Rethinking Infrastructure in Latin America and the Caribbean Spending Better to Achieve More Marianne Fay Luis Alberto Andres Charles Fox Ulf Narloch Stephane Straub Michael Slawson © 2017 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522- 2625; e-mail: pubrights@worldbank.org. Design by sonideas (www.sonideas.com) Acknowledgments T his report is a joint product of the Latin America and Caribbean Region Vice-Presidency and the Chief Economist Office of the Sustainable Development Practice Group. The task team leaders are Marianne Fay and Luis Alberto Andres. Contributors to this report include Charles Fox, Ulf Narloch, Michael Slawson, and Stephane Straub. The following World Bank colleagues and consultants contributed background papers and substantive inputs: Daniel Benitez, Diana Cubas, Steven Farji, Xijie Lu, Maria Claudia Pachon, and Tatiana Peralta Quiros (Transport); Gabrie- la Elizondo Azuela, Jiemei Liu, Farah Mohammadzadeh, and Patricia Vargas (Energy and Extractives); Aroha Bahuguna, Christian Borja-Vega, Gustavo Perochena Meza, Diego Rodriguez, Gustavo Saltiel, and Virginia Ziulu (Water); Shohei Na- kamura and Carlos Rodriguez Castelan (Poverty); Kevin McCall and Juan Jose Miranda Montero (Environment); Beatriz Eraso, Catalina Marulanda, Emanuela Monteiro, and Carlos Perez-Brito (Social and Urban); Catiana Garcia-Kilroy, Cledan Mandri-Perrot, Heinz Rudolph and Fernanda Ruiz Nunez (PPP); Jorge Araujo and Naotaka Sugawara (Fiscal); Sebastian Lopez Azumendi (Institutions); Diego Dorado, Ian Hawkesworth, and Jens Kromann Kristensen (Public Spending Effi- ciency); Cesar Chaparro Yedro, Tania Ghossein, Khasankhon Khamudkhanov, Federica Saliola, and Mikel Tejada Ibanez (Procurement). Additional background papers were commissioned from KTH Royal (Oliver Broad, Mark Howells, and Gustavo de Moura), Fernando Miralles-Wilhelm, and Daniel Nolasco. The team is grateful for guidance provided by Jorge Familiar Calderon, Vice President of the Latin America and Caribbe- an Region, and Karin Erika Kemper, Senior Regional Adviser, as well as comments and helpful suggestions from Cecilia Briceno-Garmendia, James Brumby, Uwe Deichmann, Alfonso Garcia Mora, Ejaz Ghani, Jesko Hentschel, Jens Kromann Kristensen, Antonio Nucifora, Carlos Perez-Brito, Pablo Saavedra, Jordan Schwartz, Luis Serven, Carlos Silva-Jauregui, Stephane Straub, Marijn Verhoeven, and Jan Weetjens. Paul Holtz edited the report. The findings, interpretations, and conclusions expressed in this document are those of the authors and do not necessar- ily reflect the views of the Executive Directors of the World Bank, the governments they represent, or the counterparts consulted or engaged with during the informality study process. Any factual errors are the responsibility of the team. 3 Table of contents Acknowledgments 3 Executive summary 7 What is the goal? And how to set it? 9 How to improve services as cost effectively as possible? 11 Who should pay—and what does it imply in terms of financing options? 13 Part I. Infrastructure in Latin America and the Caribbean: Modest spending, uneven results 17 How much does Latin America spend on infrastructure? 18 What is the region getting for its money? 21 Transport: Unimpressive outcomes, but infrastructure is just part of the challenge 21 Water and sanitation: Good coverage for water, but sanitation an increasingly urgent challenge 29 Energy: A sector at a turning point? 35 Part II. What lies ahead for the region’s infrastructure? 43 Inefficient public spending may limit how much more can or should go to infrastructure 43 A tight fiscal stance limits how much more could be spent on infrastructure 48 Public investments in Latin America: Rising during the boom, but still much lower than in other regions 48 A “bifurcated” fiscal panorama in Latin America 49 Climate change is creating new challenges, but possibly new opportunities 52 Infrastructure needs to be more resilient and better adapted to the changing climate 52 Pressures will mount to reduce emissions from infrastructure 57 Urbanization and changing socioeconomics are complicating matters 62 Part III. The road ahead: spending better to meet “real” infrastructure needs 65 Focusing on priorities—setting the right goals is essential 66 Improving utility performance, and focusing public and concessional finance where it is truly needed 67 Commercial financing and the importance of making judicious use of public resources 68 Corporatization and the importance of improving the performance of utilities—public and private 70 Improving public investment management and spending efficiency 72 References 75 Annex 1. Public expenditure reviews examined for this report 80 Annex 2. Procurement performance of Latin American countries: relatively good, but with wide variation across countries and indicators 83 5 Executive summary L atin America and the Caribbean (henceforth re- Mexico. Many others—Bolivia, Costa Rica, Honduras, Nic- ferred to simply as Latin America) does not have aragua, Panama, Peru—invest more than 4 percent of the infrastructure it needs or deserves given its in- GDP a year. Transport and wastewater are real challenges, come level. Infrastructure also falls short of what is but the region performs quite well in electricity and wa- needed to advance social integration and achieve ter. In fact, Latin America’s clean, sophisticated electricity higher growth and prosperity. Moreover, the region’s in- sector could become a serious competitive advantage. frastructure does not correspond to the aspirations of its growing middle class. Second, the focus should be on the service gap, rather than on a notional, and largely hypothetical, investment Many argue that the solution is to spend more. With per- gap. To the question of, “how much is needed?” the re- haps the exception of Africa, Latin America does invest sponse should always be: “for what?” And the answer the least in infrastructure among developing regions as a should lie with countries’ aspirations of economic growth share of GDP—less than 3 percent compared with 4-8 per- and their social and environmental objectives, as well as cent elsewhere (table ES.1). So the story might seem sim- with their choices on the relative roles of infrastructure ple: the region underperforms on infrastructure and has to and other investments in achieving those aspirations. spend more to narrow its infrastructure “investment gap.” Third, the investment gap approach necessarily focus- But that story would not match the facts. es attention on the question of raising more resources. But closing the service gap should not—and, indeed, TABLE ES1:  Latin America invests the least in cannot—be just about spending more. The service infrastructure among developing regions gap can be narrowed, if not closed, in two other ways: (public and private infrastructure investments, latest year available) by ensuring that spending (particularly of scarce pub-  Region Percentage of GDP lic resources) is well targeted and that it is efficient. East Asia and the Pacific 7.7 This report has one main message: Latin America can Central Asia 4.0 dramatically narrow its infrastructure service gap by Latin America and the Caribbean 2.8 spending efficiently on the right things. It remains to Middle East and North Africa 6.9 be seen whether spending better will be sufficient for the South Asia 5.0 region to fully achieve what it aspires to. But there is suffi- Sub-Saharan Africa 1.9 cient evidence that spending better and focusing scarce Source: http://Infralatam.info; ADB 2017; own estimates. public resources on what matters would significantly nar- Note: No data was available for Eastern Europe. Applying these shares to 2014 row the service gap. GDP figures suggests Latin America accounts for about $180 billion out of total developing country infrastructure spending of about $1.5 trillion. The “spend better” message is also pragmatic. Most Latin American countries have limited fiscal space to increase public investments (total, not just infrastructure-related); First, the region’s infrastructure performance varies, both these have dropped to an average of about 3.4 percent across countries and sectors. The region invests little of GDP across the region. Historically, only about a third in infrastructure on average, but this average is driven of total public investment goes to infrastructure. So, at by some of its largest countries: Argentina, Brazil, and least in the short to medium term, it is highly unlikely that 7 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE With greater efficiency, four times as many water utilities could access private financing FIGURE ES 1:  (percent of utilities) 90 80 70 60 50 40 30 20 10 0 Current level Step 1 Step 2 Non- Step 3 With Step 4 Increase with cash cost Collection rate labor cost reduction of revenue by 10% recovery > increased to reduced by 15% current level of 120% 100% NRW to 25% Source: Courtesy of William Kingdom and Alexander Danilenko (World Bank) based on IBNET data 2016. Note: NRW stands for non-revenue water, or water that is produced but “lost” before it reaches the consumer. the region’s public investment in infrastructure could rise only relies on public spending and PPIs. As such, spend- much above 1.0-1.5 percent of GDP. ing better on the right things (which includes improving the creditworthiness of public utilities and the balance Could private investments be a way to close the service between user fees and tax-payer financing of infrastruc- gap? Investments captured through the World Bank’s ture) is not just the best way to make a significant dent private participation in infrastructure (PPI) database have in the investment gap. For many countries it will be the ranged from 0.5 to 1.2 percent of GDP per year since only way. 2006. But with about one-third of this financing coming from public sources, and about half of the deals requir- Thus, this report advocates a much more careful discus- ing public guarantees, PPI expansion is constrained by sion of investment needs in Latin America, one that starts limited public finance. In other words, while PPIs may with a debate about what infrastructure is needed given help improve performance, they do not leverage sig- countries’ development priorities (which typically also in- nificant amounts of private capital and are best seen cludes prudent fiscal policies), that thoroughly examines as a complement to, rather than a substitute for, public how to achieve infrastructure goals efficiently (which re- investments. quires looking at country, sector, and project cost drivers), and that relies on well-thought rules of the game for de- As to commercial borrowing by public utilities, it is limited ciding what should be financed by taxpayers rather than by poor creditworthiness, at least in the water sector: only users. 20 percent of the Latin American water utilities includ- ed in the International Benchmarking Network for Water As such, the key questions to ask are the following: and Sanitation Utilities (IBNET) database (2016) generate enough of a surplus to mobilize commercial borrowing ››What is the goal? (assumed to be cash revenues exceeding costs by at least 20 percent). This means that 80 percent of Latin America’s ››How to achieve it as cost-effectively as possible? water utilities would have difficulties in mobilizing com- mercial (non-government guaranteed) financing unless ››Who pays? they implement significant reforms. Another way of look- ing at this is that greater efficiency could bring in much The first two questions will determine investment needs, more financing (figure ES1). while the third will determine financing options. We ad- dress these three questions in what follows—hoping to So, realistically, Latin America is unlikely to see infrastruc- offer a useful framework for countries to devise infrastruc- ture investments rise much above 2.7 percent of GDP if it ture strategies. 8 What is the goal? And how to set it? Infrastructure (defined here as power, water, sanitation, So, how to think about Latin America’s infrastructure transport, flood protection, and the backbone of tele- goals? We suggest two sets of inputs that countries can communication) is necessary for growth, poverty alle- factor in as they go about defining the infrastructure viation, social inclusion, resilience, and environmental needed to support their economic, social, and environ- sustainability. mental objectives. Again, however, these are no more than inputs to help make the needed policy choices. But it is the service that matters more than the kilome- ters of roads or pipes and cables, along with quality and affordability. In fact, better or more infrastructure is of- Stocktaking along with peer comparison or ten only one of many ways to achieve a policy goal. Take benchmarking—this suggests that Latin America transport as an example. If the objective is to improve may want to focus on sanitation and transport mobility or access to services, it may be more effective to better regulate transport services or build more clinics In comparing Latin America to its peers, there are clear and schools rather than roads and highways. And infra- areas of strength and weakness. The region scores well structure needs change over time, as economies devel- in terms of access to water and electricity. Some 94 per- op, societies urbanize, the middle class grows, the climate cent of households have access to improved water. The evolves, and technologies drive transformation on many 20 million or so households that still lack access are con- fronts. centrated in just six countries, and all but one (Haiti) are comfortably middle income, suggesting that full cover- Investment gap approaches seldom take these consider- age is well within the region’s financial capacity. The sto- ations into account. Instead, the methodologies behind ry is similar for electricity: 96 percent of households have the investment gap estimates do one of two things: access. Access rates are progressing well, helped by sup- portive policies (again with the exception of Haiti). Water ››They examine how infrastructure access or stocks and electricity utilities could do better, especially in terms have evolved historically relative to income, urban- of efficiency and cost recovery, but there are many good ization, economic structure, and other factors, then performers. estimate what it would take to maintain this relation- ship. If in the past this access or stock was suboptimal, The electricity sector has the potential to turn into a great the projected ones will be as well. Importantly, these competitive advantage for the region. It is the cleanest estimates are in no way the result of an optimization of any region, based mostly on hydroelectricity. The in- exercise: they are not investment that will maximize creased uncertainty of water precipitation associated growth or poverty reduction, but just the investment with climate change means that the region will need to needed for business as usual to continue. diversify its renewable resources. It has significant poten- tial in solar and wind, and while these remain just a sliver ››They price a goal (such as universal access to water of overall generation, investments are increasing rapid- or electricity) using either rough estimates of costs ly—not surprising in light of the high score given to the (such as average cost of a water connection) or, in investment framework for renewables by World Bank best case scenarios, an economic-engineering mod- analysis. el (a least-cost optimization plan for the electricity network). This approach is useful where goals have The policy framework for energy efficiency remains want- been set (such as universal access to water or elec- ing, however, which results in poor electricity efficiency. tricity) and can help inform policy debates by pro- This is a priority area, along with improving the regional viding estimates of the costs of different goals (such connectivity of electricity systems. The region could po- as access to water or to safely managed water; in the tentially leapfrog into the kind of systems now emerging house or within 100 meters of the dwelling) or differ- in high-income countries, with decentralized electric- ent ways of achieving them. But it cannot substitute ity production and consumers as “prosumers.” Doing so for policy choices regarding the service gap or pub- would require a substantial evolution on the part of util- lic priorities—for example in transport, where there ities, but the sophistication of the region’s markets and is no simple “universal access” goal to aspire to. regulators makes it a distinct possibility. 9 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE More challenging are sanitation and modern cooking most struggle with high congestion, pockets of inaccessibil- fuels, both of which have serious implications for public ity, inefficient and often inequitable pricing, and continuing health and human capital accumulation. Some 17 per- reliance on informal public transport providers. cent of Latin Americans have no access to a private, im- proved sanitation facility and only about a third of waste- water is treated. The dismal wastewater performance is Emerging needs and challenges: climate change, a real emergency, and one that epitomizes the potential increased demand, and urbanization for spending better. The sector is hampered by overly ambitious “imported” regulations that are unrealistic and Infrastructure is typically long-lived and influences house- leave no room for gradual improvement. Worse, legisla- holds and firms’ own investments and locational choic- tion usually precludes resource recovery even though es—choices that are difficult to reverse, creating lock-ins. wastewater plants can be designed to generate electric- As such, infrastructure decisions need to be forward look- ity for their own use or sale, and grey water and treated ing. Two trends that matter for the region’s infrastructure sludge can be used for agriculture and other purposes. choices are climate change and the combination of ur- A case in point is Lima, a city of nearly 8.5 million in the banization and changing socioeconomics. middle of a desert, that discharges its used water into the ocean and disposes of its sludge in expensive sanitary Climate change means that energy, transport, and water landfills instead of allowing it to be used for agriculture. and sanitation systems will need to be built differently to become more resilient to extreme events and better able As for modern cooking fuels, some 87 million people to respond to associated changes in demand: more elec- still lack access. This issue has not received sufficient at- tricity during heat waves, better water storage to cope tention or financial support. More than half the people with both droughts and extreme rains, and protective of Guatemala, Honduras, and Nicaragua still use solid dams and improved drainage to reduce flood risk. fuel for cooking with serious public health implications. Even in relatively wealthy Mexico, the share is still 13 Climate change will also increase the need for better re- percent. source management—it is another factor that creates pressure for greater efficiency, especially for water and en- But the transport sector is where Latin America most un- ergy. Mexico, and to a lesser extent Brazil and Colombia, derperforms its peers. This is partly due to the region’s have policies in place to boost energy efficiency. But im- low population density, which makes it extremely hard plementation has been weak, and has occurred mostly in to affordably develop a dense transport network. Latin the context of domestic energy crises. Further, fossil fuel America’s paved road density is similar to that of Africa subsidies remain high in countries such as Argentina, Bo- (as is the perception of its road quality, according to the livia, Chile, Ecuador, and Venezuela. In general, the region World Economic Forum’s rankings). This may be a normal has significant potential for improving energy efficiency. consequence of the region’s geography and need not imply that its countries need more roads. The region does Climate change also means that pressures will mount have a very high road occupancy rate, however, as well as to reduce emissions from infrastructure. As mentioned large pockets of inaccessibility. above, this could become a source of comparative ad- vantage for Latin America, which has the cleanest elec- Physical infrastructure is only part of the transportation tricity matrix in the world. But increasingly variable pre- challenge. A lack of competition in trucking and inefficient cipitation and strong popular pushback against dams are customs clearance processes are largely responsible for rel- making it increasingly urgent for the region to diversify its atively low logistics performance. The road transport indus- renewable energy sources. Transport emissions are grow- try is some 15 times more concentrated than in the United ing rapidly, and without action, emissions from infrastruc- States. In Central America, increased competition on nation- ture-related sectors are likely to increase further. A busi- al routes could reduce prices by 8 cents per ton-km as op- ness-as-usual scenario, with increased motorization and posed to only 3 cents for tackling inefficiencies such as high decreased reliance on hydroelectricity, would see ener- congestion, long waits at borders, and bribery. The region gy-related emissions more than double across the region has limited integration among different transport modes, between 2010 and 2050. especially rail and road. Ports suffer from highly congested access roads. In urban transport, a number of cities have Changing socioeconomics are also affecting the demand modern, well-functioning bus rapid transit systems, but for infrastructure services. The region’s middle class grew 10 by about 50 percent during the boom years of 2003-09. in Santiago, Montevideo, La Paz, Buenos Aires, and Brasil- The vast majority of the middle class has access to basic ia, driven by transport, land use, and housing policies. If services, but is far from saturated in terms of consumer expansion patterns continue unchanged, built-up urban durables (notably cars, air conditioners, and washing ma- areas could double in the region by 2035, pushing up chines). The combination of higher incomes with a recent infrastructure costs. A recent World Bank review of ur- boom in consumer credit could significantly raise overall banization in Mexico suggests that denser urbanization energy demand, in a way that may not be sufficiently tak- would reduce infrastructure investments and mainte- en into account in traditional energy forecasts. At the same nance costs by 41 percent in Merida and 67 percent in time, most remaining challenges in basic access to water Los Cabos. and electricity are now concentrated among the poorest decile, which may be the social group most difficult to Another critical question is whether the region can im- reach, either because of remoteness or depth of poverty. prove how urbanization is planned and managed. Un- The last mile challenge may well be more complex than planned urban expansion translates into higher costs as previous ones and require innovation in technology, deliv- slum upgrading costs two to eight times more than reg- ery, and financing. ular land development. Already some 25 percent of the region’s urban dwellers live in slums, and they account Linked with changing socioeconomics is the region’s on- for a large claim on the region’s infrastructure investment going but changing urbanization. Density has declined budget. How to improve services as cost effectively as possible? Spending better could materially reduce the cost of im- the region has been investing in recent years, with good proving infrastructure in Latin America. Take the case results on water but less so on sanitation. of electricity: availability can be increased by building more power plants or by improving energy efficiency. A simple framework is useful to think through the vari- And despite having a rather sophisticated and mature ous elements of reform needed to improve infrastructure electricity market, the region still fares poorly on en- spending efficiency. The ultimate objective is to deliver ergy efficiency. In fact, transmission and distribution infrastructure services to households and firms in a way losses are some of the highest in the world. Not sur- that maximizes some measure of welfare. Hence it will prisingly, electricity investment needs would average depend: $23-24 billion a year if South America follows the same investment path it has in the past, but perhaps as little ››on the supply side, on the cost of producing such ser- as $8-9 billion if it adopts a transformational approach vices, from the initial construction of supporting net- that favors demand-side management, energy efficien- works to the end delivery of services to users, and cy, and renewable energy solutions. (This number only includes investments and not the cost of demand-side ››on the demand side, on the availability and afford- management and energy efficiency programs. The to- ability of services, given the quality produced and the tal would be higher, but still likely to be substantially pricing schemes for users. less than under a traditional path.) (Broad and others 2016). On the supply side, key whole-of-government challeng- es have been identified as particularly problematic for Similarly, for water, the World Bank estimates that the Sus- public investment management. These include: tainable Development Goal of universal access to safely managed water and sanitation would cost between 0.1 ››Weak planning, project appraisal, and preparation and 0.4 percent of GDP a year through 2030, depend- capacity. Work by the World Bank, Inter-American ing on how it is implemented. A reasonable way forward Development Bank, and IMF shows that Latin Amer- would cost about 0.25 percent of GDP—roughly what ica has a low level of investment efficiency, with the 11 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE weakest stages of public investment management disbursement data show a chronic gap between being appraisal and evaluation. Many projects are committed and executed funds. funded that are not sufficiently prepared—either be- cause of limited capacity or oversight, or because of ››Procurement that could be improved. Inefficient parliamentary amendments to annual budget laws. procurement processes contribute to limited bud- This is a serious issue in Brazil in particular, where a get execution and excess costs. Some countries in combination of weak capacity, lax enforcement, and the region such as Chile and Mexico have seen re- pork-barrel politics results in investment projects be- markable success in driving procurement reforms. For ing included in ministerial budgets without having example, the ChileCompra electronic portal is esti- been subject to formal appraisal. mated to have generated US$280 million in savings. Mexico’s modernization of its tendering processes ››Overly rigid or myopic budgeting, which is designed generated savings of more than US$1 billion within to manage fiscal deficit and focuses on controlling three years of its 2009 start. But a detailed analysis of cash expenditures rather than promoting efficient procurement performance across the region shows spending. As a result, most countries have annual there is considerable room for improvement. The budgets that do not allow for carry-over. This results World Bank’s Benchmarking Public Procurement 2017 in rushed procurement and execution and is further report found that suppliers identify obstacles such as complicated by the fact that the fiscal year runs from excessive bureaucracy and red tape in Colombia and January to December when the right time for public Honduras; payment delays in Argentina, Jamaica, and works in most countries is the dry season of Novem- the Dominican Republic; lack of transparency and ber to March. Colombia appears to be one of the few opaque tendering processes in Brazil and Mexico; and countries in the region that has a strong planning sys- inefficiency in Barbados. tem and has introduced medium-term expenditures frameworks. Other countries have created fiduciary ››Unclear project sustainability, due to an imbalance be- funds to avoid this budget rigidity, choose public-pri- tween capital and current spending on infrastructure. vate partnerships (PPPs), or continue to cope with This can spring from overly rigid budgets and suboptimal multiannual allocations. planning—very few countries in the region link invest- ment budgets and forward expenditure estimates—or ››Difficulties with budget execution. Many countries from poor coordination between central governments also struggle with the execution of capital budgets (which fund and often manage capital investments) and (figure ES 2). This is a particular issue in Brazil, where local governments (which may lack the financial and Many Latin American countries chronically under-execute their capital investment budget FIGURE ES 2:  (disbursement as percent of commitment) 100 80 60 40 20 0 Bolivia Brazil Chile Colombia Honduras Peru 2013 2014 2015 Source: Courtesy of Diego Dorado, based on data from government budgetary reporting systems 12 technical capacity to take over these investments or cov- in the industries related to transport services. The latter er operation and maintenance costs). may allow for substantial gains to be redirected from profits in these industries to consumers of infrastruc- ››An uncompetitive construction industry. Competi- ture. For example, monopolistic freight transport ser- tion policy, to the extent that it affects competition in vices may well end up capturing most of the potential construction and related activities, is likely to have an rent created by additional physical infrastructure, re- important impact on construction costs. ducing demand and potentially nullifying the gain from the investments. Next comes a post-construction stage that relates to operational efficiency in service delivery and depends Finally, on the demand side, a number of price-relat- on the ownership structure, the quality of the regulato- ed aspects are key to driving efficiency. The first is sim- ry framework, and corporate governance. A review of ply adequate pricing, which encourages efficient use by utility performance in the region found that on average, consumers (to buy more fuel-efficient cars, use public private utilities outperform public ones, although there transportation, turn off lights, buy energy-efficient appli- are good and bad performers in both groups (the top 10 ances, fix leaking faucets, and so on). Pricing services ap- percent of public utilities outperform the average private propriately also makes it possible to attract commercial utility). Independent regulatory agencies that are trans- financing, which in turn may create additional pressures parent, accountable, and free of political interference are for efficiency. critical to improving the performance of both public and private operators—helping raise labor productivity and But adequate pricing is not simply cost-recovery tariffs. cost-recovery ratios while reducing operational expens- It needs to factor in social acceptability—which requires es and distribution losses. For state-owned enterprises, that price regulation go hand in hand with regulation of performance further depends on the existence of a cor- quality and with considerations of availability and afford- porate structure that prevents political intervention, re- ability. This last point is where equity concerns and issues wards performance, and is subject to public scrutiny. Best of externalities should be included, to ensure that policy practice corporatized frameworks—which include an in- makers consider the added social value of services (for dependent performance-driven board of directors, a pro- example in water, sanitation, or public transport) when fessional staff, transparency and clear disclosure policies, defining what should be funded by taxpayers rather than and a clear mechanism for evaluating performance—are users. associated with high levels of performance, with perfor- mance orientation and professional management being But while pricing is important, other mechanisms (incen- the most important contributors. tives, information, and nudges such as options that de- fault to more efficient settings, quotas, or performance Similarly, for transport (the one non-utility sector we standards) are available to push consumers to switch to examine) substantial efficiency gains are likely to come more efficient patterns of consumption and applianc- from regulatory frameworks that encourage both es. These can be helpful complements, especially where greater intermodal coordination and more competition price elasticities are low. Who should pay—and what does it imply in terms of financing options? Spending efficiently on the right things also means mak- treatment or public transportation, or due to equity con- ing judicious use of scarce public resources. Infrastructure cerns as in subsidies targeting the poor). is funded by either taxpayers or users. Taxpayer money is best used where it is not possible to charge the users (as But the potential for cost recovery cannot be divorced with flood protection or rural roads) or not desirable to from the efficiency with which a service is provided. Poor do so (due to environmental and social externalities that quality service will reduce willingness to pay, while high result in under-consumption of the service, as with water costs will reduce the likelihood of achieving full cost 13 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  decision-making framework to ensure the judicious use of scarce public and concessional FIGURE ES 3: A finance 1 Commercial Financing Can commercial nancing be cost-e ectively mobilized for sustainable investment? if not... 2 Upstream Reforms & Market • Country and Sector Policies Can upstream reforms be put in place to • Regulations and Pricing address market failures? If not... • Institutions and Capacity 3 Public and Concessional Resources for Risk instruments and Credit Enhancements • Guarantees Can risk instruments & credit enhancements • First Loss cost-e ectively cover remaining risk? If not... 4 Public and Concessional Financing, including Sub-Sovereign • Public nance (incl. national development: banks and domestic SWF) Can development objectives be • MDBs and DFIs resolved with scarce public nancing? Source: World Bank 2017. recovery, especially if there is a perception of inefficiency In this context, the World Bank Group, along with oth- or predatory pricing on the part of the service provider. er multilateral development banks, is suggesting an ap- And even where subsidies may be justified by external- proach that weighs the benefits and opportunity costs ities, as with wastewater treatment for example, it may of deploying public and/or concessional resources (fig- be possible to reduce the needed subsidies by reducing ure ES 3). The starting point of this approach is that any costs. Most countries find it difficult to achieve full cost investment project or program that can be financed on recovery for wastewater treatment plants using tradition- commercial terms while remaining affordable and offer- al business models, under which a water treatment plant ing value for money, should be. is a cost-center. Instead, new models are being proposed whereby water treatment plants can generate electricity Where commercial financing is not cost effective or vi- for self-consumption and even sale to the network and able due to perceived risks or market failures, efforts sell sanitized sludge for use as a fertilizer. should focus on addressing these market failures through upstream reforms to strengthen country and sector poli- Charging users, where possible and appropriate, has a cies, regulations, and institutions or with targeted public number of advantages. First, it helps manage demand. interventions (e.g., targeted subsidies or complementa- Second, it creates a market test and puts pressure on ry public investments, such as transmission lines). Where the service provider (public or private) to improve qual- risks remain high and raise the cost of commercial capi- ity. Indeed, there is some evidence that the efficiency tal beyond that afforded by project or corporate revenue gains typically associated with private participation in in- generation, it may be possible to reduce costs through frastructure are dependent on such a market test. (This risk-sharing instruments backed by public or concession- raises some concerns, given that only about 43 percent al finance. Only where commercial financing is still not of the region’s PPPs are backed by user fees or a purchas- viable or cost-effective should public and concessional ing agreement with fully private utilities.) Third, charging resources be deployed. users directly increases the revenue base for investments and creates the potential for commercial financing, re- Importantly, this framework can only be applied to ser- gardless of who operates or owns the infrastructure. And vices that can be charged to users, as user fees are what commercial financing in turn is likely to create pressure creates the basis for commercial financing options be- for greater efficiency. yond the use of general taxes.1 1 In addition, there may be potential to capture the value created through infrastructure investments in less traditional ways (land-value capture, congestion charging, parking fees) or through the commercial exploitation of infrastructure assets (advertising, real estate). 14 Equity and poverty concerns are not at odds with reliance ››Relatively high water access, though quality and safe- on commercial financing, even though they are often in- ty remain poor, with sewerage access low and less voked in arguments against full cost recovery for basic than 30 percent of wastewater being treated—an un- services such as water and sanitation, electricity, public acceptable level for a region with its levels of income transport, and modern cooking fuels. The needs of the and urbanization. poor are in fact typically best served by a combination of cost recovery tariffs and targeted subsidies and payments ››Mediocre transport services due to poor infrastruc- schemes adapted to their needs. Most of the wealthier ture and an uncompetitive transport industry, result- Latin American countries have well-developed social reg- ing in costly freight transport, congested cities, and istries and safety nets (e.g., Brazil, Colombia, Chile, Mexi- deep pockets of rural isolation. co), but targeting is likely to be a challenge for countries without these in place. Improving the region’s infrastructure performance in a con- text of tight fiscal space will require spending better on well identified priorities. Unlike most infrastructure diagnostics, In sum, the present report insists that much of what is needed lies outside the infrastructure sector and has to do with broad- Latin America spends a good deal of money on infra- er government issues—from competition policy to bud- structure. In return, it gets: geting rules that no longer solely focus on controlling cash expenditures. But quite a lot also involves sector reforms, ››High electricity access—with good prospects for with the traditional recommendations regarding indepen- closing the access gap given that the remaining un- dent, well-performing regulators and better corporate gov- connected households are concentrated in mostly ernance. We also insist on the critical importance of cost middle-income countries—but low nonsolid fuel ac- recovery where feasible and desirable, since user fees are cess, with serious health implications. the basis for commercial finance—while keeping in mind the importance of reducing costs, either through efficien- ››The world’s cleanest electricity—mostly from hydro- cy or adoption of alternative business models, such as are electricity which is challenged by increasingly fre- emerging for water treatment plants. quent droughts—and small but rapidly growing wind and solar sectors. Latin America has long been an innovator in infrastruc- ture. The report notes many challenges, but it also high- ››Some world-class utilities for both water and electrici- lights many examples of the region’s capacity for innova- ty, and a few countries with sophisticated, stable, and tive solutions, its expertise with sophisticated regulations, predictable regulations, especially for electricity. But and its experience with PPPs. Latin America has the most utilities and regulatory schemes could do better, means and potential to do better. And it can do so by with potentially significant cost and resource savings. spending more efficiently on the right things. 15 Part I. Infrastructure in Latin America and the Caribbean: Modest spending, uneven results I nfrastructure serves a number of purposes. It provides the on infrastructure. This belief comes from general dissat- underlying conditions for increased development, be it isfaction with infrastructure services in the region, and by providing more efficient transport, the last-mile water the widely held idea that increased spending on infra- supply for the poor, or energy for all. It can serve to reduce structure is key to improving the region’s growth and inequalities (by serving slum dwellers rather than favoring competitiveness. better off residents of cities), it can open access to markets to increase prosperity (transport infrastructure), and it is recog- But would increased spending solve Latin America’s in- nized as a core input into any sustainable growth strategy. frastructure challenges? To start answering that question, this section examines how much Latin America currently So what should Latin America and the Caribbean be spend- spends on infrastructure—from both public and private ing on infrastructure?2 Estimates vary widely—from 3 to 8 sources—and what it gets for its money. percent of regional GDP—though they tend to hover around 4 to 5 percent. Most of these estimates use a simplistic meth- What we find, on average, is modest spending and un- odology that measures only how infrastructure spending even results across sectors and countries. The region has would need to evolve as it has historically, along with eco- made impressive progress in some areas, such as extend- nomic growth (box 1). They do not indicate what Latin Amer- ing access to water and developing modern public trans- ica should be spending to maximize growth (though some port systems in many large cities. Some water and energy studies look at the cost of achieving a specific social, eco- utilities are world class. But too many Latin Americans still nomic, or environmental goal). Such estimates are the results suffer from poor or nonexistent services. And the region of benchmarking exercises, not optimization exercises. is far from having the infrastructure it needs to compete internationally and provide the middle-class standard of Yet there is widespread agreement in policy circles and living that many of its people aspire to, and that a mostly in the press that Latin America needs to spend more upper-middle-income region should be able to deliver. 2 Henceforth we use Latin America as short hand for Latin America and the Caribbean. 17 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE How much does Latin America spend on infrastructure? During 2008-13 (the most recent years for which data This regional average masks large differences across are available) average annual investment in infrastruc- countries, from 1.6 percent of GDP in Mexico to 6 per- ture in Latin American ranged from 2.4 to 3.2 percent of cent in Nicaragua (figure 2). Indeed, the average is largely GDP, for an average of 2.7 percent (figure 1). This is far determined by Argentina, Brazil, and Mexico, which to- from the upper-bound “needs” estimates of 8 percent, gether account for four-fifths of the region’s GDP.3 Many or even from the 4-5 percent average suggested in ta- other countries—such as Bolivia, Costa Rica, Panama, and ble B1, but in line with the lower-bound estimate of 3 Peru—invested more than 4 percent. No pattern emerges percent. about country size or income level. For example, smaller,  ow should Latin America define its needs for infrastructure investment? Box 1: H Fay (2000) and Fay and Yepes (2003) provided the first estimates of infrastructure investment needs in developing regions. Their approach was essentially a benchmarking one, using simple econometrics to examine how infrastructure spending has evolved over time along with income, population, urbanization, GDP composition, and other relevant determinants of house- hold and firm demand for infrastructure services. It then priced the investment needed to keep a similar relation as income, population, and urbanization increase over time. The main limitation of this approach is that there is no optimization. It simply prices what it would take to maintain the histori- cally observed (possibly constrained) relation between infrastructure and income, population, and urbanization. The approach does not establish a causal relationship between investments and growth, nor does it price a specific desired objective (such as the Millennium or Sustainable Development Goals). It is also very sensitive to assumptions about future growth rates. Subsequent authors (Perrotti and Sanchez 2011; Kohli and Basil 2011; Ruiz-Nunez and Wei 2015) have introduced various im- provements and updates to this original model, with estimates for Latin America ranging from 3.6 to 6 percent of the region’s GDP (box table 1). To our knowledge, however, none has been able to answer the question of optimality (the amount of invest- ment needed to achieve a particular growth path). CAF (2011), Powell (2013), and Serebrisky and others (2015) mostly draw from this body of work to conclude that the region’s infrastructure investment needs are 4-6 percent of GDP. Climate change considerations have also affected estimates and methodologies. Bhattacharya, Romani, and Stern (2012) esti- mate that Latin America would need an additional $200-300 billion a year to achieve mitigation and resilience objectives. But while reasonable methodologies exist to estimate the cost of low-carbon investment pathways, estimating adaptation invest- ment needs at the level of a country—let alone a region—is pure guesswork. Estimated BOX TABLE 1:   annual spending requirements for infrastructure in Latin America vary considerably Authors Estimates (percentage of GDP) Period of prediction Maintenance included? (percentage of GDP if provided) Fay and Morrison (2007) (a) 4-6 2010-30 Yes (1.0) Fay and Morrison (2007) (b) 3 2010-20 Yes (1.0) Perrotti and Sanchez (2011) 5.2 2006-20 Yes (2.5) Kohli and Basil (2011) 4 2011-40 Yes Ruiz-Nunez and Wei (2015) 3.6 2014-20 Yes (1.8) CAF (2011) 5-6 2010-40 Yes Bhattacharya, Romani, and Stern (2012) 6-8 2012-20 No a. Level of investment needed to bring Latin America’s infrastructure to a level equivalent to that of the Republic of Korea b. Level of investment needed given expected growth and historical co-evolution of infrastructure, GPD, and economic structure, plus cost of achieving universal coverage in water, sanitation, and electricity. 3 Brazil represents about 44 percent of the region’s GDP, Mexico about 25 percent, and Argentina 10 percent. 18  ublic and private infrastructure investments in Latin America have been fairly stable, 2008-13 FIGURE 1: P (percent of GDP) 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 2008 2009 2010 2011 2012 2013 Public With private participation Source: www.infralatam.info, downloaded on August 4, 2016. Note: Includes investments in electricity, telecommunications, transport, and water and sanitation. Data weighted by country GDP. Data on investments with private participation come from the PPI database (www.ppi.worldbank.org), represent committed rather than actual investments, and are reported in full for the year they are committed rather than when investments are disbursed. nfrastructure investment levels varied enormously across countries in 2008-13 FIGURE 2: I (percent of GDP) 7 6 5 4 3 2 1 0 zil o or a y ala ge ia e ay ica ia ru ras a a ua n il am gu xic mb liv ad Pe Bra gu Ch ra nti aR du m g ara Me Bo n ave alv Uru ra ate lo e n Pa st Arg Ho Pa Nic Co El S Co Gu al n gio Re Public With private participation Source: www.infralatam.info, downloaded on May 2, 2016. Note: See figure 1. poorer Central American countries are evenly split be- are recorded in full the year a deal is signed, rather than tween those that invested a little (El Salvador, Guatemala) spread over time as the investments take place.) The and those that invested a lot (Honduras, Nicaragua). top four destinations in terms of share of GDP include some of the region’s largest and richest countries (Bra- Public-private partnerships (PPPs) accounted for some zil, Peru) and some of its smallest and poorest (Honduras, 40 percent of the region’s infrastructure investments— Nicaragua). though with significant variation, from 4 percent in Boliv- ia to 58 percent in Honduras, again with no clear pattern During the first half of this decade PPPs in Latin Ameri- based on country size or income. (A note of caution is in ca gyrated wildly, mostly due to fluctuating investments order in interpreting these numbers as these are com- in Brazil (figure 3). Elsewhere in the region, PPPs have mitments rather than actual investments. Moreover, they grown steadily, from around 0.2 percent of GDP to 0.8 19 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE nfrastructure public-private partnerships commitments have fluctuated wildly in Latin America, FIGURE 3: I 1990-2015 80 1.8 70 1.6 1.4 Total Investment (percent of GDP) Total Investment (US$ billion) 60 1.2 50 1.0 40 0.8 30 0.6 20 0.4 10 0.2 0 0.0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Total investment (US$ billion) Total investment (percent of GDP) Source: www.ppi.worldbank.org, downloaded on April 15, 2016. Note: Data represent committed rather than actual investments, and are reported in full for the year they are committed rather than when investments are disbursed. Equity accounts for a small share of public partnership finance in Latin America, 2000-15 FIGURE 4:  (US$ billions) 35 30 25 20 15 10 5 0 2011 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2012 2013 2014 2015 Bonds Equity International nancial institution support Loans Source: Project Finance and Infrastructure Journal, www.ijglobal.com, downloaded on May 5, 2016. percent. Over the past 10 years most PPPs in Latin Amer- accounted for less than a third of PPP finance in Latin ica have been greenfield investments, mainly in the en- America (figure 4). Loans have almost always been the ergy sector. most common instrument used to finance infrastructure projects in Latin America—primarily from commercial But even if the private sector participates in a project, banks, but about a third from public sources such as de- it does not necessarily directly invest its own money velopment banks, state and national banks, export credit (equity) in the project. Indeed, during 2011-15 equity agencies, and other public authorities and companies.4 4 Authors’ calculations based on data from Ruiz-Nunez (2016). 20 Bonds remain a small share (12 percent on average over guarantees, while transport received the highest share of the last five years) even though the region leads emerg- direct government support. ing economies in the use of bonds for infrastructure financing. Government support typically declines with the maturity of the PPP market, but it does not disappear: in 2000-14 More generally, direct and indirect government support nearly 75 percent of PPPs in countries with limited expe- plays a critical role in facilitating private investment in rience with such projects received guarantees, compared infrastructure.5 From 2010 to 2014 about half of all PPP with 45 percent of PPPs in countries with stronger track deals in Latin America received some form of govern- records. These direct and indirect guarantees create con- ment support through direct or indirect contributions, tingent liabilities for the public sector even if there is still with payment guarantees being the most common type debate on how they should be accounted for in public of support. The energy sector got the highest share of accounts. What is the region getting for its money? Latin America spends a fair bit of money on infrastruc- most utilities and countries could do better, with po- ture. In exchange, it gets: tentially significant cost and resource savings. ››Mediocre transport services due to low-quality infra- structure, and an uncompetitive transport industry, Transport: Unimpressive outcomes, but resulting in costly freight transport, congested cities, infrastructure is just part of the challenge and deep pockets of rural isolation. Good transport services are a major challenge for Lat- ››Relatively high water access, though quality and safe- in America for several reasons. First, the region’s overall ty remain poor given that sewerage coverage is low, population density is low, making it difficult to design a and less than 30 percent of wastewater is treated—an dense, affordable network. Second, the region’s economy unacceptable level for a region with its levels of in- is dependent on trade—the recent boom was driven by come and urbanization. commodity exports—making its competitiveness highly sensitive to the performance of its transport sector. Third, ››High electricity access, with the remaining uncon- the region’s high urbanization and relatively high income nected concentrated in a few countries (rich and call for fairly sophisticated urban transport systems. poor) and good prospects for closing the access gap, but low access to nonsolid fuel—with serious health Rising awareness of the implications of poor transport implications. infrastructure likely explains the substantial increase in transport investments between 2002 and 2013. These ››The world’s cleanest electricity, mostly from hydro- started to pick up in 2004 and grew significantly through electricity (though threatened by increasingly fre- 2009, hovering around 1.0-1.25 percent of GDP since quent droughts), and small but rapidly growing solar 2007, more than other infrastructure sectors (figure 5a). and wind sectors. But trends varied by country. Most larger, richer coun- tries (Argentina, Brazil, Mexico) spent around 1 percent of ››Some world-class utilities for both water and electrici- GDP in 2008-13, while others (Bolivia, Nicaragua, Panama, ty, and a few countries with sophisticated, stable, and Peru) spent two to three times as much. Public invest- predictable regulations, especially for electricity. But ments dominated the period, though a recent increase in 5 Governments can offer indirect support through guarantees to reduce specific project risks—for example, payment, revenue, and exchange rate guarantees. They can also provide direct support to deals from their budgets in case user fees or power or water purchase agreements with private entities or wholesale markets are not possible or sufficient (in some cases government may collect user fees but provide availability payments to the private entities bearing the demand risk). 21 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  owns and ups in transport investments in Latin America, 2000-13 FIGURE 5: D (percent of GDP) a) After a steady increase, transport spending leveled o b) An increase in private investments o set the recent decline in public spending 3.0 1.2 2.5 1.0 2.0 0.8 1.5 0.6 1.0 0.4 0.5 0.2 0.0 0.0 11 11 00 01 02 03 04 05 06 07 08 09 10 12 13 00 01 02 03 04 05 06 07 08 09 10 12 13 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Transport Total Private Public Source: www.infralatam.info, downloaded on November 15, 2016.  oad density in Latin America is lower TABLE 1: R private investment helped offset a drop in public spend- than almost anywhere else, 2010 ing (figure 5b). 2010 Density Level (paved lane-km per km2 land) OECD Europe 2.1 Weak infrastructure and an uncompetitive industry India 1.3 result in poor performance China 0.7 Despite greater investments, transport performance re- OECD North America 0.5 mains poor. Latin America has low transport infrastruc- OECD Pacific 0.4 ture density given its income level, with paved road Japan 5.5 density similar to Africa’s and about one-quarter that of ASEAN 0.4 the next lowest region (table 1). This may however be Middle East 0.2 a normal consequence of the region’s geography, and need not necessarily imply that the region needs more Latin America 0.05 roads. Nevertheless, Latin America’s roadway occupan- Africa 0.04 cy rate is higher than that of any other region (figure Source: Dulac 2013. 6), and the region is characterized by a combination of large pockets of inaccessibility and congestion chal- lenges, particularly in large cities. For rail infrastructure, Central Asia over the last 10 years, but in Latin Ameri- density is less than 5 kilometers per thousand square ca roads and ports improved only marginally, and rail- kilometers for countries with a rail network, compared roads and airports did not improve at all (figure 7). In with 16 for OECD countries. More relevant is the fact 2016 international investors ranked the region’s road that service is limited and not an effective substitute infrastructure at the same level as that of Sub-Saharan for, or even complement to, road transport. Only Brazil Africa. Argentina, Brazil, Colombia, Costa Rica, Paraguay, and Mexico carry more than 20 percent of freight by rail Peru, and Uruguay have particularly low rankings given (CAF 2013). their income levels. Moreover, the region’s infrastructure is perceived as Latin American airports, historically superior to those being of rather low quality. Road, rail, port, and air in- of other emerging economies, have fallen behind (see frastructure improved steadily in Eastern Europe and figure 7d). Passenger demand has risen in recent years, 22  atin America has the world’s highest road occupancy levels, 2000‐10 FIGURE 6: L (vehicle-km to paved lane-km, thousand) 1,000 800 600 400 200 0 OECD N. America OECD Europe OECD Paci c China India ASEAN Latin America Africa Source: Dulac 2013. nternational investors are not impressed with Latin America’s transport infrastructure, 2006-16 FIGURE 7: I (Logistics Performance Index, 1=worst, 7=best) a) Roads: Latin America has underperformed its peers b) Ports: Latin America has improved, but less than its peers 7 7 6 6 5 5 4 4 3 3 2 2 1 1 0 0 -11 -12 -07 -08 -09 -10 -13 -14 -15 -16 -11 -12 -07 -08 -09 -10 -13 -14 -15 -16 10 11 06 07 08 09 12 13 14 15 10 11 06 07 08 09 12 13 14 15 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 c) Railroads: Latin America has increasingly fallen behind d) Air transport: Latin America has worsened while its peers have improved 7 7 6 6 5 5 4 4 3 3 2 2 1 1 0 0 -11 -12 -07 -08 -09 -10 -13 -14 -15 -16 1 2 7 8 9 0 3 -14 -15 -16 0-1 1-1 6-0 7-0 8-0 9-1 2-1 10 11 06 07 08 09 12 13 14 15 13 14 15 1 1 0 0 0 0 1 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 OECD high income countries East Asia Paci c Europe and Central Asia Latin America and the Caribbean Source: World Economic Forum 2015. Note: Country scores weighted by GDP (in constant 2010 US$) to calculate regional aggregates. High-income countries excluded from regional aggregates 23 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  atin America’s 2014 Logistics Performance Index (LPI) was dragged down by poor FIGURE 8: L infrastructure quality and slow customs procedures 4.0 +0.6 Component e ect on overall LPI for LAC +0.4 3.5 +0.2 +0.0 LPI (1-5, 5=best) -0.2 3.0 -0.4 -0.6 Ability to track and trace consignments Competence and quality of logistics services Ease of arranging competitively priced shipments E ciency of customs clearance processes Frequency with which shipments reach consignee within scheduled or expected time Quality of trade and transport related infrastructure 2.5 2.0 OCED high income countries East Asia Paci c South Asia region Latin America and Caribbean Europe and Central Asia Sub Saharan Africa Middle East and North Africa Source: http://data.worldbank.org/products/wdi Note: Regional averages weighted by current GDP. High-income excluded from regional aggregates everywhere except OECD. powered by the continued emergence of a Latin Amer-  atin America’s average export costs and TABLE 2: L ican middle class—a trend expected to continue. In times, while better than some, are higher than in freight, Latin America/Europe and Latin America/North competing regions America routes have been growing steadily and are ex- Region Cost to export Time to export pected to continue to grow by about 5 percent a year (US$) (hours) over the next decade (Boeing World Air Cargo Forecast Eastern Europe and Central Asia 1,133 110 Team 2015). Turnaround costs for planes rose 34 percent Middle East and North Africa 938 253 in real terms over the period 1995-2009, more than in Sub-Saharan Africa 914 257 most North American and European airports (Serebrisky Latin America and Caribbean 785 81 2011). East Asia and Pacific 573 56 Ports appear to have improved somewhat over the past South Asia 542 155 decade—as measured both by liner shipping connectiv- OECD High income countries 302 20 ity (an index that captures how well countries are con- Source: World Bank 2016b. nected to global shipping networks) and by internation- Note: Regional averages weighted by current GDP. High-income countries excluded everywhere except OECD. al investors’ impressions—but by less than the region’s peers (figure 7b). A particular challenge for ports is access infrastructure: surface infrastructure such as access roads would need to increase by some 15 percent within 50 return journeys—a sure way to increase costs and con- kilometers of ports and key transport centers by 2030 giv- gestion (IDB 2015). In Central America improved cost en current and predicted trade volumes (ITF 2016). efficiencies—offsetting or reducing the effects of con- gestion, long wait times at borders, and high informal The effects of generally poor infrastructure are com- payments—could cut costs by 3 cents per ton-kilometer. pounded by an uncompetitive road transport industry But simply having more competition on national routes that is some 15 times more concentrated than that of would cut prices by 8 cents per ton-kilometer. For the the United States (OECD/ECLAC 2012). Data for Argenti- region as a whole, allowing foreign companies to serve na, Brazil, Colombia, Costa Rica, El Salvador, Guatemala, national routes rather than limiting them to international Mexico, and Paraguay suggest that 44 percent of road point-to-point trips could help bring freight costs in line hauls in these countries are made without fee-earning with its peers (Osborne and others 2014). 24  ome pockets of low road access overlap with environmentally protected areas MAP 1: S (percent of rural population living within 2 kilometers of an all-weather road) Source: Authors’ estimates. Latin America ranks poorly on the World Bank’s Logistics productivity, and access to basic services. In Peru a sur- Performance Index (LPI), closer to Sub-Saharan Africa vey of 176 rural districts in the poorest and previously than to East Asia (figure 8). Disaggregating the region’s isolated areas in Andean provinces found that improved overall LPI shows that this lackluster performance is due rural accessibility significantly increased agricultural wag- to inefficient customs clearance processes and shoddy es, as well as land and housing prices, and boosted the trade and transport infrastructure. Country disparities frequency of health consultations by 70 percent (World are pronounced. Chile was the region’s best performer Bank 2015). in 2014, on par with the overall score for Eastern Europe and Central Asia. But Bolivia, Guatemala, Jamaica, and The low density of both populations and transport infra- Venezuela were at or below the Sub-Saharan average. structure in Latin America makes it a challenge to increase rural accessibility. Further, data on rural connectivity are The World Bank’s Doing Business data offer a more pos- scarce, with few countries having geo-referenced data- itive diagnostic, showing Latin America to have high- sets on roads that would enable the calculation of some er costs and longer times to export than East Asia (and type of rural accessibility measure (let alone some kind of higher costs than South Asia), but lower costs and shorter rural investment optimization). A rural accessibility index times than other developing regions (table 2). estimating the share of the rural population living within 2 kilometers of an all-weather road is available for only a handful of countries in Central and South America. The Rural areas are challenged by low transport and index indicates that poor rural access to roads is concen- population densities trated in the Amazonian basin. But there are noticeable pockets of poor access in coastal regions of Colombia Connecting rural communities to the “outside world” is and Ecuador, and access is mixed in Honduras and Nic- essential for inclusive development. Good rural access aragua as well (Map 1). The need for rural access should can raise household welfare, asset ownership, agricultural be balanced with environmental protection, given the 25 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  wnership of two-and four-wheel vehicles spiked between 1990 and 2010 FIGURE 9: O (per 1,000 people) 250 250 200 200 Four wheeler motorization rate Two wheeler motorization rate 150 150 100 100 50 50 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Brazil Chile Colombia Brazil Chile Colombia Mexico Uruguay Mexico Uruguay Source: ITF 2015. well-documented impact of roads on deforestation (Ali limited data are available on how Latin American cities and others 2015).6 fare relative to others—data coverage only includes large Brazilian cities and Mexico City. They suggest that while congestion in the region results in long travel times for Urban areas suffer from congestion and lack of cars, it is not systematically worse than in cities elsewhere accessibility despite the expansion of rapid bus of similar size (figure 10). That said, striking differences between cities of similar size in the region (such as Mex- systems ico City and São Paulo) prove that policymakers have the Urban areas, where 80 percent of Latin Americans live, power to effect change. have seen a sharp increase in the demand for and com- plexity of urban mobility. This is due to both population The past decade has seen significant investments in and income increases. Over the past 10 or so years the large-scale public transport systems in key Latin Ameri- number of people living in cities with more than 300,000 can cities where population densities justify the invest- people—and so in need of reasonably sophisticated ments. Many urban centers now have multiple public transport systems—grew by 28 percent (75 million). At transport modes. Bus rapid transport (BRT) systems have the same time, income growth and a sharp expansion of grown rapidly in the region and now serve more than 20 the middle class have caused the number of cars and mo- million passengers a day. Among Latin American cities torcycles to increase rapidly (figure 9 ): for example, from with more than 1 million inhabitants, 59 percent have about 6-7 million in 1990 in Brazil or Mexico to some 21- BRT systems, 31 percent have urban rail systems, and 25 22 million by 2010. percent have both.7 Rapid urbanization and increasing motorization have Some BRTs have become victims of their own success, proven difficult for policymakers to manage. Conges- as there is a limit on the number of passengers they can tion is common in many urban areas, resulting in fre- serve each day. When ridership approaches 700,000- quent complaints about time lost in traffic. Unfortunately, 800,000 passengers a day, urban transport systems have 6 See, for example, research cited in Ali and others (2015) that used high-resolution imagery to analyze the impact of new road construction on deforestation in Brazil (Laurance, Goosem, and Laurance 2009) and Bolivia, Panama, Paraguay, and Peru (Reymondin and others 2013). 7 Authors’ calculations based on BRTdata.org, urbanrail.net, metrobits.org, and UN (2014). 26  otorists’ morning commutes are long FIGURE 10: M  hough their number is low, bicycle FIGURE 11: T in Latin America, especially in big cities, but not lanes are expanding in Latin America, 2011 necessarily longer than in cities elsewhere (kilometers of bicycle lanes per million urban inhabitants) 120 20 Morning peak (extra travel time, percent) 100 15 80 60 10 40 5 20 0 0 5,000 10,000 15,000 20,000 25,000 0 ay ru ia na e ia r do il gu liv Pe City population (thousands) mb nti Ch a Bo ra Ecu e lo Pa Arg Co Cities in other regions Brazilian cities Mexico City Source: Hidalgo (2011). Source: Authors’ calculations based on TomTom International 2015 and United Nations Population Division 2014. Note: Extra travel time defined as percent additional time needed relative to free-flowing traffic. Red triangles are Brazilian cities, the light blue square is Mexico City, and blue dots represent cities outside Latin America. transportation for citizens who need to reach mass transit such as BRTs and urban rail systems quickly and safely. Inadequate infrastructure for non-motorized transport (such as sidewalks and dedicated bike lanes) to consider diversifying—especially if urban planning or and drivers’ noncompliance with traffic laws expose us- physical geography do not allow BRT systems to expand. ers to traffic accidents, thefts, and assaults. Most vic- That was the case in Quito, Ecuador, which is now plan- tims in traffic accidents are pedestrians and cyclists, ning its first underground rail mass transit system. Metro many of whom belong to lower-income groups. In ad- systems and subways have become increasingly viable dition, pedestrians and cyclists are heavily exposed to modes of transport in large cities with high population pollution. Encouragingly, though the stock is still low, densities. Where urban rail systems are already in place, the number of bicycle lanes has exploded in recent the density of service per resident remains low in most years (figure 11). countries. Except for Santiago, Chile, Latin American cities have substantially less urban rail infrastructure than their The quality of urban transport remains mixed across Lat- counterparts in OECD countries. in America. Some newer public transport systems are of high quality. BRT users across all Latin America cities rate Non-mass transit systems remain poor, also challenged their service a respectable 3.5 out of 5 (BRTdata.org). The by the region’s rapid urbanization. City blocks are of- cost of an average mass transit journey varies substan- ten infilled, leaving no space for pavement. This creates tially across the region, with the average fare in São Paulo problems for pedestrians, who might be forced to walk and Curitiba five times that in Quito and Guayaquil (figure in traffic, making walking dangerous during the day and 12). Whether this is due to longer transit, higher costs, or almost impossible at night. And where pavement exists, greater cost recovery in the tariffs is unclear. it is often poorly maintained. Non-motorized transport is seldom well-integrated with public transportation, and is Without targeted demand-side subsidies, public trans- too often considered a peripheral issue rather than a key port can be unaffordable for low-income groups. House- mode of transport. hold surveys from Bogota, Buenos Aires, Santo Domingo, and São Paulo found that commuting on the formal pub- As part of efforts to reduce inequality, high-quality lic transport system could eat up 20-30 percent of poor pavements and cycle lanes are extremely cost-effec- households’ budgets (Cubas and others 2015). And while tive because poor people make heavy use of these in- there is not an accepted ceiling for what households vestments. They are also essential in providing last-mile should spend on transport, a fare affordability index can 27 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  sing a fare affordability index to guide a subsidy program in Buenos Aires Box 2: U While the scale and scope of a subsidy program are political decisions based on a city’s finances and other goals, certain tools can help inform choices on who should receive funding and how much they should receive. An example is a fare affordability index to measure the financial impact of a standard basket of transit trips—say, 45 a month for each household member—on various income groups. The financial impact of this public expense is often measured as a percentage of household income. Box figure 1 shows this approach for Buenos Aires, where the affordability of public transport was measured for an average household as well as for families in the lowest income quintile.  ffordability of public transport in Buenos Aires, 2003-13 BOX FIGURE 1: A (percent of income) a) For the typical household in Buenos Aires, spending on public b) This was also true for the poorest households ( rst income quintile) transport as a share of total income fell dramatically from 2003 to 2012 10 50 8 40 6 30 4 20 2 10 0 0 2002 2004 2006 2008 2010 2012 2014 2002 2004 2006 2008 2010 2012 2014 Though this approach is helpful, there is no accepted norm for determining what share of household income spent on trans- port would be considered unaffordable. Affordability depends on the alternatives (the practicality of walking and cycling as alternatives to public transport) and other costs of living, including housing. Housing decisions, in particular, have a large impact, because most cities present a tradeoff between high accessibility (in the city center) and more affordable housing (in the metropolitan periphery). Still, the fare affordability index can help policymakers understand and articulate different sub- sidy levels in a structured manner. In Buenos Aires the understanding of affordability levels at different periods helped show that fares could be raised considerably and still be more affordable—as measured by the share of income the poorest quintile spent on a basket of travel—than a decade earlier. Source: Mehndiratta, Rodriguez, and Ochoa 2014. help guide policy decisions on existing or potential sub- subsidizes feeder transport (such as cable cars in Rio de sidies (box 2). Janeiro and Medellin that connect poorer neighbor- hoods to main transit arteries and allow a certain num- Latin American countries have struggled to devise af- ber of free trips per day; Rodriguez and others 2016). But fordable, effective subsidies for public transport. One both programs contribute to urban sprawl and suffer common approach is to set fares for cost recovery, but from errors of inclusion (wealthier persons using these offer targeted demand-side subsidies for specific pop- lower-cost options). ulation groups (Serebrisky and others 2009; Gwilliams 2012; Mehndiratta, Rodriguez, and Ochoa 2014). But ex- Encouragingly, subsidy efforts are becoming more so- perience with demand-side subsidies is mixed, with dif- phisticated as technology makes disbursement more ficulties identifying and reaching target populations and effective and efficient. In Bogota transport subsidies are potential abuse of subsidies (such as transferring them disbursed on travel cards linked to the Sistema Nacional to unintended recipients). Other approaches have in- de Selección de Beneficiarios (SISBEN), the national da- cluded Bilete Unico, operational in Rio de Janeiro and tabase of potential beneficiaries for social support pro- São Paulo, which caps the fare for multi-modal trips and grams (Transmilenio 2015, Rodriguez and others 2016). 28  he average mass transit journey costs more in Latin America than in many other places, 2009 FIGURE 12: T (U.S. dollars) Average, Sao Paulo RIT, Curitiba Transmilenio, Bogota Transantiago, Santiago Megabus, Pereira Metrobus, Istanbul SIT Optibus, Leon Macrobus, Guadalajara Metrobus, Mexico City Transjakarta, Jakarta Metrovia, Guayaquil Metrobus-Q, Quito Janmarg, Ahmadabad 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Source: Cubas and others 2015. The program disburses the subsidy as a half-off fare when the card is used to pay for public transport (capped at 40  ublic transport for all? Sexual Box 3: P trips a month). The program builds on the growing use harassment is a major issue on public of electronic fare media (smart cards) in Bogota’s public transport in Latin America transit systems and national experience with other pov- erty-targeting initiatives (such as conditional cash trans- Safety for women is one component of transport quality fer programs) that use SISBEN. Fraud attempts are de- that is often overlooked. Sexual harassment and assault terred through the use of biometric identification and are higher in Latin America’s major metropolitan areas requirements for photo IDs. than elsewhere. Women in the region consistently report feeling less safe than do women in European metropol- itan areas. In Bogota, Lima, Mexico City, and Santiago, Still, the continued strong presence of informal trans- 60-90 percent of women report experiencing sexual ha- port modes indicates that transport needs are not be- rassment or assault on public transport in any given year. ing met by formal public transport services. Informal Actual levels of sexual harassment and assault could be transport systems are especially common among poor- higher because reporting rates are low. Brazil and Mex- er households that are not served by or cannot afford ico have implemented “women only” cars on trains and formal transport. Relative to formal transport, informal subways. But their reception has been mixed; some argue transport can be more accessible, flexible, reliable, faster, that they do not address the cultural norms that reinforce and cheaper. But the vehicles are typically not suited to sexual violence, while others note that they pose issue for collective transport. They are often unregulated, in over- women traveling with boys or elderly men. supply, unsafe, and unpredictable. And even though in- Source: Balbotín and Arredondo 2015; Jaimurzina, Salas, and Sánchez 2015. formal transport is an important source of employment, it is also a major cause of congestion and pollution. The most common examples of informal transport systems (UN Habitat, 2013). In addition, women are often sexually are the vans and mini/micro buses used in cities like harassed on public transport, undermining their ability to Brasilia, San Salvador, and Santo Domingo. Other exam- safely participate in economic and social activities (box 3). ples are motorcycle taxis in Lima, Santo Domingo, and Fortaleza (Jiron 2013). Water and sanitation: Good coverage for water, More generally, safety is a major challenge, with traffic but sanitation an increasingly urgent challenge deaths a serious issue across Latin America, and not just in urban areas. When weighted by population, the death Good water and sanitation coverage is critical to public rate from road traffic accidents in the region is more than health, especially in areas with high human density, pollu- 19 per 100,000 people—three to four times that in Europe tion, or both. In addition, inadequate access usually implies 29 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE nadequate water and sanitation impose a health burden in Latin America, 2012 FIGURE 13: I (disability-adjusted life years (DALYs) lost due to inadequate water and sanitation) 300 1,000 250 200 100 DALY (100,000 people) DALYs (thousands) 150 100 10 50 0 1 Haiti Brazil Mexico Guatemala Bolivia Colombia Venezuela Honduras Peru Argentina Ecuador Nicaragua Paraguay Panama El Salvador Cuba Chile Jamaica Costa Rica Guyana Uruguay Suriname Belize DR DALY upper bound DALY lower bound DALY estimate DALY per 100,000 Source: WHO and UNICEF 2016. Note: One DALY can be thought of as one year of “healthy” life. The sum of DALYs lost across a population can be thought of as a measure of the gap between current health status and an ideal health situation where the entire population lives to an advanced age, free of disease and disability. that households have to spend more time and money years (DALYs) across the region in 2012, concentrated in getting water. While Latin America has made good prog- Haiti, Brazil, Mexico, Guatemala, and Bolivia (figure 13). ress on increasing the share of households with access to water, it has not done as well on sanitation. Many house- Water and sanitation have traditionally represented a holds lack access to improved sanitation, and only about small share of Latin America’s investments in infrastruc- 30 percent of the region’s wastewater is treated. These ture, hovering between a quarter and a third of a percent shortfalls caused a loss of 941,000 disability-adjusted life of GDP (figure 14). Funding has predominantly come  ater and sanitation investments in Latin America were modest in 2000-12 FIGURE 14: W (percent of GDP) a) Water and sanitation account for a small fraction b) Private investments have typically been limited, of infrastructure investments but seem to be on the rise 3.0 0.30 2.5 0.25 2.0 0.20 1.5 0.15 1.0 0.10 0.5 0.05 0.0 0.00 11 11 01 01 03 04 05 05 07 08 09 10 12 00 02 06 07 08 09 12 00 02 03 04 06 10 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Water and sanitation Total Private Public Source: www.infralatam.info, downloaded on November 15, 2016. 30 mpressive progress on access to water though rural and poor populations still less likely to be FIGURE 15: I served a) Latin America has fairly good overall access to an improved water source, 2015 (percent of population) OCED high income countries Europe and Central Asia Latin America and Caribbean East Asia Paci c Middle East and North Africa South Asia Sub Saharan Africa 0 10 20 30 40 50 60 70 80 90 100 Total population Rural population Urban population Source: WHO and UNICEF 2016. Note: Republic of Korea excluded from OECD High Income Countries for lack of data. b) But with wide variation across the region, especially among rural populations (percent of population) 100 90 80 70 60 50 40 mi on ean bb a na lize Gre a Pa da Uru ay ay ile Co ico a Jam a Pa a Su ma e Gu zil Lat alv a an in Am ador Re s Ve lic ela Ecu a Co or Nic bia a ru iti an ra ari ric yan b am El S mal i gu Ric aic liv Ha ad Pe b Bra u gu Ch na Cu nti x nic du zu lom na Be pu dC e rag ara Me Bo rin sta Gu ate e ne Arg H Rural Urban Do Source: WHO and UNICEF 2016. c) Wealth matters for water access, especially in rural areas, 2012 (percent of population with access, by rural and urban wealth quintile) 100 90 80 70 60 50 40 30 20 pu nic a lize o ala ay ia a ru ia ras a iti yan gu aic xic liv mb Pe gu Re omi Ha c u m Be bli ara Me Jam Bo nd Gu ra ate lo D Ho Pa Nic Co Gu Rural wealth quintiles (Q1 = poorest) Urban wealth quintiles (Q1 = poorest) Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 Source: WHO and UNICEF 2016. 31 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE from the public sector, though the past few years have  eliability of water service is an issue FIGURE 16: R seen an increase in private funding that has helped offset for many, 2008-13 a dip in public spending. (percent of households) Non continuous water Every day continuously Water: good progress on access, but no room for complacency 87 Some 94 percent of households have access to an im- proved source of drinking water, placing Latin America in the top ranks of developing regions—though still short of the 99 percent in high-income countries (figure 15a). Moreover, this average masks a large gap between rural (84 percent) and urban (97 percent) coverage (figure 15b). In fact, while coverage is lower for poor households, the rural/urban divide trumps income as a marker for lack of access to an improved drinking water source (figure 15c). 13 The region’s relatively high coverage is no cause for com- placency. More than 20 million people, mostly in rural ar- eas, lack access to an improved source of drinking water, three-quarters of them in six countries—Brazil, Colombia, 48 Ecuador, Haiti, Mexico, and Peru—that include the re- gion’s richest and poorest countries. The type of service also differs between rural and urban areas, with 94 of the urban population with access to water served with piped water on their premises, but only 68 percent in rural areas 8 (WHO and UNICEF 2016). Further, the quality of “access to an improved water 4 source” is often poor. About a quarter of those with ac- 5 cess get it by informal means (Borja-Vega and others 35 2015).8 Quality is often inadequate, with implications for public health. But data are weak, partly because measur- A few hours per day Every two or three days Once a week ing water quality is difficult and expensive at a large scale. Once every 15 days Less than once in 15 days A pilot effort to use a rapid quality assessment deployed across Nicaragua found that 16 percent of water points Source: CAF 2014, based on a survey conducted in Argentina (Buenos Aires, Cordoba, Rafaela), Bolivia (Cochabamba, La Paz, Santa Cruz), Brazil (Nueva posed high to extremely high sanitary risk. In addition, Iguazu, Rio de Janeiro, San Pablo), Colombia (Barranquilla, Bogota, Medellin), reliability is an issue even in relatively privileged urban Ecuador (Manta, Quito, Guayaquil), Peru (Arequito, Peru, Piura), Uruguay areas, with 13 percent of the population surveyed in a (Montevideo, Salto), Venezuela (Caracas, Maracaibo, San Cristobal), Panama (Panama City), and Mexico (Mexico City). sample of the region’s largest cities reporting they do not have continuous daily service (figure 16). And while Latin America’s utilities perform reasonably metering, about 30 percent non-revenue water (that is, well, most could do better, as measured by their gap with water lost or stolen, for which the utilities do not charge), the top 10 percent of performers. The middle 80 percent and can cover operational expenses but not much more in the International Benchmarking Network for Water from revenues. In contrast, the top 10 percent achieve and Sanitation Utilities (IBNET) database—which covers 100 percent metering, 15 percent non-revenue water, about 1,900 Latin American utilities—average 80 percent and full cost recovery. 8 This share is inferred by looking at the share of the urban population that lives in slums and other informal settlements that utilities may not be allowed to serve. 32 FIGURE 17: M  ost Latin American utilities perform Sanitation performance remains poor, with limited reasonably well but could do better, as evidenced access and low wastewater treatment by the top performers Some 83 percent of Latin America’s population has ac- a) Non-revenue water has fallen over time but cess to some form of sanitation, making the region’s rel- remains fairly high for most utilities ative and absolute performance much worse than for 70 water (figure 18a). Again, rural areas have worse service Di erence between water supplied and sold 60 coverage (64 percent) than do urban areas (88 percent), with poverty being more of a marker for lack of access 50 than living in rural areas (figure 18b). Colombia and Par- 40 aguay stand out, with access in the bottom rural quintile (percent) at 40 percent—half or less than the national average (fig- 30 ure 18c). 20 Wastewater treatment and reuse is also low in Latin Amer- 10 ica. Data are scarce, but it is estimated that only about 30 0 percent of the region’s wastewater is treated—with sig- nificant implications for public health and environmen- 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 tal sustainability. This average masks significant variation b) Most water utilities can cover their operational costs but not much more across countries, from 4 percent of wastewater treated in Costa Rica to 99 percent in Chile (figure 19). 2.5 Total annual operational revenues/ Total One of the most concerning aspects of the institution- 2 al environment for wastewater is the inadequate at- annual operating costs 1.5 tention paid to the regulating, monitoring, and enforc- ing provisions designed to restrict industrial discharge 1 into receiving water bodies. Enforcement of industrial pre-treatment to remove heavy metals, organic com- 0.5 pounds, and other contaminants is needed to prevent damage to pipe infrastructure, ensure that biological 0 treatment processes work effectively, and keep harmful 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 concentrations from preventing the reuse of sludge as a fertilizer. c) Metering seems to have plateaued at 80 percent for most water utilities 100 The poor performance of most Latin American coun- meter/ total number of connections (percent) Total number of connections with operating tries in wastewater treatment stands out compared 80 to other infrastructure sectors, and so merits more in- depth analysis. Background work for this report iden- 60 tifies several causes of this poor performance (Nolasco 2016): 40 ››Split responsibilities between central government 20 agencies that fund construction and local govern- ments that lack the technical and financial resources 0 to run treatment plants. In Mexico, for example, the federal government is responsible for designing and 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 constructing wastewater treatment plants, while cit- Bottom 10 percent Middle 80 percent ies are charged with operations and maintenance. But Top 10 percent most cities do not charge for sanitation, and where Source: Analysis courtesy of Luis Andres and Aroha Bahuguna using IBNET costs are passed on to users, tariffs are often insuffi- data. Dashed lines indicate fewer than ten but more than five observations. The solid lines for Latin America indicate more than 10 observations. cient to cover even operating costs. This disconnect between the central agency that designs and builds wastewater treatment facilities and the subnational 33 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  atin America has not done as well on providing access to improved sanitation FIGURE 18: L a) The region ranks near the middle among its peers, 2015 (percent of population) OECD high income countries Middle East and North Africa Europe and Central Asia Latin America & Caribbean East Asia Paci c South Asia Sub-Saharan Africa 0 10 20 30 40 50 60 70 80 90 100 Total population Rural population Urban population Source: WHO and UNICEF 2016. Note: New Zealand excluded from OECD high-income countries for lack of data. b) Access is particularly low in rural areas (percent of population) 100 90 80 70 60 50 40 30 20 10 0 bb rica Gre na Uru a Co uay a ile ba lize Gu a Ecu a mi Pa or Re y Ho lic Me s Ve ico Co ela & C tin A bia alv e Pa r Nic ma a ru Gu zil la ia iti o ura an ua El S am d yan gu Ric aic liv ma Ha ad ad Pe b Bra Ch na Cu nti Su an x zu La lom ari me na Be pu g nic rag ara Jam Bo nd rin sta e ate e ne Arg Do Rural Urban c) Sanitation coverage is much lower among the poor, especially in rural areas, 2012 (percent of population with access by rural and urban wealth quintile) 100 90 80 70 60 50 40 30 20 10 0 ia pu an a lize a o ia la ay s ru a iti ra yan aic gu xic liv mb ma Re inic Pe gu Ha du Be c Bo ara Me Jam bli Gu ra ate lo m n Ho Pa Nic Co Do Gu Rural wealth quintiles (Q1 = poorest) Rural Urban Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 Source: WHO and UNICEF 2016. 34 FIGURE 19: O  n average, about a third of bear in full. As such, public subsidies may be justified. wastewater in Latin America is treated But costs could also be substantially lowered—and (percent of wastewater treated) wastewater plants transformed from cost to revenue centers—if countries in the region could stop treat- 100 ing wastewater treatment sludge as dangerous sol- id waste and instead allow the reuse of grey waters.. 80 Lima, a large city in the middle of a desert, discharg- 60 es its wastewater into the ocean and disposes of its sludge in expensive sanitary landfills, instead of allow- Est. regional ing the agriculture industry to make use of these nu- 40 average (30) trient-rich bio solids. Similarly, efforts to use electricity 20 generated by treatment plants are seldom encour- aged. Exceptions include the Tenorio treatment plant 0 in San Luis Potosi, Mexico, which provides advanced Chile Mexico Brazil Colombia Peru Costa Rica (2012) (2011) (2012) (2012) (2011) (2013) primary treatment for 60 percent of the wastewater Lower bound Upper bound (which is then used for irrigation), and generates elec- Source: Nolasco 2016. tricity for both own consumption and for sale—sav- ing $18 million over a six-year period. agencies that operate them also exists in Argentina ››Infrastructure not adapted for poor people. Sewage and Peru. systems often suffer if they provide piping through neighborhoods, because either resources or interest ››Overly ambitious “imported” regulations that leave is lacking to make final connections between the pip- no room for gradualism. Many Latin American coun- ing and businesses. Examples abound. In Guayaquil, tries have adopted legislation developed in high-in- Ecuador, where 30,000 households remain uncon- come countries that have strong institutional and nected 30 years after sewerage pipes were installed, technical capacity and high financial support from about 31 percent of households surveyed noted that both government and users. Such legislation often money was a key impediment to connection (Poveda imposes standards that are unrealistic and unafford- and others 2014). A number of countries in the region able. To illustrate: in Cordoba, Argentina, legislation have developed schemes to tackle this issue (box 4) implemented in 2015 requires that wastewater be- ing discharged have a level of cleanliness that few treatment plants can meet even in high-income Energy: A sector at a turning point? countries. The cost of operating the Cordoba plant far exceeds the municipality’s financing capacity. Latin America’s energy sector could be at a turning point Further, in most countries regulation is binding from (World Bank forthcoming). The region is well positioned the day of its passage, instead of offering a path to close the remaining access gap—at least for electric- toward gradual improvement in treatment cover- ity, as the need for clean and efficient cooking fuel has age and quality. The region would be better served not received enough support in most countries. But shifts by focusing on a progressive array of appropriate in demand associated with urbanization and a growing treatment technologies, starting with cheaper, low- middle class, together with climate change, are creating er-cost solutions followed by upgrades to more ad- both new challenges and opportunities. vanced technologies as and when financial resourc- es allow. Investments in energy dropped dramatically in the early 2000s and have been slowly recovering since, hovering ››Limitations on resource recovery. Few utilities, even in around 0.75-1.0 percent of GDP in recent years (figure high-income countries, charge users the full cost of 20a). The drop was driven by a sharp and sudden decline wastewater treatment. This is partly because house- in private flows to the sector, from about 1.4 percent of holds are reluctant to pay for a service that they feel GDP in 2000 to about 0.2 percent four years later. Private may benefit others as much as themselves, but also flows have slowly and unevenly increased in recent years, because in low- and middle-income countries the to account for some two-thirds of energy investments in cost would be high for an average household to the region in 2012 (figure 20b). 35 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE nnovative schemes to expand sewage services across Latin America Box 4: I In Brazil, where cost is a significant barrier to the expansion of sewerage in major cities, at least three utilities have taken steps to create more pro-poor services. EMBASA, Bahia state’s utility, implemented an innovative program between 1995 and 2007 to introduce “condominial” sewerage to low-income neighborhoods. This program was a city block-based approach that relied on community consultations to design low-cost joint sewerage facilities. It placed the responsibility for final linkups on groups of households rather than individual households. Similar schemes have been pioneered by CAESB, Brasilia’s state water utility. Under its condominial scheme, connections to the main public sewers are constructed after the targeted community signs an agreement setting out the rights and obligations of all partners. The investment cost of the system is kept low by the use of small bore pipes, starting at diameters of just 100 millimeters (Shankland and others 2010). Another Brazilian utility, SABESP, targeted the “last mile” problem by offering sub- sidies and payment by installment for connecting households to sewerage services. Finally, in the state of Minas Gerais, the state water utility COPASA has established a subsidiary that offers pro-poor tariffication and technologies. Colombia, Peru, and Uruguay are also implementing projects to tackle sewerage for the poor but they have yet to develop them at scale. Elsewhere, public-private partnerships for wastewater treatment plants appear to be working well. A perfect example is the Atotonilco de Tula plant in Mexico, which treats wastewater generated in the valley of Mexico. It is 54 percent funded by the private sector and is operated by a private consortium. There are also examples of private businesses building wastewater treatment plants to make use of municipally discharged wastewater—the Enlozada-Cerro Verde treatment plant in Arequipa, Peru, was built and operated by the mining company Cerro Verde, which uses the treated water at its mine site. Source: Nolasco 2016. If given enough attention, the access gap could be living standards are associated with a lack of electricity closed and nonsolid fuels. Latin America has fairly high access to electricity and modern (nonsolid) fuels, with access rates of 96 percent The need for cleaner, more efficient cooking has not in electricity and 84 percent in nonsolid fuels (figure 21). received enough policy and financial support in the This is well above the developing country averages, esti- region. More than half of the population in Guatema- mated at 86 percent and 59 percent, and on par with the la, Honduras, and Nicaragua use solid fuel for cooking. middle-income average for electricity. Even in relatively wealthy Mexico, this share is 14 per- cent. These levels are worrisome given the association Nevertheless, 22 million Latin Americans remain without between indoor air pollution and acute lower respirato- access to electricity (more than half of them in Haiti and ry infection and chronic obstructive pulmonary disease, rural Guatemala and Peru) and 87 million lack access to 2 of the top 10 causes of death in Latin America (IHME nonsolid fuels (three-quarters of them in Brazil, Guatemala, and World Bank 2013). Haiti, Mexico, and Peru; figure 22). Electrification rates are increasing throughout the region, but for access to non- Affordability remains a major barrier to increased access solid fuels, slow progress in a number of countries (such as to electricity. Even with social tariffs, subsistence electric- Guatemala and Venezuela) means that these countries are ity consumption is not affordable to the poor in many unlikely to attain universal access any time soon. Latin American countries (figure 23). In Bolivia, Colombia, Guatemala, and Mexico the cost of subsistence electrici- A major issue is inequality of access for both elec- ty—30 kilowatt-hours a month, enough for limited light- tricity and nonsolid fuels—not only in countries with ing, a television, a fan, and one medium-power appli- lower levels of access such as Haiti, Guatemala, Guy- ance—imposes an unacceptable burden for the poorest ana, and Honduras, but also in upper-middle-income 40 percent of the population. Part of the problem is that countries such as Brazil, Chile, and Mexico. Access is electricity subsidies are not always well targeted. Elec- mostly a challenge in rural areas, but some periurban tricity subsidies are regressive in every Central American areas and slums lack access or have unreliable, illegal, country—to give an example, in Panama, for every $1 of and unsafe electricity supplies. Low quality of life, se- subsidies received by the poorest 10 percent of house- vere health problems, poor education and medical holds, the richest 10 percent gets $4.4 (Hernandez and care, and limited opportunities for raising incomes and others, forthcoming). 36  nergy investments in Latin America are rebounding, 2000-12 FIGURE 20: E (percent of GDP) a) After a dramatic drop, energy investments are slowly recovering b) Private investments, which drove the drop, are also driving the timid recovery 3.0 1.6 1.4 2.5 1.2 2.0 1.0 1.5 0.8 0.6 1.0 0.4 0.5 0.2 0.0 0.0 11 01 00 02 03 04 05 06 07 08 09 10 12 11 01 00 02 03 04 05 06 07 08 09 10 12 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Energy Total 20 Private Public Source: www.infralatam.info, downloaded on November 15, 2016.  ccess to electricity and nonsolid fuels is high, but not relative to peers and not in rural areas, FIGURE 21: A 2012 (percent of population) a) Access to electricity Europe and Central Asia OCED high income countries Middle East and North Africa Latin America and Caribbean East Asia Paci c South Asia Sub Saharan Africa 0 10 20 30 40 50 60 70 80 90 100 Total Rural Urban b) Access to nonsolid fuels OCED high income countries Middle East and North Africa Europe and Central Asia Latin America and Caribbean East Asia Paci c South Asia Sub Saharan Africa 0 10 20 30 40 50 60 70 80 90 100 Total Rural Urban Source: World Development Indicators. 37 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  ccess deficits are concentrated in a few countries, 2012 FIGURE 22: A (percent of regional total) a) Electricity b) Non solid fuels 24 21 24 30 Haiti Mexico Guatemala Brazil Peru Peru Colombia Guatemala 6 8 13 Haiti Honduras 6 Nicaragua Colombia 15 11 Others 7 Others 12 12 11 Source: World Development Indicators. FIGURE 23: I n many Latin American countries, the Sustainable Energy) index, which includes indicators to poorest cannot afford electricity assess how far countries have gone in introducing the (affordability, percent of bottom 40 percent household income) key policies, regulations, and plans needed to increase 12% energy access and attract private participation (based on current good practice), shows most Latin American 10% countries to be doing well. But Haiti, with energy access levels similar to those in Afghanistan and Somalia, is still 8% lacking the core elements of an energy access program 6% (an officially approved electrification plan, a framework for grid electrification, or even a framework for off-grid 4% electrification based on standalone systems). Honduras has also progressed slowly on all these fronts, with policy 2% and planning instruments that consider some good prac- 0% tice elements but not all. Chile Ecuador Honduras general Mexico Argentina social tari Bolivia social tari Brazil social tari Colombia social tari Dom Rep social tari Guatemala social tari A gap remains between the world’s best performers and Latin America’s major utilities Latin America pioneered electricity market reforms and 6 kWh/month 20 kWh/month 30 kWh/month what became known as the orthodoxy of the 1990s (un- Source: World Bank forthcoming; based on SEDLAC household surveys. bundle, privatize, regulate). As a result, several countries in Note: Affordability is measured in terms of the annual cost of subsistence the region have highly sophisticated electricity markets, consumption as a percent of bottom 40 percent household income. The analysis with good governance and regulatory certainty. More is based on national social tariffs (and not on special programs applying to sub-national levels). Electricity is generally considered affordable if annual generally, most countries in the region have functional expenditure on a volume of subsistence consumption is less than or equal to 5 electricity markets, and the region has built a wealth of percent of income. Preliminary work from World Bank Energy Global Practice. experience with the design and operation of electricity markets (box 5). On a more positive note, except for Haiti and Honduras, The reforms strengthened the performance of many util- countries with the lowest levels of electricity access have ities, some of which have grown to become multination- implemented policies that should increase it. The ener- al companies or at least companies able to sell shares gy access pillar of the RISE (Readiness for Investment in through stock markets to finance their investment plans 38 below 3 percent of its 264 billion gallon capacity, forc-  atin America has pioneered Box 5: L ing blackouts throughout the region (Poindexter 2015). innovations to make markets more Reservoir expansion in the region may be difficult given economically and technically efficient the mounting opposition to large hydroelectric projects, including the 11.2 gigawatt Belo Monte project in Brazil’s Successful electricity reforms Latin America are a testa- Amazon as well as Chile’s five-dam HidroAysen project ment to the sophistication of the region’s professionals (Tissot 2012). Other forms of renewable energy—geo- and institutions and their willingness and ability to em- thermal, solar, tidal, wind, biomass, and biofuels—con- brace change. Examples of adjustments and innovations tribute only a sliver of electricity production in the region that moved markets to increased competition and effi- (10 percent in 2013), and have ramped up too slowly to ciency include: offset the decline of hydroelectricity. ››Through regulatory adjustments, Colombia pio- neered the auctioning of reliability payments to en- But while non-hydro renewables are a rather new phe- sure the availability of energy during dry periods and nomenon in Latin America, such investments are rising. El Niño events. The region has seen rapid growth in solar photovoltaic (PV) plants, with a doubling of regional capacity in 2015— ››Argentina, Brazil, and Peru introduced competitive though from a low baseline. Similarly, onshore wind in- tenders or auctions for concessions to ensure the vestments have been growing since 2008 (figure 26). Sig- timely addition of transmission capacity, and success- fully separated ownership from operations, whereby nificant small hydropower and bioenergy investments are transmission belongs to multiple owners. made every year, while geothermal investment appears ir- regularly, with a limited number of projects under develop- ››Peru and Brazil from 2009 (followed by Chile in 2015 ment in the region. Private investment in renewables has and Mexico in 2016) launched auctions to scale up also gradually increased in the region, mostly to support nonconventional renewable energy. wind-based generation and mainly concentrated in Brazil, Source: World Bank forthcoming. Chile, Mexico, and Peru. (Colombia, Ecuador, Uruguay, and Central American countries also attracted modest private sector investment in renewable energy in 2008-14.) (such as ISA, the state-controlled electricity grid operator of Colombia, Empresas Públicas de Medellin). But a few With its reliance on renewables and vulnerability to climate still struggle with service quality and financial sustainabil- shocks, Latin America would benefit enormously from ity. That may be due to the reversal of reforms in some good regional interconnections. But many existing inter- countries or to the lack of managerial expertise in some connections are not used because of technical, regulatory, utilities. And overall, there is still a gap between most Latin and market barriers.9 For instance, regional power trade in American utilities and global best performers (figure 24). Central America, including interconnection with Mexico, was just 2-4 percent of potential trading capacity in 2012 and 2013. Power exchanges between Argentina, Brazil, and Electricity is clean but vulnerable to climate shocks, Uruguay have also been consistently low. Electrons are not and still has lots to gain from “negawatts” flowing across borders—or much less than they could be. With 56 percent of installed electricity generating capac- Another factor that could help the region keep its low ener- ity in renewable sources, Latin America has the world’s gy emission footprint, as well as reduce vulnerability, would cleanest electricity sector. But much of this depends be increased energy efficiency. Latin America’s energy in- on hydroelectricity, which has suffered from droughts tensity is the lowest of any region, largely due to the nature brought on by climate change and poor water resource of its economic activities. But energy efficiency, particularly management. As a result the share of electricity pro- in the use of fossil fuels, remains poor. And the region could duced from renewable sources has been falling in recent reduce transmission and distribution losses, which are the years, reaching 53 percent in 2013 (figure 25). In January highest of any region except the Middle East and North Af- 2015 São Paolo’s six-reservoir hydroelectric system fell rica (and three times those of East Asia; figure 27) 9 The region has four clusters of interconnection: Mexico-United States, Central America, Andean Community (Colombia, Ecuador, Peru, Venezuela), and Southern Cone (Argentina, Brazil, Paraguay, and Uruguay). 39 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  few Latin American utilities are among the best, but most are less reliable than the global FIGURE 24: A median performer, 2015 1,000.00 Honduras Dominican Republic 100.00 System average interruption frequency index (SAIFI) Guyana El Salvador Sao Paulo, Brazil Paraguay Nicaragua Suriname Uruguay Peru 10.00 Bolivia Chile Colombia Jamaica Argentina Panama Dominica Ecuador Grenada 1.00 0.01 0.10 1.00 Los Angeles, USA 10.00 100.00 Spain Denmark Beijing, China Global Median France Czech Republic Germany United Kingdom Shanghai, China 0.10 New York City, USA Monterrey, Mexico Moscow 0.01 System average interruption duration index (SAIDI) Source: World Bank 2016c. Note: Axes are in log scale. SAIDI is the average outage duration for each customer served on a yearly basis, measured in minutes. SAIFI is commonly used as a reliability indicator by power utilities, and is the average number of interruptions that a customer experiences in a given year. This dataset measures SAIDI and SAIFI scores for the largest utility from the largest business city, unless otherwise noted. Dark blue dots are for Latin American utilities, while light blue ones are for the country noted.  mong regions, Latin America has the largest share of electricity produced from renewables, FIGURE 25: A but this share has been declining due to droughts, 2001-13 (percent of electricity produced by renewable sources) 70 60 50 40 30 20 10 0 East Asia and Paci c Europe and Central Latin America Middle East and North Africa Sub-Saharan Africa Asia and Caribbean 2001 2006 2011 2012 2013 Source: http://data.worldbank.org/products/wdi 40 nvestments in renewables are rising rapidly in Latin America, driven by onshore wind FIGURE 26: I (US$ billions) 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Onshore wind Solar pv Bioenergy & waste Geothermal Small hydropower Marine Source: World Bank forthcoming. Note: Includes both public and private investment. Hydropower larger than 50 megawatts is not covered by the analysis, but is expected to remain the dominant technology for further renewable energy deployment in the region in the long term.  he region’s transmission and distribution losses are some of the highest in the world, 2001-13 FIGURE 27: T (percent) 20 18 16 Middle East & North Africa 14 Latin America & Caribbean 12 Europe & Central Asia 10 Sub-Saharan Africa 8 East Asia & Paci c 6 4 OECD high income 2 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Source: World Bank forthcoming. In sum, and potential to do better. And it can do so by spending One way to interpret the diagnostic laid out above is to more efficiently on the right things. argue that Latin America’s infrastructure performance is rather mediocre for an upper-middle-income region with So, where to go with this diagnostic? The next section significant growth aspirations. Another is to highlight the turns to the challenges and opportunities that will shape many examples of the region’s capacity for innovative how Latin America can or should tackle its infrastructure solutions, its expertise with sophistications, and its expe- agenda. rience with PPPs. In sum, Latin America has the means 41 Part II. What lies ahead for the region’s infrastructure? Latin America faces many challenges and opportunities have been long in the making but have reached a scale as it contemplates where and how to improve its infra- that puts the region in a different realm of demand and structure. Some of these are new, such as climate change delivery options. And others are familiar, such as the fiscal and the policy and technology shifts associated with it. space challenge and generally poor spending efficiency Others, such as urbanization and a growing middle class, that have long plagued Latin America. Inefficient public spending may limit how much more can or should go to infrastructure A recent analysis of infrastructure governance in Brazil a systemic bias against capital spending. Many of these concluded that “creating additional space for investment causes are compounded by limited coordination be- may not lead to economic growth or better infrastruc- tween sector agencies and central and subnational gov- ture services, unless the management of capital projects ernments. These points, developed below, suggest sub- is considerably improved” (World Bank forthcoming, p. stantial potential for efficiency gains. 3). And while in-depth analysis of infrastructure planning and investment frameworks is rare—in Latin America and elsewhere—it is possible to glean insights from both Weak planning, project appraisal and preparation public expenditure and investment management analy- capacity sis and the 20 or so reasonably recent World Bank public expenditures reviews that have addressed at least one in- Public investment management systems are inefficient frastructure sector. (See annex 1 for a list.) across the region, though some countries are better than others. Most countries have public investment manage- A review of these works shows that inefficient public ment systems whose original goal was to standardize spending on infrastructure in Latin America has myriad capital budget preparation by setting legal procedures causes. Many lie outside infrastructure sectors or are sys- for project identification and budgeting. A World Bank/ temic across government agencies—but all converge IMF Public Investment Management index (PIMI) (Dab- around a lack of institutional capacity for planning, capi- la-Norris and others 2012) found that Latin America has tal budgeting, and implementation. Issues include weak low investment efficiency, with significant variations appraisal and preparation capacity, overly rigid budget- across countries, mainly determined by how old systems ing rules designed to control cash expenditures rather are. The study analyzed four stages of infrastructure man- than improve spending efficiency, difficulties with bud- agement—appraisal (including planning), selection, im- get execution, inefficient procurement procedures, and plementation, and evaluation—and found appraisal and 43 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE evaluation to be the weakest stages. A 2016 study reached “the combination of weak capacity, lax enforcement, and similar conclusions (Armendaris and Contreras 2016). pork-barrel politics means that investment projects can be included in ministerial budgets without having been Longer-term infrastructure planning is undermined by subject to formal appraisal (World Bank forthcoming)” short-term pressures. The region’s public investment management systems were built with a strong focus on In terms of quality control, the situation is often worse at managing fiscal deficits and therefore strong controls on the subnational level. States and municipalities frequent- actual cash expenditure. There is a disconnect between ly face a lack human and technical capacity to manage decentralized project planning responsibilities in sector large-scale investments. The state of Rio de Janeiro has a agencies and overall fiscal targets and plans. This discon- multi-annual plan to guide its investment decisions, but nect emerges because agencies develop plans without day-to-day decisions are at the mercy of political and sec- being integrated with the budget discussions. For ex- toral considerations that lead to ad hoc budget alloca- ample, a 2015 World Bank perception survey of project tions (World Bank 2012a). Subnational levels sometimes managers in Jamaica concluded that only 13 percent follow national levels guidelines and manuals for project of government projects were aligned with a high-level appraisal without properly understanding them. In ad- strategic goal.10 In Haiti a lack of coordination among key dition, subnational levels often see the use of national agencies and the absence of an integrated budget are guidelines and manuals as a purely formal pre-requisite major challenges. While the country has a detailed, com- for applying for national funding rather than as a way to prehensive budget cycle for investments, the budget still strengthen their own decision-making process.11 fails to incorporate long-term decision-making in several ways. Overly rigid or myopic budgeting Another issue is the complex interplay between politics and planning.  Spending efficiently requires a sober as- In the region, annual budget rules limit the inclusion of sessment of any project’s net benefits to society. Projects large and long term projects. A majority of public invest- should be prioritized against each other, across sectors, ment projects are funded on an annual basis, meaning and over time—thereby turning wish lists into strategies. these projects need to be executed in less than a year That approach requires a political mandate, a process for according to tight fiscal calendars. Of over 23 Latin Amer- generating such a mandate, and institutional capacity to ican countries that had Public Expenditure and Financial manage such a process (in an infrastructure unit, the min- Accountability (PEFA) analysis done since 2006, only one istry of finance or planning, or the president’s office). Key, got an A rating (Colombia) on the PEFA indicator that however, is a robust public investment management pro- captures the existence of sector strategies with multi- cess that ensures quality. year costing of recurrent and investment expenditure. Out of a portfolio of 42,810 projects registered in Boliv- However, a lack of public investment management-based ia’s project data bank (SISIN) until 2013, nearly 70 percent quality control leads to projects being funded that are were for 2 years or less. Less than 1 percent of Peru’s near- not sufficiently prepared. A 2014 UN-ECLAC survey of 15 ly 220,000 public projects registered between 2000 and Latin American economies found that in five countries, 2015 at national and subnational levels had a duration of PPP projects can bypass public investment management three or more years. system controls, while in four more significant exceptions can be obtained (Perretti and Vera 2015). Brazil has a mul- There is growing realization that this is an issue and some tiyear planning process, but it is seen more as a bureau- countries are working on reforms. Colombia, for example, cratic burden than as an instrument for medium-term has a strong planning system and has introduced medi- planning (World Bank forthcoming). In addition, parlia- um-term expenditures frameworks. However, it still has a mentary amendments to annual budget laws can add complex process for multiannual budget allocations for substantially to planned investments (some 33 percent projects that requires approval by a specialized commit- in 2015). A recent study consequently concludes that: tee and dedicated cabinet approval. 10 Survey conducted by the World Bank in August 2015 under the Jamaica Strategic Public Sector Transformation Project (P146688). 11 Courtesy of Diego Dorado Hernandez, based on the findings of World Bank Rapid Assessment and Action Plans conducted in: Lambayeque, Perú (2013), Santa Cruz de la Sierra, Bolivia (2014), Córdoba, Argentina (2014), San Salvador, El Salvador (2014), and Colombia (2010-2013). 44 TABLE 3: O  nly one Latin American country fares information about financial flows three months or less well with respect to the multi-year budgeting of before scheduled project start. In Mexico’s water sector, projects. unwieldy budgetary rules squeeze the budget cycle in a (latest available year, 2007 – 2013) way that makes it difficult to engage in multiyear budget- Bolivia D ing and properly execute capital budgets. This is because Brazil C resources are not released until a technical agreement Colombia A is signed between the federal water agency and partic- ipating states—expected by March 31 but sometimes Costa Rica D delayed—and any resources unspent by December 31 Dominican Republic C must be returned. This tight schedule results in less com- El Salvador D petitive procurement procedures and low budget execu- Grenada D tion (World Bank 2014a). Guatemala C Finally, the fiscal calendar is often misaligned with the Haiti D project implementation calendar. The Latin American Honduras D fiscal year is the calendar year, which results in fiscal au- Jamaica C thorities requesting deliverables and payments before Paraguay D the end of the year. However, for most infrastructure proj- Peru C ects, the right time to do investment is during the dry or summer season that happens between November and Trinidad and Tobago D March. In response, some countries have created fiducia- Source: PEFA Secretariat ry funds to avoid this budget rigidity, while other chose Note: Table shows scores for ID-12 indicator (iii): Existence of sector strategies with multi-year costing of recurrent expenditure and investment PPPs or struggle with multiannual allocations. Difficulties with budget execution Further, the effectiveness of budget allocations is limited by a lack of information about and timeliness of finan- Many countries struggle with the execution of capital bud- cial flows to projects. This hampers project manager’s gets (figure 28). This is a particular issue in Brazil where dis- capacity to do financial planning during project imple- bursement data show a chronic gap between committed mentation stages. In Jamaica, the same survey of proj- and executed funds. This problem was exacerbated by the ect managers found that over 70% of projects received transformation of the Pilot Program of Investments into  any Latin American countries chronically under-execute their capital investment budget FIGURE 28: M (disbursement as percent of commitment) 100 80 60 40 20 0 Bolivia Brazil Chile Colombia Honduras Peru 2013 2014 2015 Source: Courtesy of Diego Dorado, based on data from government budgetary reporting systems (Bolivia: VIPFE; Brasil: SIOP; Chile DIPRES; Colombia: Portal de Transparencia; Honduras: SIAFI; Peru: www.mef.go.pe/estadisticas) 45 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE the much more ambitious Program of Growth Accelera- percent of its 2008 budget by taking advantage of pro- tion, which increased public investments from 0.7 of GDP curement practices such as reverse auctions, consolidat- in 2007 to 1.3 percent in 2012. In 2013 the Federal Audit ed purchases, and clearer standards (World Bank 2008). In Court found that a third of road construction and manage- Haiti excessively high thresholds for public tendering of ment projects were paralyzed for a variety of reasons, and works, goods, and services result in overreliance on direct three-quarters of projects with fully committed funds had procurement or noncompetitive invoicing (World Bank an execution rate of less than 25 percent. The court also 2016d). Guatemala has no systematic procurement re- found serious irregularities in 58 of the 102 large infrastruc- views for the procurements of goods and services, many ture projects it audited (World Bank forthcoming). of which occur in infrastructure sectors (World Bank 2013b). In Brazil adjustments of 25-50 percent of initial Low or medium level budget execution has many causes. cost estimates are allowed to preserve “economic-finan- For example, in Honduras cash rationing is considered the cial equilibrium,” encouraging underbidding and over- main driver of low budget execution in infrastructure invest- charging (World Bank 2014b). ments, as money is siphoned off to protect current expendi- tures. Furthermore, budget execution is defined as accrued Some economies in the region such as Chile, Mexico expenses rather than actual payments, resulting in perpetu- and Brazil have seen remarkable success in driving pro- ally overstated budget capacity (World Bank 2013a). curement reforms. In Chile for example, the ChileCompra electronic portal is estimated to have generated US$280 Chronic over execution is also symptomatic of poor bud- million in savings. Mexico’s modernization of its tender- geting. In Guatemala the budgetary requirements of the ing processes (which involved eliminating 586 obsolete perpetually underfunded road ministry were estimated procurement regulations and creating an online platform in 2010 to be four times its $1 billion budget. These fund- to boost transparency and ease of access) generated sav- ing requirements include projects that are approved but ings of more than US$1 billion within three years of its not executed, as well as projects executed but where the 2009 start (World Bank, 2013c). In Brazil, a reform estab- contractor has not yet been paid for its work. To com- lishing an e-procurement system led to 51 percent sav- pensate for the lack of funds, the ministry often receives ings in transaction costs and 25.5 percent reduction in transfers from ministries that have more funds than they prices between 2000 and 2006 (Hunja 2015). can fully execute. That, combined with frequent cost overruns in road projects, led the road ministry to exe- But detailed analysis of procurement performance in Latin cute 120-200 percent of its budget between 2006 and America, both general and of public private partnerships 2008 (World Bank 2013b). This effectively undermines the shows there is considerable room for improvement (see annual budget process’ ability to guide fiscal policy. Annex II for a more detailed review). The region’s reason- ably good average performance compared to others hides wide variation both across countries and across different Procurement that could be improved aspects of performance. The World Bank’s Benchmark- ing Public Procurement 2017 found that suppliers identify Inefficient procurement processes contribute to limit- obstacles such as excessive bureaucracy and red tape in ed budget execution. Capital expenditures depends on Colombia and Honduras; payment delays in Argentina, procurement policies for the timely provision of goods Jamaica, Dominican Republic; lack of transparency and and services. Some procurement or budget laws do not opaque tendering process in Brazil and Mexico; lack of ef- allow agencies to start the procurement process before ficiency in Barbados and Puerto Rico; the list goes on. a budget appropriation has been given. This causes un- necessary delays. In Jamaica, the project manager survey revealed that 76% of the projects faced delays on the im- Insufficient attention to social and environmental plementation of their procurement plans. risk management 12 In addition, better procurement could save significant Adequate social and environmental risk management resources in many Latin American countries. For exam- helps secure popular support for a project—a “social ple, it is estimated that Costa Rica could have saved 13-18 license to operate” which reduces business cost by 12 This section is based on Stapledon (2012), Geurs and others (2009) and World Bank (2006) 46  roject delays as a result of land acquisition, expropriation and regulation requirements TABLE 4: P Project and Concession Start date Total Investment at Completion by 2015 in % risk (US$ millions) Jorge Chavez International Airport 2001 1,062 30.6 Red Vial No. 6. Pucusana-Cerro Azul-Ica 2005 294 36.5 Autopista del Sol-Trujillo-Sullana 2010 330 21.2 Line 2, Lima Metro 2014 5,347 4.2 Source: Ministry of Economy and Finance, Peru. Total investment refers to the initial estimated cost of the investment. Unclear Project sustainability reducing project delays, cost overruns, and reputational risk to investors (Stapledon 2012). Data from the Min- An imbalance between capital and current spending on istry of Economy and Finance of Peru showed that de- infrastructure is a chronic problem around the world. It lays associated with land acquisition and expropriation can spring from overly rigid budgets and suboptimal significantly delayed projects. The Ministry found that planning. Of the 23 Latin American countries that had these delays caused substantial increases in project PEFA analysis since 2006, only three ever got a rate of B costs (table 4). (Guatemala, Peru, and Trinidad and Tobago) on the in- dicator capturing the link between investment budgets In Latin America, while some countries have put in and forward expenditure estimates (table 5). place relevant policies and procedures to address sus- tainability in infrastructure, there is still a long road The imbalance is also frequently due to poor coordination ahead. Yet when social sustainability dimensions are between central governments (which fund and often man- not clearly understood and properly analyzed, deci- age capital investments) and local governments (which may sion-makers tend to ignore them until they flare up and lack the financial and technical capacity to take over these directly threaten project implementation (Geurs and investments or cover operation and maintenance costs). others 2009). Another area in which social and environmental risk man-  atin American countries score poorly on TABLE 5: L agement is undervalued is feasibility studies. As in other linkages between investment budgets and forward regions, feasibility studies in Latin America tend to focus expenditure estimates. on engineering as well as economic and financial anal- (latest available year, 2007 – 2013) ysis. They seldom make use of other tools such as envi- Bolivia C ronmental and social impact assessments, stakeholder Brazil C engagement, or pay due attention to Free, Prior and In- Colombia C formed Consent rules. Costa Rica D Attention to these issues can help avoid later compli- Dominican Republic C cations, reduce negative impacts, and compensate for El Salvador C any residual impacts. In the experience of the World Grenada D Bank, feasibility studies sometimes fail to adequately Guatemala C assess project sites, unforeseen site conditions includ- Haiti D ing social and economic activities, existing utilities, and most importantly contextual risk. Moreover, feasibility Honduras C studies must integrate the projects’ unique characteris- Jamaica C tics in order to properly analyze their potential distribu- Panama D tional impacts on diverse social groups, both spatially Paraguay D and temporally. A flexible and adaptive social and en- Peru B vironmental management system in which the level of effort to manage the risks is proportionate to the de- Trinidad and Tobago B gree of the risks in the infrastructure projects improves Source: PEFA Secretariat Note: Table shows scores for ID-12 indicator (iv): Linkages between investment projects’ chances of success. budgets and forward expenditure estimates 47 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE In Honduras municipalities are legally mandated to invest Bank 2013a). Peru has the opposite problem: the fiscal re- 50-70 percent of the transfers they receive from the central sponsibility law’s requirement that mining revenues (can- government, but because transfers are unpredictable in ons) be dedicated to investments has led to an imbalance both timing and amount, municipal leaders tend to front- between capital and current expenditures in local budgets load recurrent expenditures (salaries, interest payments and and the need for the central government to step in for criti- so on) and use whatever is left over for investment (World cal current spending (World Bank 2012b). A tight fiscal stance limits how much more could be spent on infrastructure Improved efficiency is not just good policy; it may well be Many countries in Latin America are now facing per- the only option in the short to medium term to increase sistent and often sizable fiscal deficits, which is bound to resources available given the region’s tight fiscal position. impact their ability to undertake much-needed public in- This is not new. Fiscal constraints have long influenced vestments in infrastructure even with the asymmetry dis- both infrastructure investments as well as policies in the cussed above. But as discussed below, the overall fiscal region. picture is not uniform across the region. Infrastructure privatization and reforms in the 1990s were largely driven by fiscal concerns. But while the re- Public investments in Latin America: Rising forms did ease the strain on public coffers, it quickly be- during the boom, but still much lower than in came clear that public-private partnerships (PPPs) were no substitute for public investment in infrastructure. As other regions a result, by the early 2000s, it was clear that the region A narrative that has gained some traction in the region needed to invest more. But by then Latin America was states that Latin America squandered its opportunity to struggling with negative fiscal balances, which were invest during the boom years of 2004-08 and instead raising fears of a return to hyperinflation and fiscal de- mostly expanded public consumption. That is not quite bacles. Pressure to limit spending combined with severe true. Total public investment (not just infrastructure relat- budget rigidities (typically due to a large share of the ed) increased by 50 percent during 2003-08, while pub- budget allocated to constitutionally mandated expen- lic consumption held steady. The reverse has happened ditures or to entitlements) meant that investments bore since 2009, during the crisis years, with public investment the brunt of the adjustments—even as some were ar- declining while consumption spending expanded (figure guing that infrastructure investments, being growth-en- 30). However, throughout that time, public consump- hancing, should be increased. The debate raged on re- tion has remained many multiples of public investment, garding whether fiscal space could be created at least which hovered around a modest 3-4.5 percent of GDP – for carefully selected growth-enhancing investments significantly lower than among competitors and peers, (that in turn would create more fiscal space through especially Asian ones (table 6). growth). Most countries opted not to for fear of jeopar- dizing credit ratings. The implications for infrastructure investments are dire. Though PPPs have played an important role, they only Indeed, evidence shows that the evolution of the pub- represent about 40 percent of infrastructure investments lic sector’s budget constraint—which plays a big role in in the region. Further, about a third of PPP financing shaping fiscal space—matters for the level of infrastruc- comes from public institutions (and about half of them ture investment. Public investments in infrastructure in rely on some type of government support, creating con- Latin America respond to lagged changes in public sav- tingent liabilities on public coffers). In other words, pub- ings. However, this response is asymmetric, being stron- lic and private investments are complements rather than ger in good times (when lagged public savings rise) than substitutes. Without vigorous effort on the public side, in bad times (when lagged public savings fall) (Serebrisky infrastructure investments in the region are unlikely to in- and others 2015). crease. In the absence of fiscal space, this vigorous effort 48  otal public investments expanded during the boom years while public consumption remained FIGURE 29: T steady – but public consumption remains many multiples of public investment (percent of GDP) 18 16 14 5 12 4 10 8 3 6 2 4 1 2 0 0 2000 2002 2004 2006 2008 2010 2012 2014 2016f 2000 2002 2004 2006 2008 2010 2012 2014 2016f Public investment Public consumption Source: World Development Indicators, IMF World Economic Outlook database. Note: Shaded area represent the boom years of 2003-2008. Regional average is weighed by GDP shares. n Latin America, total public investment is much lower than in other regions TABLE 6: I (percent of GDP) Region 2000-2003 2004-2008 2009-2013 2013-2016 Period average (2000-2016) East Asia Pacific 17.5 15.5 14.8 12.2 15.1 Middle East and North Africa 5.7 7.5 9.1 8.9 7.8 South Asia 6.0 7.3 7.1 6.0 6.7 Sub Saharan Africa 5.3 5.2 5.7 5.6 5.4 Europe and Central Asia 2.8 4.1 4.3 4.2 3.9 Latin America and the Caribbean 3.4 3.7 4.5 4.0 3.9 Source: World Development Indicators, IMF World Economic Outlook database. Note: Regional averages are weighted using PPP GDP. 2016 is forescast. would require reducing public consumption to create (ii) the Mexico-Central America-Caribbean (MCC) group, space for public investments—something that is unlike- where net commodity importers prevail, and which is ly given existing budget rigidities and the current reces- generally following the US cycle. As figure 31(charts A-B) sion— and/or substantially increasing public spending shows, fiscal balances have deteriorated markedly in efficiency. most of South America, but not so much for the average MCC country. A “bifurcated” fiscal panorama in Latin America But even within these broad groups, there is some het- erogeneity. Fiscal deterioration has varied among South A highly heterogeneous fiscal space picture has emerged American countries. Faster growing economies, such as in the region, largely driven by trade structure (De la Torre Chile, Colombia, Peru, and Uruguay have fared better fiscal- and others 2016). We can identify two broad groups: (i) ly than slower growing ones, such as Argentina, Brazil, and South America, dominated by net commodity exporters, Ecuador (figure 30). On the whole, South American coun- which are generally following China’s ups and downs; and tries have not yet succeeded in their adjustment efforts: 49 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  atin America and the Caribbean: F FIGURE 30: L  iscal “Bifurcation” (percent of GDP) a) South America: Fiscal balances per country group b) Mexico, Central America & Caribbean: Fiscal balances per country group 2 2 1 0 0 -2 -1 -2 -4 -3 -4 -6 -5 -8 -6 2010 2011 2012 2013 2014 2015 2016f 2010 2011 2012 2013 2014 2015 2016f South America low growth South America high growth Central America Caribbean Mexico c) South America: Country scal de cits d) Mexico, Central America & Caribbean: Country scal de cits 14 8 12 6 10 4 8 2 6 0 4 -2 2 -4 0 -6 -2 -8 Brazil Bolivia Argentina Ecuador Uruguay Colombia Chile Peru Paraguay Costa Rica El Salvador Dominican Republic Mexico Nicaragua Guatemala Honduras Panama Jamaica Fiscal de cit Primary de cit Fiscal de cit Primary de cit Source: IMF World Economic Outlook database and national sources. Prepared by the World Bank Latin America and the Caribean Chief Economist Office. Note: South America low growth includes Argentina, Brazil, and Ecuador; South America High Growth includes Colombia, Peru and Uruguay. Central America includes Costa Rica, El Salvador and Guatemala; Caribbean includes Dominican Republic, Haiti, and Jamaica. fiscal deficits have widened as expenditures continued to through revenue increases. In the case of Jamaica, a mas- grow—with Ecuador a clear exception (figure 31, Chart sive reduction in its primary deficit was achieved, helping A). There are some other important differences among address the country’s hitherto unsustainable debt trajec- SA countries, attesting to the distinct nature of their fiscal tory. Nevertheless, in a few MCC countries (e.g. Costa Rica woes: While interest payments play an outsize role in Bra- and Nicaragua), the fiscal situation remains uncomfort- zil, Colombia and Uruguay, primary expenditures are the able, characterized by high and increasing primary and/ dominant factor in Argentina, Bolivia and Venezuela. or overall fiscal deficits. In contrast, there has not been a dramatic, widespread While fiscal consolidation should ultimately help countries deterioration of fiscal balances in the MCC countries – improve their fiscal stance, the composition of adjustment even as they generally started from a weaker fiscal stance matters too. It has been all too common in the past for than their South American counterparts. Indeed, there countries to disproportionately cut capital expenditures as has been some reduction in fiscal deficits, especially in a politically more expedient way of adjusting. It is still not the Caribbean (figure 31). Fiscal consolidation occurred entirely clear whether current or capital spending is bear- mostly through expenditure reductions rather than ing the brunt of fiscal adjustment in the present round. 50  atin America and the Caribbean: S FIGURE 31: L  ources of changes in fiscal deficits (percent of GDP) a) South America: Fiscal de cits – sources of change 12 8 4 0 -4 -8 -12 Brazil Bolivia Argentina Ecuador Uruguay Colombia Chile Peru Paraguay Change in revenue Change in expenditure Change in scal de cit b) Mexico, Central America & Caribbean: Fiscal de cits – sources of change 6 4 2 0 -2 -4 -6 -8 Costa Rica El Salvador Dominican Mexico Nicaragua Guatemala Honduras Panama Jamaica Republic Change in revenue Change in expenditure Change in scal de cit Source: IMF World Economic Outlook database and national sources. Prepared by the World Bank Latin Amercia and the Caribbean Chief Economist Office. Note: The changes are computed between the last observation and the point in time after 2009 when the lowest fiscal balance was recorded. If the lowest balance corresponds to the latest observation, the difference was computed with respect to the point in time when the fiscal balance was closest to zero. The values for the total and primary fiscal deficits correspond to the latest available observation. Data coming out of recent IMF Staff Reports show a mixed to ascertain clearly the composition of fiscal adjustment picture. Capital spending as a share of GDP is projected to in the region – and its implications for public investments. contract in Bolivia, Ecuador and (to a lesser extent) Jamaica The point remains, however, that few, if any, Latin Ameri- and Mexico, while remaining broadly stable in Costa Rica can country can boast of ample fiscal space with which to and Peru. More country-level information will be needed finance a significant infrastructure expansion. 51 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE Climate change is creating new challenges, but possibly new opportunities Changing climatic conditions, natural disasters, and ex- a potential competitive advantage, especially if carbon treme weather events are affecting the ability of existing taxes are imposed in some of its key export markets. So infrastructure to deliver services in Latin America. Melting infrastructure also needs to be made cleaner and more glaciers and recurrent droughts are undermining hydro- efficient. electric production and may become a serious challenge for water-constrained cities. Storms, floods, and land- slides wash away roads and bridges and destroy other Infrastructure needs to be more resilient and infrastructure. Thus infrastructure must become more better adapted to the changing climate resilient. Disaster and climate risks are expected to increase in At the same time, Latin America has the world’s cleanest the region, with implications for infrastructure develop- energy matrix and plenty of potential to further reduce ment (box 6). Energy, transport, and water and sanita- emissions associated with infrastructure—opportuni- tion systems need to be built in more resilient ways to ties that carry local and immediate benefits such as re- reduce disruptions during extreme events. In addition, duced air pollution and congestion. This offers the region climate change will alter the demand for infrastructure  ow will climate change affect Latin America? Box 6: H Latin America is exposed to many climate risks, most of which materialize through the water cycle. These risks vary by sub-re- gion but also depend on climate change. Impacts are expected to increase in intensity and severity as global temperatures continue to rise: ››Most dry regions could get drier and most wet regions could get wetter. Increased rainfall in tropical and subtropical Pacific coastlines and southern Brazil is in contrast with decreased precipitation in the Caribbean, Central America, northeast and central Brazil, and Patagonia. Rainfall could decrease 20-40 percent if global warming increases 4oC above pre-industrial temperatures (a “4oC world”). ››Seasonal distribution of stream flows could become more variable. Glaciers could disappear in a 4 C scenario and suffer o considerable loss in a 2 C scenario. This retreat could cause glacial lake outbursts and flooding, as well as reduced water o runoffs in some river basins. ››Increased risks of droughts and extreme heat could undermine seasonal water availability. Higher mean temperatures and increased rainfall variability could extend dry spells and drought conditions by up to 20 percent in a 4oC scenario. ››More intense rainfalls could increase the risks of landslides. Excessive rainfalls within short periods could overwhelm urban natural drainage systems. Landslide risks are highest in sloping terrains—where poor people often live. ››The intensity and frequency of tropical storms will likely increase. The frequency of the strongest tropical cyclones could increase by 40 percent in a 2oC world and by 80 percent in 4oC world. Storms could increase coastal flooding, mostly in the Caribbean. ››Higher sea levels could cause coastal flooding and erosion. Projections vary by zone and range from an average of 0.38 meters in a 2oC world to 1.14 meters in a 4oC world. Such extreme coastal flooding and rising sea levels could expose many zones to the risk of storm surges. Some of these risks will be conditioned by changes in ecosystems and feedback relationships. A 40 percent decrease in the Amazon’s forest area could be a potential tipping point at which forest-climate interactions could decrease precipitation. For- est cover loss is already associated with droughts in the Amazon. Source: Magrin and others 2014; World Bank 2014c. 52  isaster damages for infrastructure are highest for transport FIGURE 32: D Hurricane Felix, Nicaragua, 2008 Hurricane Wilma, Mexico, 2005 Floods, Guyana, 2005 Hurricane Ivan, Cayman Islands, 2004 Hurricane Isidore, Mexico, 2002 Hurricane Keith, Belize, 2000 Hurricane Ivan, Jamaica, 2004 Hurricane Gustave, Haiti, 2008 Hurricane Jeanne, Haiti, 2004 Floods, Santa Fe, Argentina, 2003 Floods, Mexico, 2007 Hurricanes Jeanne & Francis, Bahamas, 2004 Hurricane Stan, El Salvador, 2005 Hurricane Emily, Mexico, 2005 Tropical Storm Ida, El Salvador, 2009 La Niña, Bolivia, 2008 Hurricane Stan, Guatemala, 2005 El Niño, Bolivia, 2007 Floods, Guatemala, 2010 Hurricane Stan, Mexico, 2005 0 20 40 60 80 100 % of total damage Transport Water & sanitation Energy Other infrastructure Non-infrastructure Source: GFDRR Loss and damage database (www.gfdrr.org/damageandlosses) accessed November 2016. services: heat waves will increase electricity demand, coastal cities and on Caribbean islands, where most droughts and extreme rains will create the need for critical transport systems are in low-lying areas, these more water storage capacity, and increased flood risks impacts could be large. More intense tropical cy- will demand infrastructure solutions such as protective clones in combination with rising sea levels could ex- dams. tend port downtimes for ships, raising shipping costs and reducing trade. Extreme events already cause extensive damages to in- frastructure. Rapid loss and damage assessments done ››Rising sea levels and intense El Niño events could after natural disasters in Latin America suggest that in- threaten coastal infrastructure. Combined with frastructure bears a disproportionate share of the costs. tropical storms, these could mean that a 100-year The transport sector, in particular, bears a heavy burden flood event formerly reaching 2.8 meters could reach (figure 32). nearly 8 meters by mid-century. This is a major con- cern given that more than 7.5 million inhabitants, Climate change can interact with infrastructure in many 42,600 square kilometers, and built capital valued different ways (Magrin and others 2014; World Bank at $334 billion are situated at elevations below to- 2014c): day’s 100-year extreme sea level (Reguero and others 2015). ››Extreme events such as intense rainfalls and storms are already damaging transport systems, causing ››Rainfall variability and prolonged water stress widespread disruptions. Intense rainfalls can cause could threaten water supply in parts of the region. flash floods and landslides, which can destroy bridg- Though Latin America has relatively abundant wa- es and other critical road segments—leading to high ter resources, the spatial and temporal dimension of repair costs and severe travel disruptions. Especially in water vary (Miralles-Wilhelm 2016). Notwithstanding 53 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  number of sub-regions of Latin America show a consistent drying trend MAP 2: A Water availability, 2015 Change in water availability, 2015-50 Annual runo (km3/year) in Latin America and Caribbean Change in the annual runo (km3/year) in Latin America and the Caribbean Source: Miralles-Wilhelm 2016. Note: Projections for 2050 are for a world on a trajectory to reach 4oC warming by 2100. large uncertainties, areas in the Amazon basin, north- A wide range of options is available to increase the re- ern Mexico, northeast Brazil, the Pacific coasts of Chile silience of energy, transport, and water and sanitation and Peru, and countries in the Caribbean and Central systems, whose effectiveness depends on climate and di- America show a consistent drying trend that is also saster hazards and local contexts (table 7). These include expected to endure (map 2). Large cities like Lima and engineering options (to retrofit existing infrastructure or Mexico City suffer from water insecurity and regular design new infrastructure), technological and ecosys- water shortages. tem-based options (such as protecting systems through conservation or restoration of natural systems), and insti- ››Extreme events and changes in seasonal stream tutional actions (such as rerouting of roads and electricity flows are already changing the supply of and de- grids, reducing exposure to hazards). mand for energy services. Droughts in Brazil, Colom- bia, Costa Rica, Uruguay, and Venezuela and torrential Given the long lifespan of most infrastructure, robust de- storms and rains such as those seen in Bolivia, Chile, sign options and adaptive pathways that build in flexi- and Paraguay are threatening energy infrastructure in bility are needed to manage future uncertainties (box 7). the region. More variable river runoff could limit hy- Although resilience planning is still a nascent field—es- dropower generation, which produces the most elec- pecially in the transport and electricity sectors—there tricity in the region. Storms with high winds and in- is growing recognition that uncertainties about future tense rainfalls could damage pipelines, windmills, and climate impacts need to be incorporated in today’s in- power transmission and distribution lines, resulting in vestment decisions. A recent analysis by the World Bank power shortages. Heat waves increase peak demand, and SEDAPAL, Lima’s water utility, did just that and de- reduce thermal conversion efficiency (and thus firm veloped a robust investment strategy that ensures water capacity), diminish transmission capacity, and in- reliability across as wide a range of future conditions as crease the temperature of cooling water sources. possible (Kalra and others 2015). Other key water utilities 54 TABLE 7: W ays to make infrastructure more resilient, by approach and sector Approach Electricity Transport Water and Sanitation Engineering Adjusted design codes for power plants and Hazard-resistant construction standards and More or better-maintained water reservoirs electricity grids design parameters for roads and bridges and storage Underground cable networks Strengthened/ heightened protection walls Pumping station Increased system capacity for roads Sewerage work Increased grid integration Raised height of causeway roads in ports Drainage systems Technological Renewable technologies Use of new designs and materials Water-saving technologies Use of information technologies to control Hydrometeorological monitoring traffic Desalination technologies Ecosystem-based Water flow regulation and sediment control Slope stabilization through replanting Protection of upstream areas that provide through watershed protection water regulation and filtration Institutional Hazard mapping, vulnerability assessments, spatial planning, disaster preparedness and contingency plans Source: Adapted from IPCC 2014. in the region—such as Aguas de Manizales in Colombia billion a year (Ward and others 2010). The World Bank’s and Empresa Pública Metropolitana de Alcantarillado y Economics of Adaptation to Climate Change estimated that Agua Potable de Quito in Ecuador—have also conducted the total costs of adapting infrastructure in Latin America comprehensive risk studies to prepare for extreme events could be around $4 billion a year (World Bank 2010).13 (Balcazar 2012). But given the avoided costs from negative impacts, in- Resilient strategies need not necessarily cost more. The vesting in more robust infrastructure can have a high resilient investment plan developed by Lima’s water util- payoff. A study of Peru’s 2007 earthquake estimates that ity (and shown in box 7) resulted in changed rather than the country could have saved 27 times the recondition- more investment, with should reduce investment costs ing spent on water and sanitation infrastructure if the af- by 25 percent, thanks to an increased emphasis on de- fected systems had had proper maintenance and been mand side-management, and a focus on “no regret” in- built using earthquake-resistant materials (Cannock and vestments (Kalra and others 2015). others 2011). Also in Peru, the benefits of making critical segments of the country’s road network flood resilient But estimating an overall cost of adapting to climate exceed the costs in almost all future scenarios (figure 33). change is made difficult, if not impossible, by the fact Generally, however, benefits—like costs—are very loca- that adaptation costs are highly situation- and site-specif- tion specific and very difficult to estimate. ic. Not surprisingly then, available estimates vary greatly along with the cost elements and climate scenarios con- Cross-sectoral coordination mechanisms and integrated sidered. One study estimated that for Brazil alone some planning frameworks help make infrastructure sectors $50 billion of capital investments would be needed to more prepared and responsive. Effective solutions need ensure reliable electricity supply by 2035 due to the pro- to work across sectors and levels of governments align- jected lack of reliability of hydroelectricity (de Lucena and ing local plans with regional and national strategies. For others 2010). Background work for this report suggests example, after Chile’s 2010 earthquake the biggest chal- that providing additional reservoir capacity to meet fu- lenge for water and sanitation systems was the absence ture industrial and municipal water demand across Lat- of electricity. Chile’s experience shows that in addition in America could cost $44-57 billion between now and to improving the response capacity of individual water 2050—an annual average of $1.3-1.6 billion (Miralles-Wil- utilities, these utilities need to be integrated in a region- helm 2016), higher than the previous estimate of $1 al, multisector contingency plan (Balcazar 2012). Recent 13 The study estimated the cost to be $3.5 billion in 2005 dollars, which is about 4.2 billion in current (2016) dollars. 55 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  on-probabilistic decision-making under uncertainty methodologies Box 7: N Although the scientific evidence for a changing climate is clear, uncertainty about the speed and intensity of these changes is high. In this context it is impossible to define probabilities for future climate conditions or extreme climatic events. To account for these deep uncertainties, non-probabilistic approaches are needed—often called decision-making under un- certainty (DMU) methodologies. Rather than weighting futures probabilistically to define an optimal strategy, these methods seek to identify robust strategies— those that satisfy decision makers’ objectives in many plausible futures (Lempert and others 2013; Kalra and others 2014). These methods have been applied, for example, to define a robust portfolio of water res- ervoirs for implementing Lima’s long-term water resource plan (Kalra and others 2015). The basic steps include identifying all the possible conditions that could make a project fail, and deciding whether these are reasonable scenarios to try and protect against based on decision makers’ tolerance for risk, inputs from experts and stakeholders on whether the scenarios are worth worrying about, and the costs of robust strategies options. Such an approach can also help to define pathways that allow for flexibility and adjustment of the strategy once new infor- mation becomes available and future developments become more predictable. For example, for Lima an adaptive portfolio of water reservoirs was identified, starting with no-regret reservoirs to be implemented in the near term. In the medium term, reservoirs can be added depending on the feasibility of and necessity for more complicated projects. And in the long term, decision makers can choose those reservoirs that meet future water demand.  obust and adaptive portfolio of water reservoirs to implement Lima’s Water Resource BOX FIGURE 1: R Master Plan Near-Term Mid-Term Long-Term 50% Budget, Full Project Feasibility Mid-Term, Full Feasibility (No additional projects) • Cañete Trans/WTP Climate and • Casacancha Res Demand? 75% Budget, Full Project Feasibility • Chosica Res/Grato WTP • Chancay GW • Chancay Res/Hural WTP • Lima Sur Desal • Jacaybamba Res 75% Budget, Full Project Feasibility Full Project and Additional Actions Feasibility? • Chancay GW • Chancay Res/Hural WTP • Lima Sur Desal Near Term • Jacaybamba Res • Atarjea WTP • Additional actions beyond Master Plan Limi • Lurin WTP ted • Pomacocha Res/Huachipa Climate and WTP Demand? 50% Budget, Limited Project Feasibility (No additional projects) Mid-Term, Limited Feasibility • Chancay GW New actions beyond Master Plan • Chancay Res/Huaral WTP • Additional actions beyond Master Plan • Lima Sur Desal Future conditions Legend Wetter Implement no-regret projects Feasibility undetermined Limited feasibility Drier Monitor uncertainties Full Feasibility Less More demand demand Source: Kalra and others 2015. Note: “WTP” stands for water treatment plant, “Res” is reservoir, “Desal” is desalinization plant, “GW” is groundwater 56  lood-proofing critical road segments in Peru pays off in almost all possible scenarios FIGURE 33: F (net present value of interventions, US$ million) 4,000 Panamericana 800 Piura 600 Carretera Central 3,000 600 400 2,000 200 400 1,000 0 200 -200 0 -1,000 0 -400 -2,000 -200 -600 -3,000 -400 -800 More frequent Maintenance Increasing Redundancy Increasing redundancy (For some routes only) More frequent Maintenance Increasing Redundancy Increasing redundancy (For some routes only) More frequent Maintenance Increasing Redundancy Increasing redundancy (For some routes only) Flood-proof Road Flood-proof Road Flood-proof road Source: Briceño-Garmendia and others, 2015 Note: The three figures show for critical segments of three major roads in Peru (Panamericana, Piura, and Carretera Central) the net present value of four possible interventions: more frequent maintenance, increasing redundancy (for all or just some routes), and flood-proofing. The net present value of these interventions is estimated for hundreds of possible scenarios—depicted in the many horizontal bars—that explore uncertainties around possible future flooding, damage associated with this flooding, and time to rebuild. The net present value is calculated as the difference between the reduction in annual losses (including user costs and rehabilitation costs) due to intervention and the investment and maintenance costs of the intervention option. droughts in São Paulo have shown the need not only larger than average contributor: in 2012 transport ac- for technological solutions, but also for assigning water counted for 32 percent (compared with 23 percent global- rights in basins shared among jurisdictions (CAF 2015). ly) of energy-related emissions. Most Latin American coun- tries saw rapid increases in emissions from both electricity Resource management is also critical to improving resil- and transport: in Brazil both shot up more than 60 percent, ience. Work done for this report argues that future water while in Peru emissions from electricity generation tripled scarcity will be driven by water demand rather than cli- and emissions from transport grew 80 percent (figure 35). mate-influenced water availability (Miralles-Wilhelm 2016). Similarly, energy efficiency and demand-side management Without action, emissions from infrastructure-related in electricity will help reduce the need for more baseload sectors are likely to increase further. According to the power plants, all of which are heavily water dependent. Global Energy Assessment (GEA) model developed by the International Institute for Applied Systems Analysis, a business as usual scenario with increased motorization Pressures will mount to reduce emissions from and decreased reliance on hydroelectricity would see en- ergy-related emissions in the region more than double infrastructure between 2010 and 2050 (Vergara and others 2013). Infrastructure-related sectors will play an important role in achieving Latin America’s emission targets. Although Mitigation options in energy-related sectors could re- the region accounts for just 11 percent of global emis- duce emissions considerably. Estimates from the GEA sions, significant emission reductions are needed to re- model suggest that energy-focused mitigation strategies main within the global 2oC warming target. Emissions in in Latin America could cut annual emissions by up to 4.1 Latin America grew from 3.6 gigatons of carbon dioxide GtCO2e by 2050, reducing energy-related emissions to al- equivalent (GtCO2e) in 1990 to 4.4 GtCO2e in 2012, with a most zero. Such savings would require 60-80 percent of doubling of emissions from electricity and transport and the primary energy mix to come from renewables, 75-100 a 50 percent increase in “other energy” more than offset- percent of the electricity mix to come from low-carbon ting an 18 percent decrease in emissions from land use sources, and further improvements in energy intensity changes and forestry (figure 34). (Vergara and others 2013). More generally, mitigation strategies involve action on three fronts: decarbonizing Among energy-related sectors, transport is the fastest energy mixes, electrification (notably of transport), and growing contributor to emissions (figure 34). It is also a increasing efficiency (Fay and others 2015). 57 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  atin America’s greenhouse gas emissions have been growing driven by energy-related FIGURE 34: L emissions, 1990-2012 (gigatons of carbon dioxide equivalent, GtCO2e) 5 4 3 2 1 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Land use change and forestry Non-energy Other energy Transport Electricity Source: WRI 2016.  missions from energy-related sectors have been growing since 2000 FIGURE 35: E (megatons of carbon dioxide equivalent) Electricity Transport Other energy 700 700 700 600 600 600 500 500 500 400 400 400 300 300 300 200 200 200 100 100 100 0 0 0 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 Brazil Mexico Argentina Venezuela Colombia Bolivia Chile Peru Others Source: WRI 2016. In transport, the main mitigation options include shift- La Paz has developed a cable car to bring passengers ing to lower-carbon transport modes, switching to from the suburbs to the city center. Given the relatively low-carbon fuels, improving vehicle and engine per- clean energy mix, switching to electric cars can reduce formance, and reducing journeys and travel time. Some overall emissions, as shown in Bogota (Delgado and oth- of the greatest gains could come from developing bus ers, 2014). Similarly, enhancing the quality and quantity rapid transit (BRT) systems. These systems have main- of nonmotorized transport (such as cycling and walk- streamed bus-based mass transit by offering high ca- ing) can cut emissions. For interregional or intercountry pacity and good service at relatively low costs and have transport, the expansion of waterways, such as the Pan- been developed in cities in Argentina, Brazil, Ecuador, ama Canal, is another option for reducing emissions (de Guatemala, and Uruguay (World Bank 2013d). Important Marucci 2012). advances have also been made in upgrading and ex- panding rail-based systems, such as in the metropolitan Electricity decarbonization is possible through in- areas of Rio de Janeiro and São Paulo (World Bank, 2012). creased reliance on renewable energy. However, 58  he region is improving its business environment for renewable energy investments, though Box 8: T it remains far from the good practice frontier The Readiness for Investment in Sustainable Energy Index (RISE)—a suite of indicators that assesses the legal and regulatory environment for investment in sustainable energy, including renewable energy, energy efficiency, and energy access—gives Latin America a 52 percent score (“average/to be improved”) for renewable energy index and 41 percent for energy efficiency. This is better than Sub-Saharan Africa but below other developing regions, especially for energy efficiency. The index evaluates whether a region has introduced key measures on planning, policies and regulation, and administrative efficiency, and assess- es counterparty risk and the existence of carbon pricing and monitoring. In its first global rollout (2015), RISE covered 14 Latin American countries: Argentina, Bolivia, Brazil, Chile, Colombia, Domin- ican Republic, Ecuador, Guatemala, Haiti, Honduras, Mexico, Nicaragua, Peru, and Venezuela. Most of these countries have introduced laws, regulations, targets, and action plans to promote renewable energy development (Haiti and Venezuela are exceptions). Most have also developed resource mappings and integrated renewable energy into traditional expansion plan- ning. But it is in the details of regulations, operational rules, and planning where countries still need progress. For example, few countries in the region consider renewable energy scale-up in their transmission planning, and few produce strategic plans or provide zoning guidance on siting, which is essential to investors. Most countries also need to make regulatory and policy incentives more efficient and effective, and improve the creditwor- thiness of their utilities to make them more attractive to potential investors. Finally, credit enhancement and risk mitigation instruments—such as letters of credit, escrow accounts, and payment guarantees—are needed to attract private investment in renewable energy, but few governments in the sample are able to offer such incentives. A few countries in the region have developed policies, regulations, and institutions to boost energy efficiency (Mexico, and to a lesser extent Brazil and Colombia). But implementation has been weak, and most has occurred in the context of crises and deficits in energy supplies. In addition, fossil fuel subsidies remain high in countries such as Argentina, Bolivia, Chile, Ecuador, and Venezuela. Source: World Bank forthcoming; Banerjee and others 2017. expansion of hydroelectricity is limited by environmen- but also to reduce investment needs and local pollu- tal and social concerns and climatic variability. Envi- tion. Latin America’s energy intensity is declining, but at ronmental concerns push for run-of-the-river plants, a slower rate than in other regions (which, admittedly, which means existing reservoirs are being used in ex- have much more room for improvement) and not in all cess, compromising the flexibility they bring to the countries. Brazil has actually increased its energy intensity overall energy system (Broad, de Moura, and Howells from 3.9 megajoules per 2011 PPP$ in 1990 to 4.1 in 2012. 2016). Thus the region is losing storage capacity each Significant scope for further reducing energy intensity year, because the production of hydro plants with exists in transport, agriculture, industry, and commercial large reservoir capacity remains constant but demand services. But significant improvements in regulations and is increasing. Fortunately, the region has considerable policies—as well as their implementation—are needed potential in wind (especially in Brazil, Chile, Ecuador, (box 8). Peru, and Uruguay), geothermal (mostly in Andean countries), and solar (Argentina, Brazil, Chile, and Peru The question is whether Latin America’s power sector have areas for large-scale electricity production using is ready for the kind of transformation and new busi- concentrated solar power plants). Equally important, ness models that are emerging in advanced econ- the region has made good progress on improving the omies as a result of a number of related disruptive business environment for investment in renewables— factors: new technologies and changing cost curves, though it remains far from the good practice frontier, as climate change pressures (for both mitigation and ad- demonstrated by the Readiness for Investment in Sus- aptation), shifts in demand profiles associated with dis- tainable Energy Index (box 8). tributed generation (the consumer as producer—or prosumer), increased demand for air conditioning and The region has significant potential for improved energy cooling, and possibly demand for electric transporta- efficiency—which is critical not just to reduce emissions tion (table 8). 59 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  atin America’s disruptive challenges in the power sector TABLE 8: L Challenge Response Greening the grid and transportation • The region already has the world’s greenest energy matrix with extensive use of hydro resources, and • Expand nonconventional renewables sustain- modest use of coal, through coal remains the dominant fuel for power generation in Mexico and smaller ably—both at utility scale and distributed markets. generation. • Most countries are deploying grid-connected nonconventional renewables, pioneering the use of • Expand clean mass transport solutions and increase competitive schemes (auctions). Penetration of distributed generation still modest. market penetration of energy-efficient vehicles • Large cities are adopting simple mass transit solutions (such as bus rapid transit corridors). (cars, trucks, buses). • Vehicular compressed natural gas (CNG), blended gas, and fuel ethanol have significantly reduced CO2 and other pollution emissions. Making the power system more resilient • Distribution automation has helped make the grid more resilient, with fault location, automated crew • Increase automation and technical resilience of the dispatch, and other ways to handle catastrophic events. grid (supply side). • The combination of complementary nonconventional renewables with hydro storage has made the • Foster the adoption of decentralized solutions. system more resilient to drought and helped manage intermittency cost-effectively. • Develop a comprehensive load control and demand • Much more is needed on demand-side management, including stronger interface with clients, energy response program, particularly in load congested management initiatives, and demand response. There have been a few cases of effective power crisis areas. management using demand response (in Argentina, Brazil, Chile, Colombia, Panama), but it has not • Combine storage (thermal, electric) with load been embedded in the regulatory compact. control to mitigate renewable intermittency. New utility model and regulatory compact • Unlike in Europe, the Republic of Korea, and the United States, utilities are not yet a key vehicle for • Increasing the role of the utility (or supplier) in delivering energy efficiency and demand-side management, even though they are the most effective going “beyond the meter” in supporting energy channel for achieving significant energy and demand reduction. Only Brazil has a “wire charge” of 1 per- efficiency and distributed generation. cent of utility revenue, half of which is dedicated to energy efficiency. This is largely due to inadequate • Specific policy directives to foster energy efficiency, regulations, which do not compensate the utility for energy efficiency efforts. demand-side management, and renewable energy initiatives. • Specific regulations to encourage utilities to engage beyond the meter, such as decoupling, and shared savings Creative business and financial models • Power sector in the region still dominated by regulated utilities and independent producers. In industrial • Emergence of new players and applications, in countries there is a multitude of other players (marketers, demand aggregators, curtailment service collaboration or competition with the utility in providers, home management systems, and others) working in collaboration or in competition with providing information, load control, home man- utilities, and they facilitate the diffusion and adoption of technologies for which the traditional utility agement systems, and distributed generation. has no appetite or interest. • New value added services and players – such as • A few forward-looking utilities have established unregulated subsidiaries to provide value added district cooling facilities, energy service companies, services, such as Light ESCO (energy services in Rio), and EPM (district cooling in Medellin). and concessions for street lighting • The power sector lacks financial instruments to deal with a variety of new risks, particularly on the • New financial models and products to manage climate front. Some interesting experiences have encouraged in Uruguay (low-hydrology hedge instru- range of new risks. ment) and Colombia (auctioned Cargo por Confiabilidad). Source: World Bank forthcoming. According to recent World Bank analysis, Latin Ameri- significantly and sustainably scaling up wind and solar ca is reasonably well positioned to transform its ener- generation. gy sector and leapfrog to new business models (World Bank forthcoming). Successful reforms in the 1990s in These new technologies and a potential transformation several medium-size and large countries show that of the business model could mean that the clean way for- the region’s power sector can embrace and manage ward is the cheapest. Modeling done for this report finds change. The region has some of the most sophisticated that South America would need to invest $23-24 billion a power sectors among developing regions and is largely year in additional generation capacity under a business prepared to confront some of the challenges facing 21st as usual scenario or even under the nationally deter- century utilities. For example, Latin America pioneered mined contributions (NDCs) that countries in the region the use of competitive procurement for renewables, offered at the 2016 Paris Climate Conference (COP22) 60  he costs of a green transition could drop dramatically in South America if full use is made of TABLE 9: T new technologies and business models Model and Scenarios Investment Additional capacity / Investment / year (US$ billion) year (gigawatts) (US$ billion) SAMBA (From 2018 –31) Business as usual 303.0 9.1 23.3 RET 336.5 101.2 25. 9 INDC 310.1 9.2 23.9 Disruption 109.2 3.1 8.4 ICEPAC business as usual (From 2012-31) 315.9 7.77 24.3 Business as usual Source: Broad, de Moura, and Howells 2016; World Bank forthcoming. Note: SAMBA is the South America Electricity model developed by KTH Divisions of Systems Analysis. RET is a scenario with higher levels of renewable penetration but rapid growth in demand, INDC includes all renewable additions envisaged under countries’ Intended Nationally Determined Contributions as per the UNFCCC Paris Agreement, Disruption includes the same large-scale renewable energy as under the NDCs but much greater decentralized renewable energy and energy efficiency, along with smart grids. ICEPAC is based on national expansion plans and demand estimates (OLADE’s Super Model). Note that for ease of comparison, SAMBA prices were applied to the capacity expansion of ICEPAC (ICEPAC assumes much lower prices). (table 9). But under a disruption scenario that compares n Brazil long-term planning shifts FIGURE 36: I with the NDCs in terms of large-scale renewable energy optimal abatement strategies but makes much greater use of new technology for smart (emision reduction by 2020, megatone carbon dioxide) grids, decentralized generation, and energy efficiency by 60 both grids and end users, costs could drop to $8.4 billion a year between now and 2031. Leapfrogging, perhaps 50 not surprisingly, could save the region a lot of money. But a number of issues would need to be tackled for this 40 transformation to happen. Most utilities in the region lack the skills to interact with customers more active- 30 ly—a challenge given the role that utilities must play in implementing integrated planning, fostering energy ef- 20 ficiency, helping clients manage load, and working with end users to scale up distributed generation. Further, the 10 transformation needed will be driven by changes in pol- icies and regulations. Policymakers will need to develop 0 national energy goals and policy tools, while regulators In the 2020 strategy In the 2030 strategy must provide utilities with incentives to engage “beyond Improvement in re neries Metro, train and bullet train the meter”—moving away from pure concepts of “return Other measures on assets” and “revenues coupled to sales” to a new ap- Source: Vogt-Schilb and others 2015. proach that compensates utilities for a broader set of ser- vices provided to clients. More generally, and beyond the electricity sector, long- and improvements in refineries. Measures for 2030 are term planning is critical to greening the region’s infra- costlier but have much larger abatement potential in the structure. The optimal strategy and mix of abatement long term, such as building a new metro system. When options depend on the time horizon and end goal. An designing an emissions reduction plan, costs, time to example from Brazil shows how the planning horizon implement, and overall emissions reduction potential will affect what the “right” strategy looks like (figure 36). all need to be considered. Options in transport that take A strategy for 2020 makes more use of measures that are time to reach their full potential, but are required to reach cheap and quick to implement, such as new processes long-term goals, need to get started early. 61 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE Urbanization and changing socioeconomics are complicating matters Latin America is highly urbanized yet its cities are still see- occupation that hinders work on access roads or water, ing growth and transformation, with infrastructure both sewerage, or drainage work, this complicates infrastruc- leading and following these changes.14 Urbanization has ture provision and raises the costs. As a result estimates been one of the drivers of increased access to infrastruc- are that upgrading a slum costs two to eight times more ture services, making it easier to achieve economies of than regular land development (Abiko and others 2007). scale. But these savings depend on the density of cities: The implications are serious for infrastructure investment encouraging compact development instead of sprawl costs given that infrastructure typically represents some can reduce public infrastructure and service costs by 10- 70 percent of slum upgrading costs (Marulanda and oth- 40 percent (Marulanda and others 2015; Litman 2015). ers 2015). (There is a limit to the gains from densification, however, as the limits of bulk infrastructure can be reached above a What about changing socioeconomics? Latin Ameri- certain threshold of density; Libertun de Guren and Guer- ca’s middle class grew by about 50 percent during the rero Compéan 2015). boom years of 2003-09 (Ferreira and others 2012) reach- ing about 30 percent of the population just as the share But housing, urban, and infrastructure policies are en- of the population living in poverty fell by an equivalent couraging Latin America cities to expand at the periph- amount. The vast majority of this middle class, as well as ery, with serious implications for infrastructure service those approaching middle class status, have access to costs. The region’s built-up areas have expanded rapid- electricity. Most also have access to water, while relative- ly in recent decades. If expansion patterns continue un- ly few have access to sanitation; figure 37). Most reside changed, built-up urban areas will double in the region in urban areas where they presumably have access to by 2035. Density has declined in Santiago, Montevideo, paved roads and some type of public transportation. La Paz, Buenos Aires, and Brasilia. This increase in sprawl is largely policy-driven: in Buenos Aires, for example, trans- But this growing middle and near-middle class is far port, land use, and housing policies are driving factors of from being a saturated market in terms of consumer du- low density growth in the periphery (Inostroza, Baur, and rables. A small share of the upper income deciles own Csaplovics 2010). And while in Mexico only 14 percent of the full suite of consumer durables (refrigerators, wash- cities over 100,000 inhabitants “sprawled” between 1990 ing machines, air conditioners, computers, cars) that and 2010, most still have relatively low population den- characterize the middle classes in high-income coun- sities, with immediate implication for infrastructure in- tries and drives their energy consumption and demand vestment needs. The World Bank’s Mexico Urbanization for infrastructure services: on average, some 90 percent Review found that denser urbanization would reduce own refrigerators, though that share drops rapidly with infrastructure investment and maintenance costs by 41 other durables (70-80 percent for washing machines, percent in Merida and 67 percent in Los Cabos compared and 40 percent for cars) and for the near-middle class to business as usual urbanization (Kim and Zangerling decile (figure 38). 2016). Potentially facilitating access to durables is the boom in In addition, some 25 percent of Latin America’s city dwell- consumer credit that occurred in the last decade in Lat- ers are believed to be living in slums (UN Habitat 2013), in America, increasing from 9 percent to 20 percent of many in areas not suitable for residential construction. total credit. This boom in consumer credit, which has Slums are often located in flood-prone or environmen- led policy makers to complain that the region’s financial tally protected areas that put slum dwellers at high risk of system has been concentrating too much on financ- natural disasters or create hazards for others (for example, ing consumption at the expense of production (de La when slums are in areas that threaten urban watersheds, Torre, Ize, Schmukler 2012), is expected to continue due as in Curitiba). Combined with disorderly and dense to a combination of increased access to banking and 14 The urbanization part of this section is from Marulanda and others (2015) while the socioeconomic part is from Fay and Straub (2016), both background papers for this report. 62  ood, bad, and worst:  FIGURE 37: G almost all but  irst the fridge, then the washing FIGURE 38: F the poorest consumers in Latin America have the order of acquisition of machine, then the car:  access to electricity, access to water is less consumer durables in Latin America universal, and access to sanitation is low, even (Percent) among the middle class (Percent) Refrigerator ownership by income decile Electricity access by income decile 100 100 80 80 60 60 40 40 20 20 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Water access by income decile Car ownership by income decile 100 100 80 80 60 60 40 40 20 20 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Sanitation access by income decile Washing machines ownership by income decile 100 100 80 80 60 60 40 40 20 20 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Worst performing country Best performing country Worst performing country Best performing country Average Average Source: Fay and Straub 2016. Source: Fay and Straub 2016. Note: Deciles are defined region wide, so the figure shows how Latin Note: see previous figure. Americans earning similar incomes (within a decile) may experience wide variation in access to services, depending on the country where they live. For example, the average rate of access to water among households in Latin America’s poorest decile varies from 20% in the worst-performing country to given its greater accessibility to lower-middle-class and 100% in the best. poorer household. By one account, Brazilian families own some 181 million store cards, while Mexico’s have 9 million (Capizzani, Ramirez Huerta, and Rocha e Oliveira rapid expansion of credit cards and store cards. Three Lat- 2012). in American countries (Colombia, Argentina, and Chile) are ranked in the top 10 for the fastest growth in card The combination of increased incomes and rising ac- lending debt since 2008. Store cards (retailer credit) are cess to credit could significantly affect overall ener- a relatively new phenomenon that is taking off quickly gy demand. Consumer durables, and hence energy 63 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE consumption, do not increase one-to-one with income. a way that may not be sufficiently taken into account in Instead, it is generally believed that at low levels of in- traditional energy forecasts. come, increases in income have limited impact as cred- it-constrained households use the extra cash on basics. At the same time, most remaining basic access chal- But as income rises, purchases of durables increase rap- lenges in water and electricity are now concentrated idly—especially if credit constraints are removed (Ger- in the poorest income decile. Chances are that these tler and others 2016). Data from Latin America suggest households are located in hard to reach areas—remote that this is true for the poorest households, which exhib- rural locations or difficult urban sites. They are for the it low income elasticity for consumer durables, but that most part very poor, so more likely to face affordability consumer behavior varies both across income groups challenges whether for connections or for monthly ser- and type of durables. Still, growth in energy demand vice. Thus, the implication is that the remaining access from first-time purchases of consumer durables could challenge could be more challenging than it has been have significant impacts on overall energy demand in in the past. 64 Part III. The road ahead: spending better to meet “real” infrastructure needs P art I started by discussing Latin America’s infra- last-mile challenge as being the improvement in the elas- structure investment needs. Needs are usually de- ticity that could be achieved through additional or more fined based on the historical relationship between efficient investment. For example, improving spending to income and infrastructure growth. So let us look at increase the income elasticity to 0.2 (that is, above the the region’s road ahead if countries maintain the current regional best performer) would allow the nine historical link between infrastructure access and income best-performing countries to connect all the households growth, assuming a high-growth scenario similar to the in the first decile to water in less than 21 years. An even 2002-12 one—but go a step further, using decile-specific higher elasticity of 0.5 would allow all countries except El access rates to see how different income groups are likely Salvador to complete that task in 16 years or less. to fare given that in most countries public policies favor richer households. So what to do? Are we back to the starting point of the report, with estimates of investment needs that argue The results are sobering. If the road ahead is anything like for Latin America to spend a lot more than it does? More the one behind, it will take the region between 10 and money would make things easier. But is that an option? 50 years to achieve universal coverage in electricity, up to 90 years for water, and a staggering 200 years for sani- The discussion of fiscal space in part II makes it abun- tation (Fay and Straub 2016). The value of the elasticity of dantly clear that governments in the region have limit- access with respect to income plays a crucial role here. ed room to increase public investments. Over the last 16 Consider the three countries with about 50 percent cov- years total public investment has been 3.1-4.6 percent of erage rates for water at the bottom of the income distri- GDP. Today it is about 3.4 percent. On average, only about bution—Bolivia, the Dominican Republic, and Peru. The a third of this public investment goes to infrastructure. fact that Peru is much better than Bolivia in translating So, going forward, it is unlikely that public investment in income growth into improved access (as shown by their infrastructure will be much above 1.0-1.5 percent of GDP. respective elasticities of 0.18 and 0.08) means that Peru is likely to achieve full coverage for the poorest decile in As for public-private partnerships (PPPs), the discussion in half the time (43 years as opposed to 96; admittedly still part I showed that while PPPs account for about 40 per- a very long time). cent of Latin America’s infrastructure investments, they depend heavily on government support: about a third But these are constrained elasticities, capturing a com- of their financing comes from public sources, and about bination of constrained demand (largely due to lack of half of all deals receive some type of government guar- income) and constrained supply (due to limited or inef- antee. In other words, constrained public finance also ficient public investments). It is probably more useful to means constrained private finance for infrastructure. And think of the potential margin of adjustment to tackle the though the region has had a few “extra-ordinary” years, 65 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE private participation in infrastructure has usually hovered and improve the access of public service providers to between 0.5 and 1 percent of GDP. commercial finance. That can be done by reducing the tab associated with investment spending “needs” by So at the most, the region is unlikely to see invest- focusing on priorities, managing demand, reducing ments in infrastructure from traditional sources exceed costs and improving public spending (and public en- 1.5-2.5 percent of GDP in the near future. This means terprise) efficiency more generally. Fortunately, as not- that the main way forward, at least in the short to me- ed throughout this report, Latin America has plenty of dium term, is to make scarce resources stretch further scope to do this. Focusing on priorities—setting the right goals is essential There is no such thing as an absolute investment need. times more, or about 0.25 percent of GDP. (This 0.25 per- The amount of money needed for Latin American infra- cent happens to be what Latin America currently spends structure depends on the goals that are set and on how on water and sanitation. As such, the region should be efficient investors are at achieving them. So, a sensible able to achieve ambitious goals without necessarily in- approach to defining infrastructure investment needs creasing capital spending.) must start by defining goals—in terms of the access, quality, affordability, sustainability, and inclusiveness of How the region goes about pursuing this goal will make infrastructure services that are needed to achieve the re- a big difference. There is lots of uncertainty in these kinds gion’s development ambition. of estimates—about population, level of urbanization, costs. And there are choices, some to do with the analy- As such, the received wisdom that Latin America “needs” sis (what discount rate to use?) and some to do with the to spend 4-5 percent of GDP on infrastructure is not help- strategy, notably the path chosen to get there. The 0.25 ful. In fact, it is outright distracting. Instead, countries percent of GDP assumes that about half of households should conduct their own analyses to define their goals will go straight to a higher level of service and half will and what it would take to achieve them. Unfortunately, first pass through basic water and unimproved sanitation. few countries in the region have developed clear goals— If all first go through the low-cost technology and later and even fewer have priced them or examined whether upgrade, overall costs would be higher. Thus analysis pro- they have the resources or institutions to develop strate- vides low and high estimates: a low discount rate and a gic plans, as opposed to wish lists. population that immediately obtains the higher level of services results in an estimated 0.1 percent of GDP a year Yet doing so would be highly instructive. Take water and “investment need”, while the higher discount rate and in- sanitation. A careful analysis of water and investment direct path to the SDG would raise the estimated cost to needs done at the World Bank (Hutton and Varughese some 0.4 percent of GDP. 2016) started by defining two possible goals. One was the Millennium Development Goal (MDG) of achieving Similarly, and as discussed in the section on climate universal access to basic water, sanitation, and hygiene change, investment “needs” for electricity depend on the by 2030; the other was the much more ambitious Sus- goals set and the strategy adopted to achieve them. Ac- tainable Development Goal (SDG) of achieving universal cording to modeling done for this study, a business as access to safely managed water and sanitation services by usual investment path that follows South America coun- 2030. tries’ master plans, with demand growing as in recent years and an unchanged climate, would entail annual in- The difference in ambition between these two goals has vestments of $23-24 billion a year (Broad, de Moura, and striking implications for their cost. Providing basic access Howells 2016). A scenario with much greater penetra- to everyone in the region would cost a mere 0.05 percent tion of renewables would be somewhat more expensive of GDP a year through 2030; achieving universal access at $25-26 billion per year. But an ambitious “disruptive to safely managed water and sanitation would cost five scenario” that adopts smart grids, smart metering (and 66 hence effective demand-side management), and ambi- on infrastructure as opposed to services. Good roads, tious penetration of renewables would cost $8-9 billion a well-maintained rail tracks, modern airports, and dredged year. This does not include the cost of demand-side man- waterways are important, but competitive, well-priced agement programs so the final tab would be higher— and well-regulated trucking, bus, rail and aviation com- but the point remains that how the region goes about panies and good multi-modal integration are at least as implementing its energy strategy will have a significant critical. Good transport infrastructure is expensive—so impact on how much is needed. a focus on the service side and on spending efficiency is essential. And as the section on spending efficiency Interestingly, these numbers are well within the range of showed, transport has massive potential for increased what the region has been spending on energy (0.75-1.0 efficiency. percent of GDP), even when adding the cost of expand- ing access to electricity and modern cooking fuels: ex- Infrastructure goals are best defined in terms of the ser- pressed as a percentage of South America’s current GDP, vice needed—say, mobility—rather than the input re- the disruptive scenario would amount to 0.2 percent of quired—say, kilometers of roads. This approach opens up GDP (not counting demand-side management invest- the possibility of managing demand to reduce the cost ments) and the business as usual scenario, 0.6 percent. of satisfying this demand. Thus, as discussed earlier, dens- In other words, an ambitious, clean, and equitable energy er cities enhance mobility at much lower cost. Rough goal, well-conceived and delivered, is achievable within estimates are that encouraging compact development the current spending envelope. instead of sprawl can reduce overall infrastructure con- struction and service costs by 10-40 percent (Marulanda That leaves transport, which is far more complicated. First, and others 2015). what goals should be set? There must be at least three: one for urban transport, one for rural transport, and one Much can be achieved through demand-side manage- for freight. But how to set them? Unlike for water, sani- ment. The discussion on investment “needs” in energy tation, electricity or modern cooking fuels, there can be shows that demand side management largely deter- no presumption that universal access should be the goal mines how much investment will really be needed—and for any of these. Nor is it even clear what universal access “negawatts” (energy saved from energy efficiency and would mean for urban, rural, or freight transport. Further, demand-side management efforts) are almost always there is no good database on any of these three dimen- cheaper than megawatts. And here pricing services ade- sions to allow us to estimate how far the region is from quately have a double-whammy effect: it encourages ef- whatever goals are set. ficient use by consumers (to buy more fuel-efficient cars, use public transportation, turn off lights, buy energy-effi- As such, every country will need to define its trans- cient appliances, fix leaking faucets, and so on), and gen- port goals and develop a master plan. The few master erates resources, thereby reducing the burden on taxpay- plans that exist tend to be wish lists and overly focused ers and scarce public funds. Improving utility performance, and focusing public and concessional finance where it is truly needed Pricing services appropriately and running utilities efficient- difficulties mobilizing commercial financing unless they im- ly also makes it possible for them to attract private financing. plement significant reforms to improve cost recovery. In 2016 only 20 percent of the Latin American water util- ities included in the International Benchmarking Network But if few can attract commercial financing now, the po- for Water and Sanitation Utilities (IBNET) database covered tential for improvement is significant. Using this same their operations and maintenance costs and generated dataset, the average level of reported non-revenue water enough of a surplus to mobilize commercial borrowing (as- (water that is delivered into a network but “lost” before it sumed to be cash revenues exceeding costs by at least 20 reaches a legitimate consumer) was 31 percent. But there percent). This means that 80 percent of utilities would have were significant variations in performance: the top 10 67 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  ith reasonable progress on better management, four times as many utilities could access FIGURE 39: W commercial financing (percent of utilities considered credit worthy) 90 80 70 60 50 40 30 20 10 0 Current level with cash Step 1 Collection rate Step 2 Non-labor cost Step 3 With reduction of Step 4 Increase cost recovery > 120% increased to 100% reduced by 15% current level of NRW to 25% revenue by 10% Source: Courtesy of William Kingdom and Alexander Danilenko (World Bank) based on IBNET data 2016. Note: NRW stands for non-revenue water, or water that is produced but “lost” before it reaches the consumer. percent of performers achieved non-revenue water levels could be maintained or increased over the long term of 15 percent, while the level among the lowest decile was through good policies, governance, and incentives, the 57 percent. In terms of collection efficiency, the average ability to mobilize commercial financing starts to look like for the dataset was 75 percent of the amount billed, while a distinct possibility. And this could well lead to a virtuous the average in the top decile was 100 percent and in the circle: the lure of external financing—and the scrutiny of lowest decile, 46 percent. Assuming that utilities could external investors—would motivate utilities to be effi- achieve the performance of utilities in the top decile in cient, and the additional financing would enable them to terms of collection efficiency (Step 1 in figure 39), that invest and provide service good enough that consumers modest non-labor efficiency gains (predominantly ener- would be willing to pay for. gy) of 15 percent can be made (Step 2), that leakage lev- els could be reduced to 25% (realistic and less than the top decile) and that water saved from leaks can be sold Commercial financing and the importance of at the prevailing average tariff (Step 3) then 65 percent of making judicious use of public resources the region’s utilities in the IBNET sample could create suf- ficient surplus to mobilize commercial borrowing.15 Increased reliance on commercial financing, where pos- sible and appropriate, along with the judicious use of The implication is that reasonable progress in efficient public and concessional resources, is one of the essential management could more than triple the number of wa- ways through which Latin American countries can “spend ter utilities with potential access to commercial financ- better.” To do so, the World Bank is recommending a sim- ing. In this simplified model (up to Step 3), the financial ple sequential decision-making framework (figure 40). improvements are achieved at current tariff levels. If the utility can enhance its revenue stream from tariffs, taxes The starting point of this approach is that any investment or a mix of the two (Step 4), then 82% of utilities would project or program that can be financed on commercial achieve the 120% cost recovery level. Mobilizing funds terms while remaining affordable and offering value for up front to deliver such efficiency gains would be need- money, should be.16 Where commercial financing is not ed, as would capacity building. But if these cash surpluses cost effective or viable due to perceived risks or market 15 This discussion was developed based on Leigland and others (2016) 16 “Commercial” capital or financing is meant here as capital or financing that subjects the borrower (e.g. a utility) to the discipline of the market. It includes project and corporate finance, commercial bank financing, project and corporate bond issuance, and private equity, for example through stock exchange listings or direct investment. It would preclude financing from state-owned commercial banks, national development banks, or sovereign wealth funds investing in domestic jurisdictions (but not abroad) as well as MDB or DFI financing where there is an explicit or inferred sovereign counter-guarantee. 68  decision-making framework to ensure the judicious use of scarce public and concessional FIGURE 40: A finance 1 Commercial Financing Can commercial nancing be cost-e ectively mobilized for sustainable investment? if not... 2 Upstream Reforms & Market • Country and Sector Policies Can upstream reforms be put in place to • Regulations and Pricing address market failures? If not... • Institutions and Capacity 3 Public and Concessional Resources for Risk instruments and Credit Enhancements • Guarantees Can risk instruments & credit enhancements • First Loss cost-e ectively cover remaining risk? If not... 4 Public and Concessional Financing, including Sub-Sovereign • Public nance (incl. national development: banks and domestic SWF) Can development objectives be • MDBs and DFIs resolved with scarce public nancing? Source: World Bank 2017. failures, efforts should focus on addressing these market are more likely to accept charges for good quality ser- failures through upstream reforms to strengthen country vices and rates perceived to be fair. and sector policies, regulations, and institutions or target- ed public interventions (e.g. targeted subsidies or com- Equity and poverty concerns are not at odds with reliance plementary public investments such as transmission lines on commercial financing, even as they are often invoked or where appropriate). Where risks remain high and raise in arguments against full cost recovery for basic services the cost of commercial capital beyond that afforded by such as water and sanitation, electricity, public transport, project or corporate revenue generation, the possibility and modern cooking fuels. The needs of the poor are in of cost-effectively reducing risks with risk-sharing instru- fact typically best served by a combination of cost recov- ments backed by public or concessional finance that low- ery tariffs and targeted subsidies and payments schemes er the cost of commercial capital. If commercial financing adapted to the needs of the poor. Most of the wealthier is still not viable or cost-effective, then public and con- Latin American countries have well-developed social reg- cessional resources are the likely solution. istries and safety nets (e.g. Brazil, Colombia, Chile, Mexico) but the targeting is likely to be a challenge for countries Importantly, this framework can only be applied to ser- without. Still, efforts at targeting the poor are likely to help vices that can be charged to users, as user fees are what them more than underfunded, low-cost recovery utilities creates the basis for commercial financing options be- unable to expand coverage or provide quality service. yond the use of general taxes.17 A very rough estimate by the World Bank estimates that perhaps up to 50 percent Nevertheless, it may be difficult to reconcile the higher of infrastructure investment needs could theoretically be return expectations of commercial financing require- financed on a commercial basis, based on the feasibility ments with affordability concerns. This is particularly true and desirability of different subsectors to generate user in island states where the cost of infrastructure service fees. Feasibility depends on the ability to tie a service provision such as power generation may already be high to an individual user (easier for electricity than for rural for reasons of economic geography. It is also an issue for roads) while desirability will vary depending on external- countries and sectors with low access rates where the ities, the need for demand-side management and social consumer base is too small to bear the cost of service or political economy considerations. And of course, will- expansion, hence allow for full cost pricing (i.e. average ingness to pay will also matter for cost recovery—users cost pricing). Fortunately, this is only a challenge in a few 17 In addition, there may be potential to capture the value created through infrastructure investments in less traditional ways (land-value capture, congestion charging, parking fees) or through the commercial exploitation of infrastructure assets (advertising, real estate). 69 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE countries in the region (e.g. Haiti) and sectors with low such as ownership structure, regulatory governance and coverage (wastewater treatment and modern cooking corporate governance be addressed strategically, not in fuels). In such cases, the solution typically involves a blend isolation. Private management or ownership of a utility is of financing tools—commercial, public, and concession- no silver bullet—private participation is unlikely to yield ary—to match the conditions that the investments face. the desired efficiency gains in the absence of some type of market test—and Latin America is rife with examples Applying this framework to judiciously deploy scarce pub- of costly and failed PPPs. Similar experiences have led lic and concessional resources and ensuring commercial OECD countries to adopt a number of principles aiming financing provides value for money is not always simple to ensure affordability, value for money, and transparency however. It implies carefully weighing the trade-offs be- (box 9). 19 tween the financing of infrastructure on the back of gov- ernment taxes versus commercial financing backed by A companion report on PPPs in Latin America (Garcia-Kil- the securitization of user fees. While public and conces- roy and Rudolph 2017) finds that the region has signifi- sional financing appears cheaper than commercial sourc- cant potential to increase PPPs, but insists on the fact that es of finance, it is drawn from a pool of scarce resources PPPs are not a silver bullet. Even in advanced economies whose opportunity cost should be carefully considered with successful programs, PPPs seldom exceed 10-15 with each investment decision. As to commercial capital percent of total infrastructure finance. And PPPs should (which includes the cost of any guarantee, off-take agree- not be used as an off-balance sheet financing instrument ments, and any government-backed credit agreement), to bypass fiscal constraints.  Instead, the rationale for PPPs its higher cost must be weighed against potential effi- should be to achieve value for money for the govern- ciency gains in the provision of the asset. ment.  Finally, effective PPPs require sophisticated insti- tutions to implement them as well as a minimum level of financial sector development. Corporatization and the importance of improving The report argues that achieving Latin America’s poten- the performance of utilities—public and private tial in terms of PPPs will require tackling the following So how to obtain such efficiency improvements? Some challenges:  countries and sectors have turned to the private sector for service delivery, with the result that the number of ››Often weak underpinnings. This includes: (a) PPP connections served via PPIs increased from 11 percent frameworks, where they exist, are still facing import- in 1995 to around 60 percent by 2006 for electricity (but ant challenges on issues such as quality project se- are still only 8 percent for water).18 A review of more than lection and structuring, risk allocation between the 250 electricity distribution companies and more than public and the private sector, and procurement pro- 1700 water and sanitation utilities in Latin America found cedures; (b) project finance skills, which are at the that on average, private utilities outperform public ones, core of PPP financing, are underdeveloped among although there are good and bad performers in both banks, and financial regulations seldom acknowl- groups (the top 10 percent public utilities outperform edge the difference with corporate lending; (c) depth the average private utility) (Andres and others 2013). The and sophistication of local financial markets (banks or conclusion of the review was that, when carefully designed capital markets), whose size and sophistication are in- and implemented, private participation in service provi- sufficient to address financing gaps or facilitate ad- sion improves sector performance—specifically labor equate risk sharing between government and other productivity, efficiency, and service quality. stakeholders. The caveat about careful design and implementation is ››Low participation of foreign sponsors and financiers — important. As Andres and others (2013) point out improv- who can provide volume, competition, and knowl- ing sector performance demands that key determinants edge transfer. 18 Data from Andres and others 2013 for electricity and IBNET for water. In 2014 the private sector accounted for 65 percent of electricity generation according to industry data. 19 These principles aim to establish a clear, predictable and legitimate institutional framework; ground the selection of Public-Private Partnerships in “value for money”, and minimize minimize fiscal risks. For more details, see http://www.oecd.org/gov/budgeting/oecd-principles-for-public-governance-of-public-private-partner- ships.htm 70  he Political Economy of Reform: Conditions for Change Box 9: T A review of successful water sector reforms found that building reform momentum and sustaining service required action on three inter-related fronts: ››Build and strengthen internal capacity and culture: The technocratic and managerial skills that helped launch the reforms were then used to build strong staff and managers. Other managerial techniques – like performance-based pay, inclusive corporate strategic planning, and general transparency – also helped build and strengthen perfor- mance-based cultures. Utility leaders built internal capabilities and cultures that make the utility successful and lat- er on helps to sustain reforms. A professional culture provides a barrier against predation on the utility since, being anathema to professional culture, the organization itself will fight against it. Grooming future leaders and promoting training and development were particularly important in the cases studied. This is an indication that leadership is a critical variable, not solely at the outset of reforms, but throughout the life of a successful utility, particularly in a pre- carious macro political economy context. By creating a deep management bench, a utility reduces the risk that losing a leader will undo its success. A performance-based corporate culture, with an emphasis on transparency, also promoted reform momentum and reform resilience. A performance-based pay system – based on annual reviews against perfor- mance indicators derived from the utility’s overall mission, and well-defined job descriptions for staff – is important in many of the case utilities. Many utilities also involve all staff in strategic planning. This sets the scene for establishing ownership of the institution, with staff at all levels understanding, operationalizing and able to articulate the approach taken by the management. ››Forging and embedding alliances with external stakeholders: Alliances with customers, the Government, development partners and other stakeholders were used to build momentum for reform, and help sustain it. Utility leaders demon- strated a high degree of political savvy, understanding of the political and wider sociocultural context, and the ability to navigate competing interests in an unsettled governance and institutional environment. Alliances constructed in the mo- mentum phase were instrumental in maintaining reforms later. ››Creating and strengthening formal rules and structures (institutions): In isolation, formal rules and structures are an inadequate guarantee of sustained success. ‘Independent’ boards, for example, are routinely replaced by politicians, some- times in breach of company law. ‘Independent’ regulators may be reluctant to approve tariff increases. But when coupled with professionally capable organizations embedded in a web of stakeholder alliances, formal regulatory and governance structures can contribute to the sustainability of reforms. Source: Heymans and others 2016. ››A lack of consensus among public and private sec- But whether utilities and service providers are public, pri- tor stakeholders around PPP frameworks and pro- vate, or a mix, they need to be run as corporatized entities grams—this is critical to reach the necessary trust and regulated by independent regulatory agencies that among the different parties. are transparent, accountable, and free of political interfer- ence. Key findings and conclusions of the review of Lat- The report offers a useful reminder that PPP programs are in America’s electricity distribution companies and water not right for all countries. Only large and medium-sized and sanitation utilities (Andres and others 2013) include countries with a minimum threshold of financial devel- the following: opment would be able to afford PPP programs in local currency. For smaller countries, PPPs could be a relevant ››The presence of a regulatory agency significantly im- source of financing for select signature projects with rev- proves sector performance, raises labor productivity enues in dollars, but most would need concessional fi- and cost-recovery and reduces operational expenses nancing or guarantees from multilateral development and distribution losses. finance institutions (DFIs) to access international capital. The experience of the region has been of a small pres- ››Improving sector performance demands that key ence of multilateral DFIs. As to domestic DFIs who have determinants—such as ownership structure, regula- been active financiers of infrastructure, they have tended tory governance and corporate governance—be ad- to prioritize direct lending instead of assuming a catalytic dressed strategically, not in isolation. Private manage- role to mobilize private financing. ment or ownership of a utility is no silver bullet. 71 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE ››For SOEs, strong accountability is central to improving boost to capital stock as increasing government invest- performance. Corporate governance standards (which ment by 5 percentage points of GDP in emerging econo- are the norm for privatized or privately managed mies and by 14 percentage points of GDP in low-income utilities), performance orientation, and professional countries (IMF 2014). These types of estimates are fairly management were found to be the most important heroic, but they do suggest that public investment effi- determinants of performance. More generally, best ciency gains can have more than micro-level effects. In practices for any utility or corporatized service provider addition, more efficient countries are likely to be more includes an independent performance-driven board, a attractive to investors interested in public-private part- corporate structure that prevents political interference, nerships. In fact, evidence suggests that PPI flows are a professional staff, clear disclosure policies, and mech- sensitive to the investment climate—more so than direct anisms to evaluate and reward performance. foreign investment in general (Araya and others 2013). ››Reform success requires addressing the technical and There are many ways to improve public investment effi- financial dimensions of the reform (concession laws ciency. In order of importance, these include: and contracts that clearly assign and mitigate risks, discourage opportunistic bidding and renegotiations, Build effective public sector institutions with clear and are embedded in transparent and predictable mandates. A critical factor in a country’s ability to spend regulatory systems that promote efficiency and ac- efficiently on well-identified priorities is the strength of countability and address social concerns). its institutions (Andres, Biller, and Dappe 2016). Where de- cision-makers ignore the results of prioritization exercis- ››Reform success also necessitates addressing legit- es, or where regulatory institutions cannot weigh in with imate social concerns and perceptions. Reversal of checks and balances to block questionable projects, even policies can be triggered by popular discontent asso- the best methodologies will be of little use. So a priority ciated with a failure to help those most affected (for for infrastructure in the region is to strengthen the insti- example through targeted subsidies to protect the tutions and systems relevant to transport, electricity, and poor against necessary tariff hikes or job search assis- water and sanitation investments as follows: tance for redundant employees of a utility) but also by a failure to communicate success. ››Central of Government institutions, such as a minis- try of finance or planning, need to take responsibility There are many examples of utilities and other service pro- for facilitating an overall infrastructure strategy that is viders that have turned around and become good per- based on sound analysis, prioritization, and political formers—but this often requires a combination of a catalyt- mandate. ic event that creates a space for reform, and savvy political and technical leaders that seize the opportunity to formu- ››Sector institutions must have the capacity to identify, late a mutually beneficial partnership. Together they must prioritize and implement projects according to politi- help shape networks and alliances for change and start to cally endorsed sector priorities and visions. embed the reform legacy. But success is only possible if the balance of political economy pay-offs remains in favor ››Major administrative systems such as planning proce- of reform, and, once achieved, in favor of sustained good dures, budgeting rules, treasury operations, procure- service, even as the attractions of predation on the utility ment procedures, human resources and safeguard increase. Experience points to steps reformers can take to systems must be aligned with the needs of capital in- ensure lasting success (box 9). And while these findings vestment projects (rather than being narrowly focused were derived in the context of an analysis of utilities, they on controlling cash expenditures) to ensure efficiency, apply equally well to any infrastructure service provider. effectiveness and transparency around investment. Such public sector capacities are critical to enable the Improving public investment management and development of an overall national investment plan, a politically sanctioned short list of projects, and an overall spending efficiency policy on how to use private sector expertise and inno- According to the IMF, reducing inefficiencies in develop- vation in the development of the national infrastructure ing countries’ public investment (including but not lim- framework. Infrastructure Australia offers an interesting ited to infrastructure) by 2030 would provide the same example of such an approach (box 10). 72  ssessing needs and proposing a pipeline – the case of Infrastructure Australia Box 10: A In 2008, in response to inadequate coordination in infrastructure planning, the Australian government created Infrastructure Australia (IA), an independent statutory body, whose role is to provide independent research and advice to all levels of govern- ment as well as to investors and owners of infrastructure on the projects and reforms Australia needs to fill the infrastructure gap. Its responsibilities include auditing the country’s infrastructure needs and performance, and developing a rolling 15-year infrastructure plan that identifies Australia’s national and state-level infrastructure priorities in transport, including energy, telecommunications and water. Implementation of the plan requires that the federal government and/or the sub-national governments subsequently make it official policy. Infrastructure Australia is tasked with identifying Australia’s long-term infrastructure needs through an infrastructure audit. This audit is based on an analysis of drivers of infrastructure demand such as population and economic growth, looks out to 2031, and has served as a key input to Australia’s current Infrastructure Plan which contains a package of reforms regarding how infrastructure is financed, delivered and used. The strategy developed in the Plan considers a wide range of options and instruments, including institutional and regulatory reforms, as well as investments. The longer-term view helped move beyond the more common project-centered approach and enabled a more integrated view of how infrastructure across various sectors can contribute to the country’s development. Infrastructure Australia then developed an Infrastructure Priority List of initiatives and projects in collaboration with state and territory governments and industry that considers three dimensions: strategic fit, economic, social and environmental value, and deliverability. The plan identified a number of complex reforms that could deliver significant productivity benefits country-wide, but are politically challenging. To overcome this Infrastructure Australia proposed a three-tiered approach that called on the national government to leverage its investment in infrastructure to encourage state and local governments to implement the reforms identified in the Plan. In addition, Infrastructure Australia emphasized the need for need for early community engagement— which can improve the quality of planning and reduce opposition. Key strengths of the Australian system include: ››Insulation from political pressures. As an independent body, Infrastructure Australia is, in principle, insulated from the political process, and can therefore assess infrastructure needs and develop recommendations on the basis of objective criteria. ››A structured approach. Infrastructure Australia’s applies a sequenced and structured approach to infrastructure planning by framing investment choices within a long-term assessment of needs (the Audit) and a considered evaluation of the var- ious options for addressing those needs that is guided by a set of long-term goals (the Plan). ››An integrated strategy. By considering all infrastructure sectors within a single plan that is guided by a set of long-term ambitions, the Infrastructure Plan provides for an integrated perspective of infrastructure. Such a holistic and integrated approach encourages greater alignment across sectors and investments, and improves the scope for generating synergies. Source: OECD 2017. ››Include multiyear budgeting in project selection and can hamper ‘pork barrel’ politics. Audit and evalua- budgeting. This ensures resources are available for op- tion should include ex-post evaluation and external erations and maintenance, donor-funded projects are audits, as well as, importantly, the existence of an as- included in the budget, and that the public has ac- set register or inventory of public property. Findings cess to key fiscal information. should be made public. ››Allow for project scrutiny by the legislature and public. ››Ensure budgetary rules work to strengthen good project Greater transparency regarding the key feasibility, execution, rather than just control spending. Govern- cost/benefit analysis and tendering and contract doc- ment agencies should be allowed to carry over proj- uments, as well as the use of independent evaluators, ect funds from year to year. This would help reduce 73 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE pressures to inappropriately expedite procurement institutions and frameworks—notably budgeting and and address the fact that the dry season, which is the procurement systems—should enable the region to optimal time for most public works, typically spans substantially stretch the resources it already allocates to two fiscal years. Budgeting funds to investment pro- infrastructure. grams in large blocks also strengthens good capital management as it allows agencies to switch funds This report provides a framework for countries wishing to from projects that are going slowly to those that are improve their infrastructure performance despite fiscal con- ahead of schedule or are more pressing. Portfolio straints. It does so by recommending a careful definition of budgeting also prevents the private contractors from priorities, highlighting steps that can be taken to improve knowing what the government expects to pay for a efficiency, and noting the need to carefully chose between contract, thereby turning it into a bid floor. what should be funded by users versus tax payers. But this report could usefully be complemented by more in depth, ››Use objective criteria and analysis to help decision-mak- country or regional, analysis of the following issues: ers prioritize among the multiplicity of projects across all sectors. Unfortunately, Latin America’s govern- ››Quality of service—as Part I of this report makes clear, ments seldom have the resources and data to do a limited information is available on the quality of service, full-fledged analysis, but multicriteria analysis can be raising the question of whether infrastructure access used with some basic project appraisal data to rank a levels paint a full picture of how well the region is doing. multitude of projects (See Marcelo and others 2016, and Andres, Biller, and Dappe 2015 for methodolo- ››How to manage the region’s sprawling urban de- gies). This is likely to work best where decision crite- velopment—a forthcoming World Bank report on ria, weighting, and sensitivity analysis is decided in urbanization will shed some light on this evolution advance, and the analysis is made public and open based on new data and analysis. to third-party review (Marcelo and others 2016). Ad- equate social and environmental risk is also critical to ››Subsidies—the limited data available suggests that managing costs, ensuring popular support, and the subsidies, often poorly targeted, remain substantial in likelihood of a project’s success. a number of countries, and may be larger than what the region invests in infrastructure. In sum, ››Public spending efficiency and fiscal space available for capital spending—this report is limited to some- Good infrastructure is key to Latin America’s ambitions what anectodal evidence regarding public spending in terms of growth, inclusiveness and environmental efficiency and falls far short of a full analysis of fiscal sustainability. But money is not necessarily the missing space and of countries’ ability to rebalance public ingredient for the region to achieve its ambitions. More spending in favor of capital spending. focused goals and efficient strategies can substantially reduce financing needs. In addition, upstream reforms Public expenditure surveys focused on infrastructure will enable Latin America to both improve spending ef- would be extremely valuable tools in helping govern- ficiency and attract private financing on better terms— ments understand where institutional weaknesses lie and whether through PPPs or commercial borrowing by pub- the extent to which efficiency gains could help achieve lic enterprises. And efforts to improve public investment infrastructure goals. 74 References Abiko, Alex, Luiz Reynaldo de Azevedo Cardoso, Ricardo Rinaldelli, Climate Change Economics and Policy Grantham Research Insti- and Heitor Cesar Riogi Haga. 2007. “Basic Costs of Slum Upgrad- tute on Climate Change and the Environment. London, UK. http:// ing in Brazil.” Global Urban Development. www.lse.ac.uk/GranthamInstitute/wp-content/uploads/2014/03/ ADB (Asian Development Bank). 2017. Meeting Asia’s Infrastructure PP-infrastructure-for-development-meeting-the-challenge.pdf Needs. Manila. Boeing World Air Cargo Forecast Team. 2015. World Air Cargo Fore- Ali, Rubaba, A. Federico Barra, Claudia Berg, Richard Damania, John cast (WACF) 2014-2015. Seattle, WA. http://www.boeing.com/ Nash, and Jason Russ. 2015. Highways to Success or Byways to commercial/market/cargo-forecast/. Waste: Estimating the Economic Benefits of Roads in Africa. World Borja-Vega, Christian, Gustavo Perochena and Virginia Zuilu. 2015. Bank, Washington, D.C. http://elibrary.worldbank.org/doi/ “Infrastructure for Sharing Prosperity in Latin American and the full/10.1596/978-1-4648-0654-4_ch5. Caribbean Region (LAC) Water and Sanitation.” World Bank. Andres, Luis, Dan Biller, and Matias Herrera Dappe. 2016. “A Method- Briceño-Garmendia, C., H. Moroz, and J. Rozenberg. 2015. “Road ological Framework for Prioritizing Infrastructure Investment,” Networks, Accessibility, and Resilience: The Cases of Colombia, Journal of Infrastructure Development; 8(2)1-17 (2016) Ecuador, and Peru.” World Bank, Washington, D.C. Andres, Luis, Jordan Schwartz, and Jose L. Guasch. 2013. Uncovering Broad, Oliver, Gustavo de Moura, and Mark Howells. 2016. “Invest- the Drivers of Utility Performance: Lessons from Latin America and ment Needs in the Energy Sector along Multiple Socio-Eco- the Caribbean on the Role of the Private Sector, Regulation, and nomic Pathways”. Background paper commissioned for this re- Governance in the Power, Water, and Telecommunication Sectors. port. KTH Royal Institute of Technology, Stockholm Directions in Development – Infrastructure Series. The World BRT Centre of Excellence, WRI Ross Center for Sustainable Cities, Bank, Washington, D.C. International Energy Agency and The Latin American Associa- Araya, Gonzalo, Jordan Schwartz, and Luis Alberto Andres. 2013. tion of Integrated Transport Systems and BRT. 2016. “Global BRT “The Effect of Country Risk and Conflict on Infrastructure PPPs.” Data.” Version 0.00. http://brtdata.org/. World Bank Policy Research Working Paper No. 6569. The World CAF Banco de Desarrollo de América Latina. 2015. La infraestructura en Bank, Washington, D.C. el desarrollo integral de América Latina. IDeAL 2014. Bogota. http:// Armendaris, Edna and Contreras, Eduardo. 2016. “El gasto de in- scioteca.caf.com/bitstream/handle/123456789/746/2014_ versión pública en América Latina: cuanto y cuan eficiente” pre- principal.pdf?sequence=1&isAllowed=y sentation done in Costa Rica, April 19, 2016 at the VI Conference CAF Banco de Desarrollo de América Latina. 2014. Encuesta de ho- of the Public Investment Management Systems of LAC. gares 2013. Principales resultados. Bogota. http://scioteca.caf. Balbotin, Patricio Rozas, and Liliana Salazar Arredondo. 2015. Vi- com/handle/123456789/409 olencia de género en el transporte público: Una regulación pen- CAF Banco de Desarrollo de América Latina. 2013. La infraestructura en diente. Recursos Naturales E Infraestructura 172. United Na- el desarrollo integral de América Latina. IDeAL 2013. Bogota. https:// tions, CEPAL, Santiago. http://repositorio.cepal.org/bitstream/ www.caf.com/_custom/static/ideal_2013/assets/book_1.pdf. handle/11362/38862/1/S1500-626_es.pdf CAF Banco de Desarrollo de América Latina. 2011. La infraestructu- Balcazar, Cecilia. 2012. “Resilient Infrastructure for Sustainable Ser- ra en el desarrollo integral de América Latina: IDeAL 2011. Bogota. vices—Latin America: Mainstreaming of Disaster Risk Manage- http://www19.iadb.org/intal/intalcdi/PE/2011/09525.pdf. ment in the Water Supply and Sanitation Sector”. Water and Cannock, Graham, Silva Yon, Jessica, Jara Trujillo, César S., O’Con- Sanitation Program. Washington, D.C., World Bank. nor La Rosa. Roberto, and Saavedra Bonifaz, Fernando. 2011. Banerjee, Sudeshna Ghosh; Moreno, Francisco Alejandro; Sinton, “Economic Impact of the 2007 Earthquake in Peru on the Jonathan Edwards; Primiani, Tanya and Seong, Joonkyung. 2017. Drinking Water and Sanitation Sector: Full Study.” World Bank, “Regulatory indicators for sustainable energy: a global score- Washington, D.C. https://www.wsp.org/sites/wsp.org/files/ card for policy makers”. Washington, D.C., World Bank. http:// publications/WSP-LAC-Economic-Impact-Earthquake-Full- documents.worldbank.org/curated/en/538181487106403375/ Study.pdf pdf/112828-REVISED-PUBLIC-RISE-2016-Report.pdf Capizzani, Mario, Felipe Javier Ramirez Huerta, and Paulo Rocha e Bhattacharya, Amar, Mattia Romani, and Nicholas Stern. 2012. In- Oliveira. 2012. “Consumer Credit in Latin America: Trends and frastructure for Development: Meeting the Challenge. Centre for Opportunities in Credit and Store Cards”. IESE Business School, 75 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE University of Navarra, Barcelona. http://www.iese.edu/re- Fay, Marianne, and Mary Morrison. 2007. Infrastructure in Latin Amer- search/pdfs/OP-0200-E.pdf. ica and the Caribbean: Recent Developments and Key Challenges. Connell, William. 2014. “The Economic Impact of Late Payments.” World Bank, Washington, D.C. Economic Papers 531, European Commission, Brussels. http:// Fay, Marianne, and Tito Yepes. 2003. Investing in Infrastructure: What Is ec.europa.eu/economy_finance/publications/economic_pa- Needed from 2000 to 2010? World Bank, Washington, D.C. http:// per/2014/pdf/ecp531_en.pdf, 10/19/2016 elibrary.worldbank.org/doi/pdf/10.1596/1813-9450-3102. Cubas, Diana, Maria Claudia Pachon, Tatiana Peralta, Xijie Lu, and Ferreira, Francisco H. G., Julian Messina, Jamele Rigolini, Luis-Felipe Steven Farji. 2015. Access to Transport Infrastructure and Services. López-Calva, Maria Ana Lugo, and Renos Vakis. 2012. Econom- Background Paper Commissioned for the Study. World Bank, ic Mobility and the Rise of the Latin American Middle Class. Latin Washington, D.C. America and Caribbean Studies. Washington D.C., World Bank. Dabla-Norris, Era, Jim Brumby, Annette Kyobe, Zac Mills, and Chris http://dx.doi.org/10.1596/978-0-8213-9634-6. Papageorgiou. 2012. “Investing in Public Investment: An Index Garcia-Kilroy, Catiana and Heinz Rudolph. Forthcoming. Infrastruc- of Public Investment Efficiency”. IMF Working Paper. Interna- ture Finance in Latin America and the Caribbean:Challenges and tional Monetary Fund, Washington, D.C. https://www.imf.org/ Opportunities. World Bank, Washington DC. external/pubs/ft/wp/2011/wp1137.pdf. Gertler, Paul J., Orie Shelef, Catherine D. Wolfram, and Alan Fuchs. de la Torre, Augusto, A. Ize, D. Lederman, F. Bennett, and M. Sas- 2016. “The Demand for Energy-Using Assets among the son. 2016. “The Big Switch in Latin America: Restoring Growth World’s Rising Middle Classes.” American Economic Review: 106: through Trade”. Chief Economist Office, Latin America and the 1366–1401. Caribbean, the World Bank. Geurs, Karst T., Wouter Boon, and Bert Van Wee. 2009.“Social Impacts of de La Torre, Augusto, Alain Ize, and Sergio L. Schmukler. Transport: Literature Review and the State of the Practice of Trans- 2012. Financial Development in Latin America and the Ca- port Appraisal in the Netherlands and the United Kingdom.”Trans- ribbean: The Road Ahead. World Bank, Washington, D.C. port Reviews, 29:1, 69-90, DOI:10.1080/01441640802130490. https://openknowledge.worldbank.org/bitstream/han- Gwilliams, K. 2012. “Practical Indicators for Comparing and Evaluat- dle/10986/2380/657440REPLACEM0110USE0SAME0BOX0IN- ing Subsidy Instruments.” Mimeo. FO.pdf?sequence=1&isAllowed=y. Hernandez Ore, Marco Antonio, Luis Alvaro Sanchez, Liliana D. Sou- de Lucena, Andre Frossard Pereira, Roberto Schaeffer, and Alexan- sa, and Leopoldo Tornarolli (forthcoming). Fiscal and Welfare Im- dre Salem Szklo. 2010. “Least-Cost Adaptation Options for Glob- pacts of Electricity Subsidies in Central America. Washington, D.C., al Climate Change Impacts on the Brazilian Electric Power Sys- The World Bank. tem.” Global Environmental Change 20 (2): 342–50. Heymans, Chris, Rolfe Eberhard, David Ehrhardt, and Shannon Riley. de Marucci, Silvia. 2012. “The Expansion of the Panama Canal and Its 2016. “Providing Water to Poor People in African Cities Effective- Impact on Global CO2 Emissions from Ships.” Maritime Policy & ly: Lessons from Utility Reforms. World Bank, Washington, DC. Management 39 (6): 603–20. https://openknowledge.worldbank.org/handle/10986/25115 Delgado, Ricardo, Angela Inés Cadena, Mónica Espinosa, Catalina Hidalgo, Dario. 2011. Transporte sostenible para América Latina: Peña, and Mateo Salazar. 2014. “A Case Study on Colombian Situación actual y perspectivas. Documento de Respaldo Foro de Mitigation Actions.” Climate and Development 6: 12–24. Transporte Sostenible FTS de América Latina 2011 Bogotá Junio Dulac, John. 2013. Infrastructure Requirements: Estimating Road and 22 - 24 de 2011. EMBARQ. Washington, D.C., World Resources Railway Infrastructure Capacity and Costs to 2050. International Institute Energy Agency, Paris. https://www.iea.org/publications/free- Hunja, Robert. 2015. E-Procurement: Opportunities & Challenges. Eu- publications/publication/TransportInfrastructureInsights_FI- ropean Commission. http://ec.europa.eu/geninfo/query/resul- NAL_WEB.pdf. taction.jsp?swlang=en&QueryText=robert+hunja+procure- Fay, Marianne. 2000. Financing the Future: Infrastructure Needs in Lat- ment&sbtSearch=Search in America 2000 - 2005. World Bank, Washington, D.C. http://eli- Hutton, Guy, and Mili Varughese. 2016. The Costs of Meeting the 2030 brary.worldbank.org/doi/pdf/10.1596/1813-9450-2545. Sustainable Development Goal Targets on Drinking Water, Sanita- Fay, Marianne, and Stephane Straub. 2016. Rising Incomes and the tion, and Hygiene. Water and Sanitation Program Technical Pa- Demand for Infrastructure of Latin American Households. Back- per. Washington, D.C., World Bank. ground paper commissioned for this report. World Bank, Wash- IDB (Inter-American Development Bank). 2015. Freight Transport ington, D.C. and Logistics Yearbook 2014. Washington, D.C. https://publica- Fay, Marianne, Stephane Hallegatte, Adrien Vogt-Schilb, Julie tions.iadb.org/handle/11319/6885. Rozenberg, Ulf Narloch, and Tom Kerr. 2015. “Decarbonizing IHME (Institute for Health Metrics and Evaluation) and World Bank. Development: Three Steps to a Zero-Carbon Future.” World 2013. The Global Burden of Disease: Generating Evidence, Guid- Bank, Washington, D.C. https://openknowledge.worldbank. ing Policy—Latin America and Caribbean Regional Edition. Se- org/handle/10986/21842. attle, WA. http://www.healthdata.org/sites/default/files/files/ 76 policy_report/2013/WB_LatinAmericaCaribbean/IHME_GBD_ Kim, Yoonhee and Bontje Zangerling (Ed). 2016. Mexico Urbaniza- WorldBank_LatinAmericaCaribbean_FullReport.pdf. tion Review: Managing Spatial Growth for Productive and Livable IMF (International Monetary Fund). 2014. World Economic Outlook: Cities in Mexico. Washington, D.C. Legacies, Clouds, Uncertainties. Washington, D.C. Kohli, Harpaul Alberto, and Phillip Basil. 2011. “Requirements for Inostroza, Luis, Baur, Rolf, and Csaplovics, Elmar. 2010. Urban Sprawl Infrastructure Investment in Latin America under Alternate and Fragmentation in Latin America: A Comparison with Europe- Growth Scenarios, 2011–2040.” Global Journal of Emerging Mar- an Cities. The Myth of the Diffuse Latin American City. Lincoln In- ket Economies 3 (1): 59–110. stitute of Land Policy Working Paper. Lincoln Institute of Land Laurance, William F., Miriam Goosem, and G.W. Laurance. 2009. “Im- Policy. Cambridge, MA. http://docplayer.net/2719326-Urban- pacts of Roads and Linear Clearings on Tropical Forests.” Trends sprawl-and-fragmentation-in-latin-america-a-comparison- in Ecology and Evolution 24 (12): 659–69. with-european-cities-the-myth-of-the-diffuse-latin-american- Leigland, James, Sophie Trémolet, and John Ikeda. 2016. Achiev- city.html. ing Universal Access to Water and Sanitation by 2030: The Role of IPCC. 2014. Climate Change 2014: Impacts, Adaptation, and Vulner- Blended Finance. Water Global Practice: Discussion Paper. World ability. Part A: Global and Sectoral Aspects. Contribution of Work- Bank, Washington, D.C. http://documents.worldbank.org/cu- ing Group II to the Fifth Assessment Report of the Intergovernmen- rated/en/978521472029369304/pdf/107971-WP-P159188- tal Panel on Climate Change. Cambridge, U.K. and New York, NY: PUBLIC.pdf. Cambridge University Press. Lempert, Robert J., Steven W. Popper, David G. Groves, Nidhi Kalra, ITF (International Transport Federation). 2016. Capacity to Grow: Jordan R. Fischbach, Steven C. Bankes and Benjamin P. Bryant. Transport Infrastructure Needs for Future Trade Growth. Organisa- 2013. Making Good Decisions Without Predictions: Robust Deci- tion for Economic Co-operation and Development, Paris. http:// sion-making for Planning Under Deep Uncertainty. RAND Corpo- www.itf-oecd.org/sites/default/files/docs/future-growth-trans- ration, Washington, D.C. http://www.rand.org/pubs/research_ port-infrastructure.pdf briefs/RB9701.html. ITF (International Transport Federation). 2015. Transport Outlook Libertun de Duren, Nora, and Roberto Guerrero Compéan. 2015. 2015: Chapter 4 Preview: Urban Passenger Transport Scenarios “Growing Resources for Growing Cities: Density and the Cost of for Latin America, China and India. Organisation for Econom- Municipal Public Services in Latin America.” Urban Studies, 1–26. ic Co-operation and Development, Paris. http://www.oecd-ili- Litman, Todd. 2015. “Analysis of Public Policies That Unintentionally brary.org/content/book/9789282107782-en. Encourage and Subsidize Urban Sprawl.” Victoria Transport Poli- Jaimurzina, Azhar, Gabriel Pérez Salas, and Ricardo J. Sánchez. 2015. cy Institute, supporting paper commissioned by LSE Cities at the Políticas de logística y movilidad para el desarrollo sostenible y la London School of Economics and Political Science, on behalf of integración regional. Recursos Naturales E Infraestructura 174. the Global Commission on the Economy and Climate. London, United Nations, CEPAL, Santiago UK. http://static.newclimateeconomy.report/wp-content/up- Jiron, Paula. 2013. Sustainable Urban Mobility in Latin America and loads/2015/03/public-policies-encourage-sprawl-nce-report. the Caribbean. Thematic study prepared for Global Report on pdf. Human Settlements 2013. UN Habitat. http://unhabitat.org/ Magrin, G.O., J.A. Marengo, J.-P. Boulanger, M.S. Buckeridge, E. Cas- wp-content/uploads/2013/06/GRHS.2013.Regional.Latin_. tellanos, G. Poveda, F.R. Scarano, and S. Vivuna. 2014. “Central America.and_.Caribbean.pdf. and South America.” In Climate Change 2014: Impacts, Adapta- Kalra, Nidhi, Stephane Hallegatte, Robert Lempert, Casey tion, and Vulnerability. Part B: Regional Aspects. Contribution of Brown, Adrian Fozzard, Stuart Gill, and Ankur Shah. 2014. Working Group II to the Fifth Assessment Report of the Intergov- “Agreeing on Robust Decisions: New Processes for Deci- ernmental Panel on Climate Change: 1499–1566. Cambridge, sion-making Under Deep Uncertainty.” Policy Research Work- U.K. and New York, NY: Cambridge University Press. ing Paper 6906. World Bank, Washington, D.C. http://docu- Marcelo, G. Darwin, Xavier C. Mandri-Perrott, Ruth S. House and Jor- ments.worldbank.org/curated/en/365031468338971343/ dan Z. Schwartz. 2016. “Prioritizing Infrastructure Investment: pdf/WPS6906.pdf A Framework for Government Decision Making”. Washington, Kalra, Nidhi Rajiv, David G. Groves, Laura Bonzanigo, Cayo Ra- D.C., World Bank. http://documents.worldbank.org/curated/ mos, Brandon Enrique Carter, Iván Rodriguez Cabanillas, and en/805021467996728921/pdf/WPS7674.pdf Edmundo Molina Perez. 2015. “Robust Decision-Making in Marulanda, Catalina, Beatriz Eraso, Emanuela Monteiro, and Pablo the Water Sector: A Strategy for Implementing Lima’s Long- Gluzmann. 2015. LAC “Infrastructure Gap—A Territorial Devel- Term Water Resources Master Plan.” WPS7439. World Bank, opment Perspective”. Background paper commissioned for this Washington, D.C. http://documents.worldbank.org/curated/ report. World Bank, Washington, D.C. en/2015/10/25153429/robust-decision-making-water-sec- Mehndiratta, Shomik, Rodriguez, Camila and Ochoa, Catalina. 2014. tor-strategy-implementing-lima%E2%80%99s-long-term-wa- Targeted Subsidies in Public Transport: Combining Affordability ter-resources-master-plan. with Economic Sustainability. World Bank, Washington, D.C. 77 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE Miralles-Wilhelm, Fernando. 2016. “Water Security and Infrastruc- Ruiz-Nunez, Fernanda, and Zichao Wei. 2015. “Infrastructure Invest- ture for Development in Latin American and the Caribbean ment Demands in Emerging Markets and Developing Economies”. 2050”. Background paper commissioned for this report. World Policy Research Working Paper 7414. World Bank, Washington, D.C. Bank, Washington, D.C. Serebrisky, Tomás. 2011. Airport Economics in Latin America and the Nolasco, Daniel. 2016. Report on Current Wastewater Treatment Caribbean: Benchmarking, Regulation, and Pricing. Directions in Trends in LAC. Nolasco Associates. Water, Environment and Busi- Development. World Bank, Washington, D.C. ness Consulting Services. Serebrisky, Tomás, Ancor Suárez-Alemán, Diego Margot, and Ma- OECD. 2017. “Infrastructure Governance in Chile”. OECD, Paris. ria Cecilia Ramirez. 2015. Financing Infrastructure in Latin Amer- OECD/ECLAC. 2012. “Latin American Economic Outlook 2012.” ica and the Caribbean: How, How Much and by Whom? IDB (In- doi:http://dx.doi.org/10.1787/leo-2012-en. ter-American Development Bank), Washington, D.C. https:// Osborne, Theresa, Maria Claudia Pachón, and Gonzalo Enrique publications.iadb.org/bitstream/handle/11319/7315/Infra- Araya. 2014. “What Drives the High Price of Road Freight structure%20Financing.%20Definitivo.pdf?sequence=1. Transport in Central America?” Policy Research Working Paper Serebrisky, Tomás, Andres Gomez-Lobo, Nicolas Estupinan, and Ra- 6844. World Bank, Washington, D.C. http://documents.world- mon Munoz-Raskin. 2009. “Affordability and Subsidies in Public bank.org/curated/en/364161468212970549/pdf/WPS6844. Urban Transport: What Do We Mean, What Can Be Done?” Trans- pdf. port Reviews 29 (6): 715–39. Perretti, Daniela and Mariana Vera. 2015. “Avances y retos de los Shankland, A., K. Caplan, K. Paiva, H. Neder, H. Gómez Bruera, and L. Lupo. Sistemas Nacionales de Inversión Publica de América Lati- 2010. Global Econonic and Sector Work (ESW) on the Political Econo- na – Resultados de la encuesta 2014” at Serie Gestión Publica my of Sanitation in Four Countries. World Bank, Washington, D.C. No. 83, ILPES, CEPAL. http://repositorio.cepal.org/bitstream/ Stapledon, T. 2012. Why Infrastructure Sustainability is Good for your handle/11362/37862/1/S1500113_es.pdf Business. Brisbane: Cooperative Research Centre for Infrastruc- Perrotti, Daniel E., and Ricardo J. Sánchez. 2011. “La brecha de in- ture and Engineering Asset Management (CIEAM). fraestructura en América Latina y el Caribe”. Recursos Naturales Tissot, Roger. 2012. “Latin America’s Energy Future”. IADB Discussion e Infraestructura. United Nations, CEPAL, Santiago. Paper. Inter-American Development Bank, Washington, D.C. Poindexter, Gregory B. 2015. “Brazil’s Drought Brings Water Supply TomTom International. 2015. “TomTom Traffic Index: Measuring to Near Zero Capacity at Hydroelectric Facilities.” Renewable En- Congestion Worldwide.” https://www.tomtom.com/en_gb/ ergy World, January 29. trafficindex/list. Poveda, S. and others. 2014. “Investigacion y diagnostico social para Transmilenio S.A. 2015. “Beneficios de Transporte Para Personas la universalicion de las conexiones intradomiciliarias de alca- Sisbenizadas.” Bogota. http://www.sitp.gov.co/publicaciones/ ntarillado sanitario del subsistema sur de Guayaquil”. World beneficios_de_transporte_para_personas_sisbenizadas. Bank, Washington, D.C. UN-HABITAT. 2013. Planning and Design for Sustainable Urban Mo- Powell, Andrew. 2013. Rethinking Reforms: How Latin America and bility. Global Report on Human Settlements. Geneva, United the Caribbean Can Escape Suppressed World Growth. Washing- Nations. http://unhabitat.org/?mbt_book=planning-and-de- ton, D.C., IDB (Inter-American Development Bank). sign-for-sustainable-urban-mobility-global-report-on-hu- Reguero, Borja G., J. Losada Iñigo, Pedro Díaz-Simal, Fernando J. man-settlements-2013. Méndez, and Michael W. Beck. 2015. Effects of Climate Change United Nations Population Division, Department of Economic and on Exposure to Coastal Flooding in Latin America and the Ca- Social Affairs. 2014. World Urbanization Prospects: The 2014 Revi- ribbean. PLoS ONE 10 (7): e0133409. http://journals.plos.org/ sion. New York, United Nations. https://esa.un.org/unpd/wpp/. plosone/article?id=10.1371/journal.pone.0133409. Vergara, Walter, Ana R. Rios, Luis Miguel Galindo Paliza, Pablo Gut- Reymondin, Louis, Andy Jarvis, Carolina Navarette, Alejandro Coca, man, Paul Isbell, Paul Hugo Suding, and Joseluis Samaniego. Denny Grossman, Alberto Villalba, and Paul Suding. 2013. “Road 2013. The Climate and Development Challenge for Latin America Impact Assessment Using Remote Sensing Methodology for and the Caribbean: Options for Climate-Resilient, Low-Carbon De- Monitoring Land-Use Change in Latin America: Results of Five velopment. Inter-American Development Bank, Washington, D.C. Case Studies”. Technical Note IDB - TN - 561. Inter-American De- Vogt-Schilb, Adrien, Stéphane Hallegatte, and Christophe De Gou- velopment Bank Environmental Safeguards Unit, Washington, vello. 2015. “Marginal Abatement Cost Curves and the Quality D.C. of Emission Reductions: A Case Study on Brazil.” Climate Policy Rodriguez, Camila, Juan Miguel Gallego, Daniel Martinez, Sergio 15 (6): 703–23. Montoya, and Tatiana Peralta-Quiros. 2016. “Examining the im- Ward, Philip J., Kenneth M. Strzepek, W. Pieter Pauw, Luke M. Brand- plementation and labor 1 outcomes of targeted transit subsi- er, Gordon A. Hughes, and Jeroen CJH Aerts. 2010. “Partial Costs dies:SISBEN Subsidy, Bogota Colombia.” Mimeo. of Global Climate Change Adaptation for the Supply of Raw In- Ruiz-Nunez, Fernanda. 2016. “Overview of Infrastructure Finance in dustrial and Municipal Water: A Methodology and Application.” LAC (Selected Graphs).” World Bank, Washington D.C. Environmental Research Letters 5 (4): 44011. 78 WHO (World Health Organization) and UNICEF. 2016. Joint Monitor- World Bank. 2013a. Honduras—Managing Fiscal and Public Finance ing Project: Progress on Sanitation and Drinking Water 2015 Up- Challenges. Washington, D.C. date and MDG Assessment. Geneva: World Health Organization. World Bank. 2013b. Towards Better Expenditure Quality: Guatemala http://www.who.int/water_sanitation_health/publications/ Public Expenditure Review. Washington, D.C. https://openknowl- jmp-2015-update/en/. edge.worldbank.org/handle/10986/16085. World Bank. Forthcoming. Energy Markets in LAC: Emerging Disrup- World Bank. 2013c. Mexico Moves to Results-based Pro- tions and the Next Frontier. Washington, D.C. curement System. Washington D.C. http:// World Bank. Forthcoming. Strengthening Governance in Infrastruc- beta.worldbank.org/en/results/2013/04/03/ ture in Brazil. Washington, D.C. mexico-moves-to-results-based-procurement-system World Bank. 2017. “Infrastructure Finance: Guiding Principles for World Bank. 2013d. “Inclusive Green Growth in Latin America and the World Bank Group–A Cascade Decision-Making Approach.” the Caribbean.” Washington D.C. Washington, D.C. World Bank. 2012a. “Brazil—Third Rio State Fiscal Efficiency for World Bank. 2016a. Benchmarking Public-Private Partnerships Pro- Quality of Public Service Delivery Development Policy Loan curement 2017: Assessing Government Capability to Prepare, Pro- Project”. World Bank, Washington, D.C. http://documents. cure, and Manage PPPs. Washington, D.C. worldbank.org/curated/en/979281468229179151/pd- World Bank. 2016b. Doing Business 2016: Measuring Regulatory Qual- f/627930REVISED00Official0Use0Only090.pdf. ity and Efficiency. Washington, D.C. World Bank. 2012b. Peru—Public Expenditure Review for Peru: Spend- World Bank. 2016c. Doing Business 2017: Equal Opportunity for All. ing for Results. Washington, D.C. http://documents.worldbank. Washington, D.C. org/curated/en/225811468297875669/pdf/NonAsciiFile- World Bank. 2016d. “Better spending, better services—A review Name0.pdf. of public finances in Haiti”. Washington, D.C. http://docu- World Bank. 2010. “The Costs to Developing Countries of Adapting ments.worldbank.org/curated/en/964461467030596497/pdf/ to Climate Change: New Methods and Estimates.” Economics ACS14253-v2-ENGLISH-PUBLIC-FINAL-ENG-PER-OVERVIEW.pdf. of Adaptation to Climate Change. Washington, D.C. World Bank. 2015.“Peru—Support to the Subnational Transport Pro- World Bank. 2008. Costa Rica—Public expenditure review: enhanc- gram Project”. Washington, D.C. https://hubs.worldbank.org/ ing the efficiency of expenditures. Washington, D.C. https:// docs/imagebank/pages/docprofile.aspx?nodeid=25485799. hubs.worldbank.org/docs/imagebank/pages/docprofile. World Bank. 2014a Mexico—Oaxaca Water and Sanitation Sector aspx?nodeid=9543594 Modernization Operation Project. Washington, D.C. http://doc- World Bank. 2006. Management of Environment and Social Is- uments.worldbank.org/curated/en/451781468123256085/pd- sues in Highway Projects in India. The World Bank, Wash- f/871040PAD0P145010Box385211B00OUO090.pdf. ington, D.C. http://documents.worldbank.org/curated/ World Bank. 2014b. Ceara PforR: Full Fiduciary Systems Assessment. en/313631468042034171/pdf/39752.pdf Washington, D.C. http://documents.worldbank.org/curated/ World Economic Forum. 2015. The Global Competitiveness Report en/346981468238750993/pdf/892790WP0P12080ry0Assess- 2015-2016. Geneva: WEF (World Economic Forum). http://re- ment0final.pdf. ports.weforum.org/global-competitiveness-report-2015-2016/. World Bank. 2014c. Turn Down the Heat: Confronting the New Climate WRI (World Resources Institute). 2016. “CAIT Country Greenhouse Normal. World Bank, Washington, DC. Gas Emissions.” Washington, D.C. 79 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE Annex 1. Public expenditure reviews examined for this report World Bank. 2015. El Salvador—Estudio de gasto público social y sus hubs.worldbank.org/docs/imagebank/pages/docprofile. instituciones. Washington, D.C. https://hubs.worldbank.org/ aspx?nodeid=17817212 docs/imagebank/pages/docprofile.aspx?nodeid=25011533 World Bank. 2013. Uruguay—Public Expenditure Review: Mitigat- World Bank. 2015. Haiti—Towards Greater Fiscal Sustain- ing Fiscal Risks. Washington D.C. https://hubs.worldbank.org/ ability and Equity: A Discussion of Public Finance.  Vols. docs/imagebank/pages/docprofile.aspx?nodeid=17731695 1 and 2 of Better Spending, Better Services—A Review World Bank. 2012. Guatemala - Public Expenditure Review. Wash- of Public Finances in Haiti. Washington, D.C. https:// ington, D.C. https://hubs.worldbank.org/docs/imagebank/ hubs.worldbank.org/docs/imagebank/pages/docprofile. pages/docprofile.aspx?nodeid=16593124 aspx?nodeid=24732446 World Bank. 2012. Mexico—Social Protection System in Health and World Bank. 2014. Brazil’s Productivity Challenge: Selected Issues in the Transformation of State Health Systems. Washington D.C. Understanding and Improving Productivity. Washington, D.C. https://hubs.worldbank.org/docs/imagebank/pages/docpro- https://hubs.worldbank.org/docs/imagebank/pages/docpro- file.aspx?nodeid=17591008 file.aspx?nodeid=19717534 World Bank. 2012. Peru—Public Expenditure Review for Peru: Spend- World Bank. 2014. Honduras—Central America Social Expen- ing for Results. Washington, D.C. https://hubs.worldbank.org/ ditures and Institutional Review. Washington, D.C. https:// docs/imagebank/pages/docprofile.aspx?nodeid=17204328 hubs.worldbank.org/docs/imagebank/pages/docprofile. World Bank. 2010. Main Report Vol. 1 of El Salvador—Enhancing the aspx?nodeid=23839196 Efficiency and Targeting of Expenditures: Public Expenditure World Bank. 2014. Ceara PforR: Full Fiduciary Systems Assess- Review. Washington, D.C. https://hubs.worldbank.org/docs/ ment. World Bank, Washington, D.C. http://documents.world- imagebank/pages/docprofile.aspx?nodeid=16242844 bank.org/curated/en/346981468238750993/pdf/892790WP- World Bank. 2010. Main report Vol. 1 of Public Expenditure Tracking 0P12080ry0Assessment0final.pdf. and Service Delivery Survey: Education and health in Hondu- World Bank. 2013. Agricultural productivity and family farms in Bra- ras. Washington, D.C. https://hubs.worldbank.org/docs/image- zil: creating opportunities and closing gaps. Washington D.C. bank/pages/docprofile.aspx?nodeid=15165197 https://hubs.worldbank.org/docs/imagebank/pages/docpro- World Bank. 2010. Nicaragua—Selected Issues in Social Sectors file.aspx?nodeid=17852509 Management: Programmatic Social NLTA Implementation Re- World Bank. 2013. Belize—PPIAF SNTA Municipal Finance: Report port. Washington, D.C. https://hubs.worldbank.org/docs/im- of RESP Activity. Washington D.C. https://hubs.worldbank.org/ agebank/pages/docprofile.aspx?nodeid=16579617 docs/imagebank/pages/docprofile.aspx?nodeid=17723213 World Bank. 2010. Peru—Evaluación del Presupuesto Participativo World Bank. 2013. Nicaragua—Agriculture Public Expenditure Re- y su relación con el presupuesto por resultados. Washington, view. Washington, D.C. https://hubs.worldbank.org/docs/im- D.C. https://hubs.worldbank.org/docs/imagebank/pages/doc- agebank/pages/docprofile.aspx?nodeid=23927226 profile.aspx?nodeid=12996952 World Bank. 2013. Public Expenditures for Decentralized Gover- World Bank. 2010. Peru—The decentralization process and its links nance in Honduras: Towards Restoring Fiscal Consolidation. with public expenditure efficiency. Washington, D.C. https:// Washington, D.C. http://documents.worldbank.org/curated/ hubs.worldbank.org/docs/imagebank/pages/docprofile. en/794291468037567297/pdf/826620REVISED00PER0HN0en- aspx?nodeid=16373412 glish0FINAL.pdf. World Bank. 2009. Mexico—Agriculture and Rural Development Pub- World Bank. 2013. The Quest for Optimal Tax and Expenditure Poli- lic Expenditure Review. Washington, D.C. https://hubs.worldbank. cies for Shared Prosperity. Vol. 1 of A Public Expenditure Review org/docs/imagebank/pages/docprofile.aspx?nodeid=11616349 for Paraguay. Washington, D.C. https://hubs.worldbank.org/ World Bank. 2008. Costa Rica—Public Expenditure Review: Enhanc- docs/imagebank/pages/docprofile.aspx?nodeid=18706123 ing the Efficiency of Expenditures. Washington, D.C. https:// World Bank. 2013. Towards Better Expenditure Quality: Guate- hubs.worldbank.org/docs/imagebank/pages/docprofile. mala Public Expenditure Review. Washington D.C. https:// aspx?nodeid=9543594 80 World Bank. 2008. Main report Vol. 1 of Brazil—Toward a more in- World Bank. 2007. Honduras: Public Expenditure Review Exec- clusive and effective participatory budget in Porto Alegre. Wash- utive Summary and Main Report. Washington, D.C. https:// ington, D.C. https://hubs.worldbank.org/docs/imagebank/pag- hubs.worldbank.org/docs/imagebank/pages/docprofile. es/docprofile.aspx?nodeid=9053955 aspx?nodeid=8961697 World Bank. 2008. Nicaragua—Análisis del Gasto Público Social. World Bank. 2006. Paraguay—Public Expenditure Review: Main Re- Washington, D.C. https://hubs.worldbank.org/docs/image- port. Washington, D.C. https://hubs.worldbank.org/docs/im- bank/pages/docprofile.aspx?nodeid=16588702 agebank/pages/docprofile.aspx?nodeid=6884196 World Bank. 2008. Nicaragua—Public Expenditure Review 2001- World Bank. 2006. Mexico—Water Public Expenditure Review. 2006. Washington, D.C. https://hubs.worldbank.org/docs/im- Washington D.C. https://hubs.worldbank.org/docs/image- agebank/pages/docprofile.aspx?nodeid=9701111 bank/pages/docprofile.aspx?nodeid=17234229 World Bank. 2008. Nicaragua—Social protection public expendi- ture review. Washington, D.C. https://hubs.worldbank.org/ docs/imagebank/pages/docprofile.aspx?nodeid=16579503 81 Annex 2. Procurement performance of Latin American countries: relatively good, but with wide variation across countries and indicators A lthough reforms have been undertaken by a number of countries throughout Latin Ameri-  hat is the Benchmarking Public Box 11: W ca, several areas can be further improved. The Procurement Database World Bank’s Benchmarking Public Procurement 2017 found that suppliers identify obstacles such The Benchmarking Public Procurement 2017 (BPP) initia- tive is a global assessment of public procurement regula- as excessive bureaucracy and red tape in Colombia and tory frameworks across 180 economies—of which 30 are Honduras; payment delays in Argentina, Jamaica, Domin- in Latin America. It focuses on eight key areas of the pub- ican Republic; lack of transparency and opaque tender- lic procurement process for a tender of works: ing process in Brazil and Mexico; lack of efficiency in Bar- bados and Puerto Rico; the list goes on. ››Needs assessment, call for tender, and bid prepa- ration: indicators assess the quality, adequacy, and More generally the Benchmarking Public Procurement 2017 transparency of the information provided by the pro- data reveals that many Latin American countries still face curing entity to prospective bidders. challenges in establishing legal and regulatory environ- ments that enhance efficient and transparent public pro- ››Bid submission: indicators examine the requirements that suppliers must meet in order to bid effectively curement markets. On average, however, the region com- and avoid having their bid rejected. pares reasonably well with other regions in all but one area measured: the region’s score is less than half that of OECD ››Bid opening, evaluation, and contract award: in- countries for recurrent delays in payments (figure 41). Time- dicators measure the extent to which the regula- ly payment of suppliers encourages firms’ participation in tory framework and procedures provide a fair and transparent bid opening and evaluation process, as tenders, which in turn will lead to more competition and well as whether once the best has been identified, better value for money for the purchasing entity (Connell the contract is awarded transparently and the los- 2014). Payments take place within the recognized good ing bidders are informed of the procuring entity’s practice time frame of 30 days in only 5 economies in the decision. region (Belize, Grenada, Nicaragua, Peru and The Bahamas), and can take six months or longer in the region’s worse per- ››Content and management of the procurement: con- formers (Dominican Republic and Trinidad and Tobago). tract indicators focus on several aspects during the contract execution phase related to the modification In addition, the region could improve its performance with and termination of the procurement contract, and the procedure for accepting the completion of works. respect to online access to information and services. Digital tools can streamline public spending, make it more trans- ››Performance guarantee: indicators examine the exis- parent and evidence-based and integrate it with informa- tence and requirements of the performance guarantee. tion on market conditions. Only 13 of the 30 Latin Ameri- can countries evaluated make all procurement information ››Payment of suppliers: indicators focus on the time available online, and 7 have yet to establish an electronic and procedure needed for suppliers to receive pay- ment during the contract execution phase. portal dedicated to public procurement. The same goes for online services during the bid submission, bid evaluation 83 RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  atin America’s regional performance is on par with others, except when it comes to timely FIGURE 41: L payment of suppliers 100 90 80 70 60 50 40 30 20 10 0 OECD high income Europe and Central Asia Latin America and Caribbean South Asia region Sub Saharan Africa Middle East and North Africa East Asia Paci c Europe and Central Asia OECD high income Latin America and Caribbean South Asia region Sub Saharan Africa East Asia Paci c Europe and Central Asia OECD high income Sub Saharan Africa Latin America and Caribbean South Asia region East Asia Paci c Middle East and North Africa Europe and Central Asia OECD high income Latin America and Caribbean Sub Saharan Africa Middle East and North Africa South Asia region East Asia Paci c Europe and Central Asia Sub Saharan Africa Latin America and Caribbean Middle East and North Africa OECD high income South Asia region East Asia Paci c OECD high income Europe and Central Asia South Asia region Sub Saharan Africa Latin America and Caribbean Middle East and North Africa Middle East and North Africa East Asia Paci c Needs assessment, Bid submission score Bid opening, Content and Performance Payment of suppliers call for tender, and bid evaluation and award management of guarantee score score preparation procurement contract Source: Benchmarking Public Procurement 2017. Note: For each indicator developed, the scores of individual questions are averaged and multiplied by 100, resulting in a final score ranging from 0 to 100. The economies at the top of the range (with scores approaching 100) are considered to have a regulatory framework that closely aligns with internationally recognized good practices, whereas the economies at the bottom of the range (scores closer to 0) have significant room for improvement in the particular area measured. and award and contract management phases. Just as sup- the contract execution phase are an important tool to pliers need to access information online, they should be ensure delivery of service per contract terms and protect able to conduct the procurement process online, regard- parties in case of delays in execution. All 30 economies less of what is being procured by the government. measured in Latin America by the Benchmarking Public Procurement initiative impose a performance guarantee The availability of electronic tendering is not as widespread requirement during the contract execution phase. Perfor- for the procurement of works as it is for the procurement mance guarantees should be well-regulated in order to of goods. In Chile for example, the electronic submission of protect suppliers and not create an additional impediment bids has become the rule for the procurement of goods but for them. For example, it is important for suppliers to have not for the procurement of works. Similarly, when it comes a choice with regard to the form of performance guaran- to the bid opening session, bids are sometimes opened tees. Costa Rica and Ecuador are the economies that pro- electronically in Brazil, Costa Rica, Ecuador, Jamaica, Mexi- vide the most options for the form of the guarantee, which co and Uruguay and never in the remaining 24 economies. can include a certified check, a certificate of deposit, or a letter of credit, among others. It is also critical that the legal Performance guarantees, which are particularly important framework set a timeframe for the purchasing entity to re- for public works, is another area where the Benchmarking turn the performance guarantee after the execution of the data highlights great variation among the different players contract. Only 11 countries such as El Salvador, Mexico and in the region (figure 42). Performance guarantees during Panama impose such a time limit. 84 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Costa Rica Colombia Dominica Mexico Dominican Republic Nicaragua Haiti Peru Paraguay Panama Jamaica Ecuador FIGURE 42: V Mexico Argentina est performance. Uruguay Costa Rica Argentina Dominican Republic El Salvador El Salvador Grenada Paraguay Nicaragua Brazil Panama Haiti Bid opening, evaluation and award St. Lucia Uruguay Brazil Bolivia Ecuador Guatemala 0 10 20 30 40 50 60 70 80 90 100 Source: Benchmarking Public Procurement 2017 Antigua and Barbuda Jamaica Colombia The Bahamas Honduras Guatemala Bolivia Puerto Rico Mexico Colombia The Bahamas Brazil Guatemala Venezuela, RB Needs assessment, call for tender, and bid preparation Costa Rica Honduras St. Kitts and Nevis Peru Peru St. Lucia Ecuador Payment of suppliers Puerto Rico Trinidad and Tobago Haiti St. Kitts and Nevis Belize Paraguay Trinidad and Tobago Suriname Suriname Venezuela, RB Antigua and Barbuda Venezuela, RB Belize Dominica Honduras Barbuda, and St Lucia on the other hand, show the weak- 5 performing economies across three areas. Antigua and Ecuador, for example, shows high scores in most areas tries in the region. Out of the 30 economies measured, Regional averages hide the wide disparity between coun- measured. Similarly, Costa Rica, which appears in the top Suriname Grenada Argentina Barbados Barbados El Salvador 85 Uruguay 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 The Bahamas Ecuador Ecuador Belize Bolivia Peru Grenada Guatemala Suriname Nicaragua Jamaica Jamaica Puerto Rico Brazil Costa Rica Bid submission Bolivia Colombia Argentina Panama El Salvador Brazil Antigua and Barbuda Puerto Rico Panama Jamaica The Bahamas Guatemala Barbados Belize Dominican Republic Dominica Panama El Salvador Dominican Republic Paraguay Grenada St. Kitts and Nevis Suriname Mexico St. Lucia Argentina Uruguay Trinidad and Tobago Costa Rica Puerto Rico Honduras Paraguay Mexico Nicaragua Nicaragua Haiti Peru Honduras  ariation appears between Latin American economies across thematic areas St. Lucia Bolivia Content and management of procurement contract Uruguay Venezuela, RB Dominica Barbados Dominican Republic Colombia Haiti Belize Trinidad and Tobago Antigua and Barbuda Venezuela, RB Trinidad and Tobago Antigua and Barbuda St. Lucia St. Kits and Nevis The Bahamas Barbados Dominica Public Procurement of Public-Private Partnerships Grenada St. Kitts and Nevis in two of four thematic areas (unsolicited proposals and Overall, economies in Latin America perform well with ond globally behind the OECD high-income economies Procurement of Public Private Partnerships, placing sec- based on data from the World Bank’s Benchmarking for respect to procurement for public-private partnerships RETHINKING INFRASTRUCTURE IN LATIN AMERICA AND THE CARIBBEAN SPENDING BETTER TO ACHIEVE MORE  enchmarking PPP Procurement Scores vary by region and thematic areas FIGURE 43: B (scores) 100 90 80 70 60 50 40 30 20 10 0 OECD High income Europe and Central Asia South Asia region Latin America and the Caribbean East Asia Paci c Sub Saharan Africa Middle East and North Africa OECD High income Europe and Central Asia Latin America and the Caribbean East Asia Paci c Middle East and North Africa Sub Saharan Africa OECD High income Latin America and the Caribbean Europe and Central Asia Middle East and North Africa East Asia Paci c Sub Saharan Africa OECD High income Latin America and the Caribbean Europe and Central Asia East Asia Paci c Sub Saharan Africa Middle East and North Africa South Asia region South Asia region South Asia region Preparation Procurement Unsolved proposal Contract management Source: Benchmarking PPP Procurement 2017. Note: PPP = public-private partnership; USP = unsolicited proposal. contract management) (figure 43).20 The region’s lowest very well in that area, but its scores are not as good for the performance is for the PPP Preparation indicators. preparation stage, particularly for the concession regime. The good average performance masks wide variation across countries (figure 44). Procurement is the area Drivers of Variance in Preparation and Contract where the economies obtained more similar scores, in- Management dicating a more consistent and overall adequate perfor- mance. However even in this area there are outliers, such But the area with the widest variation in performance is as Ecuador and Jamaica, which score significantly lower Preparation of PPPs. Benchmarking PPP Procurement 2017 than the other economies in the region. data measures whether economies conduct six assess- ments, which include: socio-economic analysis, afford- Wide variation exist around unsolicited proposals, for which ability assessment, risk identification, financial viability some countries have either enacted comprehensive regu- or bankability assessment, comparative assessment and lations (Jamaica, Peru, Colombia and Costa Rica) and oth- market assessment. Some economies conduct almost ers have no regulations at all (Guatemala, Nicaragua and all of these assessments (such as Mexico and Jamaica), Panama). Similarly, wide variations in scores appear around while others conduct only one of them (Panama and preparation and contract management, with some econ- Nicaragua), or none at all (Dominican Republic). And in omies performing relatively well in both areas (Colombia, Nicaragua for example, the government is not required Mexico, Peru and Uruguay), and others lagging behind in to integrate the prioritization of PPP projects with oth- both (Nicaragua). Brazil has adopted very comprehensive er public investment projects, nor are the assessments regulations regarding contract management and scores required. 20 The Benchmarking PPP procurement data assesses governments’ capacity to prepare, procure and manage PPP projects. http://bpp.worldbank.org/~/media/ WBG/BPP/Documents/Reports/BenchmarkingPPP2017.pdf 86  ariation also emerges across the region by thematic area FIGURE 44: V Preparation score Procurement score 100 100 80 80 60 60 40 40 20 20 0 0 Colombia Peru Mexico Uruguay Jamaica Honduras Costa Rica Brazil PPP Guatemala Ecuador Brazil Concession Dominican Republic Argentina Panama Nicaragua Dominican Republic Mexico Brazil Concession Brazil PPP Nicaragua Colombia Guatemala Uruguay Argentina Peru Costa Rica Panama Honduras Jamaica Ecuador Unsolicited proposal Contract management score 100 100 80 80 60 60 40 40 20 20 0 0 Jamaica Peru Colombia Costa Rica Brazil Concession Brazil PPP Ecuador Honduras Mexico Argentina Dominican Republic Uruguay Brazil PPP Mexico Brazil Concession Peru Uruguay Costa Rica Colombia Argentina Guatemala Panama Honduras Dominican Ecuador Jamaica Nicaragua Source: Benchmarking PPP Procurement 2017 Note: The following economies do not have a regulatory framework that explicitly mention unsolicited proposals (Unsolicited proposals are not regulated, and therefore not scored in: Guatemala, Nicaragua, Panama). Our analysis also examines whether the Ministry of Fi- a PPP contract, as well as the three circumstances that nance, or a central budgetary authority, needs to ap- could be regulated (a change in the scope/object of prove the PPP project before a procurement process is contract, a change in the risk allocation of the contract, launched. According to our assessment, such an approval a change in the investment plan or duration of the con- is not required in four of the 14 economies measured (Ar- tract). The regulatory frameworks in Ecuador, Guatemala, gentina, Dominican Republic, Ecuador, and Nicaragua). Honduras and Nicaragua do not regulate the modifica- tion/renegotiation of a PPP contract. Contract Management is another area where a signif- icant difference in performance exists amongst econ- Overall, the regulatory frameworks governing the procure- omies. Although there are several factors that lead to ment of PPPs in the Latin America region are better than in the lower performance of some economies in this area, other regions of the globe. However, there is room for im- a few particular ones stand out. Jamaica and Honduras provement in all areas, in order to increase the quality of the are the economies where the regulatory framework does regulatory framework to match those that exist in the OECD not expressly establish grounds for the termination of a high-income economies. Preparation of PPPs appears to be PPP contract. Our data also measures whether the reg- the area where the region as a whole could focus its efforts ulatory framework establishes a specific dispute resolu- in order to improve regulatory frameworks for PPPs. While tion mechanism, and we find that Colombia, Jamaica, the regulatory frameworks for PPPs in economies like Co- Nicaragua, and Panama do not have such a system. Last- lombia, Mexico, Uruguay, and Peru are comparable to that ly, our data examines whether the regulatory framework of more mature markets, some other economies clearly lag expressly regulates the modification or renegotiation of behind in several aspects (Nicaragua and Jamaica). 87