39897 Risk RENEWABLE Assessment ENERGY Special Report March 2007 Methods for RISK ASSESSMENT Power Utility Planning Risk Assessment Methods for Power Utility Planning Donald Hertzmark Special Report 001/07 March 2007 Energy Sector Management Assistance Program Copyright © 2007 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Produced in India First printing March 2007 ESMAP Reports are published to communicate the results of ESMAP's work to the development community with the least possible delay. The typescript of the paper therefore has not been prepared in accordance with the procedures appropriate to formal documents. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, or its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The Boundaries, colors, denominations, other information shown on any map in this volume do not imply on the part of 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 copyrighted. Requests for permission to reproduce portions of it should be sent to the ESMAP Manager at the address shown in the copyright notice above. ESMAP encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. RENEWABLE ENERGY Special Report 001/07 March 2007 Masami Kojima Risk Assessment Methods for Power Utility Planning Donald Hertzmark Energy Sector Management Assistance Program (ESMAP) In Memoriam: Shimon Awerbuch One of the rewards of international development work is the opportunity to meet, often unexpectedly, top professionals and thinkers. The best of them can change the way we look at some part of the world. Shimon was such a professional. His zeal and verve to bring the best thinking in financial economics to bear on energy problems and his matchless ability to explain his ideas with rigor, clarity and humor marked Shimon as a unique contributors to our work. We were shocked and saddened to learn of his sudden passing in an airline accident in late February, along with his companion and child. We would like to dedicate this volume to his memory, since he had so much to do with generating the ideas that percolate through these pages. Donald Hertzmark, Tae Jung and Anil Cabraal. March 2007. Contents Foreword vii Acknowledgments ix 1. Outline of this Report 1 2. Introduction 3 What are the Issues which Need to be Addressed? 4 3. Synopsis of the Expert Consultation Workshop 7 Current Approaches to Electricity System Investment Modeling ­ What Works and What Needs Fixing 7 The Single-company Approach ­ WASP and Related Models 7 Systematic Treatment of Risk Identification and Risk Mitigation ­ an Initial Approach 10 Investment Decisions for Public and Regulated Generators 16 Alternative Formulations of Investment Decisions and Risk-mitigating Measures 20 How much Consideration of Risk can be Incorporated into a Coherent Modeling Framework? 23 Case Studies 25 Practical Applications of Multiple Model Approaches 27 4. Results of the Workshop and the Way Forward 29 Findings, Conclusions and Recommendations 29 Workshop Findings 29 Workshop Recommendations 33 Annexes World Bank/ESMAP Workshop on Electricity Investment Modeling and Risk Mitigation 37 Risk Assessment Methods for Power Utility Planning Table 3.1: Modeling Combinations: Needs and Resources 27 Figures 3.1: Wien Automatic System Planning Package (WASP) 8 3.2: Importance and Controllability of Power System Risks 11 3.3: Importance and Controllability of Power System Risks ­ Nuclear-dominated System 12 3.4: Importance and Controllability of Power System Risks ­ Gas/CCGT-dominated System 12 3.5: Importance and Controllability of Power System Risks ­ Hydro-dominated System 13 3.6: Siemen's Model for Risk Assessment 15 3.7: Modeling Trade-offs: Capabilities vs Transparency (Longer Bars are Better) 19 3.8: Capabilities and Needs in Power System Investment Planning 19 3.9: Risk and "Return" for Three-technology US Generating Portfolio Assumed Cost for Riskless Renewable: US$12/kWh 20 3.10: Wind/RE Lowers Mexico Generating Cost 25 3.11: One-step Analysis for Planners 26 4.1: Capabilities and Needs in Power System Investment Planning 30 4.2: Complex Tools Allow Earlier Recognition of Risks 31 4.3: Importance and Controllability of Power System Risks 32 Boxes 3.1: The Evolution of WASP as a Planning and Simulation Tool 8 3.2: Fitting Reality into a Four-part Grid: How to Work with the Risk Matrix 10 3.3: Portfolio Theory: What it is and What it can Do 15 3.4: What are the Decision-support Needs of a Publicly-owned Integrated Utility? 18 3.5: How does the World Energy Council see the Role of Risk Analysis in Power Generation Investments? 22 iv Acronyms and Abbreviations A&E Architecture & Engineering ANL Argonne National Laboratory CCGT Combined Cycle Gas Turbine CFE Federal Electricity Commission (Comisión Federal Electricidad) CRE Comisión Reguladora de Energía DMC Developing Member Countries DSM Demand-Side Management EMCAS Center for Energy, Environmental and Economics System Analysis ENPEP Energy and Power Evaluation Program ESMAP Energy Sector Management Assistance Program FIs Financial Institutions GE General Electric GEF Global Environment Facility HFO Heavy Fuel Oil IAEA International Atomic Energy Agency IEA International Energy Agency IPP Independent Power Producers IRP Integrated Resource Planning LCP Least Cost Planning LOLP Loss of Load Probability MAPS Multi-Area Production Simulation MARS Multi-Area Reliability Simulation MB Multiseller Multibuyer Market Structure NERA International Economic Consulting Firm NY New York PJM Regional Transmission Organization (Pennsylvania Utility Regulator) PPA Power Purchase Agreement RE Renewable Energy RoR Run of River RPS Renewable Portfolio Standards SOE State-owned Enterprise TA Technical Assistance WASP Wien Automatic System Planning Package v Foreword More than 15 years ago, the World Bank reported in a series of papers that the planning environment for the Bank's client electric utilities was becoming increasingly difficult. The papers cited the increasingly unstable nature of world commodity markets combined with the unreliability of demand forecasts and uncertainty or even systematic bias that plagued cost estimates for new power plants and called for an approach to electricity system planning which could account explicitly for risk and uncertainty. Today, the situation for power system planners is even more fraught with uncertainties about prices and costs. The continued volatility of fuel prices has increased significantly the normal risks of projecting future investment economics. In addition, the structures of the power systems themselves, formerly stable as vertically integrated monopolies, now come in a variety of states of reorganization, privatization, unbundling and market orientation. For some time, it was widely thought that the adoption of market transactions for electricity would sidestep the analytical issues raised in earlier investigations. As has become increasingly apparent, the various degrees of marketization have not resolved fundamental issues of risk assessment, with the increased attention to environmental protection and use of renewable resources in electricity generation adding new dimensions of risk to traditional investment assessments. The introduction of new parties into the power market through Independent Power Producers (IPPs), power marketing companies and privatization of distribution in some countries has introduced new and more sophisticated financial concepts into electricity investment analysis. The overall thrust of the financial community is to systematize the assessment of risk, so as to plan for its mitigation. In a sense, the current work is then a logical extension of earlier attempts to identify and mitigate risks, keeping in mind the new tools which are available today. With the traditional risks and newly identified risks in mind, the Energy Sector Management Assistance Program (ESMAP) put together an expert consultation workshop on June 27-28, 2006. The workshop brought together practitioners, analysts, regulators and software vendors to discuss the changes in both the external environment affecting the use of simulation models for investment planning and the advances in the art and science of such modeling over the past few years. The analytical and empirical presentations from the workshop, discussed in this ESMAP report, indicate renewed interest in the overall subject of investment assessment, especially in regard with the integration of risk analysis. The conclusions from this workshop point towards greater use of more detailed and integrated approaches to modeling investments in unbundled and privatized power syste. vii Risk Assessment Methods for Power Utility Planning This workshop has pointed the way towards several concrete activities. These activities entail integration of different types of analytical models to capture the differentiated functions in modern utility system, assistance to borrowers in the construction of risk assessment and mitigation simulations and provision of training to member-countries in the new techniques. Jamal Saghir Director, Energy, Transport and Water Chair, Energy and Mining Sector Board viii Acknowledgments This project was undertaken by the ESMAP unit under the overall management guidance of the Manager, Mr. Ede Ijjasz-Vasquez. The project manager was Mr. Tae Yong Jung (ESMAP) and coordination was provided by Mr. Anil Cabraal (ETWEN). This team played an essential role in providing structure and focus to the workshop and the subsequent ESMAP report. Oversight and coordination with World Bank initiatives in this area was overseen by Mr. Enrique Crousillat (LCSEG), who also gave useful comments on the selection of topics and on the earlier drafts of the ESMAP report. Special thanks to Ms. Nyra Wallace-Crawford (ESMAP) for having provided excellent assistance with the organization of the Conference and also to Ms. Ananda Swaroop and Ms. Marjorie K. Araya for the editing and production of this report. The draft was reviewed by ESMAP and World Bank specialists and I am grateful for their comments. Nevertheless, any remaining errors or misinterpretations remain the responsibility of the author. ix 1. Outline of this Report This ESMAP report summarizes an expert ability to identify, quantify and, where possible, consultation on electricity system investment mitigate risks in power systems investments. planning and modeling held at World Bank in The report is organized as follows: Washington D.C. on June 27-28, 2006. The workshop brought together practitioners, · Introduction to the issues; analysts, regulators and software vendors to · Synopsis of the workshop and summary of discuss the changes in both the external discussions; environment affecting the use of simulation · Findings and recommendations; models for investment planning and the · Annexes: World Bank/ESMAP Workshop on advances in the art and science of such Electricity Investment Modeling and modeling over the past few years. In addition to Risk Mitigation; and the focus on generation investment modeling, · CD containing list of participants, papers and a key subject of interest at the workshop was the presentationsdeliveredattheexpertconsultation. 1 2. Introduction More than 15 years ago, the World Bank reported structure of the electric power system. Some in a series of papers that the planning environment countries have devolved decision authority for World Bank's client electric utilities was becoming generation planning to the point that no one is increasingly difficult. The papers cited the responsible for system expansion planning, now increasingly unstable nature of world commodity that the central planning function of the power markets combined with the unreliability of demand company or of the government itself has forecasts and uncertainty or even systematic bias shrunk. And in most countries, the reliance on which plagued cost estimates for new power plants IPPs for future expansion of generation and called for an approach to electricity system capacity, reduces the proscriptive power of a planning which could account explicitly for risk and single generation expansion plan. It is now uncertainty. Today, the situation for power system better understood that regulated utilities in planners is even more fraught with uncertainties decentralized systems, regardless of the about prices and costs. The continued volatility of structure or ownership of the system, need fuel prices has increased significantly the normal more guidance on future investment risks of projecting future investment economics. In opportunities. At the same time, the regulators in addition, the structures of the power systems these countries are charged with the responsibility themselves, formerly stable as vertically integrated to ensure that investments in new capacity are monopolies, now come in a variety of states of "prudent" and that ratepayers and plant owners reorganization, privatization, unbundling and are not left with excessive payment liabilities for market orientation. As a final blow to the older inappropriate investments. Thus, interest in certainties of power system investment planning, planning has re-emerged as an issue in many of the increasing power of independent regulators, the World Bank's Developing Member Countries many of whom were created during the past (DMC) utilities and regulators. 10 years, means that non-specialists are increasingly relied on to vet and to approve In an effort to discern some orderly investment highly technical plans based on complex procedures on an otherwise fluid reality, and to modeling activities. assure national authorities and regulators that the ratepayers and/or the taxpayers are making In the current electricity market environment, sound expenditures, utilities try to assess what the virtually all of the key parameters on which best generation plant investments might be for generation expansion plans are based, are in play. their systems. Historically, this has involved The planning supply function, once normally the constructing simulation models of the relevant responsibility of a state-owned enterprise, has, in power generation system, extrapolating future many instances, also fallen victim to changes in the demands, and calculating the most efficient 3 Risk Assessment Methods for Power Utility Planning path to meet that demand over a specified The Wien Automatic System Planning Package planning horizon. (WASP), developed at the International Atomic Energy Agency (IAEA), has emerged as the An efficient investment path has been understood most-used simulator program of its kind. WASP historically to mean a least-cost path, subject to has been distributed to more than 75 countries certain performance constraints. Acceptance of and has become the standard approach to this approach among planners has led to the investment planning in World Bank's developing development of optimization models which will member-countries and is still used in other calculate the generation mix and investment countries as well. schedule that can meet projected demand at an acceptable Loss of Load Probability (LOLP). This expert consultation workshop is aimed at exploring the strengths and weaknesses of the The accepted way to accommodate the shifting standard, examining alternative approaches, and investment cost, fuel price and plant performance assessing the usefulness of different simulation parameters has been through the construction of models to a world of increasing uncertainty about scenarios. Each scenario, usually a high, low and costs and prices. Newer approaches, some from medium, is presented with a description of the modern finance theory, challenge the validity of essential output elements ­ plant mix, generation the types of results produced by WASP and its cost, investment needs to meet the plan. Typically, associated models. The workshop has provided the medium scenario has been constructed to a forum for a vigorous exchange of views on represent the most likely set of outcomes. these matters. It was quickly understood that a plethora of What are the Issues which Need to least-cost expansion models, while perhaps be Addressed? desirable from a scientific viewpoint, would Planning investment strategies with least cost complicate tremendously the work of those who optimization models have generally involved must ultimately approve and fund the resulting answering the question: What is the least expansion plans. Banks and regulators, in expensive way to meet future demand while particular, need to be able to compare results maintaining a LOLP no greater than some from one country to another with some confidence specific target? Today, a host of additional issues that valid comparisons can be made. Developing and concerns have been added to the country utilities cannot be expected to pay for traditional formulation. "one-off" models, and such initiatives can strain training budgets, entrench a "model priesthood" Experience with a variety of approaches to and increase the costs of data-gathering and planning has conclusively demonstrated that the data input to the model. As a result, the World era of "one-size-fits-all" modeling is over and that Bank and other international institutions have the tools used for investment planning need to be favored an approach to system expansion and tailored to the system, and for which the planners investment simulation which makes use of need to have tools which are consistent with the internationally recognized tools which can be structure of that system: readily shifted from one situation to another, where people can be trained in a standard · Is it an unbundled market or a planned manner, and where the results are believed to vertically integrated enterprise? be accurate, computationally efficient · Is trading important in the system, and does and reproducible. trading use contracts or markets or both? 4 Introduction · Do the participants in the system understand · Mitigation methods must be devised which are the consequences of planning mistakes, and if consistent with market structure, financial so, who is responsible for mitigating the sophistication and system size; and impacts of these mistakes? · Smaller power systems should be able to benefit from better tools through the The environment which surrounds investment development of either direct or proxy planning today has changed dramatically measures which permit explicit and there are new participants and new consideration of risk. goals in the planning process, even in countries with state-owned monopoly systems. To address these four categories of issues and These new participants and their goals include: problems in electricity system investment modeling, the expert consultation workshop was · Ministry of finance ­ which will wish to reduce divided into the following substantive sessions: the impact of power sector financing on the public budget; Current Approaches to Electricity System · Private investors ­ who will seek to transfer as Investment Modeling: What Works and What much risk as possible to the public sector; Needs Fixing ­ subjects covered included an · Regulators ­ who will seek to bring a broader overview of current issues and problems in and more independent perspective to the electricity generation investment modeling, as well process; and as two presentations on the WASP model and · Other parties ­ who will seek to promote other current modeling approaches. environmental, technological and other Investment Decisions for Public and interests in the power sector. Regulated Generators ­ with presentations on To support the changing technical nature of public sector perspectives on investment planning, investment analysis and planning, and the regulator's viewpoint on modeling increasingly complex information flows must techniques and results. become endogenous to the overall planning activity. Not only do the diverse points of view of Alternative Formulations of the Investment an enlarged set of participants need to be Decision and Mitigating Measures ­ covering accommodated, but these information flows a critique of optimization approaches and an present a challenge to both the planners and the introduction to risk-return methodology for users of planning output, since newer, more investment planning, and a companion complex analytical models also tend to be less presentation on risk mitigation methods in transparent to a lay audience. portfolio optimization. Risk has increasingly become a central Case Studies and the Way Forward ­ to give element in electricity system investment some flavor of how new approaches and planning. This means that it must be dealt with methods might be put into use, there was a final systematically, comprehensively and appropriately, session of case studies focusing on given the tools available. implementation of new modeling techniques in system planning in the US and a case study of · Risks must be identified and their relationships the risk-return approach in Mexico for to other risks understood; Renewable Energy (RE) project assessment. 5 3. Synopsis of the Expert Consultation Workshop Current Approaches to Electricity System The Single-company Approach ­ WASP and Investment Modeling ­ What Works and Related Models What Needs Fixing The ideal of treating system planning as a The simulation and planning tools, commonly technical matter was honored largely, in fact, and used in World Bank's DMCs, came of age in an occasionally in the breach. In many countries, the era of vertical integration and state ownership planning process was constrained to yield specific of the electricity sector. With the political system results. In particular, after the oil price hikes of the exercising supervision over the power sector 70s, many countries consciously adopted fuel through a minister of energy or a similar senior diversification goals as a specific constraint on the official, there was no perceived need for results of the planning process.1 Even when some further oversight and intervention, especially by oil use was probably least-cost (e.g., Heavy Fuel an outside and independent regulator. Indeed, if Oil (HFO) in some oil exporters during the 80s any outside considerations regarding system and most of the 90s) the power system planners expansion and investment were put into play, it would exclude such possible outcomes a priori.2 was likely from either the World Bank and As was observed during this session, legitimate regional development banks or from equipment constraints on technology and fuel choice often suppliers and Architecture & Engineering (A&E) became confounded with efforts to "fiddle" the firms. In this environment, the planning function expansion plan, clearly not the intent of serious for both generation and transmission was seen system planners. as a largely technical matter. Structural issues in the electricity sector were generally limited to System planning models became more useful as whether to separate rural service from urban they migrated from mainframes and minis to electricity supply and the necessity or wisdom of desktop computers in the late 80s, allowing more forming separate regional distribution interaction with planners and better fine-tuning of companies. Generation and transmission were the assumptions and results. As a result of the seen as indissoluble elements of the national focus on a single approach and a long-term effort economic planning apparatus. to train local engineers in the use of the WASP 1This conscious process contributed to the expansion of coal-fired generation around the world in the 1980s and 1990s when other fuel cycles might have proven competitive. 2In many oil-exporting countries, a surplus of HFO was exported at low prices due to a dearth of upgrading capacity in regional refineries in South-East Asia and the Eastern Mediterranean. In contrast, US refiners invested billions in upgrading capacity to use HFO as a feedstock, effectively eliminating the fuel from the power market on a price basis alone. 7 Risk Assessment Methods for Power Utility Planning model, executives and planners became Valoragua, must be used as well. The canonical increasingly comfortable with the results of this problem for WASP takes the general form: process. In effect, WASP became the planning framework for scores of DMC utilities and its Min (investment costjk + operating costjk), subject to output was broadly respected as appropriately Demandk > constraint accurate and complete. LOLPk> constraint Fuel pricek = price path Traditional power system planning models, à la Technological parametersj = assumed or WASP, include limited stochastic information. This calculated values; where information deals mostly with the LOLP connected j and k represent technology and time period. with the expected reserves and plant availability of a particular system configuration. This formulation, though somewhat stylized, addresses the major questions of cost, fuel To obtain a probabilistic estimate of the LOLP choice, technology and system reliability. connected with hydro plants, another model, WASP identifies the least-cost expansion plan Figure 3.1:Wien Automatic System Planning Package (WASP) INPUT OUTPUT · Load Forecast · Existing System · Candidates · Constraints: ­ Reliability ­ Implementation ­ Fuel · Build Schedule ­ Generation · Generation ­ Emissions · Costs · Fuel Consumptions · Emissions Box 3.1:The Evolution ofWASP as a Planning and SimulationTool It has been well understood from the start that system planning models represent reality, but do not reproduce reality. Due to the abstraction from certain details in the power system, it is always easy to find problems with simulation models. In response to peer reviews and critiques of various planning approaches, the model architects and software developers have put together packages which respond to the demands of their marketplaces. In the 80s, WASP was modified so that it could run on desktop computers; in the 90s, emissions of major pollutants were added to the WASP output set (as Decades, and later WASP IV). In response to criticism about how hydro plants were handled in the WASP framework, Valoragua was added to the suite. And finally, as systems have restructured and the finer resolution of plant dispatch has come to the fore as a concern, WASP has been combined with dispatch and IPP-oriented models, such as GTMax, to improve the resolution of results and identify generation-transmission interactions and bottlenecks. 8 Synopsis of the Expert Consultation Workshop from the point of view of a single decision · Technology concerns; maker (one company). · Diversification (or its absence);6 · Other operational factors; and Figure 3.1 shows the logical sequence of WASP · Market risk ­ consisting of both supply-side and data and output flows:3 demand-side risks. A presentation early in the workshop by Argonne For each of these issues a separate scenario National Laboratory specialists on WASP and other must be constructed in the context of WASP. The system models, demonstrated the evolution in the interaction of different risks (i.e., correlation) is capabilities of the WASP model and the ability to not addressed by most scenario methods, augment its output with modules which improve although a skilled analyst should be able to the ability of the model to project valid results construct reasonably integrated alternatives. for hydro plants (Valoragua), IPP and However, even the most skilled scenarist might transmission (GTMax), and market Center for have difficulty answering an increasingly Energy, Environmental and Economics System common question today: "How likely is that Analysis (EMCAS). series of events you just described to us?" Probability of risks and interaction with In systems which still conform reasonably close to investment planning is discussed in the next the one-company model, WASP can develop a section. Suffice it to say, without systematic and least-cost expansion plan as a basis for planning quantitative risk identification and and investment. Specific plant investments can be measurement, assigning probabilities to inserted or removed to assess the impact on scenarios is difficult at best. overall generation cost and reliability. In partially restructured systems, especially those making To remedy these perceived limitations while still heavy use of IPPs within a single-buyer type of using WASP as a reference scenario, it is model, WASP can provide a reference expansion possible to combine WASP output with the other plan against which future investments may be models ­ Valoragua, GTMax and EMCAS. This compared. Such an approach is still used in South approach allows some of the risks listed above Korea and Poland. to find explicit representation in the overall modeling framework. In particular, using this The Argonne presenters identified several combination of models will help to explicitly categories of risks or resolutions which are not account for plant dispatchability, transmission endogenous to WASP including: constraints, contracts for power delivery and diverse corporate goals. However, the approach · Fuel prices;4 will still require a scenario approach to grapple · Plant dispatchability and load concurrence; with risks posed by fuel prices, construction · Construction cost;5 costs and demand uncertainties. Specific · Environmental standards; modeling of risks and devising mitigation 3A. Gritsevskyi, The Needs for International Comparable (sic) Energy and Environmental Statistics, International Atomic Energy Agency (IAEA), 2006. 4Ryan H. Wiser, Lawrence Berkeley National Laboratory, The Value of Renewable Energy as a Hedge Against Fuel Price Risk, 2004; and Mark Bollinger, Ryan Wiser, and William Golove, Accounting for Fuel Price Risk, LBNL-53587, Lawrence Berkeley Lab, 2003. 5See Merrow, Edward W. and Ralph F. Shangraw Jr, Understanding the Costs and Schedules of World Bank-supported Hydroelectric Projects, World Bank, Washington, 1990; and Bacon, Robert and John Besant-Jones, Estimating Construction Costs and Schedules, World Bank, Washington, 1996. 6Crousillat, Enrique, and Sprios Martzoukos, Decision Making Under Uncertainty: An Option Valuation Approach to Power Planning, Washington, 1991; and Shimon Awerbuch, The Role of Wind Generation in Enhancing Scotland's Energy Diversity and Security, ECN, Netherlands, 2006. 9 Risk Assessment Methods for Power Utility Planning measures against these risks will still be external compare the relative riskiness of different to such models. The next subsection includes a investment alternatives. However, to call for such discussion of incorporating risk explicitly into an risk-based assessments neither identifies the risk, optimization model environment. nor its potential impact on an investment, nor a suitable mitigating strategy. Systematic Treatment of Risk Identification and Risk Mitigation ­ an Initial Approach A first step in the assessment of the relative riskiness of various investment alternatives is an Identification, measurement and quantification of understanding of two key attributes of risk: impact risk are essential elements in formulating a (or importance) and controllability. The importance strategy to incorporate both the consideration of of a given risk element in system expansion varies risk and suitable mitigating measures in an with the impact of that element on the rate of investment plan. Many energy companies use a return and the variability of that return. Fuel prices P5, P50 and P90 set of probabilities when assessing are very important risks in a thermal-dominated the attractiveness of a particular investment. By system, but far less so in one that relies more on bracketing the potential returns, the company can nuclear, hydro or wind. Conversely, plant Box 3.2: Fitting Reality into a Four-part Grid: How toWork with the Risk Matrix One way to focus on analysis and assessment of investment risks in electric power systems, is to perform a type of analytical triage. In this formulation, issues or individual risks are categorized by (i) how important they are to the success of the project; and (ii) whether the party making the analysis has some degree of control over that risk. The most important elements of this exercise are the identification of the risks and the judgment about which ones are vital to the success of the project. Risks which are important (they can make or break the project) and which are controllable (the investor can initiate activities on his own, which will mitigate or reduce a potential risk) are located in Quadrant I and are considered strategic risks. These should be tended to first, since they are controllable to some degree and not tending to them can be damaging to the overall project. The next priority focus is on those risks which are important, but not easily controlled by the investor. Somehow, the investors must find ways to reduce the unknown impacts on the project. Typically, for important risks, this can involve some type of reformulation for the overall investment project so as to circumvent that risk, mitigate portions of the risk, or devise loss-mitigation instruments to insure against the impacts of that risk. Even if nothing is or can be done to mitigate the contingency risks, it is important to understand what these are so as to reduce the potential adverse impacts on the overall project. Risks which are neither important nor controllable (Quadrant III) are best identified, so that time and energy is wasted on mitigation strategies. Finally, Quadrant IV identifies risks which can be controlled, but with little overall impact on the project. Some of these risks may fall in the general heading of "nice to have, but not entirely necessary" and can be mitigated as time and resources permit. 10 Synopsis of the Expert Consultation Workshop dispatchability may represent a significant · Tougher environmental standards; economic risk in a wind or run-of-river hydro · Technology concerns; system, but a lesser one in a coal or Combined · Diversification (or its absence); and Cycle Gas Turbine (CCGT) system. · Operational factors. The second key to understanding risk is to look at Each of these factors has a distinct locus on the the controllability of that risk. Some risks can be risk and controllability axes given the mitigated by actions of the plant owner or system characteristics of the specific risk element as well regulator, while other risks fall beyond a as the characteristics of the system where this risk reasonable attempt at mitigation. Figure 3.2: Importance and Controllability of Power System Risks IV High I Strategy Strategy Refinement Tend to these as time permits Resolve these First High Ability to High Ability to Mitigate Minor Risks Mitigate Important Risks Control Low High Importance Low Priority Issues Figure out how to circumvent, reduce or reformulate these Low Ability to Low Ability to Mitigate Mitigate Minor Risks Important Risks III II Non-strategic Low Contingency Figure 3.2 shows one approach to assessing the needs to be mitigated. Figures 3.3 and 3.4 importance and controllability of risk in the context indicate how each of these factors might look in of power system generation expansion. different types of systems. System planners and regulators will want to know In a nuclear-dominated system (Figure 3.3), fuel where different risk elements stand in the schema price risks represent, if anything, an opportunity of a particular system expansion exercise. for the plant owners. Generally, they are considered relatively unimportant, and, hence, Consider the following seven risk elements: non-strategic. However, technology and construction risks, historically two of the biggest · Fuel prices; concerns of the nuclear industry, are both highly · Plant dispatchability; strategic and relatively controllable.7 · Construction cost; 7Note that dispatchability, long a problematic feature of nuclear power, is no longer considered important, as operational improvements in plant management have raised the availability factors for such plants dramatically in the past 10-15 years. 11 Risk Assessment Methods for Power Utility Planning Any system which is dominated by a single The correlation of risks may also be significant. technology is definitionally short on diversity in For the nuclear industry, technological and its technologies and fuel cycles. Whether this construction risks have traditionally moved represents a controllable or uncontrollable risk together. This means that risk with respect may depend significantly on specific choices by to one attribute may often be exacerbated by the power suppliers vis-à-vis other technologies. risk in the other.8 Figure3.3:ImportanceandControllabilityofPowerSystemRisks­Nuclear-dominatedSystem IV High I Strategy Strategy Refinement Construction Costs Operation & Technology Dispatchability ? Control Diversification Low High Importance ? Environmental Fuel Prices Legislation III II Non-strategic Low Contingency Figure 3.4: Importance and Controllability of Power System Risks ­ Gas/CCGT-dominated System IV High I Strategy Strategy Refinement Construction Costs Fuel Prices Operation & ? Dispatchability Control Diversification Low High Importance ? Technology Environmental Legislation III II Non-strategic Low Contingency 8For the nuclear industry, both construction cost and technological risk are commonly positively correlated with regulatory risk as well. 12 Synopsis of the Expert Consultation Workshop A far different risk profile emerges in a ROR system. Technology, largely proven and not gas-dominated CCGT system: subject to significant innovation, is non-strategic. Fuel prices, which could impact a decision to invest Unlike the nuclear system, fuel prices represent a in better operational control of the peak period strategic risk ­ how controllable are they? In many output capability, assume a greater importance systems today, the very factors which render than for a nuclear system, given the trade-off construction costs and technology relatively between better water management and the use of unimportant in CCGT plants ­ a proven combustionturbinestomeetpeakdemand.Indeed, technology largely factory-built ­ may well be the fuel price and the operational risks are correlated with both fuel price and legislative risk. correlated to some degree. For example, improved Certainly in many countries, including the US, controllabilityofpeakperiodplantavailabilityfrom many CCGT plants were built as a response to ROR hydro will also make fuel costs more looming limitations on various pollutants from controllable. Success in this theater comes from other fuel cycles, especially coal. converting a contingency to a strategy element. Figure3.5:ImportanceandControllabilityofPowerSystemRisks­Hydro-dominatedSystem IV High I Strategy Strategy Refinement Construction Costs ? Control Diversification Low High Importance Fuel Prices ? Technology Operation & Dispatchability Environmental Legislation III II Non-strategic Low Contingency Finally, a well-run hydro system, dominated by The nature of risks will change over time and run-of-river (RoR) plants, generates yet a third planning must recognize that as well. Some risks risk profile: which are both important and controllable for certain types of systems (for example, those For almost any hydro-dominated system, associated with hydro in Quebec or gas supply construction costs are extremely important and in Indonesia) may not be as controllable for largely controllable. However, operational risk, other types of systems. Even within one system, especially the issue of the plant's availability during the impacts of investments and policies may peak periods, is a matter of high importance to the change the nature and controllability of certain planners, but relatively low controllability in a pure risks over time. For example, fuel price risk in 13 Risk Assessment Methods for Power Utility Planning Mexico has grown over time as the roles of · Treatment of IPPs; and gas-fired CCGT plants and imported gas have · Plant dispatch bidding. assumed greater roles in system capacity These models will clearly reflect the interests of the and output. people that pay for the results. Thus, Aurora, the How much Consideration of Risk has been product of yet another electricity modeling firm, is Incorporated into a Unified Contemporary oriented towards identifying optimal generation Modeling Framework? portfolios in a relatively deregulated US environment. The model addresses explicitly The various modules associated with WASP, Energy certain risk issues raised above, specifically fuel and Power Evaluation Program (ENPEP) for prices and diversification value. Since IPP bidding demand forecast and fuel supply, GTMax for plant is an important part of the model's output, the dispatch and IPP contract simulation, Valoragua for integrated dispatchability module helps address an hydro project simulation, all plug into the input or important risk element. output sides of WASP. EMCAS is a downstream Siemens uses a different approach, and has specific model which follows a decision-theoretic modulestoaddressvarioussegmentsofthebusiness framework for modeling the decisions of individual ­ resource planning, dispatch (Figure 3.6). agents. Specific consideration of various risk factors can then be attributed to different "agents" Key risks are addressed parametrically. For along with different risk mitigation strategies. example, models using the ProSym engine (Figure 3.6) for plant dispatch simulation, allow WASP can now incorporate some of the risks "what-if?" scenario analysis of selected variations discussed above for some generation technologies. in key parameters. One module, "Risks and For example, with the inclusion of GTMax, WASP Markets," addresses risk in a systematic manner. can now address the dispatchability issue. Along with Valoragua, the estimated firm capacity of Generally, model developers will only incorporate ROR hydro can be estimated to some degree, and endogenize feature sets which users are better endogenizing technology and operational willing to pay for. As a result, some of the risks risks in the model's framework. Most of the other noted above, but especially the correlation of risk categories must still be included through the different risks, remain outside the explicit use of scenarios with varying parameterizations. simulation engines discussed above. Other modeling systems have been developed as During the workshop, modelers from GE Power well. Both General Electric (GE) Power Systems Systems presented an approach which provides and Siemens have suites of models for utility explicit consideration of market structure, fuel system planning and investment analysis. These policy, environmental policies and fuel price models contain feature sets which correspond to volatility. To cope with a Multiseller Multibuyer the needs of the market, as does WASP. Naturally, Market Structure (MSMB), the GE Power Systems the emphasis of a given approach will vary from approach permits three different modes of pricing one company to another as regards certain key behavior in the power market: (i) price-takers; (ii) aspects of electricity investment system cost-based pricing; and (iii) market share pricing. simulation, including: Since the GE models were developed by a vendor · Endogenous plant dispatch; of several of the non-hydro technologies currently · Iterative dispatch and generation planning; available for generation, their approach 14 Synopsis of the Expert Consultation Workshop Figure 3.6: Siemen's Model for Risk Assessment System Status Static Data ­ Committed Status ­ Unit Location ­ Production Level ­ Unit Cost Data ­ Long-term Outage Schedule Inputs ­ Load History EMS Feed ­ Unit Commitment ­ Fuel Cost ­ Bid Strategy ­ Bid Library User Inputs ­ Load Update ­ Update to Fuel Cost ProSym ­ New Sales/Purchases ­ Changes to Bid Strategy ­ Update to Unit Commitment or Heat Rate Scheduling ­ Short-term System Dispatch Schedule Unit dispatch schedule ­ What if Analysis XML to feed schedule ­ Avoided costs communication system analysis to support trading declsions Source: New Energy Associates, a Siemens Energy Company, Product Descriptions (http://www.newenergyassoc.com/products/promod), 2006. specifically includes technology. These the GE modeling system is also able to produce a technological parameters include: "one-company" integrated plan, which can be used as a reference case for the market-oriented · Adoption curves for new technologies; simulation. The GE model suite stops short of an · Explicit inclusion of non-dispatchable or explicit risk-return frontier9 using portfolio theory, intermittent technologies; which will be discussed in the next subsection. · Demand-side programs; and · Transmission investments and trade-offs vis-à- In addition to the commercially available models, vis generation resources. some companies and individuals have developed a variety of simulation approaches to power Acknowledging the appeal of a reference case, system expansion. As a rule, these models are Box 3.3: Portfolio Theory: What it is and What it can Do Portfolio theory is not a system modeling approach. Rather, it is part of the tool kit of a more comprehensive approach to investment modeling and simulation. The main points of using portfolio theory are to (i) Redefine the efficient frontier of investment choices from least-cost to optimal risk-return; (ii) Endogenize the specific risk elements which need to be considered; and (iii) Identify correlated risks so as to reduce the probability of foreseeable negative outcomes.10 9Awerbuch, op cit , has focused attention on the efficient risk frontier. Internal to that concept is not only the variability and risk associated with individual parameters, but also the correlations between and among different risk factors. 10As a generic tool, the portfolio optimization approach can be used to focus on most types of quantifiable risks. However, in practical terms, the important risks would need to be identified and measured beforehand. 15 Risk Assessment Methods for Power Utility Planning developed with feature sets which correspond to (Comisión Federal Electricidad-CFE), the Mexican the perceived needs of the situation and the state electricity company (and summarized in preferences of the architect. For example, one box 3.4), and the other by the head of Mexico's model which was used for the economic electricity regulator, Comisión Reguladora de evaluation of wind in Mexico also contains gas Energía (CRE), brought focus to the matters of production, pipeline and liquefaction modules, public decision-making regarding utility as well as power system expansion modules, investments. The regulator noted that regulators in since much of the model's structure was World Bank's DMCs still tend, for now, to come developed primarily to provide decision support from the ranks of utility industry professionals. This for private investments in "midstream" gas to means that they are generally conversant with the power or gas to liquids.11 The electricity system main concepts contained in an optimal planning simulation module of this model was intended to approach and believe themselves to be proficient support Power Purchase Agreement (PPA) in interpreting the results. negotiations, and thus endogenizes several of the risk elements discussed above. Since the The flip side of this confidence in the planning model results must be expressed for a specific process is the potential for getting attached to client in terms of probabilities, the entire model approaches which may be superseded by is run using probabilistic simulation of several of technological progress or advances in the key parameters, including fuel prices, plant understanding of markets.12 As technical operations and conversion efficiency. capabilities evolve, understanding of dynamic Construction costs and operational factors for markets improves and the remit of regulators non-target plants are treated parametrically. changes, the decision-support environment will, of necessity, evolve as well. Tools which were once Investment Decisions for Public and both useful and sufficient for decision-support, Regulated Generators may fall victim to a changing environment and advancing capabilities. There are significant The second major topic of the workshop trade-offs when this occurs. Figure 3.7 illustrates concerned issues of key importance to utility some of these issues. regulators and publicly owned and vertically integrated utility companies. The focus of this Analysis and assessment techniques which can be session was: (i) what information do public sector performed on a handheld calculator, such as the regulators and company executives need to make next plant or production cost approaches, have optimal decisions; (ii) how do their limited output flexibility. However, they are highly decision-support needs differ, if at all, from private transparent and their assumptions, methods and sector entities; and (iii) what are the key results, can be explained to a lay regulatory elements of credibility in decision-support for audience quite readily. As with most regulatory bodies? decision-support systems, there are trade-offs. Simple analytical approaches are relatively Two presentations, one by the Chief of easy to explain to regulators and public Planning for Federal Electricity Commission funding authorities. 11ESMAP, Mexico: technical assistance for Long-term Program of Renewable Energy Development, Washington, D.C., February 2006, pp.85-86. This model also produced the incremental cost results for the GEF Project Appraisal, published in May 2006. 12See the Workshop paper by A. Peraza, Regulatory Tools: Promoting Renewables. 16 Synopsis of the Expert Consultation Workshop More complex tools, relying on sophisticated particular, it was noted that an increased focus on models and computational techniques, are by RE should probably have some legislative backing, definition less transparent, even as they become and not simply be a regulatory initiative. Such an more useful to specialists. Regulators will (rightly) approach would resolve at least some of the note that the analytical capabilities should not run problems which complicate investment planning ahead of the ability to explain important public under a legal mandate to obtain the least-cost sector investment choices even as more private supply mix (see Box 3.4 on Dr. Soler's comments investment brings the need for better identification in this regard). and mitigation of risk factors. Further, it was shown that the changing structure A confluence of changing industry structure, new of the utility industry, even in systems where a analytical methods and new generation vertically integrated entity still maintains nominal technologies are creating a concerted challenge to control over investment decisions, creates the least-cost method. As Dr. Peraza noted, the significant pressure for accommodating additional increased interest in renewable and intermittent points of view in the regulatory and planning energy sources in World Bank's DMCs not only processes. One of the commentators noted that creates a decision conundrum for investors and this tension was healthy, since modelers may be regulators, but also for those providing the more aware of the strengths of their approach decision-support. The discussion at the workshop rather than that of the weaknesses. A public airing supported the idea that multiple methods of of the issues might represent progress in forcing analysis, along the lines of extensions to WASP or both users of the model output, as well as the the GE Power Systems family of simulation tools, modelers themselves to think hard about what would be needed to keep up with the decision kind of information and presentation improves needs of regulators and public entities. both comprehensibility and transparency, the beginning of any resolution to what seems the A key issue noted by Dr. Soler or CFE13 is the most difficult trade-off identified in the workshop. continuingobligationofstate-ownedcompaniesto abidebyleast-costprocurementregulations,typical Consensus Viewpoints on Public Sector Issues of state enterprise regulation. It is not the intention of in Modeling the workshop to devise methods of circumventing An emerging consensus from the workshop was this wholly justified constraint on the use of the public that the output of a least-cost optimization was money;rathertheworkshopdiscussantslookedat sufficiently understood, and both its strengths and howthisfundamentalprinciplemightbeaugmented weaknesses fairly well-known to both users and to byadditionalanalyticalmethods. regulators, that continued use of a WASP type of approach had significant ongoing value. Twootherpresentationsinthissessionbyanalysts Additional decision-support for the public sector, from World Bank and the Pennsylvania Utility especially where structural change in the utility Regulator (PJM)14 added important system is ongoing, would then involve embellishments to Dr. Peraza's material. In 13 See the workshop paper, Investment Planning in the Electricity Sector: Differences from Private Electricity Companies. 14 See workshop papers by Barua, Regulators' Viewpoint on Modeling: How Does a New Approach/Tool Pass Regulatory Muster; and J. Besant-Jones, What is Required for New Planning Techniques to Become Acceptable Within the New Regulatory Framework. 17 Risk Assessment Methods for Power Utility Planning Box3.4:WhataretheDecision-supportNeedsofaPublicly-ownedIntegratedUtility? A key interest in putting together this workshop was to provide a forum for model developers and model users to reflect on their diverse viewpoints and needs. In a presentation which followed the Argonne and GE papers, the head of planning for Mexico's CFE, the state-owned electricity supplier, provided a list of key decision-support needs for his utility and generalized his observations to other similar organizations. From the point of view of an integrated state-owned utility, the economic analysis of an investment plan is considered just as important as its financial implications. This observation has direct relevance to the choice of modeling instruments, as it provides a continuing role for a "reference" investment plan that keys to social and economic opportunity costs. Another key difference, one that does not affect the choice of models, is the discount rate ­ generally higher for private companies ­ which reflects the greater risk associated with specific projects and fuel cycles. And unlike a private generator, CFE must explicitly account for the costs to society of unserved energy. These considerations, and the need for a paper (or electron) trail that can be vetted by its regulator, CRE, argue for the continued use of a WASP type of model, one that produces a reference least-cost generation expansion plan. Other considerations, especially dispatch and fuel choice, will remain subsidiary elements of the forecast. As long as the current legal and regulatory environment remains in force the construction of a risk-return frontier provides background but not dispositive calculations, which must remain with the least-cost generation plan according to the Government of Mexico's overall fuel and demand forecasts. This last point is a key one in the case of state-owned vertically integrated utilities. CRE is not free to generate its own fuel price forecasts and must use the same planning figures for prices, costs, and demand as other state entities. In such a structure, fuel price risk, technology diversification, and other parameters must be seen as scenarios or excursions on the reference case. The explicit inclusion of such parameterization of a reference case optimization was discussed on the second day of the workshop and will be included as an activity in the Global Environment Facility's (GEF) current wind project in Mexico. augmenting the least-cost output with additional which emerges from the optimization process analytical modeling activities to address the key as a recommended future generation mix; risk factors. In addition to the dispatch, · Special attention should be paid to the transmission and water submodels, participants confluence of generation with load for concluded that: evaluation of RE resources (i.e., improved temporal resolution); and · A specific measure of the riskiness should be · Spatial analysis of load confluence should be constructed for each of the least-cost scenarios made to determine whether risk in renewable 18 Synopsis of the Expert Consultation Workshop Figure 3.7: ModelingTrade-offs: Capabilities vsTransparency (Longer Bars are Better) Model Type Capabilities Complexity Transparency Addressing Risk Next Plant Production Cost Energy Balance Least-cost Model LCP + Dispatch LCP + Dispatch + Transmission Note: The use of an integrated risk assessment tool, such as risk-return modeling, will generally increase the ability to assess risk, while adding complexity and reducing transparency. Figure3.8:CapabilitiesandNeedsinPowerSystemInvestmentPlanning Complex Significant Role for LCP + Dispatch + TA and Consulting Transmission High LCP + Dispatch Capabilities Least-cost Planning Basic Sophisticated Energy Balance Production Cost Models Next Plant Heavy Use of Needs Dispatch Modeling Basic generation can be reduced from use of Resolving these trade-offs will become the key multiple locations for generation. activities of any effort to improve the decision-support for energy investments. As with For each of these recommendations ­ more the development phase of any new technology, models, more statistical analysis, more advanced financial techniques ­ the transparency and there will need to be a variety of tests, comprehensibility of the results for regulators will including: fall. Figure 3.7 illustrates some of the trade-offs between comprehensiveness and · Comparisons of optimal solutions from comprehensibility and transparency: different models; 19 Risk Assessment Methods for Power Utility Planning · Comparisons of reference solutions with · Policy-focused; market outcomes; and · Suited to current and expected future · Assessments of multi-model vs single model market structures; approaches (also, multi-model · Manageable by its practitioners; family comparisons). · Transparent; and · Participatory. Alternative Formulations of Investment Decisions and Risk-mitigating Measures Some of these criteria have been discussed above, especially transparency, market structure The third session of the workshop focused on compatibility and policy focus. The present specific suggestions for combining analytical discussion will then focus largely on how to approaches to endogenize the treatment of risk in incorporate risk identification and mitigation and the investment planning process. This session how to make sure that the planning process introduced a discussion of the particular measures incorporates the views of important stakeholders in of risk identification and mitigation which might be the system. critical elements in a new, hybrid investment modeling approach. If using point estimates of costs by themselves is no longer reasonable from a best practices Drawing upon the consensus from earlier viewpoint,15 then explicit incorporation of risk presentationsandaddingyetanotherdimensionto assessment must be ushered into both the the trade-offs discussed above, a suite of modeling process and the regulatory oversight. recommended planning techniques will need to be: The first essential step is to identify the types of Figure 3.9: Risk and"Return"forThree-technology US Generating Portfolio Assumed Cost for Riskless Renewable: US$12/kWh 0.18 M: Optimum Coal-Gas A Mix (72%-28%) A: 60% GAS 0.16 cent) K: 6% Renewables Coal-Gas Capacity Mix US M K 0.14 65%-35% per Coal-Gas Generation Mix 77%-23% (kWh 0.12 B: 100% COAL H H: 50% Renewables + 0.10 50% M ``Return" F 100% Renewable 0.08 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 RISK: Portfolio Standard Deviation Source: S. Awerbuch, "Getting it Right: The Real Cost Impacts of a Renewables Portfolio Standard." Public Utilities Fortnightly February 15, 2000. 15See the workshop paper by S. Awerbuch, Generation Investment Planning and Modeling: A Finance Theory Perspective. 20 Synopsis of the Expert Consultation Workshop risks which are relevant. Returning to the menu of Any generation paths or "portfolios" in the risks listed earlier in this report: language of finance theory, to the right of the line, are inefficient in that there exists another · Fuel prices; portfolio, which, holding either risk or return · Plant dispatchability and load concurrence; constant, will yield better results. Generation paths · Construction cost; to the left of the line are infeasible. Risk mitigation · Environmental standards; measures should aim at bringing the feasible · Technology concerns; generation paths as close to the efficient · Diversification (or its absence);16 and frontier as possible.18 · Market and demand risks. As was discussed above and in Chapter 4, These risks may be either random and, therefore, whether the risks listed above are systematic or subject to mitigation through diversification or random depends in part on the particular systematic method and, therefore, not readily generation technology. Technological risk is remediable through diversification.17 Random risks essentially random (and low) for CCGT and hydro include the generation of electricity from a plants, but is a major factor in nuclear, and run-of-river hydro plant or a single wind perhaps wind technologies. Construction cost risk generator. Systematic risks include factors which is a correlated risk factor (with technology risk) in can make a generation portfolio co-vary with a nuclear power plants, but not in hydro plants. Fuel market portfolio. An ideal generation portfolio cost risk is a systematic risk, and a powerful one from a risk standpoint is one whose systematic risk (high co-variance with market portfolio) for gas is as low as is practicable for a given return and oil plants, but a weaker one for coal plants (measured as the costs of electricity output). and still weaker for nuclear plants. It is the high risk associated with gas and oil generation For each return on the generation portfolio there technologies which has motivated much of the is an associated systematic risk, and an efficient current investigation of efficient portfolios in portfolio from the financial standpoint is one in generation investment. which there is no excess systematic risk. In other words, there is a risk-return frontier which defines A first step in the integration of risk into a set of generation paths, each providing the best investment planning for utilities is the identification ratio of risk-to-return available. Other generation of risk. This means categorizing ­ random or paths will not be on the frontier, either because systematic ­ and then determining how important they are too risky relative to the return or the various types of risks may prove to be for returns are too low relative to the risk. Figure 3.9 different generation technologies.19 Once the shows the efficient frontier for several different risks to be included in the analysis have been generation paths. identified, it will become necessary to measure 16Crousillat, Enrique, and Sprios Martzoukos, Decision-making Under Uncertainty: An Option Valuation Approach to Power Planning, Washington, 1991; and Shimon Awerbuch, The Role of Wind Generation in Enhancing Scotland's Energy Diversity and Security, ECN, Netherlands, 2006. 17This section is a highly simplified and stylistic representation of financial portfolio theory. The reader is referred to standard graduate business school financial management texts. For an application to electricity portfolios see S. Awerbuch and Martin Berger, Applying Portfolio Theory to EU Electricity Planning and Policy-making, International Energy Agency (IEA) 2003. 18Long-term mitigation measures would seek to move the efficient frontier to the left, reducing the risk of each portfolio or generation path. 19This inevitably brings up the matter of how far one should go in the measurement of second-order impacts. It was noted at the conference that since different models will treat second-order impacts with varying levels of thoroughness, the results provided by different approaches can be graded in their applicability according to how the highest priority second-level impacts are handled. 21 Risk Assessment Methods for Power Utility Planning them. This step is likely to be more difficult than it In a nutshell, this summary article links fuel may seem and is described in another document price risk, regulatory and policy risk and project from the current project, New Approaches random risks to a recommendation which to Electricity Investment Simulation and diversified efficient portfolios, combined with Assessment ­ A Proposed Way Forward (World effective and predictable regulation, remain the Bank, unpublished, 2006). only real methods of reducing systematic risk in a generation portfolio. The workshop addressed Finally, the appropriate measures of risk will need the question of how to operationalize and to be incorporated into the normal investment internalize risk mitigation analysis in analytical planning project assessment for investors and modeling of generation investments once the regulators. This need to incorporate risk into the relevant information has been identified and investment framework was featured in the recent World Energy Council report gathered. This effort was beyond the scope of (summarized in Box 3.5), and provides a the workshop, except in a general sense, and concise summary of the array of risk issues will comprise a subsequent ESMAP-funded facing power sector investors. activity in this area. Box 3.5: How does theWorld Energy Council see the Role of Risk Analysis in Power Generation Investments? In an article published in the second issue of The World Energy Book from the World Energy Council, two senior NERA (an economic consulting firm) economists examined the various uncertainties confronted by potential North American power generators while making investment decisions. The authors noted that, in addition to the commodity and regulatory risks which have been affecting the industry for years, North America is also facing uncertainty over environmental policy and the implicationsitwillhaveforrecapitalizationofthegenerationsector.Inaddition, recentexperience has highlighted the risks posed by the inherent unpredictability of fuel prices over the normal lifetime of generation assets, which increases the business risk of generation projects. While the authors acknowledge that the issues confronting the industry are challenging, especially in light of the greater levels of uncertainty, the paper offers a list of concrete steps which regulators, policymakers, and companies can take to minimize the cost and risk in future power supply choices. The recommendations can be summarized as: · Regulatory and policy stability and certainty in the rules of the game are core criteria in investment decisions. It is essential that the regulatory and policy framework is well developed, consistent, and predictable in order to remove risk (and cost) from the industry and from consumers; · Companies should realize that their operating future is inherently unpredictable. Thus, companies must choose their supply options based on portfolios which are robust and can withstand the outcomes which are guaranteed to be different than those forecast by even the best prognosticators; and · Finally, regulators, policy makers, producers, utilities and consumers will all benefit by re-embracing the development of robust and competitive markets for wholesale power. In conjunction with greater regulatory and policy certainty, trading provides tools to quantify future risk and to hedge its effects, thus reducing a primary investment disincentive. Source: Mike King and Dr. Michael Rosenzweig, The World Energy Book, World Energy Council, September 1, 2006. 22 Synopsis of the Expert Consultation Workshop Nevertheless, it is possible to provide a general dispatch modeling. This procedure, similar to outlineofcautionsregardingtheendogenizationof the one used in some of the candidate models, risk measures in generation investment models. can result in a wide array of potential These cautions are intended primarily to guide generation paths, which can then be analyzed future efforts to incorporate risk into pathways which according to their risk-and-return efficiencies. meet the guidelines listed in this chapter under the title "AlternativeFormulationsofInvestment In theory, it is possible to provide probability DecisionsandRisk-mitigatingMeasures," especially ranges for variables and for the pricing those which require interface with regulators and parameters. This is probably feasible in the context other stakeholders ­ transparency, policy-focused, of a "home-made" model for some countries. participatory.Thefollowingsubsectionindicates However, it is highly unlikely that this could work in some of the types of risks which can be internalized the context of a generally disseminated initiative in andwhatissuesmodelersmightfaceinmaintaining investment simulation and evaluation. accurate,transparentandclearmodeling Computational and data complexity would make proceduresandmethods. the results very difficult to explain beyond the "black box level of reporting." This would violate How much Consideration of Risk can be the transparency requirement and could be a fatal Incorporated into a Coherent flaw to widespread regulatory acceptance of such Modeling Framework? an approach. Technical Considerations: One of the key The final section of this report lists some of the factors facing any modeler is the question of what specific data activities which will be needed to to exclude. For example, it is easy to talk about a make this integration of financial and optimization WASP model which has endogenous Monte Carlo models a reality. However, the most important implementation of fuel price risks, technology information activities will include: risks, construction cost and timing risks, and the like. One must question, though, what such an · Country-specific computation of systematic risks increase in computation requirements would add wherever possible and proxy measures of such to the model and its associated framework in risks where measurement is not possible; complication, computing time, data needs, and · Measurement of random and systematic risks understandability of the results. Some models can associated with renewable technologies and handle add-ons better than others and the identification of potential mitigating tractability of the resulting modeling system is likely measures; and to vary from one product to another. · A systematic assessment of the importance and controllability of both systematic and random One approach to this issue is to make specific risks in RE systems, including improved risk endogenization a second stage analysis, to transmission investments, spatial separation of be undertaken once a family of least-cost wind units, short-term storage at hydro plants generation paths have been computed from the and improved dispatch response to transients in interaction of generation, transmission and renewable supplies.20 20The workshop was fortunate to have participation from 3Tier Group, who provided the paper Risk Assessment in Renewable Energy Projects. Much of the discussion of integration of risk identification and mitigation is based on that paper and the discussion following its presentation. 23 Risk Assessment Methods for Power Utility Planning Regulatory Acceptability: Regulators, even if towards private generation firms rather than they possess significant industry experience, will publicly-owned ones. A systematic assessment of tend to find more complex model outputs difficult different models and their suitability to different to assess and approve on the same basis as more market and ownership structures should be specific and focused models. It is generally better undertaken as a part of the next phase of to explain to a regulator what types of information this activity.21 a particular model cannot provide, than to claim that it can provide almost everything, but that Consistency with approaches used in other "it might take some time to go through an sectors: Some regulators, especially those in explanation." Beyond a certain point, the results multi-sector settings (gas, refined oil, telecoms), of some models, especially when computing may prefer consistent approaches across with probabilities parameters which are sectors, so that regulated rates of return in gas correlated to varying degrees, can become can be compared to the returns of other difficult to explain intuitively. Subsequent efforts regulated industries. This approach often leads in this area will have to be worked out with to the adoption of specific "home-made" regulators and other members of the public to simulation models. assess how more complex methods can be Sophistication of Public Sector Enterprises: In framed in public fora to provide acceptable and some countries, the state-owned utility may be understandable results. constrained by reasons of budget, staffing, Other key issues for regulators which relate to the procurement procedures or lack of interest in criteria appearing in this chapter under the title pursuing new approaches. In other cases, the "Alternative Formulations of Investment Decisions State-owned Electricity Generating Companies and Risk-mitigating Measures," and can affect the (SOE) may be willing to devote the time and choice of modeling suites include: internal resources necessary to take new approaches, but may be constrained by lack of Public sector goals: Demand-Side Management funds, regulatory disapproval or state-related (DSM), renewable portfolios, integrated resource purchasing difficulties. planning are all features of the current energy regulatory landscape worldwide. Even if the The tractability and coherence of a particular regulated companies cannot implement the approach is a dynamic process. As computers Integrated Resource Planning (IRP) which has been increase in power and user interfaces grow more produced, regulators of public policy reasons may sophisticated, more complex approaches become still want to see what specific goal-oriented IRPs increasingly feasible. Simply to imagine how might look like. Some models are more oriented complicated it would have been to add the towards public sector planning than are others. material balance feature of Decades to the WASP of the early `80s, is to understand how far the art Electricity system structure and ownership: of system modeling has come. Some modeling approaches may be more acceptable to integrated systems than to unbundled In similar ways, the interests of the current age ­ ones. Other models may be more oriented fuel price risks, intermittent generation/ 21A first cut at comparing several different models for their suitability to a changing business and technology environment in Mexico was made as a part of an earlier work on the Mexico GEF Wind Project. This working paper could provide the basis of a cross-model comparison. 24 Synopsis of the Expert Consultation Workshop dispatchability risks for renewable technologies, without the wind resources. Very detailed wind efficient diversification of generation portfolios ­ studies, with resolutions down to one second in some may not yet be elegantly addressable by existing cases, were combined with New York ISO hourly tools in addition to the current range of answers. loaddataandday-aheadforecasts.These How to put the appropriate analysis kit together measurement efforts were analyzed to determine without sacrificing accuracy, comprehensibility or the variance in system performance with regard to public acceptability is the challenge which future meetingloadwithandwithoutwindgeneration. efforts will need to address. Usingday-aheadwindforecasting,itwaspossibleto Case Studies reduce substantially the use of fuel to meet peak demand for power when wind was available. As was A final set of workshop presentations gave indicated by an earlier presentation by 3Tier, the assessments of wind investments and their role in a ability to predict wind even one day in advance, at a broad-basedgenerationportfolio.Bothpapers22 fairly rudimentary level of precision can result in stressed the need to integrate the analysis of a significantsavingsovertheno-predictioncase. specific investment with a more broad-based look at the risk in the system receiving the power output. The GE researchers found that it was necessary to combine production simulation, transmission The first approach, using elements of the GE system performance and wind data to effectively modelingsuite,showedhowtechnicalanalysisfrom understand the behavior of the system with a powerflow model could be integrated with additional wind generation. In addition, they found statisticalmeasuresofrisk,transmissionsystem that load varies more than does wind availability performance and bidding behavior, to give a basis from one day to another, minimizing the stability for estimating the reliability of supply with and impacts of wind on the overall system. However, Figure 3.10: Wind/RE Lowers Mexico Generating Cost Gas : $.03/kWh Coal: $.04/kWh Wind: $.05/kWh Mix S Mix Q 60% Fossil 98% Fossil 17% Hydro 0% Wind Mix P 10% Wind 2% Geo 40% Fossil 12% New Geo 2% Nudear Mix N 20% Wind 20% Wind 27% Hydro 12% Geo 2010 Mix 74% Fossil 2% Nuclear 20% Hydro 2010 Mix 2% Geo 67% Fossil 4% Nuclear 260% Hydro 2% Geo © Dr. Shimon Awerbuch. 22See Doug Welsh, The Effects of Integrating Wind Power on the New York State Power System; and Shimon Awerbuch, Generation Investment Planning and Modeling: A Finance Theory Perspective II: Portfolio-based Planning. 25 Risk Assessment Methods for Power Utility Planning Figure 3.11: One-step Analysis for Planners 1.5% Projected Mexico 1.0% 2010 Mix 0.5% +100-MW x100 Geothermal +1000-MW Gas-Turbine 0.0% Return in -0.5% +1000-MW +1000-MW Wind Power Hydropower Change +1000-MW -1.0% Oil -1.5% -3.0% -2.0% -10% 0.0% 1.0% 2.0% 3.0% Change in Risk (X100) © Dr. Shimon Awerbuch. the authors suggested that seasonal wind or further away from the efficient frontier. characteristics needed to be assessed as part of an Figures 3.10 and 3.11 show the results of such optimal generation portfolio so as not to induce the "comparative statics" analysis for Mexico. Figure underbuilding of other system resources, with a 3.10 represents Mexico's efficient frontier for consequent degradation in reliability. These results generation portfolios and Figure 3.11 is the represent the type of concrete outputs which a new comparative statics representation. modeling approach might yield for RE. The analysis in this paper shows that certain elements of risk can indeed be applied directly to Even the GE modeling effort stopped short of investment assessment modeling. Two of the comparing different generation portfolios for proposed follow-up steps for this activity are to systematic risk. A second paper by Dr. Awerbuch (i) identify and measure the risk elements; and looked at specific generation portfolios for Mexico (ii) derive a sequence of modeling activities which using the risk-return approach. can be undertaken so as to endogenize this risk analysis into the portfolio evaluation. Using generalized performance and cost data for wind, coal, CCGT-gas and hydro, the paper One of the major implications of the use of more compared a variety of portfolios for risk and cost. As sophisticated risk assessment techniques is that was predicted by the theory, ignoring the systematic single-plant evaluations become, paradoxically, risk degrades the financial performance of a more feasible if the reference case generation particulargenerationportfolioandtheassessmentof path has already undergone transformation to a future generation investments generally. risk-reward frontier format. This will mean generalizing the risk assessment of generation One of the most powerful techniques in the portfolios along the lines suggested in this finance tool kit, after the efficient frontier itself, is chapter under the title "How much the ability to assess whether any given addition to Consideration of Risk can be Incorporated into a generation capacity will move the system closer to Coherent Modeling Framework?" 26 Synopsis of the Expert Consultation Workshop Practical Applications of Multiple (i) needs; and (ii) capabilities. By needs we mean Model Approaches the entire set of issues regarding system size, structure, variety of prime movers and fuel cycles, If the era of "one-size-fits-all" modeling is over ownership patterns and the likelihood of significant for electricity system investments, then it is restructuring in the near future. By capabilities we reasonable to ask which tools are currently mean the training, size, responsibilities, available, and how they might be combined, so as sophistication of the current utility planners along to incorporate some of the important observations with the abilities of the financial sector and from this workshop. A detailed investigation into regulators to comprehend the output of a chosen how to combine two or more of these models, modeling approach in an approval vein. and the circumstances under which such combinations would be most useful, is the subject A final set of considerations is operational. of the next phase of this work. However, As noted above, modeling suites cannot become experience in some of World Bank's DMCs, and too complex in their results, especially if these in the New York wind study by GE, already point results must be presented to a public sector towards fruitful decisions regarding innovative entity for approval of prudence or finance. ways to make use of existing tools. Moreover, different modeling suites will have differing data input formats and may vary The key desiderata about the appropriate considerably from one family to another. modeling approach, shown in Figures 3.7 and However, for effective planning in utilities, 3.8, indicate the key decision considerations about especially publicly-owned ones, there should be what models to use and in what combination are: a reasonable degree of continuity from one Table3.1:ModelingCombinations:NeedsandResources Modeling Needs WASP Family GE Family Siemens Family Least-cost Model WASP (+ Valoragua) WASP + MARS Strategist LCP + Dispatch WASP + Valoragua + WASP + MARS + Strategist + ProMod GTMax MAPS LCP + Dispatch + WASP + Valoragua + WASP + MARS + Strategist + ProMod Transmission GTMax/EMCAS, or WASP + MAPS Strategist + GTMax/EMCAS LCP + Dispatch + WASP + Valoragua + WASP + MAPS + Strategist + ProMod Transmission + GTMax/EMCAS, or WASP + Proprietary Risk + Nostradamus or Risk Strategist + GTMax/EMCAS + Model PowerBase Proprietary Risk Model23 Notes: Strategist can use the WASP database. Neither Nostradamus nor PowerBase is a portfolio model in the sense that is laid out in Professor Awerbuch's papers. The choice of GTMax or EMCAS is a needs-based one. In more market-oriented systems, a user might opt for EMCAS, whereas in a largely integrated system with some market aspects, GTMax is probably more appropriate. 23One possible output of the next phase of this project is a generalized risk-return portfolio model which would stand in for the "proprietary" model now slotted in the WASP and GE family trees. 27 Risk Assessment Methods for Power Utility Planning year to the next so that model results can be energy in New York State used the results of a readily compared. least-cost generation plan + power flow simulation from MAPS + specific wind data + load variance With these cautions in mind, Table 3.1 provides a simulation to derive the expected contribution of general guide to reasonable uses and wind energy at different times of the year and combinations of models for system planning: under varying wind conditions. As Table 3.1 indicates, there is no one family Another case study, cited by the Argonne National which provides an entirely integrated solution to Laboratory (ANL) presenters, combined WASP and the more complex problems which have been GTMax to assess generation and transmission raised in this workshop. However, it is system options for South-East Europe. This study reasonable to look at whether the members of a provides a textbook-type examination of how the particular family of models can address the two models can be combined. issues which are likely to arise with additional needs in the future. The ongoing GEF project in Mexico will attempt to extend the work done by GE and ANL, by As the case studies showed, there is already a combining the least-cost planning activities with significant amount of work which combines one or hydro and wind assessments using Valoragua more of these models. The GE case study on wind and GTMax. 28 4. Results of the Workshop and the Way Forward Findings, Conclusions and Recommendations Workshop Findings On June 27-28, 2006, the World Bank's Energy The focus on risk assessment and risk mitigation in Sector Management Assistance Program (ESMAP) power systems has been motivated by several conducted a workshop on electricity system changes in the power industry environment. Power investment modeling and risk mitigation. The focus systems have been restructured, breaking the of the workshop was to find ways of integrating verticalintegrationwhichallowedtransmissiontobe appropriate risk considerations into advanced treated as a technical matter. Private power simulation models in a way which will improve the producers are more concerned about plant ability of planning and investment models to dispatchability in a competitive setting, fuel prices capture the complex trade-offs which must be haveonceagaindestabilizedinvestmentplans,and made in successful electricity investments. environmental concerns have increased awareness of the need to treat renewable generation resources Practitioners of the art of electricity planning and in a more detailed manner. And finally, estimates of modelingfromArgonneNationalLaboratory(USA), plantconstructioncostsstillloomlargeasriskfactors GE Power Systems, World Bank and Mexico's CFE forsomegenerationtechnologies. were brought together with electricity regulators, analysts,financialeconomistsandclimatescience Participants in the workshop noted that the specialists to present their views on the central theme structural changes in the power sector, among of the workshop: how can modelers improve the World Bank's DMCs, have brought into question treatment of risk in large-scale simulations of power the ability of any single approach to investment systeminvestments? planning to address appropriately the key issues To address this theme, the conference participants facing investors and regulators. In addition to presented papers on a wide variety of germane the structural and technological changes, the topics, including: following issues were identified by the conference participants as key factors in the · Currently available simulation tools; decision-support environment: · Desired characteristics of future simulationtools; · Characteristics of the system: · Regulatory issues and concerns with models Size and structure; degree of market opening; andsimulation; and interaction with other power systems in · Endogenizing financial risk assessment in neighboring countries; simulation models; and · Policy priorities ­ diversification, RE, domestic · Case studies on risks and risk mitigation in RE. resource mobilization; 29 Risk Assessment Methods for Power Utility Planning · Integration of diverse concerns of financial controversial as an accompaniment to new power community, regulators, IPP developers; sector participants, new financial constraints, and · Increased perception of the importance of emerging technologies. information flows between generation and transmission; As a part of an improved planning process, · Risk and uncertainty identification and some entity, agreed to by other market mitigation strategies; and participants, must take ownership of that · Loss tolerance ­ the ability of a power system process. A key part of the credibility of putting to absorb planning mistakes. one institution in charge is the need to safeguard the objectivity of both the process This led to the first finding of the workshop: and its results. Trade-offs are inevitable because it is clear that the more comprehensive Finding 1: There has been material change in the planning process, the more the various the utility industry environment and structure concerns can be taken into account. However, in the Bank's DMCs, vitiating many of the such richness may come at the cost of assumptions upon which the current comprehensibility, especially for some of the investment planning process operates. market participants who may not be specialists in the arcana of power system modeling and What has emerged is the desire to see, first of all, planning methods; leading to: an improved planning process. In this changed environment, the planning process must address Finding 2: New investment analysis tools and the issues of ownership, objectivity, approaches are available which address a comprehensiveness and comprehensibility of both broader range of issues than least-cost the process and its outputs. In countries as diverse planning alone can do. as Mexico, Indonesia and Jamaica, the planning process for generation has itself become The trade-offs are illustrated in Figures 4.1 and 4.2. Figure4.1:CapabilitiesandNeedsinPowerSystemInvestmentPlanning National power systems move from basic planning and investment assessment to Complex increasingly sophisticated forms Significant Role of least-cost system planning. for technical assistance LCP + Dispatch + As with most decision-support (TA) and Consulting Transmission systems, there are trade-offs. Simple analytical approaches LCP + Dispatch are relatively easy to explain to High regulators and public funding authorities. They are highly Capabilities transparent. Least-cost Planning More complex tools, relying on sophisticated models and computational techniques, are Basic Sophisticate Energy Balance by definition less transparent, even as they become more useful. Regulators will (rightly) note that the analytical Production Cost capabilities should not run Models ahead of the ability to explain important public sector Heavy Use of investment choices even as more Next Plant Needs Dispatch Modeling private investment brings the need for better identification Basic and mitigation of risk factors. 30 Results of the Workshop and the Way Forward Figure 4.2: ComplexTools Allow Earlier Recognition of Risks Model Type Capabilities Complexity Transparency Addressing Risk Next Plant Production Cost Energy Balance Least-cost Model LCP + Dispatch LCP + Dispatch + Transmission Note: The use of an integrated risk assessment tool, such as risk-return modeling, will generally increase the ability to assess risk while adding complexity and reducing transparency. Figure 4.1 shows how complexity increases as has remained the main focus of the effort. Fuel needs and capabilities evolve. Unfortunately, some price variations, rainfall (for hydro), wind of this complexity and richness of detail in output is speeds, construction cost risk along with other gained at the cost of transparency. Figure 4.2 risk factors are dealt with through the shows that as the complexity of the tool kit rises, so mechanism of defined scenarios. will the ability to identify risk earlier and more explicitly so as to better mitigate them. This For each of these issues, a separate scenario must trade-off between transparency and sufficiency of be constructed. The interaction of different risks approach will make itself felt in how regulators (i.e., correlation) is not addressed by most must grapple with investment and tariff approvals scenario methods. based on increasingly complex simulation models Such an approach has proved insufficient where and with the success of the national authorities in system deintegration, increased private investment attracting investment to replace the public funds and a constant improvement in simulation tools has previously used to pay for system expansion. raised the bar for what are considered useful and World Bank has played a key role in leading the appropriate investment planning tools and points evolution of its DMCs from basic methods to more towards the third finding of the workshop. sophisticated ones. WASP, the most widely used least-cost planning tool, has been promoted by the Finding 3: New investors and altered Bank and others as a relatively inexpensive, industry structures have brought the standardized and well-understood approach. Its explicit consideration of new types of ubiquity adds a degree of transparency to the risks to the forefront of investment planning model which other tools with similar capabilities and analysis. might not be granted by regulators and financers. To improve the usefulness of simulation and However, because such tools as WASP evolved in investment planning tools, the participants at the an era of vertically integrated and state-owned workshop agreed that models should assist in the power systems, treatment of dispatch, various identification of key risk factors and trade-offs and systematic risks and transmission was outside the also show the impacts of potential mitigation detailed economic planning process. Generation measures. One way to do that is a simple grid 31 Risk Assessment Methods for Power Utility Planning which shows the relationship between two key Finding 4: The specific characteristics of a attributes of risk: impact (or importance) and power system, including size, structure, controllability. The importance of a given risk ownership and prime mover types, will all element in system expansion varies with the impact play key roles in determining the appropriate of that element on the rate of return and the planning tools for system expansion and variability of that return. Fuel prices are very investment analysis. The "one-size-fits-all" important risks in a thermal-dominated system, but era is over. far less so in one which relies more on nuclear, hydro or wind. Conversely, plant dispatchability Throughout the workshop, the presenters may represent a significant economic risk in a demonstrated the factors which have led to calls wind or run-of-river hydro system, but a far lesser for a more inclusive approach to investment one in a coal or CCGT system. modeling. These factors include: The second key to understanding risk is to look at · Structural changes in utility systems ­ the controllability of that risk. Some risks can be unbundled systems require different analytical mitigated by actions of the plant owner or system approaches ­ call for modeling of the distinct regulator, while other risks fall beyond reasonable segments of the businesses, generation, attempts at mitigation. The risk and controllability transmission and distribution; matrix looks like the one depicted in Figure 4.3 for · Ownership changes, including IPPs in otherwise investment problems generally: integrated systems, have brought new investors, with new perspectives into the For a particular power system, the risk investment mix. The presence of new controllability matrix can point to specific participants has led to a greater appreciation problems, issues and trade-offs, as these will often for the endogenization of risk analysis in be both location- and technology-dependent, investment portfolios; implying that: · Increased capabilities in simulation modeling Figure 4.3: Importance and Controllability of Power System Risks IV High I Strategy Strategy Refinement Tend to these as time permits Resolve these first High Ability to High Ability to Mitigate Minor Risks Mitigate Important Risks Control Low High Importance Low Priority Issues Figure out how to circumvent, reduce or reformulate these Low Ability to Low Ability to Mitigate Mitigate Minor Risks Important Risks III II Non-strategic Low Contingency 32 Results of the Workshop and the Way Forward have made possible more sophisticated and · Risk factors need to be better quantified, time differentiated approaches to investment including correlation, if any, among different evaluation and modeling; and risk factors. This should include: · Greater interest in new generation ­ Fuel price risks, including gas/ technologies, especially renewable ones, has oil/coal correlations; made clear the need for approaches which ­ Construction cost and/or delay risks, can explicitly assess the costs and benefits of especially the correlations between large such technologies, not as an afterthought, hydro, coal and nuclear plants; but rather as an integral element in ­ Systemoperationrisks,includingcoincidence system planning. of expected generation from intermittent sources over the load curve; and Workshop Recommendations ­ Ability of transmission investments to reduce generation-side risks by enabling better use A new approach to the investment planning of generation resources, especially to back process must start from the understanding that up intermittent generation sources. today's investors and regulators are more · Proxy measures for risk as well as risk demanding with regard to the results of the mitigation need to be developed for smaller planning process. They are also far more aware systems where necessary data for risk analysis of the mistakes and problems have arisen in may be insufficient or unavailable, including: the past as a result of insufficient attention to ­ Wind and hydro-effective capabilities to risk factors. contribute capacity; Based on the discussions and presentations in the ­ Mitigating measures which may improve the workshop, as well as the findings presented above, ability of wind and hydro to contribute the following recommendations have emerged capacity (e.g., operational coordination from the proceedings: between wind generation and dispatch of storage hydro plants); and · Current least-cost generation planning tools ­ Role of transmission and improved should remain as a part of the overall planning dispatch procedures. process, but dispatch and transmission models should be integrated with the least-cost models Concerns and issues key to Financial Institutions where appropriate; and (FIs), IPP developers and regulators remain largely unaddressed and need to be more explicitly · For example, Mexico currently uses all three considered in investment planning process: techniques for operations or planning in its electricity system. However, these efforts are · Important sources of risk and uncertainty need not integrated, nor are the results and data to be identified, including: uniform. As a result, the insights into planning ­ Construction and technology risks; offered by dispatch modeling are not explicitly ­ Regulatory concerns and mitigating and systematically available to CFE, the measures, especially with regard to Mexican electricity operator. intermittent generating technologies; ­ Fuel price risks and methods to mitigate Least-cost generation planning should be such risks; and augmented with explicit endogenization of key risk ­ Trade-offs between expected generating factors appropriate to the system and the costs and systemic variability in technologies in use in that system; generating costs. 33 Risk Assessment Methods for Power Utility Planning · Dispatch and transmission modeling need to measures which World Bank can take in this become integral to generation investment regard include the following ones: planning process: As a first step, World Bank should canvas its DMCs ­ Integrated dispatch and LCP models need to to assess what their needs are in the new become the norm in larger systems; investment planning environment. This means ­ Transmission models need to be integrated talking to regulators, IPPs, private financial with generation planning to assess institutions and market operators and determining generation:transmission trade-offs and risk the key decision-support elements which the new reduction options; and environment requires. World Bank also needs to ­ Modeling for planning needs to be province investigate whether current planning and of a trusted party to avoid charges of bias investment activities generate systematic biases or omission. with regard to future demand, cost of new plants, · Improved planning process needs to be and the operational and dispatch characteristics of explained carefully to regulators to different types of plants. improve transparency and allay fears of "black box" syndrome. Important risk categories, such as those identified above, need to be better quantified and World Activities starting up under the GEF's Mexico Wind Bank can play a role in generating such Energy Project point the way for one possible information and disseminating it to the DMCs. In approach for organizing and implementing particular, World Bank needs to support efforts to enhanced investment planning for utilities. The gather and construct both country-specific and technical assistance for this project seeks to: proxy measure of the following risk elements: (i) quantify the risks in wind generation; (ii) identify mitigating dispatch strategies with hydro · Fuel price risks, including gas/ operations through integrated planning and oil/coal correlations; dispatch simulations; and (iii) use a posteriori · Construction cost and/or delay risks, especially analysis from the dispatch and planning models to the correlations between large hydro, coal and estimate an increasingly accurate value for nuclear plants and the correlations of such generation using wind. In addition, it will be construction risks with worldwide heavy possible, using the suite of simulation models engineering activity in mining and oil/gas proposed for the GEF Mexico project, to estimate production; closely related to construction the parameters of a more accurate and risks are sophisticated risk-return frontier, as an essential · Market measures of risk which continue to step to integrating such analysis into investment create noticeable differences between ex ante planning using optimization techniques. assessments of plant returns based on demand studies and ex post plant dispatch and In other DMCs, the issues will be different from utilization results; those in Mexico, especially as regards system · Output and operational data for wind and ROR size and ownership. If World Bank's DMCs are hydro plants to better gauge the operational able to take advantage of new methods of risks and the effective capacity contributions system planning and risk identification and associated with such plants; mitigation, then Bank will need to move · Potential role of transmission investments to affirmatively to make adoption of such new reduce generation-side risks by enabling better methods simpler and less costly, both in time use of generation resources, especially to back and money, than they now are. Specific up intermittent generation sources; and 34 Results of the Workshop and the Way Forward · Mitigating measures which may improve the of least-cost generation planning models ability of wind and hydro to contribute with other transmission and dispatch- capacity (for example, operational oriented programs. coordination between wind generation and · Specific models and their integration should be dispatch of storage hydro plants), improved investigated further to assess their ability to short-term water management in RoR hydro integrate with current approaches and units. appropriate integration programs need to be Risk analysis needs to be explicitly integrated into devised; and planning models and analyses. · WorldBankshouldhelptoimplementtheuseof modelswhichexplicitlyincludecontractswithIPPsor · World Bank should investigate how to take otherbilateralinstruments,aswellaspoolprices. the quantitative measures of risk and its proxies mentioned in the paragraph above At the same time, World Bank can help promote detailing the important risk categories, and economy of effort in simulation modeling of integrate these measures into larger investment and system expansion by supporting simulation and investment analysis models in high quality tools with training programs and data a manner which produces consistent and acquisition support which has been vetted by the robust results; testing activities and data analysis explained in the · World Bank should work with one or more of above paragraphs; and the software publishers in this business space to define appropriate ways to integrate the risk Training programs need to be designed to assist measures already enumerated; and World Bank's DMCs in making use of the new · World Bank should work with its DMC clients to investment analysis and planning methods to improve demand forecasting and prepare design and implement coherent investment methods of mitigating market-related risks. planning, analysis and risk mitigation capabilities, including the need to report results Where conditions warrant, World Bank should to regulators and investors in a manner which promote the use of integrated generation and enhances the quality of regulatory decisions, transmission planning, augmenting the use especially with regard to transparency and stakeholder participation. 35 ANNEX 1 World Bank/ESMAP Workshop on Electricity Investment Modeling and Risk Mitigation World Bank and ESMAP have embarked upon 09:30-10:15 New Approaches, WASP and a study to contribute to improving power Beyond I ­ Günter Conzelmann systems planning methodologies to better and Tom Veselka, Argonne reflect supply and price uncertainties, and National Laboratory valuation of supply diversity. To support this 10:15-10:30 Q&A with commentary by Enrique work, we are co-hosting this workshop to lay Crousillat, The World Bank the groundwork for a systematic assessment 10:30-10:45 Coffee Break of planning methods and attempt to reach 10:45-11:30 NewApproaches,WASPandBeyond consensus on a development path which will II ­ E LaRose, GE Power Systems; ultimately mainstream such approaches in the power sector. The output from this 11:30-12:15 Q&A with commentary by Shimon workshop will contribute to preparation of Awerbuch, Tyndall Center, model specifications and terms of reference University of Sussex for the development of new planning tools, 12:15-13:30 Lunch and ultimately to the development and validation of improved models for power Afternoon Sessions I: Investment Decisions for systems planning. Public and Regulated Generators ­ Session Chair Charles Feinstein, The World Bank Workshop Agenda 13:30-14:15 Public Sector Perspectives on Investment Planning, How do the Tuesday, June 27 Needs of State Enterprises Differ 08:30-09:00 Continental Breakfast at from Private Electricity Meeting Room Companies? ­ Andres Soler, 09:00-09:15 Welcome and Introduction ­ Anil Comision Federal de Electricidad Cabraal, The World Bank (Mexico), Günter Conzelmann, Argonne National Laboratory Morning Sessions: Current Approaches to 14:15-15:10 Regulator's Viewpoint on Electricity System Investment Modeling ­ Modeling, How does a New What Works and What Needs Fixing ­ Session Modeling Approach or Tool Pass Chair Anil Cabraal, The World Bank Regulatory Muster? ­ Alejandro 09:15-09:30 Issues and Problems in Electricity Peraza, Comision Reguladora de Generation Investment Modeling, Energia (Mexico), Rajnish Barua, Current Practices and Results ­ Pennsylvania Public Utility Donald Hertzmark, Consultant, Commission, John Besant-Jones, The World Bank Consultant, The World Bank 37 Risk Assessment Method for Power Utility Planning 15:10-15:30 Q&A with commentary by Charles Wednesday, June 28 Feinstein, The World Bank 15:30-15:45 Coffee Break 08:30-09:00 Breakfast Afternoon Sessions II: Alternative Case Studies and the Way Forward ­ Formulations of the Investment Decision and Session Chair Enrique Crousillat, Mitigating Measures ­ Session Chair John The World Bank Besant-Jones, Consultant, The World Bank 09:00-09:40 Managing a Transition to New 15:45-16:20 Critique of Optimization Approaches ­ E LaRose, GE Approaches and Introduction of Power Systems; Andres Soler, Risk-return Criteria for Investment CFE Planning in Power Systems ­ 09:40-10:30 Case Studies of Alternative Shimon Awerbuch, Tyndall Center, Approaches ­ Shimon Awerbuch, University of Sussex D Welsh, GE Power Systems 16:20-17:00 Risk Mitigation Methods in Portfolio 10:30-10:45 Coffee Break Optimization ­ Victor Niemeyer, 10:45-11:10 Q&A with commentary by Bart EPRI; Bart Nijssen, 3Tier Nijsson, 3Tier Environmental Forecasting Group 11:10-12:00 Round-table Discussion ­ 17:00-17:20 Q&A with commentary by Günter Summary, Next Steps, Relevance Conzleman, Argonne National to World Bank Members ­ Laboratory Moderated by Donald Hertzmark 17:20-17:45 Lessons Learned and Round-table and Andres Soler, Comision Discussion ­ Moderated by Donald Federal de Electricidad (Mexico) Hertzmark, Consultant, 12:00-12:20 Closing Remarks, Enrique The World Bank Crousillat, The World Bank 17:45-19:00 Reception 12:20 WorkshopAdjourns 38 RENEWABLE ENERGY Moving into a world with less carbon emissions, better energy security through a more diversified energy supply, and increased availability of energy in unserved areas, in particular where the poorest people live. ESMAP supports renewable energy with advice on policy formulation and development incentives adapted to local conditions. The program assists in design of Energy Sector Management Assistance Program (ESMAP) renewable energy projects suitable for 1818 H Street, NW financing by bilateral assistance, Washington, DC 20433 USA international institutions, or the Tel: 1.202.458.2321 private sector. Fax: 1.202.522.3018 Internet: www.esmap.org Email: esmap@worldbank.org The analytical work of ESMAP includes legal and regulatory frameworks for renewables, efficient integration of distributed generation in electrical power systems, and better energy access for remote and poor communities. ESMAP is a knowledge clearing house for good practice and opportunities for renewable energy ranging from large scale electricity generation to biomass serving household heating and cooking needs.