Project lenders have found it difficult to fill their books for most of 2002. Whilst some bankers would say that their deals have been under far more intense scrutiny by credit committees, there has also been a marked decline in deal flow. The blame for this situation lies in large part, at the door of constrained sponsor balance sheets ? there is simply not enough equity available to put into a power project.

A number of corporates, however, are now looking at taking some of their smaller and more digestible assets off balance sheet. Peaker plants, for instance, are about the right size for a reasonably highly leveraged financing that requires smaller infusions of equity. More significantly, most have now abandoned attempting to finance these on a merchant basis.

Peakers are usually single-cycle gas-fired plants that can be fired up quickly and efficiently with a minimum of degradation. In contrast to a baseload facility, they will be idle for much of their economic lives, and when run constantly will degrade quickly. This happened to a number of plants running during the California crisis.

The timing of such a shift in the strategy of power producers is auspicious ? utilities are now interested in securing reliable sources of power from third parties. And the availability payments provide the stable revenue streams that lenders find attractive. Many utilities want to either capture the revenue from spikes in prices or use peaker plants to drive down spikes in their own region by bringing timely capacity online, but may lack the funds.

Merchant peakers never had a strong following in the finance community. Their earnings are highly volatile, at the mercy of weather and other factors of demand, and peakers are usually the preserve of utilities anxious to maintain reliability in power supply. During the merchant power boom years, a refusal to countenance building a peaker plant was one of the few examples of an envelope that would not be pushed.

There have been signs, however, that lenders searching for new assets that fit neither the baseload or merchant profile may be exchanging one type of exposure for another. In contrast to baseload assets ? often at the mercy of wholesale power prices but usually well-located ? peaker plants will have a solid power purchase agreement (PPA) in place but be dependent on the credit quality and existence of an offtakers.

To be sure, many peakers are well-sited and integral to the functioning of their respective grids. Part of the due diligence required for any project deal would be to work out the value of peaker assets without contracts attached ? and here the thinking is less clear.

At least one possible solution is to place merchant peaker assets in a portfolio, using the same principle by which genco financings went ahead during the merchant boom years ? that of diversity. The reasoning behind this theory is that while a peaker might spent a year with out being called on, the varied land mass in the United States should not provide universally mild weather conditions and pools free of plant outages.

Anecdotally, and given a summer dominated by reports of extreme weather events, this should be a viable proposition. Finding a way of assigning probabilities to this volatility is very difficult. But NRG Energy, currently struggling to keep its head above water following a collateral call on $1.3 billion in equity for its construction revolver, may have cracked the conundrum.

For its $325 million NRG Peaker Finance deal, it tried to find a capital market interested in taking on more exposure to such a troubled generator. Before this, it needed to find a way of presenting a portfolio of peaker plants as a viable business proposition. Here, the science of modeling and forecasting travels into uncharted territory.

NRG Peaker Finance contains five plants: Rockford I, Rockford II (in construction but commercial in June 2002), Bayou Cove (in construction with commercial operation scheduled for fall 2002), Big Cajun Peakers and Sterlington. The combined capacity of the plants is 1316MW. The first two plants are located in Illinois and will accompany existing LS Power and CogenAmerica acquisitions, whilst the last three will operate alongside NRG's South Central genco capacity, of which the Big Cajun facilities are the most prominent.

The plants are subject to PPAs for the first few years of their 15-year financing. Moreover, they are an important part of NRG's existing asset portfolio, with the stress on existing, since NRG will probably sell any assets for which it receives interest if the price is right. At present, however, the advantages of central dispatch and the ability of NRG to charge a premium for reliable power should increase the portfolio's financial health.

Rather than sell this story to a syndicate of project lenders already full up on NRG paper, the sponsor, advised by Goldman Sachs, decided to deal with a single lender ? the monoline XL Capital Assurance. Since both XL and Goldman, which also underwote the bonds, have been unwilling to provide details of what they consider a proprietary structure, so it is difficult to tell how the deal made it to an investment-grade shadow rating.

Part of the key, according to deal participants, is to move from a base case understanding of the portfolio's credit profile to a cumulative model that monetises volatility. It also appears to persuade XL, duty-bound to make the payments to bondholders come what may, to put up with some lean years with the promise of fat years ahead. There, are however, a few structural enhancements.

The most important of these is the funding of a debt service reserve by the revenues under the PPAs. And even once this has been funded, excess distributions to NRG may be clawed back by XL in a sustained period of poor power revenues. This scenario is not too far-fetched ? rumours from the southern power pool suggest that independent power producers may be facing curtailment notices from incumbent utilities.

It is this unpredictable way in which utilities can behave that scares the most hard-line anti-peaker bankers. This is the worry that an offtaker will not exist a few years down the line, say in the case of a bankruptcy caused by a venture into unregulated businesses. This threat already has antecedents ? Williams and Dynegy are on the end of tolling or power purchase agreements with AES, Tenaska and Southern Power, amongst others.

This lies outside of the statistical analysis that has been pioneered for peaker plants ? one that overlays the traditional stuff of market forecasters' reports with a probability-based analysis. Where a forecast prepared for a baseload plant, one expected to dispatch all of the time and subject largely to vagaries of prices, is a good candidate for a base-case analysis, peakers require probabilities to come into play.

Mark Griffiths of NRG Peaker's market consultants, Henwood Energy, says that ?peakers are a physical expression of an option ? or an insurance policy that you don't expect to use.? And probability-based analysis has a different set of rules, leading to what Griffiths calls an ?actuarial flavour, what we call stochastic analysis?. This stage of the analysis can determine when and how often a plant may run, and is in large part a product of historical analysis. Some of the standard questions for price analysis ? reserve margins, spark spread, interconnection, fuel mix ? can then be laid over this variety of outcomes to produce a model by which volatility can be measured.

The difficulty, as Griffiths concedes, is that it requires an actuarial frame of mind to take on that form of analysis. The monolines, at the front of which XL stands, are best placed to take on this risk, although how much of this XL took on (or at least underwrote before selling on) is difficult to quantify. But priced at 107bp over the three-month Libor, the cost looks attractive. The deal was swapped into floating rate with the assistance of Goldman's derivatives arm and sold on to banks as triple-A paper.

Another example of the extreme innovation that is the virtue of peaker deals is InterGen's Wildflower refinancing. InterGen had assembled a bridge loan through Deutsche Bank It consists of two Californian plants ? Larkspur (90MW), located in San Diego, and Indigo (135MW) located in Palm Springs. The plants' five turbines were installed as a prompt reaction to California governor Gray Davis' urgent call for more capacity at the start of 2000. The two plants are located near to major load pockets and interconnection points and were the first to be licensed under the California emergency siting process.

Both came online during the middle of 2001. Joining Deutsche on the $135 million bridge financing were Fortis Bank, NordLB, Citibank and Helaba. The new package required a new lead arranger after the abrupt and lamented departure of Deutsche ? in this case Citigroup. Likewise, Scotia Capital replaced NorldLB, which has little appetite for peaker paper.

The most important aspect of the deal is that it is at first to be (at least partially) backed by a power purchase with California's Department of Water Resources (CDWR). Trying to attain a reasonable level of comfort with the CDWR contracts, signed at haste (or, as Davis might say, under duress) following the power crisis, has been difficult. The expectations that the contracts might be repudiated or modified through political pressure are still high.

Wildflower was structured so that the two elements of the plants' offtake contracts ? one directly with the Shell affiliate Coral Energy, the other through Coral with the CDWR ? were linked to separate drawdown events. Of the final $110 million debt, $85 million is linked to a Coral credit, with a further $25 million conditional on a reliability agreement. The main hurdle here is the forthcoming issue of over $11 billion in bonds to capitalise the CDWR and provide it with enough cash to honour its obligations and remain ringfenced from the State government.

Now that these have received a rating, the conditions for drawdown will be approaching. But the threat of repudiation, exacerbated by some favourable rulings over market manipulation (for instance Ferc versus El Paso Gas) remains. Moreover the financing, while at the standard length for a mini-perm (a 2006 maturity), should retire a reasonable amount of debt before maturity, and features a bank-friendly cash-sweep. Nevertheless, sources at InterGen believe that a bond refinancing may be an option before maturity.

A final category of peaker deals consists of what might be called distressed project financings. These are assets that a sponsor may wish to keep on balance sheet because the cost of non-recourse debt in financial and operational flexibility is considerable. Corporate unsecured debt, however, is very hard to raise and very expensive. Raising debt off the credit of a strong PPA and giving up security to the banks is often the easiest way to raise cash fast.

Calpine has been an early convert to this way of raising cash. It has a $100 million financing out for the Blue Spruce plant, led by Credit Lyonnais. The deal is noteworthy for being Calpine's first pure project financing in recent years, a sharp contrast with its earlier preference for corporate level debt and construction finance revolvers.

Blue Spruce is a 300MW gas-fired plant located in an industrial park near Aurora, Denver, and sells its output to Public Service Company of Colorado under a ten-year tolling agreement. This agreement gives it a comfortable tail on the six-year financing agreement. The deal is around 65% leveraged and is presently in syndication. Sources at Credit Lyonnais aim to close the book-building process by the end of October.

Calpine has also secured a letter of intent from ING Bank regarding a forthcoming sale leaseback of some of its California peaker plants. These could raise up to $500 million, and are, like other similar plants, waiting on a solid outcome to the CDWR's rating.