|
Report Contents Report#:EIA/DOE-0607(99)
Related Links |
by
|
| Historically, electricity prices have not been determined by the competing interests of suppliers and consumers in open markets. Rather, State regulators have set prices by reviewing the costs incurred by the utilities under their jurisdiction, using both equity and economic efficiency as criteria to determine how costs should be allocated to different customer groups. Restructuring of the electric utility industry has the potential to significantly alter the relationship between the prices paid by different customer classes. This paper presents a framework for evaluating price differentials among customer classes and quantifies such differences with the introduction of a regulatory preference parameter. It then explores how the prices paid by different groups may change as the generation sector of the electric power industry becomes more competitive. |
Introduction
The emergence of competitive markets for generation in the electricity industry has created the potential for a new alignment of costs and benefits among classes of utility customers. From a purely economic perspective, it is well established that the “optimal” price of electricity equals the marginal cost of generation, transmission, and distribution. In the electric power industry, however, because the costs of transmission and distribution decline with the volume of service provided, setting price equal to marginal cost may fail to generate enough revenue to recoup the total costs. To avoid this problem, rates have traditionally been set by regulators based on the average cost of producing electricity and serving the customer, including both short-run costs such as fuel and long-run costs such as plant and capital recovery.
Historically, average cost pricing has been seen as a way of ensuring that revenues cover total costs. Although optimality may be defined as efficient in a strictly economic sense, many other considerations, such as equity and simplicity, play a role in rate-setting, leading to different rates by customer class. This paper provides a framework for evaluating rate differentials under rate-of-return regulation and explores some possible outcomes when customer choice is allowed.
Background
The changing nature of the electric utility industry will undoubtedly modify the burden of costs among customer classes. The generation component of the electricity market is being separated from transmission and distribution. While nearly all customers are expected eventually to benefit from the introduction of competition in the generation function, the rate and degree of such benefits may vary by customer class.1 Regulators’ goals in terms of efficiency and equity are not expected to change, but their ability to act upon those preferences may change as prices become more market-based.2 Competition may introduce new products and services as markets are restructured, and firms may have incentives to change their pricing to meet specialized demands.
An analogy may be seen in the U.S. natural gas market for transmission services, where restructuring resulted in a wider array of options for some customers. In particular, those customers with more flexibility in their transmission and distribution requirements were in a position to find less expensive service.3 Those users, primarily large industrial consumers, benefitted most from the restructured natural gas market. Figure 1 shows the transmission and distribution markup (the difference between the wellhead and end-use prices of natural gas) by sector from 1985 to 1997, indexed to 1985.4 As shown, the average price of transmission and distribution for industrial users declined significantly more (on a percentage basis) than that for residential users.
Figure 1. Index of Real U.S. Natural Gas Transmission and Distribution Markups by End-Use Sector, 1985-1997 [source]
While a portion of these sectoral price differentials can be attributed to differences in the cost of service, a substantial fraction can only be accounted for by price discrimination. Price discrimination can be defined as selling two or more varieties of a commodity to two or more buyers at different net prices, the net price being the price paid by the buyer, corrected for cost or quality differences. Given this definition, price discrimination is a ubiquitous phenomenon in that most firms sell several varieties of products at prices that do not fully reflect their differences in quality or cost. For example, the practice of most supermarkets of providing shoppers with coupons is a form of price discrimination, because customers who save and use the coupons pay a lower price than those who do not. Another example of price discrimination is the airline industry practice of charging lower fares to passengers who stay over Saturday night in a destination city. Individuals who do not stay over are presumed to be traveling on business and thus are judged to be relatively price insensitive. Those who do stay over are presumed to be traveling for pleasure and thus to be more concerned with the price paid.
There are two main reasons for rate differentials in regulated markets. The first is that, as in most markets, customers vary in terms of their demand elasticity (i.e., their sensitivity to price), and regulators respond to the differences by offering the lowest prices to the most price-sensitive customers. A countervailing influence in a regulated market is equity considerations. Unbridled price discrimination based on demand elasticities alone may result in price differentials that are "unfair and unreasonable." To achieve a more equitable solution, regulators may order the firm to modify its pricing strategy. While both of these motivations for price discrimination may still be present in the restructured electricity industry, the changing form of the market is likely to make the rate structure more responsive to market forces.
Pricing in Fully Regulated Markets
Consider a monopolist generator/distributor of electricity that provides service to three distinct customer classes: residential, commercial and industrial, where prices are set by the regulator. We divide the cost of electricity into its three major components: production, transmission, and distribution. In many cases the marginal cost of electricity is below the average cost. That is, if every customer class were charged its incremental cost of service, a utility would not cover its total costs. It is economically most efficient to recover this deficit through a lump-sum payment from general funds. This approach would cause the least deviation from marginal costs. However, regulators typically have found it infeasible to recoup all costs in this manner. Therefore, allocation of such costs over and above the incremental costs are decided by regulators. Such allocations, translated into rates, may be considered a form of price discrimination. Price discrimination in this sense is inherent in the development of traditional electricity rates.
Given the traditional market structure of a regulated monopoly, a second-best framework is a pricing approach whereby classes of customers with inelastic demands pay a higher markup over marginal cost than those with more elastic demands. This is generally referred to as "Ramsey pricing." The goal of this pricing strategy is to recoup the fixed costs while minimizing the distortion associated with prices in excess of marginal costs. Under Ramsey pricing, customers with the most inelastic demands pay the highest markup over marginal cost. Because those customers have the fewest options, their consumption is reduced the least. That is, the least deviation of consumption from marginal cost pricing is achieved by using this general rule. Given the constraint that the firm cannot operate at a loss, economic efficiency is attained through the use of Ramsey pricing.
The theoretical optimality of the Ramsey inverse elasticity pricing rule depends on a number of assumptions. The most important is that society considers only economic efficiency and that, as a result, it is indifferent to whether the pricing structure is "fair" and "reasonable." Since, in general, equity considerations are important, policymakers may deviate from the economically efficient outcome so as to avoid imposing "unreasonably" high prices on groups with inelastic demands.
One way of incorporating such deviations is to explicitly represent the preferences of regulators in terms of equity in the Ramsey framework. We define the price-cost margin as the ratio of the price less marginal costs relative to the total price. The price-cost margin then represents the degree to which the price exceeds marginal cost. Under the Ramsey pricing model, consumer surplus is maximized when the price-cost margin is inversely proportional to the sectoral elasticity. We can extend this framework by attributing different weights to the consumer surplus of each sector. Maximizing the weighted consumer surplus, subject to the constraint that total costs must be covered, yields the implied regulatory preference.
We can draw two major conclusions from the results of this extended model. First, the price-cost margin varies inversely with the absolute value of the price elasticity of demand. That is, those customers with the least response to price have the highest price-cost margin. Second, the implied regulatory preferences will affect the difference between price and marginal cost. Specifically, if the decisionmaker attaches zero weight to the interests of a specific customer class, then that customer class faces the monopoly price. As the implied weight that the regulator attaches to the customer class increases, the price for that customer class declines, approaching marginal cost. The mathematical representation of these preferences is shown in the Appendix.
By comparing the results of such a model with historical prices, one can deduce the implied regulatory preferences. Figure 2 shows the estimates of such preferences for the residential, commercial, and industrial sectors based on historical prices by North American Electric Reliability Council (NERC) region. For each region three bars are shown, one for each sector. The scale is relative. A higher value represents a greater implied regulatory preference for that sector; lower values represent less implied preference. The preference is inversely proportional to prices, with a higher preference indicating greater regulatory preference and thus lower prices. Figure 3 compares actual prices of electricity by customer class in 1997 with an estimate of economically optimal prices. That is, the implied regulatory preference is set to unity across all sectors. The results indicate that industrial and commercial prices in 1997 were largely higher and residential prices lower than the prices associated with the economically optimal Ramsey solution.
Figure 2. Implied Regulatory Preference by NERC Region [source]
Figure 3. Actual 1997 Electricity Prices by Sector and Calculated Prices with Optimal Pricing [source]
Pricing with Competitive Generation
Assume that electricity is distributed under rate-of-return regulation as before, but the generation component is now competitively priced. Assume further that the regulatory preferences remain the same in the restructured environment. There are three customer classes: residential, commercial, and industrial, each with varying characteristics. Some sectors, such as residential, have little ability to leave the distribution system and thus have low elasticities. Others, such as industrial customers, have a much higher elasticity and the ability to bypass the system. In this sense, "bypass" refers to the ability to generate electricity on site, relocate, or otherwise withdraw from the local system. We assume that the demand for bypass is positively related to the price of distribution; that is, as the price increases the usage decreases, and the desire to bypass the system is increased.
Given these circumstances, regulators must look to modify their pricing algorithms. In a restructured environment where competition for generation is introduced, the generation component of electricity price is necessarily beyond the regulator's control; only the pricing of transmission and distribution functions remains regulated. Because transmission and distribution services are only a portion of the total price of electricity, the impact of apportioning their fixed costs among customer classes on the total price will be smaller than when all fixed costs—for generation, transmission, and distribution—are included. If the price paid by a particular class is set equal to the marginal cost, that group contributes nothing to the fixed costs. As the price rises above this level, the group contributes to the fixed costs but only to the extent that it remains on the system.
The sectoral distribution price is a function of the elasticity of the customer class and the implied regulatory preference as before. However, three additional factors must be included: the price of the generation component, the fraction of load available for bypass, and the elasticity of bypass. While generation is purchased competitively, and thus its price is set by the market rather than through regulation, its level affects the overall price and the ultimate consumption by the customer class. The elasticity of bypass represents the willingness of the customer class to explore bypass opportunities as the price rises and as the fraction of load for which bypass may be available increases. Both a larger elasticity of bypass and a larger fraction of load available for bypass would tend to cut consumption from the grid as the price rises. As long as the price of distribution is above marginal cost, it is worthwhile to continue to supply service. As the share of potential bypass candidates decreases, the price charged approaches that which might be set if no bypass were available. Similarly, the price declines as the elasticity of bypass increases. This model provides a framework for analysis, given a regulated distribution system and customers with different elasticities and abilities to avoid the distribution system. The mathematical representation of such a market structure is shown in the Appendix.
Applications in the Annual Energy Outlook 1999
The Annual Energy Outlook 1999 (AEO99) incorporated the effects of restructuring in sectoral electricity price projections. Figure 4 compares the sectoral prices projected in the AEO99 full competition and reference cases for the Electricity Cooperative Agreement for Reliability (ECAR) NERC region, which comprises the States of Ohio and Indiana, the Lower Peninsula of Michigan, and parts of Pennsylvania, West Virginia, Virginia, and Kentucky. Prices in the ECAR region decline throughout the projection period due to declining coal prices, declining capital expenditures, and improved efficiencies of new plants.
Figure 4. ECAR Electricity Prices by End-Use Sector in the Reference and Full Competition Cases, 1999-2020 [source]
In the full competition case, generation is priced on a marginal cost basis. In the early years of the projection, the marginal cost of generation is below that of the generation component of price under rate-of-return regulation. Therefore, the price of electricity falls as stranded cost recovery is completed. The reduction has the greatest impact on the residential and commercial sectors, which bear the largest absolute portion of the generation price. In the last years of the projection, the marginal cost of generation increases slightly, reflecting both the increasing price of natural gas and the greater proportion of time that the relatively more expensive natural-gas-fired plants set the margin over less expensive coal-fired units. As a result, the prices for all three sectors in the competitive case are flat over the last 10 years of the projection. The reference case prices continue to decline, dropping below those in the competitive case by 2020.
It was assumed that, through the function of an Independent System Operator (ISO) or other market structure, the generation component of price at any one instant would be equal for all customers. That is, the difference between the average yearly price of the generation component of electricity for different customer classes depends only on the fraction of their annual load purchased during high-priced periods. This had the effect of causing prices to decline most rapidly in the residential and commercial sectors under competition as compared with the reference case, as generation was a larger fraction of their total price.
The near equivalence of the generation prices is illustrated by the representative price-duration curve shown in Figure 5, which depicts the ranked hourly price of electricity from the most expensive to the least expensive hour. The key feature of the graph is that the price-duration curve is relatively flat on a per-kilowatthour basis. A flat curve shows that, except for a limited number of peak hours, the price of generating electricity is relatively constant. Accordingly, in a fully competitive market, the average annual price of electricity is similar for customers with relatively flat requirements as compared to those with large peaks in their demands. The differences reflect different levels of consumption among the customer classes in the peak periods. Generation costs for different customer classes are shown in Figure 6.
Figure 5. ECAR Electricity Generation Price Duration Curve [source]
Figure 6. Generation Component od ECAR Electricity Prices by End-Use Sector, 1999-2020 [source]
The model discussed here assumes that, while both transmission and distribution will continue to be under the purview of regulators, the unbundling of generation from transmission and distribution will provide medium and large consumers with a greater ability to obtain price concessions from the operator of the distribution system. Specifically, under the new market structure some consumers may have the ability to bypass the distribution system at relatively low cost by connecting directly to the transmission system or building an on-site generator. Concessionary pricing—i.e., changes in the allocation of fixed costs among the customer classes—may be necessary to retain those customers. The equations in the Appendix reflect the effect of concessionary pricing on the sectoral distribution of prices.
The ability to bypass the distribution system, or to credibly threaten such bypass, is the most important factor in determining the level of price concessions. The level depends on a number of factors, including the projected cost of distributed generation technologies, fuel costs at the distributed generation site, and the regulatory regime under which distributed generation may occur. Bypass has been parameterized as the fraction of load with the potential to bypass the distribution system. Figure 7 shows the national level of industrial prices under competition with several values of the bypass parameter. With bypass restricted, industrial prices may rise above those in the reference case. As more of the industrial load has the potential for bypass, the industrial price declines.
Figure 7. Industrial Prices Under Four Bypass Parameter Values, 1999-2020 [source]
Figures 8, 9, and 10 again compare the sectoral prices in the reference case with those in the full competition case, including concessionary pricing of transmission and distribution. As a counterpoint, another projection is shown, based on the assumption that industrial customers would be unable to obtain any additional concessions from the operators of the transmission and distribution system (no concessionary pricing). As can be seen, if average generation prices by customer class tend to converge, it is possible that industrial end-use prices could rise significantly above the reference case price unless there was a reallocation of costs within the regulated transmission and distribution sector. Such an increase in the industrial price would cause a concomitant decrease in residential and commercial prices.
Figure 8. Residential Electricity Prices in the ECAR Region Under Three Fixed Cost Allocation Options, 1999-2020 [source]
Figure 9. Commercial Electricity Prices in the ECAR Region Under Three Fixed Cost Allocation Options, 1999-2020 [source]
Figure 10. Industrial Electricity Prices in the ECAR Region Under Three Fixed Cost Allocation Options, 1999-2020 [source]
Given that similar efficiency improvements are assumed in the reference and full competition cases, it is not surprising that the national sectoral price projections in the two cases are similar. However, this analysis assumes that, if price discrimination is not present in the competitive generation market, larger customers will maintain their ability to achieve lower rates in the remaining regulated portions of the industry.
Restructuring will undoubtedly present new challenges to the portions of the electric utility industry that remain regulated. It is likely that the pattern of prices among sectors will change in the new environment. While the regulatory goal of fair and equitable rates remains, the ability to act on that goal may be adversely affected by restructuring. The results shown here indicate that industrial and commercial prices could be largely higher, and residential prices lower, than the prices associated with an economically optimal solution. A comparison of the AEO99 reference and full competition cases for a single region shows that, without concessionary pricing, industrial prices under competition could be significantly above those under traditional rate-of-return regulation.
Suppose that a regulator sets prices for a firm so as to maximize a weighted sum of consumer surpluses over the customer classes served by the firm. Specifically, assume that there are three customer classes, and the decisionmaker maximizes the sum of qiCSi, where CSi is the consumer surplus of customer class i and qi is the weight that the decisionmaker attaches to i’s level of welfare. The case in which the decisionmaker is indifferent to the welfare of i is represented by qi = 0. The case in which the economic interests of i are twice as important as those of j can be represented by qi = 2 and qj = 1.
Our model is then:
(1)
such that:
(2)
This merely states that the regulator attempts to maximize the weighted consumer surplus subject to the constraint that the revenues for the regulated utility must equal its costs. An unconstrained maximization problem may be formed with the addition of Lagrange multipliers, to determine a solution that satisfies the first-order conditions. The solution is:
(3)
where ei is sector i’s price elasticity of demand,
l is a scalar to preserve
feasibility of the solution, and 
is the marginal cost of production.
Now assume that the electricity is distributed by a regulated natural monopolist. We have three sectors: residential, commercial, and industrial. Each sector purchases Xi from the generating industry at the price PGi per unit. Assume that the total cost of distribution is related to the amount of power distributed to each sector, i.e., the total cost of distribution is CD(X1S, X2S, X3S) where XiS is the amount of power that the distributor delivers to sector i through its system.
In AEO99, the industrial and commercial sectors were assumed to have the
threat of bypass—i.e., the ability to leave the system through self-generation
or other means. We let XiS = Xi - XiB, where XiB is the amount of power
that bypasses the distribution system. Let h represent the elasticity of
nongrid supply, i.e., h = 
XiB/Pi) (Pi /XiB), and let SiD be the share of
total generation purchased by the sector i that uses the distribution system
(i.e., SiD = XiS/Xi), where i = 2 for commercial and 3 for industrial.
In contrast, the residential sector must purchase all its power through
the local distribution system. Assume that the amount of power bypassed
is positively related to the end-use price of electricity, i.e., XiB/Pi
> 0, where Pi is the total end-use price. Finally, assume that the end-use
price is the sum of the generation price (PGi) and the distribution charge
(PDi).
Based on the above discussion, the decisionmaker is assumed to maximize the function:
(4)
For the customers in the residential sector, XjS = Xj, and thus the following result can be obtained:
(5)
In the sectors where bypass is possible, the distribution price for sector i can be shown to equal:
(6)
We can see that Equation (6) reduces to Equation (5) when SiD equals 1, that is, when no bypass is possible.
