Water Papers Water Papers February 2012 69684 CAPITAL SUBSIDIES IMPLICIT IN CONCESSIONAL FINANCE How to Make Them More Transparent and Better Targeted William Kingdom, Axel Baeumler and Alfonso Guzman Water Papers are published by the Water Unit, Transport, Water and ICT Department, Sustainable Development Vice Presidency. Water Papers are available on-line at www.worldbank.org/water. Comments should be e-mailed to the authors. Approving Manager Vijay Jagannathan, Sector Manager, EASIN Contact Information This paper is available online at http://www.worldbank.org/water. The author may also be contacted through the Water Help Desk at whelpdesk@worldbank.org. Disclaimer – World Bank This volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The find- ings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encour- ages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone 978-750-8400; fax 978-750-4470; http://www.copy-right. com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax 202-522-2422; e-mail: pubrights@worldbank.org Disclaimer – AusAID The views expressed in this publication are those of the authors and not necessarily those of the Australian Agency for Inter- national Development (AusAID). CONTENTS 1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 Concessional Finance and Implicit Captial Subsidies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1 Quantifying Implicit Subsidies in the Water Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.2 Opportunity Cost of Implicit Subsidies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.3 Implicit Subsidies are Widespread . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3 How to Make Better Use of Implicit Capital Subsidies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 4 Operationalizing Improved Capital Subsidy Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 4.1 How to Make an Implicit Subsidy Explicit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 4.2 Cash Flow Impact of Converting to a Loan on Commercial Terms and a Grant. . . . . . . . . . . . . . . . . . . 12 4.3 Define Who Administers the Subsidy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 4.4 Determining the Subsidy Amount—the Benchmark Loan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 4.5 Evolution of Subsidy Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 5 The Way Ahead. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Selected Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Tables Table 1: Loan Terms for a Representative Water Utility in Vietnam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Table 2: Implicit Subsidies in Concessional Loans to the Water Sector in Vietnam . . . . . . . . . . . . . . . . . . . . . 6 Table 3: Examples of Concessional Finance in the Water Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Table 4: Comparison of Options for Making Capital Subsidies Explicit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Figures Figure 1: Improved Capital Subsidy Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Boxes Box 1: Sensitivity of Implicit Subsidy Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Box 2: Definition of Transparent and Well-targeted Subsidies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Box 3: Impact on Cash Flow of Concessional Loan versus Separate Commercial Loan . . . . . . . . . . . . . . 12 and Capital Subsidy Box 4: Subsidy Design Should Evolve from Water Supply to Wastewater . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 iii Acknowledgements This report was prepared by a team comprising Bill Kingdom, Lead Water and Sanitation Specialist SASDU, Axel Baeumler, Se- nior Infrastructure Economist, MNSSD, and Alfonso Guzman, Managing Director, Castalia Strategic Advisors. It draws on sector work undertaken in Vietnam to inform the design of a possible urban water supply and sanitation project. The original sector work, and the subsequent activities leading to the preparation of this report, were supported by the Water Partnership Program (WPP) financed by the Government of the Netherlands and by the Government of Australia through an AusAid Trust Fund. The sector work required extensive field studies and consultations in Vietnam. The team would like to thank the Ministry of Finance, the Ministry of Construction and the Ministry of Planning and Investment of the Government of Vietnam for their continuous support over the period of the original studies. In addition the team worked with representative Banks to seek their input on the models being developed, and in particular the Bank for Investment and Development of Vietnam (BIDV) and the Vietnam Development Bank (VDB). To ground the original studies in reality the team worked closely with the Binh Duong Water Supply Company. The report benefited greatly from peer reviews by Jonathan Halpern, Advisor MIGOP, Caroline van den Berg, Lead Water and Sanitation Specialist MNSWA, Sudipto Sarkar, Lead Water and Sanitation Specialist, EASIN. Guidance on the preparation of the final document was provided by Vijay Jagannathan, Sector Manager EASIN, Aldo Baietti, Lead Infrastructure Specialist, EASIN, and David Ehrhardt, Chief Executive Castalia Strategic Advisors. The team also benefited from the support of Quang Vinh Nguyen, Water and Sanitation Specialist, EASVS. Abstract Multilateral Development Banks (MDBs) finance a significant objective. Specifically, the paper (i) considers the extent to share of developing country water sector investments. Much which implicit subsidies exist in MDB lending for the wa- of this financing is concessional and often on-lent by nation- ter sector; (ii) identifies a possible approach to quantify the al governments, at similar concessional terms, to water utili- amount of subsidies involved; (iii) outlines an emerging ties. This concessionality carries an implicit subsidy, i.e., the framework to make subsidies more explicit as a basis for im- difference between MDB financing terms and commercial fi- proved targeting; and (iv) discusses operational implications. nancing terms priced more in line with the underlying credit By investigating these issues, the paper intends to be a first risks. As such concessional financing is most often used for step for governments and donors to evaluate how best to capital investment projects, the implicit subsidy can be con- use the implicit capital subsidies provided by concessional sidered a capital subsidy. financing in the water sector. The paper also suggests areas of future research. This working paper asks whether there is an opportunity to increase the value of concessional financing for water sec- While the focus of the paper is on the water sector, the issues tor investments by making implicit capital subsidies more and approaches discussed in the paper are generally appli- explicit and targeting them to a clearly defined public policy cable to concessional lending for infrastructure projects. v 1. Introduction M ultilateral Development Banks (MDBs) finance water the value of the subsidy, the appropriate subsidy level, or its and wastewater infrastructure in developing coun- targeting. This contrasts with the extensive literature avail- tries to a significant extent with concessional loans. able on revenue subsidies, which considers the targeting of For example, between 2000 and 2010 the World Bank pro- subsidies for customer user fees through tariff designs. vided approximately US$ 9.4 billion in loans to borrowers for water and sanitation through its highly concessional Inter- In reality, these two types of subsidies are complementary national Development Association (IDA) arm.1 Even in some and in the longer term will need to be considered as a pack- developed countries, concessional financing still assists utili- age. Capital subsidies should be designed to meet broad na- ties in meeting specific government policies and objectives, tional policy goals such as access and service quality, within for example, in the construction of wastewater facilities. In constraints of available public funds and general affordabil- addition, many developed country service providers have ity. Revenue subsidies are specifically intended to ensure ser- benefited from concessional finance in the past to allow vice affordability at the household level within any particular them to build up their comprehensive infrastructure stock. service provider. The different objectives of the two types of subsidies are also reflected in the sources of the subsidies. Thus, concessional financing plays an essential role in the ex- Central or state governments, or international donors, tend pansion and improvement of water and sanitation services. to provide capital subsidies, while revenue subsidies tend However, sources of concessional finance are finite and gov- to come from other customers (e.g. cross subsidy between ernments and donors therefore need to determine how to different categories of users), or from local governments use these limited funds for maximum benefit. Unfortunately, (e.g. direct payments towards the water bills of low-income such a rigorous determination is rare. Governments typi- households). cally on-lend concessional funds to water utilities, on simi- larly concessional terms, without overtly assessing either the This working paper on capital subsidies also provides a time- value of the subsidy or its targeting. One reason is that such ly contribution to an evolving debate within the World Bank. concessional on-lending “hides� the value of the subsidy that A report recently completed by the Independent Evaluation is implicit in the lending terms passed on to the service pro- Group (IEG) on “Water and Development� (2010) highlight- vider. If that subsidy were made explicit (i.e. transparent), it ed that the Bank appeared to have moved away from “full could trigger a debate amongst policy makers on whether cost recovery� but did not identify how any deficit between such a subsidy was appropriate both in terms of its amount “full cost recovery� and “actual cost recovery� was being fi- and its targeting. Concessional financing is a capital subsidy made available 1 Data source: World Bank Data. This paper draws primarily upon to service providers for investments to expand or improve examples from IDA lending, but the methodologies and conceptual their assets. A capital subsidy reduces the final cost of ser- frameworks presented could also apply to the International Bank of Reconstruction and Development (IBRD) lending arm. IBRD typically vice for customers by lowering the debt service component provides loans to middle-income countries at near-market terms with of a cost-recovering tariff. There are, however, only a limited longer repayment periods. However, the amount of subsidy would be number of studies on capital subsidies—whether assessing lower. 1 nanced. There are, of course, many definitional terms to be making explicit the actual value of implicit subsidies that considered but if user fees should at least recover operation are typically provided to borrowers through concessional and maintenance (O&M) costs and debt service, then the on-lending terms. Section 2 of the paper defines implicit differences highlighted by IEG may derive partly from the subsidies, shows how the amount of implicit subsidy can concessionality of the loan made to the service provider— be quantified, and considers the opportunity cost of a lack thereby making the capital subsidy part of the balancing of transparency and poor targeting. Section 3 presents an factor.2 This, coupled with the fact that concessional loans emerging framework for better capital subsidy design and will continue to play an important role in financing the the targeting of such subsidies while section 4 highlights water sector in developing countries, means that making operational implications of the framework. Section 5 re- subsidies transparent and improving their targeting could flects on the results of the study and suggests possible fu- help achieve the best value from this public money. Thus, a ture work to enhance further the sector’s understanding of better understanding of the issue may improve the sector capital subsidies.3 overall and lead to a more informed policy within the World Bank as well as with its client countries. The authors acknowledge that significant work remains to 2 Alongside sub-optimal operating expenditures (typically resulting understand capital subsidies fully —whether their amount or in deferred maintenance) and periodic cash transfers from government their targeting. However, this topic needs to be approached to support financially stressed service providers. more methodically and this paper represents a first step in 3 This paper is based on policy dialogue and project experience that direction. gained by the World Bank in South East Asia during 2008 and 2009. Since then, governments have further refined some of the outlined approaches evaluated in this study. However, despite the subsequent This paper does not discuss capital subsidies per se, given changes, these approaches remain relevant as valuable examples that their pervasiveness in the sector, but instead focuses on clearly illustrate the concepts presented in the paper. 2 | Capital Subsidies Implicit in Concessional Finance: How to Make Them More Transparent and Better Targeted 2. Concessional Finance and Implicit Capital Subsidies C oncessional loans are, by definition, loans below mar- sidy implicit in concessional lending—commonly referred to ket terms and which therefore contain implicit subsi- as the level of concessionality—for a given dollar value of a dies. This section describes how the implicit subsidies loan can be calculated as the difference between the pres- in concessional lending can be quantified and highlights ent value of the concessional and benchmark loan principal the opportunity cost associated with those implicit subsi- and interest payments.4 dies. Opportunity costs arise from a lack of transparency on the amount of subsidy provided which can limit a more in- The amount of implicit subsidy can be illustrated by an ex- formed debate about their optimal use and targeting. ample from Vietnam based on the model that was in use there until 2010. In Vietnam, where approximately 85 percent of investment in the water sector was financed with conces- 2.1 Quantifying Implicit Subsidies sional Overseas Development Assistance (ODA) loans, the in the Water Sector government set on-lending terms based on the size of the Concessional loans from MDBs are often on-lent by finance city benefiting from the loan. Cities were grouped in five cat- ministries to water utilities on terms that are more generous egories according to population size. Category One was the than would be available to those utilities in the commercial largest and included cities with a population of more than market. It is possible to calculate the subsidy implicit in gov- 500,000 while Category Five, the smallest, included cities ernments’ concessional lending by comparing the conces- with a population less than 50,000. A typical loan to benefit a sional on-lending terms to the commercial terms. A simple Category Five city would have a maturity of 20 years, a grace formula can be used to quantify the implicit subsidy. period of eight years and an interest rate of four percent. A loan that would have benefited a Category One to Category The first step is to identify a benchmark that represents the Four city would have a maturity of 20 years, a grace period of terms on which the utility could borrow commercially. After three years and an interest rate of 6.7 percent. selecting a benchmark loan, the present value of the sub- At the same time, some water utilities were also able to bor- row from commercial banks, on commercial terms. One utility in Vietnam,5 in addition to mobilizing ODA funds on- lent by the Ministry of Finance, borrowed from several com- Present Value (PV) subsidy = PV of concessional loan mercial sources, including the equivalent of US$ 520,000 principal and interest payments – PV of bench- from the Bank for Investment and Development of Vietnam mark loan principal and interest payments Where: 4 This calculation follows the principles used to calculate the level of PV indicates a present value calculation of the debt concessionality of IDA credits at the country level. service payments on the loan, discounted at the 5 This example refers to the Binh Duong Water Supply Company, lo- interest rate of the benchmark loan. cated outside Ho Chi Min City, which was analyzed in the study: De- veloping a New Financing Framework for the Urban Water Sector in Vietnam: Final Report to World Bank�. Castalia (December 2008). | 3 Table 1: Loan Terms for a Representative Water Utility in Vietnam Source Interest Rate (percent) Maturity Grace Period BIDV (Commercial) 10.2% 5 years 0 years Ministry of Finance 4.0% 20 years 8 years Source: WSC accounts. As reported in “Developing a New Financing Framework for the Urban Water Sector in Vietnam: Final Report to World Bank�. Castalia (December 2008). (BIDV). Table 1 compares the lending terms from BIDV with rate. The concessional loan with a five-year maturity and the concessional ODA on-lending terms offered by the Min- 13 percent interest rate would result in an implicit subsidy istry of Finance. of 75.6 percent as opposed to 66.3 percent—a significant difference. This comparison highlights why it is important Clearly, the Ministry of Finance terms are more generous to choose an appropriate benchmark loan. In the water than the commercial terms, as itis willing to bear more risk, sector, finding the most suitable benchmark loan is likely for lower return, than a commercial bank. By agreeing to lend to remain problematic, as there are few private sector and bear risk without a compensating return, the Ministry commercial lenders. Nevertheless, despite these chal- of Finance is providing an implicit subsidy to the water util- lenges, subsidies implicit in concessional finance can still ity—and all other utilities to which it lends at concessional be made explicit. terms. Applying the above equation—with the subsidy ex- pressed in percentage terms—shows that the concessional Implicit subsidy calculations are not only sensitive to the loan terms from the Ministry of Finance to the water utility adopted benchmark loan but also to the interest rate and imply a subsidy equivalent to 66.3 percent of the loan’s total grace period of the concessional loan. Box 1 provides fur- value. Only a few utilities, however, are able to borrow from ther discussion of the sensitivity of the implicit subsidy cal- commercial banks, or if so, only for a fraction of their required culations to the assumptions used in calculating the subsidy investment needs. Nevertheless, if utilities are able to bor- component. row money from commercial banks as well as concessional financing sources, this approach can quantify the subsidy element implicit in ODA loans. 2.2 Opportunity Cost of Implicit Subsidies Selecting a benchmark loan Implicit subsidies are, by definition, non-transparent and this However, it can be difficult to select an appropriate risk- can make them difficult to target. If governments direct sub- equivalent benchmark loan, especially in the water sector sidies in a way that cannot be easily analyzed or debated, the where very little “true� private sector commercial lending subsidies are likely to be less successful in serving the de- reflects full-risk premiums. Also, in the example of the Viet- sired public policy objectives. There is an opportunity cost namese water utility, it is likely that the lending terms from if the subsidy element in concessional lending is not effec- BIDV are a lower bound estimate of true private sector com- tively targeted—the additional benefit that could have been mercial financing as BIDV is a state owned commercial bank achieved had the subsidy been explicit and well-targeted. which may under-price some of the underlying credit risks.6 Concessional loans can both target and support public pol- The choice of benchmark loan is important, because it has icy objectives. For example, a concessional loan dedicated a significant effect on the estimate of the implicit subsidy. to small towns or to increasing access in under-served areas For example, if the Vietnamese water utility took another supports a clearly defined objective. However, because the commercial loan with a five-year maturity, no grace peri- od, and an interest rate of 13 percent, the benchmark from 6 An attempt was made to use the World Bank and IFC’s Municipal this loan would show a much higher level of implicit sub- Fund pricing as benchmarks for private sector commercial lending in sidy when compared with the implicit subsidy calculated Vietnam’s water sector. However, the Municipal Fund had not, at that using the commercial loan with a 10.2 percent interest time, completed a water sector investment in Vietnam. 4 | Capital Subsidies Implicit in Concessional Finance: How to Make Them More Transparent and Better Targeted Box 1: Sensitivity of Implicit Subsidy Calculations Variations in the repayment terms, including maturity, grace period and interest rate of a loan can have a significant effect on the level of implicit subsidy. The following table shows the sensitivity of implicit subsidy calculations to changes in these terms of a concessional loan. The implicit subsidy amounts were calculated by setting the present value (PV) of the grant plus a 10-year government bond* (assuming a bond yield of 15 percent and no grace period) equal to the PV of the loan under the relevant concessional on-lending terms. Concessional Repayment Grace period interest rate period (years) (years) 0 3 5 8 10 27.7 49.1 59.3 70.7 Implicit subsidy (as a 15 percentage of loan): 33.9 54.3 64.1 74.7 6.7% 20 38.2 57.8 67.2 77.3 10 36.5 53.8 62.3 71.8 Implicit subsidy (as a 15 percentage of loan): 44.8 60.7 68.5 77.0 4.0% 20 50.4 65.3 72.5 80.5 This table illustrates the sensitivity of implicit subsidy calculations to different loan terms. Based on the calculations above, differences in the grace periods of loans cause the most significant variations. For example, a loan with a concessional interest rate of 6.7 percent, and a repayment period of 10 years could have an implicit subsidy component that ranges from 27.7 percent with no grace period to 70.7 percent with an 8-year grace period. *Note: This example is based on the general approach adopted in Vietnam up until 2010 to illustrate the effect of different choices in relation to interest rates, grace periods and repayment periods. Source: Authors’ calculations subsidies contained within the loan are often not quanti- ential terms for smaller cities. However, the level of subsidy, fied the amount of subsidy may not be clear, i.e., the exact and the extent of the preference for Category Five cities, amount per additional connection. This lack of clarity may arbitrarily differs between programs—a Category Five city hinder a more informed and structured policy debate about that borrows under the AFD program receives only a five alternative uses of public funds. percentage point preference over a Category One city, while the differential under Decision 181 would be about To be well-targeted, a subsidy should be both directed to- 13 percentage points. Similarly, when Decision 181 targets ward a defined policy objective, and paid in a way that pro- a subsidy of 57.8 percent of total loan value for larger towns motes that objective. Yet this is not always easy to achieve, as and cities, the World Bank program offers these towns and illustrated by the following examples. cities a larger subsidy amount, equivalent to 69.7 percent of the total loan value. These discrepancies illustrate how dif- The government of Vietnam’s policy, as promulgated ficult it is to create a scheme that targets public resources through Decision 1817 on-lending terms, seemed to pro- effectively toward government objectives when subsidies mote investment in the water sector in general, and in small- remain implicit. er Category Five cities in particular. This should have resulted in a constant degree of subsidy for all Category One through Category Four cities, along with a constant preference (in the form of a higher degree of subsidy) for Category Five cities. However, studies indicate that this has not occurred. 7 At the time of this study, Decision 181 was the government regula- tion establishing what the on-lending terms should be and it intended that subsidies should vary between the categories of city. This Decision Further, as shown in table 2, the AFD and World Bank pro- has since been superseded by a new approach, which continues to de- grams subsidize all categories of cities, and have prefer- velop the sector’s on-lending mechanism. | 5 Table 2: Implicit Subsidies in Concessional Loans to the Water Sector in Vietnam Subsidy Repayment Grace Period Interest Rate Present Value Component Loan Source Town sizes period (years) (years) (percent ) (percent) (percent) Decision 181 Category Five Cities 1 20 8 4.0 19.5 80.5 Category One to Four Cities 1 20 3 6.7 42.2 57.8 Agence Française de Category Five Cities 25 3 82 0.0 11.6 88.4 Développement (AFD) for Category One to Four Cities 25 3 82 5.0 18.7 81.4 Vietnam Development Bank (VDB)-managed water revolving fund World Bank in Vietnam Urban Category Four and Five Cities 20 3 3.0 31.9 68.1 Water Supply Project I Urban Centers, including 20 5 5.4 30.3 69.7 Category One to Four Cities Bond (Benchmark) 3 10 0 15.0 100.0 0.0 Notes: 1 The specified lending rate was 33.3 percent * L (the current state investment credit rate for VND-denominated loans—which at 5 November 2008 was set at 12 percent) for Category 5 cities, and 55.5 percent * L for Category 1 to 4 cities. However, Decision 181 did not specify the repayment period or grace period for either group of cities. For both groups, the authors have assumed a repayment period of 20 years, which is the cap, as well as a grace period of 3 years for Category 1–4 cities, reflecting the construction period for new water facilities, and 8 years for Category 5 cities, reflecting their need for greater concessionality. 2 These are maximum values. 3 The 10-year government bonds (at November 2008) were issued by Vietnam Development Bank (VDB) at 100,000 dong, and have an interest rate of 15 percent per annum. The government bond was chosen as a benchmark because it was difficult to select a risk-equivalent commercial loan. Arguably, a more sophisticated methodology would show an actual rate for each borrower or a class of borrowers, but this would be highly complex. 2.3 Implicit Subsidies are Widespread In the United States, revolving funds set up at the state level provide concessional lending to local water and sanitation Concessional finance of water infrastructure is common in service providers. States receive matching federal grants both developed and developing countries. Concessional- for fund capitalization. Interest rates on loans to providers ity takes different forms ranging from subsidized ODA on- are typically 50 percent of market rates, though this varies lending to financing via tax-exempt municipal bonds. The by state. States provide an explicit grant component that is following examples illustrate how concessionality prevents transparent, and a concessional loan component, the sub- the subsidy element from being transparent, thereby mak- sidy element of which is not transparent. States are required ing it more difficult to target the subsidy to a clearly defined to disclose information to the public about their programs policy objective (see also Table 3). and the projects that they intend to fund. In the Philippines, the Local Water Utility Administration As all states which establish a revolving fund are eligible to (LWUA), a specialized lending institution that supports water receive federal funding, and all water companies willing to un- investments, typically offers concessional loans with 25-year dertake an eligible project in a state can receive funding from maturities and a four-year grace period on both interest and the revolving fund, the subsidy is not well targeted to a spe- principle. The government and the recipients are aware of cific group of beneficiaries—all projects that expand access the concession rates of the financing, but not its exact value. Any local water district utility formed according to LWUA’s rules with an eligible project can receive a loan. This means that LWUA’s lending is targeted toward encouraging local 8 Output-based subsidies have a cost—they increase the cost of fi- water utilities to corporatize and follow LWUA’s rules. How- nancing, as well as the cost administration in relation to upfront subsi- dies. Capital investments need to be financed (most likely with a loan) ever, the subsidies are not well-targeted to other govern- until the date in which the subsidy is paid. The delivery of outputs ment objectives, such as expanding access or serving poor needs to be verified by an independent agent, thereby increasing sub- households and the exact value of the subsidy is not known. sidy administration costs. 6 | Capital Subsidies Implicit in Concessional Finance: How to Make Them More Transparent and Better Targeted Table 3: Examples of Concessional Finance in the Water Sector Is the subsidy Is the subsidy well- Program Key Water Sector Financing Instrument transparent? targeted? Vietnam The Ministry of Financing in Vietnam on-lends loans from MDBs at highly concessional No Partially but not explicitly rates for water sector investments. Philippines The Local Water Utility Administration (LWUA) provides concessional loans for water No Partially but not explicitly sector investments U.S. State Revolving Funds Revolving funds set up at the state level provide concessional lending to local water and Partially Partially but not explicitly sanitation service providers. U.S. Tax-exempt Municipal Tax exempt municipal bonds used to financing infrastructure, including for water projects. No No Bonds to good quality sustainable water service are eligible for the specific sector objective nor targeted to any specific set of subsidy. States develop a ranking system for projects to deter- beneficiaries other than those that seek to fund public proj- mine funding priority and give priority to projects that serve ects through bond issuance. poor communities or address specific public health concerns. In all of these cases policy makers appear to have made an In addition, municipalities in the United States are allowed underlying, but not explicit, assumption that a subsidy is to sell municipal bonds to finance local projects, including warranted—to encourage expansion of access to water ser- water projects. Interest paid on the bonds is exempt from vices, or because poorer communities need more support federal tax (and from state income tax in the state where the than richer ones, or to encourage formation of corporate bond was issued). The level of the subsidy in the tax exemp- entities on the premise they will provide better service than tion can be quantified, but is not disclosed and consequent- non-corporatized entities—and that the political process ly is not transparent. The tax exemption represents a subsidy will result in public investments that maximize positive ex- from the federal and state governments that is not tied to a ternalities to communities. | 7 3. How to Make Better Use of Implicit Capital Subsidies G overnments could achieve more of their policy objec- a benchmark commercial loan (using the benchmark loan tives with the same level of public resources if implicit interest rate as the discount rate in both cases). subsidies were more transparent, and targeted better. This section provides an overview of an emerging concep- tual framework to improve subsidy design and presents a Step 2: Disclose the Subsidy step-by-step approach to apply the framework to an implicit If policy makers disclose the subsidies they can encourage subsidy scheme. The framework is based on the premise that debate among officials, politicians and the public about an implicit subsidy exists, and that the government wants to whether these subsidies are in the public interest, or wheth- make this subsidy explicit and better targeted. It consists of er they should be changed. Simply knowing how much four basic steps: subsidy exists provides clear value added for more informed policy debates. A key question relates to the best method to ■ Step 1: Quantify the amount of the subsidy disclose the subsidy. There are essentially two options. The ■ Step 2: Disclose the subsidy first option is for a government to disclose the value of the ■ Step 3: Define policy objectives subsidy by “administratively� indicating on a concessional ■ Step 4: Design a subsidy mechanism to achieve policy loan agreement the equivalent grant amount that the bor- objectives including: rower is effectively receiving by taking the loan on conces- a: Eligibility criteria sional rather than commercial terms. The second option is b: Payment mechanism for a government to actually separate the concessional loan c: Subsidy amount per recipient into two components—a commercial loan and subsidy el- ement. The different operational implications between the A summary of the framework is provided in figure 1 and the two options are discussed in section 4. specific steps described in greater detail below. Step 3: Define Policy Objectives Step 1: Quantify the Total Amount of After changing the implicit subsidy to an explicit subsidy, the Subsidy policy makers can then determine the most appropriate way The starting point for reform is to establish the value of to target the subsidy in order to achieve policy objectives. the implicit subsidies currently provided. The subsidy can Policy makers should first define the policy objectives that be quantified using the methodology described in sec- the subsidy will target, and then determine whether a sub- tion 2.1—comparing the present value of debt service on a sidy is necessary to meet these objectives. While policy mak- concessional loan with the present value of debt service of ers will need to determine their own objectives given the 8 | Capital Subsidies Implicit in Concessional Finance: How to Make Them More Transparent and Better Targeted Figure 1: Improved Capital Subsidy Design Concessional Loan Actual Benchmark Terms Rate 1) Quantify the Total Amount of Subsidy Transparency 2) Disclose the Subsidy 3) De ne Policy Objectives Improved Targeting 4b) De ne How Much 4a) Eligibility 4b) Payment Subsidy each Criteria Mechanism Recipient Gets New Subsidy and Financing Design specific country context, some common objectives in the ■ Income—the government may want to target low in- water sector include expanding access, improving services, come households and enhancing efficiency. ■ Utility Performance—the government may want to en- courage and invest in utilities that are financially viable, technically efficient, or well-governed. Step 4: Decide on Subsidy Allocation through Three Interrelated Selecting eligibility criteria almost always involves a compro- Design Criteria mise. Often the best performing utilities are in the richest and most developed regions of the country, creating a trade-off Step 4a: Eligibility criteria—Who will be eligible between rewarding performance and helping disadvantaged for the subsidy regions. How the government resolves this dilemma depends After the government has defined its policy objectives, it needs on the policy objectives and their relative importance. to take three interrelated steps to improve the targeting of sub- sidies. The first is to define those eligible to receive the subsi- Step 4b: Payment mechanism—What the subsidy dy—targeting recipients that are consistent with the policy ob- is for, and how it will be paid jectives. The following criteria can be used to define eligibility: Governments need to determine the most appropriate pay- ■ Location—the government may want to target rural ar- ment mechanism—what the subsidy is actually paid for. There eas or less developed regions of the country are three broad types of payment mechanisms for subsidies: | 9 ■ a subsidy for inputs, particularly for capital expenditure Step 4c: The subsidy amount per recipient and on new infrastructure; how the amount should be calculated ■ a subsidy for outputs, such as new connections in- Once the government has made the implicit subsidies ex- stalled; and plicit, defined the policy objectives and chosen the preferred ■ a subsidy for performance, such as meeting certain fi- payment mechanism, it needs to decide the amount of sub- nancial, technical, or governance targets. sidy each recipient will receive. In principle, this amount should be based on the social value of the output created. Traditional concessional loan schemes are essentially subsi- dies for inputs because concessional loans are usually pro- Some objectives, however, are more difficult to value in mon- vided to finance new capital expenditure. Governments etary terms—for example the value to society of a household may continue to provide a subsidy for capital expenditure, receiving a new water connection—but there are a number even after making the implicit subsidy explicit. However, of common ways to determine the level of subsidy. once the implicit subsidy has been made an explicit subsidy, the subsidy element can then be separated from the loan ■ Subsidize a percentage of the capital cost. element. It is then possible to pay the subsidy for reasons ■ Determine the subsidy needed on a case-by-case basis other than simply capital expenditure, for example, on an to cover the difference between the cost of the service output basis (see below). When governments find that sub- (including full capital cost recovery) and the tariffs paid sidizing capital expenditure is not the optimum method to by end-users. improve services—for example because the infrastructure is ■ Set a fixed subsidy for each unit of output and allow util- poorly planned, or is not maintained and so fails to deliver ities to decide if this fixed subsidy is sufficient to cover as expected—they may choose to switch from subsidizing their net costs of expanding service. inputs, and instead subsidize outputs. ■ Establish a challenge fund in which utilities compete based on the amount of subsidy they require, and award Governments are increasingly using output-based subsidies the subsidy to the utilities that require the least subsidy. to ensure that the subsidy delivers the desired results. Out- put-based subsidies could be given for installing new work- ing connections, or as a specified contribution toward the cost of water delivered to low-income customers. If the sub- sidy payment is based on the value of the output, and not the cost of the inputs, then output-based subsidies not only encourage results, but also provide incentives for efficiency.8 10 | Capital Subsidies Implicit in Concessional Finance: How to Make Them More Transparent and Better Targeted 4. Operationalizing Improved Capital Subsidy Designs G overnments might be able to achieve more of their pol- 4.1 How to Make an Implicit Subsidy icy objectives with the same level of public resources if Explicit the preceding emerging framework is implemented— thereby providing an opportunity to increase the value of con- There are several ways to make implicit subsidies explicit, cessional financing from MDBs for water sector investments by and each has different advantages and disadvantages (see making implicit capital subsidies more explicit and targeting table 4). This section considers the advantages and disad- them to more clearly defined public policy objectives. vantages of two possible approaches which can be used once the subsidy amount has been determined using the If policy makers adopt this emerging framework, or a similar formula presented in section 2. These are: one, they will probably encounter some practical implemen- tation issues when it is applied. Although these will be very Disclosing the concessional loan in terms of its constituent situation specific, this study has identified a number of ge- parts. This means disclosing the equivalent values of a loan neric issues that are likely to be relevant. This section outlines on commercial terms and a capital subsidy component (“dis- four of them. closing�) Section 4.1 discusses how to make the implicit subsidy ex- Separating the concessional loan into a loan on commercial plicit. Section 4.2 highlights the changes in a borrower’s cash terms and a subsidy (“separating�). flow that may result from converting a concessional loan into a commercial loan and a grant. Section 4.3 explores which While the disclosure option begins a process of raising sector government entity is best placed to administer an explicit awareness about the value and cost of the implicit subsidies subsidy. Section 4.4 outlines how the amount of the subsidy inherent in concessional loans, it is a weak signal which does could be determined while section 4.5 considers how the not fundamentally change the financing arrangements or capital subsidy framework, once established, can evolve. incentives in the sector. It represents the business-as-usual Table 4: Comparison of Options for Making Capital Subsidies Explicit Option Advantages Disadvantages Disclosing ■ Simple implementation ■ Limited opportunity to use subsidy to create incentives for borrower to achieve ■ Makes the subsidy element explicit and starts to change awareness in outputs the sector ■ Sends weak signals to the sector on financing issues Separating ■ Explicit subsidy could be used to create incentives to borrower to achieve ■ More complex to implement because it will require borrower to be creditworthy on results commercial loan component, and introduces more risks to borrower by bringing ■ Clear signal to borrower of cost of commercial funds forward loan repayments and delaying subsidy disbursements. ■ First step towards transparent sector financing that will facilitate the entry of domestic capital into the sector. Source: Authors. | 11 approach, albeit with greater transparency on the subsidy There are, however, some disadvantages if the concessional amount. loan is separated into a loan on commercial terms and a grant component. It would certainly be more complex for Governments and policy makers have sound reasons to a government to implement than simply disclosing the sub- progress from disclosing the capital subsidy to a more ex- sidy. Also, moving from a concessional loan to a commercial plicit separation of the components. If the concessional loan loan combined with a grant can affect the borrower’s cash is separated into a loan on commercial terms and an explicit flow (see section 4.2). However, if implemented carefully subsidy component this provides the clearest distinction and gradually, this approach offers a solid foundation for the between the amounts of the commercial loan and subsidy long-term financing of the sector. components. Importantly, a clearly separated subsidy com- ponent will create the opportunity for other financial insti- tutions to provide the loan component. This will allow, and 4.2 Cash Flow Impact of Converting to encourage, a gradual development of local private sector a Loan on Commercial Terms and financing for the water and wastewater sector. This should be a long-term sector goal of any government in a develop- a Grant ing country as it will release donor funds for other purposes If an existing concessional loan is separated into a loan on and introduce greater accountability and sustainability into commercial terms and a grant it may have a significant im- the sector. pact on the cash flow of the borrower, even if the net present Box 3: Impact on Cash Flow of Concessional Loan versus Separate Commercial Loan and Capital Subsidy The concessional loan in the first figure below has a maturity of 20 years and a grace period of eight years. The loan is disbursed in full to the borrower in the first two years, and repayment of the concessional loan begins in year eight. The second figure shows the disbursement and repayment of a loan on commercial terms and an explicit output-based subsidy component. The loan depicted has a maturity of ten years with no grace period and is disbursed in full in the first two years of the loan. Repayment of the loan begins after the loan is disbursed in year two and is completed in year ten. The explicit capital subsidy is disbursed based on achievement of outputs in the third, fourth and fifth years of the agreement. This approach illustrates a longer term model where the incentivizing potential of the separation model is maximized. In the short-term the capital subsidy would be paid in year one, along with the loan. Concessional Loan Repayment of Concessional Loan 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Disbursement of Concessional Loan Commercial Loan Plus Capital Subsidy Repayment of Commercial Loan 1 2 3 4 5 6 7 8 9 10 Disbursement of Capital Subsidy Disbursement of Commercial Loan 12 | Capital Subsidies Implicit in Concessional Finance: How to Make Them More Transparent and Better Targeted value of the debt service remains the same. The loan may cally on-lent by recipient countries’ ministries of finance. The originate from the government, but it may also come from government would also need to decide whether the minis- a commercial lending institution and this has implications try of finance is the appropriate administrator of the subsidy, for commercial banking relationships, as the change in cash or if other institutions are better placed to be the subsidy flow is likely to change the risk profile of the borrower in the administrator. view of potential commercial lenders. In defining who will administer the subsidy, the government The borrower’s cash flow from loan proceeds will change should consider whether there are institutions already per- because the amount of the commercial loan is less than the forming this function, if they are performing it well, and if concessional loan, but with a higher interest rate and shorter the institutions are independent of any banks that might grace period. Further, if the disbursement of the capital sub- provide the commercial loan. This independence from com- sidy is tied to the borrower achieving certain outputs, the mercial lenders is important to ensure that subsidy approvals likelihood and timing of the disbursement of the subsidy are not distorted by banks’ commercial interests if they are may also change. providing the loan component. The indicative cash flows for the concessional loan (with These criteria give the government several options. One pos- implicit subsidy) and the loan on commercial terms with an sibility is that a finance ministry continues to be responsible explicit, output based, capital subsidy are presented in the for the administration of both the lending component, on two figures in box 3 with the disbursements of the loans and commercial terms, and the grant component. Alternatively, subsidies represented by downward arrows, and the repay- the ministry could channel the commercial loan component ments of the loan represented by upward arrows. through a commercial bank, and channel the grant compo- nent through a separate grant-making facility. This facility The change in the borrower’s cash flow with the separation could be housed within an existing institution—such as a model may make potential lenders more cautious about the state development bank—and could provide grants to utili- borrower’s ability to repay the loan component. Without the ties that borrow at commercial rates from the commercial substantial grace period and lengthy maturity of the conces- bank. Clearly, many other combinations and permutations sional loan, borrowers will need to pay back more of the loan are possible and these would be determined on a case-by- earlier than under the concessional on-lending scheme. As case basis. a result, when evaluating whether to provide the borrower with a loan, the lender is likely to require the borrower to have more cash available to cover the earlier repayment of 4.4 Determining the Subsidy the loan. Amount—the Benchmark Loan Section 2 presented a methodology to calculate the value Cash flow considerations highlight the need to consider a of the subsidy contained in a concessional loan. This illus- planned and gradual roll out of any separation model. There- trated how the variables of interest rate, grace period and fore, governments may need to continue to provide conces- repayment period affected the value of the subsidy amount. sional loans to weaker service providers under a “disclosure The calculation assumes the existence of a benchmark loan model�, while moving stronger service providers into a sepa- (i.e. a loan from a commercial institution offering terms ap- ration model. propriate to the assessed riskiness of the borrowing water company), with the default proposal being the rates from 4.3 Define Who Administers the a long-term government bond in the absence of material commercial lending into the sector. Subsidy If the government separates the concessional loan into a In developing countries the commercial benchmark loan loan on commercial terms and an explicit capital subsidy, it may not exist, or its terms may be highly volatile (thus af- will need to decide how to administer the subsidy compo- fecting the value of the subsidy between one period and the nent. Concessional loans from development banks are typi- next). Consequently, a benchmark that is more stable, readily | 13 understood, and adequate for the purpose is required. That to ease implementation. Therefore, they may choose to start benchmark is likely to vary from country to country but its with the “disclosure model� for the implicit subsidy, provide selection should follow these criteria. the funds as payments for inputs, and continue with eligi- bility criteria similar to those in the existing concessional fi- It is important to note that this working paper does not nec- nance scheme. essarily attempt to develop the best or most intellectually satisfying subsidy model, but aims to improve understand- Periodically the design of the subsidy may be reevaluated, ing about the amounts and targeting of subsidies. It will be and if needed, improved to meet the government’s policy sufficient if the benchmark loan is viewed by sector profes- objectives more fully. This could mean that: sionals as a productive new way to consider subsidies. Later, as the sophistication of the sector increases and more com- ■ The method of making the subsidy explicit evolves from mercial lending occurs, policy makers can take a more nu- the “disclosure model� to the “separation model� anced approach to setting the benchmark loan. ■ The payment mechanism evolves from payment for in- puts to payment for outputs ■ The eligibility criteria evolve to allow more effective tar- 4.5 Evolution of Subsidy Design geting of the subsidy When governments take initial steps to make implicit sub- ■ Within the sector, the focus of the subsidy will shift from sidies explicit, they are likely to select a simplified approach water supply and toward wastewater (see box 4). Box 4: Subsidy Design Should Evolve from Water Supply to Wastewater Subsidies for the water sector are commonplace. Policy makers in developing countries have frequently chosen to provide subsidies for the water sector, and have used concessional finance to do so. Often, the water sector choice is subsidized at the expense of wastewater services. Consequently, many developing countries have larger, more advanced water sectors and smaller, less-developed wastewater sectors that receive comparatively smaller subsidies in dollar value terms. As the water sector matures—for example, as a region becomes more developed, levels of access to water services may increase, and higher incomes enable people to begin paying more for services—policy objectives may change. Governments may then reduce the level of subsidy to the water sector and instead target that subsidy to more ap- propriate policy objectives, perhaps by increasing the subsidy for wastewater services—where there are good public considerations, significant negative externalities, and a low willingness to pay by consumers. Therefore, the subsidy may gradually evolve a dual focus with a smaller, poverty alleviation targeted subsidy for water supply services, and a majority of the subsidy focused on improving wastewater services. Source: Authors 14 | Capital Subsidies Implicit in Concessional Finance: How to Make Them More Transparent and Better Targeted 5. The Way Ahead C oncessional finance for the water sector involves im- tion of stronger incentives (e.g. by making the subsidy pay- plicit capital subsidies that are generally not transparent ments output based) and facilitates increasing amounts of and often not well-targeted. This lack of transparency local private sector financing (e.g. by encouraging the com- and targeting means that the implicit subsidies have a signifi- mercial component to be sourced in part or wholly from lo- cant opportunity cost. There is, therefore, an opportunity to cal capital). In addition, the model can be actively managed obtain better value from existing concessional finance sourc- to allow a planned and gradual evolution of sector financing es by making implicit subsidies more transparent and then in a meaningful way (e.g. by changing the level and target- targeting them more effectively to agreed policy objectives. ing of the subsidy component). This working paper outlined an emerging framework on However, despite the potential benefits of improved trans- how to make capital subsidies more transparent and provide parency and better targeting of capital subsidies, a signifi- a basis for better targeting. Specifically, the paper suggested: cant amount of conceptual work remains, particularly in two areas. First, it will be necessary to identify a suitable bench- 1. Quantifying the implicit subsidy mark rate to calculate subsidies—on which the suggested 2. Disclosing the subsidy framework to improved capital subsidy design depends. 3. Defining policy objectives Second, it will be equally necessary to identify the best in- 4. Designing the subsidy mechanisms to achieve policy stitutional mechanisms of transmitting the subsidy and loan objectives on commercial terms. The shift from concept to reality will inevitably result in many Finally, this paper does not argue that all concessional fi- practical challenges. This paper highlights some examples of nance programs need to be redesigned. However, given the these, particularly when concessional finance is fully sepa- importance of the policy goals, and the amount of implicit rated into an equivalent loan on commercial terms and grant capital subsidies currently being deployed, it is evident that components. However, the separation model provides the disclosing the subsidies and ascertaining how well they are greatest opportunities to use the subsidies in the best long- aligned with policy goals would be a beneficial develop- term interests of the sector. This model allows the introduc- ment for concessional financing in the water sector. | 15 Selected Bibliography Castalia: Developing a New Financing Framework for the Urban Water Sector in Vietnam: Final Report to World Bank�, December 2008. World Bank: Allocating Donor Financing to Vietnam’s Water Sector, mimeo, June 2006 World Bank: Opportunities for a New Financing Framework, mimeo, October 2006 World Bank: Independent Evaluation Group: Water and De- velopment, 2010 16 | Capital Subsidies Implicit in Concessional Finance: How to Make Them More Transparent and Better Targeted WORLD BANK 1818 H Street, NW Washington, DC 20433