Utility planning has become increasingly complex. The power system has always been inherently interactive, but new technologies, an increasingly dynamic distribution system, and resource variability have made it more so. We have begun to think more holistically about generation, transmission, and distribution. For example, non-wires alternatives such as distributed storage may replace the need for distribution upgrades, and transmission connecting diverse regions may remove the need to build additional capacity. Another essential tool in this toolbox for planners to meet grid needs is electricity pricing.
In the resource planning process, utilities need to forecast the load that they will serve. Traditionally, they consider historical load patterns, weather, and economic growth projections. They also need to forecast their distributed energy resources (DERs): distributed generation, distributed storage, energy efficiency programs, demand response, and new flexible loads like electric vehicles. They might project from historical trends and add accelerated growth in key DERs where costs or policies are increasing uptake. Based on their net load data, they can optimize for least-cost resource portfolios.
But there’s a chicken and egg problem here: utilities directly influence demand through electricity pricing. For example, time-of-use (TOU) rates act like energy storage — they flatten load profiles. And critical peak pricing acts like peaking plants— they shave peak load. Brattle’s meta-study of 332 dynamic pricing experiments shows that stronger price signals result in greater reduction of peak load. So how are utilities to plan when they simultaneously have the means to justify building new generators and the means to reduce the need for new generators?
Dynamic Pricing Acts Like Other Grid Resources
Historically, regulators and consumer advocates have worried that time-varying rates were too complicated or that low-income consumers did not have the technology to manage their usage. At the same time, utilities had built-in financial incentives to invest in new generation. This resulted in limited efforts in many places, e.g., air-conditioner cycling programs or efficient lighting or appliances, rather than strong price signals to influence customer behavior.
Today we have no excuse to ignore the power of pricing. Moreover, we need this tool to help us achieve our clean energy goals. While my non-college-educated mother got on a TOU rate starting in 1992 and religiously operated her 90’s vintage appliances during off-peak hours, she may have been the exception. Brattle’s meta-study shows that customer responses to time-varying rates are significantly better when enabling technologies, like smart thermostats, are available to them. Many of today’s appliances from electric vehicle chargers to dishwashers to water heaters can be easily scheduled or controlled. For example, the CTA 2045 standard developed in the Pacific Northwest for water heaters is a poster child for a plug-and-play approach to controlling appliances. The ability to control these loads is essential for the ongoing process of electrification, as electrification is a double-edged sword: some of these high-powered loads can create problems if uncontrolled (high penetrations of electric vehicle chargers contributing to peak loading on feeders); alternatively, they can alleviate problems if controlled (scheduling charging during low prices or oversupply conditions). Also, we’re talking about a lot of new demand: studies show demand increasing 50% or even 100% depending on how far we decarbonize our energy economy.
Asking customers to reduce load through demand response or peak time rebate programs are certainly a valuable and useful way to extract flexibility. But the utility needs to establish a baseline and then monitor and verify the reduction which can get complicated and perhaps may be gamed. Pricing is simpler: send customers appropriate pricing signals and let them pay for what they use. It goes without saying that there are limits as to how much volatility and risk customers can manage. Cost-causation and a number of principles (economic efficiency, equity, revenue stability, bill stability, customer satisfaction) outlined by Bonbright 60 years ago underpin the science of rate-making. But within these principles is ample room to deliver the price signals to incent desired behavior.
We need to stop thinking of all demand as firm, or “must-give” (the way we used to think about wind and solar as being “must-take” on the generation side before we got smart about how to really extract their benefits). Instead of demand being fixed and system operators balancing it with a portfolio of generators, we could (and should) be balancing both sides of this equation. There may be times in the future when total generation is fixed and system operators balance it with responsive demand.
Dynamic Pricing Will Be Critical as We Increase Wind and Solar Penetrations
We know from many integration studies that net load duration curves typically become more “peaky” (fewer hours of peak loads) as wind and solar penetrations increase. Generation resources are less attractive investments if these peaking resources are run for fewer hours; on the other hand, responsive demand becomes more attractive if it is not called upon too frequently. Reducing total system costs by encouraging responsive demand ought to be of great interest to consumer advocates and policymakers. Pricing, and its direct effects on the attractiveness of responsive demand, will become especially important as utilities move toward 100% clean energy.
There are utilities that are getting this right. Arizona Public Service has a long history of TOU rates with about two-thirds of their customers on them. ERCOT (Electric Reliability Council of Texas) skated through a record high peak demand in the summer of 2019 with a 9 percent reserve margin using some 3GW of responsive demand. Their four critical peak (4CP) transmission charges (this program allocates transmission charges for the following year on the basis of demand during the ERCOT coincident peak 15-minute intervals for the four summer months of June, July, August, and September of the previous year) were a key driver to realizing this price-responsive demand.
***
One thing is certain. As our grid transforms, it is essential that we take a systems perspective. There’s a lot to be said about comprehensive integrated planning of generation, transmission, and distribution resources, because resources in one of these buckets may obviate the need for more expensive resources in another bucket. Electricity pricing is an important mechanism to directly influence the demand that might otherwise require investment in new resources.
Addendum:
As this goes to press, ERCOT is in the middle of an unprecedented energy crisis, with the loss of tens of GW of generating resources due to extreme cold weather. Dynamic pricing is sure to come under fire as customers on real-time prices receive bills that reflect the extremely high energy market prices. It is incumbent upon us to NOT throw the baby away with the bathwater, but rather to learn from these events to design pricing structures that provide price signals that enable economically beneficial, grid-friendly behavior while simultaneously protecting customers from rare, massive failures of our system.
Debbie Lew
Associate Director, ESIG
Robert Borlick says
Your message regarding the value of dynamic rates is spot on. I have been saying the same for more than ten years.
You mention ERCOT’s use of three GW of demand response. However, that amounts to less than four percent of the system’s peak summer demand and consists solely of large customers participating in the wholesale market. In contrast, the residential and small commercial customers within ERCOT represent about 75 percent of the summer peak demand, but less than one-half percent of them are exposed to wholesale market prices so there exists a huge amount of untapped demand response. In light of the experience in Florida, Oklahoma and elsewhere, it is not unreasonable to expect these small retail customers to contribute at least another 5 GW of demand response to the system if enabling devices were universally adopted.by those choosing to expose part of their loads to the ERCOT’s wholesale spot prices.
Taken together, the demand response available from these two groups would effectively add about nine percent to the system’s reserve margin and effectively eliminate the need for involuntary load curtailments except under the most severe Force Majeure conditions. ERCOT does not need a capacity market; all it needs is more price responsive demand.
Another collateral benefit of price responsive demand is that it discourages the exercise of generator market power because the generators must not only compete among themselves but also with the customers. When I was advising the Midwest ISO it conducted a gaming experiment designed by Nobel Laureate Vernon Smith that unequivocally demonstrated the efficacy of price responsive demand in suppressing generator market power.
You correctly mention that there are limits to how much load reduction retail customers can manage. Indeed! Given the high energy price cap in the ERCOT market virtually all customers will want to hedge most of their demand – particularly so after seeing what has just happened to the GRIDDY customers. There are ways to do this. The vehicle that I have proposed is for the Texas Retail Electricity Providers (REPs) to offer fixed-price contracts (as they currently do) but include the option for their customers to sell back the energy they choose to curtail during high-priced time intervals and be paid by the REP a high fraction of the contemporaneous wholesale spot market price. Top maximize economic efficiency the customers should receive the full wholesale market less their respective contract prices, but there has to be some profit for the REPs in these transactions to motivate them to offer this option.
Of course, this would require the use of customer baselines to estimate how much energy each customer curtailed. As you correctly point out, this measurement and verification task necessarily imprecise and is subject to gaming. Even so, the benefits outweigh the costs as has been demonstrated in the ISO wholesale markets. While I love real-time pricing, very few customers are willing to take on the risk when prices can soar to $9 per kWh.
So what will it take to get the Public Service Commission of Texas and its politicians to facilitate small customer demand response? Maybe the polar vortex will provide the needed nudge.
Robert Borlick
Senior Energy Advisor
Borlick Energy Consultancy
Russ Philbrick says
Hi Debbie,
You make some exceptionally important points; perhaps some of the most important for effective integration of renewables.
Unfortunately, the debate about retail prices has–too often–been ideological. Both sides present the other as a impractical caricature: either full exposure to wholesale prices or regulated retail rates with traditional flat pricing. As identified by your mom’s experience with TOU rates, prices do shape behavior, and there are many choices other than wholesale prices.
Individual customers do not have the expertise or bandwidth to manage the risks of wholesale prices. As demonstrated by the recent events in ERCOT, it appears that load-serving entities also lack this ability, particularly in an extremely competitive market that presents operators a choice between swimming naked or never getting in the water (i.e., selling a service with no insurance, or being undercut by the competition). This does not mean that dynamic pricing is bad or dangerous (though some will argue this); instead, it is essential.
Power markets are far from ideal, and regulations are necessary to avoid a race to the bottom / tragedy of the commons, or whatever quip fits. Extreme price caps are not sufficient to meet the needs of the system, and they increase volatility and risk.
Thank you for raising a most-important topic.
David Kline says
Debbie, you rightly focus on customers’ ability to respond to price signals, e.g. with technology. If they can’t respond, dynamic prices are pointless, and seem likely to lead to outcomes that no one likes, e.g. in Texas last week.
Good comments from Russ Philbrick. As to price caps, recall that retail price caps were a significant cause of PG&E’s bankruptcy during California’s energy crisis in 2000-2001. Clearly that’s not helpful, any more than customers stuck with bankruptcy-inducing bills in an extreme event. I would love to see a rate design that manages risk in a reasonable way to avoid these extreme outcomes, if anyone knows of one.
Michael Hogan says
To Russ Philbrick, the problem with prices is not “extreme price caps” (whatever that might mean), it’s the persistence of very high opportunity-cost-reflective prices long after the “opportunities” they are expected to incentivize have largely been acted upon. There are several ideas circulating about how we might place guardrails (a “circuit breaker”) around periods of “runaway” scarcity pricing, with Australia’s Cumulative Price Threshold being one practical example among several possibilities for reform of ERCOT’s poorly designed Peaker Net Margin mechanism. But what’s “extreme” can be quite variable – in the case of the recent Texas event, there were periods when natural gas prices drove the short-run operating cost for the system’s marginal generators (the ones that were still operational) very close to the $9,000/MWh system-wide offer cap, based simply on what it would have cost to purchase gas to run the plants – what do you think would have happened if prices had been capped below that level? (Hint: A very bad situation would quickly have gotten much worse, with a total grid collapse being the likely result.) So when we talk about “extreme” pricing we need to think beyond electricity prices – or, in the case of the Texas event, instead of electricity prices. The principal culprit in the Texas disaster is very possibly going to turn out to be the vaunted Texas oil-and-gas sector.
Russ Philbrick says
To Michael Hogan,
I agree that “extreme price” is in the eye of the beholder, but prices only make sense when a functioning market exists, meaning there is competitive and flexible supply, demand or both. In emergency situations this may not exist, and prices are driven by artificial designs. You argue that $9000/MWh prices are justified by the cost of fuel, but this is nonsense. The fuel cost was driven by scarce / inflexible supply of gas and the opportunity cost of $9000/MWh prices in the energy markets. “Circuit breakers” are another way of stating that there is no competitive market and that the system needs to be balanced by other mechanisms.
It is unavoidable that markets will fail under extreme conditions. One of the goals of effective market design should be to improve the ability to manage extreme conditions, but this must be balanced against the cost of improved reliability / resiliency. There need to be mechanisms to drive an “appropriate” level of investment in the system. However, as with prices, “appropriate” is in the eye of the beholder.
Amanda Ormond says
Give people price signals and ways to easily control demand and you have a winning combination. Seems we need work on the customer side to allow appliances and other loads to be managed from one’s phone or computer. Utilities working with technology providers that can engage customers would help move things faster.