Divide the grid by usage (em local vs. regional. Apportion costs accordingly, to energy customers by fixed charge, and power producers by flow and distance.
Traditionally, utilities have received transmission costs through an average, rolled-in access fee, or postage-stamp approach. In a deregulated environment, that approach will lead to distorted pricing.
And not just because of transmission-line congestion.
Much of the current debate over electric transmission pricing has centered on the various competing methods of congestion pricing, such as zonal vs. nodal pricing, or financial vs. physical rights. Meanwhile, a related problem receives scant attention. Simply put, congestion pricing will not collect enough revenues to fully recover the revenue requirement associated with transmission-line assets.
To recover the full revenue requirement of a transmission plant, any congestion-pricing regime should include two separate access charges. Retail consumers would pay the first charge, a flat, postage-stamp fee to recover local transmission system costs. The second charge, a regional access fee reflecting transmission usage based on flow and distance, would be assessed to all power producers that use the grid. In effect, this two-part fee divides the transmission grid into two segments: 1) Low-voltage lines that serve a local function, and 2) High-voltage lines that serve regional needs.
The logic behind this proposal rests on both engineering and economic rationales. By assigning separate, mutually exclusive access fees to end users and power suppliers, the grid will send correct, long-run pricing signals for siting generation. In this way, the electric transmission price will really consist of a combination of three competing pricing methods: 1) short-term congestion pricing, 2) long-term, postage-stamp pricing for local service, and 3) long-term flow- and distance-based pricing for regional use.
Generation vs. Load: The Myth of Common Costs
Many proposed methods of pricing electric transmission concentrate on optimizing use of the existing system and providing incentives to build new transmission facilities. These methods, however, rely on the principle of setting price equal to short-run marginal cost. %n1%n In fact, the marginal cost of transmission operation is usually zero or close to zero. %n2%n Thus, marginal-cost pricing will not lead to the full recovery of initial investments in transmission. Many authors note this fact, recommending an additional fixed charge to correct this revenue deficiency. %n3%n
How should this fixed charge be designed?
First of all, the fixed-charge component, the access fee, should be large enough to recover embedded costs of transmission, but not so large as to discourage access to end users or generators. Clearly, the collection of sunk costs should not distort operational decisions. However, this idea should not warrant abdication of the principle of cost responsibility according to cost causation. %n4%n Each set of customers should face the full cost of supplying its product.
In the past, under vertical integration, the utility planner had access to costs for both transmission and generation. To meet expected load growth, electric system planners would weigh the benefits and costs of additional resources in generation versus transmission. New plant construction could offset transmission upgrades, and vice versa. Rolled-in, bundled rates were assessed to all customers to collect expenses, depreciation and the regulated return on investment. Fixed costs related to transmission were treated historically as common costs (em costs not susceptible to attribution to an individual service. %n5%n
Today, however, the transmission grid is no longer just a black-box conduit for power supply to end users. Instead, the transmission grid is a complex transportation system for electrical power for end users that also provides the competitive generation market with access to various load centers. With its unbundling Order 888, the Federal Energy Regulatory Commission has identified transmission pricing as susceptible to manipulation by transmission-owning utilities to favor their own generation. To realize its goal of creating a fully competitive market in electric generation, the FERC has recognized the need to separate transmission-related costs from generation costs. Access, then, is no longer a common cost.
In both a financial and physical sense, the existing transmission system is now linking two sets of transmission users: end users and producers. It is important to see how these two sets of customers use the system. If transmission provides functionally different services, then each set of customers should be charged based upon their different usage. The problem now evolves into one of defining the "bright line" separating categories of transmission customers on a cost-causation basis.
With few exceptions, however, electric power transmission does not occur over discrete paths but rather diffuses throughout the network. For example, Figure 1 shows the incremental flow distribution in California for a 100-MW injection of power at Mohave in Nevada, supplying a 100-MW increase in WEPEX's wholesale load in California. Most of the 100-MW injection is transmitted directly to the load via California's utility interconnections, but 23 MW is transmitted to the load via loop flow north, south and east through the remainder of the interconnections to the WSCC system.
Looking at the historical development of power systems in the U.S., one finds that local utility power systems developed as independent entities with their own generators feeding their own load over their own lines. It was not until the late 1950s that the utilities in the WSCC developed interconnections involving long, extra-high-voltage lines. These connections offered many advantages, including continuity of service, reserve sharing and economy interchanges. The hypothesis appears obvious: higher-voltage lines serve a regional function (interconnection); lower-voltage lines serve load, a local function.
The principle of cost causation suggests that end users pay local transmission costs and producers pay regional costs. What's left is identifying those lines having local or regional functions.
Two-Tiered Pricing: The Local/Regional Split
This local-regional split between transmission facilities helps resolve the question of who should pay what fees to gain access to the grid.
For power producers, who use the regional grid at high voltage levels, a measure of usage based on MW-miles offers a consistent method to allocate revenue requirements and access fees. A simple allocator for regional revenue requirements can be constructed by using the ratio of a generator's regional MW-miles relative to total system regional MW-miles. In this way, each generator can be assigned a lump-sum regional access fee. Since generators pay for the regional access fees, there is no cost shifting between existing utility ratepayers.
For end users, who use the local grid at low voltage levels, the situation changes. As discussed (see sidebar), the accuracy of the power-flow model does not support application of a MW-mile allocator in WEPEX for voltages 230 kV and below, which tend to serve load. Instead, revenue requirements associated with these local transmission facilities should be assigned to end users on a postage-stamp basis, recognizing the local network functions these facilities provide. Existing ratepayers will see reduced transmission rates, because they pay only for local transmission.
Distinct, MW-mile usage based access fees will recover the fixed cost of regional transmission facilities. Based on a simulated "snap-shot" of power flows, these access fees are not determined in "real time." They do not reflect marginal costs of congestion. Within the context of WEPEX, access fees are not a form of capacity reservation tariffs and are not associated with retradeable rights. %n6%n
However, allocating costs solely on a single snap-shot, the summer-peak condition may not fairly represent regional transmission use. As load distributions change over load cycle, generator use of the transmission system changes. Additional power-flow simulations can further define the system's use, enabling fair allocation of regional transmission system cost among generators. One way to do this is to account for the seasonal and physical natures (revealed in low-capacity factors) of many generators when calculating the regional access fee. %n7%n
Why Not a Single Access Fee?
A single, postage-stamp access fee, as proposed by some, would fail to give proper locational signals to siting new generation facilities. When paid only by end users, it leads to cross-subsidies between customers.
Putting aside the question of ancillary services, buyers and sellers of wholesale power in WEPEX will tend to concentrate on generation costs. Delivery costs for WEPEX are not transaction-related, because the fixed costs for transmission are allocated to the end user on an average-cost basis.
Three distinct situations can illustrate the inaccuracies that would result from a single, postage-stamp access fee. For simplicity, assume all end users are residential customers. Based on the current rate design, a utility-specific access fee is calculated on a volumetric basis. The IOU's transmission revenue requirements are divided by the forecasted sales of the utility ($/kWh). Then, to ensure the total revenues collected are not more than the authorized revenue requirements (avoid "and" pricing), actual monthly congestion rents divided by actual monthly sales ($/kWh) are subtracted from the access fee.
PX Only. Assuming the simplest case, all power is bid into the power exchange. Power from all sources is commingled and establishes the market-clearing price (MCP). Losses and locational congestion fees define the marginal costs for transmission services. A single access fee, to recover all remaining sunk transmission costs, is paid by end users. Both local and non-local generators pay nothing for transmission access. Consequently, siting decisions will not include full transmission costs. In addition, even though more capital assets are used to bring non-local power to the load, higher access fees are not associated with non-local generators. Incorrect locational pricing signals are given for the location of new generation facilities. In this PX-only case, end users are not contractually linked to suppliers and all contribute to the recovery of transmission revenue requirements on the same average $/kWh basis.
PX with CFDs. Next, assume an end user enters into a contract for differences with a non-local generator. If the end user pays the same single access fee regardless of where the power is generated, then a generator would benefit from being non-local (lower wages, taxes and business costs, inexpensive fuel and less-stringent regulation). The non-local generator is given a competitive advantage and can under-cut the local generator's price. The CFD financially links the non-local supplier to the load. This supplier avoids paying its share of the full cost of delivery in the access fee charge.
PX with direct access. In a world with direct physical access, the subsidy becomes even clearer. Assume two functional transmission networks, local and regional. Assume transmission revenue requirements of $100 million (em $20 million related to regional transmission and $80 million related to local transmission. Forecasted sales total 25 billion kWh (em 10 billion kWh related to non-local generation and 15 billion kWh related to local generation. Assume the entire load is local. A single access fee paid only by end users would be 4 mills ($100 million/25 billion kWh). The two-component access fees would be 2 mills ($20 million/ 10 billion kWh) for non-local access and 3.2 mills ($80 million/ 25 billion kWh) for local access. The single access fee forces customers using local generation to subsidize customers using non-local generation by 1.2 mills (2.0 + 3.2 - 4.0). As the link between end user and generator strengthens, the subsidy related to a single access fee becomes clearer.
In the short run, our proposal to charge generators with distance- and flow-based regional access fees will not influence current operational decisions. The regional access fees are assigned to the generators as a lump-sum payment, irrespective of current output.
Similarly, the regional access charge should not influence the PX bid price in the short run, but could help level the field between local and non-local generation.
In Southern California, most local generation is either gas-fired or nuclear. With gas-fired generation dispatched as the marginal resource during peak and shoulder periods, gas-fired resources set the MCP in the PX. Once the decision to pay the regional access fee is made, the fee is no different from any other fixed cost. If local gas-fired generators try to pass on the cost of the regional access fee in their PX bid price, they may not be dispatched. The incentive remains for the local generators to bid their marginal cost of generation.
Since non-local generation consists mainly of lower marginal-cost resources like coal, hydro and nuclear, it does not set the PX price. These resources are infra-marginal. Thus, if they are paid the PX MCP, they receive a contribution to their fixed costs. The regional access fee paid by the non-local generators would transfer some of this producer's surplus from the non-local generator to end users, who are no longer responsible for the regional access fee.
For the majority of off-peak periods, however, non-local generators are on margin. They are selling into a market characterized by excess capacity. During this period, it would be hard to pass onto consumers the regional access fee. Again, the PX MCP should not increase by the regional access fee. The long-run price efficiency gained from charging the generators the regional access fee will not result in a higher short-run PX MCP.
In the long run, two major consequences arise from having the generator, instead of end users, pay for fixed costs of the regional transmission system. First, generators will see their profitability reduced. Second, the fees may influence the siting of new power plants. Two examples (reflecting congestion) can help illustrate the potential impact of assigning regional access fees to generators on power plant siting.
No Congestion. The first example, often quoted by advocates of the single (postage-stamp) access fee paid by end users, implicitly assumes that excess capacity exists on the regional transmission system. That is, no new regional network transmission facilities are needed to accommodate a new generator. It also assumes production costs of new, local generation will prove more expensive than new, non-local generation. For example, new, local generation may be a gas-fired combined cycle plant, while a new, non-local plant may be a mine-mouth coal plant. %n8%n The decision on where to site a new power plant is based on operating costs and line losses.
If a generator is assigned a distance-based regional access fee, then the decision to locate a power plant must now consider this additional expense. The generator, looking for its least-cost location, may site the plant closer to the load, and could choose a plant with higher production costs rather than one with lower production costs that is located farther from the load. This dilemma highlights an important issue: When making a long-term decision, like siting a power plant, the total cost of delivered power must be considered, not just the variable cost of producing energy.
For WEPEX, however, this is a moot point, because it is not free of congestion. Regional transmission lines leading into California are fully loaded many hours of the year. %n9%n
Congestion. The second case assumes congestion on the regional transmission path. In this case, if the generator pays for the new transmission facilities, its siting decision will reflect the total delivered cost of power. However, in most instances, an individual firm or group of firms will not find it economical (because of the free-rider problem) to build new regional network transmission facilities. As a backstop, additional transmission costs would be rolled into the existing transmission costs that are paid for by end users.
Here, again, if existing generators pay the regional, distance-based access fee, the least-cost option takes into account both transmission and generation costs, just as the system planner would do, working for the traditional, vertically integrated electric utility.
In a competitive market, a distance-based fee, paid by power producers for access to the regional grid, should align the producer's least-cost option with society's goal to minimize the total cost of delivered power.
Cliff Rochlin is a market advisor to the Energy Transportation Services Business Unit of Southern California Gas Co. Roger Clayton is a Consulting Engineer with General Electric Power Systems Energy Consulting Department. This article is a condensed ver
sion of a paper written by Clayton and Rochlin. The full the paper is available on request.
A Technical Study: Power-Flow Analysis
Transmission Line Use in WEPEX
Southern California Gas commissioned General Electric's Power Systems Energy Consulting department to analyze power flow and develop the technical basis for a flow- and distance-based transmission-pricing algorithm to use in the Western Power Exchange. The study analyzed transmission usage in WEPEX and identified a clear local and regional distinction in the use of California transmission lines. In California, regional lines tend to be above 230 kilovolts, while local lines are 230 kV and below.
The power-flow analysis was constructed to model a wholesale pool environment, where power injection at any generating node is balanced by a load change distributed equally to all load centers. %n10%n The resulting change in flow in each line can be denoted by distribution factors that allow flow on each line to be calculated for any dispatch condition. %n11%n Figure 1 shows that 34 MW of the Mohave power injection flows on the Mohave Lugo 500-kV transmission line. The proportion of Mohave power flowing on this line is a constant, assuming system linearity, and is defined by a DF of 34 percent (or 0.34 per unit). Every line in the system will have a unique DF associated with every generator in the system. The characteristics of these DF and the resulting line usage (MW-miles) are used to classify lines according to their local and regional function.
Study Assumptions. It was assumed that the electric utility structure in California would be based upon the WEPEX power exchange and independent system operator proposal by three major California IOUs. The PX will develop a day-ahead, wholesale auction, on an hourly basis for the following 24 hours. The next-day generator merit order will be based on bid price and transmission constraints. As the load changes, individual generators will be added in merit order.
Therefore, a key observation is that a unique generator/load pair can be identified for each generator. The load will be that portion of the total system load that requires the commitment and dispatch of the next generator in the merit order to supply the PX/ISO wholesale supplied load. This observation allows a unique distribution-factor matrix to be developed for each transmission line (1,056) and generator (245) in the WEPEX system.
Distribution factors for each generator will have values between 0.0 and 1.0 per unit for each line in the system. The actual line flow due to each generator can be calculated by multiplying the DF by the output of the generator. Distance is introduced by multiplying the actual flow in each line by its length to give usage in terms of MW-miles for each line and generator in the system.
MW-Mile Analysis. The WEPEX model results for summer peak conditions in 2000 are presented in Figure 2, which shows, by voltage level, a histogram of MW-mile usage as a function of distribution factor. A significant observation is that MW-mile usage of 69 kV, 115 kV and 230 kV lines are clustered at low distribution factors. In fact, using a DF of 1 percent as a cut-off criterion, %n12%n we observe that:
• 77.4 percent of all 69 kV MW-miles are at DF less than or equal to 1 percent. The corresponding usage for 115 kV and 230 kV are 62.8 percent and 30.9 percent, respectively.
• By contrast, only 1.4 percent of all 500 kV MW-miles is at DF less than or equal to 1 percent.
Therefore, we conclude that there exists a fundamental, generic difference in transmission line usage in WEPEX at voltage levels 230 kV and below and at voltage levels above 230 kV. %n13%n These results were confirmed by similar analyses of winter peak and spring run-off conditions.
The explanation lies in the nature of system development:
• Local, lower-voltage networks ultimately serve loads. The magnitude of DF on lines in these local networks will be directly in proportion to the (small) fraction of total system load they serve. These lines have a local (load-serving) function.
• Utility systems developed historically from low-voltage local distribution networks to high-voltage, regional, interconnected networks. The magnitude of DF on lines in these regional networks will be larger than on local lines since regional lines are fewer in number with each one serving a larger fraction of total system load. These lines have a regional (interconnection) function.
The success of a flow- and distance-based pricing algorithm is fundamentally dependent upon the accuracy of the model used to develop the DF. Considering its application to local lines at 230 kV and below, Figure 2 shows that 30.9 percent of all 230-KV MW-miles are at DF of 1 percent or less. %n14%n This fact implies that the model has to have an accuracy of better than 1 percent in order for it to be a reliable means of allocating transmission access fees. The argument is even stronger at the 115 kV and 69 kV voltage levels. In fact, the accuracy of the underlying data in power flow models does not support such precision.
These observations lead to the major conclusions of the power-flow analysis:
• A flow- and distance-based transmission pricing algorithm for local lines (230 kV) would be impractical in a WEPEX context.
• However, a flow- and distance-based transmission-pricing algorithm is entirely practical for regional transmission lines in WEPEX (lines >230 kV), since 98.6 percent of their MW-mile usage is at DF greater than 1 percent.
1Alfred Kahn, The Economics of Regulation, MIT Press, 1988, Book I pp. 63-86.
2EPRI, Transmission Services Costing Framework, TR-105121-V1, April 1995, p. 9-39.
3William Hogan, "Contract Networks for Electric Power Transmission," Journal of Regulatory Economics, 4:211-242, 214, 229 (1992). See also, EPRI, supra, note 2, p. 9-4.
4While the new pricing policy adheres to the Federal Power Act's requirements of just, reasonable, and not unduly discriminatory transmission rates, FERC clearly states that does not mean one transmission price for all. "Second, a utility must allocate among individual customers or classes of customers that portion of the total revenue requirement that is attributable to providing transmission services, in a manner which appropriately reflects the costs of providing transmission service to such customers or classes of customers... Different customers may pay different rates if they use the system in different ways." See, Inquiry Concerning the Commission's Pricing Policy for Transmission Services Provided by Public Utilities Under the Federal Power Act, 59 Fed.Reg. 55031-55045, Nov. 3, 1994, FERC Docket No. RM93-19-000, Oct. 26, 1994 (transmission policy statement). The FERC statement reinforces the principle of cost causation.
5Steven Parsons, "Seven Years After Kahn and Shew: Lingering Myths on Costs and Pricing Telephone Service," Yale Journal of Regulation, Vol. 11, No. 1, Winter 1994, pp. 149-70, 151.
6However, if capacity reservation tariffs or tradable rights are developed, the access fee concept developed in this paper could be readily adapted to conform to these principles.
7The physical nature of generation refers to differences in power plant design with respect to operation. For example, coal and nuclear power plants are designed as base-load facilities and operate at high capacity factors. On the other hand, many natural gas-fired and oil-fired plants are used to follow the load. As a result, these types of plants have much lower capacity factors. To compensate adequately for the physical nature leading to distinctive length of use for these facilities, a seasonal or monthly time-of-use access fee could be used. Instead of a single, annual, flow-based snapshot, multiple snapshots could be used. For example, a monthly access fee could be based on three (peak, shoulder and off-peak hours) flow-based snapshots of usage.
8A variation on this theme would be the siting of a new gas-fired combined cycle power plant. With available transmission, the plant could locate next to a gas pipeline outside the state. By locating there, the new power plant could avoid in-state natural gas transportation costs. Since WEPEX advocates the end user should pay all of the transmission access fee, except for electric transmission losses, the new plant can move its power to the in-state load over electric transmission lines for free. The siting of the power plant depends upon the trade-off between electric transmission losses and foregone intra-state gas transportation rates.
9"[S]ignificant segments of the bulk power transmission system encompassed in the CAL-ISO are presently operated at limit during periods of peak load and periods of peak interchange." Task 1 Draft Report: Current Reliability Criteria for Location-Dependent Ancillary Services for the California ISO Trust Transmission Reliability Study, Report # 32-97, Power Technologies Inc., April 27, 1997.
10The same method of spreading a change in load equally to all load centers is used by WEPEX to assign losses to each generator.
11Distribution factor (DF) = D Line flow due to generator injection of power (MW)
D Generator injection of power (MW)where:
D Generator injection of power is some fraction of a generator's power output.
12The Sacramento Municipal Utility District used a 3-percent cut-off value in a power-flow study that modeled eight bilateral contracts to discern the transactional use of California transmission facilities. The North American Electric Reliability Council (May 1995) has suggested a 2-3 percent value as a reasonable level of significance for distribution factors when evaluating transmission line ratings. The General Agreement on Parallel Paths experiment for pricing loop flow uses 5 percent of a transaction to determine the "prime path." See, Allegheny Power Srv. Corp., et al., (Order Accepting for Filing GAPP Experiment Participation Agreement), FERC Docket No. ER-97-697-000, March 25, 1997, 78 FERC ¶ 61,314.
13As the ISO examines the functional use of the 230-kV system, pursuant to the California legislation (AB 1890), one can expect the ISO will find many or all 230-kV lines running parallel to 500-kV lines provide the same regional function as the 500-kV system.
14Using a cut-off value of less than 1 percent is pushing the accuracy and credibility limits of the input data in the power flow model. For example, information concerning transmission lines may only have Å5 percent accuracy concerning knowledge of actual length, height and weather conditions. Similar variations can exist with load, generator and transformer characteristics.
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.