How does each region manage congestion, allocate losses and dispatch resources? Which players gain the most from each approach?
The United States now has six independent system operators, five approved by the Federal Energy Regulatory Commission and one approved by the Public Utility Commission of Texas. These ISOs present an astonishing array of similar and conflicting rules and philosophies by which transmission services are defined and priced.
This article aims to explain some of the key similarities and differences among the ISOs' transmission pricing schemes. In particular, we ask several questions:
* Pricing for Transmission Losses. How does each ISO assure that market participants pay for the power that is lost in transmission?
* Managing and Pricing Congestion. How does each ISO assure that flows over transmission facilities do not exceed the physical capacities of those facilities?
* Setting Access Charges. How does each ISO market allow transmission owners a fair chance of recovering their capital and operating costs?
* Winners and Losers. Which market participants are the relative winners and losers with respect to the ISOs' different transmission access and pricing schemes?
In the interest of brevity, we focus primarily on the real-time management and pricing of transmission services rather than on the pricing of transmission rights. We do, however, explain some of the ways in which transmission rights markets interact with real-time electricity markets.
Overview: Regions and Markets
Table 1 presents some key statistics of the six approved ISOs. All are large entities. They cover California (the California ISO), Texas (the Electric Reliability Council of Texas), a big portion of the Midwest (the Midwest Independent Transmission System Operator) and the whole Northeast (ISO New England, the New York ISO and the PJM Interconnection). [Note: Tables/graphics not included. See print copy]
The pricing methods of the ISOs are in transition. Major changes have been proposed or are in progress in the Electric Reliability Council of Texas (ERCOT), the Midwest ISO (MISO) and ISO New England, and lesser changes are occurring in the other ISOs. Table 2 summarizes key features of the ISO tariffs, dividing these features into those related to losses and congestion ("usage") and those related to overall cost recovery ("access"). Arrows (') indicate instances in which the market designers apparently intend to switch to a specific new pricing method.
Table 2 indicates no uniformity among the ISOs in the geographic scope of charges for losses and congestion management. California has zonal pricing, under which a generally uniform energy price is received or paid by all generators and consumers located within each of three geographic zones. ERCOT and New England have postage stamp pricing, under which a generally uniform energy price is received or paid by all generators and consumers throughout the ISO. MISO differentiates its loss charges by transaction. New York (and prospectively New England) has nodal pricing for generators - that is, the price can be different at every generator bus (node) - and zonal pricing for consumers. PJM has nodal pricing for everybody.
Table 2 also shows that five of the ISOs have, or are moving toward, some form of marginal cost-based pricing of losses. The exception is ERCOT, which is switching from distance-based, megawatt-mile pricing to uniform postage stamp pricing.
The great divide among the ISOs lies in whether they manage transmission congestion through price-based mechanisms. This divide mirrors the history and future of the power industry's restructuring. Historically, the regulated power industry found it necessary for each major service provider to have legal rights to the physical assets that provide service. Extending this historic practice, some ISOs manage transmission congestion by curtailing customers' use of transmission facilities according to the access priority assigned to each customer. In a competitive industry, by contrast, well-functioning markets should virtually always allow market participants to obtain the commodity they want on the spur of the moment, though sometimes at a high price. The differences in how the ISOs handle congestion reflect differences in opinion about how well market mechanisms can assure real-time power system reliability.
The differences in the methods for setting transmission access charges are relatively slight. Most ISOs began with access charges that, for each customer, depended upon the costs of only the transmission owner (TO) that served the customer. All ISOs but one are moving toward access charges that, for all customers, will depend upon the costs of all TOs in the ISO-controlled grid. All of the ISOs levy access charges on loads, not generators, and these access charges depend upon either the customer's total megawatt-hour consumption, their coincident peak megawatt loads, or their megawatt load reservations.
Line Losses: Pricing Methods
and Effects on Dispatch
There are three basic approaches by which the ISOs price (or plan to price) transmission losses: marginal cost pricing, postage stamp pricing and scaled marginal cost pricing.
Marginal cost pricing charges each consumer (and generator) for the cost of the change in losses that accompanies a change in their consumption (and output). New England and PJM intend to estimate the marginal costs of the losses that attend small changes in consumption at each bus.
New York has implemented an approach by which each consumer pays the cost of the marginal losses associated with a small percentage increase in all of the loads in its zone, while each generator pays for the marginal losses associated with power produced at each bus. Because the marginal costs of losses may exceed consumers' average costs, their payments for losses may exceed the total costs of losses. When there are such excess revenues, New York returns them to all customers as a uniform per-megawatt-hour credit against its scheduling, system control and dispatch service charges.
Postage stamp pricing charges all consumers for the average cost of the whole power system's losses. "Average cost" is the total cost of losses divided by total system load (including power exports). All load, regardless of location, pays the same per-megawatt-hour price for transmission losses, though this price may be different during on-peak periods than during off-peak periods. The postage stamp approach initially is being followed by New England, PJM and implicitly California, and is about to be adopted by ERCOT.
Scaled marginal cost pricing sets loss charges according to a mixture of marginal costs and total costs. MISO and (soon) California use two very different versions of this approach.
MISO uses a flow-based method to create a matrix of the estimated losses that accompany transactions from power resources in each zone to loads in each zone. This matrix is updated approximately every six months. The loss factors are intended to provide full recovery of the energy lost in transmission. For any given transaction, the same loss factor applies to all hours.
In California, each generator must provide energy to cover its allocated responsibility for losses. For each generator, this responsibility will be proportional to the marginal losses associated with its output. The generators' responsibilities will be scaled so that all generators together provide just enough energy to cover total losses.[Fn.1]
On the surface, the scaled marginal cost approach seems to be fairer than the other approaches. Unlike the pure marginal cost approach, the scaled approach does not levy loss charges that exceed the total costs of losses. Unlike the postage stamp approach, the scaled approach distinguishes loss charges according to how market participants cause losses.
But the MISO and California approaches differ in two significant ways. First, MISO differentiates losses according to the locations of both the generators and the consumers who are parties to a transaction, while California differentiates losses only according to the generators' locations. Second, MISO looks at the loss effects of transactions over a relatively long period of months, while California intends to consider loss effects over periods of hours.
It might seem that the New York and California approaches are similar because both reimburse market participants for the amount by which marginal cost-based loss revenues exceed the actual costs of losses. These two approaches really are quite different, however. New York returns the excess loss revenues to all consumers, as a credit against its scheduling and dispatch charge, while California returns the excess loss revenues only to the generators that cause losses. The result is that market participants that cause losses in New York see prices that reflect pure marginal costs, while those in California are charged prices reflecting a mixture of marginal costs and total costs.
To understand the differences among the three methods for pricing losses, consider the example of a simple power system with two busses connected by a single transmission line. Because transmission losses are always roughly proportional to the square of line flows, we assume for this example that the losses in this line are determined by the relationship
Losses = FLOW2 x 10-4
where FLOW is the average megawatt flow in the line. Table 3 illustrates, for two levels of flow, two key implications of this relationship - implications that are characteristic of transmission losses in general. First, average losses rise with flows. Second, marginal losses are exactly double average losses. The marginal loss rate of 5 percent, for example, means that if flow increases 1 MW, from 250 MW to 251 MW, losses will increase by 0.05 MW.
Now suppose, as illustrated in Figure 1, that this simple power system has a generator and a load at each of the two busses, none of which exercises market power. The western generator can produce up to 1,000 MW at a constant incremental cost of $20 per megawatt-hour, while the eastern generator can produce up to 200 MW at a constant incremental cost of $21 per megawatt-hour. Western and eastern loads are 500 MW and 392 MW, respectively.
Given that losses depend upon flows, Figure 1 shows the least-cost system dispatch: The western generator produces 753J MW while the eastern generator produces 145J MW. That results in total generation costs of $18,110.
Figure 1 shows that the three pricing approaches, not surprisingly, lead to different pricing results. Marginal cost pricing leads to prices that, for both generators and consumers at each bus, equal the incremental cost of generation, including the incremental costs of losses. As planned for implementation in PJM, consumer prices for losses will vary by bus as illustrated in Figure 1. In New York, which has zonal pricing for loads, the charge for losses in each zone equals a load-weighted average of the incremental costs of the losses associated with serving load at each bus. In other words, New York uses marginal cost pricing of losses between zones, but a version of postage stamp pricing of losses within zones.
Marginal cost pricing of losses can lead to over-recovery of the costs of losses. In Figure 1, marginal cost pricing yields total consumer payments of $18,232, leading to an excess recovery of $122.[Fn.2] In New York, this excess would be used to reduce scheduling, system control and dispatch service charges by $122.
The postage stamp and scaled marginal cost approaches lead to prices that are not consistent with efficient dispatch. Under the postage stamp approach, all consumers pay a price equal to the average cost of bringing power to each consumer. In Figure 1, consumers in the West and East pay the same power price, so that consumers in the west pay for losses that they clearly do not cause, while consumers in the east pay for less than half of the losses that they do cause. But at the $20.30 price, no consumer willingly would pay $21 to the eastern generator. Instead, all consumers would make bilateral arrangements to buy power from the western generator.
The scaled marginal cost approach produces a similar result. Applying the California version to the example of Figure 1, this approach would find that the western generator is responsible for 8K MW of losses, while the eastern generator is responsible for reducing losses by 2J MW. Together, they together account for the 6G MW lost in transmission. For output actually delivered to consumers, the western generator's break-even price must exceed its $20 incremental cost by enough to cover losses, while the eastern generator, which is credited with reducing losses, would have a break-even price less than its $21 incremental cost.[Fn.3] Because consumers are not directly responsible for the costs of losses, consumer prices will tend to be the same under the scaled marginal cost approach as under the postage stamp approach. Again, there are incentive problems that, under scaled marginal cost pricing, induce consumers to buy all of their power from the western generator.
When consumers buy all their power from the western generator, dispatch changes as illustrated in Figure 2. The western generator provides all output while the eastern generator stands idle. This dispatch results in total costs of $18,160, $50 more than the efficient dispatch shown in Figure 1.
Figures 1 and 2 illustrate the general result that marginal cost pricing has incentives that are always consistent with efficient dispatch, while postage stamp and scaled marginal cost pricing generally provide incentives that are not consistent with efficient dispatch.
Price vs. Non-price Methods
Transmission congestion occurs when flows through certain transmission facilities are at or near the physical limits of these facilities. To manage this common occurrence, system operators must reduce generation upstream from the congested facilities and increase generation (or reduce loads) downstream from the facilities. The California, New York and PJM ISOs accomplish this balance primarily through price-based mechanisms. The ERCOT, MISO and New England ISOs manage congestion primarily through non-price mechanisms. However, ERCOT is considering adopting an approach similar to that of California, MISO includes a price-based redispatch (for new transmission customers) that is similar to California's, and New England is in the late stages of developing an approach similar to that of New York.
Price-Based Methods. The ISOs have two different schemes for using price to manage transmission congestion. The New York and PJM ISOs take voluntary bids by which individual generators (or loads) offer to increase or reduce output for the purpose of reducing flows over constrained transmission facilities. These bids serve as the basis for calculating locational prices throughout the power system.
The California ISO, by contrast, takes voluntary bids by which merchant firms agree to redispatch their generation and loads for the purpose of reducing flows. Each merchant firm expressly is prohibited from offering to change only the output of a single generator or only the load of a single group of consumers. Instead, the merchant firm must offer the ISO a balanced redispatch in which, for example, the merchant increases output from one generator while equally reducing output from another generator. Thus, each merchant firm's resources fully serve its load both before and after the ISO conducts its redispatch. This balanced redispatch intentionally precludes the ISO from implicitly arranging trades among merchants. Because such trades almost always will allow cost reductions, merchants have the opportunity to trade power among themselves after the ISO conducts its redispatch.
To understand the differences between the New York-PJM and California methods of managing transmission congestion, consider the example presented in Figure 3. In this figure, the power system consists of three busses that are connected by identical transmission lines. Each line is able to handle up to 600 MW of flow, and no power is lost in transmission. The physics of this simple system are such that if 1 MW of power were to be produced by a generator at bus 1 to serve a load at bus 3, two-thirds of that energy would flow over the direct line, 1-3, while one-third would flow over the longer routes of lines 1-2 and 2-3.
We assume that there are three merchant firms, A, B and C, each of which serves 300 MW of load at each bus; so each merchant serves 900 MW of load and each bus has a 900-MW aggregate load. The merchant firms each have one generator at each bus, with the italicized capacities and per-megawatt-hour incremental costs shown. For simplicity, we assume that the merchant firms do not exercise market power and, therefore, that they bid their incremental costs.
If there are no transmission constraints, each generator's energy output would be as indicated to the right of its incremental cost. Every generator with a cost below $18 per megawatt-hour would operate at maximum, every generator with a cost above $18 per megawatt-hour would produce no power, and C1, the $18 per megawatt-hour generator, would provide exactly the quantity of power that is required to balance supply and demand. If price equaled $18 per megawatt-hour at each bus, then all generators willingly would dispatch themselves, as shown in Figure 3.
Because so much cheap power is available at bus 1 and power is so expensive at bus 3, the unconstrained dispatch leads to 700 MW of flows through line 1-3. Given that we assume loads to be fixed, some action must be taken to redispatch generation so that the line flow is brought down to its 600-MW limit.
The New York and PJM ISOs use bids from market participants to find the least-cost generation dispatch. They then set the price at each power system bus equal to the bid-based "cost" of getting one extra megawatt to that bus, given that all transmission constraints are respected. The price at each bus thus equals the best market price available to serve that bus while respecting constraints.
Figure 4 shows the resulting dispatch and prices, the latter of which are determined from the incremental costs of the two marginal generators: unit B1 at $15 per megawatt-hour and unit A2 at $25 per megawatt-hour. Because a load increment at bus 1 would be served by unit B1, the price at bus 1 is $15 per megawatt-hour. For a similar reason, the price at bus 2 is $25 per megawatt-hour. Bus 3 is trickier. Because of how generator dispatch affects loads over the constrained line 1-3, a 1-MW load increment at bus 3 is most cheaply served by increasing output by 2 MW at bus 2 while reducing output by 1 MW at bus 1. That produces a bus 3 price of $35 per megawatt-hour (= 2 x $25 - 1 x $15).[Fn.4]
Not coincidentally, the prices shown in Figure 4 induce the lowest-cost-possible dispatch of this power system. At each bus, the lowest-cost generators operate at maximum, the highest-cost generators produce no power, and the generators with incremental costs equal to price provide exactly the quantities of power that are required to both serve load and meet the transmission constraint.
An important feature of congestion pricing is that it almost always results in the total spot value of consumption being greater than the total spot value of generation. In Figure 4, for example, the value of consumption (measured by multiplying the load at each bus times the price at each bus) is $67,500, while the value of generation (measured by multiplying the generation at each bus times the price at each bus) is $49,500. The $18,000 difference represents congestion charges that, as discussed below, are indirectly returned to consumers through reduced transmission access charges.
Although the Northeastern ISOs all have nodal pricing of generation, they do not all have nodal pricing of loads. In particular, because of implementation challenges posed by nodal pricing, New York currently has zonal pricing of loads. For example, if Figure 4 had a "western zone" that included busses 1 and 2, all loads in the western zone would pay a price of $20 per megawatt-hour, which is a load-weighted average of the nodal prices in that zone. The disadvantage of zonal pricing is that it can lead to efficiency losses. When zonal pricing is applied to loads only, the efficiency losses will be slight if load is unresponsive to price (as is usually the case). If zonal pricing were applied to generation, however, the efficiency losses could be great because generation is usually very responsive to price. In Figure 4, for example, an eastern zone price of $20 per megawatt-hour would radically change dispatch at busses 1 and 2 and violate the constraint on line 1-3.
California finds a different solution to the problem of managing congestion among zones. Assuming that merchant C contracts to purchase 100 MW from merchant A, each merchant begins with a balanced schedule that has 900 MW of load and 900 MW of resources. To manage congestion, the ISO accepts bids for balanced redispatch by each merchant firm. Each bid indicates the resources that each merchant will redispatch.
Table 4 shows the merchant bids that correspond to Figure 3. Each bid is an offer to redispatch generation in a way that will reduce flows on line 1-3. The table's top line, for example, indicates an offer by merchant B to increase generator B3's output by up to 200 MW while reducing generator B1's output by the same amount. If merchant B exercises no market power, it will bid $24 for each megawatt-hour redispatched, which is the difference in the incremental costs of B1 and B3. Because of the power system configuration, each megawatt of the output shift from B1 to B3 will reduce the flow of line 1-3 by two-thirds of a megawatt - hence the power distribution transfer factor (PDTF) of -0.6667. Merchant B's 200-MW output shift therefore would reduce the line 1-3 flow by 133.33 MW (200 MW x 0.6667). The effective price per megawatt-hour of line flow reduction is $36 ($24/0.6667).
The ISO needs to find bids that give a total of 100 MW of congestion relief - the 700-MW flow before redispatch minus the 600-MW flow limit. The ISO picks the bids that are cheapest in terms of their effective cost per megawatt-hour of line flow relief. In this example, the ISO fully relieves congestion by accepting part of merchant B's cheapest bid. That bid, which is the marginal source of congestion relief, sets the congestion price of $36 per megawatt-hour of flow over the constrained line 1-3.
Figure 5 shows the resulting dispatch by the ISO. This dispatch respects the 600-MW line flow constraint and allows each merchant to maintain a balanced schedule. This dispatch costs $1,200 more than the efficient dispatch found by the Northeastern ISOs. Power traders can easily correct part of this inefficiency. Without affecting total quantities of power produced at each bus, merchants B and C can shift 150 MW of output from generator C1 to generator B1, thus reducing costs and increasing their profits by $450 (150 MW times the $3 difference in generator incremental costs).
The merchants will have a difficult time finding the remaining $750 of cost reductions, however. Part of the reason is that the needed redispatch involves plants at multiple busses and with multiple owners; the additional cost savings require that generator A2 increase output by 300 MW while generators B1, B3 and C1 decrease output by 100 MW, 150 MW and 50 MW, respectively. But the merchants sometimes will not want to undertake such trades because of the incentives created by the pricing of the transmission constraint between busses 1 and 3. For example, while Table 4 indicates that the California price for this constraint is $36 per megawatt-hour, the efficient price is only $30.[Fn.5]
In all fairness, California's main transmission constraint ("Path 15") does not have the sort of loop flow problem that appears in the example of figures 3 through 5. In California, therefore, dispatch generally will be almost as efficient, and sometimes equally efficient, as dispatch of transmission in the New York and PJM markets.
Regional power systems that are considering imitating the California approach to transmission pricing should be aware, however, that loop flows can compromise the efficiency of this pricing approach. There nonetheless is a simple remedy for the inefficiencies of the California approach, a remedy that would improve market efficiency with or without loop flows. This remedy is to allow the ISO to accept unbalanced bids for congestion redispatch, so that the ISO, for example, can reduce the generation of one merchant firm while increasing the generation of another merchant firm. This remedy has been explicitly rejected in California because of a belief that such voluntary bids from voluntary bidders would allow the ISO to "force" trades among market participants.
Non-Price Methods. ERCOT, MISO and New England primarily rely upon two non-price methods to assure that transmission limits are respected. The first method has the ISO (or some other party) pay generators or loads for redispatch; the costs of this redispatch are then recovered from designated customer groups. The second method has the ISO curtail use of transmission facilities according to the access priority assigned to each transmission customer. California, New York and PJM, by contrast, rely upon non-price methods only when price-based methods are not sufficient to fully manage congestion.
The different priority schemes used by ISOs display a common pattern. Customers taking network service and firm point-to-point service share an access priority that is higher than that of non-firm point-to-point service. Among the customers taking the same kind of service, those with longer-duration contracts (e.g., a year) have higher access priority than those with shorter-duration contracts (e.g., a week). Within each of these duration subcategories, priority usually is given to the customers who signed contracts first, though New England also gives higher priority to customers who pay higher prices for their service.
Following such a priority scheme, ERCOT manages congestion by first curtailing non-firm transactions, then short-duration firm transactions and then longer-duration firm transactions. If flows are still too large, ERCOT asks merchant firms to redispatch their planned resources. Each merchant's redispatch responsibility is proportional to its load on the downstream side of the transmission constraint, and each merchant bears the costs of its own redispatch. If further congestion relief is required, ERCOT orders merchants to dispatch their unplanned resources, the costs of which are shared by all customers that benefit from the redispatch.
Through its congestion management scheme, ERCOT assigns redispatch responsibilities without regard to cost. Because merchant firms can have very different costs of redispatch, redispatch costs could be reduced (to the benefit of merchants and consumers) if merchants could trade their redispatch responsibilities among themselves. Without such trades, ERCOT's redispatch scheme could impose unnecessarily high risks on merchant firms, with the costs for some merchants being much higher than the costs for others.
MISO has a priority scheme that curtails non-firm service customers first, then redispatches resources and then curtails firm service customers. MISO requires that network customers and generation owners redispatch their resources at MISO's request. This redispatch is supposed to be undertaken at least cost. The costs of this redispatch are distributed among all the transmission owner's customers, including bundled native load. If the redispatch results in greater available transmission capability, however, the customers using this capability pay for the redispatch. MISO itself does not set generation prices.
Although ISO New England has a comparable priority scheme, it manages congestion primarily through a two-step redispatch process that is similar to the way California manages congestion within zones. First, it determines the uniform market-clearing price that would prevail if there were no transmission constraints. Second, it pays generators to redispatch themselves to relieve the actual transmission constraints. For increases in output, generators are paid their bid price, which always exceeds the uniform market-clearing price. For reductions in output, generators are paid the amount by which the uniform market-clearing price exceeds their bids, thus compensating them for the profits they would lose by reducing output. This system is problematic because it overpays generators upstream from transmission constraints and it encourages gaming by generators located downstream from the constraints. Recognizing such problems, New England plans to switch to price-based congestion management during the next year or so.
A key feature of all the priority schemes is that there is little or no relationship between a customer's access priority and the value a customer derives from transmission service. Because of the small role of price in assigning priorities, the market participants with the firmest service, in many hours, may have lower-valued transactions than market participants with less-firm service. Furthermore, market participants often lack convenient means for trading transmission priorities. The result is that high-valued transactions can be curtailed while low-valued transactions continue.
MISO provides an example of the ways that non-price congestion management schemes are accompanied by restrictions on secondary trades in transmission rights. MISO has two particularly important restrictions that by no means are unique to MISO.
First, there are ceilings on the prices at which transmission rights can be traded. A purpose of these ceilings is to control market power, so MISO intends to eliminate the ceilings for the sellers of those rights who demonstrably lack market power. Until they are eliminated, however, the ceilings will inhibit transmission rights trades for precisely those hours when transmission is most valuable and most congested.
Second, MISO requires that the physical feasibility of serving the designated resources and loads of the transmission rights buyer be roughly the same as that of serving the resources and loads of the transmission rights seller. This seemingly reasonable restriction means that, in MISO, transmission rights are non-homogeneous commodities that must be traded on a case-by-case basis. Indeed, MISO can require that individual trades be preceded by system impact studies. Note that, by contrast, the ISOs with price-based congestion management have turned transmission rights into a homogeneous commodity that is freely traded every hour.
Transmission Access: Price vs. Value
For all the ISO markets, transmission access pricing boils down to dividing some measure of transmission system costs by some measure of transmission system use. The differences among the access charges reflect variations in how the different markets define costs and use.
With respect to costs, ERCOT, MISO and New England have access charges that basically cover transmission revenue requirements. California, New York and PJM design their access charges to recover only that portion of transmission revenue requirements that are not recovered through congestion charges. To foster retail rate stability during a transition period, all ISO markets except ERCOT initially have set different access charges for the customers of each transmission owner. However, all ISO markets (except New York) plan to move toward system-wide access charges based upon system-wide costs.
With respect to use, all the ISO markets measure transmission use according to loads (including exports), but they use different measures of load. California and New York base access charges on the customer's megawatt-hour consumption. ERCOT charges customers taking firm service according to their coincident peak loads in the summer months, and requires customers to purchase firm service to assure that aggregate transmission revenues will be sufficient to cover transmission costs. In MISO and New England, access charges for network service depend upon the customer's hourly energy load at the time of their transmission owner's monthly peak, while access charges for point-to-point service depend upon the customer's megawatt capacity reservation. PJM's access charges for network service depend upon the customer's hourly energy load at the time of its transmission owner's annual peak, while access charges for point-to-point service depend upon the customer's megawatt capacity reservation.
ERCOT initially recovered 30 percent of transmission costs through megawatt-mile pricing. This system charged transmission customers according to the incremental power flows attributable to their transmission transactions and according to the distances over which their power traveled. The rationale for this system was that costs may be related to the volumes and distances of transmission flows. However, the distance that a transaction covers is not necessarily related to the power losses created by that transaction, to the congestion created or alleviated by that transaction, or to the capital expenditures that might ultimately be undertaken to support similar future transactions. ERCOT recently replaced this megawatt-mile pricing approach with the simpler postage stamp pricing method.
An important feature of transmission access charges is that they bear little relationship to the value of transmission access. They are related neither to the costs of serving particular locations nor to the costs of serving different time patterns of transmission use. A customer's willingness to pay for access is related to its ability to commit to its future use of the transmission system and to the customer's intolerance for the risks of service curtailment. But that ability to commit and intolerance for risk are not necessarily related to the value that the customer might derive from transmission use in each hour. All that transmission access charges surely do is recover transmission costs that are not recovered through other means. They do not allocate scarce transmission resources to their highest-valued uses.
Winners and Losers
The various transmission access and pricing schemes offer different relative benefits to different market players.
Power traders sometimes gain from market inefficiencies, particularly in the pricing of transmission congestion and access. For example, California's congestion management approach offers power traders profit opportunities that don't exist in the markets run by the Northeastern ISOs. Non-price congestion management arbitrarily can confer benefits and impose costs on different market players. Some congestion management procedures can create financial risks that power traders should avoid.
A generator's interest in transmission access and pricing schemes depends upon its costs and location. High-cost generators can benefit from artificial barriers to trade (such as can be created by distance-sensitive pricing) that keep faraway competitors out of their local markets. Low-cost generators can benefit from minimal barriers to their entering faraway markets. Generators in high-value locations generally will get higher profits from locational prices than from postage stamp prices. Generators in low-value locations, far from load centers, might like the relatively high prices that they can get under postage stamp pricing.
Generators with local market power also will tend to profit from postage stamp pricing. Local market power problems originate from transmission constraints that limit local competition. Locational pricing merely makes local market power more obvious than does postage stamp pricing. But locational pricing also has the effect of attracting competition from both potential new generators and potential curtailable loads. Postage stamp pricing, by contrast, spreads the costs of local market power over the whole power system's load, thereby reducing the incentives for individual market participants (and regulators) to creatively respond to the high costs of market power.
Consumers generally benefit from transmission pricing that results in low-cost dispatch. Dispatch costs are likely to be lowest when transmission facility use is allocated by price. For consumers, distance-sensitive pricing and non-price congestion management can be costly.
Consumers' location-based interests in transmission access and pricing schemes are almost exactly opposite those of generators. Consumers in low-cost locations can benefit from locational pricing, while consumers in high-cost locations can benefit from postage stamp pricing. Consumers near cheap generation can benefit from policies that deny access to faraway consumers, while consumers who are far from cheap generation can benefit from better market access.
In spite of the foregoing differences, there are two characteristics of transmission pricing that would be beneficial for the vast majority of market participants.
Transmission congestion should be managed through price-based mechanisms. For consumers, price-based mechanisms keep the costs of congestion management down. For power traders and generators, price-based mechanisms can provide profit opportunities. Non-price congestion management mechanisms, by contrast, can be costly for consumers and impose unnecessary financial risks on traders and suppliers. Non-price mechanisms should be used only as a last resort, in those circumstances in which price-based congestion management is insufficient to resolve congestion.
Transmission rights should be tradable in well-organized secondary markets. There are many different ways that transmission rights can be initially assigned: Customers can choose their tariff options, rights can be auctioned, transmission rights can be granted to investors in new transmission facilities or native loads can be given grandfathered rights. The initial assignment is important because of its wealth implications and because of the role that transmission rights can play in reducing market participants' risk. But once those rights are granted, they should be tradable so that transmission is allocated, in or near real-time, to its highest-valued uses.
Well-organized secondary markets in transmission rights are good for the holders of transmission rights because such markets allow the rights holder to profit when it can sell its rights to others who value transmission service more highly. They are good for generators and consumers who spontaneously can buy transmission access when they need it. And they are good for power traders who thereby can have one additional (and very important) tool for facilitating their trades and managing their risks.
Laurence D. Kirsch is a senior economist at Laurits R. Christensen Associates Inc. in Madison, Wis. He may be contacted at firstname.lastname@example.org.
1 In California, each generator's responsibility for losses equals a "marginal loss rate" times a "loss scale factor." The marginal loss rate is defined as the power that would be lost if the generator provided 1 MW that served all power system loads on a pro rata basis. The loss scale factor is the ratio of actual losses to the losses that would be recovered if the marginal loss rates were applied to all generators.
2 $18,232 = $20 x 500 + $21 x 392. $122 = $18,232 - $18,110.
3 The figures for the scaled marginal cost approach are derived as follows. Without losses, the 892-MW system load is met by 744I MW and 147G MW from the western and eastern generators, respectively. The marginal loss rates for these generators are 0.02197 (=0.05 x 392/892) and -0.02803 (=-0.05 x 500/892), respectively, where 0.05 is the marginal line loss for this case in which line flows average 250 MW. Total marginal losses are 16.33 MW (=0.02197 x 744I) and -4.13 MW (=-0.02803 x 147G), respectively. Scaled losses are 8.36 MW (=6.25 x 16.33/(16.33-4.13)) and -2.11 MW (=6.25 x -4.13/(16.33-4.13)), respectively. Break-even prices are $20.22 = $20 x (753J)/ 744I and $20.70 = $21 x (145J)/147G, respectively.
4 Two-thirds of any load increment at bus 1 flows to bus 3 over line 1-3. One-third of any load increment at bus 2 flows to bus 3 over line 1-3. Decreasing bus 1 output by 1 MW therefore decreases flow on line 1-3 by two-thirds of a megawatt, while increasing bus 2 output by 2 MW increases flow on line 1-3 by that same two-thirds of a megawatt. In simultaneously decreasing bus 1 output by 1 MW and increasing bus 2 output by 2 MW, flow on line 1-3 is left unchanged, while supply increases by the 1 MW needed to serve a 1-MW load increment at bus 3.
5 In the context of the efficient dispatch of Figure 4, a 1-MW increase in the capacity of line 1-3 would allow generator A2 to reduce output by 3 MW while generator B1 increased output by 3 MW. The cost change attributable to a 1-MW relaxation of the line 1-3 constraint thus is -3 x $25 + 3 x $15 = -$30.
Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.