44903 WORKING PAPER no. 6 The Rise and Fall of Brownfield Concessions But Some Signs of Recovery After a Decade of Decline James Leigland Helping to eliminate poverty tHrougH private involvement in infrastructure WORKING PAPER NO. 6, 2008 The Rise and Fall of Brownfield Concessions But Some Signs of Recovery After a Decade of Decline James Leigland Public-Private Infrastructure Advisory Facility The findings, interpretations, and conclusions expressed in this Working Paper are entirely those of the authors and should not be attributed in any manner to the Public-Private Infrastructure Advisory Facility (PPIAF) or to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Neither PPIAF nor the World Bank guarantees the accuracy of the data included in this publication or accepts responsibility for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this report do not imply on the part of PPIAF or the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is owned by PPIAF and the World Bank. Dissemination of this work is encouraged and PPIAF and the World Bank will normally grant permission promptly. For questions about this report including permission to reprint portions or information about ordering more copies, or for a complete list of PPIAF publications, please contact PPIAF by email at the address below. PPIAF c/o the World Bank 1818 H. Street Washington, DC 20433 Fax: 202-522-7466 www.ppiaf.org Email: ppiaf@ppiaf.org PPIAF produces three publication series: Trends and Policies Working Papers Gridlines They are available online at www.ppiaf.org TABLE OF CONTENTS Acknowledgements vi Foreword vii Executive Summary viii 1. The Brownfield Concession Concept 1 2. Boom and Bust 3 3. Empirical Evidence of Problems 7 4. Underlying PPP Vulnerability to Cash Flow Stress 9 and Profitability Problems 5. Cash Flow Vulnerability Particular to Brownfield Concessions 11 Brownfield Concessions and Cash Flows 11 Insufficient Return on Investment 12 The Failure of Risk Mitigation 14 6. A Critical Underlying Problem: Poor Project Preparation 16 Low Cost (and Low Quality) Preparation 16 Difficulties in Undertaking Feasibility Studies 17 Shortcomings of Cost-Benefit Analysis 17 7. PPP Alternatives to Concessions 19 Overview 19 Brownfield Concessions 20 Divestiture 21 Greenfield 23 Leases and Management Contracts 25 8. How Governments Are Reducing Private Partner Risks 27 on Brownfield Concessions The Changing Nature of the Private Sector's Risk-Reward 27 Requirements Government Assumption (or Sharing) of Investment Risk 27 More (or Guaranteed) Remuneration for Private Partner 29 Assumption of Investment Risks Government Assumption (or Sharing) of Demand Risk 30 Government Ownership and Control of Project Preparation 31 9. Private Sector Actions to Strengthen Profits and Cash Flows 32 More Due Diligence, Less Opportunistic Bidding 32 Maximizing Remuneration from Available Sources 32 Concessionaire Self-Selection: Better Matching of Needs, 33 Skills, and Opportunities 10. Signs of Recovery in 2006? 35 11. Conclusions 37 iii List of Figures Figure 1. PPI Investment in Government-Owned Facilities via Brownfield Concessions Figure 2. Investment in Government-Owned Facilities via Brownfield Concessions as a Percentage of all PPI Investment Facilities Figure 3. Brownfield Concessions Versus Other Kinds of PPI Figure 4. Investment Facilities by PPP Contract Type ­ 1990-96 Figure 5. Investment in Facilities by Brownfield Concession Subtype ­ 1990-96 Figure 6. Sectoral Investments by Brownfield Concession Subtype ­ 1990-06 Figure 7. Investment in Facilities by Divestiture Subtype ­ 1990-06 Figure 8. Sectoral Investments by Divestiture Subtype ­ 1990-06 Figure 9. Investment in Facilities by Greenfield Subtype ­ 1990-06 Figure 10. Sectoral Investments by Greenfield Subtype ­ 1990-06 Figure 11. Lease/Management Contracts as a Percentage of All PPI Projects ­ 1990-06 Figure 12. Sectoral Distribution of Lease/Management Contracts ­ 1990-06 Figure 13. Sectoral Distribution of Investment ­ Brownfield Concessions: 1997 and 2006 iv ACRONYMS AND ABBREVIATIONS BOO Build, own, operate BOT Build, operate, transfer BROT Build, rehabilitate, operate, transfer PPI Private participation in infrastructure PPIAF Public-Private Infrastructure Advisory Facility PPP Public-Private Partnership RLT Rehabilitate, lease, transfer ROT Rehabilitate, operate, transfer SPV Special purpose vehicle v ACKNOWLEDGEMENTS The author, James Leigland, is the team leader Caribbean Region, World Bank and Chris for the Sub-National Technical Assistance Shugart, a consultant for the World Bank. Paul program at PPIAF. James joined PPIAF in Reddel and Joel Kolker, regional team leaders 2005 as the regional program leader for east for PPIAF for east and southern Africa and East and southern Africa in Nairobi. In August 2007 Asia and the Pacific respectively also provided he started as the team leader for the Sub- input as well as Rosalind Thomas from the National Technical Assistance Program. Before African Development Bank. joining PPIAF, Jim served as a senior municipal infrastructure advisor and the acting The findings, interpretations, and conclusions chief executive officer of South Africa's expressed in this report are entirely those of the Municipal Infrastructure Investment Unit. He authors and should not be attributed in any also served as senior urban policy advisor for manner to the Public-Private Infrastructure Southeast Asia at the U.S. Agency for Advisory Facility (PPIAF) or to the World International Development. He is a former Bank, to its affiliated organizations, or to faculty member at the University of Kentucky's members of its Board of Executive Directors or Martin School of Public Administration and the countries they represent. Neither PPIAF nor Columbia College in New York. Jim holds a the World Bank guarantees the accuracy of the PhD in political science from Columbia data included in this publication or accepts University. responsibility for any consequence of their use. The author is very grateful for comments and The material in this report is owned by PPIAF suggestions from Stephan von Klaudy, lead and the World Bank. Dissemination of this infrastructure specialist, Finance, Economics work is encouraged, and PPIAF and the World and Urban (FEU) Department, World Bank, Bank will normally grant permission promptly. Jeff Delmon, senior infrastructure specialist, For questions about this report or information FEU, the World Bank, J. Luis Guasch, senior about ordering more copies, please contact adviser from the Latin America and the PPIAF by email: ppiaf@ppiaf.org vi FOREWARD Nothing better reflects changes in public-private This paper attempts to explain why some kinds partnerships (PPPs) in developing countries over of brownfield concessions do seem to be on the the last decade than the precipitous decline in verge of extinction and how others have evolved the use of brownfield infrastructure concession into sturdier, more sustainable arrangements, contracts-- long-term PPP arrangements sometimes by being blended with other forms of whereby private companies manage and improve PPP to form hybrid structures. In some cases, existing infrastructure systems, such as water however, these hybrid arrangements bring with distribution networks and roads. In the early them their own special challenges and risks, 1990s, the development community had high which need to be well understood by the parties hopes for these kinds of contracts because they involved if contracts are to be sustainable in offered a solution to the most difficult of all ways that avoid conflicts of interest and other infrastructure problems--what to do with problems. We hope that this paper will provide existing, but badly dilapidated, government- useful insights into these issues for policy owned infrastructure assets that were difficult to makers, development agencies, prospective fully privatize or close down. The use of these private partners, and other stakeholders engaged kinds of concessions, however, went into sharp in rehabilitating existing government-owned decline with the Asian Crisis in 1997, and this infrastructure systems in developing countries. dramatic drop in popularity was an important factor in the sharp downturn of the entire PPP Laszlo Lovei market beginning at that time. Since then, the Director use of these kinds of concessions has remained Finance, Economics and Urban Department at low levels even though other elements of the World Bank market have demonstrated increasingly strong performance. Modest increases in the use of Jyoti Shukla brownfield concessions in 2006 and 2007 reflect Program Manager changes in how these agreements are being Public-Private Infrastructure Advisory Facilit structured in some sectors and suggest that some forms of brownfield concession may also finally be on the road to recovery. vii EXECUTIVE SUMMARY In 1990 brownfield infrastructure concessions Asian crisis. Other forms of PPI barely suddenly captured the attention of development registered the effects of the crisis, and all have professionals with a sevenfold increase in long ago surpassed their precrisis investment number over the previous year. For public highs. service providers as well as private operators and financiers brownfield concessions were an Only brownfield concessions have never attractive option, embodying almost all the most recovered. A 2006 surge in popularity pushed beneficial qualities associated with public- investment through brownfield concessions to private partnerships in infrastructure. Perhaps about 40 percent of its 1997 peak. The surge most important, they were seen as a solution to (probably followed by another in 2007) suggests one of the most difficult infrastructure problems that some kinds of brownfield concessions may facing the developing world--what to do with finally be poised for recovery. badly dilapidated infrastructure service systems, such as water delivery facilities and roads, that What happened to brownfield concessions? could not be shut down or sold off. No single factor accounts for the rise or decline in the use of a PPI mechanism in all situations. The concept was simple: private companies In Latin America, for example, many of the would take over badly maintained government- most attractive opportunities for brownfield owned infrastructure service systems, improve concessions were taken up in the early 1990s. efficiency, make needed investments, and After the Asian crisis public opposition to recover all their costs--plus make reasonable privatization may have combined with the profits--over the long term (20­30 years) of the eventual financial recovery of some contracts. Best of all, because of greater governments to diminish the attractions of operational efficiency, carefully targeted and turning infrastructure service provision over to managed investments, and more realistic pricing, private operators. these operators would deliver better services while still recovering costs. Thus the new But Latin America is also the source of some of arrangements would be largely self-supporting, the most compelling empirical evidence on other in dramatic contrast to the huge budget deficits key reasons for the steep decline in the use of that had resulted from public subsidies for brownfield concessions in the late 1990s. A inefficient service provision. recent study of concession contract renegotiations in Latin America suggests that But the track record of brownfield concessions is cash flow problems and low profitability were one of boom and bust (figure 1). Indeed, the common in these arrangements (Guasch 2004). sudden unpopularity of the brownfield conces- Using a definition of concession that includes sion almost single-handedly accounted for what some greenfield projects and divestitures, the is normally thought of as a sharp decline in study shows that concessions in Latin America private participation in infrastructure (PPI) had a high incidence of renegotiation--about 42 following the Asian crisis. Data from the percent, with renegotiation happening on average after only 2.2 years of operation. The PPI Project Database show that if brownfield results tended to favor operators, mostly through concessions are excluded, the PPI market, improvements to cash flow and profitability. buoyed by privatizations and greenfield projects, demonstrated few of the "crash" characteristics A second study looked at the profitability of commonly associated with the aftermath of the infrastructure concessions in Latin America viii during the late 1990s, again using a broad show early, dramatic results, but such invest- definition of concession (Sirtaine and others ments often were not optimally targeted or timed 2004). The study suggests that, on average, because the operators lacked experience with the projects became profitable only after about 10 systems. And many of the real problems years. Until then concession shareholders earned involved sector and policy issues (tariffs, labor negative returns, even when such things as productivity, corruption) rather than day-to-day management fees, estimated accumulated capital operations. gains, and potential investment markups were included. But this same study found that 40 Many operators of retail operations faced severe percent of the concessions in the sample--and currency mismatch problems, with revenues in 50 percent of those in energy and transport--did local currency and debt service payments in hard not appear to have the potential to ever become currency. In addition, many governments profitable. required brownfield concessions to pay debt service for outstanding loans used in initially Problems with cash flows and long-term developing the facilities. The need to pay off the profitability were clearly among the most initial investment on top of the new investment important reasons that brownfield concessions put more, and often unsustainable, pressure on became so unpopular so quickly. These projects cash flows. must be able to weather years of negative cash flows and constant uncertainty about long-term On the other side of the ledger, revenues were profitability. That so many contracts were often less than expected, particularly for retail renegotiated after only a few years, long before service operations that were supposed to recover they could confirm their profitability, suggests full costs. Raising tariffs to cover the full costs that cash flow problems were probably critical in of operation turned out to be impossible, or at precipitating many renegotiations. Even if least wholly impractical, in many situations, estimates of long-term profitability are positive, particularly in poor areas. Indeed, full cost a project that early on generates cash flows too recovery for essential infrastructure services small to service debt is not viable without cash such as water supply and sanitation is rarely inflows from other sources or contract renegotia- attempted even in developed economies. tion to adjust existing flows. Why weren't the problems better Why so susceptible to problems? anticipated? The PPI Project Database confirms that brown- The potential for cash flow and profitability field concessions were far more likely to problems should have been apparent during experience these kinds of contractual distress project appraisal and design. Why did so many than other forms of long-term PPI. In 1990­98 contracts reach financial closure before these the share of brownfield concessions that were weaknesses were noticed? The quality of canceled or became distressed was 41 percent preparation often seems to have been very poor, higher than that for greenfield projects. for several reasons. Why would brownfield concessions be more First, governments, as well as donors and prone to problems with cash flows and development agencies, often were unwilling to profitability? The answer is simple: as business spend time or money preparing brownfield transactions, many brownfield concessions concessions--doing feasibility studies, turned out to be far less profitable than expected. examining the true cost of the services, assessing The assets were often in much poorer condition contracting options, and the like. For many of than expected and required more basic the contracts signed in the early 1990s all this rehabilitation and investment before they could work was assumed to be the responsibility of start generating higher revenue. Concessionaires potential private partners--part of their normal and governments often wanted to start the due diligence--because if the project failed, it investment programs as soon as possible, to would be at their sole cost. ix We now know that for existing, poorly main- private funding options, and "political economy" tained facilities, governments need independent, issues. comprehensive assessments of the condition of the infrastructure so that they can identify the Quantitative estimates of the financial costs and objectives and the investments needed in brown- benefits of these projects were also often wildly field concessions and can evaluate bids on the inaccurate. In several studies of transport basis of consistent operating and investment projects Flyvbjerg (2005) found massive projections. Leaving such assessments to bidders underestimation of costs and overestimation of who put different amounts of time and resources demand. into feasibility studies and asset reviews, led to bids that were often difficult to compare or Fourth, the concessions often lacked settled based on incomplete or inaccurate views of regulatory or contractual arrangements for investment needs. Perhaps the first notable increasing tariffs or coping with unexpected example of the problem of low-cost preparation changes. Bidders were often prepared to commit was the Buenos Aires water concession--one of to concessions without such arrangements on the the first large brownfield concessions--signed in basis of government reassurances that such December 1992. A defining feature of the tender issues would be readily resolved. Often the process was poor information. rhetoric failed to match the reality, and concessionaires faced severe hurdles in securing, Second, even where one party or another was for example, a contractually mandated tariff willing and able to undertake full feasibility increase. studies, the task often turned out to be far more difficult and expensive than expected. What about risk mitigation? Management information and basic record The emergence of cash flow and profitability keeping were often outdated or nonexistent. problems was supposed to trigger risk mitigation Historical performance data were sometimes mechanisms agreed to at the outset. The most inaccurate or unavailable, and the condition of important risk mitigation instruments, structured the infrastructure, such as underground pipes, as off-take agreements and project guarantees, impossible to evaluate. Even customer records were to be provided by state-owned enterprises, were often incomplete or missing. As a result, utilities, or the governments themselves. But the there was often no way to tell, for example, how Asian crisis in 1997 forced many governments many end users were connected to water to recognize that they had an inaccurate systems, much less paying their bills. understanding of how brownfield concession contracts, as well as other kinds of public- Third, the preparation of brownfield concession private partnerships, were supposed to work. projects was probably affected by several weak- nesses now widely recognized in the project The contingent liabilities associated with the risk appraisal techniques used to help anticipate and mitigation instruments had been ignored or avoid problems with cash flows and misunderstood by the governments. Whether profitability. We now know that such techniques they knew it or not, the public sector retained were often not used at all because expensive massive contingent liabilities. But under the analysis was thought unnecessary in situations intense pressure of the Asian crisis, governments where remedial options seemed obvious. When simply repudiated these obligations, forcing the techniques were used by or on behalf of many projects into renegotiation or collapse. The government partners, they often served to justify PPI Project Database confirms that this rather than independently assess projects. Other happened more often with brownfield preparation techniques, such as economic cost- concessions than with any other form of PPI benefit analysis, seem to have been generally contract. overwhelmed by bad data, complexities, public- x The future of brownfield concessions demand risk associated with user payment for Data for 2006 from the PPI Project Database services. Under such an arrangement (generally suggest that some kinds of brownfield conces- captured by the PPI Project Database as a sions are becoming more popular now because management contract rather than a concession), governments are more aggressively structuring the government assumes most of the investment the arrangements to reduce the risks for private and demand risks. partners. In toll road projects, for example, governments are reducing investment risk by Conclusion providing capital grants or financing guarantees, The brownfield concession is not an inherently and reducing demand risk by using shadow tolls flawed mechanism--its track record in or guaranteeing part of the revenue through developed countries is reasonably successful. minimum traffic assurances. The key challenge But many of the conditions for success have in using these contracting arrangements is to proved difficult to achieve in developing find ways of maintaining performance incentives countries, where preparation is especially time for the private partners. consuming and expensive and the assets are in particularly poor condition. But we now have a Some governments are adopting hybrid arrange- much better understanding of the risks and ments to mitigate risks. In high-risk sectors such problems in dealing with existing, dilapidated as retail water distribution in Africa, projects infrastructure assets in developing countries. that once would have been implemented through Brownfield concessions as structured in the brownfield concessions are being unbundled. early 1990s may be an endangered species, but Private operators implement management the needs that drove their initial widespread use contracts and receive compensation through a still exist, and refinements to the concession flat fee rather than from user fees. Operators mechanism--along with new investment issue and collect bills, fix leaks, or manage arrangements and hybrid contract forms--are equipment. Governments and donors supply emerging to deal with these problems. funding for capital investment and take on the xi 1 THE BROWNFIELD CONCESSIONS CONCEPT By the end of 2006, private participation in entities rehabilitate existing assets and, in some infrastructure projects in developing countries cases, also build add-ons to existing facilities, seemed to be recovering after almost a decade of but all assets remain under government decline following the Asian Crisis in 1997. In ownership or revert to government ownership at nominal terms, PPP infrastructure investments in the end of the contract. So, under this definition, 2006 were close to their 1997 peak, and the these contracts include arrangements like market had by then demonstrated increased "rehabilitate, operate, transfer" (ROT), "build, investment in three consecutive years-- rehabilitate, operate, transfer" (BROT), and something not seen since the early 1990s. The other, similar mechanisms. A second important characteristics of the market restructuring that element of this definition is that the primary had taken place since the end of the 1990s were focus of the private partner's responsibilities is becoming fairly clear. Telecommunications had on existing assets--this distinguishes brownfield become the sector demonstrating the most concessions from the many varieties of dramatic recent increases in investment, "greenfield" projects, such as "build, operate, "greenfield" projects had come to transfer" (BOT) arrangements (sometimes overwhelmingly represent the most widely used referred to as "greenfield concessions"). To form of PPP vehicle for infrastructure clearly distinguish concessions involving mostly investment, and management and lease contracts existing assets from those involving mostly new also appeared to be on the increase, often in assets, this paper refers to the former tandem with increased government investment.1 concessions as "brownfield" and the latter as "greenfield." But perhaps the most telling lesson from the last decade has to do with the performance of The PPI Database's reservation of the term brownfield infrastructure concession contracts. "concession" only for brownfield projects no The PPI Project Database maintained by the longer is a widely used definition.3 But the World Bank and the Public-Private database's approach is understandable because it Infrastructure Advisory Facility (PPIAF) simply captures what in the late 1980s and early 1990s refers to brownfield concessions as was seen to be an important, groundbreaking "concessions," and defines them as contractual form of PPP, one that offered a solution to the arrangements whereby "a private entity takes severe problems associated with badly over the management of a state-owned dilapidated retail infrastructure service systems, enterprise for a given period during which it also such as water delivery networks and roads, that assumes significant investment risk."2 Capital could not easily be privatized or replaced. For a investment, along with operating and time, this kind of PPP arrangement was expected maintenance responsibilities, is one important to become the signature contract of the PPI element of the definition. Pursuant to contracts movement, but has now fallen largely into that are normally long term in character, private 1In this paper, "private participation in infrastructure" 3It is difficult to find this narrow definition in use (PPI) and "public-private partnerships" (PPPs) in anywhere but in connection with the PPI Project Database. infrastructure are used interchangeably. See, for example, the much broader definition used by the 2For all of the definitions used in database, see the World OECD (2007). The EBRD also uses a much broader Bank and PPIAF, Private Participation in Infrastructure definition: see (PPI) Project Database (ppi.worldbank.org). http://www.ebrd.org/country/sector/law/concess/core/indes. htm. 1 disuse in the developing world, at least Asian Crisis, greenfield projects and compared to other forms of PPP. management contracts are dominating the world of PPPs. In 2006, investment via brownfield In 1990, the brownfield infrastructure concessions demonstrated only the second concession suddenly captured the attention of annual increase since 2007, reaching a level first development professionals with a sevenfold achieved in 1993 (about 40 percent of the 1997 increase in the number of such projects over the peak). The 2006 performance reflects major previous year. It was an undeniably attractive changes in how brownfield concessions are contract form, embodying almost all of the most being structured in some sectors, and may signal beneficial qualities associated with PPPs for a comeback of sorts for the mechanism. infrastructure: private companies would take over badly maintained, government-owned In the sections that follow, recent research and infrastructure service systems, improve anecdotal evidence are used to piece together efficiency, make needed investments, and reasons why the brownfield concession now recover all of their costs--plus make reasonable seems to be an endangered species of PPP, how profits--over the long term of the contracts (20 some kinds of brownfield concessions have to 30 years). Best of all, because of operational evolved in order to survive, and what other types efficiencies, carefully targeted and managed of arrangements may be serving as substitutes investments, and more realistic pricing, these for these contracts. Note that the conclusion of contractors would deliver services on a cost- this paper is not that this contracting mechanism recovery basis. Instead of the huge budget is inherently flawed, but that the use of this deficits that had resulted from public particular form of concession in developing subsidization of inefficient service provision, countries is much more fragile than many these new arrangements would be largely self- market participants assumed in the early 1990s supporting. and that a complex set of conditions must be in place for this kind of mechanism to be By the mid-1990s, these agreements were being sustainable. widely promoted by sponsors, governments, and development agencies as win-win arrangements Section two reviews the global track record of with plenty of benefits to spread around to brownfield concessions over the last 15 years. everyone involved: government assets would The evidence suggests that the precipitous benefit from private involvement, but not decline in the use of these kinds of concessions actually undergo privatization, so all assets beginning with the Asian Crisis was the most would remain under government ownership.4 important single factor in the sharp decline of The assets would be rehabilitated at no cost to the entire PPP market, which began at that time. the government; private companies would make profits; customers would receive better, less Section three reviews the conclusions of recent expensive services. The many risks associated empirical studies, which suggest that these kinds with these projects would in effect be minimized of concessions have experienced problems with because, in the parlance of project finance, they profits and cash flows. would be "allocated to the parties best able to manage them." The arrangement seemed almost Section four looks at the project finance too good to be true, and that turned out to be mechanism that underlies concessions and other exactly the case. A decade after the onset of the forms of long-term PPP, and argues that despite the success of project finance in so-called developed countries, the mechanism is 4 Private involvement without privatization was an particularly vulnerable to the kinds of cash flow attractive feature of these arrangements. See Pierre and profitability problems common in Guislain and Michel Kerf (1995), "Concessions--The Way developing countries. to Privatize Infrastructure Monopolies," Note No. 59, Public Policy for the Private Sector, World Bank. 2 Section five turns to the brownfield concession Section eight reviews efforts by governments to mechanism as a particular form of project reduce some of the risks facing private partners finance, arguing that for various reasons its use in brownfield concessions. in the developing world was accompanied by an underappreciation of the risks that threaten Section nine discusses efforts by private profitability and cash flows of these deals. concessionaires to strengthen profits and cash flows generated by these projects. Section six argues that an underlying cause of these problems was inadequate project Section ten assesses whether or not PPI data preparation. from 2006 suggests that brownfield concessions may be poised for some kind of recovery. Section seven examines how other traditional forms of PPP functioned, relative to brownfield concessions, before and after the Asian Crisis. 3 2 BOOM AND BUST The 1997 Asian Crisis marked the end of a very concessions have nearly been eclipsed by other brief golden age of the brownfield concession as arrangements. Figure 2 shows the percentage of a mechanism for facilitating private investment total PPI investment in facilities accounted for in government-owned infrastructure facilities.5 by brownfield concessions. As Figure 1 indicates, the rise and fall of this mechanism were both sudden and dramatic. The Finally, and perhaps most importantly, the use of brownfield concessions, as measured by declining use of brownfield concessions for associated infrastructure investments, reached its investment in government facilities can now be lowest point in 2003 and has demonstrated low seen to have been a key factor in the overall PPI levels of use since then, with a sharp increase in market decline in such investments since 1997. 2006. The 2006 investment total matches the As Figure 3 shows, without brownfield level first achieved in 1993. concession contracts, as defined by the PPI Project Database, the track record of PPI-related But it is not just the declining investment investments demonstrates few of the "crash" associated with brownfield concessions that is characteristics commonly associated with PPI dramatic. As a PPI tool for generating over the last decade. In nominal terms, infrastructure investment, these kinds of investment figures for 2006 are 67 percent higher than for 1997. 5All investment figures in this note are nominal amounts representing millions of U.S. dollars. They include only capital investments in government-owned facilities or assets and exclude other forms of government revenue generated by PPPs, including licensing fees, lease payments, etc., which may or may not contribute to investment in infrastructure. 4 Figure 1: PPI Investment in Government-Owned Facilities via Brownfield Concessions (US$ billions) 25 20 15 10 5 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: World Bank and PPIAF, PPI Project Database. Figure 2: Investment in Government-Owned Facilities via Brownfield Concessions As a Percentage of all PPI Investment in Facilities 35% 30% 25% 20% 15% 10% 5% 0% 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: World Bank and PPIAF, PPI Project Database. 5 Figure 3: Brownfield Concessions Versus Other Kinds of PPI (US$ billions) 90 80 70 Brownfield 60 concessions 50 40 All PPI excluding 30 brownfield 20 concessions 10 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: World Bank and PPIAF, PPI Project Database. 6 3 EMPIRICAL EVIDENCE OF PROBLEMS No single factor accounts for the rise or decline brownfield variety. The study suggests that on in the use of a PPI mechanism in all situations. average, projects demonstrate profitability only In Latin America, for example, many of the after about 10 years. Up to that point, most attractive opportunities for brownfield concession shareholders earn negative returns on concessions were taken up in the early 1990s. their investments, even when adding in things After the Asian Crisis, public opposition to such as management fees, estimated privatization may have combined with the accumulated capital gains, and potential eventual economic recovery of some Latin investment markups. But this same study found American countries to diminish the attractions that 40 percent of the concessions in the sample for government officials of turning infrastructure did not appear to have the potential ever to service provision over to private operators. become profitable, with that percentage increasing to 50 percent for concessions in But Latin America is also the source of some of energy and transport sectors. the most compelling empirical evidence available on other key reasons why the use of The studies by Guash and Sirtaine suggest that brownfield concessions went into a steep decline lack of long-term profitability is not the only in the late 1990s. A recent study of concession problem with these concessions. The other contract renegotiations in Latin America by problem is the constant potential for cash flow Guasch (2004) suggests that cash flow problems crisis that most of these kinds of projects must and low levels of profitability are common live with over at least the first decade of their features of these arrangements. Using a existence. Sirtaine estimates that over the definition of concession that includes some history of these concessions, on average they greenfield projects and divestitures, the study have been unable to generate sufficient annual demonstrates that concessions in Latin America operating income, after taxes, to cover all of have registered a high incidence of their financial obligations. Only by adding in renegotiation, about 42 percent, coming after what the authors refer to as "indirect forms of only 2.2 years of operation. On average, the dividends" such as investment markups, transfer results of renegotiation favored operators, fees, and payments for capital appreciation paid mostly with improvements to cash flow and at the end of the concession period can the profitability such as delays in investment concession make a profit. In other words, these obligations (69 percent), reductions in projects must be able to weather years of cash investment obligations (62 percent), tariff flow stress before long-term profitability can be increases (62 percent), increased pass-through to achieved. For some large international tariffs of cost items (59 percent), changes in the operators, that kind of stress may be tolerable, asset base to impute rate of return (46 percent), but only as long as profitability over the longer and extension of the contract period to increase term seems likely and cash flows never become the present value of future cash flows (38 too negative. Sirtaine's study suggests that percent). many operators live through years of negative cash flows and constant uncertainty about long- Another recent study by Sirtaine et al. (2004), term profitability. For smaller, local or regional looks at the profitability of infrastructure operators, cash flow problems likely play a concessions in Latin America during the late leading role in contract distress and cancellation. 1990s, again using a broad definition of The fact that so many contracts are renegotiated concessions to include more than just the after only a few years, long before the contracts can establish their profitability, suggests that 7 cash flow problems may be important in unusual. They include things like unexpected precipitating renegotiation. decreases in revenues (e.g., unwillingness or inability of customers to pay bills), unforeseen Theoretically, profitability and cash flows are increases in investment needs, nonpayment of two different things. A business can be subsidy commitments by governments, or even "profitable," but if it does not have enough cash delays in disbursements of loan funding. All of on hand to pay bills, it experiences a cash flow these developments can precipitate cash flow crisis that can collapse the enterprise. Such a crises, particularly when large debt service crisis, rather than lack of profitability, is the payments are already straining cash balances. most frequent cause of business distress and These cash flow stresses are well documented in failure. For difficult and sometimes connection with various types of concessions; controversial infrastructure concessions, it is for example, in the African rail sector (Pozzo di now clear that disturbances to cash flows are not Borgo et al. 2006). 8 4 UNDERLYING PPP VULNERABILITY TO CASH FLOW STRESS AND PROFITABILITY PROBLEMS The vulnerability of concessions to problems protecting the sponsor's balance sheet from with cash flows and profitability is not too those responsibilities. If the SPV defaults on surprising. All long-term infrastructure PPPs loans, it is solely responsible for the involving significant levels of private investment consequences--the SPV's separate legal suffer to some extent from such vulnerability. It personality means that lenders have recourse to may be useful to explain this vulnerability whatever assets it may own, including equity before addressing concessions specifically. investments by the sponsors, but not to the sponsor's assets. This kind of limited recourse The project finance principles that underlie to the assets of sponsors and operators protects concessions as well as other forms of long-term them from possible bankruptcy caused by the infrastructure PPPs were well developed and financial problems of their large infrastructure highly successful long before the PPP boom of concession projects and shifts a significant share the 1990s. Project finance was already of project risks to lenders. Not only does the beginning to be used extensively on SPV mechanism encourage borrowing to fund infrastructure projects in the 1970s, particularly projects, but some sponsors believe that the after the viability of the approach was more money is borrowed, the more lenders must demonstrated in British Petroleum's North Sea be cooperative in times of stress. As the chief Forties oil field project in 1972 (De Lemos et al. financial officer of one private power generator 2000). The mechanism has continued to enjoy put it recently in a conference with potential success, but almost exclusively in the developed investors: "The legal ability to `walk away' from world. And this seems to be a key to a project that is significantly underperforming understanding the declining use of the gives substantial leverage in mechanism beginning with the Asian Crisis: its renegotiating/restructuring funding use in the developing world was compromised arrangements" (Williamson and Moore 2006). by serious misconceptions about underlying risks, perhaps because those structuring deals Debt service costs are prime factors in were overly optimistic in assuming that risk increasing the cash flow sensitivity of an levels were closer to developed country enterprise. Debt has important cash flow standards than they turned out to be. When consequences, because unlike dividends earned these risks materialized as serious problems, the by equity investors, debt service payments basic PPP structure did not cope well. (principal and interest on debt) must be paid according to preagreed schedules--they One reason for this vulnerability is the natural represent a constant, predictable drain on the tendency to try to maximize the debt-equity ratio cash of the enterprise. Dividends may be in PPP project financing. Because these are delayed as needed, for months or even years, in risky projects, for many reasons, sponsors and order to sustain the cash flow needed to allow operators needed a way to insulate themselves debt service payments in support of growth- from the financial liabilities of their projects and enhancing investments. This appears to have to share substantial levels of risk with other been common practice in the infrastructure parties. A project finance approach to capital concessions in Latin America referred to above. financing makes this possible through the But debt service payments are more difficult to creation of special purpose vehicles (SPVs). An delay--normally they must be made in a timely SPV is a separate corporate entity with a legal fashion in order to avoid default on obligations personality that allows the company to be fully to lenders. With most debt of this kind, default responsible for its own assets and liabilities, thus triggers payment acceleration whereby the total 9 amount owed becomes immediately payable. advantage of the boom in infrastructure Default on obligations to one lender also often development. Sometimes sponsors borrowed triggers "cross-defaults," or defaults and money to make equity investments, forcing their acceleration on debts owed to other lenders. dividend revenue to cover debt service payments Almost immediately the amount owed to lenders (Ehrhardt and Irwin 2004). Sometimes sponsors will exceed the value of the SPV, making it were permitted to count expected profits as their vulnerable to a legal finding of bankruptcy and, initial equity contributions. This sort of "in theoretically, liquidation. kind" equity could not provide financial support to the enterprise in the same manner as risk- Some experts have argued that a variety of bearing equity capital. Estache and Strong factors incentivized project sponsors or (2000) argue that obviously high-risk projects operators to use higher levels of debt in PPP took on heavy debt burdens and made too much projects than they would have otherwise--that use of the riskiest kinds of debt, including short- "too much" debt was used in the early 1990s. term debt, floating rate debt, and foreign debt for Tax regimes typically favor debt over equity projects generating local currency revenues. investment (with interest payments tax deductible, but not dividends). Regulators are To summarize, as a general rule the project reluctant to allow project companies to declare finance mechanism encourages sponsors and bankruptcy when long-term PPPs experience concession companies to maximize borrowing as distress, out of fear of service disruption, but a financing tool in order to maximize equity also because lenders may be antagonized and returns and facilitate competitive, lowest-cost privatization programs discredited (Ehrhardt and bidding, as well as control risks to their own Irwin 2004). Whatever the reason, high levels balance sheets. Of course, many factors are of debt clearly make it easier for concessionaires considered in reaching optimal debt-equity to extract favorable decisions from regulators targets for particular projects, and actual (de Fraja and Stones 2004). Termination decisions on concession funding and risk scenarios in concession contracts also tend to management frameworks are the result of favor only lenders, rather than equity investors, negotiations involving a number of key with total compensation in cases of premature players--sponsors do not necessarily have termination.6 Finally, higher debt-equity ratios advantages in the process that allow them may have also made some projects seem less unilaterally to determine the financing structure expensive and thereby helped sponsors compete that best suits them. Lenders aim to be effectively for projects awarded competitively appropriately remunerated for their share of on a lowest-cost-bid basis. risks; they use required debt-equity ratios as a risk mitigation tool, and also attempt to share Whether or not these particular incentives these risks with others. Above all, lenders try to actually resulted in more borrowing than would ensure sustainable cash flow for debt servicing have otherwise taken place is open to question, and often are willing to relax things such as debt because borrowing must be negotiated with service payment schedules if necessary to avoid lenders who have an interest precisely in cash flow crises. But in cases where the key ensuring that "too much" borrowing does not players, including lenders, share mistaken take place. But other experts have argued that assessments of risks, then "too much" borrowing the early 1990s PPP boom did in fact fuel ever- can lead to debt service costs that are larger shares of debt in project capital structures. instrumental in pushing concessions and other More debt and less equity meant that sponsors forms of long-term infrastructure PPP into cash could benefit from more projects, taking more flow crises and the urgent need to restructure contracts and borrowing arrangements. 6These termination scenarios are a matter of contract negotiation, but this kind of lender protection was a common feature of concession arrangements during the 1990s. 10 5 CASH FLOW VULNERABILITY PARTICULAR TO BROWNFIELD CONCESSIONS The PPI Project Database confirms that public or private company, infrastructure assets brownfield concessions were far more likely to (such as water pipes or rail or electricity lines) experience these kinds of contractual distress are usually difficult if not impossible to dispose than other form of long-term PPP. In terms of of for cash. numbers of projects from 1990­98, the percentage of cancelled or distressed brownfield This means that, practically speaking, there are concessions is 41 percent higher than that for no inventories, fixed assets, or intangibles to be greenfield projects, like BOTs (10.7 percent sold to enhance liquidity or pledge for additional versus 7.6 percent). In terms of affected loans. More aggressive marketing strategies are investment, brownfield concessions are three rarely helpful in selling infrastructure services. times as likely to experience these problems as Because service levels and quality are often greenfield projects (27.9 percent versus 8.6 stipulated in the concession contracts, these percent). In other words, from 1990­98, above- cannot be reduced without defaulting on average investment in brownfield concessions contractual obligations. Many brownfield was many times riskier in terms of contract concessions begin with severe cost-cutting distress or cancellation than for other forms of measures that governments may not be willing long-term PPPs. And the more investment, the to take responsibility for, so it is often difficult more risk when compared with other kinds of to make additional reductions--for example, in PPPs. the workforce--when cash crises occur after initial retrenchments have been made. And, of But why would brownfield concessions be more course, ultimate cash flow remedies such as prone to problems with cash flows and selling the business or merging with another profitability than other forms of long-term PPI? company are not available to infrastructure concession companies that do not own any Brownfield Concessions and Cash Flows company assets. A factor that adds to the potential cash flow problems faced by concessions, as defined here, Privatized infrastructure service operations is the fact that they have access to fewer of the actually own their assets, so asset disposal and remedies for dealing with the problem than do other cost-cutting measures can be done more other kinds of business enterprises or even other readily. Greenfield projects, such as BOTs, forms of PPP. This is because, as Guasch points involve new, more valuable assets, over which out, cash flow is the only asset "owned" by a private partners sometimes have considerable concessionaire, and only for the lifespan of the control. In terms of asset disposal and cost- concession contract. Unlike a privatized cutting flexibility, virtually every other form of business or a BOT project, a brownfield long-term PPP is in a better position to deal with concession involves transfer only of the right to cash flow problems than brownfield use the assets and operate the enterprise. Even concessions. in cases where contracts allow concession companies to own newly created assets, disposal Brownfield concessions also tend to be at a of these assets to help with cash flow problems disadvantage when it comes to the ability to almost always must be agreed to by government increase normal operating revenues. Although partners and done in a way that does not not a defining characteristic, most of the compromise service quality. Of course, for any brownfield concessions registered in the PPI database involve facilities that distribute services 11 (water, electricity, toll roads) to end users who before major tariff restructuring can be pay via service fees--privatized facilities and considered have proved to be problematic, greenfield projects are much less likely to particularly during the first five years of generate most of their operating revenues in this implementation. manner. Price increases are the most powerful tools for strengthening brownfield concession Again, the dependence of many brownfield cash flows, but they are unpopular during the concessions on service fees paid in local early years of these projects and often encounter currency by end users meant that exchange rate political opposition, even when they are risk was a particular problem for such projects. permitted by contracts. Operators and sponsors often had to borrow in foreign currency to raise investment funds The tendency of brownfield concessions to because so few of the beneficiary countries had involve retail service distribution means that domestic capital markets that could finance large they tend to be more sensitive to government projects. Price adjustments for exchange rate regulation of tariff setting than most greenfield changes often encountered political opposition. projects. Estache, Guasch, and Trujillo (2003) Contractual provisions designed to mitigate this have argued that widely used approaches to the risk by indexing prices to inflation were difficult economic regulation of infrastructure services to enforce under the conditions of extreme may have also contributed to the cash flow exchange-rate volatility that occurred during the problems of concessions. Estache maintains that Asian Crisis. the "fixed-price" or "price cap" regulation of infrastructure services has allocated more risk to Lenders can deal with cash flow problems by concessionaires, put more stress on cash flows, exercising step-in rights and arranging the and increased the chances of project collapse. transfer of a concession operation to a new Under a price cap approach, regulated user private company. But for concession operators tariffs are set in advance in real terms, typically and sponsors, short of an infusion of more debt for periods of five years. During each period, or equity, contract renegotiation to delay the regulated concession company bears the risk investments or increase tariff revenues are the of all cost increases (except for that of general principal ways in which cash flow problems are inflation). Every five years the tariff framework addressed. can be "reset" to account for intervening changes in service costs and demand. Price cap Insufficient Return on Investment regulation is strongly associated with the In addition to cash flow sensitivity, concessions renegotiation of concession contracts (Guasch often simply lacked the ability to generate 2004), and this may be because this regulatory positive cash flow and profits. The study by approach increases risks for operators as well as Sirtaine et al., mentioned above, concludes that pressures on cash flow. of the 50 percent of concessions that finally achieve some profitability after about 10 years The principal alternative form of regulation, of implementation, most do so only by cost-plus or rate-of-return regulation, allows supplementing normal operating revenues with price adjustments much more frequently, often things such as management and technical allowing for more pass-through costs. Often assistance fees and accumulated capital gains referred to as "low-powered regulation," this (which operators are eventually compensated approach causes much less stress on cash flows for), as well as markups on investments. A and is much less strongly correlated with hallmark of brownfield concessions in contract renegotiations. Pure forms of either of developing countries has been the fact that much these two types of regulation are rare--price cap more investment in assets is typically made than regimes usually allow for more frequent can be adequately supported by project revenues adjustments for inflation as well as some pass- generated by subsequent improvements in through costs. But any regulatory regime that infrastructure service delivery. makes operators wait significant periods of time 12 Too much non-revenue-generating invest-tment. 2000). Finally, to make matters even worse, the A fundamental problem for many brownfield most visible problems with infrastructure service concessions was that the assets turned over to provision, such as the poor condition of assets, operators were in extremely poor condition, inadequate management, generally poor service, poorer than many of the parties to the etc., often were manifestations of more profound transactions realized. Many governments problems that brownfield concessions were not became desperate enough to engage private designed to directly address, such as lack of companies in infrastructure service delivery only sectoral and institutional reform and corruption. after assets were no longer capable of generating In hindsight, it was a fundamental mistake to robust cash flows without massive investment or assume that these kinds of concessions--or any replacement. The decision to engage the private kind PPP contract--could solve such problems.8 sector was almost never taken because assets could no longer generate surpluses for Inadequate revenue. The other side of the government--it was normally taken because problem of investments generating too little these facilities operated in a state of constant operating revenue had to do with the fact that cash flow crisis, with revenues not even revenues were simply not robust enough in covering operations and maintenance expenses, many infrastructure sectors to support much less capital investment liabilities or investment without some kind of subsidization. obligations. One reason why asset disrepair was We now know that proponents of the brownfield often underestimated was the fact that typically concession mechanism, as it was used in very little performance data was available on developing countries, were unrealistic in their these systems, the status of asset repair, and expectations that user fees could cover the full even the age of assets. Information about how costs of operations and investments. User fees well they provided services, or even to whom represented a largely untapped source of services were provided, was often missing.7 revenue--something new like this was necessary if a government was going to turn over operation In many cases operators knew what to fix, and in (but not ownership) of a service system to a what sequence, only after months of operation. private company and expect the company to But meanwhile on many brownfield concessions improve and run the system without any operators were incentivized to make major government financial support. The brownfield investments as early as possible in the lifespans concession mechanism was largely associated in of these projects. This was done, often at the the early 1990s with projects involving insistence of government partners, so that distribution of infrastructure services to end concessionaires could begin showing positive users who would pay for the services via fees. results as soon as possible. Sponsors and Over two-thirds of the concession-related operators also wanted the maximum amount of investment commitments made in 1997, the peak time allowed under the contract to recoup year for concessions, involved projects of this investment costs from operating revenues. But kind.9 this meant that most of the investment was made at a time when demand risk was high, and the Many infrastructure services possess a high project was generating much less revenue than it degree of what economists refer to as would be expected to do later, after investments "excludability," i.e., the degree to which a had improved cash flows (Estache and Strong 8This is one of the key lessons learned in a recent 7This is now such a widely understood aspect of doing assessment of the characteristics of well-performing water infrastructure PPPs that private companies are in danger of utilities. See Baietti, Kingdom, and van Ginneken 2006, p. accusations of negligence if they take at face value 25. government data on public enterprise performance. See the Lawyers' Environmental Action Team (2007). 9This is an estimate by the author, based on a review of projects included in the PPI Project Database. 13 potential user can be excluded from access to the caused severe cash flow problems for services for not meeting conditions set by the companies, including infrastructure SPVs, which supplier. In the early 1990s, mainstream were massively overleveraged, often with economists were advocating the use of foreign currency debt (Pomerleano 1998). As excludability to ensure that "prices be based on the crisis unfolded it underscored all of the the most economically efficient way to recover above-mentioned weaknesses and problems, total costs" (Israel 1992, p. 27). Subsidies, and including the fragility of the concession even cross-subsidization, were discouraged as mechanism and inadequate returns on being costly and inefficient. Even as the Asian investment, as well as debt that under the Crisis was beginning to unfold, experts on circumstances turned out to be "too much." concession contracts were still making the same basic arguments in favor of cost-covering tariffs, The materialization of these risks was supposed and against subsidies and cross-subsidization to trigger risk mitigation mechanisms, the most (Kerf et al. 1998, p. 34). But as has now become important of which were those provided by evident, cost recovery in many infrastructure state-owned enterprises and utilities or the sectors, particularly water, has not been governments themselves, principally in the form attempted, much less achieved. Even high- of off-take agreements and project guarantees. income countries routinely subsidize water costs But the crisis also forced many governments to (Foster and Yepes 2006). The promotion of full realize that their understanding of how PPP cost recovery from user fees, as a fair and contracts were supposed to work was not practical solution, was fundamental in helping to accurate. In the early 1990s, many governments make the brownfield concession concept seem were convinced to use these mechanisms workable. A legacy of this kind of overselling is because it seemed that they could turn over all the large number of anti-privatization critics responsibility for infrastructure provision to who continue to equate PPPs with full cost private partners who would cover all costs and recovery, private operator profits, and accept all related business risks. The nature of exploitation of the poor.10 governments' contingent liabilities associated with these contracts was often completely The Failure of Risk Mitigation ignored or misunderstood (Pomerleano 1998). Whatever the vulnerability of brownfield Governments in fact retained considerable concessions to cash flow and profitability contingent liabilities for these operations, via problems, these and other risks associated with off-take agreements and outright guarantees, as these projects should have been identified, well as under routine early-contract-termination allocated to whomever was best positioned to scenarios. But under the intense pressure of the manage them, and appropriately mitigated when Asian Crisis, many governments simply the time came. The Asian Crisis and the repudiated these obligations. international aftershocks that followed demonstrated the failure of risk mitigation for a A managing director of Moody's Investors variety of PPPs, but particularly for brownfield Service notes that this repudiation was most concessions. The PPI Database statistics cited at clearly evident in Asia, where the beginning of section 3 indicate that by 1998, these kinds of concessions had experienced far "... structures set up to mitigate these higher levels of distress and collapse than any economic risks unraveled almost as soon as other forms of PPP. These economic shocks they were put to the test. Counterparties who had accepted risk repudiated that risk when it occurred. Legal systems simply proved 10Some private sector operators still fan the flames of this unable to enforce the contracts as agreed. debate by arguing for the necessity of "full cost recovery" Regulatory approvals that had been held out and the elimination of cross-subsidies in developing as straightforward proved anything but when countries, even in sectors such as water supply (Martin projects moved to the operations stage. And, 2005). above all, the political will that had been so 14 apparent getting the projects financed but with unhappy private partners tolerating evaporated when the projects needed support what they perceived to be less than satisfactory in a time of crisis" (Marshella 2001, p. 31). profits and cash flows rather than face penalties for default and early termination. The off-take agreements, guarantees, and other legal protections were the risk-mitigation The failure of these backstop mechanisms meant mechanisms that were supposed to provide that most private players no longer viewed sponsors, operators, and lenders with at least brownfield infrastructure concessions--as they some protection from many of the other had been developed and implemented in the problems outlined in earlier sections of this early 1990s--as a viable project type in these paper. Instead, the economic crisis led to countries. The precipitous postcrisis drop in the hundreds of project defaults, renegotiations, use of brownfield concessions registered by the restructurings, and rating downgrades. PPI Project Database seems to be, in large part, a Anecdotal evidence suggests that many projects reflection of the private sector's new perspective continued as partially restructured arrangements, on these projects. 15 6 A CRITICAL UNDERLYING PROBLEM: POOR PROJECT PREPARATION problems and adjust the contract after a period The failure of risk mitigation became apparent of implementation. Occasionally, bidders only as a result of the fallout of the Asian Crisis. purposely underbid for projects, assuming that Cash flow vulnerabilities of the brownfield contracts would later have to be renegotiated. concession mechanism and the potential for Without independent assessments done by or on inadequate returns on investment, however, behalf of governments in preparing the early should have been apparent much earlier. Why brownfield concessions the bidders and were so many of these contracts completed government officials had different opinions before these weaknesses were noticed? A large about how much investment was going to be part of the answer seems to be that the needed to turn service delivery into a preparation of these projects, from commercial proposition. conceptualization, feasibility assessment, structuring, and negotiation was incomplete or Perhaps the first notable example of the low-cost of very poor quality. preparation problem was the Buenos Aires water concession--one of the first large brownfield Low-Cost (and Low-Quality) Preparation concessions--signed in December 1992. One result of the notion that their problems were According to Alcazar, Abdala, and Shirley being handed over to private companies for (2000), a "defining feature" of the tender solution was that governments (as well as donors process was poor information. Because of a and development agencies) often were unwilling perceived need to accelerate the procurement to spend time or money preparing brownfield process, the government collected very little concessions, doing feasibility studies, assessing prebid information on the condition of the PPI contracting options, etc. In many of the service system, and officials refused to take any contracts signed in the early 1990s, all of this responsibility for the quality or accuracy of the work was assumed to be the responsibility of information provided to bidders. The winning potential private partners; part of their normal bidder used the first four pages of the bid due diligence, because if the project ultimately document to describe the serious lack of failed, it would be at the sole cost of the private information made available for bidding partners. But for existing facilities that often purposes. The poor quality of the information were over 50 years old, governments needed made the information content of bids unreliable independent, impartial assessments of the and may have led bidders to prepare bids under condition of the existing assets in order to be the assumption that they would be allowed to able to identify the objectives of brownfield renegotiate their contract as new information concessions and evaluate bids. By leaving such came to light. The concession agreement stated assessments to bidders who spent different that the operator would assume all information amounts of time and resources on feasibility risks, and that "claims based on defects of studies and asset reviews, bids were often information provided will not be allowed" difficult to compare or were (Aguas Argentinas S.A. Concession Contract simply based on incomplete or inaccurate views 1993, p. 8). In fact, the operator did eventually of infrastructure investment needs. Sometimes make such claims based on what they said was the intense competitive pressures of lowest-cost new information about the condition of the procurement forced bidders to reduce assets. A protracted series of contract preparation costs as much as possible in order to renegotiations ensued. win the project, hoping to be able to identify Buenos Aires was not an isolated example of this approach to project preparation. By the time 16 a 30-year brownfield concession contract was customer records were incomplete or missing-- signed for water services in Nelspruit, South there was often no way to tell how many end Africa, in 1999, the Buenos Aires approach to users were connected to water systems, for preparation was still being intentionally copied, example, much less how many were paying their in the interest of cost and speed (Kotze, bills. Ferguson, and Leigland 2000). No formal feasibility studies or asset reviews were done by Many private operators now agree that because or on behalf of the municipality. Prospective of the poor quality of government information concessionaires were told to base their bids on the condition of these service systems, exclusively on their own due diligence. The experience actually running the facilities being eventual contract specified that the arrangement taken over is the only reliable way of generating was entered into on a "voetstoots" basis (an reliable information about the condition of assets Afrikaans term meaning, "as is"), with the and the need for investment, as well as the municipality taking no responsibility for willingness and ability of customers to help pay information regarding the "nature, condition, costs. Investments can be planned and fitness of purpose, merchantability or suitability implemented more prudently and sometimes of any existing asset" (Nelspruit Concession reduced through a combination of efficiency Contract 1999, p. 74). A series of gains and demand management. For example, renegotiations, not dissimilar to the ones in the potential for exploiting efficiency gains as an Buenos Aires, followed. alternative to solutions involving capital investment seems apparent among African water Another notable aspect of low-cost preparation, utilities (Estache and Kouassi 2002). PPP water particularly in the early 1990s, involved the projects in Maputo, Mozambique, and Zambia's willingness of public and private partners to Copper Belt have demonstrated the value of accelerate deal closure by delaying agreement arrangements whereby private operators gain about things like regulatory or contractual insights about investment needs while operating arrangements for increasing tariffs or coping the facilities via management contracts or other with unexpected changes to the concession. arrangements that do not require operators to Bidders often were prepared to commit to make large investments while assuming concessions in the absence of such arrangements unknown risks.11 based on reassurances from governments that such issues would readily be resolved. Often, Shortcomings of Cost-Benefit Analysis the rhetoric failed to match the reality and It is likely that the preparation of concession concessionaires faced lengthy hurdles in, for projects was likely also affected by several example, securing contractually mandated tariff limitations in the way in which project appraisal increases. techniques were applied to anticipate and avoid problems with cash flows and profitability. Difficulties in Undertaking Feasibility Studies Even in cases where one party or another was First, quantitative techniques for assessing willing and able to undertake full feasibility financial and economic costs and benefits of studies to accurately determine the condition of infrastructure assets to be concessioned and the key elements needed to make a concession 11This idea is not at all new, but it raises procurement viable, the task often turned out to be difficult if issues regarding how to structure the initial not impossible. As noted above, not only were competition for the management contract, and how to the assets dilapidated and neglected, but handle the involvement of the incumbent operator management information and basic when enough information finally becomes available recordkeeping was also often nonexistent. to bid out longer-term contracts involving significant Historical performance data was inaccurate or capital investment. For an early summary of some of unavailable. The condition of underground these issues, see Klein 1998. pipes was impossible to evaluate. Even 17 prospective projects were in many cases simply poor governance and the behavioral responses to not used. A government's decision to grant a controversial projects by beneficiaries and brownfield concession, and the view that no bureaucrats. Bourguignon speculates that the practical alternative existed for solving problems inability to cope with these problems accounts with existing, badly maintained facilities, often for the declining use of cost-benefit techniques made it appear unnecessary to undertake sound in infrastructure project appraisals since the quantitative analysis of financial, much less 1980s--the techniques simply became economic, costs and benefits or comparisons of associated with unrealistic evaluations and failed PPP options as compared to traditional public projects and stopped being used extensively. options for project financing and management. This was particularly true in cases where the Third, Flyvbjerg (2005) documents the huge private sector was expected to carry out project mistakes routinely made in the estimation of preparation as part of their normal due diligence. financial costs and benefits associate with To the extent that techniques were used by or on infrastructure projects. In a survey of 58 rail behalf of government partners, they often served projects, he found that costs were to justify rather than evaluate projects.12 To the underestimated by an average of 45 percent and extent that such techniques were actually used demand forecasts overestimated by an average by bidders, the analysis may in some cases have of 51 percent. Flyvbjerg concludes that such served the needs of opportunistic bidding-- consistently large mistakes in cost-benefit allowing the preferred bidder to win the project estimations cannot be attributed exclusively to with an unrealistically low bid, with the faulty techniques, but can only be explained by expectation that contract terms could be things like "optimism bias" and "strategic renegotiated later. In any case, for projects like misrepresentation." the Nelspruit water concession mentioned above, the assumption was that because the In retrospect, it seems likely that many model was informed by an earlier, successful brownfield concessions were not viable to begin concession such as the Buenos Aires project, no with, but were nevertheless accepted as win-win real cost-benefit analysis needed to be carried solutions by many politicians and policy makers, out by, or on behalf of, the municipality. as well as some representatives of donor organizations and international financing Second, Bourguignon (2006) conjectures that institutions. In the rush to confirm that already when economic cost-benefit analysis was approved concessions were cost-effective applied in the preparation of these projects, the solutions to huge infrastructure problems, techniques may have been overwhelmed by a analytical techniques were overwhelmed with combination of bad data and the complexities of project complexities, coopted to rubber-stamp mixed public-private funding options, as well as political decisions, or simply not used at all. so-called "political economy" issues such as 12In his seminal studies of transit forecasting conducted in the late 1980s, Wachs found "nearly universal abuse" of this kind (1990). 18 7 PPP ALTERNATIVES TO CONCESSIONS Overview The Asian Crisis underscored for lenders (and Sponsors of course did not respond to lenders' everyone else) that the risks associated with concerns by compensating with more equity brownfield concessions, which they shared investment--the Asian Crisis also underscored along with sponsors and operators, had not been more clearly than ever before the risks inherent correctly assessed. By 2000, Estache and Strong in contract structures that are highly sensitive to had concluded that a key reason why cash flow problems. Problems with regulatory infrastructure concessions had become less bias also remained in many markets, and these common since the Asian Crisis was because added to the already strong reluctance of lenders had increased the price of such debt in sponsors and operators to increase equity levels order to be compensated for their revised in these projects. perception of the risks associated with these deals. Lenders also began requiring more and As Figure 4 illustrates, even before the Asian stronger security features, as well as larger Crisis, other forms of PPI, particularly shares of sponsor equity. greenfield projects, were already generating significant levels of investment in infrastructure Security features providing some recourse to facilities. After 1997, as investments associated sponsors and operators were already being used with brownfield concessions began to decrease, in the early 1990s, including sponsor support these other forms of PPI took on even larger agreements (e.g., to top-up operating revenues) roles in such investments. Greenfield projects, and technical support agreements (whereby involving construction of new facilities, have lenders could call in technical support--at generated far more investments since 1990 than sponsors' expense--when they thought it either of the other two major forms of PPI. necessary). Various covenants were used to Divestiture, involving full or partial sales of protect lenders by forcing shareholders to bear existing facilities, is of course better known as a responsibility for financial underperformance. way of generating revues from the sale of assets "Permitted distribution" tests precluded dividend (and sometimes the licensing of privatized distributions unless financial performance tests facilities) rather than for infrastructure were met. "Additional indebtedness" tests investment. Briefly, during the mid-1990s, restricted debt-financed activities as a way of revenues from sales and licensing far exceeded keeping projects from risking cash flow investment in facilities. But beginning in 1998, problems by overleveraging (Forsgren et al. the latter has exceeded the former by about 7 2006). After the Asian Crisis, all of these kinds percent, demonstrating at least that this of features were strengthened and more mechanism is emerging as a way to stimulate consistently used at the insistence of lenders. infrastructure investment as much as it is a mechanism for generating revenues that can be used for any government purpose. 19 Figure 4: Investment in Facilities by PPP Contract Type--1990­96 (US$ billions) 70.0 60.0 50.0 40.0 Concession Divestiture 30.0 Greenfield 20.0 10.0 0.0 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: World Bank and PPIAF, PPI Project Database. Brownfield Concessions Figures 5 and 6 tell the story of brownfield Figure 6 indicates the sectoral focus of these concession investment since the early 1990s. Of brownfield projects, largely transport, followed the brownfield subtypes, BROT peaked in 1997 by water and sanitation. Of the various transport at over twice the total annual investment in subsectors, brownfield concessions were most facilities generated by the next most widely used closely associated with toll roads (78 percent of subtype. These were the classic brownfield brownfield transport concessions). In the water concessions, involving the construction of some sector, 87 percent of brownfield concessions new infrastructure assets to improve or extend involved distribution, arguable the riskiest existing facilities, in addition to rehabilitation. activity for an operator. With the decline in As noted earlier, three-fourths of the investment facilities-related investment illustrated in Figure in government facilities associated with 5, the toll roads and water distribution systems brownfield concessions involved the distribution dramatically lost a major source of investment of infrastructure services to end users who paid after the Asian Crisis, particularly for system for operating costs via fees. extension. 20 Figure 5: Investment in Facilities by Brownfield Concession Subtype--1990­06 (US$ billions) 18.0 16.0 14.0 12.0 BROT 10.0 RLT 8.0 6.0 ROT 4.0 2.0 0.0 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: World Bank and PPIAF, PPI Project Database. Figure 6: Sectoral Investments by Brownfield Concession Subtype--1990­06 (US$ billions) 80 70 60 WSS 50 Transport 40 Telecom 30 Energy 20 10 0 BROT RLT ROT Source: World Bank and PPIAF, PPI Project Database. Divestiture representing a minority stake in a state-owned From 1990 through 1998, divestiture of enterprise, just edges out full divestiture as a infrastructure facilities primarily served its mechanism for generating facilities investment traditional purpose of generating government over the 1990­05 period (see Figure 7). Partial revenues from sales or licenses. Over that divestiture implies continuation of some period, these kinds of revenues were 56 percent government ownership and often a retention of higher than associated investment in facilities. some control (through regulation if not via share But beginning in 1999, the emphasis reverses; ownership). As such, partial divestiture takes from 1999 to 2005, investment numbers are 35 full advantage of the attributes that have long percent higher than sales revenues. Partial been viewed as making it an alternative to divestiture, for example via a sale of shares brownfield concessions, because it is a 21 reasonable way of involving the private sector in offered for either sale or concession via the same investment and management of existing facilities procurement process (Leigland 2000). that remain to some extent under governmental ownership and control (Kerf et al. 1998). This is Figure 8 illustrates the focus of divestiture on particularly true in partial share sales involving telecommunications (60 percent of which strategic equity partners. The question regarding involves mobile telephony in the case of partial whether to sell or concession an existing facility divestiture). In energy, 92 percent involves is not an uncommon one addressed as the PPI electricity, and about a third of that involves options for facilities are considered. This is generation only. Virtually no water or transport especially true in the case of facilities for which facilities benefit from investments related to government ownership is not considered a divestiture, although water-utility-share sales strategic necessity. In some cases, facilities are increasingly are being discussed in East African countries such as Uganda and Kenya. Figure 7: Investment in Facilities by Divestiture Subtype--1990­06 (US$ billions) 14.0 12.0 10.0 8.0 Full 6.0 Partial 4.0 2.0 0.0 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: World Bank and PPIAF, PPI Project Database. 22 Figure 8: Sectoral Investments by Divestiture Subtype--1990­06 (US$ billions) 100 90 80 70 WSS 60 Transport 50 40 Telecom 30 Energy 20 10 0 Full Partial Source: World Bank and PPIAF, PPI Project Database. Greenfield Projects This is not particularly surprising considering Figure 9 indicates that "merchant" projects have the fact that such projects involve conventional been the most powerful greenfield generators of private investment and do not rely for facilities investment since 1997 when both sustainability on government-implemented risk "build, own, operate" (BOO) and BOT projects mitigation techniques such as off-take began to decline with the Asian Crisis. The PPI agreements and project revenue guarantees. database defines merchant projects in the Thus the failure of such risk mitigation measures following way: during the Asian Crisis had virtually no impact on merchant projects. A private sponsor builds a new facility in a liberalized market in which the government Figure 10 indicates that most merchant project provides no revenue guarantees. The private investments have been in telecommunications developer assumes construction, operating, (with most of that investment going into mobile and market risk for the project.13 telephony), with a small fraction going into energy, typically in the form of merchant power This subtype of PPI lost pace only marginally projects. after the Asian Crisis and has proved to be far more resilient since then, when compared to any About three-fourths of facilities investment other form of PPI arrangement.14 generated by BOO projects has gone into energy; 86 percent of that has gone into electricity, almost all of which (96 percent of the 13World Bank and PPIAF, Private Participation in electricity total) has involved generation. BOO Infrastructure (PPI) Project Database (ppi.worldbank.org). of course is the "public-private partnership" option normally thought of as involving 14Although merchant plants are included in the PPI Project Database as an important form of "private participation in infrastructure" (or PPI), they are not classifiable under most definitions of "public-private partnership" because of the absence of risk sharing between private and government partners. 23 the most extreme degree of private involvement, investment levels. Most BOT investment has and is virtually indistinguishable from other been in energy, with 88 percent in electricity forms of conventional private investment, generation. Transport follows, with several including merchant projects, except that it subsectors benefiting, including toll roads (57 normally involves government support via percent), ports (18 percent), and rail (17 guarantees or offtake agreements. Again, as percent). In other words, energy and especially with divestiture, BOOs generate virtually no transport facilities involve assets that investment in water or transport facilities. governments want eventually to be transferred back to their control. Again, water and As Figure 10 indicates, unlike BOOs, BOT sanitation register very small amounts of BOT- projects suffered a decline with the Asian Crisis related investment. and only in 2006 regained their 1997 peak Figure 9: Investment in Facilities by Greenfield Subtype--1990­06 (US$ billions) 40.0 35.0 30.0 Merchant 25.0 BOO 20.0 BOT 15.0 BLO 10.0 5.0 0.0 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: World Bank and PPIAF, PPI Project Database. 24 Figure 10: Sectoral Investments by Greenfield Subtype--1990­06 (US$ billions) 250.0 200.0 WSS 150.0 Transport 100.0 Telecom Energy 50.0 0.0 BOO BOT Merchant Source: World Bank and PPIAF, PPI Project Database. Leases and Management Contracts Figure 11 indicates some growth since the Asian but usually not enough to generate the cash Crisis in the use of nonequity contracts such as flows needed to back borrowing for capital management contracts and leases (Figure 12 investment. There is, however, a possibility that indicates the sectoral distribution of such efficiency levels are so low in some regions and contracts). But they still play limited roles in the sectors that with significant efficiency overall PPI picture, probably because they lack improvements revenue flows would increase associated investment by private partners. The enough so that much of the currently estimated same is true for corporatization and borrowing needs would disappear (Estache and commercialization efforts, now increasingly Kouassi 2002). There also seems to be used, for example, in the African water sector as increasing use of management or lease contracts an alternative to PPP mechanisms such as in tandem with capital investment programs brownfield concessions. Management efficiency funded and implemented by governments and can be improved, donors. These contracts are discussed in the next section. 25 Figure 11: Lease/Management Contracts as a Percentage of All PPI Projects--1990­06 (numbers of projects) 12.0% 10.0% 8.0% Lease 6.0% Mgmt Contract 4.0% 2.0% 0.0% 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: World Bank and PPIAF, PPI Project Database. Figure 12: Sectoral Distribution of Lease/Management Contracts--1990­06 (numbers of projects) 120 100 80 WSS Telecom 60 Transport 40 Energy 20 0 Lease Mgmt. C. Source: World Bank and PPIAF, PPI Project Database. 26 8 HOW GOVERNMENTS ARE REDUCING PRIVATE PARTNER RISKS ON BROWNFIELD CONCESSIONS the least likely to be implemented over the near Traditional brownfield concessions will avoid term. extinction to the extent that they can satisfy the risk-reward requirements of potential private At the other end of the risk-reward spectrum are partners. Those requirements have changed over projects involving one or more of the opposite time, and in some cases have been characteristics that effectively eliminate (or accommodated by adjustments to the basic reduce) these risks: (1) the construction and use brownfield concession model. of new assets, as opposed to taking over existing assets; and (2) the provision of wholesale as The Changing Nature of the Private Sector's opposed to retail services, with payments made Risk-Reward Requirements by governments backed by government- The Asian Crisis sharpened investors' guaranteed offtake agreements. Thus, stand- understanding of the risk-reward ratios alone greenfield independent power producer associated with different government (IPP) projects, backed by take-or-pay offtake infrastructure sectors, largely because it brought agreements, represent lower-risk projects than, into clearer focus the cash flow implications of for example, retail electricity distribution risks associated with brownfield concessions. In projects. It is not surprising that IPPs now any sector, the risks are now viewed as account for three-fourths of all energy projects particularly high for projects involving involving significant private sector investment. contractual commitments by private partners to assume either or both of the following risks: Cellular telephony further illustrates how the risk-reward calculus works. The fact that these · Investment risk associated with are mostly retail operations is more than made investment in badly maintained up for by the profits resulting from the government assets, particularly when the popularity of the service, the widespread condition and performance of assets is willingness to pay for the full cost of service, not well documented. and the fact that infrastructure investment is relatively low compared with other · Demand risk associated with the infrastructure sectors. In fact, the potential willingness or ability of end users to pay rewards of investment in cellular telephony so for infrastructure services, particularly outweigh the risks that private operators and where the willingness to pay is volatile, their financial backers are increasingly willing to governments are unwilling to entertain pursue such projects on a merchant basis where the idea of subsidies for lowest-income possible, without government mitigation of users, and the possibility is strong that risks. By not sharing project risks with regulatory processes may be subject to government partners, operators are not obligated political capture. to share returns, and can avoid the uncertainty sometimes associated with revenue or regulatory support from government partners. Projects that combine these characteristics, such as water concessions involving investment in Government Assumption (or Sharing) of existing facilities and the provision of retail Investment Risk distribution services to end users, represent the In order to deal with the risks associated with riskiest end of this continuum of projects, and private investment in badly run-down government assets, governments are responding 27 by becoming financiers of brownfield network.16 With this approach, the concessions or otherwise reducing the costs and challenge is to manage the private partner's risks of borrowing for private concessionaires. incentives so that investments are cost- Some of the ways in which this is done include effective. the following: · Government provision of capital finance. · Government implementation of capital A variation on this theme, which has been projects. used frequently with brownfield rail Government (often with the help of donors concessions in Africa, involves sourcing of or multilateral lenders) may assume the investment funding by governments, who investment risk associated with a project by then lend to the concessionaires, or in some sourcing the financing and managing the cases make capital grants. This kind of improvements. In such a situation, the funding can be sourced as relatively private partner may enter into what normally inexpensive sovereign loans from bilateral would be considered a lease or management donors or international financing contract to operate the facility. Sometimes institutions. Concessionaires usually benefit the role of the private partner is extended to from below-market interest rates, longer include the identification of investment repayment periods, and, sometimes, grace needs and management of the use of the periods. All of this can substantially investment funding.15 A recent example of improve project cash flows, especially in the this approach in Africa is the management early years of these concessions when contract for water services in Ghana. The investments often are at their highest levels management contract involves operations and concessions most in danger from cash and maintenance by the contractor, but is flow crises. In the 1990s, this kind of on- paralleled by a $120 million capital lending was usually done with a markup to investment program. The contractor the concessionaire. As commercial lending prepares and submits a "Suggested Capital for brownfield concessions began to Investment Report," which the government diminish after the Asian Crisis, the markups may or may not accept as guidance for were slashed, and in some cases have capital investments. Sometime before the disappeared entirely (Pozzo di Borgo et al. expiry of the five-year management 2006). In effect, this kind of government contract, the government intends to involvement amounts to project competitively bid out a long-term lease subsidization and can easily distort the contract, presumably to manage rollout of economics of the respective projects. This the larger investments and operate the same kind of mechanism can operate in expanded system (Ghana Management cases of partial divestiture, where Contract 2005). Another recent example of governments may make shareholder loans to this approach is the Kenya Power and enterprises that are managed by strategic Lighting Company electricity distribution equity partners. India has made notable use management contract signed in 2006, paralleled by a multidonor investment program focusing on the country's electricity transmission and distribution 16A third noteworthy example of this approach is the Dar es Salaam 10-year lease contract signed in 2003 and canceled in 2005. The fate of this contract underscores the fact that government financing of 15Note that without substantial direct investment by capital improvements, in parallel with private the private partner in such deals, these contracts are management of operations (including billing and usually classified in the PPIAF-World Bank PPI collection), maintenance, and new construction does Project Database as management or lease contracts, not relieve private partners of all project risks and rather than concessions. responsibilities. 28 of capital grants to cover significant example, such a commitment was an percentages of construction costs. important element in the 1997 Manila water concession contracts, in which the As governments begin to assume or share government promised concessionaires project investment risks they also begin to "expiration payments" at the end of the experience conflicts of interest. Can contract period equal to the net present value governments act simultaneously as of the unamortized investments. But the financiers interested in the financial Asian Crisis raised uncertainty in the minds sustainability of projects, and regulators of investors about the willingness and ability charged with protecting the interests of of many governments to make good on users? Will a government allow a potentially large commitments of this concession company, in which it has nature.17 This kind of concern, in turn, invested substantial amounts of capital, to raises the possibility that concessionaires go into bankruptcy? Professional economic might slow down the pace of asset regulators are charged with balancing the investment during the last years of a interests of different stakeholders, but the contract. Governments and multilateral independence of such regulators from lenders found a new way to deal with this political capture has turned out to be much problem in the Kenya-Uganda rail less easy to secure than hoped (Eberhard concession, which closed at the end of 2006. 2007). It is clear that when concessions The end-of-contract payment commitments involve government finance, renegotiation is by the two governments were backed by more likely (Guasch 2004). This may be guarantees made available by the World because operators correctly realize that Bank. Guarantees of this kind may become renegotiation in such situations has a better more widespread, but require sophisticated chance of success because of the conflicting monitoring of asset amortization over the interests of government partners. Finally, lifetime of the contract, which in the Kenya- the use by private companies of debt for Uganda case was done using a specially which governments ultimately are designed "conceded assets account" responsible also raises questions about (Matsukawa and Habeck 2007). efficiency incentives for private partners, not to mention the possibility of excessive · End-of-contract payments for the use of profits for concessionaires. government assets. In some cases governments may be willing More (or Guaranteed) Remuneration for to reduce the pressure on concession cash Private Partner Assumption of Investment flows and at the same time improve or at Risks least rationalize remuneration for investment In addition to sharing investment risk, risk by changing the fee structure of the governments are also finding better ways of project. Fees are often used in brownfield remunerating concessionaires for taking on risks concessions to reflect the cost to the related to investment. government of providing assets for use by the concessionaire. These kinds of fees are · Remuneration for capital improvements. usually structured as lease payments and One example of this involves guaranteed often reflect the debt service costs to remuneration for unamortized investment when the concession comes to an end. Sirtaine et al. (2004 argue that fair 17In fact, in the Manila case the government never remuneration for capital improvements at expected to make these expiration payments. Government the end of the contract term is essential in officials intended to bid out the project again, after the first making such arrangements profitable. The contracts terminated, and have the new concessionaires promise to do this has long been a standard make the expiration payments to the previous contractors feature of brownfield concessions. For (Dumol 2000). 29 government associated with the assets · Government pays for bulk services, but involved. But concessionaires often manages distribution. complain that such fees unfairly saddle them First governments can manage retail with investment risks that should remain distribution themselves, and make regular with governments, because the original payments (sometimes referred to as loans were too expensive, the investments "unitary" payments) to private partners for were not cost-effective or efficiently the provision of bulk services provided via managed, and the assets created by the greenfield infrastructure projects. investment often do not generate the Performance penalties or bonuses are easier expected revenues needed to pay off the to manage with this form of remuneration. loans. In some cases, governments are This is of course the IPP model used waiving such fees in favor of a payment extensively in the electricity sector. Its use made by the concessionaire at the end of the in the water sector may also be on the contract for the amortized value of assets increase, with greenfield water and sewage transferred to private operation by treatment plants selling to single governments, and used to help make profits government customers, who back their for the concessionaire. The conceded assets commitments via "take or pay" account mentioned above in the case of the arrangements (Marin and Izaguirre 2006). Kenya-Uganda rail concession also serves this function on that project. Payments by · Government pays for distribution, the concessionaire at the end of that contract managed by private partner. are expected to be more than offset by Second, private partners can manage retail government payments for the unamortized distribution, and even billing and collection, value of assets provided by the but are remunerated via government concessionaire. payments based on (1) availability of services irrespective of usage (i.e., · More reasonable approaches to profit government takes on all demand risk), or (2) sharing. usage, as in the case of "shadow tolls," But fees may also be intended simply to based on the volume and composition of allow governments to share in the overall traffic, but not paid by motorists (here, financial benefits of concessions. In this demand risk is shared by public and private case, fees might be better linked to partners). profitability than just revenues, which frequently is the case and often puts · Private partner manages distribution, significant pressure on project cash flows. paid for by end users, but government Governments may also consider setting fees mitigates demand risk. after a careful assessment of their other Third, private partners can also receive sources of financial benefits from remuneration principally via user fees, concessions, such as taxes, which often backed by a government commitment to generate more revenue for governments over "top up" with subventions when demand the lifetime of a concession than do fees falls below preagreed levels. If demand (Pozzo di Borgo et al. 2006). exceeds estimates, private partners may be expected to share revenue surpluses with Government Assumption (or Sharing) of government. See Vassallo (2006) for a Demand Risk detailed discussion of various mechanisms Governments can do several things to reduce the used by governments to mitigate traffic risk demand risk faced by private partners handling in highway concessions. retail distribution of infrastructure services. Two issues arise when governments consider assuming or sharing of demand risk. First, such risk sharing can weaken private partner 30 incentives to maintain customer satisfaction with implementation of concessions and other PPP services provided. Alternate means of projects that present significant contingent incentivizing quality and reliability of service liabilities for government owners. The sort of are usually incorporated into these agreements. low-cost project preparation that was Second, although greenfield projects involving characteristic of brownfield concessions in the bulk supply and guaranteed government off-take early 1990s is largely a thing of the past, at least agreements are effective ways of generating in most countries with PPP experience over the private sector investment, the benefits of the decade. Many governments now require the use investment can be wasted if government service of technical feasibility studies, PPP option provision is compromised by weak institutions, assessments, and rigorous cost-benefit analysis policies, and procedures. One of the original before government agencies may proceed with reasons for using the brownfield concession PPPs such as concessions. In addition, the kinds mechanism was precisely the need to strengthen of considerations mentioned above regarding management of operations as well as increase government assumption of risks associated with investment. Because of this, in some countries investment and service delivery can raise even the use of greenfield projects is matched with higher the stakes for government involvement in other arrangements to improve service delivery. PPP arrangements, particularly if government must supply financing for implementation. New Government Ownership and Control of kinds of private partners and new ways of Project Preparation making money from concessions (discussed With end of the 1990s, many governments in below), also require much more government developing countries have followed the lead of attention to contract compliance monitoring, countries such as the United Kingdom, accounting, and reporting.18 For a review of Australia, South Africa, etc., in establishing how PPP units incorporate these features, see specialized units for developing, supervising the PPIAF (2007). development of, or monitoring the 18A relatively well-established state of the art in government assessment and management of these kinds of issues is now widely available from PPP units in countries such as Ireland, Scotland, Australia, Canada, and South Africa. Much of this guidance material is available from Internet Web sites. South Africa's PPP guidelines, one of the most extensive and sophisticated approaches to government regulation of PPP project preparation in the developing world, can be downloaded from http://www.finance.gov.za/organisation/ppp/PPP%20Manu al/default.htm. 31 9 PRIVATE SECTOR ACTIONS TO STRENGTHEN PROFITS AND CASH FLOWS Maximizing Remuneration from Available As governments act to make risk-sharing Sources arrangements on brownfield concessions more New kinds of concessions, some involving attractive to potential private sponsors and limited involvement by private partners in operators, private players are also changing their distribution (described above) highlight the approaches to these projects to strengthen their possibility that sponsors and operators will more cash flows and remuneration. carefully explore all available sources of remuneration associated with these projects. More Due Diligence, Less Opportunistic Bidding · Technical assistance fees. First, of course, fewer of the large international Technical assistance, a common feature of service providers are bidding frequently on brownfield concession projects, is used to brownfield concessions. The intentional help transfers skills between expatriate and underbidding of projects with the expectation of local staff of the concession company. Such renegotiation later no longer appears to be a fees are usually treated as costs, changed widely used technique for dealing with under separate management or technical inherently risky projects. Potential private assistance contracts. As such, they reduce partners are now more frequently willing to the profits of the concession company, but propose "negative" concessions when they bid have the effect of transferring cash to the (involving fees or other support paid to the concession shareholders, and as a result may concessionaire by governments, rather than the boost a project's effective overall other way around as with traditional brownfield profitability. In the railway sector, such concessions). In some recent cases, even the use costs normally account for less than 1 of lower-risk lease or management contracts, percent of the concession company's combined with donor funding for capital turnover, and decline quickly over time. But investment, have not evoked positive responses in some cases technical assistance costs have from potential private partners as they once been significantly higher. In the Abidjan- might have.19 Ouagadougou rail concession in West Africa (Sitarail), these costs have accounted for 5 percent of turnover (Pozzo di Borgo et al. 2006). · Transfer pricing. Similarly, when concessionaires build new repairs and maintenance, they often use 19In late 2004, the Zambian government issued a related businesses to undertake the work or prospectus for a PPP lease arrangement for the operation of source the equipment, rather than use public Lusaka's water distribution company. The lease was to be competition to find the lowest cost provider. paralleled by a donor-funded investment program, which The resulting transfer prices represent would also pay for operating costs in the early years of the income for the larger group of companies of contract (Government of the Republic of Zambia 2004). which the concessionaire may be a part. But the project attracted little interest from private sponsors Like technical assistance fees, such transfers and operators and was later abandoned in favor of a are treated as costs and therefore reduce the performance contract between the utility and the concessionaire's income, even though their government. real impact is more like that of a shareholder 32 dividend.20 In fact, Sirtaine et al. (2004) particularly active in South Asia and Sub- concluded that management fees and Saharan Africa, and historically have been investment markups may be essential in most prominent in transport sectors. The most cases for concession to become researchers speculate that as concession-type profitable in fact, if not in a strict accounting projects become smaller, these kinds of sense. private partners are able to satisfy project financing needs by taking advantage of the Transfer pricing and technical assistance fees deepening capital markets in developing have long been a concern of government owners countries. Some local and regional of concession projects. Sirtaine et al. suggest operators have built on their experiences as that audits focusing on transfer pricing issues minority partners (often as construction have been standard features of more developed contractors) with developed country PPP markets such as Latin America. But Pozzo investors in PPP projects. Finally, as local di Borgo et al. imply that they are used much firms, many of these operators and less frequently in regions like Africa where investments may be better positioned than monitoring efforts are weak and the kinds of firms from developed countries to deal with information that concessionaires must routinely political economy issues affecting PPPs in report often are not specified in contracts. As developing countries. But Ettinger et al. sponsors and operators continue to struggle to also caution that all of the implications of make traditional brownfield concession the involvement by these kinds of players profitable, these kinds of audits are likely to are not yet clear. In particular, consumers in become more routine requirements of these countries are often concerned that concession contracts. local firms may be more likely to be involved in corrupt activities because of Concessionaire Self-Selection: Better their close connections to government. Matching of Needs, Skills, and Opportunities Another way to maximize remuneration for risks · Companies seeking control over as well as risk mitigation is for would-be distribution chains. concessionaires to aim for the most lucrative Other new kinds of operators and investors matchups between their own skills, experience, are likely classifiable as subsets of the and situation and the opportunities for making general category of local and regional money on different projects. Private firms that companies described above. One type can find these matchups are winning contract identified by Pozzo di Borgo et al. (2006) is awards and weathering stresses associated with a company or group of companies trying to cash flow and profitability stresses. capture a dominant position in vertical distribution chains of productive and · Local and regional sponsors and transport sectors. In other words, they want operators. to control a significant part of the Ettinger et al. (2005) document the distribution chain to capture overall benefits emergence of local and regional investors that would outweigh poor rates of return and operators in all forms of PPP projects. associated with operation of any specific But their strongest involvement has been in activities in the chain. Tariff discounts or concessions, where they have accounted for other benefits for affiliated companies 54 percent of concession-related investment operating elsewhere in the chain might be in 1998­2004. These kinds of players are one advantage of this approach. So a rail concession might be part of a bigger 20The Sitarail case demonstrates this impact on overall distribution chain that would include profitability: categorizing the technical assistance as a cost agricultural production, freight forwarding, results in average annual return on equity at just over 9 seaport operations, shipping lines, etc. Low percent. Categorizing it as profit increases return on equity to 31 percent (Pozzo di Borgo et al. 2006). rates of return for the concession would be tolerable as long as cash from other 33 operations was available to forestall cash rather than through cash flows. In many flow crises. Obviously, concerns about this cases these probably are local companies role would include potential undue market that began their PPP careers as construction and pricing power in the transport logistics subcontractors to international firms on chain, and profit transfers from one larger brownfield concessions or greenfield subsidiary to another. projects. As operators, these firms may be willing to rely heavily for their remuneration · Construction management companies. on consulting fees charged to concession Another new kind of concessionaire, again SPVs and transfer pricing for equipment and an important subset of new local and investments, particularly in the form of regional operators, is characterized by a investment markups, as described in the company or group of companies with a section above. This interest of these kinds focus on a single sector or subsector, such as of concessionaires in managing large rail or ports, but whose interest lies in the projects raises questions about what keeps financial benefits associated with managing the size of these investment projects in large investment programs--especially check. those financed by governments or donors 34 10 SIGNS OF RECOVERY IN 2006? In 2006, infrastructure investment associated any previous year). Not surprisingly, most of with brownfield concessions demonstrated its the brownfield concession investment in 2006 sharpest increase since the 1990s, more than was in the transport sector, principally roads and doubling the level of 2005. The 2006 figure airports. Airports are well known as strong brings total investment to a level first achieved commercial opportunities for private only in 1993, but may nevertheless prefigure a concessionaires, with the number of projects recovery of sorts for the mechanism. This is limited in the past largely by governments' because the kinds of brownfield concessions unwillingness to turn over facilities to private concluded in 2006 reflect the changing nature of partners, which are both strategic and capable of investor risk-reward requirements outlined generating at least some revenues for above. Figure 13 illustrates this by comparing government owners. Globalization and the sectoral distribution of concession-related heightened security concerns have underscored investment in 1997 with that of investment in the need for more investment in and professional 2006. The sharp increase in private-sector risk management of such facilities, even in aversion is strongly apparent in 2006. developing countries. Very little investment is registered in water and Toll roads, even those generating revenues via sanitation and virtually none in energy (where user-pay arrangements, are also widely private operators and sponsors continued to recognized as having long-term commercial focus on less risky greenfield IPPs backed by value in many developing countries, particularly off-take agreements). No concession-related as toll road concessions have been strengthened investment in telecommunications is registered by government sharing of financing and demand here because private companies were optimizing risks (Fitch Ratings 2007). Cost coverage returns and minimizing uncertainty by focusing mechanisms such as availability payments, on merchant projects (more of this kind of capital grants, and shadow tolls have been cited investment was committed to in 2006 than in as principle factors in the growing private sector involvement in toll roads in the developing world (Queiroz and Izaguirre 2008). 35 Figure 13: Sectoral Distribution of Investment--Brownfield Concessions: 1997 and 2006 (US$ millions) Subsector Segments 1997 % of total 2006 % of total Electricity Distribution 3,048 22 Distribution, generation, transmission 268 0 Generation 1,678 22.2% 0 0.2% Natural Gas Transmission 600 0 Distribution 0 2.7% 0 0.0% Airports Runway & terminal 242 4,394 Terminal 8 1.1% 0 41.7% Railroads Fixed assets & freight 188 0 Fixed assets, freight, passenger 79 400 Freight 1,100 6.1% 0 3.8% Roads Bridge 81 0 Bridge & highway 49 1,558 Highway 6,487 29.4% 2,364 37.2% Seaports Channel dredging & terminal 214 70 Terminal 721 4.1% 706 7.4% Water, sewerage Potable water treatment plant 206 7 Water utility with sewerage 7,045 998 Water utility without sewerage 529 19 Sewerage collection & treatment 0 34.5% 11 9.8% Totals 22,542 100% 10,548 100% Source: World Bank and PPIAF, PPI Project Database. 36 11 CONCLUSIONS In many respects, the sharp decline in the global The brownfield concession is not an inherently PPP market, registered by the PPIAF-World flawed mechanism--the track record of this Bank PPI Project Database beginning in 1997, contract form in developed countries is reflected problems with a single PPP reasonably successful. The mechanism, mechanism, the brownfield concession contract. however, is more fragile than many market If the data on brownfield concessions is removed participants assumed as the PPP boom began in from the database, PPP-related investments in the developing world during the 1990s. The government infrastructure facilities demonstrates projects must be carefully prepared, and a host none of the "crash" characteristics commonly of reliable risk management mechanisms need to associated with PPI over the last decade. In be in place for the arrangement to be sustainable. nominal terms, investment figures for 2006 were Many of these conditions for success have 67 percent higher than for 1997. In the early proven to be difficult to achieve in developing 1990s, the development community hoped that countries. Brownfield concessions will continue the brownfield concession would become the to be used in the relatively narrow set of signature contract of the PPI movement, circumstances where it now appears that these embodying virtually all of the most beneficial contracts can be sustainable in the developing qualities associated with public-private world. A decade after the onset of the Asian partnerships for infrastructure. But the Asian Crisis, the brownfield concession may seem like Crisis triggered a precipitous decline in the use an endangered species of PPP, but the needs that of brownfield concessions, which to date have drove the initial widespread use of this recovered only to levels first achieved in the mechanism--those intractable and highly visible early 1990s. problems with existing infrastructure service systems--still exist. Governments and their No authoritative studies have been done private sector partners are responding with new showing how the mechanism was promoted in contracting arrangements that add concession- the developing world in the early 1990s, but type features to other traditional PPP options anecdotal accounts suggest that it may have been such as management contracts, leases, and even oversold to (and by) governments who saw it as partial privatizations. To a large extent, the a win-win solution to some of the most challenge in assessing the ongoing track record intractable and highly visible infrastructure of brownfield concessions will be to ensure that problems they faced. But those same monitoring and data collection categories are characteristics that made the mechanism almost flexible enough to capture these evolving too good to be true also made the resulting contract forms. An important objective of this project arrangements highly vulnerable to cash kind of monitoring work will be to learn lessons flow stresses and profitability problems. about the liabilities and risks associated with these hybrid arrangements in time to pass useful guidance on to practitioners. 37 REFERENCES Aguas Argentinas S.A. Concession Contract. 1993. Concession Agreement entered into between the Argentine State and Aguas Argentinas S.A., April 28, 1993. Alcazar, Lorena, Manuel A. Abdala, and Mary M. Shirley. 2000. "The Buenos Aires Water Concession,=." Policy Research Working Paper No. 2311, The World Bank, Washington, DC. Baietti, Aldo, William Kingdom, and Meike van Ginneken. 2006. "Characteristics of Well-Performing Public Water Utilities." Water Supply & Sanitation Working Notes, No. 9, The World Bank, Washington, DC. Bourguignon, Francois. 2006. "Rethinking Infrastructure for Development," Closing Remarks, Annual Bank Conference on Development Economics, Tokyo, Japan, May 30. De Fraja, Gianni, and Clive Stones. 2004. "Risk and Capital Structure in the Regulated Firm." Journal of Regulatory Economics Vol. 26, No. 1 (July): 69­84. De Lemos, Teresa, Martin Betts, David Eaton, and Luis Tadeu De Almeida. 2000. "From Concessions to Project Finance and the Private Finance Initiative." The Journal of Project Finance Vol. 6, No. 3 (fall): 19­37. Dumol, Mark. 2000. The Manila Water Concession: A Key Government Official's Diary of the World's Largest Water Privatization. Directions in Development, World Bank, Washington, DC. Eberhard, Anton. 2007. "Matching Regulatory Design to Country Circumstances." Gridlines Series, no. 23, May, PPIAF, Washington, DC. Ehrhardt, David, and Timothy Irwin. 2004. "Avoiding Customer and Taxpayer Bailouts in Private Infrastructure Projects: Policy Toward Leverage, Risk Allocation, and Bankruptcy." Policy Research Working Paper 3274, The World Bank, Washington, DC. Estache, Antonio, J. Luis Guasch, and Lourdes Trujillo. 2003. "Price Caps, Efficiency Payoffs, and Infrastructure Contract Renegotiations in Latin America." Policy Research Working Paper No. 3129, The World Bank, Washington, DC. Estache, Antonio, and Eugene Kouassi. 2002. "Sector Organization, Governance, and the Inefficiency of African Water Utilities." Policy Research Working Paper No. 2890, The World Bank, Washington, DC. Estache, Antonio, and John Strong. 2000. "The Rise, the Fall, and ... the Emerging Recovery of Project Finance in Transport." Policy Research Working Paper No. 2385, The World Bank, Washington, DC. Ettinger, Stephen, Michael Schur, Stephan von Klaudy, Georgina Dellacha, and Shelly Hahn. 2005. "Developing Country Investors and Operators in Infrastructure." Trends and Policy Options Series No. 3, PPIAF, Washington, DC. 38 Fitch Ratings. 2007. "Global Toll Road Rating Guidelines." Global Infrastructure and Project Finance, Criteria Report, March 6. Available at www.fitchratings.com. Flyvbjerg, Bent. 2005. "Policy and Planning for Large Infrastructure Projects: Problems, Causes, Cures," Policy Research Working Paper 3781, The World Bank, Washington, DC. Forsgren, Kurt, Paul Calder, Robert Bain, Ian Greer. 2006. "Global Infrastructure Assets and Highly Leveraged Concessions Raise New Rating Considerations." Standard & Poor's Rating Direct, Sept. 14, Standard & Poor's, New York. Foster, Vivien, and Tito Yepes. 2006. "Is Cost Recovery a Feasible Objective for Water and Electricity? The Latin America Experience." Policy Research Working Paper 3943, The World Bank, Washington, DC. Ghana Management Contract. 2005. Management Contract for Ghana Urban Water between Ghana Water Company Ltd. and Vitens Rand Water Services BV and Aqua Vitra Ltd., October. Government of the Republic of Zambia. 2004. "PSP in Lusaka Water and Sewerage Company: Prospectus." Prepared by C. Ricketson Consulting Ltd., September. Guasch, J. Luis. 2004. "Granting and Renegotiating Infrastructure Concessions: Doing It Right." WBI Development Studies, The World Bank, Washington, DC. Guislain, Pierre, and Michel Kerf. 1995. "Concessions--The Way to Privatize Infrastructure Monopolies." Note No. 59, Public Policy for the Private Sector, The World Bank, Washington, DC. Israel, Arturo. 1992. "Issues for Infrastructure Management in the 1990s." World Bank Discussion Paper 171, The World Bank, Washington, DC. Kerf, Michel, R. David Gray, Timothy Irwin, Celine Levesque, and Robert R. Taylor. 1998. "Concessions for Infrastructure--A Guide to Their Design and Award." Technical Paper No. 399, World Bank, Washington, DC. Klein, Michael. 1998. "Rebidding for Concessions." Public Policy for the Private Sector, Note no. 161, November, World Bank, Washington, DC. Kotze, Roelf, Andrew Ferguson, and James Leigland. 2000. "Government Facilitation of Public-Private Infrastructure Projects: Lessons from South Africa." Journal of Project Finance Vol. 6, No. 1 (spring): 61­70. The Lawyers' Environmental Action Team, the Legal and Human Rights Centre, the Tanzania Gender Networking Programme, the Center for International Environmental Law, and the International Institute for Sustainable Development. 2007. Amicus Curiae Submission in Case No. ARB/05/22 before the International Centre for Settlement of Investment Disputes between Biwater Gauff Ltd. and United Republic of Tanzania." March 26. Leigland, James. 2000. "Concession or Sale: PPPs to Solve South Africa's Airport Problems." The PFI Report, November: 13-15. Marin, Philippe, and Ada Karina Izaguirre. 2006. "Private Participation in Water: Toward a New Generation of Projects?" Gridlines series, no. 14, PPIAF, Washington, DC. 39 Marshella, Tom. 2001. "Debt Financing of Projects in Emerging Economies: Lessons from Asia." The Journal of Project Finance Vol. 7, No. 2 (summer): 30­34. Martin, Lloyd. 2005. "Public-Private Partnerships in the Water Sector of the EECCA Region: Update and Recommendations for the Improvement of Service Delivery." Paper presented at the Conference of EECCA Ministers of Economy/Finance, Environment and Water, OECD Task Force for the Implementation of the Environmental Action Programme for Central and Eastern Europe, Caucasus and Central Asia Yerevan, Armenia, November 17­18. Matsukawa, Tomoko, and Odo Habeck. 2007. Review of Risk Mitigation Instruments for Infrastructure. Trends and Policy Options Series No. 4, PPIAF, Washington, DC. Nelspruit Concession Contract. 1999. Concession Agreement entered into between Nelspruit Transitional Local Council and the Greater Nelspruit Utility Company Ltd., April 21, 1999. OECD. 2007. "Concessions." Directorate for Financial and Enterprise Affairs, Competition Commission, Global Forum on Competition, April 20. Pomerleano, Michael. 1998. "Corporate Finance Lessons from the East Asian Crisis." Public Policy for the Private Sector, Note No. 155, October, The World Bank, Washington, DC. Pozzo di Borgo, Pierre, Alain Labeau, Raphael Eskinazi, Julien Dehornoy, Alan Parte, Marouane Ameziane. 2006. "Sub-Saharan Africa: Review of Selected Railway Concessions." Africa Transport Sector, Report No. 36491, The World Bank, Washington, DC. PPIAF. 2007. "Public-Private Partnership Units: Lessons for Their Design and Use in Infrastructure." World Bank and PPIAF, Washington, DC Queiroz, Cesar, and Ada Karina Izaguirre. 2008. "Worldwide Trends in Private Participation in the Road Sector." Gridlines series, PPIAF, Washington, DC. Sirtaine, Sophie, Maria Elena Pinglo, J. Luis Guasch, and Vivien Foster. 2004. "How Profitable Are Infrastructure Concessions in Latin America? Empirical Evidence and Regulatory Implications." PPIAF Trends and Policy Options, The World Acronyms and abbreviations Bank, Washington, DC. Vassallo, Jose M. 2006. "Traffic Risk Mitigation in Highway Concession Projects: The Experience of Chile." Journal of Transport Economics and Policy Vol. 40, Part 3, September: 359-381. Wachs, Martin. 1990. "Ethics and Advocacy in Forecasting for Public Policy." Business and Professional Ethics Journal Vol. 9, Nos. 1 & 2: 141­157. Williamson, Mark, and Tony Moore. 2006. "International Power: Capital Structure and Project Finance,." Presentation at International Power Investor Conference, London, September 19. 40 PPIAF Program Management Unit c/o The World Bank 1818 H Street, NW Washington, DC 20433 USA Telephone: 202.458.5588 Fax: 202.522.7466 Email: ppiaf@ppiaf.org World Wide Web: http://www.ppiaf.org