The North American Electric Reliability Corporation (NERC)¹ was formed after the 1965 Northeast blackout.² It was a voluntary organization whose mission was to ensure that the bulk power system in North America is reliable, adequate, and secure. Ten Regional Reliability Councils were established on the basis of the physical design of transmission systems and historic cooperation among various utilities.³ The Reliability Councils established an obvious framework for the evolution of power pools, since the type of coordination that would increase reliability, especially planning transmission to allow access to external reserves in emergency situations, was also conducive to more routine energy transactions.

The FERC's predecessor, the FPC, had actively encouraged utilities to form power pools. The FPA, as a matter of general policy, held that power pooling arrangements were in the public interest. As of 1970, there were twenty-two formal power pools nationally, which had 60 percent of the nation's generating capacity.⁴ Section 202(a) of the FPA directed the Commission to divide the country into regional districts for the voluntary interconnection and coordination of facilities for the generation, transmission, and sale of electric energy. In enacting this section, Congress was “confident that enlightened self-interest will lead the utilities to cooperate…in bringing about the economies which can alone be secured through…planned coordination.”⁵

An effective pooling agreement must coordinate all major design, construction, and operating decisions of the contracting utilities, while dividing the gains of trade among the various participants in a mutually satisfactory manner. Cost-of-service regulation of utility rates reduced the incentive for firms to enter into cost-saving pooling contracts, because regulators confiscated most of the increased profits achieved through pooling. Regulators and courts frequently complicated the negotiating process by mandating that all utilities in a region be allowed to participate in the pooling arrangement, including smaller firms that had little to contribute in terms of transmission or generation to provide cost savings to other members.⁶

Agreements not to compete were common in pooling contracts, because each party to the agreement was generally unwilling to risk losing part of its market to other participants. While the FPC was not formally bound by the antitrust laws, the agency's duty to regulate in the public interest required it to consider the possible anticompetitive consequences of a proposed pooling agreement. The FPC should determine whether the cost savings from the pooling contract would outweigh the adverse effects of any anticompetitive restrictions.⁷ However, the decision in Central Iowa Power Coop v. FERC showed that the court would accept FERC rulings that allowed reasonable restrictions in pooling agreements.⁸

In the 1980s, there were two types of tight power pools: contractual pools between different companies, and holding company power pools. Centrally dispatched pools accounted for about one-third of installed generating capacity in the United States in 1989. The four major contractual pools were the Michigan Electric Coordinated System, New England Power Pool (NEPOOL), New York Power Pool (NYPP), and the Pennsylvania New Jersey-Maryland Interconnection (PJM). The Great Northeast Blackout of 1965 had prompted the Northeast's power companies to form power pools to ensure a dependable supply of electricity. The five large holding company pools were the Allegheny Power System, American Electric Power System, Middle South Utilities (renamed Entergy), Southern Company, and Texas Utilities Company (renamed TXU). Among the centrally dispatched pools, all but NYPP utilized centralized unit commitment. Brokered pools provided a managed market for power, using auction markets or electronic bulletin boards to facilitate market transactions among their members. Brokered pools did not engage in centralized dispatch. Brokered pools included the WSPP, the MidContinent Area Power Pool (MAPP), and the Florida Coordinating Group.⁹

These pools gradually solved many of the technical problems that needed to be addressed to make centralized spot markets a reality. The most important issue was extending coordinated operation from a single utility control area across a region with multiple, interconnected utility systems, independent generators, and other entities. A control area is a geographic region with a control center responsible for operating the power system within that area. System control consists of the control area operator functions that schedule generating units, transmission resources, and transactions before the fact and monitor and control transmission resources and some generating units in real time to maintain reliability. This service can also include after-the-fact accounting and billing.¹⁰

Control center equipment and procedures are typically organized into three somewhat overlapping systems that are generally integrated in an energy management system (EMS). The Automatic Generation Control (AGC) system coordinates the power output of generators, the supervisory control and data acquisition (SCADA) system monitors and controls transmission line equipment and generator voltages, and the unit dispatch software monitors and evaluates system security and performance and dispatches generation units. Unit dispatch software was designed to minimize generation costs subject to reliability constraints. Telemetry is used to transmit data and commands to and from the various components of the interconnected grid system and the control center. A State Estimator program gathers all available telemetry data on the system and gives a real-time picture of system status. Control centers must ensure that generating units will be ready when needed to follow the daily load cycle (accounting for ramping constraints), that the transmission system is capable of carrying the loads, and that backup generating capacity is available in case of equipment failure.

Transmission networks can be viewed as consisting of combinations of nodes (“buses”) and links between the nodes (“flowgates”). Transmission lines can be direct current (DC) or alternating current (AC). Power in North America is largely generated and delivered via AC systems. Power system frequency is the number of sine wave cycles that the alternating current completes each second. In North America, the power system frequency is 60 cycles per second, or 60 hertz (Hz) (most of the world uses 50 Hz). Electrical current flows proportionally through the conducting paths in inverse proportion to the paths' resistance. The resistance of a transmission or distribution line will be a function of the line's length and thickness. Through the interactions of Kirchoff's laws,¹¹ a line limitation affects every other power flow in an AC network. Power flows in networks follow Kirchhoff's laws and cannot be directly controlled. Therefore, it is impossible for a buyer or seller of electricity to specify the route the electric power follows.

Power generation, load, and flow in an AC system are divided into both active and reactive power components. Active power is measured in watts while reactive power is measured in volt ampere reactive (Var).¹² Reactive power supports the magnetic and electric fields necessary to operate the power system equipment. Reactive power is never consumed but is constantly exchanged (at twice system frequency) between devices that produce reactive power and those that store reactive power. Voltage can be affected by both active and reactive power loads, and the interaction between the two is critical in determining limits on real power flows.¹³ During peak load conditions, generators are usually operated to supply reactive power to the grid to help maintain adequate voltage levels. During light load conditions, generators may be used to absorb excess reactive power from the grid to prevent voltages from becoming too high.¹⁴

The thermal capacity of a transmission line sets an upper limit on the flow of power on that line. Owing to electrical resistance of the conductors (the wires, usually made of stranded aluminum woven around a core of stranded steel to provides structural strength), a small portion of transmitted power is converted into heat. If the power flow is too large, the wire will expand and eventually sag too close to the ground, causing a short circuit. Every transmission line is designed to carry a certain maximum amount of electric current, which, if exceeded for an extended period, could damage the transmission line. A normal rating is the level of power flow that the line can carry continuously. An emergency rating is usually defined as the levels of power flow the line can carry for a limited period of time.¹⁵

Even when power flows do not approach the thermal limits of the system and transmission lines appear to have excess capacity, other factors can constrain the transfer capacity. Reactive power must be taken into account when calculating thermal limits on transmission links. In an AC transmission system, the voltage is maintained at nearly same average level at all points in the system. To induce power to flow from one point to another, there must be a “phase difference” between the alternating voltage at the point of generation and the load. However, this phase difference cannot become too large without threatening voltage collapse. Stability limits on power lines specify a maximum power flow. Thermal limits dominate over shorter transmission lines; at intermediate distances the limits are related to voltage drop; and beyond roughly 300–350 miles, stability limits dominate.¹⁶

A key task of the control area operator is to prevent overloading of transmission lines, accounting for both power flows and reactive power. This was traditionally performed using generation shift factors, which express the change in flow on a particular flowgate that results from increasing generation at a node. Shift factors are meaningful only when considered in source-sink pairs, because power injected at one location must be matched by power removed at another location. Generation shift factors are used to identify which generator pairs can influence a particular flowgate.¹⁷ A nodal model automatically solves for the optimal set of shift factors.

Transactions were initially scheduled by local utility control authorities who monitored actual interchange flowing over the tie lines. Utilities monitored the power flowing across interconnection lines to record power transactions for accounting purposes and to enable their control centers to make corrective actions in order to balance the area's total generation requirements. The difference between the total required generation (i.e., real-time load) and actual generation is called the area control error. AGC is used to adjust generator output to regulate area control error. The information needed to implement AGC includes tie line flows, system frequency, and capability of generating units on line. As attempts are made to adjust the area control error for each area, errors inevitably accumulate because the actions of generators differ from what was instructed; thus the term “uninstructed deviations.” Consistent errors in one direction or the other, along with unexpected changes in demand, give rise to frequency drift, which must be corrected with frequency regulation.

Large central generators are driven by prime movers, which may be steam or hydro turbines. Under steady-state operation of a generator, the net mechanical torque supplied by a prime mover is equal to the net electrical torque. Mechanical torque generated by a prime mover can only be changed with some time delay, due to the mechanics of the machine. As a utility's system load varies throughout a day, there will be imbalances between the mechanical power input and the electrical power output. As a result, changes in system frequency will occur continuously. System frequency is maintained close to the nominal value of 60 Hz to prevent damage to rotating equipment. Large steam turbines are especially vulnerable to off-speed operation, and the units will be tripped as soon as abnormal frequencies are sensed by over/under-frequency relays. One effect of poorer-quality frequency regulation is that the frequency excursions around the nominal frequency become greater, and last for longer periods, increasing the potential for generation unit tripping and even cascading failures.¹⁸

Electricity systems rely on ancillary services to provide protection against inevitable errors and contingencies. Operators attempt to maintain system frequency at 60 Hz using various types of ancillary services to balance generation and load. The key ancillary services are Regulation, Spinning Reserves, and Nonspinning Reserves. Most generators have automatic control systems that sense local frequency deviations and respond in a sub-minute time frame to increase or decrease generator output in response, without signals from a control authority. Regulation is the use of online generating units equipped with governors and AGC that can change output quickly in response to control authority signals to balance fluctuations in customer loads and unintended deviations in generation. Spinning Reserve is the use of generating equipment that is online and synchronized to the grid that can begin to increase output immediately and be fully available within ten minutes to correct for generation load imbalances. Loads under the control of the system operator can also provide this service. Nonspinning Reserve is comprised of generating equipment and interruptible load that can be fully available and synchronized with the grid within ten or thirty minutes.

In general, nonutility generators (NUGs) did not follow the stringent operating guidelines adopted by virtually all electric utilities. Permitting access by numerous NUGs to transmission systems increased the number of transactions to be monitored, and thus the complexity of maintaining coordinated control of the interconnected bulk transmission system.¹⁹

By the end of the 1980s, there were more than 140 control areas. Because most systems were interconnected with neighboring utilities, each control area had to match its load to its own internal generation plus power exports (or interchanges to other control areas) less power imports. Utilities also belong to an interconnected network. There are three such networks in the United States: the Eastern Interconnection (which extends nearly to the Rocky Mountains), the Texas Interconnection, and the Western Systems Coordinating Council (WSCC). These interconnections extend into Canada, which has a fourth system, the Hydro Quebec System. Within each interconnection, all connected generators must be synchronized. There were about ninety-nine control areas in the Eastern Interconnection, thirty-four in the WSCC, and ten in the Texas Interconnected System. Because of interconnection, each control area must satisfy more stringent requirements for generation control, frequency control, and tie line flows than would be needed for an isolated system. Connections between networks are accomplished through DC interties to avoid synchronization problems.²⁰

Within an interconnection, inertia due to the mechanical torque from a larger number of generators can provide additional time to resolve contingencies that require immediate action to restore frequency by bringing on additional generation and/or shedding load. When a major generation unit (or transmission line from a generator) trips, the effect is to lower system frequency. The frequency decline is initially arrested because many electrical loads (e.g., motors) are frequency responsive; that is, their demand varies with system frequency. Once the frequency decline exceeds the deadband of the generator governors, the governors sense the frequency decline and open valves on the steam turbines, which rapidly increases generator output. After a few more seconds, generator output declines slightly because the higher steam flow through the turbine is not matched by the steam flow from the boiler to the turbine. At this point, operating reserves, in response to AGC signals from the control center, hopefully kick in and increase generation to match load, restoring system frequency.²¹

Centrally dispatched pools became “super” control centers, working with each utility's control area to coordinate their activities. Over time, as pools became tighter, these central dispatch controls gradually superseded the activities of local control areas, treating the pool as one big control area. Independent generators operating in these control areas were required to coordinate with the pool's control center and implement load frequency controls.

Ancillary services were produced by traditional utilities as part of the bundled electricity product they provide to their customers. They were also joint products of the overall design of power plants, optimized to meet all of the technical specifications (including voltage control and dispatch flexibility) at minimum total cost. So determining costs attributable to voltage control, or the flexibility to provide other ancillary services separate from the costs attributable to real-power production, required some arbitrary assumptions. Determining the cost of these services would also require calculation of opportunity costs (for example, to provide some ancillary services, a generation unit may have to operate at less than full capacity, which might affect heat rates as well as the net costs from having to run another unit to supply additional output) and increased operations and maintenance costs from frequent changes in output.²² Once pool control centers spanned multiple utility control areas, ancillary services gradually became products purchased in markets, allowing payments to suppliers of these services and a means to allocate costs between utilities, and eventually among all market participants.

The need for coordination over a larger region led to recognition that there must be a system operator coordinating use of the transmission system, which also implied control of the dispatch, at least at the margin. This is because adjusting dispatch of generation units is the principal means of affecting the flow of power on the grid. Available capacity on transmission systems is difficult to determine because it depends on changing factors, such as the mix of generators and load, as well as the interaction between transmission flows, known as loop flow.²³ Engineering power flow models are used to determine the available transmission capacity (ATC) of a transmission interface based on a variety of assumptions about system conditions and reliability. The rights to use the ATC over a contract path from a set of injection points to one or more withdrawal points on the network can be sold or assigned to generators or loads, assuring that sold/purchased power can be delivered.

The economic power granted the system operator through control of dispatch in turn suggested that the system operator should be independent of the participating utilities. This independent system operator (ISO) would be providing a service, but would not be competing in the energy market. An ISO with no economic interest in which generator is dispatched, or the cost of transmission and associated congestion, other than minimizing total cost and maintaining reliability, lacks the economic incentive that an integrated utility would have to manipulate ATC calculations to favor its own generation. The question then became whether the ISO should be completely independent from the spot market, which would be organized in a power exchange that was strictly separated from the activities of the ISO. The ISO would focus only on reliability and not receive any bidding information, perform any economic dispatch, or determine any spot market prices.²⁴ However, separating economic dispatch from reliability redispatch was fraught with complications because it required constant iteration from one subsystem to another (remember, SCED solved for economic dispatch given reliability constraints).

Developments in transmission and coordination technology had led to increased interconnection between independent IOUs, with joint planning and operation of facilities owned and operated by proximate utilities. Short-term purchases and sales of electricity by interconnected integrated utilities provided additional resources for reliability and to exploit opportunities for both parties to lower costs. These wholesale transactions gradually expanded to include longer-term purchase and sale contracts to take advantage of surplus generation stemming from unexpected declines in demand growth. However, state regulators were reluctant to allow utilities to use these contracts to meet capacity requirements for reliability.²⁵

Despite the advantages of strong pools, in terms of operating efficiencies and reliability, most regions never progressed further than informal arrangements. The reasons for reluctance to engage in pooling included the reduction of the scope of managerial control due to participation in a pool. Many utilities expressed strong reservations about centralized dispatch and felt that most of the benefits could be obtained with more informal cooperation. The appropriation of potential benefits by regulators reduced any incentive senior management might have to relinquish managerial control.²⁶

Regional transmission groups (RTGs) were coalitions of transmission-owning and transmission-using entities in an identified geographic region, which adopted voluntary terms of transmission access and expansion. A number of utilities had voiced concerns that formation of RTGs would violate the antitrust laws. The Department of Justice expressed support for the use of RTGs as a means of encouraging efficient planning and use of the transmission system, alleviating concerns that these organizations might raise antitrust issues.²⁷

During the final stages of congressional consideration of the Energy Policy Act there was support for legislation that would encourage RTGs, but the legislative clock ran out before a bill could be finalized. FERC issued a policy statement in response to this congressional interest, in which it formally stated its approval of RTGs and set out some guidance for their formation. The Commission believed that RTGs could be vehicles for promoting competition in generation. RTGs could provide mechanisms for encouraging negotiated agreements and resolving transmission issues and reduce the need for litigation before the Commission. As a voluntary association of transmission owners, users, and others with differing interests, an RTG could not insulate transmission utility members from proceedings under FPA section 211. However, the Commission promised an appropriate degree of deference to decisions of RTGs, depending on the degree to which an RTG agreement mitigated the market power of transmission owners and provided for fair decision making.²⁸ FERC approved a number of RTG agreements, including the Northwest Regional Transmission Association,²⁹ the Southwest Regional Transmission Association,³⁰ and the Mid-Continent Area Power Pool.³¹

The Commission initiated a notice of inquiry on alternative power-pooling institutions in 1994. A major focus of this inquiry was on power pooling to facilitate short-term transactions. FERC envisioned power pools operating spot markets while RTGs would focus on long-term and regional transmission planning. Anticipating the development of ISOs, FERC noted that the same institution could perform both functions.³²

The initial impact of EPAct 1992 was to open up electricity markets for IPPs and utilities wishing to establish unregulated subsidiaries to generate and sell power. Many utilities began to issue request for proposals for power purchases as new suppliers provided a credible alternative to generation investment. Wholesale power bids fell as low as 3 cents per kWh on long-term contracts, based on low natural gas prices.³³ The Commission began to formulate policies to deal with the brave new world of power pools, NUGs, and EWGs. Utilities had begun to sell excess generation in bilateral deals and in power pools, but needed “rules of the road” that defined permissible prices in lieu of directly regulated rates. It was too complicated and time consuming to go to FERC every time a utility wanted to sell surplus power to a wholesale customer in order to determine a “just and reasonable” rate for the transaction.³⁴

FERC had begun experimenting with rates for wholesale electricity transactions as early as 1983. In Ocean State, the Commission set forth its general approach to market transactions. The Commission can rely on market-oriented pricing for determining whether a rate is just and reasonable when a workably competitive market exists, or when the seller does not possess significant market power. A seller lacks significant market power if the seller is unable to increase prices by restricting supply or by denying the customer access to alternative sellers. Lack of market power is the key prerequisite for allowing market-oriented pricing.³⁵ Despite the assertion that the decision applied only to the specific circumstances presented by Ocean State, Commissioner Charles Trabandt pointed out that the decision established a new market pricing policy. The majority had adopted “marginal cost” pricing as the new ratemaking approach under the FPA, modeling it on the avoided cost methodology in section 210 of PURPA for qualifying facilities.³⁶ Trabandt prophetically asserted the need for adequate monitoring to ensure prompt Commission action if market forces are inadequate to constrain rates.³⁷

The legal basis for relying on market-based rates was established in the Commission's natural gas proceedings. As long as FERC exercised its authority to assure that a market rate was just and reasonable, the Commission may rely on market-based prices in lieu of cost-of-service regulation to assure a “just and reasonable” result. However, the Commission must confirm that either there is no market power or that it has established a mechanism to monitor and check market power.³⁸ Where there is a competitive market, the Commission may rely on market-based rates in electricity markets.³⁹

Once the Commission committed itself to permitting some sort of market-based ratemaking, it faced the requirement that it develop a methodology for analyzing market power. At first, the Commission seemed uncertain, relying on an “I know it when I see it” approach:

The Commission required evidence as to whether the seller was a dominant firm in generation in the relevant market; whether the seller controlled transmission facilities that could be used by the buyer to reach alternative generation suppliers; and whether the seller was able to erect other barriers to entry.⁴¹

The FERC's increased willingness to accept competitive procurement was encouraged by the DOJ, which urged the FERC to approve negotiated rates rather than to impose rates for wholesale bulk power where purchasers have competitive alternatives. DOJ encouraged FERC to reconsider a decision it had reached disapproving a wholesale power contract because the DOJ Antitrust Department felt that eighty proposals totaling 2,697 MW from seventy-four generating units constituted a competitive market for a 30 MW purchase proposal.⁴²

As long as market-based sales were on the margin, it really did not matter if FERC developed internal capabilities to analyze, monitor, and intervene in market transactions. However, as electricity markets developed, the lack of economic expertise inside of FERC would leave the agency incapable of dealing with the radically new environment. Sales of surplus electricity from and to IOUs where the bulk of generation and sales were executed under cost-based regulation could have little impact on prices and consumers. In organized electricity markets, any divergence from “just and reasonable” prices could have large impacts because a market price applied to all sales.