How Two Simple Fixes Reduce the Cost Shift By Two-Thirds

California is the leading state in the nation in terms of solar capacity, with about 44 GW installed of which about 19 GW is rooftop solar. The State aims to completely decarbonize the electrical grid by 2045 but is currently not on track to meet key climate goals. Given the successful track record, sunny weather, and ambitious climate goals, one would think that California would welcome any and all new solar installations.

However, many utilities see distributed solar less as a climate and grid benefit and more as a threat to their profits. In 2013, Edison Electric Institute released a report that characterized distributed solar as a primary disruptive threat to the traditional electric utility business model. In response, utilities mounted coordinated policy, regulatory, and public relations efforts to curb distributed solar installations. Thus, the "cost shift" argument—a narrative claiming that solar ownership is a burden to other ratepayers—was born. In California, any solar policy discussion is quickly dominated and stalled by the Public Advocates Office (PAO)'s estimated financial impact of $8.5 billion from solar Net Energy Metering (NEM) customers. But this widely cited number rests on a fundamentally flawed methodology: it treats electricity that solar owners legally generate and consume on-site as a cost to utilities. Correct that error and apply less biased inputs, and the figure falls dramatically from $8.5 billion to $2.5 billion — and is likely lower.

In 2024 the PAO estimated that NEM customers cost-shifted $8.5 billion. The California Public Utilities Commission (CPUC) later adjusted that number to $7 billion due in part to corrected capacity factor modeling after feedback from the California Energy Commission (CEC); PAO used a capacity factor of 20%, although the California average is several percentage points lower. Severin Borenstein at the Energy Institute at Haas, UC Berkeley estimated a 2024 residential cost shift of $3.8 billion. Countering that, the environmental consulting firm M.Cubed found a $1.5 billion net benefit caused by distributed solar that same year.

Solar cost shift numbers get widely cited, but it's important to understand how cost shift is defined and estimated before accepting a number at face value. According to PAO's 2024 cost shift fact sheet, the formula they use is:

In this context, Bill Savings represent what the PAO characterizes as the "total amount that NEM 1.0, 2.0, and NBT customers are compensated" (NEM stands for Net Energy Metering and NBT for Net Billing Tariff). The Avoided Cost component is determined using the CPUC's Avoided Cost Calculator (ACC), a complex modeling framework developed by the consulting firm E3, which creates a potential conflict of interest, as E3 also serves major utilities. Critics argue that the ACC is structurally biased toward utility-scale planning while discounting the localized benefits of distributed generation.

The Fundamental Flaw: Redefining the Cost Shift

Not only does Bill Savings not accurately reflect solar customer compensation—it positions a customer's reduced electricity bill as a cost to the utilities. We never treat other forms of reduced demand such as energy efficiency improvements (like switching to LED bulbs or adding insulation) or simple conservation measures (like using the dryer or air conditioning less) as shifting costs to non-participating ratepayers. In a free market, a company does not count reduced sales as a business expense. Rooftop solar is a form of demand reduction, and treating it differently imposes an artificial cost. While utilities are accustomed to guaranteed profits, even monopolies selling an essential good cannot legally compel their customers to purchase a minimum amount of their product.

The cost shift calculations made by PAO, CPUC, and Borenstein makes the egregious assumption that all generation, including self-consumption, counts as costs to the grid. This erroneous inclusion treats electricity produced and immediately consumed on-site—energy that never touches the grid—as a direct loss of utility revenue. Borenstein justifies it by claiming that "so long as a solar system is connected to the grid, there is no real distinction between self-consumption and grid supply," a claim that defies the physical reality of on-site consumption. The inclusion of self-consumption by CPUC is questionable given their long history of utility-favorable decisions.

This raises the fundamental question: if a rooftop solar owner doesn't own the electrons generated by the system they purchased, what does ownership even mean?

The Legislature's NEM statute explicitly classifies rooftop solar owners as "eligible customer-generators," framing the entire policy around their role as independent producers. By characterizing on-site usage as a financial drain on utilities, the current methodology ignores the fact that owners of rooftop arrays are established as the rightful producers of their own power.

Given the legal reality that solar owners own their generated power and the economic fact that reduced sales is not a business expense, a fair reformulation of the cost shift must reject bill savings as a metric. Instead, we should replace it with export generation cost, which quantifies what utilities pay for the excess power exported by these customer-generators. The cost shift is thus defined as how much more rooftop solar generation costs utilities compared to other electricity generation purchases, where the avoided cost should take into account all costs avoided due to the addition of rooftop solar.

If we change nothing in the PAO calculations except for excluding self-generation from the cost shift, the figure drops from $8.5 billion to $3.8 billion for residential and commercial solar, similar to Borenstein's estimate for residential solar only. Avoided cost is calculated for export generation only. Under the corrected methodology, current NBT customers show zero cost shift since they are credited at the avoided cost rate. That's the first correction. The choice of inputs is a second problem entirely.

Correction 2: Questionable Valuation Inputs and Avoided Costs

Even after the CPUC adjusted the estimate from $8.5 billion to $7 billion—in part by correcting the inflated capacity factor—the underlying valuation inputs remain highly questionable. For instance, the CPUC estimates residential NEM customer compensation at approximately $0.45/kWh, significantly above the EIA-reported average residential retail price of $0.32/kWh in 2024. Additionally, there is no justification for PAO's differing residential export rates (50% for Pacific Gas & Electric (PG&E)/Southern California Edison (SCE) vs. 60% for San Diego Gas & Electric (SDG&E)).

To avoid another raging debate over minute details while addressing the manipulation of inputs, we swap some of PAO's chosen parameters for those used by cost shift proponent Borenstein because they are less extreme than PAO's original inputs. Capacity factor, residential compensation rates, avoided cost, and the non-bypassable charge rate were all replaced with Borenstein's corresponding inputs. This change, coupled with the self-consumption correction, further reduces the cost shift to $2.5 billion.

Can More Corrections Be Made?

This analysis demonstrates how two common-sense corrections reduce the widely cited solar cost shift by well over two-thirds, beginning to dismantle a persistent utility narrative. However, even this adjusted figure likely overstates the financial impact. Our estimate excludes critical mitigating factors such as avoided generation capacity, historical grid infrastructure savings, reduced line losses, decreased California Alternate Rates for Energy (CARE) subsidies for solar adopters, and the broader benefits of resilience, reduced emissions, and increased public health. Furthermore, NBT customers receive only ~$0.08/kWh in credit for excess energy that utilities immediately resell to neighbors for ~$0.30/kWh which raises the question if they're not a net benefit to the grid. A Berkeley Lab review of the value of solar studies found that among four states included, only one—the only utility-commissioned study—quantified a large negative impact on ratepayers. Not surprisingly, the utility-commissioned study was the only one that included lost utility revenue as a cost.

While solar users should contribute to the grid costs they incur, experts such as M.Cubed argue that when additional factors are included, rooftop solar provides a net benefit to all ratepayers. Below is an overview of assumptions made in the different cost shift scenarios mentioned.

Table 1. Comparison of several key assumptions used in the different 2024 cost shift estimation methods.

Putting The Numbers in Perspective

The corrected solar cost shift of $2.5 billion is still a lot of money, but bear in mind that it's calculated using methodologies and inputs from cost shift proponents — all we did was correct the flaw in the equation that disregards ownership and updated inputs where applicable. It's likely much lower since we have not accounted for all of solar's myriad benefits.

Still, it is dwarfed by forces actually driving California's electricity affordability crisis. California's three major Investor Owned Utilities (IOUs) spent $9 billion on wildfire-related costs in 2024 alone — more than three times the corrected cost shift. PG&E and SCE made record profits in 2024; together with SDG&E their combined profits were $5.4 billion, and that number went up even further the next year to $8.7 billion in 2025 for the three IOUs according to FERC Form 1 filings. Figure 1 below compares the different 2024 cost shift estimates in light purple and puts them in relation to 2024 wildfire spending and 2024 and 2025 net profits totaled for PG&E, SCE and SDG&E in dark purple.

The Bottom Line

SolarWAVE Action's $2.5 billion cost-shift estimate remains conservative; the true figure is likely much lower as our calculations exclude vital benefits like historical grid savings from reduced infrastructure upgrades. By applying two common-sense corrections—acknowledging rightful solar ownership and aligning inputs—we identify $6 billion in phantom costs currently attributed to solar adopters. Future analysis will further evaluate the additional mitigating factors that the PAO's methodology consistently overlooks.

Utilities have long argued that because grid infrastructure is a fixed cost, NEM customers rely on the grid without paying their fair share via volumetric rates. However, this conflates a rate-design issue with the "cost shift" narrative. Even before California adopted fixed charges to recover grid infrastructure costs, using the former to justify the latter was fundamentally flawed. Treating grid cost recovery and solar self-generation as a single problem allows utilities to effectively double-dip: by advocating for fixed-charge mechanisms to recover grid costs while simultaneously citing inflated "cost shift" figures to penalize solar adoption.

Ultimately, the "cost shift" narrative functions more as a distraction from the true drivers of California's electricity affordability crisis than as a legitimate economic metric. To build a resilient and affordable energy future, we must move beyond the utility-driven status quo and embrace an accounting reality that recognizes the legitimate value of distributed energy.