Trusting Capacity Markets


Does the lack of long-term pricing undermine the financing of new power plants?

Fortnightly Magazine - December 2011

Capacity markets were implemented in a number of restructured power markets to meet resource adequacy requirements through a market-based solution. However, after several years of experience with capacity markets, their performance is questioned by many market participants.

While Brattle Group’s recent review of PJM’s capacity market, the Reliability Pricing Model or “RPM,” concludes that the market is performing very well, discussions with a number of stakeholders reveal considerable doubt that capacity markets can support new generation investments.1 The general concern is that prices are too uncertain because capacity auctions establish the price only for one year at a time, and buyers aren’t willing to sign long-term contracts bilaterally. However, different stakeholders also offer different perspectives.

State regulators in eastern PJM express the concern that after several years of operation and high prices, the capacity market hasn’t led to significant new construction of power plants. Regulators and a number of market participants also point to a lack of long-term contracting that could support the financing of new generation additions in eastern PJM. Some of the generation developers specifically note that the three-year forward capacity prices under PJM, which are locked in for only one delivery year at a time, can’t support the financing of new generation projects. They suggest that prices would need to be locked in for 10 years or more to support financing of new generation projects, a notion that’s echoed by some financial industry participants.2 They also stress that buyers are unwilling to enter into longer-term contracts.

Public power companies, such as cooperative utilities, indicate a strong interest in signing long-term contracts, but note that they’ve been unable to find willing suppliers. Stakeholders agree that long-term bilateral contracting currently isn’t available for more than the next three to five years.

The regulators’ and generators’ concern that an absence of long-term bilateral contracts undermines the financing of new plants seems inconsistent with public power stakeholders’ concern that suppliers have been generally unwilling to offer such long-term contracts. This apparent inconsistency, however, is explained largely by current market fundamentals and economic conditions that haven’t justified the addition of new generating capacity.

That means it’s premature to conclude that capacity markets need major changes to support new investment; merchant generation investment and longer-term contracting are likely to reemerge when market fundamentals support the addition of new generating resources.

The Role of Market Fundamentals

Relatively few new power plants have been built in eastern PJM since RPM has been implemented. However, that doesn’t mean no new resources have been added in eastern PJM. Since 2007, capacity uprates of existing plants (2,210 MW), reactivations (360 MW), export reductions (930 MW), increased demand response (6,550 MW), and approximately 2,040 MW of new generation capacity has been committed through RPM in eastern PJM. And another 650 MW of new generation offers have been submitted but failed to clear because sufficient capacity was offered at prices below the cost of new generation.

The relatively modest level of new generation construction in eastern PJM hasn’t led to resource adequacy shortfalls, as some stakeholders believe. Rather, reserve margins have remained at or above target levels due to the combination of entry by these new generation units, demand response resources, upgrades to existing capacity, deferred retirements, planned transmission upgrades, and the economic slowdown. Moreover, RPM has maintained resource adequacy at prices that are generally below the cost of new generating plants.

State regulators are correct that market prices for capacity in eastern PJM have been significantly higher than in the remainder of PJM in most of the recent capacity auctions. However, even the higher eastern PJM capacity prices have generally remained below the cost of new plants. Prices will remain below the cost of new plants until new generation is needed and capacity prices rise to clear new offers.

The underlying fact under these market fundamentals is that new generation simply isn’t cost-competitive with lower-cost options—such as uprates, deferred retirements, and demand response. That’s likely the primary reason why more new generating plants haven’t been built in eastern PJM. That capacity prices will remain below the cost of new plants through 2015 and possibly for several more years is likely also the primary reason that some developers’ new generation projects can’t be financed without long-term contracts that cover project costs. Current market conditions simply don’t support long-term contracts at prices high enough to finance new plants because rational buyers prefer to satisfy their capacity requirements at market prices that are below the contract cost of a new plant.

Under these market conditions, when few or no new plants are needed, the only way to finance additional new generation would be through above-market long-term contracts. Such above-market contracts have recently been offered through a legislative mandate, which procured capacity for three new plants under fixed-price 15-year contracts whose costs aren’t public but that are estimated at $270 to $350 per MW-day.3 In comparison, RPM prices in have been much lower at $136 to $225/MW-day in PJM’s most recent capacity auction. Nevertheless, in October 2011, the utilities similarly issued RFPs for new generation under direction of the Maryland PSC.

In short, the lack of long-term contracts and financing for new plant construction is a consequence of the fact that investments in new generation are currently unprofitable and not a least-cost option to ensure resource adequacy. Under these circumstances, a well-functioning market won’t reward investments in new generation. The absence of new construction is a sign that the market is working.

Current market fundamentals are also the likely reason that public power entities looking for long-term capacity contracts haven’t found willing suppliers. First, given that capacity prices might remain below the cost of new plants for a number of years, buyers interested in long-term contracts won’t be willing to sign long-term contracts priced at the full cost of new power plants. Thus, developers of new power plants will be unwilling to offer long-term contracts at prices acceptable to buyers. Second, even owners of existing generating capacity will be unwilling to sign long-term contracts at prices equal to current market prices if they anticipate that RPM prices will increase over time. It’s likely, however, that buyers’ and existing generators’ interest in longer-term contracting will increase during the next several years as excess capacity diminishes and capacity market prices rise to the cost of new generation.

Financing Power Plants

Without a need for new plants, financing for such plants won’t be available unless supported by above-market long-term contracts.4 However, this doesn’t mean that financing is unavailable for sound investments at costs that are consistent with market fundamentals. In fact, there’s been keen interest in plant acquisition. Several major transactions of power plants in eastern PJM demonstrate the availability of financing for generation investments.

A notable example in eastern PJM is Calpine’s 2010 acquisition of 4,490 MW of Conectiv Energy power plants from Pepco Holdings. The $1.63 billion purchase included some existing forward capacity and energy sales commitments as well as a six-year tolling agreement with Constellation Power for the Delta power plant that was under construction at the time. Importantly, it was financed with $1.3 billion of seven-year debt and $100 million of three-year debt.

Many generation developers prefer to build new power plants through highly leveraged project finance arrangements, which require long-term power purchase agreements. Project finance refers to the use of project-specific debt, also called “non-recourse” debt that isn’t backed by a guarantee from a larger parent company. Project finance often is the only available option for small project development companies that don’t have a significant portfolio of other assets or for companies with weak balance sheets and poor credit ratings.

Such non-recourse debt is secured solely by the revenues and asset value of the specific power plant. It’s more risky to the lender and consequently more expensive than corporate debt, which is secured by the more diversified revenues and assets of the parent company. However, while more expensive than corporate debt of companies with investment-grade credit ratings, non-recourse debt is still attractive to developers because it’s less expensive than equity and reduces the potential liability to the parent company if the project proves to be a bad investment. Non-recourse debt also can be less expensive than the corporate debt of companies with poor credit ratings.

To reduce financing costs, project developers also will prefer to lever up their investments by using higher levels of debt and less equity. However, such reductions in financing costs are possible only if project risks are reduced through long-term power purchase agreements that shift market risks from the generation owner to the buyer of the power. In fact, by assuming project risks through a long-term contract, the buyer is reducing (and essentially subsidizing) the financing cost of the new plant. Financing projects with high levels of debt (e.g., 70 to 80 percent debt) can reduce the levelized annual investment cost of a project by 10 percent to 20 percent compared to merchant plant financing, which may allow financing with only 30 percent to 50 percent non-recourse debt (backed solely by the project) or 50 percent to 60 percent corporate debt (backed by the entire parent company).

A range of financing arrangements will exist in well-functioning markets. Buyers can assume risks under a long-term contract—which supports more highly leveraged financing by the developers—or developers can assume these risks—which requires financing with more equity—depending on risk sharing preferences and the financial conditions of the counterparties.

Until new generation is actually needed, it isn’t desirable to enable uneconomic investments in new generation through long-term power purchase agreements (PPAs) when those developments are more costly or more risky than capacity from market-based resources, including from existing generation supplies and demand response when new generation isn’t needed.

Reforming Default Service

Longer-term contracting should be expected to increase as capacity market prices reach and sometimes exceed the cost of new generation. It’s conceivable, however, that market or regulatory barriers could prevent an outcome in which an efficient level of longer-term contracting can be achieved.

The current nature of retail services in restructured states may represent such a barrier that might inhibit reaching optimal levels of long-term capacity contracting. That is because a significant portion of retail load is supplied under regulated default service arranged by electric distribution companies and overseen by utility commissions. In restructured eastern PJM states, such as New Jersey and Maryland, the distribution companies are required to procure bundled energy and capacity supplies for these default service obligations. The contracts for such default service procurement have durations of only three years or less. This sole reliance on short- or intermediate-term contracts under state-regulated default service procurement appears to deviate significantly from the procurement and risk management practices of large competitive retail service providers.

Many competitive retail service providers secure a meaningful portion of their supplies through long-term contracts or even the acquisition of generating assets. Such actions are designed to counter the effects of perceived broken linkages between competitive retail and wholesale markets by reducing the transaction costs of securing long-term contracts and effectively vertically re-integrating load-serving responsibilities with merchant generation. For example, Constellation’s NewEnergy retail supply business obtains energy from a portfolio of various sources, including its own generation assets, contractually-controlled generation assets, exchange-traded bilateral power purchase agreements, unit-contingent power purchases from generation companies, tolling contracts with generation companies, and spot purchases from the regional power markets.5 This portfolio balances retail sales contracts that are reported to extend from one to 10 years and beyond, although these generally won’t be exactly matched by long-term capacity procurement contracts. Constellation Energy explicitly stated that its that its strategic objective for retail service operations is to buy generation assets in regions where the company doesn’t have a significant generation presence and to enter into longer-term agreements with merchant generators. In fact, this objective was a primary reason for Constellation’s purchase of generating plants in Texas as well as its recent acquisition of 2,950 MW of generating plants in New England, which “improved [Constellation’s] net load to generation ratio to approximately 55 percent.”6

Direct Energy, another retail service provider, appears to have started pursuing a similar strategy through long-term contracting power from generation suppliers, buying physical generation assets, and even acquiring natural gas production, storage and transportation.7 Similarly, NRG’s recently announced acquisition of Energy Plus Holdings was explained as an effort to “expand its retail marketing presence in the Northeast and Mid-Atlantic” to give the company “more of a retail presence to offset its generation assets in periods when wholesale power prices are depressed.”8 NRG’s announcement also marked another retail acquisition following Constellation Energy Group’s purchase of StarTex Power and its planned acquisition of MXenergy, and Direct Energy Services’ purchase of Gateway Energy Services.

It’s unclear what fraction of total retail load should be supplied through long-term contracts or physical plant ownership. Such decisions will depend upon a company’s tolerance for risk and expectations regarding future market conditions. While long-term contracts and physical plant ownership will stabilize procurement costs, they also create the risk that costs will be above market. However, it’s possible that the most efficient amount and duration of long-term contracting may exceed the amount realized for load under current default service procurement. This potential concern over whether the short- and medium-term nature of default service procurement creates a barrier to efficient contracting should primarily be a matter for state commissions and state legislatures, which should examine it in the context of improving retail choice and default service regulations. The best way to realize an efficient level of long-term contracting and asset ownership among retail providers might be for the states to reform or reduce their reliance on default service. That would foster increased interaction between retail service providers and customers—would allow market participants to determine the most efficient retail supply portfolio. Reduced reliance on default service, for example, exists in Texas where most retail customers are served by competitive suppliers after default service was eliminated in 2007—although a provider of last resort service is still available to customers who lose their competitive service providers.9 A second option that states could pursue would be to review default service procurement practices to determine the extent to which longer-term contracts, procured on a non-discriminatory basis from existing or new resources, should be part of default service procurement.

Only if states fail to pursue these options and generation investment lags, even as market prices reach or exceed net CONE [cost of new entry], might it be necessary to add mandatory long-term procurement to the current capacity market designs. However, this is a far less desirable option and worth pursuing only if it becomes clear that a review and revision of default service procurement is unlikely—and then only if it can be determined with sufficient confidence that longer-term contracts will actually be needed within the RTO capacity market design to assure resource adequacy at reasonable costs.

Managing Investment Risks

There’s a perception that new generation can’t be built without long-term PPAs of close to 10 years or more. As noted, this perception is largely created by current low-priced market fundamentals and the preference among developers to lay off risks onto contract counterparties. Reliance on long-term contracts is also rooted in the regulated past of the industry, including qualifying facilities (QF) under the Public Utility Regulatory Policies Act (PURPA). However, a number of observations about customer preferences and contracting practices in other capital-intensive industries suggest that widespread current perceptions might overstate the need for long-term contracting as the industry evolves.

First, most retail customers are unwilling to commit to long-term contracts. The reluctance isn’t unique to restructured electric power markets. This is also the case for most energy commodities sold in retail markets, including commodities with even higher price uncertainty, such as gasoline. If fixed-price contracts are signed in other retail market segments, they rarely go beyond the next season (e.g., heating oil), or the next two years (mobile telecom service). In fact, long-term contracts between retail customers and suppliers are uncommon even in the most risky and capital intensive portions of the energy industry—such as oil and natural gas exploration—and remain uncommon despite the high and unpredictable nature of risks, such as oil price movements based on a wide range of geopolitical influences, including cartel behavior.

Second, other capital-intensive industries with significant price risks generally require that investments are backed by companies with sufficient equity. However, such “balance sheet financing” of major investments is less common in the electric power industry.10 While numerous examples of balance-sheet financing and generation investments without long-term PPAs or other long-term price hedges exist—including merchant wind power development—project financing arrangements supported by long-term PPAs remain the first choice of most power plant developers.11

The lower reliance on balance sheet financing in the power industry doesn’t mean that project developers in other industries wouldn’t prefer the lower risk and financing costs they would be able to achieve if they had long-term sales agreements. Nor does it mean that power industry developers are unable to develop projects without long-term sales agreements. Rather, the relatively low levels of balance-sheet financing in the power industry appear to be an artifact of industry evolution. Specifically, the merchant generation sector has evolved based on: 1) long-term PPAs with regulated utilities, starting with mandated QF contracts in the late 1980s and early 1990s; 2) project development efforts by small companies without much equity; and 3) a reliance on highly leveraged financing arrangements.

Third, large competitive retail electricity providers and companies in other capital-intensive industries, including in oil and gas, also tend to be partially (but not fully) vertically integrated to manage risks and reduce transactions costs. Many competitive retail providers have bought physical assets or signed a portfolio of contracts to manage overall supply obligations and associated risks. Such a partial vertical re-integration appears to be becoming more prevalent in electricity markets. In the United Kingdom, for example, all major retail suppliers have re-integrated into the generation business.12 Similarly, generation owners are integrating vertically into retail sales, as noted in the above discussion of NRG, Constellation, and Direct Energy, and with Exelon’s proposed merger with Constellation as another recent example.

A transition to a partially integrated industry structure has a number of potential advantages and will reduce the need for, or compensate for the lack of, extensive bilateral contracting.13 Competition will be maintained or enhanced because the companies have a reduced ability and incentive to exercise market power and, unlike in non-restructured markets, aren’t fully integrated and don’t enjoy exclusive service franchises.14

Consistent with these observations, the deregulated electricity industry likely will migrate naturally to a partially vertically integrated structure that, over time, will rely less on long-term PPAs to underwrite new generation development. These trends reflect an efficient response to deregulation, which shifts the risks of potentially uneconomic generation investments away from customers and toward developers. As increasingly large and diversified companies, these developers will be in a better position to evaluate, manage, and bear these risks. Regulatory or legislative intervention to force long-term contracting in restructured markets—even if accomplished via the typical capacity market design—carries with it the risk of interfering with the natural evolution of the industry and threatening adverse long-term consequences for future capacity expansion.

Market Fundamentals

There may be many generation projects that can’t get financed and built under current market conditions. However, while some project developers might cast this as a market failure caused by the inadequacies of capacity markets or state retail choice constructs, the primary reason that these projects can’t get financed and built is that they are currently uncompetitive with alternative sources of capacity.

The main reason for the low activity of new power plant construction in eastern PJM, for example, is the fact that new plants haven’t been needed—and won’t be needed for several more years—due to a combination of economy-related decreases in load growth, transmission upgrades, and the availability of lower-cost supply options, such as deferred retirements, demand response, and upgrades to existing units. In fact, PJM’s capacity market has successfully ensured resource adequacy at prices below the cost of new plants. In other words, the lack of long-term contracts for new generation is explained by market fundamentals and the simple fact that new plants have been out of the money.

These market fundamentals also explain the lack of long-term contract offers from existing generation. Suppliers of existing capacity are unwilling to enter long-term contracts at low current prices because they expect prices will rise. At the same time, buyers are unwilling to pay higher prices or even the cost of new generation when there are less expensive options currently available in the market.

When new plants are needed for resource adequacy, capacity market prices will rise and will make these investments attractive—both on a merchant and long-term contracting basis. It’s likely that the need for and reliance on very long-term PPAs and project financing ultimately will diminish as the industry evolves and an increasing share of new plants are developed by larger, partially vertically-integrated companies with load-serving responsibilities, a portfolio of merchant generation, and sufficiently strong balance sheets to finance the needed investments.

It’s possible, however, that merchant investors remain wary of market volatility and vulnerability to regulatory intervention, in which case more long-term contracting will be needed. In that case, some load-serving entities will also be interested in procuring more of their resources under long-term contracts or by owning physical generation to hedge the price uncertainty. Yet several secondary factors could still create barriers to long-term contracting, such as the structure of default service procurement in retail access states. If these barriers turn out to be significant—which is difficult to determine under current market conditions—modifying how default service procurement is regulated at the state level may be the most effective way to address these barriers.

If merchant investment and long-term contracting are both impeded even as market fundamentals become tighter—which isn’t yet evident—it might be worth considering policy options to force long-term contracting. One option would be for RTOs to add longer-term procurement to the current capacity market designs. Such decisions shouldn’t be made prematurely, however, because it shouldn’t be the role of an RTO to force long-term contracting for capacity resources when load-serving entities don’t see the risk management benefit of entering into such contracts bilaterally. Nor would an RTO be able to readily determine the amount of long-term contracting or contract terms that optimally balance risks. Mandating too much long-term contracting would inefficiently expose suppliers to delivery and credit risks, while exposing buyers to larger risk premiums and the potential for stranded costs.



1. Pfeifenberger, Newell, Spees, Hajos and Madjarov, “Second Performance Assessment of PJM’s Reliability Pricing Model,” Aug. 26, 2011.

2. See also letters from Credit Agricole and Union Bank attached to LS Power Associate Comments on New Jersey Electric Power and Capacity Needs, Submitted in State of New Jersey Board of Public Utilities, Docket No. EO 09110920, July 2, 2011.

3. Approximate procurement prices were calculated in “Comments of the New Jersey Electric Distribution Companies on Agent’s March 21, 2011 Report,” March 24, 2011.

4. See also B. Chin, “Capacity Issues Technical Conference: State of New Jersey,” Citi Investment Research, June 24, 2010, noting that “in our view, energy/capacity markets are providing a signal that capital shouldn’t be deployed to [new] generation at this time, unless subsidies are enacted.”

5. See Constellation’s 2010 10-K filing, Part 1, Item 1, pp. 4-5.

6. “Constellation Energy Signs Agreement to Acquire the 2,950 MW Boston Generating Gas Fleet in New England,” Constellation Press Release, August 9, 2010.

7. “Direct Energy Corporate Fact Sheet.” May 2011.

8. Megawatt Daily, “NRG to buy Energy Plus Holdings for $190 mil,” Aug. 17, 2011.

9. Kiesling and Kleit (2009), Electricity Restructuring: The Texas Story, Chapter 8, AEI Press, Washington, D.C.

10. The use of balance sheet financing doesn’t mean that medium- or long-term contracts are eliminated for these projects. Rather, it simply means that the role of medium or long-term contracts is reduced because at least some projects can be built with less of the project costs hedged through long-term contracts. Projects may be built without PPAs, shorter-term PPAs, or PPAs that cover only a portion of the project’s expected sales.

11. For example, the DOE reports that in 2009, 38 percent of all new wind generation capacity was from merchant or quasi-merchant projects that relied on short-term contracts or hedged wholesale spot market sales rather than long-term PPAs. See Wiser and Bolinger (2010), 2009 Wind Technologies Report, DOE Energy Efficiency and Renewable Energy, August 2010, p. 34.

12. In the U.K., for example, restructuring in the early 1990s resulted in completely vertically unbundled industry structure. Today, the six largest competitive retail suppliers (supplying 99 percent of retail load) also own approximately 70 percent of the installed generating capacity. Note, however, that such partial integration by large companies will also tend to make it more difficult for smaller and non-integrated suppliers to enter and compete in the market. See Ofgem, Liquidity Proposals for the GB wholesale electricity market, February 2010.

13. For a discussion of the implications of vertical re-integration of competitive retail service and generation companies, see Meade and O’Connor (2009) “Comparison of Long-Term Contracts and Vertical Integration in Decentralised Electricity Markets,” Larsen Working Paper No. 26, October 2009; and Mansur (2007) “Upstream Competition and Vertical Integration in Electricity Markets,” 50 J. Law & Econ. 125.

14. See, for example, Bushnell, J. B., Mansur, E. T. & Saravia, C. (2008). “Vertical Arrangements, Market Structure, and Competition: An Analysis of Restructured U.S. Electricity Markets.” American Economic Review, 98, 237-266.