When Calpine last spoke to Project Finance, it was on the back of borrowing close to $3.5 million using an untested revolving credit structure. In a November 2000 interview, its CFO, Bob Kelly, declared that "the old recourse model is dead", and said that Calpine had been able to commoditise the business of building power plants. In meetings, bankers that mentioned the phrase 'project finance' were fined, and loved Calpine for it. One bank called them ?a whole new kind of power company.?

That was then, this is now, and while four years on Calpine may not be the largest generator in the US, it is still standing. In that space of time, the only sponsor to copy the construction finance revolving credit template, NRG Energy, has entered and left Chapter 11. Calpine has resorted to a number of high-priced refinancings, and structured transactions, some of them involving contracted power.

This activity has had one unintended effect - it has made Calpine the number one sponsor of power project finance deals over the last twelve months, as ranked by Dealogic ProjectWare. According to the data provider, in the year leading up to 30 June, Calpine sponsored eight deals, a total of $5.42 billion, and accounted for 11% of the global power project finance market.

In the intervening period it has perfected such structures as contract monetization, the use of hedging technology in non-recourse finance, and the use of a monoline in US power. The Calpine Power Income Fund in Canada has been one of the few Canadian buyers of US assets. And it has been one of the few non-renewable developers regularly calling on such European project finance banks as remain in the US.

But when asked if he's mellowed on project finance, Kelly is adamant: "I'm still down on the product, probably more so than ever before." Moreover, Kelly still has little time for the mindset of project finance lending, saying, "banks still want to intervene, but we had the ability to survive because we didn't suffer under project finance constraints."

"We have been successful in attracting new contracts - between 15 and 20 years - and we've turned to the lenders to finance those. They're still the only source of development finance." Calpine has financed Blue Spruce, Rocky Mountain and Riverside Energy Centers in the bank market in the last two years (see www.projectfinancemagazine.com for details).

Term B temptations

But it does not see the bank market as long-term partner, and is anxious to refinance as soon as possible. Calpine, like most of its peers, has been a large user of Term B debt, which accounts for the bulk of its non-recourse borrowing for the last twelve months. Calpine's first visit to the market was in July 2003, when it closed $3.3 billion in secured notes, of which some were B loans. This debt, while secured over certain specific assets, was raised at the corporate level.

Term B debt has become a crucial part of the non-recourse debt product range, but had its roots in the above corporate borrowings, and made its name as the market to which distressed corporates could turn. Its investors are a mix of loan funds, hedge funds, and other non-bank investors that require a liquid market, a rating, and security over assets.

Some sponsors see B loans as high-priced debt from a lending community that is constantly shifting and has little at stake in a borrower's future. But it is open to all borrowers - while a project needs a rating, it does not have to be a good one. Calpine went from using the debt as a short-term fix to viewing it as a long-term solution.

The first appearance at a project level was for the refinancing of the 1999 construction revolver, which closed in August 2003. This $750 million deal, led by Goldman Sachs, also used a hedge that guaranteed a certain spark spread for the Calpine Construction Finance Company's output. Goldman Sachs' J Aron commodities division was the counterparty (for more details, see Project Finance, September 2003).

At the time, Calpine maintained that it could have rolled over the debt in the project finance market without too much difficulty but liked the B loan lenders' terms. If this is true, then subsequent financings bear it out. A further refinancing, for the CalGen, or renamed CCFCII, portfolio followed in March 2004, this time for $2.4 billion. Morgan Stanley acted as bookrunner and counterparty for the spark spread hedge, after an attempt to use Deutsche for a vanilla debt issue failed.

The two construction portfolios, bereft of long-term contracts (beyond those with Calpine) may have been natural fits in the Term B market. Then Calpine announced on 13 May 2004 that it would refinance the Riverside and Rocky Mountain plants in the institutional loan market, three weeks before the plants came online and a mere three months after closing the Rocky Mountain project deal. The $661.5 million B loan, led by Credit Suisse First Boston, closed on 24 June.

Riverside, a 600MW gas-fired plant located in Beloit, Michigan, raised $230 million from Credit Lyonnais, CoBank, Bayerische Landesbank, HypoVereinsbank and Nord/LB in September 2003, and Rocky Mountain, a 600MW plant in Weld County, Colorado, raised $250 million from Credit Lyonnais, CoBank, DZ Bank, HSH-Nordbank, HVB and GMAC. The bank deals offered all-in margins of roughly 485bp over Libor, while the Term B deal priced at 425bp but was sold at a 0.5% discount.

Kelly, however, certainly doesn't miss the banks: "I think it's a breath of fresh air to be dealing with people without regulatory constraints, that don't care about ratings. We'll have gone from 10 banks on CCFC I to zero in September. I don't think that banks are thinking on a commercial basis, they've got rating agencies to care about." As far as the extra cost of term B debt is concerned, Kelly is happy to pay the premium: "It's cheap insurance, and they like the power business. I've said all along that power has been a good business for 100 years, and that someone would finance it if the banks wouldn't."

Contract contradictions

The second main thrust of Calpine's financing activity has been in raising cash against power contracts, particularly those with the California Department of Water Resources (CDWR). Calpine, based in San Jose, can claim to be a local producer, but still signed lucrative contracts for reliable power with the state in the aftermath of the 2000 price spikes and blackouts. Calpine renegotiated the contracts, which still periodically attract the ire of local politicians, but still earns attractive margins.

Or earned, since as part of the capital raising process of 2002 and 2003, Calpine embarked on a debt financing that securitized the margins between the price at which these were struck and the price that an investment bank's trading operations would offer for this power. The Power Contract Financing deal, led and backed by Morgan Stanley, raised $802 million from a CDWR contract.

The deal, which closed in August 2003, borrowed from an earlier deal by Morgan Stanley for El Paso Energy (search PCF, and UCF, on www.projectfinancemagazine.com for details). Morgan Stanley Capital Group can, in part thanks to its A+ rating, buy power for $20.10 per MWh cheaper than Calpine's contract with the CDWR. Calpine gets the cash upfront, but forgoes any of the profit that might be earned in excess of this margin.

Calpine followed this with a $301 million bond issue for the Gilroy project, a monoline-wrapped deal that used a CDWR contract that was transferred to a vehicle owning a 525MW peaker plant. To enable this deal to reach investment grade, Calpine placed a majority equity interest to third-party investors through CIT. The eight-year bonds priced at 4%.

In May, it sold a power contract to Morgan Stanley Capital Group for $101 million, and retired $79 million in non-recourse debt. The contract related to its 118MW Parlin and 58MW Newark plants, both in New Jersey. And it has signed power purchase agreements with such diverse players as the Long Island Power Authority (LIPA), Dow chemical, and the Snapping Shoals Energy Membership Corporation.

These deals are far removed from a merchant-based vision of Calpine's operations, but Kelly stands by the contract deals, saying, "we start from a business model whereby we optimize power, gas and capital on an ongoing basis. With PCF we hedged our cost of power and bought back debt that was priced at 12%."

Nevertheless, Kelly views the arrival of the investment banks to the power business as a mixed blessing. "What we had been doing was marrying up commodity risk and capital, and what investment banks do is package commodity risk in the capital markets. We based CCFC I and II - $5 billion in total - on their products. They have the skills and the cost of capital to make it work, and whoever can make it work will make money. It doesn't surprise me that TXU and CSFB are getting together. But the banks could be important competitors."

Cross-border corrections

But the most startling recent development for Calpine has been its halting first steps outside North America. Its first purchase was Saltend, a 1,200MW plant located near Hull in the UK, from Entergy. The price was $562 million, and much of the plant's output is subject to a 15-year contract with BP. The deal went through in August 2001, and since then the UK market has hit bottom and, tentatively, recovered.

Kelly defends the purchase, saying that Saltend is a fine plant, and the UK is a good market, although he acknowledges that development is tricky. "The oligopoly is winning, so we're not able to invest in price signals. As for buying plants there, we're disappointed that we ran out of cash in the US at a time when the market hit bottom." Despite Kelly's hints that Calpine has been approached to develop plants elsewhere in Europe, Saltend remains its only asset outside of the Americas.

But Calpine has begun to invest in Mexico, forming a partnership with Mitsui of Japan to develop the Valladolid III project. Calpine swapped a turbine from its inventory for a 45% stake in the project, a development that ratings agencies viewed as largely positive, since the plant has a 25-year contract with Mexico's state-owned power company, the CFE.

The structure of the deal, and the fact that Calpine has taken some write-downs against its turbine purchases - it had at one time the largest order book in the industry - have led many to see this investment as largely an opportunistic turbine play. Says Kelly, "it isn't a turbine play or we'd have been doing it for years. There's deregulation and strong demand for power, we can marry low-cost financing from Japanese sources, a tolling agreement and a low-cost construction contract with Calpine-Mitsui."

Calpine's search for low-cost financing has taken it to Canada, where it has been a pioneering user of Canadian income trusts to buy Canadian and, more recently, US, assets. The Calpine Power Income Fund, of which the US parent is a manager and has a 30% subordinated interest, owns three plants - 225MW Island in British Columbia, 300MW Calgary in Alberta, and a 50% interest in a participating loan to the 50MW Whitby plant in Ontario.

But it recently completed the purchase of the King City plant in California, as part of a restructuring of King City's leveraged lease. The purchase of the plant was largely through a C$100 million ($120 million) unit issue underwritten by Scotia Capital, TD Securities, National Bank Financial, CIBC World Markets, Canaccord Capital, Dundee Securities, FirstEnergy Capital, and Raymond James, as well as a C$53.4 million promissory note from Calpine Power Canada to smooth out distributions. The fund is now in discussions to buy Calpine's Deer Park plant in Texas.

How much further the structure can be adapted is a difficult question - such funds had been tipped as a cost-efficient way to purchase distressed power assets. But the Canadian funds are likely to be restricted in terms of the number of US assets they can buy, and, as Kelly notes, the US version, the Income Deposit Security, is the subject to a number of legal obstacles.

He says that Calpine has thought of using the structure in the UK, and adds, "in general I'm a proponent of using [such structures] on mature portfolios - they give cash back to the equity, and they're low cost. Eventually Calpine will be long on cash, and by 2007 or 2008 we'll start paying that out." Calpine is unlikely to spend it on renewable development, since, geothermal aside, "we don't think it is a good idea to buy power that is not yet economic simply because it is green."

Back to the banks

But Calpine is not yet done with the project market, Kelly's views aside. It has at least five plants with contracts that Calpine will probably finance through to operations in the bank market. The Fox Energy Center, a 235MW facility with a contract with Wisconsin Public Service that Calpine bought from MidAmerican Energy, is believed to be a Union Bank of California/Bayerische mandate, although market rumour suggests financial close has been delayed.

Kelly also mentions potential financings for Otay Mesa, a San Diego plant that has a 10-year 570MW contract with San Diego Gas & Electric, and that it bought from PG&E in 2001; Dow Freeport, with a 25-year 250MW Dow contract; Bethpage 3, an expansion of a Long Island asset with the 20-year, 80MW LIPA contract; and Mankato, a 365MW plant with a 250MW contract with Xcel Energy.

And Kelly is very proud of his following in the bank market, noting that "as a construction and operations and maintenance manager we're one of the best rated in the market, we're just not rated by the rating agencies. Banks' decisions have more to do with what agencies say than what they see at our plants." Nevertheless, "Scotia, ING, Bayerische, TD, Calyon and Union Bank of California, which have been with us from the beginning, and Goldman and Morgan Stanley on the investment banking side," are his keenest bank supporters. It looks like Calpine is not quite ready to let go.