By Jeff Bodington

Sales of merchant power plants have more than quadrupled so far in 2004. The backlog of merchant plants for sale is being worked down, and buyers and sellers are closing the spreads that led to much talk but few actual sales. Some participants have questioned whether or not the very low price Duke received for its portfolio of merchant plants in the southeastern United States is a benchmark for pricing other merchant plants that are still on the market.

This article summarizes the merchant sales activity and puts the Duke deal into perspective.

In brief, while the Duke sale now stands with sales of the PG&E National Energy Group portfolio for low value, none of these sales should set a mark for the value of all merchants.

Recent Sales

More than 18 transactions involving more than 100 merchant power projects are now pending or have closed. Net installed capacity sold totals over 14,350 megawatts. Reliant’s sale of the Orion portfolio to Brascan included 72 hydroelectric projects and, without this transaction, the total merchant sales to date would be 30 projects with an aggregate net capacity of approximately 14,000 megawatts. Nearly all of these are natural gas-fired combustion turbine-based projects constructed when the merchant business model appeared to thrive. These sales include 30% to 50% of the total merchant capacity built during the last five years. While the merchant sector is far from sold off or abandoned, these sales show that substantial progress has been made.

Sellers are primarily the developers and lenders who invested heavily in merchant generation. Developers are responding to pressure from Wall Street to repair their balance sheets. Lenders are responding to pressure from both Wall Street and the federal Office of the Comptroller of Currency, or “OCC.” In at least a few cases, the OCC is forcing writedowns that make sales a less-painful alternative. Buyers are diverse; utilities, independent power companies and private equity funds. Utilities of various types account for most of the transactions. Investor-owned utilities, municipal utilities and other entities whose ratepayers will be at risk if the new owners cannot make a go of the plants account for more than 70% of the sales by number of transactions, 45% of the sales by generating capacity and 50% by the value exchanged. Among independent power companies, Calpine has been both a buyer and seller.

Private equity firms have spent much time looking at merchant acquisitions; however, few have become buyers. The Duke sale of its merchant portfolio in the southeastern US is an example of an unusual closing. Most private equity buyers have been more successful in pursing generation that involves less risk than merchant operations: either regulated utilities or non-merchant independent power. Texas Pacific Group formed Oregon Electric Utility Company to pursue acquiring Portland General Electric from Enron. Kohlberg Kravis Roberts & Company, Blackstone Group, Texas Pacific Group and Hellman & Friedman have joined to purchase the former Reliant unit Texas Genco. AIG, Algonquin Power, Arclight Capital Partners, Goldman Sachs, Harbert and many others pursue projects whose revenues are secured by long-term contracts.

Focusing on Duke, Duke Energy North America developed numerous merchant projects and had a portfolio of eight projects in four states located within an area of the United States called the “Southeast Electric Reliability Council,” or “SERC.” All eight plants are natural gas-fired and most of their 5,280-megawatt combined capacity went into service during 2002. As Duke’s heavy investment in merchant generation failed to yield current earnings, asset sales began. Lackluster bids forced Duke to write down the value of the plants three times, and the portfolio was ultimately sold to KGen Partners. KGen is owned by MatlinPatterson, a firm that focuses on distressed debt and that was founded by distressed-debt specialists David Matlin, Mark Patterson and Lap Chan. The three founders were with Credit Suisse First Boston. Their first fund was $2.2 billion, and the second recently limited funding at $1.66 billion. The firm has invested in WorldCom/MCI, Huntsman, Oxford Automotive and now electric power by buying debt of NRG Energy and purchasing the Duke projects.

Value of Duke Southeast

Merchant sales have been painful experiences for sellers. While $/kW is a very rough guide to value, the range of prices is approximately $90/kW to $790/kW. The average for gas-fired projects is approximately $225/kW. High-value projects tend to be combined-cycle facilities with relatively low heat rates purchased by ratepayer-at-risk entities. Purchases by Avista, GenTex, Puget Sound Energy and the City of Brownsville are examples. Low-value projects tend to be combustion turbine peakers with heat rates over 11,000 Btu/kWh in regions such as SERC and ECAR (Delaware, Maryland, parts of Pennsylvania, West Virginia, Tennessee, Kentucky, Ohio, Indiana and Michigan) that have ample reserve margins and substantial coal and nuclear generation. The Duke GE 7EAs located in Georgia, Kentucky and Mississippi purchased by KGen are an example. Details of the Duke portfolio show more about why its value is not a benchmark for merchants in general.

Duke’s southeastern merchant portfolio included three combined-cycle projects and five peakers. Combined-cycle projects accounted for 2,360 megawatts of the 5,280 megawatts in total. The results of a multiple-round auction were announced on May 4, 2004, and the acquisition with KGen closed four months later on August 5.

The transaction had three key components: cash, a high-yield note and a power purchase agreement. Total cash was $425 million. Regarding the high-yield note, Duke holds a $50 million receivable from KGen. This note bears interest at LIBOR plus 14.5% and is secured by a fourth lien on KGen’s owner. Interest compounds quarterly, and both interest and principal are due in a balloon payment after 7.5 years. The transaction included a seven-year power sales agreement between KGen and Georgia Power for output from one of the plants: the Murray combined-cycle facility. Duke operates this project under a long-term operations and maintenance agreement. As part of this agreement, Duke arranged a $120 million letter of credit to secure the obligations of KGen to Georgia Power, and KGen has an obligation to reimburse Duke for LC-related expenses and drawings. While these details show that the transaction was more complex than the often-quoted figure of $475 million, they also show that there may be additional value to each party embedded in terms of the transaction. The high rate on the note may add value for Duke. The LC obligation impairs Duke’s balance sheet, and the arrangements concerning Murray have benefits and costs for both parties.

The economic logic of the price lies in both the characteristics of the projects and the regional market for power. While the projects are new and efficient, it is the nature of the regional market for electricity that led to a relatively low price for the merchant assets. Key aspects of these factors appear in the graph below.

The top line in the graph shows the number of days during the last year on which the average daily wholesale electric power prices in the Entergy region of SERC exceeded the price on the left axis. The curve shows only weekdays; low weekend loads mean that merchants are usually idle when regional capacity margins are high. For example, this “price duration curve” shows that the price was at least $20/MWh at all times and at least $50/MWh for about 50 days of the year. The curve shows the potential gross revenue available to a merchant. Compared to many other regions, the graph shows that potential gross revenue in SERC is not substantial. Non-gas generation dominates supply, and much new capacity is under construction. Seventy five percent of the 182,000 megawatts in generating capacity available in SERC during 2004 is from coal, nuclear, hydro and pumped storage. Further, while the capacity margin reported on the Forms 411 filed with the federal Energy Information Agency is currently approximately 15%, the margin including projects with signed interconnection agreements is forecasted by SERC to exceed 30% through approximately 2010.

The top line in the graph addresses gross revenue. The bottom two lines address net revenue after fuel costs. This net, “energy operating margin” is what funds are available to cover non-fuel operating costs and then yield some capital value. The bottom two lines show that the three combined-cycle combustion turbine projects in the Duke portfolio could operate profitably about 230 days of the year. However, the peaking combustion turbines, due to their higher heat rates and thus fuel costs, could never make money on a daily average basis. Profitable operations on only a few high-load afternoons may be feasible. Future increases in SERC’s increasing capacity reserve imply that near-term improvement in this situation is not likely.

The present value of the operating margin in the graph for the combined-cycle plants justifies a price of approximately $50/kW for the entire 5,280-megawatt portfolio purchased by KGen. The actual price of approximately $90/kW shows that the buyer expects some combination of the Murray contract, peak-hour operations for the combustion turbines, load growth in SERC and lower natural gas prices to add value. Holding these projects for response to future requests from regional utilities and adding steam cycles to some of the peakers are examples of additional potential sources of value. For other merchants in other regions, a similar calculation supports prices close to those at which projects actually trade.

In sum, this brief analysis shows that KGen did not make an obviously terrific buy, and what KGen paid does not mean that many other merchants are worth as little money as KGen paid.