This piece is part of a series. Read more.

The avoided-cost calculator: The controversial metric at the center of California’s solar net-metering fight

Tying rooftop solar payments to climate and grid benefits makes sense in theory. But in the real world, it could get very messy.
By Jeff St. John

  • Link copied to clipboard
(H. Armstrong Roberts/ClassicStock/Getty Images)

Welcome to the first installment of Down to the Wire, my new column for Canary Media that will tackle the more complicated challenges of decarbonizing our energy systems. Clean energy journalism can’t just focus on technological breakthroughs; it must also dig deep into the details of utility regulation, energy markets, and politics and policy from the local up to national and international levels.

Every few weeks, we’ll hack our way into the weeds on what it will take to build a clean power grid and electrify transport, heating and industry. How can we achieve breakneck growth of wind and solar power and energy storage at gigawatt scale? How can we balance distributed solar and storage, electric vehicles and grid-interactive buildings? And how well do theoretical concepts work out when put to the test in the real world? 

I’m looking forward to your feedback and suggestions — contribute comments below, tweet at me or send me an old-fashioned email (jstjohn at canarymedia dot com). Let’s get started.

Subscribe to receive Canary's latest news

Back in 2016 when the California Public Utilities Commission made its last big decision on net metering, it promised that the next time it set the values for rooftop solar that’s exported to the grid, it would go beyond simply basing it on utility retail electricity rates.

The CPUC’s 2016 order said it would find a way to link the value of the state’s rooftop solar to some measure of the benefits it provides to the electricity system and society at large. These include reducing the need for power plants or batteries to supply peaks in grid demand, cutting down on grid infrastructure upgrade costs and slashing greenhouse gas emissions.

Well, now the time has come for net metering to be reformed yet again — and the CPUC has finally settled on the metric it will use to determine the holistic value of net-metered solar. It’s called the avoided-cost calculator, or ACC, and while it’s been used for decades to measure the cost-effectiveness of energy-efficiency investments and programs, this will be the first time it’s been put to use to directly compensate rooftop solar for its grid value.

This change comes amid an all-out battle over net-metering reform in California. On Monday, the CPUC released its long-awaited proposal for updating net-metering rules, which I wrote about earlier this week. It was immediately blasted by solar industry groups for drastically reducing the economic value of customer-owned rooftop solar systems. Solar companies and their allies are pushing hard for the CPUC to release an alternative proposal before its commissioners vote on final reforms early next year.

The ACC has not gotten much attention in this fight so far, but using it to determine the rates paid for rooftop solar could make a bad situation even worse for owners of home solar systems. It’s theoretically a good thing to match solar export credits to the value that rooftop solar provides to the grid. But in practice, the ACC system the CPUC has proposed could make it next to impossible to predict the value of clean energy generated by rooftop systems, turning the country’s biggest rooftop solar market into an economic minefield.

California law requires the CPUC not just to balance the costs and benefits of net-metered solar between those who own rooftop systems and those who don’t, but also to ensure a sustainable” growth path for the rooftop solar industry. While using the ACC to put a value on exported solar may meet the first requirement, it’s not clear if it will meet the second, according to solar industry groups.

They’re asking the CPUC to avoid imposing too direct a link between how much rooftop solar systems are paid and the hour-by-hour values derived from the ACC’s models and calculations. Those ACC values — which are reassessed every year — are simply too complicated, volatile and unpredictable to give solar customers, developers and their financial backers the confidence they need to make investments with predictable payback periods, solar companies and their allies say.

The values in this calculator are the result of a series of very complex calculations that connect with predictions about the future of clean energy penetration on our grid,” said Susannah Churchill, Western senior director for nonprofit group Vote Solar. Given that complexity, I don’t think anybody has a full picture of how it exactly works,” she said.

What is the avoided-cost calculator? 

Let’s walk through the issues surrounding the avoided-cost concept and the avoided-cost calculator, which will be increasingly important in California energy policy and politics. First, what exactly are we talking about?

Avoided cost in this case is defined by the CPUC as the marginal costs a utility would avoid in any given hour” if a distributed energy resource provided power instead of the utility. Avoided cost is a common concept in the utility world and has been used for decades to establish such metrics as the value of non-utility power plants under the federal Public Utilities Regulatory Policies Act of 1978 — a subject worth a book or two all on its own.

The ACC — essentially a complicated set of formulas calculated on spreadsheets — was created for the CPUC in 2006 by consultancy E3 to assess energy-efficiency programs, i.e., the value of reducing energy demand. But it has since expanded to encompass the value of all distributed energy resources, including those that inject power to the grid, like solar panels and batteries.

Mohit Chhabra is a senior scientist with the Natural Resources Defense Council, a prominent environmental group that’s criticized California’s existing net-metering regime for unfairly imposing undue costs on lower-income residents. He explains the concept with the following thought experiment — what would your solar panel’s energy be worth if you moved it from your roof to somewhere on the utility grid? If you were to take a solar panel and, instead of putting it behind the meter, put it in front of the meter, how much would you expect to get paid?”

That value would start with the wholesale energy market price at the nearest pricing node of state grid operator CAISO, he said — a price well below the retail energy price charged by utilities. It would add in the value of how much that incremental unit of energy is worth in reducing transmission and distribution grid costs, since energy produced at the edge of the grid doesn’t need to be carried from far-off power plants with the resulting line losses and wear and tear on the grid itself.

The value would also depend on when the solar energy was produced and whether that hour-by-hour unit of energy helped meet demand at peak times of the day or year. This peak-demand-matching value is generally called capacity, or, in CPUC’s parlance, resource adequacy.

Then it would incorporate the value of reduced carbon emissions, based on whatever generation source the rooftop solar was replacing — typically natural-gas-fired power plants, plus the methane leakage from the natural-gas networks serving those power plants. This is changing, though, as California now plans to rely on gigawatt-scale batteries instead of gas plants to meet its future capacity needs.

Because avoided cost is a general value applied across the entire California grid, the ACC can’t be used to capture locational benefits down to the level of individual substations and grid circuits, Chhabra noted. Methods to capture those values are being developed in another sprawling CPUC proceeding aimed at preparing the grid for a high distributed energy resource future.”

But the ACC does break down values across the territories of the state’s big three utilities and across the California Energy Commission’s 16 climate zones,” which help determine the values of investments in building energy efficiency that affect cold-weather and hot-weather energy demands in different ways.

And while the ACC values are calculated over time horizons that stretch into the decades to capture the costs and benefits of long-lasting energy-efficiency investments, they can also be broken down into hourly increments for all 8,760 hours of the year, Chhabra said.

All of these particulars make the ACC a much more fine-grained way to establish the value of a kilowatt-hour of exported rooftop solar, he said.

The downside for net-metered solar is that the hours when solar arrays are generating most of their electricity are not the hours that carry the most value. California’s increasing number of solar panels, both on roofs and in utility-scale projects, saturate the grid at midday hours, and then their production drops off dramatically just as demand peaks in the evening.

This chart shows that relationship, with the dotted line showing a typical day’s solar production and the vertical bars representing the value of energy as calculated by the ACC, in this case in 2019. Additional units of solar become less and less useful for meeting the grid’s evening peak needs.

(E3)

In contrast, retail electricity rates, which rooftop solar system owners earn for the power they feed into the grid under California’s current net-metering system, are a more blunt instrument. A host of fixed” utility expenses — like building and maintaining the grid and subsidizing the bills of lower-income customers — are built into the volumetric,” or per-kilowatt-hour, price that customers pay for electricity, and that many net-metered solar customers avoid paying.

In this sense, retail-rate net metering is kind of a dumb, static policy,” said Rob Rains of Washington Analysis, who has analyzed California net-metering policies. 

Regular retail rates vs. time-of-use rates vs. ACC-based rates

Critics of net-metering policy have found plenty of faults with it since it was first introduced. Various attempts to create more direct links between exported solar’s value and how much it’s paid for that value have been developed in different places around the world, as well as a handful of U.S. jurisdictions. Those include feed-in tariffs that meter and pay for exported solar separately from a customer’s grid consumption, and value-of-solar tariffs” that use a case-by-case mix of metrics to set exported solar’s price.

But net metering does have simplicity in its favor. Reading a customer’s meter to determine how much energy they used and how much they exported is easier and cheaper than installing a second meter to measure exported energy or requiring more sophisticated metering to measure moment-to-moment fluctuations in energy exports and imports. That was particularly true before smart meters became widespread in California.

Now, though, the state’s big three investor-owned utilities, Pacific Gas & Electric, Southern California Edison and San Diego Gas & Electric, have installed smart meters for all their customers, so they’re able to measure energy consumption in hourly or 15-minute increments and set rates and bill customers on an hour-to-hour basis.

This has allowed the CPUC to order the three utilities to create time-of-use” rates that give power a higher value during on-peak” late afternoon and evening hours when solar power is scarce and a lower value during off-peak” mornings and midday hours when solar is plentiful. New net-metered solar customers have been required to sign up for these TOU rates since 2017, and all residential customers of California’s three big investor-owned utilities are being transitioned to them now.

The following graphic from SDG&E shows how electricity prices differ over the course of a day.

(San Diego Gas & Electric)

How the CPUC, utilities and their allies want to use the avoided-cost calculator

When the CPUC launched its latest net energy metering (NEM) proceeding, colloquially known as NEM 3.0,” early this year, it declared that it intended to use the ACC to value solar exports, but it didn’t get into specifics. It asked organizations that would be participating in the proceeding to propose new ways to use the ACC to value rooftop solar fed to the grid. Consultancy E3 got the ball rolling in January of this year with a report on alternative ratemaking methods that used the ACC to set the assumptions for the avoided costs that should be assigned to solar power flowing back to the grid.

California’s three big utilities then put forward net-metering reform proposals that would radically shrink the value of solar exported to the grid as well as impose fees on new solar systems, with both changes linked to ACC values. Importantly, the utilities are asking to adjust rates and fees annually as ACC values are recalculated, which would mean prospective rooftop solar customers would not be able to lock in” a rate that would allow them to calculate the lifetime value of their investment.

Other groups that are seeking to reduce net-metering compensation have also asked the CPUC to tie solar export values to the ACC. The consumer advocates at the CPUC’s Public Advocates Office have also proposed a rate that would change every year as the ACC is recalculated. The Utility Reform Network, a utility-customer watchdog group, has proposed a hybrid that would set export rates based on both the current ACC value and CAISO energy market prices.

In the proposed net-metering reform plan it released Monday, the CPUC itself got more specific about how it would like to use the ACC. Its new proposal embraces the ACC as the core metric for setting hourly solar export compensation rates, with different rates for each of the state’s 16 climate zones. Those hourly values would be locked in at an average rate over the first five years of a new rooftop system’s interconnection to the grid, and then would be adjusted every year based on ongoing revisions to the ACC. This, the proposed decision states, would help ensure that benefits are approximately equal to the total costs.”

How solar companies and their allies want to use (or not use) the ACC

None of this sits well with solar groups, for a number of reasons. The first problem they cite is that the ACC substitutes a novel and poorly understood measure of value for the retail rates that California solar customers, installers and financiers have used for more than a decade.

In its net-metering reform proposal submitted earlier this year, the California Solar and Storage Association asked the CPUC to gradually decrease solar export rates over time as a percentage of existing retail rates. We still propose that it be based on [retail] rates, because that’s something that customers understand,” said Brad Heavner, the group’s policy director. These types of gradual rate reductions are referred to as glide paths.”

Specifically, solar industry groups have asked the CPUC to rely on time-of-use rates as well as electrification rates” being proposed by utilities to encourage customers to adopt electric vehicles and electric heating, which also vary by time of day. Both of these retail rates are much more predictable than ACC values, which shift annually. The groups say using the more straightforward rates would make it much simpler for customers to understand and compare the likely costs and benefits of installing a new rooftop solar system.

The joint net-metering reform proposal from advocacy group Vote Solar and the Solar Energy Industries Association trade group, for instance, aims to reduce compensation for clean energy exports, but in a way that’s predictable and pegged to more stable retail rates,” Vote Solar’s Churchill said. This glide path” proposal would use the ACC to set the target for where solar export rates would end up over the course of five to seven years or so, but it would still set that compensation as a declining fraction of existing retail rates, rather than basing it on the ACC.

The second big problem for solar groups is that the ACC itself can change in ways that are hard to predict. That’s why solar groups asked the CPUC to allow customers to lock in export rates that use values averaged out over much longer periods of time than from year to year. The CPUC’s new proposal calls for locking in rates for five years, but solar companies would like to see that timeline lengthened.

In its proposals to CPUC, the California Solar and Storage Association also offered an alternative glide path” under which export rates would not rise as a percentage of retail rates, acknowledging that retail rates are almost certain to increase notably over the coming decade. But the proposal asked that lower rates be locked in over the lifetime of a rooftop solar system. The following chart shows the downward-trending steps” in this alternative glide path, which would keep export rates well below those that would be earned under the current NEM 2.0” system and, decades down the line, even below the rates proposed by the CPUC’s Public Advocates Office, referred to in the chart as Cal Advocates.” 

(California Solar and Storage Association)

That’s sensible both from the customer perspective and the utility perspective,” Heavner said. Your later-year benefits are lower, but you’ve already paid back the system” through locked-in rates in earlier years, so that’s OK.”

What won’t work, solar groups say, is to change what net-metered customers earn from their exported solar every year based on adjustments to the models, methods and regulatory and legal disputes over the ACC.

That would leave customers with zero certainty on whether their investments would be worthwhile,” Heavner states in a joint CPUC filing he co-wrote with Joshua Plaisted, a principal with consultancy Flagstaff Research. This single factor would be a poison pill for the entire market.”

Levelizing” ACC values — calculating their average value over a decade or more and then locking them in place for customers based on the year the solar system is installed — is a more appropriate use of the ACC, Heavner said. After all, the ACC was developed to guide the cost-effectiveness of long-term investments like energy-efficiency retrofits, and solar is a 25- to 30-year resource” with an equivalent lifespan, he said.

If ordinary Californians are going to invest thousands of dollars on rooftop solar and batteries, they need to reasonably predict how much they can save per month on their utility bills,” said Churchill.

Rooftop solar financiers need to be able to predict it too. The chain of financing stretches from customers and solar installers to solar loan providers and buyers of solar-backed securities on Wall Street.

Without some stable and predictable revenue stream, solar companies that install [distributed energy resources] will not be able to provide customers with any assurance as to their future economics,” Tom Beach, principal consultant at Crossborder Energy, states in a CPUC filing submitted on behalf of SEIA and Vote Solar. Nor will lenders have confidence that solar customers will save enough money to pay back the loans used to finance systems, and today’s robust market for solar financing in California may dry up,” he wrote.

But Monday’s proposed decision from the CPUC ignores or rejects these assertions from the solar industry. It states that ACC values have consistently reflected the value of exported energy year after year” and that changing those values every year brings the cost of the tariff closer to its value.”

That time when the ACC lurched around wildly

But the risk of relying on the ACC was made crystal-clear earlier this year, solar groups say, when the CPUC conducted its annual revision of ACC values.

While the details are highly complicated, the outcome of the revision process was a significant reduction in the value of net-metered solar as well as all other forms of distributed energy, from behind-the-meter batteries to energy-efficiency improvements.

First, a bit of background. In 2020, the CPUC made some major revisions to the ACC designed to better adapt it to California’s accelerated carbon-reduction policies set in place by SB 100, the 2018 bill that put the state on a 2045 deadline for a zero-carbon economy.

Among the big changes was an acknowledgment that California would no longer be relying on building new natural-gas-fired power plants to provide electricity in the years to come. Instead, its long-range energy plan now calls for gigawatts of lithium-ion batteries capable of storing solar power to serve evening demand peaks.

That means that when rooftop solar replaces grid power, it will likely not be replacing dirty gas power but rather clean power that’s been stored and then released by utility-scale batteries. The computer models that provide the underlying data for the ACC were redone with batteries as the marginal” resource that could be replaced by distributed energy over the coming decades.

Still, that 2020 revision ended up shifting the ACC’s values in ways that largely boosted the capacity and greenhouse-gas-reduction values that distributed energy could provide, a change that was largely lauded by solar groups at the time.

But it was also the first time the CPUC and E3, the consultant doing the modeling, had revised the ACC using the new paradigm set by SB 100 — and in the months after it was formalized in the summer of 2020, the CPUC reported finding significant flaws in its own methodology.

What’s more, in the time between the 2020 ACC revision and the 2021 revision, the CPUC ordered a massive procurement of renewable energy, batteries and other resources to meet its medium-term energy needs. Plugging that projected growth in clean utility-scale resources into the models also reduced the projected future value of distributed energy; essentially the new resources shrunk the gap between the present energy system and the expected future energy system that distributed energy could have otherwise helped fill.

This combination of factors, among others, led to the 2021 ACC revision essentially clawing back many of the increased values from the 2020 revision. 

To understand the changes, let’s return to this graph, which shows the 2019 ACC hourly values and the bell curve of daily solar production.

(E3)

Compare it to the three graphs below, illustrating the average hourly ACC values for the region surrounding the state capital, Sacramento, for 2019, 2020 and 2021.

(E3)

You can see how the 2020 change increased ACC values across many hours of the day, and then how the 2021 changes reduced them — quite drastically in some cases.

For rooftop solar, the impact was particularly harsh. According to testimony Beach gave to the CPUC, the change between 2020 and 2021 amounted to a 74 percent reduction in the utility bill savings from a typical rooftop solar system, from 11.7 cents to 3.1 cents per kilowatt-hour — enough to erase any likelihood of a rooftop solar system paying off its purchase price over its lifetime.

This kind of year-over-year change might not have garnered such scrutiny if it hadn’t coincided with the CPUC’s net-metering proceeding and its decision to link the future value of exported solar to the ACC. But that link led to an outcry from pro-net-metering groups, highlighting the risks that linking net-metered solar revenue to such changeable valuations could pose to the market.

That’s not to say that the changes from 2020 to 2021 were wrong, NRDC’s Chhabra said. He highlighted analysis that indicates the changes helped correct an overestimation of distributed energy’s value made in the 2020 ACC revision.

It behooves us to have the most accurate updated cost,” he said. The CPUC made an error and clawed it back really quickly.”

Still, the outcry over the 2021 revision points to a rocky road ahead for the CPUC’s next ACC revision in 2022. The annual ACC updates, whether major or minor, may become contentious battles over modeling, particularly if ACC avoided costs are used directly to compensate NEM 3.0 customers,” Beach stated in his testimony.

Tying net-metered solar export values to annual changes in the ACC could also lead to legal challenges to the underlying long-term energy planning proceedings at the CPUC that yield the values that underpin those calculations, he warned. Advocates may have little choice but to continuously litigate…modeling issues that directly impact [distributed energy resource] valuation and compensation through the ACC,” Beach wrote — something that could destabilize other parts of California’s energy policy in addition to its rooftop solar industry.

The bottom (power)line

California’s net-metering policy has become the most explosive issue for the state’s clean energy future, one that will be bitterly fought over until the CPUC makes a final decision on policy reforms early next year, and then bitterly fought over some more.

Bringing the ACC into this morass, with its esoteric metrics derived from complex modeling exercises and utility- or regulator-devised formulas, will only add to the confusion and conflict. There’s a clear allure to drilling down to find the most accurate way to assess what a kilowatt-hour of rooftop solar is worth when it’s exported to the grid. In a perfect world, using these values would create an optimally organized system, with customers, solar companies, utilities and the state as a whole getting the biggest bang for the buck out of each unit of solar deployed.

But we don’t live in a perfect world — we live in a world that faces a tight deadline to deploy as much solar energy as possible. Weighing the risks of overcompensating solar owners for investing in clean energy generation capacity on one hand, and stifling that needed investment on the other hand, it’s hard to see the benefits of choosing the latter over the former. Regulators may have, with the best of intentions, positioned a more perfect method of valuing rooftop solar as the enemy of good solar policy.

Jeff St. John is director of news and special projects at Canary Media. He covers innovative grid technologies, rooftop solar and batteries, clean hydrogen, EV charging and more.