The Faulty Diablo Canyon Study that Started it All

How Friends of the Earth and a Prominent Renewable Energy Lobbyist Hoodwinked California Policy-Makers.

The Faulty Diablo Canyon Study that Started it All

Since PG&E negotiated an agreement with anti-nuclear environmental groups in 2016 to shutter the Diablo Canyon Nuclear Power Plant, proponents of closing the plant have consistently asserted a preposterous claim, one that until recently has been broadly accepted by California legislators and regulators with little dissent. They claim that it will be cheaper to close the Diablo Canyon plant, which has already been paid for by ratepayers and provides 10 percent of California’s electricity, and replace it with new sources of renewable energy and energy efficiency investments, instead of continuing to operate it.

The claim is based upon a single study, commissioned by Friends of the Earth, an environmental nonprofit foundationally opposed to the power plant’s continued operation, and conducted by a well-known renewable energy lobbyist, V. John White of the Center for Energy Efficiency and Renewable Technologies (CEERT). With the California Legislature currently debating Governor Gavin Newsom’s proposal to keep Diablo open, a reckoning of how CEERT concluded that it was cheaper to close Diablo, how its flawed analysis came to be accepted by PG&E, the California Public Utility Commission, and the California Legislature, and the costs that CEERT’s deeply flawed analysis has imposed upon California ratepayers and taxpayers is long overdue.

How Did CEERT Skew the Numbers?

Let us count the ways.

1. Inflated Cost Categories

CEERT inflated cost categories associated with Diablo operations well above costs anticipated in PG&E’s planned relicensing efforts and then inflated them further, well above the general rate of inflation. These included costs such as refueling operations and maintenance (O&M), non-refueling O&M, medical benefits to employees, and property insurance, in some cases escalated by as much as triple the general rate of inflation. The O&M costs were not only well above the general rate of inflation but also consistently above the cost estimates of PG&E in a data request routinely cited by CEERT.

For example, CEERT inflates refueling O&M at 4.28 percent, almost double CEERT’s general inflation rate, citing the PG&E data. In actuality, the PG&E cost estimates escalated at an average of 2.36 percent per year. This unsupported assumption alone inflated CEERTs estimated cost of Diablo’s continued operation by $790 million above PG&E’s estimates.

In all, our reanalysis of CEERT’s inflation of Diablo operating costs above the general rate of inflation suggest that these manipulations inflated the cost of continuing operations of Diablo in CEERT’s base case scenarios by almost 25%, or $4.1 billion in real 2024 dollars.

Figure 1: Annual impact of added inflation of future Diablo Canyon Power Plant operating costs above the general inflation rate, as assumed in the CEERT study.

2. Outsized Fuel Costs

CEERT assumed an initial cost for Diablo’s fuel supply that was above historical fuel costs for both Diablo Canyon and the US nuclear industry writ large. Since the study’s publication, nuclear fuel costs for DCPP and the industry overall have remained static in nominal terms, if not declined, consistent with historic costs and trends. CEERT’s estimated initial nuclear fuel cost of $8.35/MWh in 2014 was $1.56/MWh above the actual Diablo Canyon Power Plant (DCPP) fuel cost in that year, which is a 23 percent addition, as seen in Figure 2.

Figure 2: Comparison of historical and current nuclear industry fuel costs per unit of electricity generation, versus assumed future fuel costs as projected by CEERT, Stanford/MIT, and the US Energy Information Administration’s 2022 Annual Energy Outlook. Inspiration for this figure came from an analysis by Parsons, one of the Stanford/MIT authors.

3. Capital Costs

CEERT projected future annual capital costs for DCPP operations based, in part, on one time capital expenditures, such as new steam generators and reactor pressure vessel heads, made by PG&E ahead of its anticipated license renewal to keep Diablo Canyon open. These sunk capital investments, which will be stranded should Diablo close as currently planned, would have been amortized over the planned license extension period of twenty years. Instead, CEERT replicated them as annual capital expenditures.

4. Misleading LCOE Metric

In order to further the impression that the cost of operating Diablo were much higher than alternatives, CEERT used a highly misleading nominal LCOE metric, in addition to representing Diablo costs in real dollars, to assess its estimate of base case costs associated with continued operation. When used to compare long-lived energy generation infrastructure with shorter lived sources, nominal LCOE significantly overstates the relative cost of the former.

The use of nominal LCOE increased the cost of continued Diablo operations by approximately 18 percent, from roughly $83/MWh to about $87/MWh (in 2024$) for the low base case and from just over $97/MWh to roughly $103/MWh for the low base case in nominal terms. CEERT’s nominal and real LCOE are shown in comparison to estimates from a recent Stanford/MIT study and by Roth and Jaramillo.

Figure 3: Comparison of estimates of the total per-unit cost of electricity from Diablo Canyon Power Plant, as projected by CEERT in real and nominal dollars, Roth and Jaramillo 2017, and the Stanford/MIT analysis. Figures in real dollars were inflated to 2024 dollars assuming an inflation of 1.174 percent through 2020 and 2 percent after, as is done in the CEERT study.

5. Sensitivity Analysis

CEERT claimed to conduct a sensitivity analysis on its Diablo cost projections but in fact did no such thing. CEERT’s sensitivity analysis was actually a series of further inflations to its high and low base case analyses, assuming that the general rate of inflation was 0.5 percent higher than its already inflated base case assumptions, assuming that the capacity factor of Diablo’s two reactors would decline by 0.5 percent per year in contravention of well established nuclear industry trends, and assuming that it's already very high assumed capital cost would be higher still. CEERT explored no cases in which any of its assumed costs would be lower than those used in its base case scenarios, nor did it conduct an actual sensitivity analysis of its own methodology, which almost certainly would have demonstrated that CEERT’s conclusions were highly sensitive to its inflated cost assumptions.

6. Heroic Assumptions

CEERT made heroic assumptions in its alternative scenarios, assuming that in the ten years between the publication of its study and the closure of DCPP, California would vastly increase its current pumped storage capacity. In the years since, California has not done so, a development that was easily anticipated. In the same year that CEERT released its Diablo study, the state energy commission released an assessment of the potential for new pumped storage in the state, concluding that there was little prospect of adding more.

CEERT also assumed that PG&E would save an additional 20 TWh per year through new energy efficiency investments, misciting a California Energy Commission study that actually identified less than 5 TWh of additional annual savings in its mid-high energy efficiency scenario, the efficiency scenario that CEERT claimed had identified an additional 20 TWh in savings above its mid-mid scenario. In actuality, since 2006, all of California’s public owned utilities have spent $2.1 billion on energy efficiency programs and only managed to achieve 7.5 TWh in annual energy savings.

Taken together, these assumptions resulted in a wildly misleading picture of the relative costs of closing Diablo versus continued operations, arbitrarily inflating cost categories with no precedent before and since, transforming one-time capital costs into annual capital expenditures, and then representing Diablo generation costs with nominal LCOE that inflated its base case scenarios well beyond the already high real dollar LCOE costs.

The resulting analysis projected future Diablo operating costs at roughly double those projected by independent analyses conducted by Stanford/MIT and Roth and Jaramillo. By contrast, present operations of Diablo cost a bit over $40/mwh. CEERT pushed every input into their analysis that they conceivably could in order to claim that over the next two decades it would rise to over $140/MWh.

How Did The FOE/CEERT Analysis Come To Be Adopted By the California PUC and State Legislature?

By its own acknowledgment, Friends of the Earth was founded in 1969 with the explicit objective of stopping the construction of the Diablo Nuclear Power Plant. That effort failed but the dream of closing the plant never waned. After forcing Southern California Edison to close the San Onofre Nuclear Power Plant in 2012, FOE set its sights on closing Diablo Canyon.

Diablo’s two reactors would need to be relicensed in 2025 and a 2010 State Water Board decision that mandated that PG&E alter the plant’s once-through cooling system threatened to substantially increase the cost of relicensing. But unlike San Onofre, which had a troubled history, Diablo is profitable, with an exemplary safety record. It provides 10 percent of California’s electricity and 17 percent of all clean electricity in the state. Replacing it would be costly and the closure of San Onofre had demonstrated that the clean electricity produced by the state’s nuclear plants was likely to be replaced with fossil energy. While the State Water Commission had mandated that Diablo abandon its once-through cooling system, it had already been forced to grant annual waivers to four very dirty natural gas plants that also utilized once-through cooling in order to maintain grid stability in Southern California in the wake of the San Onofre closure.

Enter V. John White and CEERT. White was the dean of renewable energy lobbyists in Sacramento, a well connected Democratic and environmental lobbyist with a long list of renewable energy clients and decades-long relationships with key Democratic power brokers in the state legislature. White had been deeply involved in the closure of the Rancho Seco Nuclear Power Plant near Sacramento and along with his long-time collaborator, Ralph Cavanagh of the Natural Resources Defense Council, had been the architect of the state’s renewable energy policies.

In 2014, Friends of the Earth retained CEERT, White’s non-profit consulting firm, to model the costs of closing Diablo Canyon. As noted above, White and his team at CEERT duly delivered the analysis that FOE was looking for, purporting to show that PG&E could shutdown Diablo, replace it with new renewable energy generation, energy efficiency investments, and pumped electricity storage, without increasing carbon emissions and at less cost than continuing to operate Diablo.

Armed with the CEERT study, FOE set about convincing PG&E and other skeptical partners who would become parties to the 2016 agreement to shutter Diablo that all could benefit from such a decision—most especially, PG&E, which had long been publicly on the record that Diablo was critical to grid stability, maintaining affordable electricity service, and achieving the state’s climate objectives.

PG&E’s Vice President for policy and federal affairs told the Utility Dive that CEERT’s study and subsequent meetings with representatives of Friends of the Earth “were useful in helping them reach the decision to not pursue a license extension.” That decision, however, appears to have been far more determined by internal calculations of the regulatory risks of continuing to operate the plant than it was by the FOE/CEERT analysis.

In addition to the State Water Board’s verdict, the state legislature had refused to value Diablo’s clean energy production under the renewable portfolio standard established in 2015, had put in place policies that encouraged community aggregation, which was eroding PG&E’s rate base, and was heavily subsidizing rooftop solar generation, which was eroding the value of Diablo’s baseload generation. This, together with then cheap natural gas generation, which could serve to balance wind and solar generation but was entirely inconsistent with the state’s climate commitments, made relicensing Diablo Canyon a risky bet for PG&E.

In short, there was little real basis, at the time or since, to think that closure of DCPP wouldn’t cost ratepayers and increase emissions. But the negotiated agreement allowed PG&E to recover all of its stranded capital and relicensing costs, avoid continuing costly and time consuming haggling with the State Water Board and the environmental community over relicensing, and reap a bonanza of worker retention funding provided by the CPUC and the state legislature. The FOE/CEERT study simply offered PG&E, along with the other parties to the negotiated settlement, the ability to claim that doing so would not increase costs to ratepayers or undermine California’s climate commitments.

Having cut the deal to close the plant, PG&E then completely reversed its public position on the costs and emissions implications of closing Diablo. PG&E hired a second consulting firm, M.J. Bradley & Associates (MJB&A) to reproduce the CEERT findings and justify the joint proposal to close Diablo. The resulting report, unsurprisingly given PG&E’s turnabout, concluded that DCPP was not needed and should be retired. Like the CEERT study, the MJB&A relied upon a range of assumptions about continuing Diablo operations that were unsupported by either historical data or subsequent operating experience. The MJB&A report, like the CEERT study, also made unrealistic assumptions about the availability and capability of energy efficiency programs, distributed energy resources like rooftop solar, and new energy storage capabilities to provide flexible generation to balance growing shares of variable renewable energy generation on the grid. In reality, the primary flexible resource across PG&E’s grid that balances variable generation is, and, as multiple analysis and planning documents from the state PUC suggest, is likely to remain natural gas powered generation.

Now, ostensibly replicated by PG&E’s consultants, the CEERT study became the technical basis for the joint agreement and for the CPUC’s approval of PG&E’s DCPP closure plan. In 2017 testimony to the CPUC, James Caldwell, Jr., one of the CEERT report authors presented findings from the CEERT report that the projected levelized cost of electricity for DCPP would be $97/MWh and $102/MWh in nominal dollars for the low and high base case scenarios, respectively, compared with present costs of $42/mwh and projected costs by independent analysts of around $50/mwh.

CEERT subsequently entered the report in the CPUC rulemaking as official testimony. In 2018, the California Public Utilities Commission (CPUC) formally approved the plan to shutter DCPP when its initial licenses ran out in 2024 and 2025, and PG&E withdrew its application to the NRC for license renewal.

Following the CPUC approval of the agreement, the California legislature then passed SB 1090, nominally requiring the state to “avoid any increase in emissions of greenhouse gases as a result of the retirement of the Diablo Canyon nuclear power plant,” appropriating $85 million in state funding to make up for the shortfall in funding for local services that would result from the closure of Diablo, and approving an additional $430 million in ratepayer charges for employee retention and community impact mitigation programs.

Who will pay for Friends of the Earth and CEERT’s Modeled Falsehoods?

Opponents of Governor Gavin Newsom’s proposal to extend operation of Diablo for another decade have zeroed in on a $1.4 billion loan that the Governor proposes to provide to PG&E to restart relicensing and preparations for continued operation, claiming that the state would be better off spending that money on more renewable energy, energy storage, and energy efficiency. But the proposed investments in those alternatives could dwarf the loan proposed to keep Diablo open and require direct subsidies from California taxpayers to alternative energy developers, whereas the loan for Diablo will be repaid assuming that the federal government provides funds recently appropriated in the Bipartisan Infrastructure Law for the purpose of keeping existing nuclear power plants open.

By contrast, Diablo opponents propose to spend $1.4 billion, including $450 million on zero-carbon generation and storage subsidies, $400 million in subsidies for new transmission lines, and $100 million for demand response programs. Potential additional spending could reach $3.7 billion including $380 million in subsidies for additional battery storage, $45 million to accelerate development of new offshore wind generation, and more. Whether or not those investments are wise, it should be clear at this point that additional investments in renewable energy generation, energy storage, energy efficiency, and new transmission lines are far better directed towards displacing natural gas, which still produces 38% of the state’s electricity than replacing DCPP, California’s largest source of clean power.

In the event that opponents succeed in thwarting the governor’s Diablo proposal, it is almost certain that the costs will fall heavily on California ratepayers. Over the six years since the negotiated settlement to replace Diablo was unveiled in 2016, claims about how Diablo’s firm low carbon electricity generation would be replaced have continually shifted. In the original CEERT study, the plan was to predominantly replace it with pumped storage. That source of firm capability has never materialized because it was clear from virtually the start that it didn’t actually exist and what effort has been made to develop new pumped storage in the state has predictably been opposed by the state’s environmental community.

From the moment that the PUC approved the negotiated settlement, the PUC’s mid-term planning documents have focused instead on replacing Diablo’s power with geothermal energy and battery storage. But neither has materialized at anything approaching the needed scale. The PUC itself does not anticipate major new geothermal development in the state to come on line until 2028 or later.

Now, many Diablo opponents propose to replace the plant through massive development of offshore wind along California’s central coast. Offshore wind typically has higher capacity factors than onshore. But it is still variable and cannot provide the firm, 90 percent capacity factor generation that Diablo presently provides. Under the best case, offshore wind development in California will not begin to come online until later this decade. That assumes, of course, that the state can expedite permitting and construction, which would require substantial changes or exemptions from the California Environmental Quality Act, the California Coastal Act, and a range of other laws that neither the state’s environmental community nor its Democratic leadership has thus far been willing to consider.

Absent legislation to extend Diablo operations, the likely outcome is continuing dependence on natural gas power plants to firm variable renewable generation, at a time that gas prices have hit record levels, thereby increasing grid instability, raising electricity prices, and limiting progress toward the state’s climate goals.

The alternative has always been obvious, despite the false claims of Friend of the Earth, CEERT, and the state environmental community. Keep Diablo open until such time as the state has demonstrated that it can do without its continuing heavy reliance on natural gas to keep the lights on. Neither the state PUC nor Diablo opponents have ever demonstrated that this outcome is possible. Until they do, California ratepayers, and its environment, will be best served by keeping Diablo open and ignoring the imaginary solutions that Diablo opponents continually promote to suggest that the plant can be closed without substantial economic and environmental costs.