPP Mu- Municipal Municipal Electric ESCO Centralized Self-Funding Transfer of Group nicipal Consor- Funding Utility (s) Procure- Luminaires tium Programs ment with PPP A B C D E F Source: World Bank Group; Pezco Consultoria Key: = suggested; = possoble The following section provides further details on investing in the modernization of municipal street the financing mechanisms that could be used for lighting networks under the various business models. 76 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 77 6. Financing and credit enhancement mechanisms 78 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 6 - Financing and Credit Enhancement Mechanisms T o take advantage of the benefits of to the private sector, others such as M3 depend converting to the new lighting technology, heavily on the existence of lines of credit targeted financing models are needed to facilitate at the public sector. the investments required for a city’s chosen business model. The major financial instruments are described below, together with their applicability to the Each model contains specificities that require different business models. At the end of the different financing options and models. While section, Table 35 summarizes application of the models such as M1 and M5 are more attractive instruments to the various business models. 6.1 - Financing Mechanisms 1) COSIP is a charge or tax levied by the energy utilities on consumers and transferred to the municipality to cover the costs of running and expanding its public street lighting system. It is a ring-fenced (earmarked) fund - endorsed by the Constitution and created via municipal legislation - that cannot be used for other purposes. As such, the resources arising from this tax must be used for modernizing the public lighting sector. This ring-fenced and earmarked source of revenue is a differentiating factor that can attract private capital to the sector, given that its shielded status reduces municipal credit risk. As discussed in Section 4.3.3, most Brazilian municipalities charge COSIP, and others are preparing to do likewise. In the absence of COSIP, municipalities can resort to using municipal budget funds, although these are subject to other contingencies and compete with other social priorities. Given its benefits and widespread implementation, COSIP is a key source of financing in all the business models (M1 - M8). 79 6. Financing and credit enhancement mechanisms Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment COSIP or Minicipal Budget Key: full circle = most applicable; empty circle = not applicable 2) Loans from public banks - Banco do Brasil (BB), Caixa Econômica Federal (CEF). Although these institutions do not provide specific financing lines for energy efficiency or public lighting projects, such lines could be created or projects could be tailored to their general lending requirements. Both institutions are extremely well-known to municipal governments, many of which are active clients in infrastructure and other areas. CEF possesses qualified and decentralized technical staff that could assist smaller municipalities with their public street lighting projects. Both BB and the CEF would be best placed to collaborate in M3 public sector projects and M6 municipal procurement consortia. Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment Loans from Public Banks (BB, CEF) Key: full circle = most applicable; empty circle = not applicable 3) BNDES-Finem. This is a credit line dedicated to energy efficiency projects, replacing the previous BNDES PROESCO line, which completed 27 financing operations worth around R$510 million78. Market practitioners emphasize that this line was difficult to access because of the requirement for private guarantees or solid balance sheets. Its successor, BNDES-Finem79, accepts operations equal to or greater than R$5 million and finances up to 70% of eligible items, at a cost equivalent to the TJLP (Long-Term Interest Rate) + 2.5% to 5.7% per annum. The portfolio of this new line has twenty projects already approved or being evaluated80. Subnational agencies can pass on financing. The BNDES-Finem line is likely to apply to smaller lighting projects where an ESCO is responsible for implementation (M5). Other possibilities might be projects carried out by municipalities, especially when they need counterpart funds for future concessionary programs (M4), or direct financing for equipment and services (M3 and M6). 78 Revista Brasil Energia, Issue No. 423, February 2016, p. 16. 79 BNDES credit line for project financing. In general, BNDES-Finem finances projects worth over R$20 million, covering 50% of the financeable items, but these parameters are more favorable (R$5 million and 70%) for EE efficiency projects. 80 Revista Brasil Energia, Issue No. 423, February 2016, p.15. 80 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment BNDES - FINEM Key: full circle = most applicable; empty circle = not applicable 4) Private Equity. This entails private capital investment as equity as a shareholder in a Special Purpose Vehicle (SPV) for financing a public street lighting modernization project. SPV involvement can range from an administrative concession granted by a municipality (M1) or PPP consortia (M2), to simpler structures where an ESCO is responsible for conversion to LED technology and receives payment to cover capital costs (M5). Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment Private Equity Key: full circle = most applicable; empty circle = not applicable 5) Mechanisms for improving credit risk. Forms of credit enhancement may be essential for financing public street lighting projects. Even when the COSIP tax is well-formulated and offers robust guarantees for lenders, there might be a perception of residual risk arising from insufficient COSIP resources and, thus, exposure of investors to municipal credit risk. To mitigate this risk, mechanisms (including forms of guarantee) may be required to ensure that these schemes effectively enhance financing capacity and generate the desired levels of investment. Multilateral banks have products available for private partners to invest in public street lighting projects which could help to mitigate risks. One example is the World Bank’s partial payment and partial loan guarantee schemes. Municipalities could also offer shares in municipal companies or guarantees to mitigate municipal credit risk. However, there are very few examples of these types of mechanisms being used as collateral, principally due to the limited availability of suitable assets even in larger municipalities. Credit guarantees apply to the business models with private sector investment, mainly M1–M3. They can also apply to the M4 and M5 models. 81 6. Financing and credit enhancement mechanisms Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment Credit enhance- ments (i.e. World Bank) Key: full circle = most applicable; empty circle = not applicable Figure 23 shows a potential scheme for the use of collateral in a PPP for street lighting, including a partial guarantee of payment of the municipality’s payments in the consortium. It is important to note that the cost of such risk reduction mechanisms will ultimately be borne by the municipality through contracts with the consortium. Guarantees can be applied to other business models below (e.g., municipal funding, ESCOs), in addition to the PPP model. The role of the World Bank guarantees in this example is to ensure that the municipality’s escrow account contains sufficient funds for remuneration of the public street lighting consortium (the concessionaire). A reserve account, backed by a letter of credit from a local financial institution and – ultimately – by the World Bank, is created to provide funds to the escrow account in the event of shortfalls. The amount to be allocated to the reserve account should be negotiated between the parties. This mechanism, with the sovereign guarantee of the National Treasury, is normally formulated to cover the liabilities of the account for a period of between three and six months. 82 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Figure 23 - Example of the use of World Bank guarantees in the PPP model National Treasury World Bank Federal transfers to Partial municipalities Payment Financier(s) (collateral) $ Guarantees Local financial Municipality institution Loans, debentures Repayments $ and/or grants $ Additional contributions $ Letter of Credit EE lightining service EE public Municipal Account Reserve Account lightning consortium / Escrow Account $ Payments COSIP $ Technical / operational Premiums insurances Electricity Utility Financial flows Physical flows Insurers COSIP $ End Users 6) Development banks (domestic and multilateral) a) National development banks: These are federal, state or multilateral development institutions that provide financing instruments for energy efficiency projects, potentially including public street lighting projects. Examples of Brazilian development banks (BNDES, DesenvolveSP, AgeRio, BDMG, etc.) are listed in Table 34. The BNDES has shown interest in supporting specific PPP projects to modernize public lighting, such as those in São Paulo and Belo Horizonte. The Desenvolve-SP, AgeRio and BDMG agencies on-lend funds from BNDES or can offer their own specific financing lines for EE and street lighting, given that they have a mandate to invest primarily in infrastructure projects that have economic benefits, such as climate change mitigation projects. Many of the above-mentioned institutions have good access to smaller municipalities and to knowledge of local realities, and generally offer lower interest rates than the capital markets. Such sources of financing would be more appropriate for the M1, M2 and M3 models, and potentially for the M5 and M6 models. b) International development banks: The World Bank, CAF and IDB and other multilateral banks can in principle finance projects for public entities, such as for larger municipalities using the M2 model. The private arms of these institutions, such as the World Bank Group IFC can finance PPP SPVs and ESCOs, and provide financing and technical assistance for certain public-private partnerships (M1, M2, M5), as well as possibly providing mechanisms to mitigate project risk. 83 6. Financing and credit enhancement mechanisms Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment Development Banks (domestic, multi- lateral) Key: full circle = most applicable; empty circle = not applicable Table 34 below provides examples of the financing lines already available in development banks. Table 34-Details of financial instruments of development banks Modality Description Cost81 BNDES – Climate Fund82 Credit line for projects that increase the sustainability of Cost: 1.9% + credit risk (1% for public cities, including public street lighting. sector and 3.6% for remainder). BNDES –support for Credit lines available for specific PPPs (e.g. São Paulo and Cost: Terms and conditions set out in specific projects. 83 Belo Horizonte). the Public Notices. Desenvolve SP – Green Credit line for financing EE projects. Cost: from 6.55% per annum + IPCA. Economy Line 84 Desenvolve SP –Public Credit line for financing public street lighting services up to Cost: 9.5% per annum. Street lighting85 100% of the value of the project. AgeRio Pró-Urbano86 Credit line for urban infrastructure. Public street lighting is Cost: Selic rate+ 4% per annum. one of the items eligible for financing. This line can finance up to 100% of the project, with a ceiling of R$20 million. BDMG Credit line with funding from the French Development Cost: SELIC and interest rates of 5 6% Agency (AFD), for EE projects with budgets of R$50 million. per year. Multilateral institutions Loans for large municipalities. On a case-by-case basis. (e.g., World Bank Group) Loans for electric utility companies or ESCOs. Shareholder participation in SPVs. Technical assistance and support for structuring PPPs. 7) Sectorial financing (PEE, Reluz, PROCEL). While sectorial funds have played an important role in the modernization of public street lighting systems, legal changes have reduced the amount of funding available to this sector. Although the funds for the public street lighting sector in Brazil will depend on the implementatioan of the new legislation to transfer resources from PEE to PROCEL or any other new policies and regulation emerge, it is expected that the utilities will facilitate or coordinate energy efficiency-related resources (M5). Sectorial funding for small- or medium-sized public or private projects is also a possibility, depending on the rules to be established for managing the PEE funds. 81 Figures from April 2016. 82 Available at: 83 Available at: 84 Available at: 85 Available at: 86 Available at: 84 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment Sector Financing Lines (PEE, Reluz, PROCEL) Key: full circle = most applicable; empty circle = not applicable 8) Loans from private banks. This involves Brazilian or foreign, private commercial banks lending directly to municipal governments (M3, M6) or agents acting on their behalf (M1, M2 and M5) for street lighting modernization. The terms and conditions of such loans in principle reflect market conditions and take into account the project and credit risks of the municipalities. Loans by private banks to the public sector must also take into account the fact that municipalities are subject to indebtedness limits. Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment Loan from Private Banks Key: full circle = most applicable; empty circle = not applicable 9) Debentures, FDIC, FIP, Green Bonds. These are mechanisms that allow larger and more cost-competitive funds to be obtained in the Brazilian or international financial market for large-scale projects. They are mentioned here because they involve “securitization” of the revenues of projects that receive investment. These instruments are briefly described as follows: a) Debentures are debt instruments (bonds) issued by an SPV (M1, M2) or by a municipality (M3). Bonds are nowadays the most readily-accepted securitization instrument in Brazil’s capital markets. Infrastructure debentures can be acquired by private investors or development banks (e.g., BNDES) and are particularly attractive because they can also offer substantial tax relief. b) FIDCs are mutual funds that invest at least 50% of their net assets in debt securities. They are a suitable tool to securitize the future cash flows from infrastructure projects, such as COSIP or even debentures. FIDCs can acquire debentures issued by SPVs (M1, M2 and M5) or municipalities (M3), thus diversifying the risks for investors. As with debentures, FIDCs also offer considerable tax advantages. 85 6. Financing and credit enhancement mechanisms c) FIPs (Private Equity Investment Funds) are closed funds that invest in stocks, debentures and convertible debt instruments of any publicly or privately held Brazilian company providing it is a corporation. This is most applicable to M1 and M2 models. Financial institutions, insurance companies and pension funds are among the possible investors for these funds, which can also be tax exempt in certain cases. FIPs could be used in public street lighting projects requiring funding by institutional investors, and are most applicable to the M1 and M2 models. d) Green Bonds are credit instruments similar to debentures in the Brazilian market, which can be issued by private, public or multilateral institutions. The primary purpose is to invest in projects of environmental interest, mainly focused on climate change mitigation. In principle, energy efficiency projects like LEDs for public street lighting systems would qualify for Green Bonds, since such projects reduce the GHG emissions linked to electricity production. Due to the high transaction costs (including because investors may require verifiable benefits), the Green Bond instrument would be most suitable for large-scale projects by the private sector (M1, M2, or M5) or large municipalities (M3). Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment Debentures, FIDC, FIP, Green Bonds Key: full circle = most applicable; empty circle = not applicable 10) FI-FGTS. Although FGTS resources have been used to support certain urban infrastructure projects, few have been in the street lighting area unless they form part of a broader urbanization project. While the electricity sector has historically been the main manager and financier of public street lighting projects, a sensible initiative would be to treat public street lighting modernization projects as part of the urban infrastructure, given that all municipalities now must take ownership of these assets. The allocation of FI- FGTS funds requires the intervention of the Ministry of Cities and approval by the FGTS Board of Trustees. Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment FI-FGTS Key: full circle = most applicable; empty circle = not applicable 86 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 35 summarizes applicability of each financial instrument to each business model. Table 35 - Financial instruments suitable for use in the different business models Financial Instru- M1 M2 M3 M4 M5 M6 M7 M8 ment / Business PPP Municipal Municipal Electric ESCO Central- Self- Transfer of Model Municipal Consor- Funding Utility (s) ized Funding Luminaires tium Programs Procure- with PPP ment COSIP ou Municipal Budget Loans from Public Banks (BB, CEF) BNDES - FINEM Private Equity Credit enhance- ments (i.e. World Bank) Development Banks (domestic, multilateral) Sector Financing Lines (PEE, RELUZ, PROCEL) Loans Private Banks Bonds, FIDC, FIP, Green Bonds FI-FGTS Key: full circle = most applicable; empty circle = not applicable 87 6. Financing and credit enhancement mechanisms 6.2 - Risks and Mitigation Mechanisms In addition to the credit risks mitigated by the mechanisms discussed in the above section, other key risks need to be considered and mitigated in public street lighting projects. These include: the risk of technical and/or operational performance, political risk and non-manageable risks. These are summarized in Table 36. Table 36 – Other risk mitigation mechanisms Risks Mitigation mechanisms Examples of mitigation products Technical performance risk Manufacturer´s or factory guar- Guarantees supplied by large manufacturers (cost embed- antee ded in the price of LED lamps) Technical and/or operational Insurance guarantee Reassurance company (e.g., MunichRe) products, notwith- performance risk standing the fact that these still lack market maturity Political risk Multilateral financial institution MIGA (World Bank) insurance Non-manageable risks Federal Government guarantee for Infrastructure Guarantee Fund (FGIE), tends to focus on infrastructure (ABGF) larger projects, and is not applicable to the majority of municipalities. Source: Compiled by Pezco Consultoria based on stakeholder consultations and public sources. Many of these mechanisms will be needed to support projects in which the private sector is involved as a concessionaire, an ESCO or a private financial institution (i.e., applicable to models M1 to M5). 88 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 89 7. Other considerations in the design of a LED project 90 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 7 - Other Considerations in the Design of a LED Project A 7.1 - Automation, lthough this report has covered options for business models and financing instruments that municipalities and other key stakeholders need to take into account when designing a modern Remote public lighting project with LED technology, many other details and decisions are involved. It is out Management and of the scope of this report to detail many of these other factors; nonetheless, this section provides a Ancillary Services brief overview of possible options that need to be considered, such as: (i) the level of automation of the public street lighting system; and (ii) other civil works that are included within the scope of a project Public street lighting modernization projects can to improve the infrastructure of a public lighting enable cities to take advantage of the structure system and of other related sectors. related to the installation of LED luminaires to benefit a city in terms of automation and remote maintenance and provide complementary services to benefit the population. One possibility is to rent out the available infrastructure for the private sector to increase the range of telecommunications and other services for the public. The layout of poles, access to reliable power supplies, granularity, and “on-board” communication electronics can facilitate the delivery of these services at a relatively low incremental cost. Various levels of automation and service delivery are described in Table 37. 91 7. Other considerations in the design of a LED project Table 37 – Automation levels of public street lighting projects Level of automation Description Low automation No control over the status of lamps (burnt out, off at night, on in daylight hours). No individualized measurement of electric energy and savings are based on estimates (deemed). Quality control is done by ad-hoc maintenance crews, and the model includes a customer call center. Operational automation The model includes control over the status of the lamps, and energy inputs are measured at some points. This information is sent remotely to a control center, which may also possess two-way com- munication to operate shutdowns, dimming and other features. The communication structure only serves operational requirements. Automation to support a In addition to the above-mentioned control facilities, the model may also contain a system of sen- public good sors for transmitting key data for managing different urban facilities such as traffic control, signals, security cameras, electric vehicle recharging points, etc. This structure can also support Wi-Fi points for the convenience of the population. These services are of a public nature and most of them are difficult to monitor. The cost of the services are borne by the municipality. Automation to support New private businesses can also use the public lighting system´s granularity and “ “footprint” for new businesses data sending and control. This can include data on parking management, vehicle sharing, etc. The most promising and viable application could be the use of public lighting points for improving 4G cellphone reception. While increased automation can benefit a costs and benefits of automation and develop a step- municipality’s street lighting and other key sectors by-step strategy, possibly by designing a simpler such as transportation and security, it also involves project to begin with that is “future proofed,” i.e., higher costs due to project design delays and sufficiently flexible to accommodate more complex complex contract negotiations with the private automation in due course. Box 9 contains ideas of sector. Municipalities are therefore well-advised at how to incorporate this concept in a PPP concession the project planning phase to carefully evaluate the for public street lighting. Box 9 - Future-Proofing a PPP Concession Contract Although LED technology is advancing rapidly in terms of luminous efficiency (lumens per watt) and cost per watt (i.e., a significant reduction of US$ per lumen), some obsolescence of equipment can be expected during the term of a concession agreement. While the equipment is unlikely to become dysfunctional or fail to generate the expected benefits during its lifetime, the expansion of a public street lighting system will in the future benefit from cheaper and even more efficient technologies. It may even be feasible, on energy-saving grounds, to justify replacement of LED lamps before the end of their useful life, in which case it may even be possible to replace the chip in the luminaire, retaining the electronic and optical components. Concession contracts will need to be sufficiently flexible to accommodate these technological innovations. Contracts will also need to contain provisions for reviewing the economic/financial 92 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning balance of projects if future technological developments justify modifications to the existing street lighting system. A review could, for example, be held in anticipation of the end of the lamps´ useful life, when they will need to be replaced. In this case, concession contracts should include generic clauses governing how the additional future efficiency gains due to technological innovation will be shared between the municipality and the concessionaire. It is likely that major changes in the value of a concession will be due to the incorporation of new private services that can be monetized by the city or concessionaire. One example would be the installation of a 4G cellular network benefiting from public street lighting infrastructure (posts, energy supply, control systems, etc.). This technology is not yet widely used but has already been laboratory tested, with trials in cities such as Berlin, Germany. According to experts, this technology is promising; however, other options may exist that do not depend on the public lighting network. Given these uncertainties and the high added value they could involve, the ideal way to “future- proof” would be to exclude from the public lighting concession any reference to future services that may be developed, leaving these for the local authorities to handle under a future regime. At present municipalities such as São Paulo and Belo Horizonte have concession contracts that set conditions for sharing revenues from future technological innovations and the ancillary services that may be developed by the current utility companies. Other municipalities may find it worthwhile to compare the advantages and disadvantages of this approach. Given the uncertainties and risks of future ancillary services, bidders for the project may be heavily discounting their potential value to the detriment of municipalities. For the time being, clearer information is needed on technological innovations before granting rights to use public lighting networks for undefined ancillary services. 7.2 - Modernization to public lighting (electrical energy distribution, transport, security, telecommunications, etc.). of the Public Street Cities will need to carefully consider the level and scope of their street lighting modernization Lighting System in projects: fiscal savings, improved service quality, compliance with technical regulations, a City Infrastructure aesthetic/commercial improvements, etc. Table 38 summarizes some of the different approaches to modernizing the public street lighting The modernization of public lighting systems infrastructure. can involve a wide range of different activities in addition to converting to LEDs, such as other civil infrastructure works in other sectors linked 93 7. Other considerations in the design of a LED project Table 38 – Levels of modernization of the public street lighting infrastructure Modernization level Goals Description Simple modernization • Improve energy efficiency, using the simplest and • Substitute existing lamps with more efficient most economical design possible. models to improve compatibility of the system with the quality standards established by national • Offer positive effect on service quality given that: norms. (i) burned-out lamps are replaced; (ii) some LED luminaires would help to partially light darker • Continue to use the points to which luminaires areas between poles. are fixed (poles, buildings) with no need to pro- vide extra points. Intermediate modern- • All of the above, plus: • All of the above, plus: ization • Ensure basic compatibility of the system with • Change existing points and/or installation of ex- quality standards established by national norms. tra points in order to comply with national public lighting norms. • Inspect cabling (including grounding) to ensure functioning of the lamps including in stress situa- tions, e.g., frequent lightning episodes. Complete moderniza- • All of the above, plus: • All of the above, plus: tion • Ensure the compatibility of the system with quali- • Inspect cabling (including grounding) through- ty standards established by national norms. out most of the city. • Implement other improvements for aesthetic rea- • Install remote management systems to support sons and/or to benefit other city facilities (including other public service areas (security, transport, infrastructure for private sector initiatives). etc.) and/or for creating new private sector busi- nesses (WiFi, 4G, etc.) as described in Table 37. There is no single or perfect model for cities. the public lighting services, available resources and Selecting one will depend on the strategic goals of the option of joining forces with utility companies to each modernization scheme, the initial quality of install an underground cabling network. 94 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 95 8. Gaps in the market and recommendations 96 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 8 - Gaps in the Market and Recommendations Most of the business models and financial services rendered to the community, etc.). instruments presented above require some form of institutional support to ensure countrywide It is therefore vital to identify gaps and seek speedy implementation. solutions to include the largest number of Brazilian cities in this technological revolution. This study shows that, without some kind of Table 39 lists the gaps and barriers that impede government intervention in the public street the overall development of the sector in six lighting market, the most probable scenario will areas: (1) the regulatory framework; (2) the legal be for 90% of the cities to use the self-financing framework (3) public policies; (4) financing; (5) business model (M7) representing 50% of the capacity-building; and (6) technology. The table light points (i.e., Groups D, E and F). This would also includes a number of recommendations imply a very slow retrofit of Brazil’s public street to overcome the various gaps and barriers and lighting sector reducing the economic and social suggests various key stakeholders who could play benefits that LED technology can provide (e.g., a leading role in the respective areas. energy savings, public safety, improvement of 97 8. Gaps in the market and recommendations Table 39 - Gaps, barriers, and recommendations Gaps/barriers Recommendations Key players Lack of a regulator and reg- 1) Identify possible actors with regulatory knowledge State regulatory agencies; ulatory capacity for the public to fill regulator gap, or at least to provide regulatory Ministry of Cities; Brazilian lighting sector advice to municipal governments when necessary Associations of Municipalities; consortia formed between mu- nicipalities for this purpose Legal and regulatory framework Lack of clarity in the legal 2) The federal government interpret and clarify what National Audit Court (TCU); interpretation of COSIP are the legal uses of COSIP (e.g., for expenditures elec- public ministry; legal depart- tricity, O&M, expansion, modernization). ments of the justice courts (TJ); 3) Standardization of the legal framework cities can use specific rulings by the Supreme to earmark COSIP funds Court (STF) Gaps in the legislative frame- 4) Revise the legislative framework for public consortia Ministry of Cities; Brazilian As- work to enable public consor- to allow direct financing of the public consortium´s sociations of Municipalities; PPI tia to be a viable mechanism legal entity for aggregation Lack of clarity in the appli- 5) Establish a uniform approach for oversight agencies Congress, via specific legisla- cation of Law 8.666 and the verifying the legal provisions applied to different pro- tion; PPI PPP law and their possible curement bidding modalities. inconsistencies Lack of national leadership for 6) Identify and designate federal, state, and/or munici- Ministry of Mines and Energy; efficient public street lighting pal leaders to be responsible for leading coordination of Ministry of Cities; ELETRO- Public Policies EE street lighting initiatives and for helping municipali- BRAS; public banks; PPI Execu- ties to overcome barriers tive Secretary Absence of a national strate- 7) Federal government should design a national strate- Ministry of Mines and Energy; gy/policy gy for EE street lighting, including targets for convert- Ministry of Cities ing the national public street lighting network to LEDs 98 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Insufficiency of public or sec- 8) Create new credit lines and/or instruments for ELETROBRAS; energy distrib- torial lines of subsidized credit municipalities unable to attract (sufficient) private utors via current PEE; FGTS investment. Board of Trustees Lack of standardized EE 9) Design investment funds offering new, standardized Public banks, multilateral de- investments to attract large EE asset class to attract a variety of investors at scale velopment agencies, bilateral scale, private sector financing development agencies, DFIs, private sector Fiscal Responsibility Law lim- 10) Pass new legislation that allows exceptions to be Federal government its municipal financing (to 16% made in municipal debt thresholds in order to enable Financing of Net Current Revenue). greater investments in EE public lighting (similar to exceptions that were provided to debt under PROCEL program). Municipal credit risk discour- 11) Create instruments for mitigating municipal credit Federal government; multilat- ages private investment risk. eral agencies Lack of incentives for energy 12) Provide electric utilities different incentives to ANEEL, energy distributors distributors to invest in energy promote EE street lighting by altering the form of efficiency in the public lighting compensation for energy distributors, namely through sector decoupling sales (and MWh) and revenues Currency exchange risks 13) Explore new instruments that may reduce FX risk Federal government; multilat- that are affordable and available over the longer term eral agencies; financial sector (e.g., partial indexation in US$) Insufficiency at municipal level 14) Create national or state technical assistance pro- ELETROBRAS; development of technical and/or manage- grams banks; public banks; rial capacities for managing 15) Create standardized project evaluation tools public lighting 16) Introduce standard templates for EE public street SENAI; lighting contracts 17) Create guidelines for implementing or adjusting Brazilian Association of Mu- COSIP nicipal Administration; private 18) Standardize business models sector; Associations of Brazilian 19) Standardize financial contracts/instruments Municipalities Training Lack of common understand- 20) Ensure efficient and effective flow of information Ministry of Cities; Brazilian ing of the sector by the differ- between the municipal audit courts to promote consis- Associations of Municipalities; ent Audit Courts (TCs) tency 21) Arrange seminars for exchanging ideas between the Brazilian Institute of Municipal audit counts on the most controversial subjects affect- Administration; municipalities ing public lighting contracts engaging with TCs from the start of the process Lack of clear information 22) Introduce new databases for providing key data FNP (to increase the data about the sector related to EE street lighting projects for cities and po- available in www.comparabra- tential investors (city street lighting inventories, COSIP sil.com); ELETROBRAS; energy collection information, etc.) distributors; cities 99 8. Gaps in the market and recommendations Lack of information about 23) Standardization/certification of LED equipment Brazilian Association of Techni- product quality to allow valid cal Norms; INMETRO; certifica- comparisons to be made tion companies Perception of technological 24) Standardize equipment guarantees provided by Brazilian Association of Technology performance risks of LED suppliers Technical Norms; INMETRO; luminaires by municipalities 25) Increase use of other technology risk mitigants insurance companies; Ministry available in the market (e.g., insurance) of Industry and Commerce High cost of LEDs resulting 26) Create an industrial policy for national production Minister of Industry and from low level of national of LEDs (review import taxes, etc.) Commerce; Ministry of Finance; production 27) Increase national production by creating credit lines BNDES for the national production of LEDs Source: World Bank. As shown in Table 39, the vast majority of gaps and correspond to the list in Table 39 above. barriers require public sector involvement. The different levels of government will play an important This analysis shows that there are a handful of high- role in filling gaps such as: the lack of a national priority activities that could be implemented in the strategy and/or policy, lack of clarity in the legal short-term to help catalyze the LED street lighting interpretation of COSIP, the lack of mechanisms to market in Brazil. These include (i) identifying and mitigate municipal credit risk, municipal capacity- designating national leaders for leading coordination building, etc. Leadership at the national level will of EE street lighting initiatives, (ii) creating new public also be indispensable to prioritize and coordinate the credit lines, (iii) design investment funds offering aforementioned actions to ensure the efficient use of standardized EE asset class to attract private-sector resources and avoid duplication. investors at scale, (iv) creating instruments for mitigating municipal credit risk, and (v) finalizing Not all of the above recommendations can the standardization/certification of LED equipment. realistically be completed in the short term, and Other important steps will need to be taken in the some are more critical to the development of the medium and long term to expand the market, such LED street lighting market than others are. In order as standardization of business models and contracts, to facilitate stakeholder uptake and implementation establishment of new industrial policies to promote of the above recommendations, this report also domestic LED production, and revision of the categorizes the recommended actions in terms of legislative framework for public consortia to increase their indicative timeframe for implementation and their feasibility (e.g., allow them to take on financing their level of priority, as shown in Table 40 below. obligations directly). The numbers at the end of the recommendation 100 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 40 – Prioritization of recommendations and indicative implementation timeframe Short-term (<1 year) Highest priority recommendations Important recommendations Identify and designate federal, state, and/or municipal Create standardized project evaluation tools (#15) leaders to be responsible for leading coordination of EE street lighting initiatives and for helping municipalities to over- come barriers (#6) Create new credit lines and/or instruments for municipalities Create guidelines for implementing or adjusting COSIP (#17) unable to attract (sufficient) private investment. (#8) Design investment funds offering new, standardized EE asset Arrange seminars for exchanging ideas between the audit counts class to attract a variety of investors at scale (#9) on the most controversial subjects affecting public lighting contracts (#21) Create instruments for mitigating municipal credit risk (#11) Introduce new databases for providing key data related to EE street lighting projects for cities and potential investors (city street lighting inventories, COSIP collection information, etc.) (#22) Standardization/certification of LED equipment (#23) Medium-term (1–3 years) Highest priority recommendations Important recommendations Identify possible actors with regulatory knowledge to fill Standardization of the legal framework cities can use to earmark regulator gap, or at least to provide regulatory advice to COSIP funds (#3) municipal governments when necessary (#1) The federal government interpret and clarify what are the Establish a uniform approach for oversight agencies verifying legal uses of COSIP (e.g., for expenditures electricity, O&M, the legal provisions applied to different procurement bidding expansion, modernization). (#2) modalities. (#5) Standardize business models (#18) Federal government should design a national strategy for EE street lighting, including targets for converting the national public street lighting network to LEDs (#7) Standardize financial contracts/instruments (#19) Explore new instruments that may reduce FX risk that are afford- able and available over the longer term (e.g., partial indexation in US$) (#13) Create an industrial policy for national production of LEDs Create national or state technical assistance programs (#14) (review import taxes, etc.) (#26) Introduce standard templates for EE public street lighting con- tracts (#16) Ensure efficient and effective flow of information between the municipal audit courts to promote consistency (#20) Standardize equipment guarantees provided by suppliers (#24) Increase use of other technology risk mitigants available in the market (e.g., insurance) (#25) Increase national production by creating credit lines for the national production of LEDs (#27) 101 8. Gaps in the market and recommendations Long-term (>3 years) Highest priority recommendations Important recommendations Revise the legislative framework for public consortia to allow Provide electric utilities different incentives to promote EE street direct financing of the public consortium´s legal entity (#4) lighting by altering the form of compensation for energy distrib- utors, namely through decoupling sales (and MWh) and revenues (#12) Pass new legislation that allows exceptions to be made in municipal debt thresholds in order to enable greater invest- ments in EE public lighting (similar to exceptions that were provided to debt under PROCEL program) (#10) 102 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 103 9. Conclusions 104 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 9 - Conclusions In an increasingly urbanized world, global solutions benefits associated to better lighting quality such largely depend on locally designed public policies as reduced crime and the increased perception of and projects. Energy efficiency projects in cities, security by residents of communities. They also which now account for more than two-thirds herald the integration in the lighting services of energy consumption and more than 70% of network infrastructure of “smart city” services. emissions, can have a substantial positive impact at Meanwhile, cities such as New York and London, the global level. and even entire countries such as India, are making significant progress in implementing large-scale In the case of public street lighting in Brazil, there LED projects, thus helping to reduce further the is huge potential for increasing energy efficiency costs of this technology. with projects that can significantly reduce energy consumption and also meet the country’s climate Brazilian cities, although aware of the benefits, are change mitigation goals. However, implementing still on the threshold of this technological change this kind of project is not a simple task, and there and are faced with a series of economic, financial is no single recipe applicable to all cities. Solutions and regulatory barriers to developing their public such as adopting PPPs could benefit groups with street lighting projects. This is the case not only the appropriate scale and minimum financial/ at the local level, with the imbalance between the institutional capacity, while other groups will need amount of investment needed (CAPEX) and the the support of specific public policies. It is clear restrictions imposed by the Fiscal Responsibility that within these groups each municipality´s needs Law or the lack of access to credit, but also at the can vary. The complex nature of the public street macroeconomic level where extremely high interest lighting services in Brazilian cities presents many rates prevail and where deprecation of the national challenges and opportunities that need to be faced currency is an ever-present concern. in the quest for more efficient technologies. At the same time, Brazilian municipalities are in a The new LED lighting technologies are now relatively favorable position to make investments widely accepted throughout the world as a great in the sector, given the mounting costs of energy, opportunity for cities to reduce the operational and the declining costs of LED technology and, in most maintenance costs of public street lighting. These municipalities, the existence of ring-fenced funds technologies are also important on account of the that are available through the COSIP mechanism. 105 9. Conclusions Moreover, the 40%+ municipalities that only instruments can be tailored to the business models. recently became owners of their public lighting assets now have incentives to invest in energy The report also identified several challenges and efficiency and may wish to identify alternatives to opportunities related to energy efficient investments paying electricity utilities to operate and maintain in the public lighting sector in Brazil, which have their street lighting networks. been discussed and validated with public and private stakeholders. The report has also served to trigger With this initial survey of Brazilian municipalities dialogue among the interested parties, providing (in terms of size, fiscal management and public material for decision makers to focus on concrete lighting characteristics), it has been possible to agenda to modernize the public street lighting sector. identify groups with different capacities and needs. The classification of Brazil´s 5,570 municipalities in In order to speed up the implementation of energy six groups is a first step towards finding solutions efficiency projects in public street lighting, the tailored to their needs. We are aware that there is report has identified gaps and barriers in the current room for improvement in the proposed classification market, and proposed a series of “next steps” to fill by, for example, incorporating other relevant data the gaps and overcome barriers. Given that most that we have been unable to confirm, such as of these recommendations require governmental accurate data on the COSIP revenues or the level of intervention, it is important that a national level municipalities´ technical and institutional capacities. entity begins to exercise leadership in the sector. The changes that have occurred at the federal The eight business models presented above take government level over the past few months, account of the characteristics of the groups of including the creation of an Executive Secretariat in municipalities identified, and propose structures the Investment Partnerships Program (PPI) and the and sources of financing to enable the entire public introduction of Law 13,280/2016, which enhances the street lighting system to be scaled up. These models role of ELETROBRÁS (PROCEL) in the area of energy contain a complete range of financing sources - from efficiency, could well provide opportunities for federal models that are operated and financed directly by agencies to take a leadership role in promoting the the private sector through to national development energy efficient public street lighting agenda. programs or operations financed with municipalities´ own revenues. The World Bank Group stands ready to support the Brazilian government and other key stakeholders to Before selecting a particular business model, it is accelerate investments in modernizing the public important for municipalities to carry out a critical lighting sector and increasing energy efficiency. assessment of their own needs and capacity for designing and/or managing each business model. Once the business model has been chosen, the next step is to identify the best sources of finance to take the project forward, including, whenever necessary, additional risk mitigation mechanisms, such as guarantees for mitigating municipal credit risk. This report outlined 10 potential financing and credit enhancement instruments, and showed how these 106 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning 107 ANNEXES 108 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning ANNEXES ANNEX 1 – Summary of Business Models Model M1: Public-Private compliance with a range of responsibilities (installation, O&M) related to the system over the Partnership (PPP) PPP contract life. Some public lighting projects in Brazil are able to • SPV is formed by the winning consortium (for attract private capital through a PPP. This applies example, an operator, lender and manufacturer), especially for large cities with good credit standing. to be responsible for raising finance for the project. The following is an outline of the PPP model. • The city remunerates the concessionaire through monthly payments using COSIP (or, if Main Characteristics insufficient or nonexistent, the municipal budget). • The existence of a concessionaire. The • COSIP is collected by the electric utility and municipality awards a concession involving transferred to a municipal or escrow account. Figure 24 - Example of the structure of the PPP model Financier(s) Municipality Loans, debentures and/or grants $ Reembursement $ Additional Efficient street contributions lightning services SPV of EE public street Municipal Account / lighting consortium Monthly payments $ Escrow Account COSIP $ Manufacturer Operator Electricity Utility COSIP $ Financial flows Physical flows End Users 109 ANNEXES Groups Mitigating Factors • Groups A and B represent a total of 135 • Capacity-building for municipalities, municipalities (2.4% of Brazil’s municipalities) standardization of contracts. • These municipalities contain 40.9% of the • Implementation of COSIP and provision of total population and R$10.1 billion of investment credit guarantees. (42.4% of the required investment for the entire • Public Audit Court (TC) involvement in the country). initial phases of the project. Advantages • Performance guarantees provided by the • The public sector transfers most of the manufacturer and/or concessionaire. performance risk to the private sector, which has Figure 25 shows a potential scheme for the use of better ability to manage this risk. guarantees in a public lighting PPP to mitigate the above-mentioned risks. This scheme includes the Disadvantages creation of a reserve account as a partial guarantee of payment by the municipality to the public lighting • Transaction costs involved in a PPP consortium of the PPP. This mechanism is usually preparation can be significant. formulated to cover obligations from the Escrow Account during a term of three to six months. The Risks mechanism includes a sovereign guarantee by the National Treasury. • Lack of qualified municipal personnel. • Municipal credit risk, lack of funding. • Lack of regulatory framework. 110 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Figure 25 – Example of use of guarantees in the PPP model National Treasury World Bank Federal transfers to Partial municipalities Payment Financier(s) (collateral) $ Guarantees Local financial Municipality institution Loans, debentures Repayments $ and/or grants $ Additional contributions $ Letter of Credit EE lightining service EE public Municipal Account Reserve Account lightning consortium / Escrow Account $ Payments COSIP $ Technical / operational Premiums insurances Electricity Utility Financial flows Physical flows Insurers COSIP $ End Users Model M2: PPP with Main Characteristics Municipal Consortium • Municipal consortium formed by small- or medium-sized municipalities, with the creation of Given that it is not feasible for a large number of a SPV. municipalities to individually grant concessions using the PPP model, municipal consortia are a possible • Very similar to M1, although a municipal solution to generate the scale necessary for the consortium can be the granting authority. implementation of a PPP. International experience • SPV of the municipal consortium reimburses shows that there are huge economies of scale to be the concessionaire via monthly payments from derived from the procurement of luminaires. This COSIP (or, if insufficient or nonexistent, from the model is summarized below. municipal budget). • COSIP is collected by the electric utility and transferred to a municipal or escrow account. 111 ANNEXES Figure 26 - Example of the structure of the PPP/Municipal Consortium model End User Cosip $ Financial flows Physical flows Electricity Utility Cosip $ Municipal Consortium Financier(s) Municipality 1 Municipality 2 Municipality 3 Loans, debentures, $ $ and/or grants $ R Efficient public lightning Service SPV of EE public SPV of Consortium lighting consortium of Municipalities Manufacturer Operator $ Payment $ Escrow Account Groups • Lack of clear governance of the municipal consortium. • Most suitable for municipalities in Group • Municipal credit risk, lack of funding. C, due to small scale, typically less than 20,000 luminaires. • Lack of regulatory framework. Advantages Mitigating Factors • Increases the feasibility of the PPP model to a • Training for municipalities and the consortium; larger number of municipalities. standardization of contracts. • Diversification of municipal credit and political • Implementation of COSIP and provision of risk. credit guarantees. Disadvantages • Audit Court involvement in the initial phases of the project. • Consortium governance is more complex; transaction costs and perception of risk can be high. • Performance guarantees provided by the manufacturer and/or concessionaire. Risks • Lack of municipal consortium qualified personnel. 112 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Model M3: Municipal Finance Main Characteristics The large amount of capital needed to implement LED • The municipality takes on loans or issues technology is the main challenge to the conversion of bonds. the public lighting network, considering the relatively • The municipality reimburses the lender high upfront outlay. In Model M3, the municipality through monthly payments using COSIP (or, if itself is responsible for raising finance and making insufficient or nonexistent, from the municipal investments. Below is a summary of the model. budget), collected by the electric utility and transferred to a municipal or trustee account. • The municipality is responsible for undertaking O&M services either on its own, or outsourcing and overseeing them. Figure 27 – Example of the structure of municipal financing model Financier(s) Funds $ Reimbursement $ Equipment Services Manufacturer Municipality Outsourced Company Additional contributions $ COSIP$ Payment $ Payment $ Municipal Fund COSIP $ Electricity Utility COSIP $ Financial flows Physical flows End Users 113 ANNEXES Groups • Consortia experience in the solid waste treatment sector is an advantage. • The relatively small municipalities in • Implementation of COSIP, credit Group B (88 total) and Group C (329 total), guarantees. have good fiscal management and scale, but perhaps this scale is not sufficient to justify the Model M4: Electric transaction costs associated with structuring a PPP. Utility Programs As previously described, in both the national Advantages programs dedicated to the public lighting sector (PROCEL-Reluz and PEE), electricity utilities • Project structuring less complex (fewer have an important role in the financing and/ players, routine bidding process already set or facilitation of investments for large-scale forth in Law 8.666/1993). public street lighting projects. Given their lack of resources, these two programs currently have little Disadvantages potential for increasing the volume of investments, but it is possible to envisage a scenario where • Restricted ceilings of municipal utilities could play a major role in modernizing indebtedness disqualify many municipalities Brazil’s public lighting network, as well as being • Municipalities could use their available tax involved in a wider range of energy efficiency resources for investments in other areas that projects. Below is a description of the model. could not be undertaken by the private sector (opportunity cost). Main Characteristics • The public sector assumes most of the responsibility for project performance. • Represents an expansion of the PEE program, with some design changes. Risks • The electric utility grants loans to municipalities; these are recovered from two • Lack of technical and human know-how to sources: from the municipality (COSIP or manage the process. municipal budget) and from increased electricity • Municipal credit risk, investors´ lack of tariffs for end users (as already occurs under the interest PEE program). • Net receipts from the program would be Mitigating Factors transferred to a Revolving Fund for financing public lighting projects. • Training in best national/international practices; competitive biddings; adequate • The municipality is responsible for O&M technical performance guarantees; specialized services (on its own or outsourced). advisory services for the municipalities. 114 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Figure 28 - Example of structure of electricity utilities program model Revolving Fund Funds $ Repayment $ Electricity Utility End Users COSIP $ Loans $ Repayment $ COSIP $ Municipalities Municipal Account (s) $ Equipment Payments $ Services Financial flows Manufacturer Outsourced Company Physical flows Groups an environment where the trend is to reduce involvement in the sector by electricity utilities. • Municipalities in Groups D and E (>4.200 municipalities; 75% the population); in certain Risks cases, some municipalities from Groups C and F. (represent R$9.4 billion in investment). • Lack of interest by electricity utilities, or non- approval by ANEEL of the project concept. • Relatively small-scale, with relatively good fiscal management • Municipal demand for funds exceeds supply. • Municipalities lack qualified personnel to Advantages implement the project. • Loan costs below capital market levels. Mitigating Factors • Benefits from funding centralization and better risk diversification • Strong engagement with ANEEL over the benefits of the program; or, in the short term, the • Option for municipalities with few options to utilities would earmark more PEE resources for raise funds the public lighting sector87. Disadvantages • To restrict number of eligible municipalities (e.g., small–medium size) and provide capacity- • Requires regulatory change by ANEEL in building for municipalities. 87 This may be possible for concessionaires operating in areas with few low-income consumers. There is a regulatory obligation to invest 60% of the PEE funds in this group of consumers, but in some concession areas the market is already saturated and there is not sufficient demand in this sector for concessionaires to invest. A recent legal change (Law No. 13.280/2016) establishes a maximum of 80% of PEE investments in low-income areas, with no minimum limit. 115 ANNEXES Model M5: The ESCO Model Main Characteristics This model involves off-balance sheet financing • The company or consortium of companies for the municipality, where the investment is (SPV) raises funds, purchases and installs LED made by a third party (“third-party financing”). luminaires in exchange for regular payment In this way the investments for modernizing the by the municipality using COSIP and/or the public lighting network do not affect municipal municipal budget. indebtedness limits. Below is a summary of the • The municipality is responsible for O&M ESCO model. services (on its own or outsourced). • Two modalities: (1) the ESCOs share the efficiency gains; or (2) the ESCOs receive a fixed payment for the investment made and present a product performance technical guarantee. Figure 29 – Example of structure with ESCO model Equipment Repayment $ Manufacturer ESCO Financier Funds Payment $ Repayment $ Equipment Outsourced Services Municipalities Company Additional Payment $ COSIP $ contributions $ Municipal Account COSIP $ Electricity Utility COSIP $ Financial flows Physical flows End Users 116 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Group Risks • This model is applicable to Group C, • Lack of qualified municipal personnel to which totals 329 municipalities, characterized manage the project. by relatively small-scale, but with good fiscal • Lack of ESCOs´ financial capacity to ensure management. financing, thus reducing the scope and scale of the project. Advantages • Municipal credit risk. • Provides an off-balance sheet financing option for smaller municipalities, without requiring Mitigating Factors the same level of involvement by the regulator or political changes as model M4. • Capacity-building for municipalities. • Increase financing lines for ESCOs; focus on Disadvantages ESCOs that are subsidiary companies of electricity utilities; concentrate on small-scale projects. • Smaller coverage compared to M4. • Implement COSIP, credit guarantees. • Higher costs for municipalities given that ESCOs have less access to the municipal resources directly linked to consumer electricity payments. 117 ANNEXES Model M6: Centralized Procurement Main Characteristics This model seeks to capture economies of scale • Establishment of a consortium or other related to joint equipment (and possibly services) mechanism (or process) to undertake centralized procurement by multiple municipalities without the procurement. responsibility for raising finance. Below is a summary • If an SPV is established this would raise of this model. funds for the municipality; if not, municipalities are responsible for raising financing themselves. • The municipality is responsible for O&M services (on its own or outsourced). Figure 30 – Example of structure of centralized contracting model End Users COSIP $ Electricity Utility COSIP $ COSIP $ Funds $ Funds $ Financier(s) Financier(s) Municipality 1 Municipality 2 Municipality 3 Repayment $ Repayment $ COSIP and COSIP and other other sources $ sources $ Contract Equipment Payment $ and services Financial flows Physical flows Manufacturer + outsourced company 118 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Groups Risks • This model is applicable to municipalities in • Municipalities‘ lack of qualified personnel to Groups C and D (1.216 municipalities; 21.8% of the prepare technical specifications. population). • Complex coordination involved in raising • In some cases it could also apply to Groups E finance for several municipalities. and F. • Municipal credit risk; investors‘ lack of interest. Advantages Mitigating Factors • Potential for lower transaction costs compared to model M2, if SPV not established. • Capacity-building for municipalities. Disadvantages • Standards covering legal and financial issues for municipality consortia. • If a SPV is not established, municipalities’ • Implementation of COSIP, credit guarantees. difficulty to raise financing is not resolved. 119 ANNEXES Model M7: Self-Funding Main Characteristics When municipalities have no other options to raise • Municipalities undertake public lighting funds for their projects, one alternative is to self- modernization investments using public fund projects over a longer timescale reflecting the lighting revenues pari passu with corresponding municipality´s financing availability. A summary of expenditures and investments. this model is below. • Involves slow pace of investment (and a longer period of lighting retrofit), basically depending on whether COSIP revenues are subsidized or not with funds from the municipal budget. Figure 31 – Example of structure of self-funding model End Users COSIP $ Electricity Utility COSIP $ Municipality $ Equipment (+services) Manufacturer (+ outsourced company) Financial flows Physical flows Groups Advantages • This model is applicable to municipalities in • Low transaction costs due to not requiring Groups C and D (1.216 municipalities; 21.8% of the financial or institutional arrangements. population). • This could be one of the few viable options for • In some cases the model could apply to some cities unable to obtain government support. Groups E and F. 120 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Disadvantages • High cost per lamp due to small-scale purchasing; lack of resources to implement the • In municipalities with small scale and low project. staff qualification there is less space for creating a COSIP that would produce enough revenue to Mitigating Factors finance investments. • Capacity-building for municipalities. • Greater performance risk for municipalities. • Robust COSIP implementation to produce Risks surplus funds for the retrofit project. • Lack of municipal qualified personnel to prepare the technical specifications. 121 ANNEXES Model M8: Transfer of Luminaires Main Characteristics This model involves the transfer of HPS (or mercury • Interim transfer system of stocks of HPS from vapor) lamps, together with complete luminaires, municipalities that have been retrofitted with LEDs from cities that have been LED-retrofitted, to other to municipalities unable to afford LED lighting in cities that use less efficient lamps but which are at the short or medium term. present unable to modernize their public lighting • Creation of a coordinating structure to networks with LAD due to lack of finance, etc. handle bilateral transactions using transparent auctions organized by public or private agents. The model presupposes a massive conversion in the next few years of HPS and mercury vapor • The buyers would be municipalities or lamps to LEDs88. This would result in a large stock outsourced companies; the sellers would be of secondhand HPS lamps that will be available for municipalities or private entities that have acquired transferring to locations that cannot yet afford to the materials from the municipalities. invest in LED lighting and which still use other less efficient technologies. Figure 32 – Example of structure of luminaire transfer model Municipalities A Municipalities B Municipalities C Municipalities D HPS and mercury vapor $ Technical assistance / financing Coordinating Agent Público Entity $ HPS and mercury vapor Financial flows Municipalities E Municipalities F Physical flows 88 In 2012 the installed network of HPS lamps amounted to 11.4 million units, far more than the 1.2 million light points currently existing in the Group F municipalities. Metallic multivapor lamps totaled 201,000, and mercury vapor lamps (candidates for substitution) 3.8 million. There is also an operational public street lighting network using incandescent lamps (188,000), mixed lamps (283,000), fluorescents (160,000) and halogen lamps (10,900). 122 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Groups Risks • Groups E and F, with 4.219 municipalities • Complex coordination resulting in high (approximately R$1.6 billion) transaction costs. • Lack of interest or capacity of public and/or Advantages private agents in the scheme. • Provides an opportunity to improve energy • Lack of municipal resources or qualified efficiency and lighting services in municipalities personnel. with low buying power or financial/institutional difficulties. Mitigating Factors Disadvantages • Involvement of public sector agency to supervise the process and of private sector actors • The implementation of a sale/exchange/ interested in providing services. donation system may be operationally complex. • Prices charged to municipalities must • Possible shorter service life of the re-located adequately reflect the technical risks. equipment. • COSIP implementation consistent with creating a surplus to finance lighting modernization. 123 ANNEXES ANNEX 2 - S / P Correction between Scotopic Vision (S) and Photopic Vision (P) Table 41 - S / P correction factor between scotopic vision (S) and photopic vision (P) Low pressure sodium 0.25 High pressure sodium (HPS) 250 W clear 0.63 HPS 400 W clear 0.66 HPS 400 W coated 0.66 Mercury vapor (MV) 175 W coated 1.08 MV 400 W clear 1.33 Incandescent 1.36 Halogen headlamp 1.43 Fluorescent Cool White 1.48 Meetal halide (MH) 400 W coated 1.49 MH 175 W clear 1.51 MH 400 W clear 1.57 MH headlamp 1.61 Fluorescent 5000 K 1.97 White LED 4300 K 2.04 Fluorescent 6500 2.19 Source: Outdoor Lighting: Visual Efficacy. Volume 6, Issue 2. January 2009. Alliance for Solid-State Illumination Systems and Technologies. Lighting Research Center. Rensselear Polytechnic Institute 124 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning ANNEX 3 - Methodology for Generating Clusters and Grouping of Municipalities B razil has 5,570 highly heterogeneous municipalities. Designing financing models for the public street lighting of these requires the prior grouping of municipalities with similar characteristics in order to facilitate the segmentation of the modeling task and tailor the distinctive features of each group to a different model. This section presents the approach and procedures employed in the construction of a taxonomy of Brazilian cities. The study involved: database construction; cluster analysis; regrouping; and a representative analysis of the groupings thus obtained. 3.1 - Summary The first challenge of the project is therefore to identify groups of municipalities with similar of Approach and features, prior to suggesting the business models best suited to different local circumstances. In Methodology short, it is necessary to form substantial groups of municipalities to develop effective solutions for public lighting projects. The overall goal of the project was to identify business models for public street lighting projects The approach was based on three principles: in Brazilian municipalities. The municipalities are highly heterogeneous, both in terms of i. For the analysis input, a set of variables needed socioeconomic characteristics (size, income levels to be considered to mirror the underlying fiscal and development), and in their physical and and socioeconomic features of the municipalities demographic features, all of which are relevant while respecting data quality and availability to the modernization of the public lighting constraints. network in terms of demographic density, degree ii. The methodology of the analysis was based of verticalization of the urban mesh, etc. This on accepted scientific models. degree of heterogeneity makes it impossible to design a general model to apply to all Brazil´s iii. Regarding analysis outcomes, the number municipalities. of resulting groups needed to be small, so that 125 ANNEXES business models could be effectively applied based on the instruments and actors available 3.2 - Database in the Brazilian market. Note that two grouping procedures were performed and Selection — a statistical cluster analysis, followed by the identification of homogeneous clusters — for two of Variables main reasons: The dataset for the study is formed by Brazil’s existing 5,570 municipalities. In the selection of variables, we • The cluster model allowed the insertion of considered those that could potentially be used in the a set of key elements in the analysis such as financing models but which had not been considered the physical configuration and density of the previously, since the grouping and modeling needs network, as well as fiscal aspects, thereby to address different characteristics, although related producing a model far superior to a simple visual to the common goal (public lighting modernization). stratification with only two or three variables; There was a simple reason for this: to use a variable • Large cities, especially São Paulo and in the statistical model that had already been used Rio de Janeiro, tend to form clusters with a for the initial segmentation would have weakened single municipality, which could lead to a its predictive power, since the segments had already very heterogeneous grouping. In these cases become homogeneous in relation to this particular the cluster analysis needed to allow for the characteristic. formation of more homogeneous groups. It is not possible to design a generalized business After the clusters were identified, and in order model for public lighting projects. The primary to include in the analysis other aspects relevant challenge was therefore to group the municipalities to the development of the public lighting in terms of their similarities to develop solutions that network, clusters of municipalities were grouped could be adapted to each type of municipality. according to the variables network size and fiscal management. The first data source consisted of a research survey carried out in over 300 Brazilian municipalities with The two-step strategy allowed for consideration different demographic and economic profiles for us of multiple key factors to be considered (size, to gain a better understanding of the overall context development, fiscal situation, network density, of the country’s public street lighting sector. sectorial indicators) while maintaining a small number of clusters. Without these two steps it After tabulation and revision, the data obtained from would have been necessary to initially include this exercise were consolidated and calculated for a specific number of clusters in the statistical each of the states, for the geographic regions, and analysis, or to choose a type of stratification based for Brazil as a whole, with weightings based on the on few indicators. In the event we were able to number of municipalities per state. The precision obtain a small number of groupings that met the index for the survey results are shown in Table 42, key requirements of the study. by region. 126 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 42 - Precision index by region Region Precision index (%) North 86.3 Northeast 90.3 Center-West 81.3 Southeast 89.7 South 85.7 BRAZIL 94.5 Source: World Bank; Castagnari Consultoria Ltda. With this research data and a quantity of were drawn directly from the aforementioned sample socioeconomic information obtained on all the survey. Brazilian municipalities, the team was able to group the municipalities according to their Table 43 summarizes the characteristics considered capacity to develop different business models to most relevant for selecting well-delineated groups enable investments targeted at lighting sector — obtained from public data sources and which modernization. were intuitively related in some way to a potential financing outcome (default). The variables were The set of socioeconomic characteristics included the derived from public sources (IBGE, Firjan, National economic status of the municipalities, their physical Treasury, Ministry of Cities), with the exception of data size, the size and density of the lighting network, on the public street lighting situation in Brazilian levels of socioeconomic development and their fiscal municipalities, provided by the World Bank Group situation. Other sectorial indicators such as type of sample survey of 300 municipalities. lighting technology and existing network coverage 127 ANNEXES Table 43 - Database characteristics Variable Definition Objective/logic for using Unit Year of data GDP per capita GDP divided by number of inhabitants of Proxy of the municipality´s level of R$ 2012 municipality development Consumer units per Measurement of the water supply network Proxy of the level of “verticalization” Ratio 2013 water supply connec- density. A connection consists of the of the municipality, thus of the public tion consumer unit/units linked via a single street lighting network density. The extension to the distribution network. 89 public water supply system often covers the same area as the public lighting network. IFGF Index for measuring how taxes paid by Proxy of the level of the municipality´s Index 2013 community residents are administered by fiscal management. the municipal authorities. Composed of five indicators: Own Revenue, Personnel Costs, Investments, Liquidity, and Debt Cost. NCD/NCR Ratio between NCD (Net Consolidated Debt) To measure municipality´s indebted- Ratio 2015 and Net Current Revenue (NCR). 90 ness level. Number of light points Estimate of the number of light points in Proxy of the size of municipalities, and Number 2014 the public street lighting system. 91 to estimate the size of the public street lighting system - also a proxy for as- sessing the level of investment needed for a public lighting retrofit project. More than 20% mer- Indicator (yes/no) to estimate whether the Efficiency level of the current network Indicator 2014 cury vapor lamps used municipality uses more than 20% of inef- in order to gauge the possible level of in the public street ficient mercury vapor lamps in its public energy savings via a modernization lighting network 92 street lighting network 93 program Percentage of the Percentage of the municipality not covered Proxy for the amount of investment Percentage 2014 municipality not cov- by public street lighting.94 needed to modernize the existing net- ered by public street work versus extension of the network lighting to unserved areas. Source: IBGE; FIRJAN; National Treasury; Ministry of Cities, World Bank Group Note: The population was not considered, in view of the high correlation with GDP (95.6%). The size of municipalities was calculated on the basis The database included aspects of the configuration of the total number of existing light points in use. of the public lighting network of each municipality. Meanwhile, a municipality’s development level was To assess the extension and density of public assessed by its per capita GDP measured in monetary lighting networks, we researched the public water units. supply service, which essentially covered the same 89 For example, a building with ten apartments can have a single connection serving ten consumer units, therefore for this building the ratio consumer units/per connection = ten. 90 There are cases where the municipality may have negative NCD (i.e., cash availability exceeding financial liabilities). In this case a negative NCD/ NCR index indicates how much cash the municipality has in relation to its Net Current Revenue. 91 Using current available data and estimates using a regression model when data unavailable. 92 Technology used in the network estimated from the use of a proportion of more than 20% mercury vapor lamps. 93 Using current data when available and estimates using a regression model when they are not available. 94 In these variables, we used data from a sample survey done by the World Bank Group with Brazilian municipalities, with no data available for the entire universe of municipalities. 128 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning geographic area as that of the public street lighting negative NCD/NCR index indicates how much service. We observed, for example, the density cash the municipality possesses in relation to Net (consumer unit per connection)95 of the water Current Income (NCI). supply network, which also gave an idea of the To calculate the size of the public street lighting ‘verticalization’ of the urban space (i.e., a greater network, the number of light points were counted density of population per light point can make a directly or estimated for each municipality96. The public lighting project more feasible). obsolescence of the network was assessed on the basis of the use of 20% or more mercury vapor lamps The following two variables were used to show in the municipal system. Network coverage was the characteristics of the municipalities from the determined by the percentage of the municipality standpoint of their fiscal (tax) management and not covered by public lighting. For these variables, indebtedness: we used data from the above-mentioned World Bank survey. • Firjan Index of Fiscal Management (IFGF), which aims to measure how a population’ s taxes are We sought to include other data such as the value handled by the municipal governments. This index of the COSIP that might have been relevant to comprises five indicators: own revenue, personnel business models, but the databases consulted, such expenses, investments, liquidity and debt cost. as FINBRA, provided little information in this respect. • Ratio of Net Current Debt (NCD), which is We decided therefore not to take this variable into the amount of the Consolidated Debt of the account in the grouping. municipality deducted from the balance of financial assets, and Net Current Revenue (NCR - The matrix of the cross-correlations of the selected the total of tax revenues accruing from property, variables is shown in Table 44. It can be seen that agricultural, industrial, service-provision activities the level of correlation between pairs of variables is as well as other current transfers and receipts). relatively low, which indicates that all of them have a There are cases where the municipality may key explanatory power for the selection of the groups have negative Net Debt, i.e., with cash availability (i.e. the “redundancy” level is low)95. exceeding its financial liabilities. In this case, a Table 44 – Matrix of cross-correlations of selected variables Correlations a) GDP per b) Cons. c) IFGF d) NCD/NCR e) No. Light f) MV > 20% g) % w/out cap Units Points PL a) GDP per cap 1.000 0.124 0.323 –0.152 0.305 –0.087 –0.237 b) Consumer Units 0.124 1.000 0.163 –0.015 0.401 –0.077 –0.139 c) IFGF 0.323 0.163 1.000 –0.349 0.316 –0.132 –0.143 d) NCD/NCR –0.152 –0.015 –0.349 1.000 0.524 –0.028 –0.019 e) No. Light points 0.305 0.401 0.316 0.524 1.000 –0.088 –0.098 f) MV > 20% –0.087 –0.077 –0.132 –0.028 –0.088 1.000 0.024 g) % without public –0.237 –0.139 –0.143 –0.019 –0.098 0.024 1.000 lighting 95 A connection is an extension connected to the distribution network. For example, a building with ten apartments can have a single connection serving ten consumer units, therefore for this building the ratio consumer units/per connection = ten. 96 The model used for the estimated calculation of light points was derived from the data obtained from the sample survey of the socioeconomic characteristics of Brazilian municipalities. 97 The GDP variable was considered, but given the fact that it had a high correlation with the number of light points in the municipality (0.912) it was not used in the analysis. 129 ANNEXES 3.3 - Generation Net Current Revenue. Of the municipalities with available data, there is no public lighting in 7.45% of Clusters of their territories. • Cluster 2 (number of municipalities = 887): The purpose of cluster analysis is to group sample Comprises municipalities with an average per units of interest (i.e., individuals, companies, capita GDP of R$16,000. These possess a average of cities, countries, etc.) into categories, so that the 2,400 light points. While these municipalities have components have a high degree of affinity or fairly low economic standing, they are nevertheless homogeneity while maintaining a clear distinction well-managed (average IFGF of 0.63), with have an between classification groups. The cluster analysis average cash flow of 11.5% of net revenues. All the technique is used when the classification category municipalities with available data use more than structure is not known in advance and when there 20% mercury vapor lamps and on average 6.45% is little or no information about it. A single list of of their territories have no public lighting. observations has unknown grouping categories, and the core objective is to discover a classificatory • Cluster 3 (number of municipalities = 73): The structure that fits the available data. municipalities in this cluster are also not high- income (average per capita GDP of approximately The number of possibilities for grouping a small R$16,000) and an average of only 2,000 light sample of observations into a small number points. Moreover , a common feature is that they of groups can be very substantial. Most of have no information on the number of units per these possibilities are irrelevant or are simply extension. However, their fiscal management is variations that rival other possibilities of greater reasonable (IFGF average of 0.55), although this interest. Complex cluster analysis algorithms cluster includes one of the leaders in this respect: enable structures and relationships to be found Alvorada de Minas (MG), with an IFGF of 0.91, between observed data that are not evident from and an average cash flow of 6.4% of Net Current a quick visual inspection. The method allows these Income. The only municipality with appropriate possibilities to be reduced to the relevant choices data available has 80% of its territory with no for analysis of the problem. public lighting. The 18 clusters (homogeneous groups) obtained • Cluster 4 (number of municipalities = 1): An in the analysis are randomly described below as exclusive group of the city of Rio de Janeiro (RJ) follows98: on account of its unique characteristics compared to the rest of Brazil. Its development is reflected in • Cluster 1 (number of municipalities = 110): its income indicators (per capita GDP of R$34,000) The municipalities of this group have prominent and high verticality (2.31 units per extension — verticality (1.54 consumer units per connection) one of the highest in Brazil) and a large-scale among non-high income municipalities (average public lighting network (426,000 light points). GDP per capita R$ 19,500). They also have, on average, fewer light points (7,300), but with • Cluster 5 (number of municipalities = 872): These reasonable to good fiscal management (average municipalities are characterized by their small size IFGF of 0.59) and an average cash flow of 7% of (average of 2,700 light points) and average income The data presented here should be considered at the aggregate level. 98 130 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning (per capita GDP of R$20,000). Fiscal management is not covered by public lighting. is low (average IFGF of 0.48), and the average cash flow represents 3.6% of NCR. • Cluster 10 (number of municipalities = 4): This group comprises the municipalities of Fortaleza • Cluster 6 (number of municipalities = 97): The (CE), Salvador (BA), Belo Horizonte (MG) and municipalities in this group are characterized Brasília (DF)—with some of the highest verticality by their considerable size (an average of around in Brazil (an average of 1.59 units per extension and 37,000), reasonable income (average per capita an average IFGF of 0.66), although average income GDP of R$27,700), some verticality (average of 1.3 in these municipalities does not significantly units per extension) and good fiscal management exceed that of the rest of Brazil (per capita GDP of (an average IFGF of 0.61). R$28,500). • Cluster 7 (number of municipalities = 832): This • Cluster 11 (number of municipalities = 7): The group comprises the municipalities with the worst municipalities in this cluster are characterized indexes (an average of less than 1,000 light points), by the highest verticality indexes in the country income (average per capita GDP of R$6,800 ) and (average of 2.93 units per branch). In general, poor fiscal management (average IFGF of 0.23), they have a considerable public lighting structure as well as having average debts corresponding (average of 29,300 light points) and relatively to 6% of NCR. All municipalities for which data low incomes (per capita GDP between R$4000 is available use more than 20% mercury vapor and R$87,000), good fiscal management (with an lamps and on average 13.8% of the territories are average IFGF of 0.67) and an average cash flow of not covered by public lighting. 5.4% of net revenue. • Cluster 8 (number of municipalities = 7): • Cluster 12 (number of municipalities = 1): This Although these municipalities have a modest scale cluster contains only the city of São Paulo (SP), of public lighting (average of 21,000 light points), Brazil’s largest and richest city with more light they nevertheless have Brazil’s highest per capita points than any others: 600,000. GDP per capita is GDP - average of R$262,000. In general, they have also high (R$41,000), and considerable (1.59 units good fiscal management (average IFGF of 0.65) per extension). and an average cash flow (negative NCD) of 1.5% of NCR. • Cluster 13 (number of municipalities = 92): As with cluster 3, the municipalities in this group have • Cluster 9 (number of municipalities = 1,169): no data on consumer units per extension. Their This group covers municipalities that are more development indexes are similar to clusters 7 and developed than those in group 7 but with small- 9, with average per capita GDP of approximately scale public lighting (average of 1,000 light R$10,000, around 1,700 light points on average points), income (GDP per capita of R$8,100) and and critical fiscal management situation (average management (average IFGF of 0.43). Most of them IFGF of 0.34). Furthermore their debts correspond have no debt (average cash flow of 0.78% of net on average to 12% of revenue. revenue) and, among the municipalities with available data, 4% use more than 20% mercury • Cluster 14 (number of municipalities = 17): vapor lamps. An average of 47% of their territories This cluster contains a concentration of large 131 ANNEXES similar to clusters 7, 9 and 13, by low development indices: average of 1,500 light points, per capita GDP of R$8,100 and critical fiscal management status (0.33 of IFGF). Unlike the municipalities in clusters 7, 9 and 13 this group has debts corresponding on average to 59.6% of net revenue. • Cluster 17 (number of municipalities = 181): This cluster contains small-scale municipalities (less than 5,000 light points on average), but with a relatively high per capita GDP (average of R$44,500). They possess reasonable fiscal management (average IFGF of 0.55) and a cash flow of 2.6% of revenues. Of the municipalities with data available, they use more than 20% mercury vapor lamps in their public street lighting networks. • Cluster 18 (number of municipalities = 39): The municipalities in this group have a high per capita GDP (average of R$100,000) and a fairly large public lighting network (an average of 17,000 light points). They also have good fiscal management, municipalities with an average of 103,000 light with an average IFGF of 0.60 and average cash flow points), average per capita GDP of R$41,800 and of 7.8% of revenues. high verticalization (an average of 1.42 units per extension). In general they possess good fiscal management (average IFGF of 0.64) but with average debt levels of 12% of Net Current Revenue. 3.4 - Technical • Cluster 15 (number of municipalities = 460): Details of Cluster This group consists of municipalities with low incomes and a low scale of public street lighting Analysis99 (1,200 light points) and average per capita In practice, we compare individuals according to GDP of R$10,000. However, they are somewhat observable characteristics, and each characteristic of better-managed (average IFGF of 0.54) than the interest defines a variable of the process. Through municipalities in other low development clusters. the observed variables of each individual cluster, The municipalities with NCD / NCR data possess a analysis determines the closeness or distance that considerable cash reserve - an average of 109% of separates individuals, so that nearby individuals are net revenue. placed in the same cluster (small distance within the groups), and distant individuals (large distance • Cluster 16 (number of municipalities = 721): between groups) in different clusters. The municipalities in this group are characterized, 99 This section presents a more technical approach to cluster analysis. For readers interested in a less technical approach this section can be avoided without loss of understanding of the context of the analysis. 132 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Figure 33 – Schematic diagram of cluster analysis Minimize distance within clusters Maximize distance between clusters By analyzing a set of p dimensions or attributes of a the groups is then calculated and the observation is particular individual (with real values), we can see the allocated to the nearest centroid group. individual as a point in a p-dimensional space. Thus it is possible to define the proximity between two points with Once all the observations have been allocated and the real value characteristics when using a metric of the groups formed, the algorithm recalculates the centroids distance between both points (in this case the Euclidian of these and restarts the algorithm, calculating the metric): distances between the observations and the centroids. The algorithm ceases to be used when there is no further group change between the observations. p d(X,Y) = (xi -yi ) 2 Grouping largely depends on the initial seed choice — two different choices can lead to completely different i=1 groupings (MILLIGAN, 1980). Fortunately there are seed selection techniques that ensure that the resulting where X and Y represent two different points as groups will be well-discriminated. For more details on p-dimensional vectors. such methods see Pavan et al. (2011). We used the k-means algorithm (MACQUEEN, 1967), Although there are no totally satisfactory methods for which is initialized with the desired number of groups determining the optimum number of groups (EVERITT, (i.e., the number of clusters must have been previously 1980; Harkan, 1985; BOCK, 1985), the decision criterion established). For k clusters, k initial points known as used in this analysis is based on two metrics known seeds are defined in the p-dimensional space of the in the scope of the cluster analysis. The first is the data that will function as the centroids of the k groups. pseudo-F statistic (CALINSKI; HARABASZ, 1974), which The distance between each point and the centroid of is a ratio between the sum of squares between groups 133 ANNEXES and within groups, weighted by their respective levels choosing the number of clusters as well as the pseudo-F of freedom (which are functions of sample size and the (high CCC values = good groupings). number of clusters). 50 independent cluster analyses were performed, The idea of using pseudo-F statistics is to evaluate each with a number of clusters ranging from 1 to whether the groups formed have very concentrated 50. The pseudo-F and CCC statistics were computed elements (i.e., low variance in the group) while, as for each cluster, and the former led to the graph in groups, remaining distant from each other (i.e., Figure 34. Since there are several peaks in a region high variance between groups). We can analyze the where the pseudo-F is high (between 6 and 18 evolution of pseudo-F statistics as the number of groups), the approach was to establish a threshold clusters gradually increases, and we have to consider value for the statistic and treat all clusters above the the extent to which the pseudo-F statistic increases in threshold as potential candidates. These indicate the relation to the previous number of clusters (MILLIGAN, quantities of groups in which the discrimination of COOPER, 1985). the municipalities (in terms of reduction of group variance) is optimal. In this study a Harper value of The second metric is the Cubic Clustering Criterion 1,400 was chosen (Figure 34), and the 8, 15 and 18 (CCC), created by the SAS software (SARLE, 1983). This groups stood out as potential candidates. is a statistic that compares the variance between the groups obtained with the variance of these groups in the In order to make the final decision on the number of hypothetical case of the observations being distributed groups, the CCC statistic (Figure 35) was compared uniformly (plus making cluster analysis impracticable). for the numbers of groups exceeding the specified The positive values of this statistic indicate that the ceiling, where it is possible to observe that the group variance between the groups formed is greater than it of 18 clusters has an advantage, with a much greater would be if the data had a uniform distribution, thus CCC than the analyses with the 8 and 15 clusters. Given indicating in this case that Cluster Analysis is necessary. the criteria outlined above, and also to enhance the The author also indicates the statistic as a criterion for granularity of the groups, 18 groups were considered. Figure 34 - Evolution of pseudo-F statistics for different amounts of groups 1600 1400 1200 1000 Pseudo F - 800 600 400 200 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Number of clusters Source: World Bank Group; Pezco Consultoria 134 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Figure 35 - Evolution of CCC statistic for different amounts of groups 300 250 200 150 ccc 100 50 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Number of Clusters -5 0 Source: World Bank Group; Pezco Consultoria After forming the 18 groups, Table 45 shows for each by the clusters. The index also reveals the relevance of the variables a comparison between the variances of the variable in the formation of groups (different within and outside the groups, resulting in an index from statistical significance insofar as no hypothesis that determines how well the variable is broken down test is performed). Table 45 - Relevance of variables in the formation of groups Correlations R²/(1 - R²) Per capita GDP 3.80 Extensions/consumer unit 7.26 IFGF 3.57 NCD/NCR 1.57 No. light points 10.53 MV > 20% 8.80 % with no public lighting 2.29 It can be seen that the variables reflecting the number of light points) the proportion of non-coverage by of light points, the utilization of mercury vapor public street lighting differs very little between the lamps and the consumer units per per connection groups. are the most outstanding features of the groups, while the NCD / NCR ratios and the percentage of The dispersion diagrams (Figures 35 and 36) show territories not covered by public lighting are lower the standardized averages of the clusters obtained in the ranking. This means that although there are and therefore give an idea of their location. Note that clearly distinct groups (e.g., in terms of the number the graph in Figure 35, involving two key variables for 135 ANNEXES the cluster, shows groups that are distant from each not clearly show where the data should be divided to other. This graph, however, does not show groups 4, form the groups — the distance between the averages 10 and 12 due to the scale of the graph (they have of the clusters here is considerably smaller (note standardized average light points of 30, 14 and 42 the scale of the graph). It is also worth considering respectively). The distance between the averages of that the two graphs are comparable because the the groups shows that the two variables show clear variables observed have been reduced to the same divisions n the data (hence the need for a breakdown). scale. In this way they possess an average of 0 and a Figure 36 shows two less relevant variables and does variance of 1. Figure 35 - Dispersion diagram of the clusters averages of the variables GDP and consumer units per connection Cluster averages (standardized) 9 8 11 Consumer units per extension 7 6 5 4 3 1 2 6 14 1 2 18 7 17 8 0 5 16 -1 15 -1 90 1 2 3 4 5 6 7 8 -2 -3 -4 3 -5 13 Number of light points Source: World Bank Group; Pezco Consultoria Figure 36 - Diagram of dispersion of the averages of the clusters relative to the NCD / NCR variables and % of non-coverage of public lighting Cluster averages (standardized) 5 3 4 % of unserved areas 3 9 2 1 7 5 8 4 2 0 13 1 14 16 -5 15 -4 -3 -2 -1 0 1 2 18 17 6 10 11 -1 NCD/NCR Source: World Bank Group; Pezco Consultoria 136 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 46 shows, for each of the 18 groups, the averages street lighting. Using this criterion, the higher the of the variables considered in the study, while Table number, the worse the group is in relation to the 47 lists the groups within each variable (from best to specific variable. If two or more groups have the worst average). For example, a municipality with a same average, the group with the lowest standard high use of mercury vapor lamps is interesting from deviation is favored. the standpoint of the need for it to invest in public Table 46 - Average variables in each group Cluster No. of mu- Number Per capita Econ/Lig IFGF NCD/NCR No. of MV > 20% nicipalities of light GDP (Brazil- light points ian R$) points 1 110 7,297 19.476,50 1,54 0,59 –7,07 0,00 7.45 2 887 2,442 16.090,86 1,07 0,63 –11,52 1,00 6.45 3 73 1,953 15.889,60 0,00 0,55 –6,43 0,00 80.00 4 1 426,609 34.232,33 2,31 0,82 58,69 0,00 10.00 5 872 2,755 20.109,47 1,06 0,48 –3,57 0,00 3.97 6 97 36,876 27.785,22 1,30 0,61 6,39 0,07 3.07 7 832 907 6.866,14 1,02 0,23 6,31 1,00 13.83 8 7 21,249 262.525,95 1,08 0,65 –1,55 – – 9 1,169 1,053 8.100,13 1,03 0,43 –0,78 0,04 47.25 10 4 206,000 28.518,58 1,59 0,66 25,65 0,00 1.33 11 7 29,388 35.982,75 2,93 0,67 –5,43 0,00 0.00 12 1 600,000 41.978,81 1,59 0,77 185,74 0,00 7.00 13 92 1,763 9.985,46 0,00 0,34 12,00 0,00 7.80 14 17 103,018 41.797,52 1,42 0,64 12,61 0,00 4.22 15 460 1,224 10.096,27 1,03 0,54 –109,00 0,00 3.45 16 721 1,504 8.177,45 1,04 0,33 59,57 0,00 9.02 17 181 4,978 44.574,25 1,06 0,55 –2,60 1,00 1.67 18 39 16,925 100.173,63 1,12 0,60 –7,79 0,00 0.00 Source: World Bank Group; Pezco Consultoria 137 ANNEXES It is important to note, among other things, that lamps. Groups 7, 9, 13 and 16 have a negative profile groups 4, 10 and 12 are prominent from the in terms of development (number of light points, GDP standpoint of scale (number of light points) and per capita) and fiscal management. Table 48 assigns fiscal management (IFGF), groups 8 and 18 by income A, B, C and D ratings to the groups by considering the (GDP per capita), groups 4 and 11 by verticalization averages of each variable, while respecting some of (consumer units per extension), and groups 2 and 7 the thresholds suggested in Table 48. by the number of municipalities using mercury vapor Table 47 - Classification of clusters in each variable Cluster No. munic. Per capita GDP (Brazilian Econ/Lig IFGF NCD/NCR No. light MV > 20% % without R$) points IP 1 110 R$ 19,476.50 1.54 0.59 –7.07 7297 0.00 7.45 2 887 R$ 16,090.86 1.07 0.63 –11.52 2442 1.00 6.45 3 73 R$ 15,889.60 0.00 0.55 –6.43 1953 0.00 80.00 4 1 R$ 34,232.33 2.31 0.82 58.69 426609 0.00 10.00 5 872 R$ 20,109.47 1.06 0.48 –3.57 2755 0.00 3.97 6 97 R$ 27,785.22 1.30 0.61 6.39 36876 0.07 3.07 7 832 R$ 6,866.14 1.02 0.23 6.31 907 1.00 13.83 8 7 R$ 262,525.95 1.08 0.65 –1.55 21249 . . 9 1169 R$ 8,100.13 1.03 0.43 –0.78 1053 0.04 47.25 10 4 R$ 28,518.58 1.59 0.66 25.65 206000 0.00 1.33 11 7 R$ 35,982.75 2.93 0.67 –5.43 29388 0.00 0.00 12 1 R$ 41,978.81 1.59 0.77 185.74 600000 0.00 7.00 13 92 R$ 9,985.46 0.00 0.34 12.00 1763 0.00 7.80 14 17 R$ 41,797.52 1.42 0.64 12.61 103018 0.00 4.22 15 460 R$ 10,096.27 1.03 0.54 –109.00 1224 0.00 3.45 16 721 R$ 8,177.45 1.04 0.33 59.57 1504 0.00 9.02 17 181 R$ 44,574.25 1.06 0.55 –2.60 4978 1.00 1.67 18 39 R$ 100,173.63 1.12 0.60 –7.79 16925 0.00 0.00 138 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning Table 48 - Criteria for assigning score for each variable Score Per capita GDP Econ/Lig IFGF NCD/NCR No. light MV > 20% % without IP (in R$1000s) points (thou- sands) D < R$ 15 N/D < 0.4 > 100 <2 — > 15 C [R$ 15 ; R$ 30 ) < 1.2 [0.4; 0.6) (10; 100] [2 ; 20) < 1 ou. (8; 15] ou. B [R$ 30 ; R$ 60 ) [1.2; 1.8) [0.6; 0.8) (–10; 10] [20 ; 100 ) [1;50) (4; 8] A >= R$ 60 >= 1.8 >= 0.8 <= –10 >= 100 >= 50 <= 4 3.5 - Regrouping Reducing the number of clusters involved carrying out a new procedure consisting of grouping 18 clusters into a total of six larger groups by making The clusters analysis method produced a number a qualitative evaluation of characteristics that of clusters that cannot be easily identified with had similar prospects for the implementation business models (i.e., nonoperational). In principle, of appropriate business models. The regrouping the analysis should result in a small number of mainly took into account scale and fiscal clusters, so that business models can be effectively management, as can be seen in Table 49. applied based on a set of instruments and actors currently available in Brazil Table 49 - Regrouping of clusters according to main characteristics Group No. of mu- Clusters incorpo- Size of municipality100(num- Fiscal management (IFGF average; NCD/ nicipalities rated ber of light points) NCR) A 47 4, 10, 12, 14* Very large >50.000 Relatively well-managed IFGF average >0.6 NCD/NCR average > 0 B 88 6, 8, 11, 18 Large Relatively well-managed IFGF average >0.6 20.000–50.000 NCD/NCR average > 0 C 329 1, 17** Medium-sized Relatively well-managed IFGF average = 0.6 <20.000 NCD/RC average L < 0 D 887 2 Small Relatively well-managed IFGF average >0.6 Average< 5,000 Average NCD/NCR < 0 E 3.406 3, 5, 7, 9, 15 Very small Moderately managed Average < 2,000 IFGF average >0.4 NCD/NCR average < 0 F 813 13, 16 Very small Limited management Average < 2,000 IFGF average <0.3 NCD/NCR average > 50 *Includes municipalities in Group B with more than 50,000 light points ** Includes municipalities in Group B with less than 20,000 light points Source: World Bank Group; Pezco Consultoria Thresholds of 20,000 and 50,000 light points were identified through consultations with market agents. 100 139 ANNEXES 3.6 - Relevance municipalities, size of resident population, number of light points and the amount of capital expenditure of Clusters (CAPEX) required to convert the public lighting network to LED technology. The relevance of groups formed by clustering is shown in Table 50—in absolute numbers of Table 50 - Relevance of groups of municipalities, absolute numbers Groups No. of municipalities Population Light points (estimated) Capex (estimated), in R$ A 47 59.878.706 5.052.440 6.547.962.240,00 B 88 23.763.643 2.771.156 3.591.418.176,00 C 329 14.706.648 2.073.597 2.687.381.712,00 D 887 23.033.096 2.161.580 2.801.407.680,00 E 3.406 64.445.770 5.085.685 6.591.047.760,00 F 813 18.613.109 1.246.586 1.615.575.456,00 BRASIL 5.570 204.440.972 18.391.044 23.834.793.024,00 Source: World Bank Group; Pezco Consultoria Table 51 presents the relative importance of the terms over national aggregates). clusters in the Brazilian context. (in percentage Table 51 - Relevance of groups of municipalities,% Groups No. of municipalities Population Light points (estimated) Capex (estimated), in R$ A 0,84% 29,29% 27,47% 27,47% B 1,58% 11,62% 15,07% 15,07% C 5,91% 7,19% 11,28% 11,28% D 15,92% 11,27% 11,75% 11,75% E 61,15% 31,52% 27,65% 27,65% F 14,60% 9,10% 6,78% 6,78% BRASIL 100,00% 100,00% 100,00% 100,00% Source: World Bank Group; Pezco Consultoria 140 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning ANNEX 4 - International Standards and Specifications for High Performance Luminaires The Global Lighting Challenge (GLC), a Clean Energy Deployment (SEAD)101 initiative has established Ministerial (CEM) initiative, aims to install 10 billion quality standards for industrial and street lights that high-efficiency lights such as LEDs, and make them represent typical standards for the top 2–3% of the available to a large part of the population. One of best products on the market. These products are the basic principles of the initiative is to use products eligible for the “Global Efficiency Medal”. that have high standards of efficiency, performance and durability. By way of example, the following table lists some of the standards required for high quality external The GLC’s Super-Efficient Equipment and Appliance luminaires used in public street lighting networks. Features Standard recommended for external luminaires 1) Minimum efficiency of luminaire at ground level • 120 lumens per watt 2) Color Reproduction Index (CRI) • 70 3) Power factor • 0.90 4) Total harmonic distortion <20% 5) Operating temperature From –30 oC to +50 ºC 6) Early faults <5% faults during first 6.000 hours 7) Useful life expectancy 50.000 hours 8) Presence of toxic elements Meets requirements of the RoHS (Restriction of Hazardous Substances). European Union Ministerial Directive 2002/95/EC. Source: Global Lighting Challenge (GLC), 2016. Available at: 101 141 ANNEXES REFERENCES AGENCE FRANÇAISE DE DÉVELOPPEMENT (AFD); INSTITUTO DE PESQUISA ECONÔMICA APLICADA (IPEA); FUNDACIÓN CIUDAD HUMANA (FCH). O financiamento da cidade latino-americana: instrumentos a serviço de um desenvolvimento urbano sustentável. Série Savoirs Communs, n. 16, Mar. 2014. ARVATE, Paulo et al. Lighting and violent crime: evaluating the effect of an electrification policy in rural Brazil on violent crime reduction. Discussion text No.408 – CMICRO, n. 33, Jan. 2016. São Paulo: Escola de Economia de São Paulo da Fundação Getúlio Vargas (FGV-EESP). BOCK, H. H. On some significance tests in cluster analysis. Journal of Classification, n. 2, pp. 77–108, 1985. CALINSKI, T.; HARABASZ, J. A dendrite method for cluster analysis. Communications in Statistics, n. 3, pp. 1–27, 1974. CARVALHO, A.X.Y.; DA MATTA, D ; RESENDE, G.M. Clusterização dos municípios brasileiros. In: CARVALHO, A.X.Y. et al. Dinâmica dos municípios. Brasilia: IPEA, 2008. CATTELL, R. B. A note on correlation clusters and cluster search methods. Psychometrika, n. 9, pp. 169–184, 1944. CENTRO DE ESTUDOS E PESQUISAS EM ADMINISTRAÇÃO MUNICIPAL (CEPAM). Iluminação pública: Guia do gestor. São Paulo. Fundação Prefeito Faria Lima. July, 2013 DIETRICH, J. L. Mecanismos de financiamento para o saneamento básico. In: PHILIPPI JR., A.; GALVÃO JR., A. C. Gestão do saneamento básico: abastecimento de água e esgotamento sanitário. São Paulo: Manole, 2012. EUROPEAN COMMISSION. Lighting the Cities: Accelerating the Deployment of Innovative Lighting in European Cities. European Commission - Directorate-General Communications Networks, Contents & Technology. June, 2013. EVERITT, B. S. Cluster analysis. 2nd. ed. London: Heinemann Educational Books, 1980. GOLDMAN SACHS WEBSITE. The Low Carbon Economy: GS SUSTAIN: Equity investor’s guide to a low carbon world, 2015–25. New York: Equity Research, 30 Nov. 2015. HARTIGAN, J. A. Statistical Theory in Clustering. Journal of Classification, n. 2, pp. 63–76, 1985. IPWEA, 2014. Practice Note11: Towards More Sustainable Street Lighting (July 2014). Prepared by Next Energy and Strategic Lighting Partners for the Institute of Public Works Engineering Australasia (IPWEA) and the Australian Centre of Excellence for Local Government (ACELG), University of Technology, Sydney (UTS). 142 Lighting Brazilian Cities: Business Models for Energy Efficient Public Street Lightning MACQUEEN, J. B. Some methods for classification and analysis of multivariate observations. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, n. 1, p. 281-297, 1967. MILLIGAN, G. W. An examination of the effect of six types of error perturbation on fifteen clustering algorithms. Psychometrika, n. 45, p. 325-342, 1980. MILLIGAN, G. W.; COOPER, M. C. An examination of procedures for determining the number of clusters in a data set. Psychometrika, n. 50, p. 159-179, 1985. NASCIMENTO, E. R.; DEBUS, I. Lei Complementar n. 101/2000: Entendendo a Lei de Responsabilidade Fiscal. Brasilia: National Treasury Secretariat, 2002. PAVAN, K. K. et al. Robust seed selection algorithm for k-means type algorithms. Journal of Computer Science & Information Technology (IJCSIT), v. 3, n. 5, p. 147-163, 2011. Makumbe, P.; Weyl, D.K; Eil,A; Li,J . Proven Delivery Models for LEDs Public Lighting. World Bank Group, ESMAP, 2016. SAADI, M.. A figura da Cosip e as PPPs no setor de iluminação pública. Revista Brasileira de Infraestrutura – RBINF, Belo Horizonte, Yr. 3, n. 6, p. 169-183, Jul./Dec. 2014. SARLE, W. S. The cubic clustering criterion. SAS Technical Report A-108. Cary, NC: SAS Institute Inc., 1983. SINGH, Jas. Why energy efficiency matters and how to scale it up. Live Wire, World Bank Group, 2016. SOKAL, R.; SNEATH, P. Principles of numeric taxonomy. San Francisco: W. H. Freeman, 1963. THE CLIMATE GROUP. LED adoption by utilities, 2016. Available at: Lighting The Clean Revolution. The rise of LEDs and what it means for cities. June, 2012. Available at: . Accessed 26/10/2016. The Big Switch: Why it’s time to scale up LED street lighting. Sep. 2015. Available at: . Accessed 26/10/2016. TRYON, R. Cluster analysis. New York: McGraw Hill, 1939. TUROLLA, F. A.; GABRIELLI, M. F.; GONDIM, I. J. C. Crédito e financiamento à infraestrutura no Brasil. Revista Tecnologia de Crédito (Serasa-Experian), v. 1, p. 1, 2013. WELSH, Brandon; FARRINGTON, D. Effects of improved street lighting on crime. Campbell Systematic Review, v. 13, n. 1, 2008. 143 1