Electrification is a top priority for just about every state and province in North America and for many countries around the globe. The urgency comes from the need to mitigate climate change.^[1]

The biggest barrier facing electrification is the high cost of electricity in states^[2] such as California and New York, and in the New England states.^[3] Similar barriers exist in the Canadian province of Alberta.

While rebates from utilities and income tax credits from governments provide substantial financial incentives that lower the capital and installation cost of electrification technologies, such as heat pumps for HVAC^[4] and water heating and electric vehicles (EVs), those financial incentives don’t lower their operating costs. Thus, heat pumps are not being adopted by customers as fast as policy makers had hoped they would.^[5]

EV sales continue to grow despite high electric rates in certain regions because gasoline prices are equally high. In my case, during the past 12 months, I spent US$1,095 to charge my Model 3 Tesla and saved US$578 by not driving a comparable gasoline car.^[6] However, as electric rates continue to climb, the driving cost advantages of an EV over conventional, internal combustion vehicles will diminish, slowing the rate of adoption of EVs.

At the same time, we are seeing the accelerated deployment of time-varying rates in many regions of North America, enabled in part by the widespread installation of advanced metering infrastructure (AMI) and in part by the encouraging results that several rate design pilots have yielded. But time-varying rates by themselves cannot lower the rate level, which is the primary barrier to electrification.

There is no easy way to lower the rate level overnight, because that will erode utility revenues and create financial turbulence. We need a new rate design paradigm that satisfies three conditions:

1. Makes electrification affordable
2. Recovers the utility’s revenue requirement
3. Does not unleash a public outcry

It’s a time-honoured principle in public utility economics to price energy services at marginal costs.^[7] In theory, electrification can be encouraged by setting the energy charge equal to the marginal cost of energy in cases where marginal cost is lower than average cost. But that will create a significant deficiency in revenues for public utilities. Thus, some have suggested recovering the revenue deficiency through a fixed charge. However, in states and provinces with high electric rates, this will yield really high values for the fixed charge when the energy charge is dropped to the marginal cost of energy, which will be a lot lower than the average cost.

To deal with that adverse effect, some have suggested that the fixed charge should vary based on income, being lower for low-income customers and higher for all other customers. That was the genesis of California’s income graduated fixed charge (IGFC). In the original conception, written by three academics at U.C. Berkeley, energy prices would be set equal to marginal costs. The revenue shortfall of some US$4 billion would be recovered through a fixed charge.^[8] In the case of PG&E, the fixed charge would end up being US$74.02, much higher than anywhere else in the country and infinitely higher than the existing fixed charge of US$0.

Thus, the paper proposed to divide customers into five income tiers, and graduate the fixed charge across the income tiers, with customers in the lowest tier paying the lowest charge and customers in the highest tier paying the highest fixed charge. For customers in the highest income tier, the fixed charge would be US$186.

Even PG&E realized that such a high fixed charge would not be approved by the California Public Utilities Commission (CPUC). Thus, it proposed four income tiers with customers in the highest income tier paying US$92 a month. The average fixed charge would be US$53.

Even that proposal elicited a public outcry. It was criticized by several state legislators and members of California’s Congressional delegation, and in several major newspapers, not only in the state, but nationally. It also came in for severe criticism from the public on social media.

When the CPUC ultimately ruled on the matter, the number of income tiers had been reduced to three with those in the highest income tier paying US$24.15 a month. The average fixed charge computes to roughly US$18. Energy prices are going to be lowered only by 5-7 cents/kWh across the three investor-owned utilities.^[9]

However, even the CPUC’s approved IGFC rate design suffers from several limitations.^[10]

• It will raise bills for energy consumers who are frugal, efficient or green, without any fault of Several of them have spent thousands of dollars to lower their electric bills. Their investments will be laid waste.
• It will lower bills for consumers who use large amounts of electricity, regardless of whether they have electrified their homes and their vehicles. Many of them live in large homes that use a lot of electricity.
• There is no empirical evidence it will promote electrification, because whatever customers save on their energy charges they will lose through their fixed charge. All it will do is to create winners and losers among customers and stir public anger.
• It may even be challenged in court for being an income tax in disguise.

Clearly, the IGFC is not a Pareto optimal rate design. It’s simply an income redistribution program being carried outside the tax code.

A NEW RATE DESIGN PARADIGM

Under the existing paradigm, rate design should not be technology-specific. However, the new circumstances require us to change that paradigm, since mitigating climate change via electrification is the new reality.

Under the new paradigm being proposed in this paper, marginal cost pricing would only be applied at the margin for incremental consumption associated with the installation of heat pumps for HVAC and water heating, EV chargers and other electrification technologies such as induction stoves.

This approach is not without precedent in the world of tariff design. Today, a few utilities allow customers with EVs to be charged a rate that is specific to that vehicle if they install a separate meter. But that can be expensive. Other utilities are examining the use of telematics to bill EV customers for charging their vehicles at home, which would eliminate the need for separate metering.

As for HVAC heat pumps, under the traditional metering and pricing paradigm, this would have required end-use metering, which can be very expensive. That’s no longer necessary. Artificial Intelligence (AI) may act as a perfectly acceptable alternative.^[11] AI can infer the incremental load associated with electrification to which marginal cost pricing would be applied.

If there is some hesitation in using AI, an alternative would be to initially apply marginal cost pricing to all incremental changes in load shape. The notion of applying marginal cost pricing for incremental load shapes is not as radical as it sounds. It is not without precedent.

Georgia Power has implemented marginal cost pricing in this fashion since the 1990s for commercial and industrial customers. The pricing design they offer is real-time pricing (RTP). Both day-ahead and hour-ahead versions are provided, depending on the size of the load.^[12] The logic behind the rate is discussed in a paper by Michael T. O’Sheasy that was published in 1998 in The Electricity Journal.

O’Sheasy concludes the paper by saying:

In summary, two-part RTP affords the customer the luxury of buying above the CBL [customer baseline load] when the price is low and selling back effectively below the CBL when the price is too high. RTP customers have demonstrated an uncanny ability to do just this resulting in remarkable reductions in their cents/ kWh [effective average rate]. Since these changes are performed at a price reflecting the utility’s marginal cost, the utility benefits likewise. Imagine a classic win-win whereby the seller and the buyer are in perfect accord to buy low and sell high.^[13]

This paper applies that idea to residential customers with one modification. It proposes that marginal cost pricing be only applied for incremental loads, as in Georgia Power’s case, but limits the new rate design to those households who have either already electrified their homes or their cars or are considering electrifying their homes or their cars. Households would have to furnish proof that they have acquired either heat pumps or EVs or both. They could do that by providing a copy of their utility rate application or copy of their income tax credit application. Of course, the household would need to have a smart meter in place for this to work but 80 per cent of US households today do have that capability in the US.

Marginal cost pricing does not have to be full-blown hourly pricing. That idea will have to wait until technology advances enable prices to be sent directly to devices with the customer’s consent and foreknowledge.^[14] Marginal cost pricing could take the form of any time-varying rate, including time-of-use rates, critical-peak pricing rates, peak time rebates or real time pricing.

It could also include a capacity cost element if electrification in certain zones bumps into distribution capacity constraints.

A CASE STUDY

Consider the case of Pacific Gas & Electric Company, which serves more than 5 million customers in northern California. The average residential rate currently stands at 42 cents/ kWh.^[15] Using the E-1 tiered rate as a point of reference, the price of electricity has doubled over the past seven years, far exceeding the rate of inflation. In the seven years prior to 2017, it had only grown by 23 per cent. As a point of reference, in 2008 the rate stood at 16.4 cents/ kWh.^[16] More increases are expected to occur at year end, with the average rate possibly reaching 50 cents/kWh.

History of the average residential rate (cents/kWh)^[17]

Electric vehicle (EV) rate plans^[18]

One of the popular rates being used by its EV customers is EV2-A.^[19] The rate features three pricing periods. During the summer, the off-peak rate is 31 cents/kWh.^[20] If EV load is priced at the marginal cost of electricity, the price may drop to 10 cents/kWh. A typical household whose EV load is 3,000 kWh a year would see their annual EV driving costs drop substantially from US$930 to US$300. This would substantially enhance the appeal of EVs to drivers who are in the market for a new car, and probably accelerate the EV adoption rate.

In the areas that lie east or south of San Francisco, or in the Central Valley, summers are hot and winters are cold. A heat pump for heating, ventilating and air conditioning (HVAC) may consume 3,500 kWh a year. In the summer, a heat pump in the cooling and ventilation mode is likely to run for several hours a day, spanning the off-peak, mid-peak and peak periods. It is likely to run most intensely in the late afternoon and early evening periods. In the winter, in the heating and ventilation mode, it is likely to run mostly in the mid-peak and off-peak periods.

If the year-round peak period price averages 55 cents/kWh, the mid-peak averages 49 cents/kWh and the off-peak price averages 31 cents/kWh, then a weighted average price of 45 cents/kWh may be used to get a rough estimate of the annual operating cost of a heat pump.

With the existing rate, that would amount to roughly US$1,575. If a marginal price of 15 cents/kWh is used, the cost would drop to US$525, making it a substantially more attractive investment for customers, and probably accelerating the adoption rate. In both cases, operating costs fall by two-thirds, as brought out in the figure below.

Reduction in operating costs of electrification: Whole house v end use TOU rates^[21]

It should be noted that operating costs with whole house TOU are likely to already be lower than with a flat rate, by 5–10 per cent. These savings are dwarfed by the savings shown above. A recent paper by ESIG examined the reduction in operating costs of heat pumps associated with three different whole house rate designs: TOU, higher fixed charges, and demand charges. In all three cases, the reduction in operating costs was less than 20 per cent.^[22]

Similar calculations can be performed for other electrification technologies, such as heat pump water heaters and induction stoves.

In some areas, electrification might run into distribution capacity constraints, requiring capacity expansion. In such cases, estimates of marginal capacity costs would be added to marginal energy costs. In addition, electrification-focused marginal cost pricing should feature time variation in energy rates to avoid creating new peaks and to facilitate load flexibility.

PILOTING THE NEW RATE DESIGN

As with any new rate design, it would be good to test the empirical effectiveness of electrification-specific marginal cost pricing through carefully designed pilots before proceeding with full scale implementation.

Lessons learned from designing and evaluating pilots with time-varying rates could be harnessed to assist in pilot design.^[23] At a high level, pilots should seek to imitate the best features of medical clinical trials that are used to test the efficacy of new drugs. These clinical trials feature randomized selection of treatment and control groups, where treatment refers to the new rate design paradigm. Such designs are called Randomized Control Trials (RCT). A good example is the rate design pilot that was carried out by SMUD, the municipal utility that serves half a million customers in the Sacramento area.^[24]

If the RCT design is not possible, for budgetary or ethical reasons, pilot designs should seek to approximate such designs to the best extent possible.  Examples  include  Randomized Encouragement Design and Random Sampling with Matching Controls.

In all designs, a sufficient number of participants have to be selected so that the statistical test of differences have sufficient power.^[25]

The pilots should be designed to measure the impact of the new rate on three variables: customer adoption rates for electrification technologies, changes in customer load shapes and changes in customer bills.

To get an accurate idea of the effectiveness of the new rate design, all three variables should be measured before and after the treatments are offered. When all is said and done, the impact of the treatment will be measured as a difference-in-differences. In other words, any pre-existing differences between the treatment and control groups will be netted out of the measured difference between the two groups after the treatment has been offered. Pilots should run for more than a year.

CONCLUSION

High electric rates in many regions of North America pose a major barrier to electrification, which is viewed by many policy makers as an important enabler of reaching climate mitigation roles. While marginal cost pricing at the whole house level can lower operating costs of electrification equation to some extent, the reduction is not sufficient to accelerate customer adoption of these technologies.

It’s imperative to change the existing rate design paradigm, which argues that rates should only be applied at the whole house level and extend the rate design paradigm to allow marginal cost pricing to be applied at the technology level.

Technology-based marginal cost pricing can lower the operating cost of electrification without triggering a redistribution of wealth among customers, which would precipitate a public backlash. Rates for existing load shapes would remain unchanged. Utilities will still recover their revenue.

Society as a whole will benefit through the reduction of carbon emissions that will accompany electrification. Climate change will be mitigated. Customers who electrify will see lower bills compared to what they would be paying with gas furnaces and internal combustion engine vehicles. There will be no losers. No one will see higher bills, unlike the situation in California with the IGFC.

In other words, such a rate design will yield a Pareto-optimal outcome.

* The author is an Economist-at-Large who retired from The Brattle Group in December 2021. He has been working worked on rate design issues since 1979 when he joined EPRI’s Electric Utility Rate Design Study. He has helped design and test the impact of various forms of time-varying rates and testified in several jurisdictions in North America on modernizing rate designs. He has consulted with clients and spoken at conferences on six continents and published widely in a variety of academic and trade journals. He holds a doctorate in economics from UC Davis.

1. Masahiro Sugiyama, “Climate change mitigation and electrification” (2012) 44 Energy Pol’y 464, online: <www.sciencedirect.com/science/article/abs/pii/S030.142.151200033X>.
2. U.S. Energy Information Administration, “Rankings: Average Retail Price of Electricity to Residential Sector, July 2024 (cents/kWh)”, online (pdf): <www.eia.gov/state/rankings/?sid=US#/series/31>.
3. Hawaii has the highest electric rates in North America but since natural gas is not widespread there.
4. See, for example, Consolidated Edison Company of New York, “Get Thousands Off an Air-Source Heat Pump”, online (pdf): <www.coned.com/en/save-money/rebates-incentives-tax-credits/rebates-incentives-tax-credits-for-residential-customers/electric-heating-and-cooling-technology-for-renters-homeowners/save-on-a-central-air-source-heat-pump>. Financial incentives in the $10,000 range are being offered in New York. Even then, the all-in installed costs of a heat pump HVAC system in states such as California can exceed $25,000.
5. Another option being considered by policy makers in a few states is to ban natural gas in new buildings or to put a tax on the consumption of natural gas. The experience of the City of Berkeley in California is discussed in this blog by Professor Severin Borenstein. See Sanem Sergici et al, « Heat Pump–Friendly Cost-Based Rate Designs” (January 2023), online (pdf): <www.esig.energy/wp-content/uploads/2023/10/Heat-Pump–Friendly-Cost-Based-Rate-Designs.pdf>.
6. This is based on information from the Tesla app under Charge Stats. It assumes that an equivalent gasoline vehicle would have 30 mpg and gasoline would be priced slightly above $5 a gallon.
7. Paul L. Joskow, “Contributions to the Theory of Marginal Cost Pricing” (1976) 7:1 The Bell J of Econs 197, online (pdf): <www.jstor.org/stable/3003196>.
8. Leo Borenstein, Severin, Meredith Fowlie and James Sallee, Designing Electricity Rates for An Equitable Energy Transition, Next10 (2021), online (pdf): <www.next10.org/publications/electricity-rates>.
9. Stephanie Wang, “California approves restructure to income-based electric bills, cut to residential electricity prices” (22 May 2024), online: <www.dailycal.org/news/state/california-approves-restructure-to-income-based-electric-bills-cut-to-residential-electricity-prices/article_664622d2-181e-11ef-ba29-4fb356af5997.html>.
10. Ahmad Faruqui, Jim Lazar & Richard McCann, “New electricity rate reform in California: A rejoinder to Meredith Fowlie” (2023) 11:4 Energy Regulation Q, online: <energyregulationquarterly.ca/articles/new-electricity-rate-reform-in-california-a-rejoinder-to-meredith-fowlie>.
11. For EVs, there is another option, which is to use the telematics in the car for metering their usage.
12. See Georgia Power, Electric Service Tariff: Real Time Pricing – Day Ahead, online (pdf): <www.georgiapower.com/content/dam/georgia-power/pdfs/business-pdfs/rates-schedules/RTP-DA-5.pdf>. See also Georgia Power, Electric Service Tariff: Real Time Pricing – Day Ahead, online (pdf ): <www.georgiapower.com/content/dam/georgia-power/pdfs/electric-service-tariff-pdfs/RTP-HA-6.pdf>.
13. Michael O’Sheasy, “How to buy low and sell high” (1998) 11:1 The Electricity J, online: <www.sciencedirect.com/science/article/abs/pii/S1040619098800201>.
14. OGE in Oklahoma has been doing that for almost a decade. It sends four-tiered critical-peak pricing rates directly to smart thermostats of the roughly ten percent its customers who have enrolled in such rates. The critical-peak pricing rates have four tiers, depending on the nature of the power-supply balance. OGE does not control the thermostats but merely sends the signal to them. It is up to the customer to program their thermostats to respond to the price signal. Similar ideas have been expressed for charging EVs. The key is gaining the driver’s assent and communicating successfully with the EV charger. Of course, the drivers have to keep their EV plugged in for “managed charging” to work.
15. The rate has been discounted temporarily during the summer months. Earlier, it had reached 46 cents/kWh.
16. See Pacific Gas and Electric Company, “Electric rates: Current historic electric rates” (last visited 23 October 2024), online: <www.pge.com/tariffs/en/rate-information/electric-rates.html#accordion-a84c67dc1e-item-e10eec0cc5>.
17. Based on annual data for the E-1 tariff provided by PG&E. Pacific Gas and Electric Compagny, “Electric rates: Current and historic electric rates” (last visited 29 October 2024), online: <www.pge.com/tariffs/en/rate-information/electric-rates.html>.
18. Pacific Gas and Electric Compagny, “Electric Vehicle (EV) rate plans” (last visited 29 October 2024), online: <www.pge.com/evrates>.
19. Allen Meredith, Electric Schedule EV2: Residential Time-of-use: Service for Plug-in Electric Vehicle Customers, Pacific Gas and Electric Company (2023), online (pdf): <www.pge.com/tariffs/assets/pdf/tariffbook/ELEC_SCHEDS_EV2%20(Sch).pdf>.
20. It was 17 cents/kWh in 2019, just five years ago.
21. Author’s computations.
22. Supra note 6.
23. See Ahmad Faruqui & Sanem Sergici, “Household response to dynamic pricing of electricity–a survey of 15 experiments” (2010) 38 J of Regulatory Econs 193. See also Sanem Sergici, Ahmad Faruqui & Sylvia Tang, “Do Customers Respond to Time-varying Rates: A Preview of Arcturus 3.0” (2023) Brattle, Working Paper, online (pdf): <www.brattle.com/wp-content/uploads/2023/02/Do-Customers-Respond-to-Time-Varying-Rates-A-Preview-of-Arcturus-3.0.pdf>.
24. Meredith Fowlie & al, “Default Effects and Follow-On Behavior: Evidence From An Electricity Pricing Program” (2020) Energy Inst At HAAS, online (pdf): <haas.berkeley.edu/wp-content/uploads/WP280.pdf>.
25. Eduardo Hariton & Joseph J Locascio, “Randomised controlled trials — the gold standard for effectiveness research” (2018) 125:13 An Intl J of Obstetrics & Gynaecology 1716.