Demand charges vs. TOU rates: The great Arizona rate design experiment
Rate design experts debate APS's plan to be the first utility with mandatory demand charges for residential customers
To cover a utility's fixed costs, are demand charges or time-of-use (TOU) rates superior?
It's a question on the cutting edge of utility rate design discussions across the country, and one Arizona regulators are addressing on the ground today.
In its newly-initiated $3.6 billion general rate case (GRC) for 2016 to 2018, Arizona Public Service (APS) is asking state regulators to approve the first-ever mandatory demand charge for an investor-owned utility.
APS Director for State Regulation and Compliance Greg Bernosky thinks a demand charge is the best way to manage the utility’s peak demand and its costs.
“If we have a price signal that gets customers to scale back their energy use during our peak demand period between 3 p.m. to 8 p.m. on weekdays, we can spend less on infrastructure and fuel costs,” Bernosky told Utility Dive on the sidelines of the recent Solar Power International 2016 conference.“We have to provide a precipitous amount of energy in that peak demand period and that is a real cost driver because that system peak costs a lot of money.”
But rate design experts who joined Bernosky for a panel discussion at the annual conclave say there are many questions yet to be answered about demand charges, and one is whether time-of-use rates are superior to reduce peak demand and fit consumer needs.
“If the utility’s real goal is to find a rate tied to cost recovery that provides price signals customers can understand and act on, it needs to look at more than one rate,” said Rick Gilliam, distributed generation program director at Vote Solar, an industry advocacy group.
Gilliam, representing Vote Solar, was party to the recently-resolved Colorado proceeding in which Xcel Energy and renewables advocates agreed to a settlement that included both a demand charge pilot and a TOU rate trial.
“We are not opposed to the APS demand rate proposal but the mandatory nature is problematic,” said. “The utility is being honest that it has a [fixed cost] problem to solve, but the way it is trying to solve it is to impose what amounts to a fixed charge on customers who would be unable to address the new rate effectively.”
Demand charges vs. TOU rates
Utilities have long used demand charges to reduce consumption from commercial and industrial customers, but their application to residential consumers has only caught on in the past two years. As regulators across the country scaled back utility proposals to increase fixed charges, sector leaders began to look at new options to cover system costs.
Both demand charges and time-of-use rates aim to use price signals to reduce consumer usage and shift it to off-peak hours, noted a recent report from Rocky Mountain Institute (RMI).
TOU rates make electricity consumed during high-demand hours — such as the late afternoon — more expensive, since it costs more for the utility to generate and deliver power at those times. This better aligns the utility rate structure with the costs of service, and helps shift consumer usage away from those hours.
Demand charges, by contrast, impose a per-kilowatt charge for all energy consumed during a customer’s highest demand hour each month. The cost for that hour would be added to customers’ volumetric and fixed bill charges, motivating the customer to use practices and technologies to bring it down.
When applied successfully, either rate structure can reduce peak load enough to allow utilities to defer or cancel costly investments in grid infrastructure, Dan Cross-Call, a manager in RMI’s electricity practice and co-author of the report, "A Review Of Alternative Rate Designs," recently told Utility Dive.
Demand charge appeals to utilities because they are seen as providing a more certain way to cover flat or declining revenues, he said. To make them more palatable to consumers, demand charges are typically introduced along with decreases to volumetric and/or fixed fees.
The new rate structure APS proposed in its GRC includes such lower volumetric rates, along with a lower net energy metering credit to solar owners, higher fixed charges, and three demand charge rates — R-1, R-2, and R-3.
Typical residential customers, classified as R-1, would have their basic service charge increased to $24/month. They would be subject to a $6.60/kW demand charge during both the summer and winter.
A second (R-2) class of residential customers could choose to limit their fixed fee increase to $14.50/month by taking a higher summer and winter demand charge of $8.40/kW.
Higher peak demand consumers would make up a third (R-3) class of residential customers. They would have the $24/month fixed fee. Their summer demand charge would be $16.40/kW and their winter demand charge would be $11.50/kW.
To protect smaller consumers, a flat per-kWh rate option would be available to the 200,000 of the utility’s 1.1 million customers who use 600 kWh per month or less, according to Bernosky, but they would have to opt-in to the flat rate.
“The APS case is a very big deal because it would be the first investor owned utility to use a mandatory demand charge,” Autumn Proudlove, senior policy analyst at the North Carolina Clean Energy Technology Center. “There have been 14 pending or decided demand charge proposals since the fourth quarter of 2014 in the U.S. but no commission has approved one for an IOU as long as we have been tracking them.”
Still, fixed charge proposals remain relatively uncommon in the sector. During the same period, Proudlove said her group tracked 81 proposals for fixed charge increases, a more traditional utility response to cover system costs.
Proudlove suggested three reasons utilities and commissions have likely been reluctant regarding demand charges.
First, they are unfamiliar, especially for residential customers, so planners and regulators lack experience with them. Second, they are seen as complicated for residential customers.
Third, many utilities don’t have the advanced metering infrastructure (AMI) in place that is necessary to make demand charges actionable for customers.
“That introduces the questions of who would pay for the AMI and whether it is worth the cost,” Proudlove said.
AMI is fully deployed in the APS territory, Bernosky said, but critics say the utility would be wise to use that technology to roll out time-of-use rates, rather than mandatory demand charges.
Peak demand impacts
For APS, the demand charge is “a better price signal than a time-of-use rate because it aligns the cost-causing with the way we collect our costs from customers,” Bernosky said.
“We believe the new structure captures more of the fixed costs and the costs of meeting system demand.”
As the energy provider of last resort, the utility has to design its system around the peak need and maximum usage of its customers, Bernosky went on, and the TOU rate “does not quite get you to what the peak period inside the time-of-use window means from a system cost perspective.”
The objective is to lower the aggregate customer peak demand spike so the utility can avoid or defer capital expenditures for generation facilities and fuel contracts. For that, the demand charge is a more specific signal, Bernosky said.
“Our spike is the customers’ aggregate spike so if they understand how to manage their peak they will be lowering their aggregate peak usage and that lowers the utility’s peak.”
APS, like other utilities pushing demand charges, is doing so after regulators scaled back requests for hefty hikes on fixed monthly charges to rooftop solar customers. Bernosky now says those charges are not the correct way to cover fixed costs from the entire rate base.
“Fixed charges are a blunt instrument," Bernosky said during the panel discussion. "They are a way to collect costs but they do not send a price signal to the customer. The demand charge more effectively sends a price signal for customers to reduce their demand when it is needed.”
The one-hour peak demand charge window was selected as “the sweet spot between bigger and smaller windows and as a reasonable proxy for what the system needs because we plan our procurement around hourly needs and our trading floor is set up around hourly increments,” Bernosky said.
But it also allows customers to smooth any 10 minute or 15 minute spikes in usage during the hour, he added.
“We are trying to roll up our sleeves and think about the impacts on solar customers and non-solar customers, storage developers, people with load control and smart thermostat technologies,” Bernosky said. The demand charge “better informs customers who are considering adopting the new technologies that will help them help us by reducing their peak period usage.”
Vote Solar’s Gilliam did not disagree. “We are in a phase where new technologies are really impacting the revenue of utilities," he said. “The key to the rate design puzzle, whether it is demand charges or time of use rates, is to make sure that when revenue is eroded, costs are reduced at the same time, not just today but over the next five years or ten years.”
But, Gilliam added, TOU rates are tied to the utility’s peak, not just those of individual customers.
“What individual customers will have to pay is based on their individual peaks and those could be very different than the APS peak,” he said.
A demand charge, based on the customer’s one hour of highest usage, ends up being “roughly 60% of their total cost of service, and that is an enormous impact on the bill for one hour of usage which may have nothing to do with when the utility has its peak.”
The TOU rate, by contrast, provides a price signal to customers to keep their usage lower during the utility’s entire on-peak window.
“That spreads the impact of a single hour of excessive use over the entire set of hours in the peak windows for the month,” Glliam said. “The effect on the customer is diminished but they are reducing their usage during the time when the utility expects to see its peak demand. I fail to see how that does not tie usage to cause with a price signal more effectively than demand charges.”
During the panel, Gilliam cited an NV Energy assessment of when its costs are incurred. It concluded the demand charge does not effectively address peak load because the customer’s peak demand can come at any hour, and any rate design that makes the demand charge more specific becomes more like a TOU rate.
“The answer seems to be to assign transparent prices to each hour,” Gilliam said. Peak period hours can be priced “to get more of a reaction from customers,” he added. “But tie the price signal to the customer to the time period when the costs of the utility are being incurred. That is what demand charges do not do.”
Vote Solar program director Briana Kobor, who will participate in the GRC proceeding, believes the goal of rate design should be managing demand.
“But even if it is increasing fixed cost recovery, it is easy to see that with a four-hour TOU peak you can incent customers to do something with a softer hand than with a demand charge,” she said. A TOU rate “gets to the same thing as a demand charge without the significant unavoidable increases on customer bills.”
APS is confident its customers can benefit from what Bernosky calls a TOU rate with a demand component.
Of the approximately 120,000 of the utility’s customers enrolled in its voluntary residential demand charge program, about 60% have learned to use them to lower their demand period usage and overall energy usage, Bernosky said.
Over 90% of the 60% who lowered their usage had average bill savings of 9% in an APS study of a 12-month period between 2012 and 2014, he added.
During the summer, when there are more options for energy management, those customers averaged a monthly demand period usage reduction of 3% to 4%, he said.
“The demand charge is a kind of fixed charge, but it is one you can control,” Bernosky said. “If we can get people to shift their use, it will lower the highest demand that the system is being built around and the savings will flow back to customers.”
Customer impact matters to APS, Bernosky said. The cost shift now imposed by owners of distributed generation “is a material impact to 96% of our customers.”
The new rate design was included in the GRC proposal “because we don’t like to push more costs on customers than we have,” he added. “That is why we want to thoughtfully apply this concept more broadly. We think customers can and are materially saving when they are on the demand rate.”
By reducing the volumetric per-kWh rate at the same time the demand charge is introduced, most customers “will not notice a difference,” Bernosky said. “But there is a bill component the customer can control.”
Owners of distributed generation will have to manage their costs, he acknowledged. But the battery storage and load management technologies industries are waiting for this type of price signal.
The demand charge is “a real opportunity because customers will begin to understand what we see as a trend toward more DER,” Bernosky said.
Smart thermostats, connected appliances, battery storage and other DERs are not yet mainstream, but in three to five years they are likely to be, he added. “This is a rate design that sends a price signal to the customer that this is good direction to go in.”
APS certainly has more experience than most utilities with demand charge implementation, Proudlove said.
Its proposed one hour peak demand charge period “is definitely better for the customer than 15 minute or 30 minute periods because usage spikes will have less of an impact,” she noted.
“But the customers that are on the rates now may have self-selected, knowing they could save money,” she added. “I wonder if the utility has considered whether the results will be the same with all customers.”
VoteSolar’s Kobor has looked at that question.
The participants in the APS voluntary demand charge program are self-selected customers who had used the APS-provided tool and determined they could benefit from the demand charge rate, she said. “Yet 40% increased their peak demand, exactly the behavior the charge is supposed to discourage.”
That sample of customers, expected to be more knowledgeable than the average ratepayer, demonstrates that many will likely have trouble dealing with the new rates if they are imposed mandatorily, Kobor has concluded.
“For most residential customers, a demand charge will function similarly to a fixed charge because what defines a customer’s peak load in a typical household are not things people have a lot of control over,” Kobor said. “Even with significant customer education, and that is a big hurdle, I don’t think APS customers will be able to react to the price signal in a demand charge.”
Gilliam expanded on her point. “The demand rate that is now voluntary started as mandatory in 1980, but there were so many complaints that the commission changed it to an optional rate after three years,” he said. “After over 35 years, almost 90% of their customers have opted not to use it.”
Solar advocates also point out one other example of a mandatory demand charge, imposed only for solar customers by Arizona public utility Salt River Project, did not immediately lead to an uptick in DER adoption. Quite the opposite.
After imposing the rate in December 2014, solar interconnection applications fell from 3,467 in the last quarter of that year 64 in in the first quarter of 2015, according to data from Lawrence Berkeley National Laboratory (LBNL) Research Scientist Galen Barbose.
New applications for Q2 2016 were up to 173, far from 2014’s per quarter average of 1,930.
The GRC proposal seems aimed mainly at solar owners and at driving them toward additional major costs for storage and other DER technologies, Gilliam said.
“If that is the direction they want to go, why don’t we work together to develop policies, incentives, and programs to move us in that direction together?" he asked. "You drive the market by gradually working toward a solution, with adjustments and refinements along the way, not by a Draconian overnight change for 800,000 customers.”
Wanted: More data
Too often, Gilliam said, utility rate design proposals are made that are not supported by independent data that allows comparing them in a statistically valid way to other rate concepts.
“In Colorado, we will collect data and study and compare the results of the demand charge pilot and the time of use trial so the final choice between them is reasoned and data-driven,” he said.
APS stands by the numbers from its voluntary program. “We have had several decades of customers on this kind of rate and we have been able to go back and look at what the impacts have been,” Bernosky said.
There are unlikely to be other studies relevant to the APS rate plan, he acknowledged. “But that does not mean our proposal is not a sound one. We have a unique precedent within our system. We feel good about the data we have seen and we think there is something here that folks can work with.”
Galen Barbose, who has studied the impacts of a separate demand charge program from Arizona public utility Salt River Project, linked both Gilliam and Bernosky's points.
“There is analysis of customer price response to other kinds of [rate] designs but I don’t know if it can be transferred to mandatory demand charge rates because that is a very different kind of price signal,” he said. “It is an area ripe for further research and I hope APS and the Arizona Corporation Commission use this as an opportunity to do some useful and public analysis.”