Demand a Better Utility Charge During Era of Renewables:

Getting Renewable Incentives Correct With Residential

Demand Charges

This paper is based upon a presentation by its author at a meeting of the New York Association for Energy Economics (USAEE) in New York City, Oct. 9, 2014.

Mark B. Lively
Consulting Engineer
Utility Economic Engineers
(Gaithersburg, MD)


A growing interest in rooftop solar and other forms of distributed generation has created some worries in the electric utility industry about disappearing throughput.  Under cost of service regulation, reduced throughput normally means higher prices for the remaining throughput.  The higher prices for the remaining throughput encourages customers to add more distributed generation, which reduces throughput, raises prices, etc., etc., etc.  Twenty years ago this death spiral concern was sparked by some exceedingly expensive power plant investments.  Now the spark seems to be a reduction in the effective cost of rooftop solar, as well as the removal of various institutional barriers.

The death spiral can be arrested, or at least slowed, by utilities changing their price structure for residential customers, putting into place a three part rate structure.  A two part rate structure includes a fixed customer charge and an energy charge.  A three part rate structure differs in that it includes a demand charge, where demand should be read in the power context, that is, as the speed with which electric energy is taken by the customer

The concept of a demand charge in the electric power industry can be illustrated by reference to the telecommunications industry.  Cell phones have progressed through 2G, 3G, and now 4G.  The cable television industry has progressed through coaxial cable, DSL, and fiber optics.  These various technologies differ in how fast content can be downloaded.  The faster the download speed, generally the higher the price for the service.

The language of the telecommunication industry is instructive for understanding the death spiral issues of the electric power industry.  Not only is service speed an important concept, but there is also the concept of separating the content provider from the service provider.  Generally we can use a cell phone or our internet connection to download content from an entity other than the connectivity provider.  In much the same way, we often get electric energy from an entity other than the wires company.

The electric utility industry long had the content provider being the same entity as the service provider.  The electric utility industry called the concept vertical integration.  Some parts of the U.S. still have vertical integration of the electric utilities.  However, most electricity service providers do not have an economic interest in the electricity content providers, the power plants that produce the electricity we want for our homes.

The disconnect between electricity service providers and electricity content providers has almost always existed to some extent but the pervasiveness of this disconnect is a recent phenomenon.  Municipal and cooperative utilities were often too small to take economic advantage of the economies of scale associated with owning larger power plants.  And when these small utilities outgrew their initial power plants, buying electricity from the local investor owned utility made economic sense.  Some municipal and cooperative utilities formed their own production companies, such as the municipal joint action agencies or generation and transmission cooperatives, creating an indirect interest in the content providers.

More recently, various states have required their utilities to be restructured with the generation owned by a separate party.  This separate ownership has been facilitated by the creation of Independent System Operators which run a competitive market for such generation and by the creation of Regional Transmission Operators which perform the same function in regard to transmission lines.  So, many of the utilities that provide service through distribution lines no longer have an economic interest in generation, the content providers for the electric power industry, content providers that compete with rooftop solar.

The electric power industry has long had a pricing concept called a demand rate, which charges customers a fee based on the speed with which energy is used.  The demand charge is assessed against the peak demand, which is the fastest rate at which a customer takes energy.  For instance, some utilities have installed Advanced Metering Infrastructure (AMR) which can record and report the energy through the metering during each 15 minute interval during the month.  The utility typically looks at all of the intervals during the month and uses the maximum to set the demand fee for the month.  Some utilities use 30 minute intervals, other use 1 hour intervals, and at least one utility has used a full month to determine the demand charge for the following year, eliminating the need for interval-meters in regard to demand billing.

The concept of customer demand can again be illustrated by comparison to the telecommunications industry.  Telecommunications customers are interested in the speed with which they can download content, the number of digital packets that can be downloaded per second.  For power customers, the concept is how many KWH can be downloaded in a given time.  The speed with which electricity can be downloaded is capacity.  The amount of electricity actually downloaded in a given time is called demand.

Historically, the cost of an interval-meter was extremely expensive, especially when compared to the cost of an energy meter.  Another expensive option was the thermal demand meter.  To cut down on the cost of serving residential customers, most electric utilities only installed energy meters on residential customers, placing interval-meters only on industrial and large commercial customers.

Just as competition and changes in the electronics industry have brought down the cost of rooftop solar, so has the cost of interval-meters dropped.  Indeed, many, if not most, electric utility residential customers in the U.S. have their monthly consumption determined using AMI.  So, interval data are now available on which an electric power utility can assess a demand charge, based on the maximum speed with which a residential customer uses electricity.

But just because an electric power utility can bill using a demand charge, should it?  Customers desire access to the wires of the electric power utility in order to download content from the grid, and sometimes to upload content to the grid when the customer has surplus generation to send to the grid.  The collective amount of connectivity determines the size of wires that the electric service provider needs to own and operate.  The amount of connectivity relates to either how fast the customer uses the system or wants to use the system.

Customers "download" less electricity from the grid as they install rooftop solar or other forms of distributed generation.  But they still want connectivity to the grid for reliability and their remaining electrical needs.  The decrease in the amount of energy downloaded is relatively large compared to the decrease in the demand that the customers place on the utility, as is demonstrated in Figure 1, a histogram of customer load factors for two groups of residential customers.

Electric utilities use the term load factor to refer to the ratio between the amount of energy that a customer used during a time period, like a month or a year, versus the maximum amount of energy that the customer could have used during that time period if the customer had used its peak demand throughout that time period.  High load factors refer to high amounts of energy relative to the peak demand.  Low load factors refer to low amounts of energy relative to the peak demand.

Figure 1 shows that customers with rooftop solar typically have lower load factors than do other residential customers, at least for the customers in the two groups included in Figure 1.  The first group includes all rooftop solar residential customers for a specific utility.  The second group is a randomly selected sample of all other residential customers of that utility.  The skew between the load factor histograms of the two groups of customers suggests that as the rooftop solar customers have reduced the amount of energy downloaded from the grid that they have not proportionately reduced their demand on the utility. The electric grid, like the telecommunication network, is designed and built to meet peak loads on the system. Utility costs to maintain the grid will therefore not be reduced in proportion to any energy reduction by solar customers, and could potentially increase if high levels of solar penetration cause operational issues.

Utilities set their prices using the concept of a revenue requirement, the amount of revenue that the prices are meant to provide.  The revenue requirement is developed from the cost of service, the costs that the utility is incurring to provide service to consumers. The utility then allocates the cost of service to rate classes, such as residential, developing the revenue requirement for that class or group of customers.  In their simplest form, rates are then determined by dividing the class revenue requirement (as appropriately classified between customer, demand, and energy) by the billing determinants of the class (such as the number of customers in the class, the total class billing demand, and the total class energy).  The process is illustrated in Table 1.

The analysis associated with determining electric utility rates includes the classification of costs between energy, demand/capacity, and customer, as is shown in Table 1.  For illustrative purposes, the residential class is divided into two identical subclasses, one with rooftop solar and the other subclass being all other residential customers.  It should be noted that Table 1 shows the rooftop solar class before the introduction of rooftop solar, incurring the same costs as do all other residential customers.  Notice that the rates resulting from the assumptions in the table are identical across the columns for the different subclasses, $10/customer and $100/MWH.  Both subclasses have the same average realization of $117.50/MWH.

It is important to note that the reported average realizations are for the subclasses as a whole.  An individual customer within the subclass will have a much different average cost for electricity.  For instance, a customer using 10 KWH would have a bill or $11.00.   This reflects an energy charge of $1.00 and a customer charge of $10.00.  Such a customer would have an average cost of electricity of $1.10/KWH or $1,100/MWH.  The same high realization often occurs for cell phones.  How many times have you paid $50 a month for cell phone service and only used the phone once, for one minute?  That is a cost of $50/call or $50/minute.  And what is the average cost per call when you make no calls during a month?

After customers have installed rooftop solar, the amount of energy used by the rooftop solar customer drops, 90% in this series of examples, as shown in Table 2, from 20,000 MWH in Table 1 to 2,000 MWH in Table 2.  The rooftop customers still impose the same customer and demand costs on the utility; but, the division to produce rates shows recovering energy and demand costs through the energy price produces different prices for the two subclasses of residential customers in Table 2.

If separate prices were charged to the two subclasses, then the rooftop solar subclass would have an energy charge of $550/MWH versus a charge of $100/MWH for all other residential customers.  However, the standard approach of charging all residential customers the same, a requirement in some states, would increase the energy charge from $100/MWH in the left column of Table 1 (as well as in the right column of Table 2) to $140.91/MWH in the left column of Table 2.  In Table 2, the realizations vary greatly, $117.50/MWH for all other residential customers, $725/MWH for rooftop solar customers, and $172.73/MWH for the class as a whole.  As stated previously, such differentiation between the residential subclasses is not allowed in some states.

An alternative rate form is the demand charge, based on the customer peak usage.  Tables 3 and 4 show the result of implementing a three part rate form that includes a demand charge.  Tables 3 and 4 parallel the conditions of Tables 1 and 2.  Tables 1 and 3 are before the installation of rooftop solar.  Tables 2 and 4 are after the installation of rooftop solar, where the installation of rooftop solar reduces the energy delivered to the customer by 90%.

In both Table 3 and Table 4 the two subclasses of customers would experience uniform prices of $10/customer, $50/MWH, and $20/KW.  The rates in Table 4 are the same as the rates in Table 3 despite the assumed introduction of rooftop solar, an introduction that reduced the content moving across the electric wires.

In Table 3, the realizations are a uniform amount of $117.50/MWH.  In Table 4, the realizations vary in the same manner as they do in Table 2, $117.50/MWH for all other residential customers, $725/MWH for rooftop solar customers, and $172.73/MWH for the class as a whole.  This variation is despite each subclass of customers experiencing the same prices for the three different cost classifications.  In both Table 3 and Table 4, the energy price is $50/KWH, half of the energy price without a demand charge.

Adopting a three part rate for residential customers corrects the economics of customer supplied generation.  In the example presented in Tables 1, 2, 3, and 4, the inclusion of a demand charge reduces the energy price from $100/MWH to $50/MWH.  Customer supplied energy is thus priced at the value of energy on the system.  Reforming half of the former energy charge into a demand charge effectively treats a portion of the cost as a reliability payment.  Customers who want to eliminate the demand charge need to have dispatchable generation, generation that can be expected to provide electricity reliably.  Customers who have such reliable distributed generation can effectively be a micro-grid, operating independently of the utility grid.

Adopting a three part residential rate with a demand charge will correct the economic incentives associated with customer owned generation.  Energy produced by the customer owned generation will reduce the energy portion of the customer bill.  Reliable operation of the customer owned generation can help the customer avoid part or all of the reliability based demand charge.  A three part rate can correct these economic incentives without having a general rate case or an increase in the prices paid by all other residential customers, though some regulatory proceeding would be required.

The example presented in the four tables are based on self generation reducing energy consumption by 90%.  For the histogram in Figure 1, the self generation customers reduced their energy consumption, on average, relative to their demand consumption by 19% to 23% depending upon whether the average is weighted based on demand or a simple average.  These energy consumption reductions are utility specific and other utilities will certainly experience a different impact when their customers install roof top solar.  It is likely that the 19% to 23% energy reduction calculations have been mitigated by some reduction in the customer demands, since solar power will have some coincidence with customer peak.  Thus, if this had occurred in the four tables, the allocation of the demand costs in Tables 2 and 4 would have changed instead of remaining unchanged.

Click to view a printable version of this article.