The "duty to connect" demands definition - such as the optimal investment in local wires, and who should pay for it.

As the electric utility industry continues its slow but inexorable transformation into a more "competitive" industry, there has been a notable absence of discussion concerning continued regulation of local distribution utilities, or discos.

This glaring omission is problematic, as the overall success of restructuring efforts will depend to a critical degree on maintaining a safe and reliable "obligation to connect." If that obligation is not met, or if the costs of meeting it increase, the benefits of competition will not be realized.

And the matter is even more complicated. A fundamental regulatory conflict may exist between regulated discos and unregulated retail energy suppliers. This conflict may well require action from utility regulators.

The Conflict That Arises

Suppose a large, electricity-intensive industrial customer wishes to expand its manufacturing operations in a particular disco's service territory.fn1 If the expansion takes place, it will mean additional secondary growth in electricity demand as new suppliers, businesses and households locate in the area. If not, local area demand will increase very slowly.

The manufacturer wants the best possible deal for energy and will examine competing offers of suppliers. Those suppliers will have to meet all applicable regional transmission system requirements (e.g., installed capacity, spinning reserve, etc.) to ensure that the energy can be reliably delivered to the manufacturer's local disco without adversely affecting the regional transmission system. The disco will remain responsible for delivering the electricity to the customer, ensuring sufficient distribution capacity to meet the increased demand from secondary sources.

But what happens if there isn't enough distribution capacity at the local level to deliver the power to the manufacturer? The manufacturer won't be able to construct its own set of "wires and poles" in order to increase available distribution capacity to its facilities, or otherwise bypass the disco, because the disco will have an exclusive distribution franchise. This problem illustrates a part of the fundamental tradeoff: A disco must satisfy an "obligation to connect" in exchange for gaining an exclusive local distribution monopoly.

Utility regulators will want the disco to meet this new duty to connect in an efficient and cost-effective manner. Regulators may impose incentive programs that reward the disco for greater efficiency, or require specific planning methodologies be used to evaluate distributed resource options that can meet the increased demand for distribution capacity, whether through new investments in "wires and poles," distributed generation or demand-side management (DSM).

The potential for conflict among regulators (utility and environmental), the disco, the customer and the retail energy provider is very real. The retail provider wants to maximize the profits of its electric sales to the manufacturer. The manufacturer wants to pay as little as possible to the disco for delivering the electricity it buys, and not absorb the entire cost of new distribution system upgrades that may provide benefits to other customers in the area. The disco cannot refuse to provide service to the manufacturer, even if it means making new capacity investments. Utility regulators will not want to burden other ratepayers with the costs of expanding distribution capacity to benefit the manufacturer, but will have to work within the parameters of existing cost-allocation and rate design rules. Lastly, environmental regulators may not like some capacity-increasing alternatives such as distributed generation because of the local environmental impacts, even though utility regulators may wish to promote those same alternatives in order to defer the need for new poles and wires. The result is four interest groups with four potentially conflicting goals.

For the disco to meet its obligation to connect efficiently and equitably, two difficult questions must be answered:

1. What is the optimal amount of local distribution capacity or its equivalent for the disco to satisfy its obligation to connect?

2. How should the resulting stream of current and future costs be allocated among different customers, including (possibly) retail electric suppliers?

The first question requires the disco to meet uncertain future local area capacity demand. This will be all the more difficult because, unlike the existing system where the disco and the electric supplier are one and the same, the disco will probably not know the specific terms of the sales transaction between the retail supplier and the manufacturer. Although it may be possible to offer interruptible power contracts to the manufacturer to reduce peak capacity and delay or obviate the need for new distribution system investments, such arrangements will also have a cost, and may not coincide with what the retail energy supplier wishes to offer. The disco will need a method for determining which investments in area capacity it should make, if any, and regulators will need to approve of such methods.

The second question, while less technical, will be more contentious. In the past, distribution expansion costs have rarely been allocated to specific customers because of the public good nature of the distribution system. Instead, those costs have traditionally been allocated among all customer groups, based on standard cost-allocation methodologies for integrated utilities. That probably will have to change.

Doing so will raise several issues. How will customer groups be defined, by local area or across different areas? Will individual customers be required to pay the full costs of major distribution system upgrades, even if the benefits accrue more widely? What will happen if forecast demand for that capacity does not materialize? How will regulators ultimately define "prudent" disco investments in area capacity when future demand is uncertain?

What is "Least-Cost" Distribution?

When a disco invests in new local area capacity, its regulators and customers will want those investments to be efficient to keep distribution costs low. Overbuilding distribution circuits or investing in overly expensive distributed generation and DSM in anticipation of local demand that never materializes will result in other customers paying more than they might otherwise. Conversely, under-building may require expensive "emergency" investments or, in some cases, an inability for customers to obtain the electricity they want. Thus, knowing when to invest and in what to invest will be an important component of disco life in a restructured environment.

A few regulators already have recommended approaches, usually by extending the concept of avoided costs to determine cost-effective distribution investments.fn2 Applied to distribution investment, the avoided cost approach begins with a "base case" distribution investment plan and then determines whether specific investments, such as a new substation or distribution circuit, can be deferred by distributed resources like DSM and local generation. The deferral benefit is simply the present value savings associated with deferring an investment for a given amount of time.

For example, suppose that constructing a new substation will cost $11 million. If the utility's weighted cost of capital is 10 percent, then the deferral benefit from delaying construction of the substation exactly one year will equal $1 million, based on a simple present-value analysis: $11 million - ($11 million/1.10) = $1 million. Now suppose further that peak loads will increase exactly 1 megawatt during the year. Then the "avoided distribution cost" will be just $1,000 per kilowatt (or $1 million/MW). Using avoided costs, distributed generation and DSM costing less than $1,000/kW would be a cost-effective alternative. This reasoning is straightforward, compelling and, unfortunately, completely wrong.

If this method is used, alternative investments will be chosen to defer traditional transmission and distribution investments for as long as possible without adding additional costs. That is the fundamental flaw in the avoided-cost approach: What is cost-effective is not necessarily least-cost, unless the objective is to maximize the amount of deferral provided by the alternative investments, or what is the same thing, to maximize the penetration of alternative distribution system investments (DSM, distributed generation, etc.) beyond the point at which they are part of the optimal solution.

The avoided-cost approach makes the wrong marginal comparisons for distribution investment; it may defer traditional distribution capacity investments for too long or too little, but never the correct amount. Most importantly, the method does not incorporate future uncertainties, particularly uncertainty about future local distribution capacity demand. Thus, as a method for determining appropriate investments in new local area capacity, the avoided cost approach will lead to higher than necessary costs for consumers. This mistake reduces the overall economic benefits of restructuring and may lead to additional conflicts between retailers, customers and regulators.

So how should local distribution capacity planning be performed? First, utilities and regulators must recognize that uncertainties at the local level are likely to be magnified relative to overall system load uncertainties. The effects of one industrial customer's expansion plans are likely to be far greater in a specific local area than over a disco's entire service territory. Ignoring uncertainty, which most avoided cost methods do, will result in higher than necessary costs.

But if utilities and regulators address uncertainty, they need to do so correctly. The local area distribution capacity investment problem is strategic, in that the investments are long-lived, and dynamic, in that decisions and outcomes are interrelated. Addressing the problem with an avoided cost approach, however, treats the problem as a tactical, short-term exercise with deterministic solutions.

The economic benefits of new technologies such as microturbines and fuel cells, as well as targeted local DSM efforts, must be evaluated fairly if those resources are to be part of "least-cost" distribution systems and compete on a "level-playing field." That requires the use of newer analytical techniques that explicitly account for the strategic and dynamic characteristics of distribution system investment planning. Fortunately, these methods exist, and have been successfully tested at several utilities, including Green Mountain Power, Ontario Hydro and Wisconsin Electric Power.fn3 These same methods can be applied to other distribution investment decisions, such as whether to repair or replace aging infrastructure.

Local Capacity Investments: Who Pays?

Determining appropriate local capacity investments is a technical issue that can be solved with enough computing horsepower. Determining who should pay for those investments is tougher. One reason is that distribution systems have qualities of what economists call "public goods" and "club goods." Public goods are those where consumption by one individual does not affect consumption by another, while club goods are those that are uniformly provided by a "club" that an individual must belong to. A disco is a local club that buyers (local electricity consumers) must join (pay an access fee) to receive club benefits (electricity bought on the market from retail suppliers.)

Because individual disco customers will have different demands for the specific public goods supplied, including peak capacity, voltage quality, reliability, etc., "free-rider" issues will need to be addressed. For example, slight voltage spikes are probably of little importance to residential customers, but can severely damage some industrial production facilities, imposing significant costs. If voltage spikes are reduced by new system investments at the behest of the industrial customer, residential customers will benefit as well, even though they have not paid for the benefits. The existence of free-rider effects is one reason that distribution system improvement costs often are not paid for by individual customers.

The same arguments follow for increases in local area capacity. Simply forcing individual area customers to pay for incremental increases in local area capacity may not meet existing regulatory definitions of equity. Nor may it be efficient, because individual demand for new capacity may be less than the cost of supplying large fixed increments of capacity. Observed increases in local area peak demand may be the result of random effects, such as weather, rather than actual shifts in customer demand because of new electricity-consuming equipment. Changes in observed peak demand may also be the result of changing rate structures.

Distribution system costs historically have been allocated among all customers of a utility by customer class, rather than on an individual customer or individual area basis, in part to address these complexities. But as retail competition takes hold, this allocation system is likely to be challenged.

Cost-Allocation Schemes

There appear to be three alternative cost-allocation options for new distribution system capacity investments:

1. Preserve the existing system by which capacity investments are allocated among all distribution customers using standard cost-allocation principles;

2. Identify customers responsible (individual users) for increased local area peak demands and require them to pay all incremental expansion costs;

3. Require some customers to pay for new local area capacity, but allocate costs across all distribution system customers for others. (The actual payment mechanism, such as alternative distribution tariffs for individual customers or local area-specific tariffs, is another issue, which I do not consider here.)

1. Preserve the Existing System. Under the current system of allocating costs by customer class, investments in new capacity are rolled into the overall class distribution rates for the disco's customers. Disco customers outside the specific area of growth are affected adversely. However, in the new restructured world, the disco will be responsible for meeting customer distribution capacity needs but will have no influence over the terms and conditions for energy sold to customers. That will complicate attempts to apply the existing cost-allocation system. Furthermore, the existing allocation methods are based on traditional definitions of customer classes, but those customer class distinctions may no longer persist in a restructured world. Thus, it isn't clear how efficient and equitable the treatment will be if regulators maintain the status quo in allocating costs for distribution capacity investment.

2. Allocate to Specific Users. An alternative is to allocate distribution capacity investment costs to the individual customers who impose the need for such investments.

In some circumstances, it may be possible to identify specific distribution investments that provide benefits to only a subset of clearly identified customers, usually large industrial or commercial ones. As restructuring takes place and separate discos are formed, there is likely to be greater regulatory pressure for those discos to identify specific customers who are driving the need for new local area capacity investments. Regulators may face political pressure not to impose new distribution costs on customers in general, so as to deliver on promises of lower cost electricity service to all classes of customers. Unfortunately, in many cases, investments will provide joint benefits to a variety of customers in an area.

That pressure may be exacerbated by the structure of distribution rates developed by the regulators. To the extent that fixed distribution costs are allocated on a volumetric (per kilowatt-hour) basis, consumption decisions will be affected more than if these costs are allocated on a fixed, ready-to-serve charge basis. Because many of the distribution costs are fixed and the marginal cost of providing off-peak local area capacity is, for all intents, zero, recovering distribution capacity costs on a volumetric basis will be inefficient. Such inefficiency will exacerbate inequities and complicate the cost-allocation question.

3. A Mixed Method. A third option is for regulators to assign distribution system capacity investment costs directly to certain individual customers, but allocate other costs to groups of customers. This option may be thought of as a subset of existing allocation systems, except replacing groups of multiple customers with groups of individual customers.

This "mixed" option represents a theoretical ideal in terms of equity: Allocate non-joint costs to the customers driving the need for new investments but allocate joint costs more globally. Thus, if there are no other customers in the same area as the industrial facility, then asking that customer to pay the expansion costs is reasonable and appropriate. This solution, however, assumes away the very difficulties likely to be encountered: multiple beneficiaries and joint costs. Lastly, any mixed cost-allocation scheme will be scrutinized by those identified as specific beneficiaries, much as there often are disputes about cost allocations among different customer classes.

A Workable Solution

In the context of what is equitable, unique cost-allocation solutions simply do not exist. What is fair for one local area customer may not be perceived as fair by another, especially if those customers consume fundamentally different disco services.

Historically, regulators used traditional fully allocated cost (FAC) rules to share investment costs among customers. Regulators further addressed equity concerns by adjusting prices; commonly, "ready-to-serve" charges were set much lower than their "true" cost, because minimum bills several times larger than current levels have not been politically viable.

Yet, cost-allocation decisions can and should be separate from rate structure decisions. Thus the first step in any cost-allocation scheme is to adjust prices correctly: Ready-to-serve charges should be raised to their real levels, and regulated volumetric charges, if any, should reflect only those disco services that truly vary with consumption. This will allow consumers to make efficient choices about their electric consumption and allow discos to plan more effectively for future capacity needs.

If electric restructuring is to be successful, any chosen cost-allocation scheme should be judged by its affect on overall economic efficiency. The best allocation scheme will meet specific and well-defined equity goals with the least adverse impact on economic efficiency. Although the existing FAC system has a proven track record, it was designed for a fully regulated and integrated system. As that system is replaced with a mixed competitive-regulated framework, it will be faced with new types of costs, customers and suppliers, none of which may fit within the traditional framework.

Lastly, an allocation mechanism also should account for transactions costs. Cost allocations should be as straightforward as possible, so as to minimize the regulatory and legal costs necessary to determine them.

Allocating distribution costs in the newly restructured world will require a careful balancing act that recognizes the conflicting incentives of retail energy suppliers, customers and discos. Not to give additional attention to these issues will be a disservice to customers and retail suppliers, and likely will reduce the potential benefits of competitive generating markets.

Jonathan Lesser, Ph.D., is a senior project manager at REED Consulting Group. Previously, Lesser was manager of economic analysis at Green Mountain Power Corp. He helped develop an innovative new approach to distribution system investment planning with staff from the Electric Power Research Institute, and directed all of GMP's "distributed utility" planning efforts. Lesser has authored and co-authored numerous publications, which have been published in The Energy Journal, The Electricity Journal, Public Utilities Fortnightly and the Journal of Regulatory Economics, among others. He also is the author of a textbook, Environmental Economics and Policy, published in 1997 by Addison Wesley Longman.

1 This example is adapted from a case of a major expansion at a ski area service territory, which the author addressed while employed at Green Mountain Power Corp., as part of GMP's distributed utility planning efforts.

2 See, for example, Vermont Department of Public Service, Statewide Energy Efficiency Plan, Appendix 5, May 1997.

3 The following articles present aspects of these new methods, all of which rely on decision analysis and dynamic programming techniques: 1) Feinstein, C. and Lesser, J., "Defining Distributed Resource Planning," The Energy Journal, Special Issue. Distributed Resources: Toward a New Paradigm of the Electric Business, at 41, 1998; 2) Feinstein, C., Morris, P. and Chapel, S., "Capacity Planning Under Uncertainty: Developing Local Area Strategies for Integrating Distributed Resources," The Energy Journal, Special Issue. Distributed Resources: Toward a New Paradigm of the Electric Business, at 85, 1998; and 3) Lesser, J. and Feinstein, C., "Electric Utility Restructuring, Regulation of Distribution Utilities, and the Fallacy of 'Avoided Cost' Rules," Journal of Regulatory Economics 15(1) at 93, 1999.

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.