Charge for Less : An Analysis of Hourly Electricity Pricing for Electric Vehicles

By motivating Electric Vehicle (EV) owners to charge their vehicles when power supply exceeds demand, dynamic pricing can improve system load shape and capacity utilization, reduce consumer costs, and cut pollution. We compare what perfectly rational EV drivers would pay to charge their vehicle on ComEd’s hourly pricing program with costs associated with the utility’s flat-rate energy price. We find that ComEd’s hourly pricing program would have saved EV owners significantly over its flat-rate tariff in both 2016 and 2017, with cost reductions from 52 percent to 59 percent. Using price signals to manage charging is almost certainly one of the best (and cheapest) strategies to implement in order to achieve the traditional regulatory goals of a safe, reliable, and affordable service while advancing system efficiency, enhancing environmental sustainability, and facilitating the integration of distributed energy resources.

cording to the U.S. Department of Energy's National Renewable Resources Laboratory, millions of EVs on the road could increase overall U.S. electricity demand by 38 percent, or up to a sustained 80 terawatt hours per year. 2 If not managed appropriately, such an increase in usage could require costly expansion of electric system delivery and generation capacity.
Yet the Rocky Mountain Institute shows that increased power usage associated with transportation electrifi cation could be largely accommodated without additional power plants or grid expansion if EVs are charged at optimal times. 3 How can we make sure that EVs charge at the right times? While multiple strategies may be required, time-variant rates are almost certainly the cheapest way to accomplish this aim. 4 By motivating EV owners to charge their vehicles when power supply exceeds demand, dynamic pricing can improve system load shape and capacity utilization, reduce consumer costs, and cut pollution.
Particularly in states that have deployed smart meters, implementing that simple policy option can make EVs a substantial source of system benefi t, even for those who don't drive or own an EV.
Some utility EV programs to date have assumed that EVs will be price-responsive without necessarily putt ing into place measures that guarantee price-responsiveness. 5 There are several reasons for this -including the fact that we are still in the early stages of EV deploy-
While we disagree with some of her conclusions, dynamic pricing critic Barbara Alexander is correct when she says that it is "poor public policy to leap into (new methods of pricing) electricity service to residential customers without a careful analysis and access to factual information on the impacts of such proposals on customer bills." 6 In this paper, we att empt to fi ll this information gap within the realm of EVs by comparing what customers of Illinois utility Commonwealth Edison would have paid in 2016 and 2017 to charge their vehicle under average rates compared to its hourly pricing program.
We use three representative batt ery ranges and four representative daily driving amounts to do so. We fi nd that hourly prices would have yielded energy cost savings ranging between 52 and 59 percent, depending upon the circumstances, for drivers using a Level 2 charger.
The savings are even greater for Level 3 DC fast chargers. Because Level 3 charging occurs during the daily hour with the lowest priced energy, every vehicle saves 59 percent over fl at-rate energy pricing under all driving scenarios.
We then supplement these empirical fi ndings with a normative recommendation -policymakers should implement "opt-out" dynamic rates for EV charging and charging infrastructure, as none of the relevant conditions typically invoked to support fl at-rate pricing are present in the case of EVs.
With the aid of the sophisticated sensor and data-analysis capabilities prevalent in vehicle charging technology, utilities could isolate EV-related consumption, making a separate opt-out policy feasible should policymakers decide to preserve the consumer's prerogative to opt-in to hourly pricing for other forms of household usage.
We conclude by outlining why hourly pricing has several key advantages over time-of-use rates if the goal is (as it should be) to "charge for less" in both the economic and environmental sense of the term.
In this paper, we use actual 2016 and 2017 PJM locational marginal prices (LMP) to compare what perfectly rational EV drivers would pay to charge their vehicle on ComEd's hourly pricing program with costs associated with the utility's fl at-rate energy price for both Level 2 and Level 3 DC fast charging.
We started by choosing three representative electric vehicles: the 2018 Toyota Prius Prime, the 2018 Chevy Bolt, and the Tesla 3 Long-Range (Fig. 1). These vehicles off er a range of batt ery sizes, power effi ciencies, and maximum A/C charging rates, and serve as good examples of products currently on the market.
In the next step, we chose off -the-shelf representative Level 2 and Level 3 chargers to estimate the maximum achievable charge rate. Fig. 2 summarizes the specs for the two selected products from ChargePoint.
Next, while the model was constructed to allow testing of any driving level, we picked four typical daily driving amounts to simplify presentation: 15 miles (light driver); 30 miles (average driver); 50 miles (heavy driver); and 100 miles (Lyft/Uber driver). 12 In the end, then, we ran the model quantifying the results for twelve overall cells (Fig. 3).

Fig. 4: Sample
Week, July 10-16, 2017 13 The IPA procures energy for eligible retail customers in monthly on-and off -peak blocks, according to ComEd's load projections. For summer and non-summer seasons, ComEd calculates Purchased Energy Charges (PECs) equal to the weighted average cost of that energy.
"Eligible retail customers" refers to residential and small commercial customers not taking energy supply from an alternative retail energy supplier or through a municipal aggregation agreement. Summer months run from June through September; non-summer months include October through May.

Fig. 5 DHC: Daily Hourly Charges ($)
The dollar value of electric supply charges resulting from battery recharge under scenario and vehicle conditions using optimized hourly charging.

VCR: Vehicle Charge Rate (kW)
The maximum hourly charging rate for test vehicle.

T: Charging Hours (H)
The total number of hours required to recharge battery under test conditions, rounded to the next whole hour.

CHR: Charge Required (kWh)
The total amount of energy required to recharge battery under test conditions.

LMP n LMP during nth lowest ranked hour of day ($/kWh)
14 For more information on ComEd's hourly pricing program, see htt ps://hourlypricing.comed.com/ Illinois is the only state in the USA where the two largest utilities (ComEd and Ameren Illinois) off er comprehensive, "opt-in" real-time pricing programs to all residential customers. 15 htt p://www.pjm.com/markets-and-operations/energy/real-time/ lmp.aspx 16 As a PHEV, the Prius Prime has a signifi cantly smaller batt ery; for daily driving amounts above the electric only range it was assumed the batt ery was fully depleted.
With these assumptions in place, we calculated what EV drivers would pay to charge their car on ComEd's fl at-rate energy tariff to meet their daily driving needs. Because this tariff includes recovery of capacity costs and certain administrative costs, it was necessary to isolate the energy-supply only component of the fl at-rate charge to allow for an accurate comparison with hourly pricing.
These Purchased Electricity Charges (PECs) were calculated by combining Illinois Power Agency (IPA) procurement results for the study delivery years, and taking the seasonal weighted average price of energy for each month. 13 Daily fl at-rate charges were determined by multiplying the total energy required for batt ery recharge by the prevailing PEC for that month. Consumers on ComEd's hourly pricing program are charged PJM's real-time ComEd Zonal Residual LMP for their hourly volumes. 14 To calculate the costs of charging vehicles on hourly pricing, we took the hourly prices for each day in 2016 and 2017 from PJM, and placed them in ascending rank order. 15 Fig. 4 summarizes the process for the week of July 10-16, 2017.
The required daily recharge consumption is determined by each vehicle's kWh/mile drive effi ciency, divided by the number of miles in a given driving scenario. 16 For Level 2 charging, the hourly recharge consumption is equal to the vehicle's maximum A/C charge rate, and the number of charge hours equals the total kWh recharge volume divided by the hourly rate. For Level 3 charging, the recharge rate depends on the charger's rating, rather than the vehicles; in this case, the cars recharged at 50 kW per hour, for less than an hour, in all scenarios.
From this, an optimal daily charging amount was

Results
calculated as the sum of the minimal amount of charging consumption needed to meet daily driving needs multiplied by LMP during the required number of charging hours, starting with the lowest priced LMP hour and moving to the next rank-ordered LMP hour if necessary. More specifi cally, the respective vehicle's kW charging rate was multiplied by the LMP for each day's lowest ranking LMP hours up to the total number of required charging hours less one, with the fi nal hour being assessed the remaining kWh required (Fig. 5).
Once optimized hourly and fl at-rate charging costs were calculated, we fi nally compared the total charging costs for each car and driving scenario by summing the daily costs for both rate options in 2016 and 2017 and then calculating the diff erence between the two total cost summations.
ComEd's hourly pricing program would have saved EV owners signifi cantly over its fl at-rate tariff in both 2016 and 2017, with cost reductions from 52 percent to 59 percent, equaling as much as $389 over the twoyear study period. Fig. 6 summarizes the results for the 12 scenarios in the case of Level 2 Charging.
Given the daily driving amounts tested and the 50 KW charge rate, every vehicle saves 59 percent with hourly pricing over fl at-rate pricing using Level 3 DC charging.
Because this analysis assumes a perfectly rational consumer who only charges in the cheapest hour(s) needed to meet her driving needs, by defi nition Level 3 charging occurs during the hour with the lowest priced energy, and thus every vehicle and driving scenario has the same percentage savings.
Total two-year cost savings ranged from $54 to $389 depending upon the circumstances. Fig. 7 summarizes the fuel cost results of the overall analysis.
A few notes are in order. First, this is an energy-only analysis and thus does not include the costs of electric distribution, transmission, capacity, and taxes, surcharges, and fees. This approach has no material impact on the comparison between charging costs on hourly-and fl at-rate energy pricing, but it does mean that it would not be 'apples to apples' to compare the fuel costs above with the gasoline cost needed to power a traditional internal combustion vehicle.
Second, as stated previously, our model is an optimization analysis that assumes a perfectly rational charging strategy, where EVs are charged only the minimum number of hours needed to meet daily driving needs and are charged at the lowest-cost times. This is an idealized assumption, and a diffi cult strategy to implement fl awlessly even in a world with increased automation.
Nevertheless, the data reveals ample opportunity for savings even under sub-optimal conditions. More than 81 percent of the hours in 2016 and 2017 were below ComEd's fl at-rate energy price, and 23 percent of the total hours were less than 2 cents/kWh. Finally, while the total dollar amount of savings through hourly pricing (max. $389) is small in comparison to the fuel-cost savings achieved simply by switching from an internal combustion engine vehicle to an EV, this analysis does not take into account the substantial grid and environmental benefi ts inherent in price-responsive demand when targeted at reducing peaks and improving load shape.
The fact that optimized hourly pricing cut EV charging bills at least in half in the two study years without consideration of these additional benefi ts strongly indicates that dynamic pricing can play a key role in maximizing social welfare.

If the goal is to "charge for less," dynamic pricing is essential to EV charging.
Transportation electrifi cation presents a rare opportunity for all stakeholders aff ected by electricity regulatory policy to benefi t. The right set of policies can help achieve the traditional regulatory goals of safe, reliable, and aff ordable service while advancing system effi ciency, enhancing environmental sustainability, and facilitating the integration of distributed energy resources.
But to achieve these aims, we need to ensure that EVs charge at the most optimal times for the power grid. While there are other possibilities, and while multiple approaches may be needed, using price signals to manage charging is one of the best (and cheapest) strategies.
Time-based rates are eff ective at incentivizing EV owners to charge their vehicles when it will not burden the utility system. 17 And as this analysis shows, they also provide a route for EV drivers to unlock savings at the same time. For these reasons, we recommend that policymakers implement opt-out dynamic pricing for EV charging.
One rate structure is usually applied to all usage in a home, but it need not be in the case of electric vehicles, as the chargers (and/ or cars) have sophisticated sensor and data-analysis capabilities. Although we generally believe that the risks of dynamic pricing -and the concomitant benefi ts of traditional, average utility rates -are overstated, separately calculating EV charging costs can be a boon to adoption by customers who may fear having all their household usage priced under time-variant rates. 18 Because it is vital that regulatory policy get out ahead of transportation electrifi cation to maximize consumer and environmental value, we do not want to see opt-out dynamic rates for EV charging stalled because of controversies surrounding whole-home dynamic pricing.
Will EV-only, opt-out time-variant rates also prove controversial? Perhaps. But it is worth noting that none of the arguments typically made against dynamic pricing apply in the case of electric vehicles.
Consider, for example, the claim that dynamic pricing is problematic because not all consumers can respond to price signals. 19 EVs are simply diff erent than other appliances because: • they have batt eries; • batt ery capacity means even heavy drivers do not need to charge very often; • the charging process itself can be easily scheduled through automation; • EV operating costs can be reduced signifi cantly by charging in low-cost hours. In fact, electric vehicles have the ideal type of load and load shape for dynamic pricing, from both an individual owner and a societal welfare point of view. For these reasons, it is critical to utilize this kind of rate design.
Automated charging has the potential to further expand the base of customers who could realize these benefi ts when combined with machine learning. Moving from the retrospective optimization model, which relies on perfect pricing information, to a model that employs pricing algorithms to make charging decisions would allow EV owners to put this strategy into practice using a "set it and forget it" approach.
This would make the potential of realizing the full cost-savings accessible to all customers. Further research into optimized charging models should incorporate pricing models with the option to utilize strategies such as inter-day price arbitrage, skipping a day of charging, or even selling energy power as behind-the-meter generation, should a particular day's LMPs exceed expected levels.
This discussion raises the question of whether a time-of-use (TOU) or hourly-pricing rate structure is preferable. Our view is that either can work and that the primary issue is gett ing as many EVs as possible on time-variant rates aimed at ensuring charging occurs when it is most advantageous for consumers, the grid, and the environment.
That having been said, as transportation electrifi es and there are millions of EVs on the road, hourly pricing may prove the bett er alternative. To maximize the public interest, we will want to incorporate distribution system and environmental att ributes in price signals and also be prepared to respond rapidly when (and if) the peak starts to change. He is a board member of the Illinois Environmental Council.

Jeff Zethmayr, Senior Policy Analyst
As senior policy analyst, Jeff Zethmayr is in charge of writing complex testimony on key Illinois energy issues, and also analyzes the consumer impacts of energy-related legislative and policy proposals in Illinois. He was the lead author of a pioneering study on a dynamic pricing model: "The Costs and Benefi ts of Real-Time Pricing: An empirical investigation into consumer bills using hourly energy data and prices," representing the nation's most comprehensive dynamic-pricing analysis of smart meter data.
Before coming to CUB in 2014, Mr. Zethmayr provided fi nancial and strategic consulting for a number of energy-related clients, including providing a United Nations agency with a benchmarking analysis on disclosure practices in the hydraulic fracturing industry. In addition, he worked for Grassroots Campaigns, Inc., in New York, as a fi eld manager. He has extensive experience in energy/green-related grassroots campaigns. He earned a master's degree in public administration from Columbia University.
The Citizens Utility Board is Illinois' leading nonprofi t utility watchdog organization. Created by the Illinois Legislature, CUB opened its doors in 1984 to represent the interests of residential and small-business utility customers. Since then, CUB has saved consumers more than $20 billion by helping to block rate hikes and secure refunds. For more information, call CUB's Consumer Hotline at 1-800-669-5556 or visit citizensutilityboard.org.
Charging at night in Illinois because of wind -or during the day in California because of the duck curve 20 -is an easy rule-of-thumb now, but that may change as EV deployment scales. The inherent fl exibility of hourly pricing provides an advantage over administratively set TOU rates. Thus, we recommend that hourly pricing be off ered as an alternative for all EV drivers, even in states where policymakers choose an opt-out TOU structure.
Transportation electrifi cation is in its infancy, but the wheels are beginning to pick up speed and are unlikely to stop. To preserve this momentum, stay current with the evolving market, and ensure that it delivers system benefi ts requires proactive regulatory policies. Opt-out dynamic pricing must be one of those tools.
We encourage all states to seize the moment and open proceedings as soon as possible to start moving in this direction, as there are many logistical and strategic implementation questions to answer. For example, will states need to reconsider 'meter grade' billing requirements and other potential regulatory hurdles? It is possible.
Also, should third parties, such as a pharmacy or shopping center, be able to off er charging rates that diff er from the dynamic rate? We think the answer is probably yes, provided the third party (or an entity it has a business relationship with) pays the actual time variant-price.
But there are many complex questions involved here and it's important that they be carefully considered in a stakeholder process. In the fi nal analysis, if the goal is to "charge for less" in both the economic and environmental sense of the term, it is imperative that dynamic pricing is required of EV drivers.