Globalizing the Cost of Capital and Capital Budgeting at Aes
Globalizing the Cost of Capital and Capital Budgeting at AES In June 2003, Rob Venerus, director of the newly created Corporate Analysis & Planning group at The AES Corporation, thumbed through the five-inch stack of financial results from subsidiaries and considered the breadth and scale of AES. In the 12 years since it had gone public, AES had become a leading independent supplier of electricity in the world with more than $33 billion in assets stretched across 30 countries and 5 continents.Venerus now faced the daunting task of creating a methodology for calculating costs of capital for valuation and capital budgeting at AES businesses in diverse locations around the world.
He would need more than his considerable daily dose of caffeine to point himself in the right direction. Much of AES’s expansion had taken place in developing markets where the unmet demand for energy far exceeded that of more developed countries. By 2000, the majority of AES revenues came from overseas operations; approximately one-third came from South America alone.Once a critical element in its recipe for success, the company’s international exposure hurt AES during the global economic downturn that began in late 2000. A confluence of factors including the devaluation of key South American currencies, adverse changes in energy regulatory environments, and declines in energy commodity prices conspired to weaken cash flow at AES subsidiaries and hinder the company’s ability to service subsidiary and parent-level debt.As earnings and cash distributions to the 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 begin the process, the CEO and board of directors asked Venerus, as director of the new group, to revalue the company’s existing assets, which required creating a new method of calculating the cost of capital for AES businesses.Central to the questions facing Venerus was the international scope of AES, as he explained: “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 but overwhelming challenge.As he prepared his materials for the board, Venerus wondered if his new approach 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 class discussion. 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 be reproduced, 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. 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES he need for a simple, straightforward process that could be implemented accurately and consistently throughout the organization. AES Corporation1 Roger 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 that became known as the Public Utility Regulatory Policy Act (PURPA). The legislation was part of the United 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 their new power needs through qualified cogenerators and small independent power producers, provided that the power generated by independents cost less than if the utility were to produce the power itself. Sant and Bakke recognized that in shielding small independent power producers from costly state 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 cost of production. The company initially struggled to raise financing but after the construction of its first cogeneration facility in Houston, Texas, in 1983 and the subsequent development of a profitable cogeneration facility in Pittsburgh, Pennsylvania, in 1985, AES experienced rapid growth. By the time the company went public in 1991, revenues had grown to $330 million and net income had soared to $42. million from $1. 6 million just three years earlier. In the early 1990s, AES began to shift its focus overseas where there were more abundant opportunities for the company to apply its nonrecourse, project finance model to the development of contracted generating facilities. In addition, foreign governments often provided incentives to attract foreign direct investment in infrastructure projects like power plants.The willingness of international development banks to invest alongside AES in volatile parts of the world helped mitigate the risk of expropriation, and the increased breadth of the global financial markets provided greater access to capital. AES initiated its international expansion in 1991–1992 with the purchase of two plants in Northern Ireland. The following year, AES began what would become a massive expansion into Latin America with 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 Chinese development projects. As the pace of deregulation quickened around the world, AES was presented with an abundant supply of capital and a wealth of opportunities for investments in energy-related businesses, some of which were more complex than AES’ portfolio of contract generation projects. In addition to expanding its line of business profile, it continued its geographic expansion and between 1996 and 1998 the company acquired several large utility companies in Brazil, El Salvador, and Argentina.By this time the company was spending an estimated 80%–85% of its capital investment overseas 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. 2 1 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 AES 204-109 AES in 2002 By 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 four separate lines of business: Contract Generation, Competitive Supply, Large Utilities, and Growth Distribution. 3 Contract generation In 2002, AES’s Contract Generation business accounted for approximately 29% of AES revenues and consisted of generation facilities, which sold electricity under long-term (five years or longer) contracts.The term of the contracts allowed AES to limit its exposure to volatility in electricity prices. The resulting stable production requirements enabled AES to accurately predict supply needs and enter into similarly long-term agreements for coal, natural gas, and fuel oil, thereby limiting its exposure to fuel price volatility. Facilities varied considerably in size, with plants as small as the 26 MW Xiangci-Cili hydro plant in China to the enormous 10-plant 2,650 MW Tiete hydro complex in Brazil. Competitive supply Accounting for 21% of AES revenues, the Competitive Supply line of business old electricity directly to wholesale and retail customers in competitive markets using shorter-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, oil and other raw materials. AES’s margin in U. S. dollars was influenced by a host of factors including weather conditions, competition, changes in market regulations, interest rate and foreign exchange fluctuations, and availability and price of emissions credits.Such price volatility had recently damaged several Competitive Supply businesses including the Drax plant in the U. K. , the largest plant in AES’s Competitive Supply fleet. 4 Large utilities By the end of 2002, the Large Utility business included only three major utilities, 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. La Electricidad de Caracas in Venezuela (EDC). These utilities combined generation, transmission and distribution capabilities and were subject to local government regulation and price setting.All three enjoyed regional monopolies and in total accounted for 36% of AES revenues. U. S. energy regulations 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 distribution Growth Distribution businesses offered AES significant potential growth due to their location in developing markets where the demand for electricity was expected to grow at considerably faster rates than in developed countries.However, these businesses also faced notable risks related to operating difficulties, less stable governments, and regulatory regimes, and differing cultural norms regarding basic principles such as payment conventions and safety regulations. 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 Annual Report (Arlington: AES Corporation, 2002) and AES Corporation, 2002 Annual Report (Arlington: AES Corporation, 2003). Energy companies typically refer to generation companies not as members of a “portfolio” but members of a “fleet. ” 3 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES Recent Difficulties AES’s placement in foreign markets as well as poor performance at several new U. S. businesses nearly crippled the company during the global economic slowdown that began in 2001. AES’s market 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 (see Exhibit 4).Wall Street began to question the company’s ability to weather the storm, and one analyst wrote, “It is clear that AES’s current stock price is reflecting the scenario that the company will not survive. ”5 The collapse of the stock price and the subsequent $3. 5 billion loss that included a substantial write-off in 2002 were brought on by several factors, the effect of which was amplified by AES’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 protect itself against.Currency Devaluations During 2001, a political and economic crisis in Argentina brought about a significant devaluation of most South American currencies against the U. S. dollar. In December, the newly elected government abandoned the country’s fixed dollar-to-Argentine-peso exchange rate (1:1) and converted 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 at a rate of 3. 32 pesos to the U. S. dollar and had been as high as 3. pesos. 7 The currencies in Brazil and Venezuela—equally important markets for AES—followed suit, with the Brazilian Real and the Venezuelan Bolivar each depreciating approximately 50% against the U. S. dollar during the same period (see Exhibit 5). As a result, AES recorded foreign currency transaction losses of $456 million in 2002. Several of AES’s subsidiaries in South America defaulted on their debt and were forced to restructure. The debt was nonrecourse to the parent, AES Corporation, so AES was not obligated to service the subsidiary debt.However, the parent company did suffer from cash flow shortfalls as a result of lower-than-expected dividends back from the subsidiaries. The impact of devaluation was increased when foreign businesses were paid in local currency but had obligations to repay debt denominated in U. S. dollars. Adverse Regulatory Changes During the late 1990s, the regulatory agencies in Brazil had failed to produce a market structure sufficiently attractive to encourage domestic construction of new generation assets. Demand exceeded 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 to the loss of sales volume, the decline of the Brazilian real against the dollar triggered a regulatory 5 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. nvestext. 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. 4 Globalizing the Cost of Capital and Capital Budgeting at AES 204-109 conflict concerning the applicable exchange rate for the real-to-dollar energy-cost pass-through provisions in AES’s contract.In effect, the government of Brazil required AES to purchase energy in dollars 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 on Eletropaulo, one of its major Brazilian businesses. Commodity Prices Decline A 2001 change in the regulatory regime in the U. K. also adversely impacted AES by increasing competition and reducing prices in its generation markets. That, along with an unusually warm winter in the U. K. brought wholesale electricity prices down approximately 30%. 9 These pressures caused several counterparties to default on their long-term purchase agreements. This counterpart risk, coupled with changes in the commodity markets, enhanced the financial pressure on AES facilities, and those that could not sell electricity above their marginal costs were taken off-line or shut down. Above and beyond the currency and regulatory difficulties at AES, the company was forced to take significant impairment charges on unprofitable or discontinued businesses.In 2002, the company took after-tax charges of $465 million on development and construction projects, $301 million on discontinued operations, and a massive $2. 3 billion in asset impairments associated with several large utility and generation businesses. 10 AES Reaction In response to the financial crisis, AES successfully refinanced $2. 1 billion of bank loans and debt securities. The refinancing arrangement came through the day before AES was to pay down $380 million of its outstanding debt. A group of 63 banks and investment funds agreed to provide $1. billion in new loans, and AES secured a two-year extension on another $500 million in notes due in 2002. 11 AES also secured agreements to sell a number of its assets. Total proceeds from the sales were expected to be approximately $819 million. Proceeds from sales in 2003 were expected to be approximately $310 million. 12 Capital Budgeting at AES Historically, capital budgeting at AES was fairly straightforward. When AES undertook primarily domestic contract generation projects where the risk of changes to input and output prices was minimal, a project finance framework was employed.Venerus explained that this framework consisted of a fairly simple set of rules—all nonrecourse debt was deemed good, the economics of a given project were evaluated at an equity discount rate for the dividends from the project, all 9 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 at 11 “AES Stock Shoots Up as Refinancing Keeps Bankruptcy at Bay,” The Washington Post, December 17, 2002, available from Factiva, http://www. factiva. om, accessed July 17, 2003. 12 AES Corporation, 2002 Annual Report, p. 36. 5 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES dividend flows were considered equally risky, and a 12% discount rate was used for all projects. In a world of domestic contract-generation projects where most risks could be hedged and businesses had similar capital structures, Venerus felt that this model worked fairly well. Beginning in the early 1990s, with AES’s international expansions, this model of capital budgeting was exported to projects overseas.Early on, the model worked well (as it had with the initial expansion in Northern Ireland), because this project had many of the characteristics of domestic opportunities. Venerus explained that the model became increasingly strained with the expansions in Brazil and Argentina because hedging key exposures such as regulatory or currency risk was not feasible. In addition, the financial structure of a going-concern business like a utility is notably different than that of a limited-lifespan asset like a generating facility. Nonetheless, in the absence of an academic or other alternative, the basic methodology remained intact.Another factor that created fundamental difficulties for transporting this model to overseas settings was the ever-increasing complexity in the financing of international operations. As one example of this, Venerus described how international operations would be evaluated and financed. Exhibit 6 illustrates the typical structure: subsidiary A and B were financed with debt that was nonrecourse to the parent. The subsidiaries’ creditors had claims on the hard assets at the power plants but not on any other AES affiliate or subsidiary.The local holding company, which often represented multiple subsidiaries, also borrowed to finance construction or acquisitions and received equity in the various subsidiaries it held. In addition, the holding company had debt that was nonrecourse to the parent, secured by dividends from the operating company. Finally, AES borrowed once again at the parent level in order to contribute equity dollars into holding companies and subsidiary projects. At the end of 2002, AES had $5. 8 billion in parent company (recourse) debt and $14. 2 billion in nonrecourse debt.Using this subsidiary structure, the parent company received cash flows in the form of dividends from each subsidiary (some of which were holding companies) and, because the structure of every investment 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 and holding company level effectively increased, and the subsidiaries struggled to service their foreign currency 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 coverage ratios 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 over 80% without any hedging of foreign exchange for any meaningful duration; this is where the model broke down. Venerus’s solution to the problem had to be consistent, transparent, and accessible. He knew his olution would have to account for changes in required returns due to leverage, incorporate some understanding of a project’s risk profile, potentially include country risks, and still provide values that were consistent with market behavior, including trading multiples. Globalizing the Cost of Capital To overhaul the capital budgeting process and evaluate each investment as a distinct opportunity with unique risks, Venerus knew he would have to calculate a cost of capital for each of the many diverse AES businesses. As a starting point, he considered the 15 representative projects shown in 6Globalizing the Cost of Capital and Capital Budgeting at AES 204-109 Exhibit 7a and, using the financial data in Exhibit 7b, he endeavored to derive a weighted average cost of capital (WACC) for each project using a standard methodology: WACC = E D re + rd (1 ? ? ) V V In order to calculate each WACC, Venerus knew he would have to measure all of the constituent parts for the 15 projects: the cost of debt, the target capital structure, the local country tax rates, and an appropriate cost of equity. In order to find the cost of equity, he would first have to estimate a reasonable equity beta.Venerus questioned whether the traditional CAPM model could help him calculate all of the necessary 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 market portfolio of equities. AES owned businesses in poorly integrated capital markets, so Venerus feared the use of a World CAPM might yield artificially low costs of capital due to the low (or in some cases negative) correlation of developing economies with the world market.For example, a world CAPM might generate the unreasonable result of a WACC lower than the U. S. risk-free rate due to its negative correlation with the world market portfolio. Similarly, Venerus did not advocate the use of a “Local CAPM” where beta measured the covariance of a project’s returns with a portfolio of local equities. Countries such as Tanzania or Georgia, 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 a high level, Venerus developed an approach with two parts. First, he calculated a cost of debt and cost of equity for each of the 15 projects using U. S. market data. Second, he added the difference between the yield on local government bonds and the yield on corresponding U. S. Treasury bonds to both the cost of debt and the cost of equity. Venerus believed that this difference or “sovereign spread” 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 Debt To estimate an equity beta for each project, Venerus first had the Corporate Analysis & Planning group take unlevered equity betas from comparable U. S. companies. They averaged the betas to yield one unlevered beta for each of the four lines of business. Since the equity betas reflected not only the market risk associated with each company, but also the differential effects of leverage, the group relevered the equity betas at indicative capital structures for each of the 15 projects using the following equation: levered = ? unlevered E V Using the relevered equity betas, Venerus had the group calculate the cost of equity for each project using the traditional CAPM equation: Cost of Equity = r f + ? rm ? r f ( ) 7 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES Finally, an appropriate cost of debt needed to be calculated. Given the significant regulatory and market changes impacting AES over the previous two years, Venerus decided not to use the historical cost of debt which might reflect market conditions that no longer existed.Instead, he attempted to estimate the return on debt demanded by investors given the cash flow risks of a given project. To do so, he applied the following equation: Cost of Debt = r f + Default Spread The estimation of “default spread” was based upon the observed relationship between EBIT coverage ratios for comparable energy companies and their cost of debt (shown in Exhibits 9a and 9b). The group estimated the appropriate EBIT coverage ratio for each project given its volatility of cash flows and leverage. Then, using the observed relationship, they assigned the commensurate cost of debt.For example, a project with a target EBIT coverage ratio of 3. 0x was assigned a default spread of approximately 300 bp. Adding the Sovereign Spread13 Before plugging the cost of equity and cost of debt into the WACC equation, Venerus wanted to account for country-specific market risk. He believed that risk could be captured in the difference between local government bond yields and the corresponding U. S. Treasury yields, or the “sovereign spread. ” Thus, he added the spreads found in Exhibit 10 to both the cost of equity and cost of debt and used those values to generate a WACC for each project.WACC Adjustments for Unsystematic Risk Venerus knew the above CAPM-based sovereign spread approach could provide AES with a useful WACC reflecting the systematic risk associated with each project according to its local market. However, was the approach reasonable in developing markets where access to capital was limited and information was less than perfect? Venerus believed that company-specific risk could not be easily diversified away in such markets. Moreover, AES—as an “investor” looking for potential projects—could not diversify in the same way a portfolio manager might diversify.Perhaps most importantly, Venerus was concerned that calculating expected cash flows by a probability-weighted average of various outcomes would be extremely difficult, if not impossible, to do accurately or consistently across the entire AES portfolio, even without the urgency of his present task. He felt budgeted cash flows would be more readily available. Thus, he believed the appropriate discount rate for AES businesses should account for some level of project-specific risk. Even if expected cash flows were available, Venerus felt that some degree of project-specific risk deserved consideration.Venerus illustrated his point with an example: Consider two hydro plants in Brazil that are identical in every respect except the hydrological risk of the rivers that feed them. Both plants have the same probability-weighted expected value cash flows. The hydrology of plant #1 produces cash flows that can vary by plus or minus 50% in a given year. The hydrology of plant #2 produces cash flows that can vary by plus or minus 10% in a given year. If both these plants are financed with 100% equity and pay no taxes, CAPM tells us that these plants are worth the same amount.That, to me, is unconvincing. 13 Also referred to as the “country spread model” or the “Goldman Model. ” See Jorge O. Mariscal and Rafaelina M. Lee, Goldman Sachs, “The Valuation of Mexican Stocks: An Extension of the Capital Asset Pricing Model,” 1993. 8 Globalizing the Cost of Capital and Capital Budgeting at AES 204-109 In order to compensate for this “undiversifiable project-specific risk,” the Corporate Analysis & Planning group created a risk scoring system designed to supplement the initial cost of capital. First, seven categories of project-level risk were identified.Each category was ranked and weighted according to AES’s ability to anticipate and mitigate certain risks. For example, because AES was unable to hedge changes in currencies in certain markets, “currency risk” received a high weight and rank. In contrast, AES felt it could control for most technical or plant-related problems and, as such, “operational risks” received a relatively low weight. See Exhibit 11 for the seven risks and examples for each. Second, projects were graded on their level of exposure to the seven categories of project risk.For each category, a project was assigned a grade between 0 (lowest exposure) and 3 (highest exposure). Next, the grades were multiplied by the respective weights and the seven categories added together to yield a single business-specific risk score. For example, Table A shows how the Lal Pir project, a contract generation business in Pakistan, might be assigned grades that translated into a businessspecific risk score of 1. 41. Table A Risk Score Calculation for Lal Pir Project Grade for Lal Pir 1 1 2 0 1 2 2 Risk Scores (grade x weight) 0. 035 0. 070 0. 210 0. 00 0. 180 0. 430 0. 500 1. 425 Categories of Risk Operational/Technical Counterparty Credit/Performance Regulatory Construction Commodity Currency Contractual Enforcement/Legal Sum of individual scores = business-specific risk score Source: Company document (actual assessments disguised). Weight 3. 5% 7. 0% 10. 5% 14. 5% 18. 0% 21. 5% 25. 0% Finally, the business-specific risk scores were used to calculate an adjustment to the initial cost of capital. The lowest business-specific risk scores (score = 0) received no adjustment to the calculated cost of capital.For projects with the highest business-specific risk scores (score = 3), the cost of capital was increased by 1500 bp. The relationships between business-specific risk scores and adjustments to the cost of capital were linear. Thus, a business-specific risk score of 2 would yield an adjustment to WACC of 1000 bp, and a business-specific risk score of 1 would yield an adjustment of 500 bp. 14 Preparing for the Board Venerus reviewed his methodology and considered the mandate he had received from the board. In order to refine the capital budgeting process at AES, he ad to devise a coherent and practical way to define cost of capital in all of AES’s international markets. In his own mind, he went over the steps 14 AES also considered a more complicated non-linear algorithm to generate the WACC adjustment from the business-specific risk score. 9 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES in his process: calculate the cost of equity and the cost of debt using U. S. market data, add the sovereign spread to each, calculate WACC using a target capital structure, and finally, add a business- specific risk adjustment to WACC.Still, questions lingered in his mind. He reviewed the project cash flows for the AES Lal Pir contract generation plant in Pakistan presented in Exhibit 12 as a way of gauging the effect of his new methodology. In doing so, he considered the differences in value created by each of the adjustments to the discount rate. Was his discount rate an actual representation of the risk associated with the project? Did it yield the correct value? More generally, did the sovereign spreads accurately capture the market risk specific to a given country?Had he used the appropriate risk categories and suitable weights to reflect AES’s appetite for risk? It was time for him to decide. Should he move forward with the addition of the business-specific risk score or should he simply use the traditional sovereign spread model? The board’s reaction was impossible to predict. What if the results were inconsistent with observable trading multiples? Would they accuse him of creating an over-complicated method, or would they applaud the new technique as a pragmatic way to calculate the cost of capital in an international context? 0 Globalizing the Cost of Capital and Capital Budgeting at AES 204-109 Exhibit 1 AES Consolidated Income Statement 2002 $4,317 4,315 8,632 (3,627) (3,086) (6,713) (112) (2,031) 312 219 (87) (1,600) (612) (456) (203) (2,651) (27) (34) (2,590) (573) (3,163) (346) $(3,509) 2001 $3,255 4,390 7,645 (2,416) (3,052) (5,468) (120) (131) (1,575) 189 116 (65) 18 (30) 176 755 206 103 446 (173) 273 $273 2000 $2,661 3,545 6,206 (2,093) (2,210) (4,303) (82) (79) (1,262) 201 51 (52) 143 (4) 475 1,294 368 120 806 (11) 795 $795Amounts in millions except per share figures Revenues Regulated Non-regulated Total revenues Cost of sales Regulated Non-regulated Total cost of sales SG&A expenses Severance and transaction costs Interest expense Interest income Other income Other expense (Loss) gain on sale of investments and asset impairment expense Goodwill impairment expense Foreign currency transaction loss Equity in pre-tax (loss) earnings of affiliates (Loss) income before income taxes and minority interest Income tax (benefit) expense Minority interest (income expense) (Loss) income from continuing operations Loss from operations of dicontinued businesses (net of income tax benefit of $90, $10 and $5, respectively) (Loss) income before cumulative effect of accounting change Cumulative effect of change in accounting principle (net of income tax benefit of $72) Net (loss) income BASIC (LOSS) EARNINGS PER SHARE (Loss) income from continuing operations Discontinued operations Cumulative effect of accounting change Basic (loss) earnings per share Source: $(4. 81) $(1. 05) $(0. 65) $(6. 51) $0. 84 $(0. 32) $$0. 52 $1. 67 $(0. 01) $$1. 66 AES Corporation, 2002 Annual Report (Arlington: AES Corporation, 2003). 11 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES Exhibit 2 AES Consolidated Balance Sheet 002 $780 211 1,264 384 218 1,492 4,349 194 23,050 (4,204) 18,846 1,403 8,984 33,776 2001 $802 215 1,137 468 215 1,855 4,692 3,031 21,127 (3,015) 18,112 2,433 8,544 36,812 2000 $950 1,297 1,566 569 1,193 209 5,784 3,122 21,874 (2,632) 19,242 2,248 2,642 33,038 1999 $669 164 936 307 327 184 2,587 1,575 14,210 (763) 13,447 1,904 1,367 20,880 1998 $491 35 383 119 155 71 1,254 1,933 6,029 (525) 5,504 1,490 600 10,781 Amounts in millions, as of December 31 ASSETS Cash & Equivalents Other Short-Term Investments Accounts Receivable Inventory Prepayments & Advances Other Current Assets Total Current Assets Long-Term Investments Property Plant & Equipment Accum Depr. & Amort. Property Plant & Equipment, Net Goodwill/Intangibles Other Long-Term Assets Total Assets LIABILITIES & SHAREHOLDERS’ EQUITY Accounts Payable Short-Term Debt Curr.Long-Term Debt and CLOs Other Current Liabilities Total Current Liabilities Long-Term Debt Total Long-Term Debt Deferred Taxes Other Long-Term Liabilities Total Liabilities Stockholder’s Equity Common Stock Additional Paid in Capital Retained Earnings Treasury Stock Other Equity Total Shareholders’ Equity Total Liabilities + Shareholders’ Equity Shares Outstanding Source: “AES Annual Balance Sheet,” http://www. onesource. com. December 1,139 3,341 2,031 6,511 17,684 17,684 981 8,941 34,117 727 2,449 1,752 4,928 17,406 17,406 627 8,312 31,273 743 2,462 1,834 5,039 17,382 17,382 1,863 3,212 27,496 381 1,216 973 2,570 12,136 12,136 1,787 1,750 18,243 215 1,413 348 1,976 5,791 5,791 268 952 8,987 6 5,312 (700) (4,959) (341) 33,776 558 2003, 5 5,225 2,809 (2,500) 5,539 36,812 533 available from 5 5,172 2,551 (507) (1,679) 5,542 33,038 509 OneSource 4 2,615 1,120 (1,102) 2,637 20,880 414 1,243 892 (343) 1,794 10,781 361 Services, Information 12 Globalizing the Cost of Capital and Capital Budgeting at AES 204-109 Exhibit 3 AES 2002 Revenues by Line of Business and Geographic Region Line of Business $1,180 $3,137 Large Utilities $1,837 Contract Generation Competitive Power Supply Growth Distribution $2,478 Geographic Region $1,568 $2,783 South America North America Europe/Africa $1,739 Carribean $2,091 Source: AES Corporation, 2002 Annual Report. 13 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES Exhibit 4 AES Stock Price History, March 1996 through December 2002 $80. 00 $70. 00 $60. 00 $50. 00 $40. 00 $30. 00 20. 00 $10. 00 $12/1/96 12/1/97 12/1/98 12/1/99 12/1/00 12/1/01 12/1/02 3/1/96 6/1/96 9/1/96 3/1/97 6/1/97 9/1/97 3/1/98 6/1/98 9/1/98 3/1/99 6/1/99 9/1/99 3/1/00 6/1/00 9/1/00 3/1/01 6/1/01 9/1/01 3/1/02 6/1/02 9/1/02 Source: Note: Created by casewriter. Stock prices adjusted for splits. 14 Globalizing the Cost of Capital and Capital Budgeting at AES 204-109 Exhibit 5 Selected South American Exchange Rates (2001–2002) (local currency per U. S. dollar) 4. 0 3. 5 3. 0 2. 5 2. 0 1. 5 1. 0 0. 5 Jan-01 May-01 Brazilian Real (left) Source: Bloomberg LP. 1,600 1,400 1,200 1,000 800 600 400 200 Sep-01 Jan-02 May-02 Sep-02 Venezuelan Bolivar (right)Argentine Peso (left) 15 204-109 -16- Exhibit 6 Typical Structure of an AES Investment AES Parent Corporation Assets Liabilities US Bank Debt Equity subsidiary Equity holding co. Corporate Debt Local AES Holding Company Assets Liabilities $-denominated debt Equity subsidiary (non-recourse to parent) AES Subsidiary A Assets Liabilities Fossil fuel power plant $-denominated debt (non-recourse to parent) AES Subsidiary B Assets Liabilities Hyrdo power plant $-denominated debt (non-recourse to parent) Source: Company documents and casewriter analysis. 204-109 -17- Exhibit 7a Risk Scores AES Project Data Line of Project Description Spread 3. 57% 8. 3% 3 3 Spread 300 MW gas fired combined cycle plant currently under construction 30 km east of Santo Domingo 123 MW hydroelectric power plant located on the San Juan river in western Argentina Largest coal-fired power station in western Europe. It can produce enough electricity – about 4000 MW- to meet the needs of approximately four million people Distribution company that serves a population of 14 million in Sao Paulo 277 MW fossil fuel plant located in Tocopilla, 1500 km north of Santiago 360 MW gas turbine facility located 25 kilometers southeast of Dhaka, capital of Bangladesh 600 MW coal fired power plant 337 MW coal fired power plant 210 MW Oil-fired facility supplying the capital city of Santo Domingo Joint Venture with the Government of Orissa. Two 210 MW P. C. oal-fired units Oil fired 140 MW cogeneration facility – under contracts of up to 10 years, electricity, steam, compressed air, dematerialized water and nitrogen to three chemical facilities adjacent to the plant 832 MW natural gas-fired plant Distribution Company serving 380,000 customers Distribution Company serving Tbilisi, the capital of Georgia. 600 MW gas-fired combined cycle power plant 7. 9% 25. 0% 23% 35. 1% 28. 7% 25% 32. 9% 0. 0% 33. 3% 2. 5x 17. 0% 35. 2% 2. 5x 4. 34% 34. 0% 30. 0% 3. 5x 2. 89% 0. 0% 29. 5% 3. 0x 3. 57% 0. 00% 35. 0% 40. 8% 3. 0x 3. 57% 16. 25% 3 25. 0% 35. 1% 3. 0x Tax Rate Debt to Cap. Coverage EBIT Default Sovereign Construction Operation/ Technical Regulatory Currency Counterparty Contract enf. / Legal 3 3 3 3 2 2 2 2 2 2 3 3 1 3 1 1 1 1 1 1 2 2 1 2 2 1 2 3 3 3 3 3 2 3 2 1 3 3 1 2 1 3 3 3 3 3 3 2 3 Andres Dominican Republic CG Caracoles Argentina CS 2 DraxUnited Kingdom CS – 2 Eletropaulo Brazil LU 8. 93% – 1 Gener Chile CG 1. 73% – – Haripur Bangladesh CG 4. 34% 5. 23% 2 – Kelvin South Africa CG 2. 5x 3. 0x 4. 0x 4. 34% 3. 57% 1. 85% 3. 14% 9. 90% 8. 93% 1 – 1 3 Lal Pir Pakistan CG Los Mina Dominican Republic CG OPGC India CG 30. 4% 3. 0x 3. 57% 3. 60% – 1 Ottana Italy CS 35. 0% 42. 5% 2. 5x 4. 34% 0. 14% – – Red Oak USA CG 37. 5% 30. 0% 20. 0% 34. 0% 39. 5% 36. 5% 26. 1% 32. 2% 3. 0x 2. 5x 4. 0x 4. 0x 3. 57% 3. 57% 1. 85% 1. 85% 0. 00% 9. 98% 9. 98% 8. 93% – 2 2 – Rivnoblenergo Ukraine GD Telasi Georgia GD Uruguaiana Brazil CG Source: Company document. Project descriptions taken from http://www. aes. om/businesses/default. asp. Commodity 3 1 3 2 2 1 1 3 2 3 2 Business / Project Country Business 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES Exhibit 7b AES Selected Financial Data Select Financial Information 10-Year U. S. Treasury Bond U. S. Risk Premium Unlevered Equity Betas by Line of Business Contract Generation Large Utility Growth Distribution Competitive Supply 4. 5% 7. 00% 0. 25 0. 25 0. 25 0. 50 Source: Company document. Exhibit 8 Summary of WACC Calculations for AES Step 1. Calculate unlevered equity beta. 2. Relever equity betas at target capital structure. 3. Calculate cost of equity for each AES business.Required Information • • • • • Betas at comparable U. S. companies Target capitalization ratios Risk-free rate Equity risk premium Relevered equity beta Risk-free rate Default spread Approach Unlever and average equity betas for comparables in each AES line of business Estimated by project using cash flows to calculate desired EBIT coverage 10-Year U. S. Treasury Note Long-term avg. difference between S&P 500 and U. S. Treasuries 10-Year U. S. Treasury Note Observed relationship between EBIT coverage ratios for comparable companies and their costs of debt The difference between local government dollardenominated bond yields and the corresponding U. S.Treasury Note 4. Calculate the cost of debt. • • 5. Add country specific risk to the cost of debt and cost of equity. • Local sovereign spread Source: Company document and casewriter analysis. 18 Globalizing the Cost of Capital and Capital Budgeting at AES 204-109 Exhibit 9a EBIT Coverage Ratios and Default Spreads 25. 0x 12. 0% 20. 0x 10. 0% 8. 0% 15. 0x 6. 0% 10. 0x 4. 0% 5. 0x 2. 0% Baa1 Baa2 Baa3 Caa1 Caa2 Caa3 Ba1 Ba2 Ba3 B1 B2 Aa1 Aa2 Aa3 A1 A2 Aaa A3 B3 – EBIT Coverage Ratio Source: Company documents. Default Spread 19 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES Exhibit 9b EBIT Coverage Ratios and Default SpreadsCredit Rating Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa1 Caa2 Caa3 EBIT Coverage Ratio 21. 1x 15. 1x 10. 9x 8. 1x 6. 3x 5. 2x 4. 6x 4. 2x 3. 9x 3. 6x 3. 2x 2. 6x 1. 9x 1. 0x 0. 8x 0. 6x 0. 4x 0. 1x 0. 1x Default Spread 0. 2% 0. 3% 0. 4% 0. 6% 0. 7% 0. 9% 1. 2% 1. 5% 1. 9% 2. 3% 2. 9% 3. 6% 4. 3% 5. 2% 6. 2% 7. 4% 8. 6% 10. 0% 11. 4% Source: Company documents. 20 Globalizing the Cost of Capital and Capital Budgeting at AES 204-109 Exhibit 10 Credit Ratings and Sovereign Spreads Used by AES US (AAA) Australia (AAA) Bahamas (n/a) Canada (AAA) UK (AAA) Italy (AAA) Spain (AAA) Netherlands (AAA) Hungary (A-) Chile (A-) Qatar (A-)Czech Republic (A-) Mexico (BBB) China (BBB) Oman (BBB) South Africa (BBB-) India (BB) Bangladesh (n/a) Sri Lanka (n/a) El Salvador (BB+) Kazakhstan (BB) Panama (BB) Brazil (BB) Dominican Republic (BB-) Bolivia (B) Georgia (n/a) Pakistan (B) Ukraine (B) Venezuela (CCC+) Argentina (D) Cameroon (n/a) Nigeria (n/a) Tanzania (n/a) Uganda (n/a) 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% Source: Company document; Standard and Poor’s and Lehman Brothers. 21 204-109 Globalizing the Cost of Capital and Capital Budgeting at AES Exhibit 11 Project Specific Risk Categories and Weightings Risk Category Operational Example An AES plant may fail to operate at capacity or fail to produce sufficient electricity to meet contractual obligations.AES has offtake agreements that—like futures and other derivative instruments—require credit; the counterparty may either fail to post additional collateral as required or fail to pay. Contract settlement processes in a foreign country may change after AES has made investments in a generation facility. Regulatory agencies may choose not to adjust rates for a utility after inflation goes up or other market dynamics change. Construction of a specific plant is complete but the plant may not perform they way it is supposed to (the heat rate is too high, the output too low, the availability too low, etc. ). Prices of coal, oil, or other fuels may spike.