Merchant Plants Start to Sell
Buyers and sellers of merchant power projects in the United States are beginning to overcome the many obstacles that have slowed sales of such projects during the last several years. Deals are nearing close and, pending various approvals, buyers are findings ways to manage market risk, and the gap between bidding and asking prices is closing.
Standing back, few sales during 2002 and 2003 involved projects with merchant risks. To date during 2004, several are pending and many are under negotiation. That said, most of these sales provide little or no market data on merchant plant values. Assignments to lenders, foreclosures, sales to contractors and contract buyouts imply little about the value of a project to an owner who must pay cash and take market risks.
A few arm’s-length merchant project transactions have now been announced that provide early data on what such projects may be worth.
The number of merchant projects and megawatts that are now and potentially will be for sale is unclear. Lenders have taken over 14 projects with a combined capacity of more than 12,000 megawatts, and the sale of only one of these has been announced. Adding those still under the control of owners who have declared bankruptcy brings the totals to over 25 projects and 19,000 megawatts. Adding further those owned by companies under some financial pressure or looking for a strategic exit who have announced their intention to sell brings the total for sale to more than 33,000 megawatts.
Public data on three pending transactions illustrate the varied history of merchant projects and show a wide range in potential value.
One sale announced during 2003 involves an “accidental merchant” with a storied history. Frederickson was developed during the early 1990s by Tenaska and was supported by a power sales contract with Bonneville Power Administration. BPA terminated the contract while the project was under construction, years of litigation and restructuring ensued, BPA became the owner, and EPCOR Power Development of Alberta ultimately purchased the project from BPA. EPCOR completed construction and operations began during September 2002. Last year, Puget Sound Energy announced its intention to purchase an interest in this project for $76.4 million, and PSE will contribute another approximately $4 million for upgrade costs. Closing is contingent upon timely approval of full cost recovery by the Washington State Utilities and Transportation Commission. Subject to that now-controversial approval, Puget Sound Energy’s ratepayers will bear the risks associated with the acquisition, power values and fuel costs. Ratepayers will pay a total of $584/kW for 137 megawatts of Frederickson’s capacity, and this is approximately 79% of actual original cost.
Mirant and CLECO began development of a peaker and combined cycle project in Perryville, Louisiana during the late 1990s. With the peaker in operation and the combined-cycle project under construction, Mirant sold its 50% interest in the project entity to its partner, CLECO. Construction was completed during June 2002; however, tolling contract and other difficulties plagued the projects, and the project entity, Perryville Energy Partners, filed for protection from creditors under chapter 11 of the US bankruptcy code. Then, during 2003, CLECO announced the sale of 100% of the 718-megawatt project to Entergy Louisiana for $170 million. The buyer plans to use 25% of the power for its own customers and to sell 75% under a new power purchase agreement with Entergy Gulf States. The transaction is pending, and Entergy awaits regulatory approval from the Louisiana commission to pass all costs through to ratepayers. A project that cost $451/kW was sold for 52% of actual cost on a net basis and, if approved, ratepayers will bear all risks.
Finally, Brazos Valley provides another example of merchant plant value. The project was developed by NRG Energy, and construction was well underway during 2002 when NRG became unable to meet its equity funding commitments. The lender group foreclosed during January 2003 and, following much evaluation and a restructuring of the group’s interests, the lenders funded the completion of construction during mid-2003. An extensive marketing effort and several false starts ultimately led to Calpine’s announcement during February this year that it would purchase the 570-megawatt project for $175 million, or $307/kW and approximately 68% of actual cost.
In contrast to these values, the average price paid for projects with contract-secured revenues during 2003 was approximately $750/kW. This average includes all technologies. The average for sales that involved primarily natural-gas-fired assets with contract-secured revenues was approximately $700/kW. This gas-only value is slightly below the average for all transactions due to the high prices paid for geothermal, wind and hydroelectric projects that have no fuel costs.
While $/kW figures are useful for comparison of a project’s change in value and broad price levels, they are at best a signpost to the value of an individual asset. The cost figures in the table below are estimates because some costs are closely guarded and not all companies record and report costs the same way. Further, these $/kW sales values do not imply that all merchant projects have values in the same range. Only a valuation method such as discounted cash flow can be tailored to an asset’s unique characteristics and then yield a well-supported, rational value. The high uncertainty associated with power values and fuel costs means that probabilistic analysis is often necessary and discounted cash flow provides a framework within which risks can be considered. Under these circumstances, rate-of-return requirements become fluid. A 10% after-tax return on equity may be acceptable for both an average-case project with contracted revenues and a conservative-case project that involves some merchant risk. As assumptions about performance for the merchant become more aggressive, return requirements can rise over 20%.
While each project is different and valuation demands careful analysis of many factors, the logic behind the prices reported in this article is evident after doing a few order-of-magnitude calculations. Assuming, for example, a risk-burdened 12% weighted average cost of capital and a 20-year time horizon, the present value of a net operating margin averaging $10/mWh is approximately $75. For a power project with an 80% capacity factor, that totals approximately $550/kW and is within the range of what a combined-cycle project costs to build. For a project that runs just half the year at rated capacity, the total is approximately $325/kW. This latter figure is consistent with forecasts of near-term gluts, thin margins and is within the range of what buyers appear willing to pay in troubled markets.
The three transactions reported in this article illustrate two important trends. First, the Puget and Entergy transactions show that regulated utilities are potential buyers of some of the unsold merchant projects. Dominion and Southern California Edison are further examples of regulated utilities that are purchasing either uncontracted assets or buying into projects with which they have contracts. (See related article, “FERC Restricts Power Plant Sales to Utilities,” starting on page 8 of this issue.) Second, merchant assets, for now, appear to be worth 50% to 70% of original constructed cost. Purchasing these assets and working to arrange power contracts or selling to regulated utilities are examples of ways to add substantial value.