Globalizing the Cost of Capital and Capital Budgeting at AES

Globalizing the Cost of Capital and Capital Budgeting at AES

Globalizing the Cost of Capital and Capital Budgeting at AES

9-204-109REV: OCTOBER 23, 2006MIHIR DESAIGlobalizing the Cost of Capital and CapitalBudgeting at AESIn June 2003, Rob Venerus, director of the newly created Corporate Analysis & Planning group atThe AES Corporation, thumbed through the five-inch stack of financial results from subsidiaries andconsidered the breadth and scale of AES. In the 12 years since it had gone public, AES had become aleading independent supplier of electricity in the world with more than $33 billion in assets stretchedacross 30 countries and 5 continents. Venerus now faced the daunting task of creating a methodologyfor calculating costs of capital for valuation and capital budgeting at AES businesses in diverselocations around the world. He would need more than his considerable daily dose of caffeine topoint himself in the right direction.Much of AES’s expansion had taken place in developing markets where the unmet demand forenergy far exceeded that of more developed countries. By 2000, the majority of AES revenues camefrom overseas operations; approximately one-third came from South America alone. Once a criticalelement in its recipe for success, the company’s international exposure hurt AES during the globaleconomic downturn that began in late 2000. A confluence of factors including the devaluation of keySouth American currencies, adverse changes in energy regulatory environments, and declines inenergy commodity prices conspired to weaken cash flow at AES subsidiaries and hinder thecompany’s ability to service subsidiary and parent-level debt. As earnings and cash distributions tothe parent started to deteriorate, AES stock collapsed and its market capitalization fell nearly 95%from $28 billion in December 2000 to $1.6 billion just two years later.As one part of its response to the financial crisis, AES leadership created the Corporate Analysis &Planning group in order to address current and future strategic and financial challenges. To beginthe process, the CEO and board of directors asked Venerus, as director of the new group, to revaluethe company’s existing assets, which required creating a new method of calculating the cost of capitalfor AES businesses. Central to the questions facing Venerus was the international scope of AES, as heexplained: “As a global company with operations in countries that are hugely different from the U.S.,we need a more sophisticated way to think about risk and our cost of capital around the world. And,frankly, the finance textbooks aren’t that helpful on this subject.”The mandate from the board of AES to create a new methodology presented an interesting butoverwhelming challenge. As he prepared his materials for the board, Venerus wondered if his newapproach would balance the complexities of the unique business situations around the world with________________________________________________________________________________________________________________Professor Mihir Desai and Research Associate Doug Schillinger prepared this case. HBS cases are developed solely as the basis for classdiscussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management.Copyright © 2004 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685,write Harvard Business School Publishing, Boston, MA 02163, or go to http://www.hbsp.harvard.edu. No part of this publication may bereproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or by any means—electronic, mechanical,photocopying, recording, or otherwise—without the permission of Harvard Business School.204-109Globalizing the Cost of Capital and Capital Budgeting at AESthe need for a simple, straightforward process that could be implemented accurately and consistentlythroughout the organization.AES Corporation1Roger Sant (MBA ’60, HBS) and Dennis Bakke (MBA ’70, HBS) founded AES Corporation(originally Applied Energy Services) in 1981 shortly after the adoption of federal legislation thatbecame known as the Public Utility Regulatory Policy Act (PURPA). The legislation was part of theUnited States government’s reaction to growing concern over American dependence on foreign oil.The act sought to diminish this dependence by requiring that electric utilities source some of theirnew power needs through qualified cogenerators and small independent power producers, providedthat the power generated by independents cost less than if the utility were to produce the poweritself. Sant and Bakke recognized that in shielding small independent power producers from costlystate and federal regulation, PURPA actually created a market for a new private sector power market.In practice, the act almost ensured that independent power producers could undercut a utility’s costof production.The company initially struggled to raise financing but after the construction of its firstcogeneration facility in Houston, Texas, in 1983 and the subsequent development of a profitablecogeneration facility in Pittsburgh, Pennsylvania, in 1985, AES experienced rapid growth. By thetime the company went public in 1991, revenues had grown to $330 million and net income hadsoared to $42.6 million from $1.6 million just three years earlier.In the early 1990s, AES began to shift its focus overseas where there were more abundantopportunities for the company to apply its nonrecourse, project finance model to the development ofcontracted generating facilities. In addition, foreign governments often provided incentives to attractforeign direct investment in infrastructure projects like power plants. The willingness ofinternational development banks to invest alongside AES in volatile parts of the world helpedmitigate the risk of expropriation, and the increased breadth of the global financial markets providedgreater access to capital.AES initiated its international expansion in 1991–1992 with the purchase of two plants in NorthernIreland. The following year, AES began what would become a massive expansion into Latin Americawith the acquisition of the San Nicolas generation facility in Buenos Aires, Argentina. A year later,AES created a separately listed subsidiary, AES China Generating Co., to advance Chinesedevelopment projects. As the pace of deregulation quickened around the world, AES was presentedwith an abundant supply of capital and a wealth of opportunities for investments in energy-relatedbusinesses, some of which were more complex than AES’ portfolio of contract generation projects. Inaddition to expanding its line of business profile, it continued its geographic expansion and between1996 and 1998 the company acquired several large utility companies in Brazil, El Salvador, andArgentina. By this time the company was spending an estimated 80%–85% of its capital investmentoverseas in places as diverse as Australia, Bangladesh, Canada, Cameroon, The Dominican Republic,Georgia, Hungary, India, Kazakhstan, the Netherlands, Mexico, Pakistan, Panama, Puerto Rico,Ukraine, The United Kingdom, and Venezuela.21 Much of this overview comes from Paula Kepos, ed., International Directories of Company Histories, Volume 10 (Detroit: St.James Press, 1995), pp. 25–27.2 Paula Kepos, ed., International Directories of Company Histories, Volume 53. (Detroit: St. James Press, 1995), p. 17.2Globalizing the Cost of Capital and Capital Budgeting at AES204-109AES in 2002By 2002, AES was one of the largest independent power producers in the world. (See Exhibits 1,2, and 3 for AES consolidated financial statements.) The company was organized around fourseparate lines of business: Contract Generation, Competitive Supply, Large Utilities, and GrowthDistribution.3Contract generationIn 2002, AES’s Contract Generation business accounted forapproximately 29% of AES revenues and consisted of generation facilities, which sold electricityunder long-term (five years or longer) contracts. The term of the contracts allowed AES to limit itsexposure to volatility in electricity prices. The resulting stable production requirements enabled AESto accurately predict supply needs and enter into similarly long-term agreements for coal, naturalgas, and fuel oil, thereby limiting its exposure to fuel price volatility. Facilities varied considerably insize, with plants as small as the 26 MW Xiangci-Cili hydro plant in China to the enormous 10-plant2,650 MW Tiete hydro complex in Brazil.Competitive supply Accounting for 21% of AES revenues, the Competitive Supply line ofbusiness sold electricity directly to wholesale and retail customers in competitive markets usingshorter-term contracts or daily spot prices. Competitive Supply businesses, sometimes called“merchant plants,” were highly susceptible to changes in the price of electricity, natural gas, coal, oiland other raw materials. AES’s margin in U.S. dollars was influenced by a host of factors includingweather conditions, competition, changes in market regulations, interest rate and foreign exchangefluctuations, and availability and price of emissions credits. Such price volatility had recentlydamaged several Competitive Supply businesses including the Drax plant in the U.K., the largestplant in AES’s Competitive Supply fleet.4Large utilitiesBy the end of 2002, the Large Utility business included only three majorutilities, each in a different country: Indiana Power and Light Company in the U.S. (IPALCO),Eletropaulo Metropolitana Electricidade de Sao Paulo S.A. in Brazil (Eletropaulo), and C.A. LaElectricidad de Caracas in Venezuela (EDC). These utilities combined generation, transmission anddistribution capabilities and were subject to local government regulation and price setting. All threeenjoyed regional monopolies and in total accounted for 36% of AES revenues. U.S. energyregulations had required AES to sell a fourth such company, Central Indiana Light and Power(CILCORP), when AES purchased IPALCO, a sale that was completed near the end of 2002.Growth distributionGrowth Distribution businesses offered AES significant potentialgrowth due to their location in developing markets where the demand for electricity was expected togrow at considerably faster rates than in developed countries. However, these businesses also facednotable risks related to operating difficulties, less stable governments, and regulatory regimes, anddiffering cultural norms regarding basic principles such as payment conventions and safetyregulations. Two new Growth Distribution businesses in Ukraine (Kievoblenergo and Rivoblenergo)and one in Cameroon (SONEL) were acquired as recently as 2001.3 The description for these lines of businesses comes largely from AES’s annual reports; see AES Corporation, 2001 AnnualReport (Arlington: AES Corporation, 2002) and AES Corporation, 2002 Annual Report (Arlington: AES Corporation, 2003).4 Energy companies typically refer to generation companies not as members of a “portfolio” but members of a “fleet.”3204-109Globalizing the Cost of Capital and Capital Budgeting at AESRecent DifficultiesAES’s placement in foreign markets as well as poor performance at several new U.S. businessesnearly crippled the company during the global economic slowdown that began in 2001. AES’smarket value started to fall slowly in 2001 but fell precipitously in 2002. Having traded for more than$70 per share in October 2000, AES stock hovered around $1 per share in the same month of 2002 (seeExhibit 4). Wall Street began to question the company’s ability to weather the storm, and one analystwrote, “It is clear that AES’s current stock price is reflecting the scenario that the company will notsurvive.”5 The collapse of the stock price and the subsequent $3.5 billion loss that included asubstantial write-off in 2002 were brought on by several factors, the effect of which was amplified byAES’s capital structure. Among these factors were adverse shifts in foreign exchange markets,regulatory policies, and commodity prices; many of these were factors AES could not fully protectitself against.Currency DevaluationsDuring 2001, a political and economic crisis in Argentina brought about a significant devaluationof most South American currencies against the U.S. dollar. In December, the newly electedgovernment abandoned the country’s fixed dollar-to-Argentine-peso exchange rate (1:1) andconverted U.S. dollar-denominated loans into pesos. On its first day of trading as a floating currency,the peso lost 40% of its value against the U.S. dollar.6 By the end of the year, the peso was trading ata rate of 3.32 pesos to the U.S. dollar and had been as high as 3.9 pesos.7 The currencies in Brazil andVenezuela—equally important markets for AES—followed suit, with the Brazilian Real and theVenezuelan Bolivar each depreciating approximately 50% against the U.S. dollar during the sameperiod (see Exhibit 5). As a result, AES recorded foreign currency transaction losses of $456 millionin 2002.Several of AES’s subsidiaries in South America defaulted on their debt and were forced torestructure. The debt was nonrecourse to the parent, AES Corporation, so AES was not obligated toservice the subsidiary debt. However, the parent company did suffer from cash flow shortfalls as aresult of lower-than-expected dividends back from the subsidiaries. The impact of devaluation wasincreased when foreign businesses were paid in local currency but had obligations to repay debtdenominated in U.S. dollars.Adverse Regulatory ChangesDuring the late 1990s, the regulatory agencies in Brazil had failed to produce a market structuresufficiently attractive to encourage domestic construction of new generation assets. Demandexceeded supply, causing shortages. The majority of Brazil’s generation capacity was hydroelectric,and energy deficiencies were exacerbated in 2001 and 2002 by below-average rainfall. In response,the Brazilian regulatory authorities began rationing energy consumption in June 2001.8 In addition tothe loss of sales volume, the decline of the Brazilian real against the dollar triggered a regulatory5 Ali Agha and Ed Yuen, Banc of America Securities, “AES Corporation, Analysis of Sales and Earnings,” October 25, 2002,available from The Investext Group, http://www.investext.com, accessed July 15, 2003.6 “Argentina’s Peso I Expected to Face Pressure This Week,” The Wall Street Journal, January 14, 2002, available from Factiva,http://www.factiva.com, accessed July 7, 2003.7 AES Corporation, 2002 Annual Report (Arlington: AES Corporation, 2003), p. 38.8 Ibid., p. 20.4Globalizing the Cost of Capital and Capital Budgeting at AES204-109conflict concerning the applicable exchange rate for the real-to-dollar energy-cost pass-throughprovisions in AES’s contract. In effect, the government of Brazil required AES to purchase energy indollars while reimbursing the costs using an earlier period exchange rate, which lagged the deflation.In the fourth quarter of 2002, AES took a pretax impairment charge of approximately $756 million onEletropaulo, one of its major Brazilian businesses.Commodity Prices DeclineA 2001 change in the regulatory regime in the U.K. also adversely impacted AES by increasingcompetition and reducing prices in its generation markets. That, along with an unusually warmwinter in the U.K., brought wholesale electricity prices down approximately 30%.9 These pressurescaused several counterparties to default on their long-term purchase agreements. This counterpartrisk, coupled with changes in the commodity markets, enhanced the financial pressure on AESfacilities, and those that could not sell electricity above their marginal costs were taken off-line orshut down.Above and beyond the currency and regulatory difficulties at AES, the company was forced totake significant impairment charges on unprofitable or discontinued businesses. In 2002, thecompany took after-tax charges of $465 million on development and construction projects, $301million on discontinued operations, and a massive $2.3 billion in asset impairments associated withseveral large utility and generation businesses.10AES ReactionIn response to the financial crisis, AES successfully refinanced $2.1 billion of bank loans and debtsecurities. The refinancing arrangement came through the day before AES was to pay down $380million of its outstanding debt. A group of 63 banks and investment funds agreed to provide $1.6billion in new loans, and AES secured a two-year extension on another $500 million in notes due in2002.11AES also secured agreements to sell a number of its assets. Total proceeds from the sales wereexpected to be approximately $819 million. Proceeds from sales in 2003 were expected to beapproximately $310 million.12Capital Budgeting at AESHistorically, capital budgeting at AES was fairly straightforward. When AES undertook primarilydomestic contract generation projects where the risk of changes to input and output prices wasminimal, a project finance framework was employed. Venerus explained that this frameworkconsisted of a fairly simple set of rules—all nonrecourse debt was deemed good, the economics of agiven project were evaluated at an equity discount rate for the dividends from the project, all9 AES Corporation, 2002 Annual Report, p. 21.10 Ibid., p. 37.Eletropaulo.The $2.3 billion in asset impairment charges included the $706 million after tax impairment charge at11 “AES Stock Shoots Up as Refinancing Keeps Bankruptcy at Bay,” The Washington Post, December 17, 2002, available fromFactiva, http://www.factiva.com, accessed July 17, 2003.12 AES Corporation, 2002 Annual Report, p. 36.5204-109Globalizing the Cost of Capital and Capital Budgeting at AESdividend flows were considered equally risky, and a 12% discount rate was used for all projects. In aworld of domestic contract-generation projects where most risks could be hedged and businesses hadsimilar capital structures, Venerus felt that this model worked fairly well.Beginning in the early 1990s, with AES’s international expansions, this model of capital budgetingwas exported to projects overseas. Early on, the model worked well (as it had with the initialexpansion in Northern Ireland), because this project had many of the characteristics of domesticopportunities. Venerus explained that the model became increasingly strained with the expansionsin Brazil and Argentina because hedging key exposures such as regulatory or currency risk was notfeasible. In addition, the financial structure of a going-concern business like a utility is notablydifferent than that of a limited-lifespan asset like a generating facility. Nonetheless, in the absence ofan academic or other alternative, the basic methodology remained intact.Another factor that created fundamental difficulties for transporting this model to overseassettings was the ever-increasing complexity in the financing of international operations.As one example of this, Venerus described how international operations would be evaluated andfinanced. Exhibit 6 illustrates the typical structure: subsidiary A and B were financed with debt thatwas nonrecourse to the parent. The subsidiaries’ creditors had claims on the hard assets at the powerplants but not on any other AES affiliate or subsidiary. The local holding company, which oftenrepresented multiple subsidiaries, also borrowed to finance construction or acquisitions and receivedequity in the various subsidiaries it held. In addition, the holding company had debt that wasnonrecourse to the parent, secured by dividends from the operating company. Finally, AESborrowed once again at the parent level in order to contribute equity dollars into holding companiesand subsidiary projects. At the end of 2002, AES had $5.8 billion in parent company (recourse) debtand $14.2 billion in nonrecourse debt.Using this subsidiary structure, the parent company received cash flows in the form of dividendsfrom each subsidiary (some of which were holding companies) and, because the structure of everyinvestment opportunity was essentially the same, all dividend flows were evaluated at the same 12%discount rate. This had the benefit of making similar projects seemingly comparable. However,when subsidiaries’ local currency real exchange rates depreciated, leverage at the subsidiary andholding company level effectively increased, and the subsidiaries struggled to service their foreigncurrency debt. Venerus recalled how the model started to crumble in early international investments:Imagine a real devaluation of 50%. That cuts EBITDA in dollar terms by 50% and coverageratios deteriorate by more than 50%. The local holding company cannot service its borrowing,and dividends to the parent are slashed. Ultimately the consolidated leverage was well over80% without any hedging of foreign exchange for any meaningful duration; this is where themodel broke down.Venerus’s solution to the problem had to be consistent, transparent, and accessible. He knew hissolution would have to account for changes in required returns due to leverage, incorporate someunderstanding of a project’s risk profile, potentially include country risks, and still provide valuesthat were consistent with market behavior, including trading multiples.Globalizing the Cost of CapitalTo overhaul the capital budgeting process and evaluate each investment as a distinct opportunitywith unique risks, Venerus knew he would have to calculate a cost of capital for each of the manydiverse AES businesses. As a starting point, he considered the 15 representative projects shown in6Globalizing the Cost of Capital and Capital Budgeting at AES204-109Exhibit 7a and, using the financial data in Exhibit 7b, he endeavored to derive a weighted averagecost of capital (WACC) for each project using a standard methodology:WACC =EDre + rd (1 ? ? )VVIn order to calculate each WACC, Venerus knew he would have to measure all of the constituentparts for the 15 projects: the cost of debt, the target capital structure, the local country tax rates, andan appropriate cost of equity. In order to find the cost of equity, he would first have to estimate areasonable equity beta.Venerus questioned whether the traditional CAPM model could help him calculate all of thenecessary ingredients for AES businesses in emerging markets. He did not advocate the use of a“World CAPM” where beta measured the covariance of a project’s return to the world marketportfolio of equities. AES owned businesses in poorly integrated capital markets, so Venerus fearedthe use of a World CAPM might yield artificially low costs of capital due to the low (or in some casesnegative) correlation of developing economies with the world market. For example, a world CAPMmight generate the unreasonable result of a WACC lower than the U.S. risk-free rate due to itsnegative correlation with the world market portfolio.Similarly, Venerus did not advocate the use of a “Local CAPM” where beta measured thecovariance of a project’s returns with a portfolio of local equities. Countries such as Tanzania orGeorgia, where AES had projects, did not have any meaningful equity markets or local benchmarks.Still, he knew he had to find a way to capture the country-specific risks in foreign markets. At ahigh level, Venerus developed an approach with two parts. First, he calculated a cost of debt andcost of equity for each of the 15 projects using U.S. market data. Second, he added the differencebetween the yield on local government bonds and the yield on corresponding U.S. Treasury bonds toboth the cost of debt and the cost of equity. Venerus believed that this difference or “sovereignspread” approximated the incremental borrowing costs (and market risk) in the local country.Exhibit 8 summarizes Venerus’s approach.Calculating the Cost of Equity and the Cost of DebtTo estimate an equity beta for each project, Venerus first had the Corporate Analysis & Planninggroup take unlevered equity betas from comparable U.S. companies. They averaged the betas toyield one unlevered beta for each of the four lines of business. Since the equity betas reflected notonly the market risk associated with each company, but also the differential effects of leverage, thegroup relevered the equity betas at indicative capital structures for each of the 15 projects using thefollowing equation:? levered =? unleveredEVUsing the relevered equity betas, Venerus had the group calculate the cost of equity for eachproject using the traditional CAPM equation:(Cost of Equity = r f + ? rm ? r f)7204-109Globalizing the Cost of Capital and Capital Budgeting at AESFinally, an appropriate cost of debt…Read the case study then do thecalculation and answer the questions. There are 11 parts to this. The first 10steps require math calculations (you must put answers directly into this excelsheet with formula). However step 11 requires you to answer questions about thefirst 10 steps (basically explain the implications of the numbers youcalculated/what do the numbers mean)1)Does each step make sense? 2) If youthink a step is not right, how would you change it? 3)Thisstep 11 will result in a few recommendations to AES CEO.

 

CLICK TO ORDER A CUSTOM PAPER

We are the best paper writer / paper help in the world