A purposeful approach to setting energy prices.
Philip Q Hanser is a principal with The Brattle Group. He acknowledges the contributions of Brattle colleagues Ryan Hledik and Ahmad Faruqui, as well as Ken Costello of the National Regulatory Research Institute. He also acknowledges editorial assistance from Heidi Bishop and Shannon Wentworth at Brattle. The opinions expressed in this article are Hanser’s and don’t represent those of The Brattle Group or its clients.
The design part of rate design is “More honor’d in the breach than the observance.” Save for dynamic pricing, many energy companies’ rate designs are simply inherited from some rate case in the dim, distant past. And with the exception of legacy rates, which were designed to encourage consumption and fell out of fashion many moons ago, and the newer dynamic pricing rates, much of rate design goes little beyond making sure the appropriate level of revenue is collected from each rate class, subject to the constraints of the metering technology at the customer’s site. Sometimes there are arguments made about a customer charge being too high or the appropriate consumption level at which a blocked rate begins or ends, but for the most part those discussions largely lack any content on the actual impacts of such changes, and they certainly don’t relate explicitly to any goals that the proposed rate designs might have.
This situation extends beyond integrated utilities to local distribution companies, generation and transmission (G&Ts) cooperatives, energy marketers, and gas distribution companies—almost any provider of energy services. G&Ts provide energy and transmission services to their retail member cooperatives, and often those rates suffer from various cost misalignments, outdated legacy provisions, riders, and exceptions, which don’t necessarily reflect the current market realities they face. For example, G&T credits for retail cooperatives’ electric heater water controls might reflect their value to the G&T before market restructuring, when it relied solely on its own generation resources to meet its retail cooperatives’ loads. Today, however, market restructuring might have reduced the value of those controls and drastically changed when awarding those credits makes sense. Similarly, municipal utilities can find themselves with very different market economics, particularly when their long-term contracts end and they are exposed to restructured markets. Marketers tend to provide rates that are similar in structure to the utility whose customers they serve. Those legacy rate structures, however, often had hardly any notion of goals in their design.
There are exceptions, and at various times there have been quite vigorous discussions of the goals of rate design. Bonbright’s Principles of Public Utility Rates, for example, wasn’t merely an academic exercise, although its scholarship is quite evident. It was written in response to the discussions about rate designs that were active at the beginning of the second half of the last century. Similarly, the Public Utility Regulatory Policies Act of 1978 (PURPA) made explicit the need for public utility commissions to review rate designs of utilities within their states. Again, this was due to the widely held perception that there was something amiss in the price signals customers were given.
Now, it’s time to revisit the rate design process for a number of reasons. First, the electric industry has gone through an enormous period of upheaval because of the restructuring of its markets. Although there are likely to be more structural changes, the industry is showing some signs of settling down. Second, it appears that the combination of environmental regulations, changes in fuel prices, and renewable portfolio and clean energy standards will yield a sea change in the underlying generation mix. This will drastically change the industry’s underlying cost structure. The cycling capacity needed to support renewables could yield a situation where peak avoided costs are low, but the value of saving energy is relatively more important. Third, metering technology has changed drastically, with a much broader spectrum of customers having significantly enhanced measurement capability of the level and timing of their consumption. Fourth, the penetration of new customer technologies as well as utility and regulatory programs that are viewed as publicly beneficial depend crucially on customer economics, and customers’ calculations depend crucially on their rates. Finally, rates are a means by which energy companies communicate their value proposition to their customers, and not merely the process by which they collect revenues.
If you had a blank slate for designing rates, you likely would follow standard business practice and outline a process to design those rates. The process would begin by defining goals or objectives, followed by the development of metrics, preferably quantitative, to determine how well a particular design meets those goals. Then, you’d design a particular rate or set of rate designs and evaluate each against your goals by calculating a scorecard based on the metrics. Finally, you likely would tweak and re-evaluate, alternating between the design process and the evaluation process until you finally settled upon a design. If you had followed this process, then you would have accomplished rate design by objective. This is the methodology that companies arguably should adopt today. (See Figure 1).
Generally, the issue of rate design goals arises in the context of changing rates. Thus, the objectives or goals that you wish to achieve through rate design depend on what you wish to accomplish, where you want to eventually be, and when you want to get there. The rate design objectives in Bonbright’s Principles or those of PURPA are a good starting point because of their familiarity and regulatory acceptability. Bonbright’s principles are: 1) The related, practical attributes of simplicity, understandability, public acceptability, and feasibility of application; 2) Freedom from controversies as to proper interpretation; 3) Effectiveness in yielding total revenue requirements under the fair return standard; 4) Revenue stability from year to year; 5) Stability of the rates themselves, with a minimum of unexpected changes seriously adverse to existing customers; 6) Fairness of the specific rates in the apportionment of total costs of service among the different customers; 7) Avoidance of undue discrimination in rate relationships; 8) Efficiency of the rate classes and rate blocks in discouraging wasteful use of service while promoting all justified types and amounts of use—including in the control of the total amounts of service supplied by the company, and in the control of the relative uses of alternative types of service.
Usually, these objectives are summarized as revenue adequacy; fair apportionment of costs among the beneficiaries of the service; and optimal efficiency or customer rationing to discourage wasteful use, while promoting all economically justified use. Bonbright’s objectives have generally been well-received by regulators and represent a reasonable balance between the financial requirements of the utility as a firm, shareholders’ and customers’ interests, and the role that rates serve as prices in a market.
PURPA’s objectives by contrast are: 1) Conservation of energy by users; 2) Efficient use of facilities and resources by utilities; and 3) Equitable rates to consumers. These, too, have been well-received by regulators. Indeed, discussions about rate objectives cite these two sets of objectives in the same breath.
So, now, how do you choose your goals? The goals depend on your circumstances. One situation that utilities sometimes find themselves in is that their rate offerings have become so complex and have so proliferated over time that their rate listings compare favorably in size to that of a dictionary. For those utilities, simplicity becomes an overarching goal. For energy companies that face drastic changes in their underlying cost structure, such as the G&T cooperatives and municipals newly entering restructured markets, re-orienting them to the new cost structure will be a primary goal. Generally, this goal would fall under the rubric of efficiency because the new rates will hew more closely to the company’s costs and that will induce efficient choices by its customers. The secondary goals, however, also are important. For example, a simplified rate structure that doesn’t also have the capability to generate a sufficient and stable revenue requirement isn’t likely to be acceptable. Alternatively, a rate focused primarily on an efficiency goal, but which is understandable only to economists, isn’t an acceptable situation either. In addition, rather than making a massive rate change in one fell swoop, the energy company might wish to change its rate offerings in a series of steps over time to reduce the sense of unease that customers might experience from a single massive rate shift. That goal would come under the heading of rate stability.
The energy company might wish to have an ensemble of rate offerings that relate to a set of varying objectives. For example, it might want to have rates that differ in their profiles in regards to price risk in order to provide a spectrum of rates differentiated along that dimension. Thus, the company might offer the following rates: 1) a simple monthly charge that doesn’t vary even as the customer’s monthly usage varies; 2) a flat rate; 3) a time-of-use rate; and 4) a time-of-use rate with a critical peak pricing (CPP) component.
Finally, the objectives could be defined with regard to customer segments within a rate class. For example, for residential customers, a real-time pricing rate with an in-home display networked to the customer’s appliances could be a design for the segment of early technology adopters and could be used to meet a company goal of becoming more involved with customer technology choices. For the commercial sector, a back-up generator rate, which provided incentives to oversize the back-up generator technology and permitted the company’s use for local generation reliability, could meet the goals of enhancing commercial customer reliability for the segment desiring high reliability and the company’s goal of improving local distribution system reliability.
Metrics and Scorecards
Goals and objectives are all well and good, but unless there is some way to translate them into something measurable, even if it’s only a ranking, they end up being largely symbolic. It’s unlikely that you can quantify exactly how each of the rate goals you set should be measured. Indeed, you expect some goals to have multiple ways of being assessed. The point is that if you let exactness of goal measurement be a hurdle that prevents any quantification at all, then you will have no metric at all in the end. The statistician John W. Tukey has probably said it best, “Far better an approximate answer to the right question, which is often vague, than an exact answer to the wrong question, which can always be made precise.”
In the context of rate objectives, this means that sometimes the energy company might need to accept approximate metrics for measuring its goals, but that these are better than having no measurements at all. Some metrics are relatively easy to conceptualize. For example, if a goal is to provide a price signal that induces efficient consumption, then how closely the rate hews to the energy company’s marginal costs is a potential metric. Alternatively, if the energy company is concerned about inducing energy efficiency investments, then one possible metric could calculate how the rate affects cost-effectiveness tests. If there’s a question of bill continuity when changing a rate, then the total bill to the customer could be calculated at various usage levels. Also, a finer metric can be made of this by calculating the components of the bill also at various usage levels. Customer satisfaction with a rate might be approximated by reviewing the customer service database for rate-related complaints for the rate. An even more approximate answer is simply a count of issues raised with a rate at the most recent rate case.
Hypothetical Rate Design
To demonstrate how rate design by objective works in practice, take the hypothetical example of Hoople Gas and Electric (HG&E). HG&E is looking to abandon its flat residential rate as a result of suggestions from commission staff that its rates could use a re-examination. HG&E finds that inclining block rates look interesting because providing better efficiency incentives is its primary goal and, although it hasn’t done a cost-benefit analysis of their implementation, it doesn’t want to make the capital investment in metering and information technology that dynamic pricing rates would entail. (See Figure 2).
HG&E’s first task is to design the pair of inclining block rates. In designing an inclining block rate, three decisions must be made. First, how many tiers? Second, where should the tier prices be set? Third, what should be the tier cut-off—that is, the highest consumption before the customer is bumped up to the next pricing tier? HG&E decides to test an alternative that’s as simple as possible and another with slightly greater complexity, so it decides that one alternative will have two tiers and the other will have three. Other utilities have had as many as five tiers.
Having decided on the number of tiers, HG&E must next decide on the cut-off point for the tiers. There are many possible choices for tiers. For example, the first tier could align with the average monthly usage level or it could approximate the usage level that can be met from baseload generation. HG&E could set the cut-off at some efficiency target—say, some percentage of a prior year’s consumption. HG&E decides to set the first tier at the average level of usage and will do so for both of the inclining block rates. For the third tier of the three-tier alternative it sets the cutoff at approximately 80 percent of average usage as an efficiency incentive.
HG&E’s next decision is the price level for each tier. Again, there are many choices. If it were to base the prices on its resource costs, the initial tier could be its off-peak marginal costs, and the final tier could be HG&E’s peak marginal costs. If it had a more policy orientation to the rate’s design, it could include any rate increase in only the upper tiers. HG&E opts for a relatively conservative choice for its two-tiered option. For the first alternative, it maintains the same customer charge, sets the rate of the second tier at 80 percent of marginal costs, but then reduces the first tier rate below that of the current flat rate so as to maintain the same level of revenues for the rate. For the second alternative, HG&E wishes to account for the policy goal of assisting customers with fixed incomes, so it reduces the customer charge by $2.00 per month. HG&E then sets the final block price at full peak marginal costs and sets the prices for the other tiers as a decreasing percentage of marginal costs.
Since it’s in a regional transmission organization (RTO), these calculations are pretty straightforward for HG&E because of locational marginal pricing. If it hadn’t been in an RTO, it would have needed to estimate its marginal costs through a study. Also, since HG&E wants the rate to remain in place for at least a couple of years, it does its price calculation based not only on historic prices, but also uses futures prices to project forward marginal costs, creating an average marginal cost for the next two years. The original flat rate and the two alternatives are illustrated in Figures 3 and 4, as is the associated customer data.
HG&E decides that its rate objectives, besides its overall goal of improving efficiency, are: 1) revenue sufficiency; 2) rate continuity; and 3) equity. Rate simplicity also could have been a goal and it will certainly come into play in any final decision of HG&E. While none of these rates are dynamic, it’s possible such rates could be developed in the future once the customers have the right metering technology. Until then, HG&E must determine how to measure its objectives.
For the first objective, providing better incentives for efficient use, viewing the rate as a portion its marginal energy costs serves as one metric. Alternatively, HG&E also could have estimated how much peak reduction the rate produces and assessed the related capacity savings from the rate. That calculation would require load shape data—which HG&E has as a result of the metering it does for rate class cost allocation, but HG&E opts not to calculate that measure. HG&E, does, however, calculate the impact the change in rate has on the paypack period for some selected energy efficiency measures, and does so for customers whose monthly consumption is below 800 kWh, 800 to 1,600 kWh, and over 1,600 kWh.
For the second objective of revenue sufficiency, HG&E calculates the class billing determinants for each of the rates and then determines the percentage of its revenues in the highest—or tail—block. HG&E has designed the three-tier rate so that ex ante it collects slightly more than the revenue requirement for the flat rate. It expects that customers will respond to the higher price and reduce their consumption in the final tier ex post, that is, after the rate is implemented. In this calculation, it could have applied a short-run price elasticity to the change in price in the tail block, but HG&E uses a considerably less formal approach and guesses what the response will be. Because a substantial portion of its revenues are in the tail block, HG&E was worried about the stability of its revenues and doing the price elasticity calculation would have reduced some of the uncertainty. Thus, if the chill of winter or heat of summer drives consumption into the tail block because a utility has a high penetration of electric heat or air conditioning, customers will face a significantly higher price in that block, and, taking into account their price elasticity will be important.
For the third objective, rate continuity, HG&E calculates how much the bills will change for customers within various levels of average monthly consumption. There’s only a single calculation here, but it would be useful to do some variations. For example, a summer-peaking utility might do this calculation for the distribution of consumption in its hottest months. Finally, for equity, HG&E calculates the average change in monthly bills for customers that are receiving bill assistance from the utility. Another measure that HG&E could have used is the average change for customers that participate in the state’s low income home energy assistance program (LIHEAP). Figure 5 shows the scorecard for the two rate alternatives.
By design, the inclining block rates are an improvement in efficiency as measured by their price as a percentage of marginal costs. Note that for customers in the lower tiers, the reduction in rates reduces their incentive to invest in energy efficiency devices, but the reverse is true as consumption rises into the upper tier(s). The increase in revenue recovered through the tail block for the three-tier inclining block rate of 42 percent compared to the current flat rate isn’t unexpected, although it causes concerns for HG&E. The rate continuity measure suggests that customers at higher-than-average levels of consumption might find the three-tier inclining block rate troublesome—but in a way, that’s the goal of that rate. Note also that although HG&E reduced the customer charge for the three-tier rate, without additional help from HG&E, customers on bill assistance still would experience an increase in their bills for both of these rates, and substantially so for the three-tier rate.
In the real-life example from which this caricature was drawn, the energy company decided to institute the single-block inclining block rate to permit customers to become accustomed to the change in rates. It surveyed its customers’ appliance holdings before the rate was put into place, and plans to perform a follow-up survey after the rate has been in place for two years to assess initial changes before and after the rate. In addition, since the utility observes strong peaks during the summer, it wanted to measure the responsiveness of the customers to the new price signal to assess its revenue stability aspects. Going forward the utility is considering implementing either the three-tier inclining block rate or acquiring more advanced meters for their potential use in developing more dynamic rates.
This example is meant to illustrate the wide range of possibilities that rate designs offer. HG&E could have considered many other design aspects. For example, the rate could be split between seasons, with the third tier only in effect during the peak season. In the example, HG&E ruled out in advance of a cost-benefit analysis the potential for more dynamic prices, but it could have analyzed what would happen if it moved from a flat rate to time-of-use (TOU) or a dynamic rate for residential customers. It also could have considered the impact of changing a general service rate with a ratcheted peak demand charge to TOU or interruptible rate, or redesigning a G&T’s rate to its retail co-ops, etc.
Charting a Path
The restructuring of electric markets and the changes in the mix of the market’s resources has left many rates out of sync with the new realities. For some utilities, this can lead to loss of customers who switch to competitors with rates more in tune with the new market realities. For other energy companies, including G&T cooperatives and municipal utilities, poorly designed rates can increase how much some customers subsidize other customers’ energy consumption, an inequitable outcome.
Rates are a key element of energy companies’ interactions with their customers. They provide customers information and incentives on how to use energy. Thus, the choices that customers make as to when they will use energy and how much energy they will use—including investments in more efficient equipment and appliances—depend on those rates. It’s a very large lost opportunity for the energy company when its rates don’t align with the company’s goals for helping customers make those choices. Finally, newly available metering technology won’t yield its full benefits unless the utility carefully reflects on what it would like to achieve with the technology.
Rate design by objective isn’t only a means of assuring that the energy company’s rates are properly calibrated to its goals, but also a means of communicating. Rates communicate the company’s intentions to its customers—preferably improving its relations with them; to its regulators, with the goal of reducing the process costs of interacting with them; and even to the company itself, helping its employees better understand what the company views as its current and future direction. It’s much easier to arrive at your destination if you have a roadmap helping guide you. Rate design by objective can help chart the way to success.
1. The utility industry calls it “rate design,” but the rest of the commercial world would call it “pricing.” Also, although this article mostly has an electric market orientation, the same principles apply to non-electric utilities and energy companies.
2. Shakespeare, William, Hamlet, Act 1, Scene 4.
3. Arguably, the responsibility for this doesn’t rest solely with the utilities, but also falls on the regulators for failing to articulate how they trade-off one or more goals against others. See Ken Costello’s “Decision-Making Strategies for Assessing Ratemaking Methods: The Case of Natural Gas,” National Regulatory Research Institute Report 07-10, September 2007. See also Ahmad Faruqui and Robert Earle, “Rate Case Mania,” Public Utilities Fortnightly, February 2006.
4. Bonbright, James, Principles of Public Utility Rates, Columbia University Press, New York, 1960. There have been excellent subsequent revisions to this classic, but it’s worth a bit of extra effort to find a copy of the original to be exposed to Bonbright’s prose. It’s also worth reviewing the Electric Utility Rate Design Study (EURDS), which initially attempted to codify the rate design by objective process. See EURDS 85 Costing for Ratemaking by J. Robert Malko, Darrell Smith, and Robert G. Uhler, Electric Power Research Institute, Palo Alto, Calif., 1981, particularly sections III and IV.
5. The author’s impression is that the states accepted the PURPA money to do the rate reviews, but few changed their rates as a result.
6. Douglas Hubbard cogently argues that for many items that some businesses find immeasurable—i.e., customer satisfaction—at least approximate metrics can be developed. See Hubbard, Douglas W., How to Measure Anything: Finding the Values of Intangibles in Business, John Wiley, New York, 2010.
7. Tukey, John W., “The Future of Data Analysis,” Annals of Mathematical Statistics 33 (1), 1962, p. 13. Ken Costello has suggested that the question for rates might be “Is it better to use marginal cost in ratemaking, which is theoretically the right concept in determining price but has imprecision in measurement, or embedded cost, which is the wrong concept, but can be measured with greater precision?”
9. The example has been taken from an actual case the author was involved with, although the data and the complexity of the problem have been greatly simplified. Also, this is meant only to be viewed as an example and not a policy prescription. See also Ahmad Faruqui, “Inclining Toward Efficiency,” Public Utilities Fortnightly, August 2008.
10. In the words of Monty Python, “Hint, hint, wink, wink, nudge, nudge.”