WATER P-NOTES ISSUE 29 FEBRUARY 2009 47362 Toward Cleaner, Cheaper Power Streamlined Licensing of Hydroelectric Projects in Brazil B razil is confronted with steadily increasing de- and transparent process for licensing and regu- mands for electricity. The country has the ability lating hydroelectric generation projects that will to meet that demand by developing its consider- not discourage the investment needed to satisfy able hydropower potential, but the regulatory pro- future demand in a sustainable way. The history of cess that governs the approval of new hydroelectric Brazil's energy industry casts long shadows over plants imposes unnecessary delays that push up its ability to keep up with demand. For years the project costs and increase uncertainty. The process, industry was controlled by the government; when among other reasons, has created a shortage of in- its financial capacity collapsed in the late 1990s vestment in otherwise viable hydropower projects in the era of public and private partnership began. favor of less efficient and more harmful technologies. Much of the current legal, regulatory, technical, Brazil's electricity sector serves roughly sixty and institutional framework was put in place at that million residential and commercial customers time--and in a rather ad hoc fashion. Fortunately, and generates revenues of US$20 billion. With some relatively simple changes in this structure can demand growing at a rate of 4.4 percent annu- bring a more unified, holistic approach to this criti- ally, an additional 3,000 megawatts of generating cal energy sector, in which Brazil has such great power will be needed by 2015. The cost of the potential. Other pieces of the puzzle will require new power plants needed to provide that power is more ambitious reforms. estimated at US$40 billion. Presently, five-sixths of the country's power needs are met by hydroelec- tric plants, though in recent years only half of the Costs of the licensing process new plants receiving licenses to begin construc- tion have been hydroelectric. The other half of the The costs of obtaining a license for a new hydro- licenses have been issued for coal, diesel, and electric plant in Brazil are such that projects must nuclear plants that provide electricity at higher unit promise an annual return on investment of 15 costs than hydroelectric plants and have greater percent, almost double the breakeven rate in well- adverse effects on people and the environment. regulated countries such as Chile. What makes prof- The seeming anomaly can be explained by the itability so elusive in Brazil are the substantial social fact that the licensing process for thermal plants is costs assessed as part of the licensing process, as simpler and more predictable than that for hydro- well as the uncertainty and opportunity costs associ- electric plants. ated with the regulatory regime. The excessive costs cannot be imputed to environmental mitigation, as Escaping this predicament will depend on the the environmental costs of hydroelectric plants are country's ability to devise a modern, predictable, much lower in Brazil than in many countries. This note highlights important information from "Environmental Licensing of Hydroelectric Projects in Brazil: A Contribution to the Debate," Volume 1, World Bank Report 40995-BR, March 28, 2008. Readers may download the complete paper from www.worldbank.org/water. WATER P-NOTES Table 1. Environmental licensing costs as a share of total costs of hydroelectric generating projects Cost US$ per installed kilowatt Share of total (percent) Direct costs 133 14.1­14.7 Social 94 10.0­10.4 Environmental 19 2.0­2.1 Regulatory uncertainty 20 2.1­2.2 Indirect (opportunity) cost 10­50 1.1­5.4 Total costs of environmental licensing 143­183 15.2­20.1 Total project costs 906­946 100 Note: The total cost per installed kilowatt provided by Eletrobrás is US$876. This excludes the regulatory uncertainty and opportunity costs shown here, which do not directly affect plant construction costs. Almost three-quarters of the direct costs of the Legal and regulatory issues licensing process stem from the various social con- tributions that developers must make in exchange Other aspects of the licensing process are more for their license (table 1). Those contributions fund susceptible to quick change than are the social road and school construction and the distribution costs just reviewed. of goods to the poor. Whether such costs should be borne by developers alone is a political matter that The legal framework surrounding Brazil's hy- deserves careful scrutiny. Government could help droelectric power industry has remained largely make hydroelectric projects more attractive to inves- unchanged since the 1980s. One of the biggest tors by sharing the burden of social-improvement problems is jurisdictional overlap and lack of activities tied to the projects, justifying the cost-shar- coordination between the states and the federal ing on the grounds that successful projects benefit government. Investors often are unsure which level everyone. has jurisdiction in a given situation, and state and federal requirements sometimes conflict. Clarifying The direct costs paid by developers during the jurisdictional responsibilities at every step would licensing process make hydroelectric projects less ec- greatly improve the licensing process and along onomically viable, but the indirect costs of delaying, with the adoption of alternative dispute resolution postponing, or rejecting the projects are borne by mechanism to avoid protracted court battles. the country as a whole. Since demand for electricity is ever-increasing, the lengthy timeline for licensing Public prosecutors in Brazil have the authority such projects means that less-desirable alternatives to intervene in any stage of the process, introduc- must fill the gap to stave off shortages. Plants that ing yet another level of uncertainty. Although the burn diesel or coal, though generally less expensive public prosecutors have qualified staffs, ample re- to build, are more expensive to operate, resulting in sources, and a mandate to ensure favorable social electricity costs that are roughly 10 percent higher outcomes, they have tended to act in an adversarial than with hydroelectric plants. In addition, the envi- rather than a problem-solving manner. Too often, ronmental damage associated with coal, diesel, and disputes land in the courts, stalling projects and nuclear generation is much greater than with hydro- raising doubts about whether they will emerge intact electric generation, and the total long-term costs are from the legal proceedings. difficult to accurately assess. A significant increase in If Brazil's prosecutors and its environmental and the efficiency of the hydroelectric licensing process energy agencies were to cooperate to establish ex- would quickly reduce reliance on these more prob- plicit guidelines for hydroelectric projects, problems lematic sources of electricity. would be less likely to arise, as developers would 2 ISSUE 29 · FEBRUARY 2009 understand upfront what they must do to meet so- state and federal government experts, not only to cial goals. The creation of an overarching body to reduce the cost to industry of preparing the reports, handle disputes, one comprising experts from all but also to increase confidence in their conclusions. relevant sectors, would also help to ensure that the Ideally, the government should take a systematic in- interests of all stakeholders were further protected. ventory of the biological and physical environment The structure of the licensing process itself of the country. Given the current lack of informa- could benefit from revision. Brazil stands alone tion, officials tend to be suspicious of information among nations in having a three-stage process, provided by developers and apply the "precau- each with separate requirements. During the proj- tionary principle" of rejecting projects where the ect planning stage, an environmental impact as- benefits seem even remotely questionable. That the sessment must be submitted before a preliminary legal system holds officials personally liable for the license is issued. In the second stage, as a condi- negative consequences of their decisions encourag- tion for receiving a second license to begin con- es excessive caution; changing this law to absolve struction, developers must submit an environmental individual bureaucrats of legal responsibility, except impact report that encompasses the engineering in cases of gross malfeasance, would mitigate this aspects of the project and its ramifications for the tendency. local population. In the third stage, once all condi- Better communication between governing bod- tions for the project are met, an operating license ies would further improve the licensing process. It is granted. is understandable that different agencies should In a multi-stage process, disputes can and do focus on different aspects of a particular project, occur repeatedly, and the initial conditions under yet a more holistic approach would enable poten- which a project is undertaken often change-- tial developers to fine-tune their projects from the sometimes dramatically. What is needed is a more start. The establishment of independent consultative coherent approach in which all project conditions panels with multi-sectoral expertise could provide are specified at the outset and government agencies much-needed guidance for developers, expertise work closely with developers to resolve disputes as that should be codified in an operations manual they arise. to assist both developers and officials. Building the capacity of agency staff is needed where agencies are understaffed and lack qualified specialists to review proposals in a timely manner. The paucity Technical and institutional issues of explicit decision-making criteria, already men- tioned, compounds the effect of capacity shortages One of the biggest challenges in preparing the and increases the likelihood of subjective, unpre- environmental impact assessment and report is that dictable agency findings that raise project costs and the terms of reference that governments issue to cause delays. guide project proposals tend to be overly general and not site-specific. When factors that might affect Perhaps a significant improvement would come the decision to issue or refuse a particular license from creating a governing council that engaged are omitted from the terms of reference, they tend all levels of government and the various agencies not to be addressed in the environmental impact and sectors involved in the planning and approval assessment and report. Delays ensue when the process. Such a body would provide a unifying licensing body must then request the missing infor- presence that would inevitably lead to greater un- mation. This situation could be avoided by devel- derstanding and awareness of the multiple needs oping site- and region-specific terms of reference, that projects must address and greater certainty in with each of the agencies involved in the licensing the likely outcome of the licensing process. A gov- process specifying in detail what it will require to erning council would also be able to remove po- approve the project. Notwithstanding, the quality tential obstacles from the beginning and serve as a of reports prepared by developers is generally very forum for the resolution of problems that might oc- poor. cur along the way. What is needed is a new spirit of cooperation, not only between the public and Baseline information for environmental impact private sectors, but--importantly--within the public assessments and reports should be provided by sphere. 3 WATER P-NOTES Statistics, scenarios and hypotheses Basic relevant statistics 1. The average total time for a UHE to go on line in Brazil, considering the sample employed, is 6.5 years, albeit with high variances. This average is approximately 30% longer than those incurred in the USA. 2. Issuing a Preliminary License (PL) takes 2.5 years on average, with1.1 years (44%) spent on preparation of the TOR by IBAMA. Developers take an average of only 6 months to submit an EIA/RIMA. 3. Issuing an Installation License (IL) (including the time required to issue the PL) takes 3.4 years. 4. According to ANEEL (the energy regulatory agency), of the 66 hydro plants of the portfolio examined, only 27 have exceeded the dates planned for their contracted start-up. Scenarios and hypotheses for delays 1. We have assumed that all delays beyond the contract end-date were due to delays in environmental licensing, which clearly overestimates the opportunity costs of such delays. 2. Only 8 units remained delayed at the time data collection was finalized for this study. Two possibilities are considered for the start-up dates: (i) using the past average of 78 months; or (ii) devising scenarios for these data beginning in September 2007 and splitting them into consecutive 6-month periods (March 2008, September 2008, etc.). 3. Since some power units have started operations earlier than foreseen, we again adopted two alternatives: ignoring the schemes completed ahead of schedule (adiantamentos)--(which provide for an 11.8-month average delay per plant), or compensating the latter with the delays incurred by the others (thereby reducing total delays by 25%). 4. Another hypothesis in the analysis (of a static type) was to ignore (or not) grossly delayed plants such as Cubatão and Monjolinnho. These alone were responsible for 22.5% of total delays. By excluding these two units, the average delay was reduced from 11.8 to 9.2 months. Scenarios and hypotheses for the marginal costs of alternative energy 1. Two different approaches for assessing the substituted energy cost were adopted: the first hypothesis is that the `substitute' plants would also be hydropower plants. In the second hypothesis, we assumed that substitute energy would be provided by `thermal' plants. In the first case, the change in total costs per MW of installed capacity was estimated by using a multiple linear regression based on the size of the plant and on temporal values. For thermal alternative energy the difference in unitary prices between thermal and hydropower energy was assumed to be that used in the 2005 auction (19.8%) and we applied it to all the units in our data base. Total opportunity costs of licensing delays These costs depend on three elements: the monthly opportunity cost of a delayed hydropower plant, the type of substitute source, and a composite factor reflecting both the cost increase over time (temporal value) and the number of months by each unit was delayed. The first was a three-factor product: the cost of installed capital per kW for the delayed plant, its size, and a monthly capital remuneration factor (assumed to be 1% per month). Assuming that the other plants substituted for the plants delayed by licensing problems, and assuming that the `before schedule' units were ignored, we obtained an estimated total opportunity cost for delays of 6.6% of the installed capacity costs. Other hypotheses could, if taken into account, reduce this percentage to a mere 1.0%. The Water Sector Board Practitioner Notes (P-Notes) series is published by the Water Sector Board of the Sustainable Development Network of the World Bank Group. P-Notes are available online at www.worldbank.org/water. P-Notes are a synopsis of larger World Bank documents in the water sector. 4 THE WORLD BANK | 1818 H Street, NW | Washington, DC 20433 www.worldbank.org/water | whelpdesk@worldbank.org