Skip to content

California’s Billion Dollar Energy Bill Question

A looming problem and a proposal.

California is staging a COVID-comeback. Vaccination rates are on the rise. Public schools are slowly opening up.  My favorite little league baseball team is back on the field. 

Go Pirates!

I can see light at the end of my pandemic tunnel. But I’m going to make it through this with my family, my job, and my health intact (sanity not so much). Others are not as fortunate.

Unpaid energy bills are one measure of the economic pressure that some households are under.  Since the pandemic took hold, these bills have been piling up. The graph below tracks arrears – wonkspeak for unpaid electricity and gas bills – households owe to California’s largest utility PG&E. 

The blue line tracks residential arrears as reported by PG&E. The red line plots the pre-pandemic average by month (using data from the previous two years). Data are available here.

Across California’s three investor-owned-utilities, residential arrears now exceed $1 billion. That’s up from $412 million pre-pandemic. Dividing these total arrears by the number of accounts in arrears works out to over $500/customer.

Households that have lost jobs or worse will struggle to pay off this debt. I understand that unpaid utility bills are a symptom of deeper problems that energy bill assistance cannot fix. But if we want to target relief to households that need it, utility debt forgiveness seems like a shovel-ready assist.

The catch, of course, is that these overdue bills need to be paid by someone. Shifting these costs elsewhere is particularly fraught in California where we typically pay for energy assistance programs by pushing high retail energy prices even higher. We need a relief mechanism for customers who have not been able to keep up with their bills during the pandemic. We also need a better way to pay for it.

The Fraught Economics of Energy Assistance

In California, discussions about post-pandemic energy debt assistance are playing out against the backdrop of energy assistance programs we already have in place. The most important is CARE (California Alternate Rates for Energy) which offers discounted energy prices to qualifying households. The graph below shows that CARE customers get a significant discount on electricity prices. 

The graph plots discounted CARE electricity prices, standard residential rates, and our estimate of the social marginal cost for PG&E over time. Details can be found here.

The graph also shows that CARE prices, although discounted, are still really high! We’ve estimated that the CARE price is now double the social marginal cost of supplying electricity. In other words, low-income customers are paying twice what it marginally costs — including climate change impacts — to get electrons to their homes.

 We pay for this CARE discount (and other energy assistance programs) by raising the energy prices other utility customers pay. This is inefficient because it moves consumer prices even farther above the true incremental cost. This can also have unintended impacts on affordability for lower-income households who do not qualify for CARE but do struggle to pay energy bills. These data show that over half of the arrears that have accumulated over the pandemic are owed by non-CARE households. 

 I see social value in shifting energy debt burdens off of households hit hard by the pandemic. But I also see problems with paying for this energy assistance via higher energy prices. 

A Better Way to Pay?

 Are there better ways to pay for post-pandemic debt relief programs? One option would be to get in line for a piece of that $15 billion California budget surplus. But that line is long and the prospects uncertain.

 Here’s another idea. Every year, revenues from cap-and-trade permit sales are transferred to residential utility customers in lump-sum payments. The original purpose of these climate credits was to encourage public support for the cap and trade program and reduce adverse impacts on low-income households. But all customers of a utility get the same credit payment. My guess is that many don’t even see these climate credits on their bills. So why not target these transfer payments at folks who notice and need them?

A half-baked proposal: Households that are already participating in bill assistance programs such as CARE would continue to receive the climate credit by default. Non-CARE households could be defaulted out of the credit program – but could opt back in if they want to keep the credit.  

Source: iStock

How much revenue could this redirect into bill assistance? When utilities have changed program defaults in the past, this has had a huge impact on customer participation. Multiplying the number of non-CARE households by the average carbon credit value from recent years for PG&E, SDG&E, and SCE respectively yields approximately $630 million in revenues. That’s approximately equal to the increase in residential arrears owed to these three utilities since the pandemic took hold.

These calculations are rough. And there are important implementation details that would need to be worked out (e.g. how should this climate credit value be allocated across utilities? How can households who need the transfer easily opt back into receiving it?). But if the alternative is raising retail energy prices for everyone, these details are worth negotiating.

Targeting Benefits and Costs

As regulators scramble to respond to the accumulating mountain of energy debt, there has been an appropriate focus on how to target relief at households who really need it. But the success of an energy assistance program will also hinge on how we pay for it. In California, we have a habit of paying for energy assistance – among other things –  with higher energy prices. Here’s an opportunity to break that habit. As we crawl out from under this pandemic, we should be looking for ways to pull costs out of California’s energy prices, versus pushing more costs in.  

Keep up with Energy Institute blogs, research, and events on Twitter @energyathaas.

Suggested citation: Fowlie, Meredith. “California’s Billion Dollar Energy Bill Question” Energy Institute Blog, UC Berkeley, April 5, 2021,



7 thoughts on “California’s Billion Dollar Energy Bill Question Leave a comment

  1. This is an interesting exchange.

    Dr. Fowlie uses a marginal cost of less than ten cents/kWh, and of that, only about seven cents is related to the bulk “clean power” function (generation plus line losses and emissions). Three cents is for T&D. I strongly disagree with the methodology that E3 has used to develop these short-run marginal costs (more on that later), but I’ll use them for now.

    Carl Wurtz advocates retaining a resource that would cost about $100/MWh, and is an inflexible resource that requires a huge investment in storage (the Helms pumped storage plant) to make the power useful. Nuclear is a lot like wind and solar: a lot of the power comes at times when you really don’t want or need it. Ten cents is a lot more than seven cents. And, when new wind and solar are costing less than five cents, the advantage is obvious.

    One solution is advocated to divert the carbon credit from the general electric bill of all consumers to support low-income energy services. That turns the cap and trade program into a carbon tax for many people. That, of course, is a darling of most economists (including myself). The tax is used to help poor people. That, of course, is a darling of most liberals (including myself). But it does break faith with the legislative intent and discussion when AB32 was passed, and the extensive negotiations (of which I was a part) on implementation.

    One interesting aspect of this is that while the IOUs include the cost of carbon allowances in the price of electricity, and then rebate the funds in a lump sum, the consumer-owned utilities mostly rebate it directly, in the form of lower prices. The current system for the IOUs was adopted at the suggestion of economists, including some at Haas, to avoid distorting the price in a way that would encourage more consumption. Now that beneficial electrification is the rage (and I co-authored a series of papers on BE for space heat, water heat, and transportation), those same economists are concerned that we have driven up the per-kWh price of electricity, and need to undo their work. OK. I’ve made a few mistakes myself, and some things have turned out differently from what I expected. We should be ready to adapt and change course where necessary.

    Now, briefly back to the marginal cost issue. I’ll focus on one sliver of the calculation, transmission. California has added a lot of new transmission in the years since AB32 passed. I helped one COU decide to NOT join the CAISO, because the transmission charges (two cents/kWh, double the COU average transmission cost at the time) were so high that they would dwarf the dispatch cost savings (about 0.3 cents/kWh).

    Those transmission costs have doubled in the intervening years, approaching four cents/kWh. If adding transmission causes the average cost per kWh to go up, then (by definition) the marginal cost exceeds the average cost. If the average cost is (now) four cents, then the marginal cost must be more than that. But the reference Dr. Fowlie directs us to, the heavily-critiqued “Next10” report, shows marginal transmission costs of only about 0.3 cents/kWh. If indeed that were true, then the CAISO transmission access charge would have declined sharply over the past decade, and the COUs that stayed out of the CAISO would now be seeking to join. Simply stated, the Next10 estimate of transmission marginal costs does not pass the laugh test.

    Frankly, it’s quite the same for generation capacity marginal cost. As California is adding expensive batteries to supply peaking capacity, the IOU rates are going up. By definition, again, if adding capacity increases the average cost, then the marginal cost must exceed the average cost. But the Next10 study also shows generation capacity at a fraction of a cent per kWh. If that were true, the IOU rates would be going DOWN, not up, as they add this new capacity.

    Yes, California IOU rates are too high. There are many reasons. The CARE program costs are one. The lost revenue due to customer self-supply is one. The framework of NEM2, with near-retail rate credits for backfeed is probably obsolete.

    But the CPUC allowing some of the highest rates of return on equity, some of the highest equity capitalization ratios, and some of the highest executive compensation in the nation is definitely also a part of it, and that part is within the control of the CPUC.

    Additionally, while other states require utilities to absorb part of the cost of resources that prove uneconomic with technological evolution, California has allowed full recovery of these stranded costs. It did so with OII-2 cogen resources (for which I was a consultant to the CEC) in the 1980s, again with the transition charges allowed in the 90’s restructuring experiment, with the DWR contracts in the 2000’s, and now with the PCIA charges to the CCAs for uneconomic resources acquired since then. It’s a VERY generous regulatory jurisdiction.

    Why can utilities from Blaine, Washington to Tallahassee, Florida provide retail service, including generation, transmission, and distribution for an average of $0.13/kWh (the national average residential electricity price) and California utilities have trouble providing service at nearly twice that cost? Why can public power provide service at average prices 30% lower than the IOUs? Santa Clara, right in the heart of PG&E’s service territory, has rates 40% lower. The small IOUs in California (Pacific Power, Liberty) have rates have rates about half those of PG&E.

    2019 Average Residential Rates (EIA Form 851 data)

    SDGE: $0.2578
    PG&E: $0.2235
    SCE: $0.1621
    SMUD: $0.1491
    Santa Clara: $0.1293
    Pacific Power: $0.1287
    Liberty Utilities: $0.1342

    These averages all include the effect of the low-income discounts; CARE customers pay less than average, other customers pay more than average. But the COUs have low-income assistance programs as well.

    Moving away from volumetric recovery of all systems costs is a serious mistake. It is assertively anti-competitive. It is precisely why regulation was created: to prevent discriminatory pricing (originally of railroads and grain elevators discriminating between farmers with access to only one railroad, and those with access to competing lines).

    Regulation is supposed to emulate the effect of competition on pricing. The supermarket does not care if I grow some of my own tomatoes. If I want to buy from them, the price is the same regardless of whether I do or do not. I pay no fixed charge, demand charge, or grid access fee. I just walk in the door, buy what I want, and the store recovers the costs of production, transmission, distribution, overhead, profit and taxes in the $1.99/lb I am charged at the register. If I buy one tomato per year, I pay $1.99/lb. If I buy five pounds of tomatoes every week, I pay $1.99/lb (seasonally differentiated, to be sure). That’s how competition works. [And, to be fair, if I do grow my own tomatoes, and bring them my surplus tomatoes, they don’t pay me retail for them. NEM has no parallel under competition — it was an infant industry support program like railroad land grants and space contracts for semiconductors. If the industry is no longer an infant, it’s time to review the policy.]

    We have a failure of regulation, not a failure of rate design. The CPUC needs to go back and re-read its own Order in Market Street Railway. The Commission ruled, and the US Supreme Court affirmed, that a public service company has no right to recover costs that become uneconomic due to the forces of competition. They have no right to stranded cost recovery. That’s what equity holders are paid to do: take risks, and bear the consequences of mistaken decisions.

    • Jim, I’d like to clear up some confusion on what it costs to generate electricity at Diablo Canyon Power Plant. On PG&E’s Form 1 filing to the Federal Energy Regulatory Commission, downloadable here:

      Click to access FERCForm1.pdf

      the company is required to give a detailed accounting if how much it costs to generate each MWh of electricity at each of its steam-electric generating plants. Page 402 (PDF page 329), line 35 lists the total “Expenses per Net KWh” at Diablo Canyon Units 1 & 2 as $.0291, or $29.10/MWh. Not “over $100”, as another poster has claimed on this thread while padding it with mythical “compliance” costs. For comparison, other columns give the following costs for generating each MWh at PG&E’s natural gas plants:

      Colusa – $38.40
      Gateway – $38.60
      Humboldt – $71.40

      As you can see, Diablo Canyon generates electricity cheaper than any other source in PG&E’s fleet – 25% cheaper than any other dispatchable source, and 48% cheaper than any utility-scale renewable source.

      I’m not sure why you believe Diablo Canyon requires “a huge investment in storage (the Helms pumped storage plant) to make the power useful.” Every MWh of electricity generated at the plant provides valuable baseload electricity to the grid; Helms hasn’t been used for decades to store power from Diablo Canyon. On sunny or windy days, excess renewable generation is sometimes used to store water at Helms at the same time it’s being drained. Though wasteful, the technique helps CAISO engineers smooth some of the irregularities in electricity generated by California’s solar and wind farms.

      “Nuclear is a lot like wind and solar: a lot of the power comes at times when you really don’t want or need it.”

      Nuclear is nothing like wind or solar, and that’s a good thing. As I’ve already pointed out, every MWh, bar none, provides valuable baseload electricity to the grid. It’s never required curtailment, while solar and wind require curtailment of unwanted power (at added cost to customers) nearly every day of the year.

  2. “As we crawl out from under this pandemic, we should be looking for ways to pull costs out of California’s energy prices, versus pushing more costs in.”

    Agree, Meredith. Two of the most obvious ways for PG&E to reduce energy costs would be:

    1) California’s PUC should to approve the utility’s application to reduce the obscene subsidy paid to net-metered residential solar owners, and approve the $70 monthly distribution fee to which PG&E is entitled. As things stand, wealthy customers with solar-paneled homes are being cross-subsidized by customers who are least able to do so.

    2) Reconsidering its decision to close Diablo Canyon Power Plant in 2025 would free ratepayers of $900 million in added debt for decommissioning, and make $3 billion already paid refundable.

    There is no justifiable reason for closing Diablo Canyon – a state-of-the-art, impeccably-maintained nuclear power plant that generates 8% of California electricity with no carbon emissions. Capital costs were paid off by PG&E’s 16 million customers over a span of 35 years, for a power plant that was expected to last at least 80. That PG&E’s former boardmembers, seeking to profit from sales of natural gas, were permitted by a corrupt CPUC to prematurely plow customers’ investment into the ground is nothing short of criminal.

    • The issue of Diablo Canyon has already been settled as I’ve pointed out to you many times here. The cost of relicensing and complying with environmental regulations will drive the cost of Diablo Canyon to over $100/MWH. Your dismissal of those requirements are unsupported and irrelevant. Diablo Canyon power can be replaced with renewables at $55/MWH or less.

      You haven’t shown how PG&E shareholders would profit from natural sales from the retirement of Diablo Canyon. All of the information you need are in PG&E’s SEC and FERC Form 1 filings. Rather than speculate, provide real evidence. Note that shareholders would make an ongoing profit from Diablo Canyon through the annual capital additions at the plant related to refueling and maintenance. PG&E invests several hundred million dollar per year in cap adds according to PG&E’s FERC Form 1.

      As to solar panel “subsidies” PG&E’s own analysis shows that the most the NEM subsidy might be is an average of $25/month based on its distribution marginal cost workpapers filed in its 2020 GRC Phase II case. That ignores any of the benefits created by rooftop solar such as avoided transmission costs (I’ve calculated in PJM that it’s close to $40/MWH) and GHG reductions and avoided REC purchases.

  3. I like this idea. I’m pretty sure that the average customer doesn’t notice the ~$30 once a year rebate. It’s a different form of the carbon tax and rebate proposal.

    • The climate credit has historically been distributed twice a year. Tyler Hodge of the EIA discussed the effect of the distributions on AVG kWh costs a few years back- “A Today in Energy article from a few months ago discusses this effect on Pacific electricity prices (

      The value of the climate credit (how many kWh can one buy with the credit) has crashed over the years due to the changes in the rate structures. .

  4. even better would be to redirect the 630M$ of the cap n trade to improve the building shells and appliances in the low income households–and also to revamp the current low income energy efficiency programs so that they go beyond those actions and help develop community solar systems (more cost-effective than rooftop solar) in low income areas to reduce their bills (and society’s) going forward.

%d bloggers like this: