Author Archives: Richard McCann

About Richard McCann

Partner in M.Cubed, an economics and policy consulting firm.

A new agricultural electricity use forecast method holds promise for water use management

Agricultural electricity demand is highly sensitive to water availability. Under “normal” conditions, the State Water Project (SWP) and Central Valley Project (CVP), as well as other surface water supplies, are key sources of irrigation water for many California farmers. Under dry conditions, these water sources can be sharply curtailed, even eliminated, at the same time irrigation requirements are heightened. Farmers then must rely more heavily on groundwater, which requires greater energy to pump than surface water, since groundwater must be lifted from deeper depths.

Over extended droughts, like between 2012 to 2016, groundwater levels decline, and must be pumped from ever deeper depths, requiring even more energy to meet crops’ water needs. As a result, even as land is fallowed in response to water scarcity, significantly more energy is required to water remaining crops and livestock. Much less pumping is necessary in years with ample surface water supply, as rivers rise, soils become saturated, and aquifers recharge, raising groundwater levels.

The surface-groundwater dynamic results in significant variations in year-to-year agricultural electricity sales. Yet, PG&E has assigned the agricultural customer class a revenue responsibility based on the assumption that “normal” water conditions will prevail every year, without accounting for how inevitable variations from these circumstances will affect rates and revenues for agricultural and other customers.

This assumption results in an imbalance in revenue collection from the agricultural class that does not correct itself even over long time periods, harming agricultural customers most in drought years, when they can least afford it. Analysis presented presented by M.Cubed on behalf of the Agricultural Energy Consumers Association (AECA) in the 2017 PG&E General Rate Case (GRC) demonstrated that overcollections can be expected to exceed $170 million over two years of typical drought conditions, with the expected overcollection $34 million in a two year period. This collection imbalance also increases rate instability for other customer classes.

Figure-1 compares the difference between forecasted loads for agriculture and system-wide used to set rates in the annual ERRA Forecast proceedings (and in GRC Phase 2 every three years) and the actual recorded sales for 1995 to 2019. Notably, the single largest forecasting error for system-wide load was a sales overestimate of 4.5% in 2000 and a shortfall in 2019 of 3.7%, while agricultural mis-forecasts range from an under-forecast of 39.2% in the midst of an extended drought in 2013 to an over-forecast of 18.2% in one of the wettest years on record in 1998. Load volatility in the agricultural sector is extreme in comparison to other customer classes.

Figure-2 shows the cumulative error caused by inadequate treatment of agricultural load volatility over the last 25 years. An unbiased forecasting approach would reflect a cumulative error of zero over time. The error in PG&E’s system-wide forecast has largely balanced out, even though the utility’s load pattern has shifted from significant growth over the first 10 years to stagnation and even decline. PG&E apparently has been able to adapt its forecasting methods for other classes relatively well over time.

The accumulated error for agricultural sales forecasting tells a different story. Over a quarter century the cumulative error reached 182%, nearly twice the annual sales for the Agricultural class. This cumulative error has consequences for the relative share of revenue collected from agricultural customers compared to other customers, with growers significantly overpaying during the period.

Agricultural load forecasting can be revised to better address how variations in water supply availability drive agricultural load. Most importantly, the final forecast should be constructed from a weighted average of forecasted loads under normal, wet and dry conditions. The forecast of agricultural accounts also must be revamped to include these elements. In addition, the load forecast should include the influence of rates and a publicly available data source on agricultural income such as that provided by the USDA’s Economic Research Service.

The Forecast Model Can Use An Additional Drought Indicator and Forecasted Agricultural Rates to Improve Its Forecast Accuracy

The more direct relationship to determine agricultural class energy needs is between the allocation of surface water via state and federal water projects and the need to pump groundwater when adequate surface water is not available from the SWP and federal CVP. The SWP and CVP are critical to California agriculture because little precipitation falls during the state’s Mediterranean-climate summer and snow-melt runoff must be stored and delivered via aqueducts and canals. Surface water availability, therefore, is the primary determinant of agricultural energy use, while precipitation and related factors, such as drought, are secondary causes in that they are only partially responsible for surface water availability. Other factors such as state and federal fishery protections substantially restrict water availability and project pumping operations greatly limiting surface water deliveries to San Joaquin Valley farms.

We found that the Palmer Drought Stress Index (PDSI) is highly correlated with contract allocations for deliveries through the SWP and CVP, reaching 0.78 for both of them, as shown in Figure AECA-3. (Note that the correlation between the current and lagged PDSI is only 0.34, which indicates that both variables can be included in the regression model.) Of even greater interest and relevance to PG&E’s forecasting approach, the correlation with the previous year’s PDSI and project water deliveries is almost as strong, 0.56 for the SWP and 0.53 for the CVP. This relationship can be seen also in Figure-3, as the PDSI line appears to lead changes in the project water deliveries. This strong relationship with this lagged indicator is not surprising, as both the California Department of Water Resources and U.S. Bureau of Reclamation account for remaining storage and streamflow that is a function of soil moisture and aquifers in the Sierras.

Further, comparing the inverse of water delivery allocations, (i.e., the undelivered contract shares), to the annual agricultural sales, we can see how agricultural load has risen since 1995 as the contract allocations delivered have fallen (i.e., the undelivered amount has risen) as shown in Figure-4. The decline in the contract allocations is only partially related to the amount of precipitation and runoff available. In 2017, which was among the wettest years on record, SWP Contractors only received 85% of their allocations, while the SWP provided 100% every year from 1996 to 1999. The CVP has reached a 100% allocation only once since 2006, while it regularly delivered above 90% prior to 2000. Changes in contract allocations dictated by regulatory actions are clearly a strong driver in the growth of agricultural pumping loads but an ongoing drought appears to be key here. The combination of the forecasted PDSI and the lagged PDSI of the just concluded water year can be used to capture this relationship.

Finally, a “normal” water year rarely occurs, occurring in only 20% of the last 40 years. Over time, the best representation of both surface water availability and the electrical load dependent on it is a weighted average across the probabilities of different water year conditions.

Proposed Revised Agricultural Forecast

We prepared a new agricultural load forecast for 2021 implementing the changes recommended herein. In addition, the forecasted average agricultural rate was added, which was revealed to be statistically valid. The account forecast was developed using most of the same variables as for the sales forecast to reflect similarities in drivers of both sales and accounts.

Figure-5 compares the performance of AECA’s proposed model to PG&E’s model filed in its 2021 General Rate Case. The backcasted values from the AECA model have a correlation coefficient of 0.973 with recorded values,[1] while PG&E’s sales forecast methodology only has a correlation of 0.742.[2] Unlike PG&E’s model almost all of the parameter estimates are statistically valid at the 99% confidence interval, with only summer and fall rainfall being insignificant.[3]

AECA’s accounts forecast model reflects similar performance, with a correlation of 0.976. The backcast and recorded data are compared in Figure-6. For water managers, this chart shows how new groundwater wells are driven by a combination of factors such as water conditions and electricity prices.




Why are real-time electricity retail rates no longer important in California?

The California Public Utilities Commission (CPUC) has been looking at whether and how to apply real-time electricity prices in several utility rate applications. “Real time pricing” involves directly linking the bulk wholesale market price from an exchange such as the California Independent System Operator (CAISO) to the hourly retail price paid by customers. Other charges such as for distribution and public purpose programs are added to this cost to reach the full retail rate. In Texas, many retail customers have their rates tied directly or indirectly to the ERCOT system market that operates in a manner similar to CAISO’s. A number of economists have been pushing for this change as a key solution to managing California’s reliability issues. Unfortunately, the moment may have passed where this can have a meaningful impact.

In California, the bulk power market costs are less than 20% of the total residential rate. Even if we throw in the average capacity prices, it only reaches 25%. In addition, California has a few needle peaks a year compared to the much flatter, longer, more frequent near peak loads in the East due to the differences in humidity. The CAISO market can go years without real price deviations that are consequential on bills. For example, PG&E’s system average rate is almost 24 cents per kilowatt-hour (and residential is even higher). Yet, the average price in the CAISO market has remained at 3 to 4 cents per kilowatt-hour since 2001, and the cost of capacity has actually fallen to about 2 cents. Even a sustained period of high prices such as occurred last August will increase the average price by less than a penny–that’s less than 5% of the total rate. The story in 2005 was different, when this concept was first offered with an average rate of 13 cents per kilowatt-hour (and that was after the 4 cent adder from the energy crisis). In other words, the “variable” component just isn’t important enough to make a real difference.

Ahmad Faruqui who has been a long time advocate for dynamic retail pricing wrote in a LinkedIn comment:

“Airlines, hotels, car rentals, movie theaters, sporting events — all use time-varying rates. Even the simple parking meter has a TOU rate embedded in it.”

It’s true that these prices vary with time, and electricity prices are headed that way if not there already. Yet these industries don’t have prices that change instantly with changes in demand and resource availability–the prices are often set months ahead based on expectations of supply and demand, much as traditional electricity TOU rates are set already. Additionally, in all of these industries , the price variations are substantially less than 100%. But for electricity, when the dynamic price changes are important, they can be up to 1,000%. I doubt any of these industries would use pricing variations that large for practical reasons.

Rather than pointing out that this tool is available and some types of these being used elsewhere, we should be asking why the tool isn’t being used? What’s so different about electricity and are we making the right comparisons?

Instead, we might look at a different package to incorporate customer resources and load dynamism based on what has worked so far.

  • First is to have TOU pricing with predictable patterns. California largely already has this in place, and many customer groups have shown how they respond to this signal. In the Statewide Pilot on critical peak period price, the bulk of the load shifting occurred due to the implementation of a base TOU rate, and the CPP effect was relatively smaller.
  • Second, to enable more distributed energy resources (DER) is to have fixed price contracts akin to generation PPAs. Everyone understands the terms of the contracts then instead of the implicit arrangement of net energy metering (NEM) that is very unsatisfactory for everyone now. It also means that we have to get away from the mistaken belief that short-run prices or marginal costs represent “market value” for electricity assets.
  • Third for managing load we should have robust demand management/response programs that target the truly manageable loads, and we should compensate customers based on the full avoided costs created.

The State Water Board needs to act to start Flood MAR pilot projects

I recently presented to CDWR’s Lunch-MAR group the findings for a series of studies we conducted on the universe of benefits from floodwater managed aquifer recharge (MAR) and the related economic and financing issues. I also proposed that an important next step is to run a set of pilots to study the acceptability of on-farm floodwater recharge projects to growers, including how do they respond to incentives and program design, and what are the potential physical consequences.

The key to initiating these pilots is getting a clear declaration from the State Water Resources Control Board that excess floodwaters are surplus and available. Unfortunately, the Water Board has not provided sufficient clarification on how these projects can take “advantage of seasonal or occasional flood waters that overtop the banks of a stream and are then directed into a designated recharge area.” Instead, the Board’s website says that such diverted floodwaters cannot be stored for future beneficial use–which obviates the very purpose of retaining the floodwaters in the first place.

The Board should be at least issuing temporary use permits for floodwaters above certain designated levels as being available for pilot projects on the basis that non-use of those floodwaters constitute a surrender of that right for the year. Then those agencies interested in flood MAR can design projects to experiment with potential configurations.

Can Net Metering Reform Fix the Rooftop Solar Cost Shift?: A Response

A response to Severin Borenstein’s post at UC Energy Institute where he posits a large subsidy flowing to NEM customers and proposes an income-based fixed charge as the remedy. Borenstein made the same proposal at a later CPUC hearing.

The CPUC is now considering reforming the current net energy metering (NEM) tariffs in the NEM 3.0 proceeding. And the State Legislature is considering imposing a change by fiat in AB 1139.

First, to frame this discussion, economists are universally guilty of status quo bias in which we (since I’m one) too often assume that changing from the current physical and institutional arrangement is a “cost” in an implicit assumption that the current situation was somehow arrived at via a relatively benign economic process. (The debate over reparations for slavery revolve around this issue.) The same is true for those who claim that NEM customers are imposing exorbitant costs on other customers.

There are several issues to be considered in this analysis.

1) In looking at the history of the NEM rate, the emergence of a misalignment between retail rates that compensate solar customers and the true marginal costs of providing service (which are much more than the hourly wholesales price–more on that later) is a recent event. When NEM 1.0 was established residential rates were on the order of 15 c/kWh and renewable power contracts were being signed at 12 to 15 c/kWh. In addition, the transmission costs were adding 2 to 4 c/kWh. This was the case through 2015; NEM 1.0 expired in 2016. NEM 2.0 customers were put on TOU rates with evening peak loads, so their daytime output is being priced at off peak rates midday and they are paying higher on peak rates for usage. This despite the fact that the difference in “marginal costs” between peak and off wholesale costs are generally on the order of a penny per kWh. (PG&E NEM customers also pay a $10/month fixed charge that is close to the service connection cost.) Calculating the net financial flows is more complicated and deserve that complex look than what can be captured in a simple back of the envelope calculation.

2) If we’re going to dig into subsidies, the first place to start is with utility and power plant shareholders. If we use the current set of “market price benchmarks” (which are problematic as I’ll discuss), out of PG&E’s $5.2 billion annual generation costs, over $2 billion or 40% are “stranded costs” that are subsidies to shareholders for bad investments. In an efficient marketplace those shareholders would have to recover those costs through competitively set prices, as Jim Lazar of the Regulatory Assistance Project has pointed out. One might counter those long term contracts were signed on behalf of these customers who now must pay for them. Of course, overlooking whether those contracts were really properly evaluated, that’s also true for customers who have taken energy efficiency measures and Elon Musk as he moves to Texas–we aren’t discussing whether they also deserve a surcharge to cover these costs. But beyond this, on an equity basis, NEM 1.0 customers at least made investments based on an expectation, that the CPUC did not dissuade them of this belief (we have documentation of how at least one county government was mislead by PG&E on this issue in 2016). If IOUs are entitled to financial protection (and the CPUC has failed to enact the portfolio management incentive specified in AB57 in 2002) then so are those NEM customers. If on the other hand we can reopen cost recovery of those poor portfolio management decisions that have led to the incentive for retail customers to try to exit, THEN we can revisit those NEM investments. But until then, those NEM customers are no more subsidized than the shareholders.

3) What is the true “marginal cost”? First we have the problem of temporal consistency between generation vs. transmission and distribution grid (T&D) costs. Economists love looking at generation because there’s a hourly (or subhourly) “short run” price that coincides nicely with economic theory and calculus. On the other hand, those darn T&D costs are lumpy and discontinuous. The “hourly” cost for T&D is basically zero and the annual cost is not a whole lot better. The current methods debated in the General Rate Cases (GRC) relies on aggregating piecemeal investments without looking at changing costs as a whole. Probably the most appropriate metric for T&D is to calculate the incremental change in total costs by the number of new customers. Given how fast utility rates have been rising over the last decade I’m pretty sure that the “marginal cost” per customer is higher than the average cost–in fact by definition marginal costs must be higher. (And with static and falling loads, I’m not even sure how we calculated the marginal costs per kwh. We can derive the marginal cost this way FERC Form 1 data.) So how do we meld one marginal cost that might be on a 5-minute basis with one that is on a multi-year timeframe? This isn’t an easy answer and “rough justice” can cut either way on what’s the truly appropriate approximation.

4) Even if the generation cost is measured sub hourly, the current wholesale markets are poor reflections of those costs. Significant market distortions prevent fully reflecting those costs. Unit commitment costs are often subsidized through out of market payments; reliability regulation forces investment that pushes capacity costs out of the hourly market, added incremental resources–whether for added load such as electrification or to meet regulatory requirements–are largely zero-operating cost renewables of which none rely on hourly market revenues for financial solvency; in California generators face little or no bankruptcy risk which allows them to underprice their bids; on the flip side, capacity price adders such as ERCOT’s ORDC overprices the value of reliability to customers as a backdoor way to allow generators to recover investments through the hourly market. So what is the true marginal cost of generation? Pulling down CAISO prices doesn’t look like a good primary source of data.

We’re left with the question of what is the appropriate benchmark for measuring a “subsidy”? Should we also include the other subsidies that created the problem in the first place?

AB1139 would undermine California’s efforts on climate change

Assembly Bill 1139 is offered as a supposed solution to unaffordable electricity rates for Californians. Unfortunately, the bill would undermine the state’s efforts to reduce greenhouse gas emissions by crippling several key initiatives that rely on wider deployment of rooftop solar and other distributed energy resources.

  • It will make complying with the Title 24 building code requiring solar panel on new houses prohibitively expensive. The new code pushes new houses to net zero electricity usage. AB 1139 would create a conflict with existing state laws and regulations.
  • The state’s initiative to increase housing and improve affordability will be dealt a blow if new homeowners have to pay for panels that won’t save them money.
  • It will make transportation electrification and the Governor’s executive order aiming for 100% new EVs by 2035 much more expensive because it will make it much less economic to use EVs for grid charging and will reduce the amount of direct solar panel charging.
  • Rooftop solar was installed as a long-term resource based on a contractual commitment by the utilities to maintain pricing terms for at least the life of the panels. Undermining that investment will undermine the incentive for consumers to participate in any state-directed conservation program to reduce energy or water use.

If the State Legislature wants to reduce ratepayer costs by revising contractual agreements, the more direct solution is to direct renegotiation of RPS PPAs. For PG&E, these contracts represent more than $1 billion a year in excess costs, which dwarfs any of the actual, if any, subsidies to NEM customers. The fact is that solar rooftops displaced the very expensive renewables that the IOUs signed, and probably led to a cancellation of auctions around 2015 that would have just further encumbered us.

The bill would force net energy metered (NEM) customers to pay twice for their power, once for the solar panels and again for the poor portfolio management decisions by the utilities. The utilities claim that $3 billion is being transferred from customers without solar to NEM customers. In SDG&E’s service territory, the claim is that the subsidy costs other ratepayers $230 per year, which translates to $1,438 per year for each NEM customer. But based on an average usage of 500 kWh per month, that implies each NEM customer is receiving a subsidy of $0.24/kWh compared to an average rate of $0.27 per kWh. In simple terms, SDG&E is claiming that rooftop solar saves almost nothing in avoided energy purchases and system investment. This contrasts with the presumption that energy efficiency improvements save utilities in avoided energy purchases and system investments. The math only works if one agrees with the utilities’ premise that they are entitled to sell power to serve an entire customer’s demand–in other words, solar rooftops shouldn’t exist.

Finally, this initiative would squash a key motivator that has driven enthusiasm in the public for growing environmental awareness. The message from the state would be that we can only rely on corporate America to solve our climate problems and that we can no longer take individual responsibility. That may be the biggest threat to achieving our climate management goals.

Is the NASDAQ water futures market transparent enough?

Futures markets are settled either physically with actual delivery of the contracted product, or via cash based on the difference in the futures contract price and the actual purchase price. The NASDAQ Veles California Water Index future market is a cash settled market. In this case, the “actual” price is constructed by a consulting firm based on a survey of water transactions. Unfortunately this method may not be full reflective of the true market prices and, as we found in the natural gas markets 20 years ago, these can be easily manipulated.

Most commodity futures markets, such at the crude oil or pork bellies, have a specific delivery point, such as Brent North Sea Crude or West Texas Intermediate at Cushing, Oklahoma or Chicago for some livestock products. There is also an agreed upon set of standards for the commodities such as quality and delivery conditions. The problem with the California Water Index is that these various attributes are opaque or even unknown.

Two decades ago I compiled the most extensive water transfer database to date in the state. I understand the difficulty of collecting this information and properly classifying it. The bottom line is that there is not a simple way to clearly identify what is the “water transfer price” at any given time.

Water supplied for agricultural and urban water uses in California has many different attributes. First is where the water is delivered and how it is conveyed. While water pumped from the Delta gets the most attention, surface water comes from many other sources in the Sacramento and San Joaquin Valleys, as well as from the Colorado River. The cost to move this water greatly varies by location ranging from gravity fed to a 4,000 foot lift over the Tehachapis.

Second is the reliability and timing of availability. California has the most complex set of water rights in the U.S. and most watersheds are oversubscribed. A water with a senior right delivered during the summer is more valuable than a junior right delivered in the winter.

Third is the quality of the water. Urban districts will compete for higher quality sources, and certain agricultural users can use higher salinity sources than others.

A fourth dimension is that water transfers are signed for different periods and delivery conditions as well as other terms that directly impact prices.

All of these factors lead to a spread in prices that are not well represented by a single price “index”. This becomes even more problematic when a single entity such as the Metropolitan Water District enters the market and purchases one type of water which they skews the “average.” Bart Thompson at Stanford has asked whether this index will reflect local variations sufficiently.

Finally, many of these transactions are private deals between public agencies who do not reveal key attributes these transfers, particularly price, because there is not an open market reporting requirement. A subsequent study of the market by the Public Policy Institute of California required explicit cooperation from these agencies and months of research. Whether a “real time” index is feasible in this setting is a key question.

The index managers have not been transparent about how the index is constructed. The delivery points are not identified, nor are the sources. Whether transfers are segmented by water right and term is not listed. Whether certain short term transfers such as the State Water Project Turnback Pool are included is not listed. Without this information, it is difficult to measure the veracity of the reported index, and equally difficult to forecast the direction of the index.

The housing market has many of these same attributes, which is one reason why you can’t buy a house from a central auction house or from a dealer. There are just too many different dimensions to be considered. There is housing futures market, but housing has one key difference from the water transfer market–the price and terms are publicly reported to a government agency (usually a county assessor). Companies such as CoreLogic collect and publish this data (that is distributed by Zillow and Redfin.)

In 2000, natural gas prices into California were summarized in a price index reported by Natural Gas Intelligence. The index was based a phone survey that did not require verification of actual terms. As part of the electricity crisis that broke that summer, gas traders found that they could manipulate gas prices for sales to electricity generators higher by simply misreporting those prices or by making multiple sequential deals that ratcheted up the price. The Federal Energy Regulatory Commission and Commodity Futures Trading Commission were forced to step in and establish standards for price reporting.

The NASDAQ Veles index has many of the same attributes as the gas market had then but perhaps with even less regulatory protections. It is not clear how a federal agency could compel public agencies, including the U.S. Bureau of Reclamation, to report and document prices. Oversight of transactions by water districts is widely dispersed and usually assigned to the local governing board.

Trying to introduce a useful mechanism to this market sounds like an attractive option, but the barriers that have impeded other market innovations may be too much.

ERCOT has the peak period scarcity price too high

The freeze and resulting rolling outages in Texas in February highlighted the unique structure of the power market there. Customers and businesses were left with huge bills that have little to do with actual generation expenses. This is a consequence of the attempt by Texas to fit into an arcane interpretation of an economic principle where generators should be able to recover their investments from sales in just a few hours of the year. Problem is that basic of accounting for those cashflows does not match the true value of the power in those hours.

The Electric Reliability Council of Texas (ERCOT) runs an unusual wholesale electricity market that supposedly relies solely on hourly energy prices to provide the incentives for incenting new generation investment. However, ERCOT is using the same type of administratively-set subsidies to create enough potential revenue to cover investment costs. Further, a closer examination reveals that this price adder is set too high relative to actual consumer value for peak load power. All of this leads to a conclusion relying solely on short-run hourly prices as a proxy for the market value that accrues to new entrants is a misplaced metric.

The total ERCOT market first relies on side payments to cover commitment costs (which creates barriers to entry but that’s a separate issue) and second, it transfers consumer value through to the Operating Reserve Demand Curve (ORDC) that uses a fixed value of lost load (VOLL) in an arbitrary manner to create “opportunity costs” (more on that definition at a later time) so the market can have sufficient scarcity rents. This second price adder is at the core of ERCOT’s incentive system–energy prices alone are insufficient to support new generation investment. Yet ERCOT has ignored basic economics and set this value too high based on both available alternatives to consumers and basic regional budget constraints.

I started with an estimate of the number of hours where prices need the ORDC to be at full VOLL of $9000/MWH to recover the annual revenue requirements of combustion turbine (CT) investment based on the parameters we collected for the California Energy Commission. It turns out to be about 20 to 30 hours per year. Even if the cost in Texas is 30% less, this is still more 15 hours annually, every single year or on average. (That has not been happening in Texas to date.) Note for other independent system operators (ISO) such as the California ISO (CAISO), the price cap is $1,000 to $2,000/MWH.

I then calculated the cost of a customer instead using a home generator to meet load during those hours assuming a life of 10 to 20 years on the generator. That cost should set a cap on the VOLL to residential customers as the opportunity cost for them. The average unit is about $200/kW and an expensive one is about $500/kW. That cost ranges from $3 to $5 per kWh or $3,000 to $5,000/MWH. (If storage becomes more prevalent, this cost will drop significantly.) And that’s for customers who care about periodic outages–most just ride out a distribution system outage of a few hours with no backup. (Of course if I experienced 20 hours a year of outage, I would get a generator too.) This calculation ignores the added value of using the generator for other distribution system outages created by events like a hurricane hitting every few years, as happens in Texas. That drives down this cost even further, making the $9,000/MWH ORDC adder appear even more distorted.

The second calculation I did was to look at the cost of an extended outage. I used the outages during Hurricane Harvey in 2017 as a useful benchmark event. Based on ERCOT and U.S. Energy Information Reports reports, it looks like 1.67 million customers were without power for 4.5 days. Using the Texas gross state product (GSP) of $1.9 trillion as reported by the St. Louis Federal Reserve Bank, I calculated the economic value lost over 4.5 days, assuming a 100% loss, at $1.5 billion. If we assume that the electricity outage is 100% responsible for that loss, the lost economic value per MWH is just under $5,000/MWH. This represents the budget constraint on willingness to pay to avoid an outage. In other words, the Texas economy can’t afford to pay $9,000/MWH.

The recent set of rolling blackouts in Texas provides another opportunity to update this budget constraint calculation in a different circumstance. This can be done by determining the reduction in electricity sales and the decrease in state gross product in the period.

Using two independent methods, I come up with an upper bound of $5,000/MWH, and likely much less. One commentator pointed out that ERCOT would not be able achieve a sufficient planning reserve level at this price, but that statement is based on the premises that short-run hourly prices reflect full market values and will deliver the “optimal” resource mix. Neither is true.

This type of hourly pricing overemphasizes peak load reliability value and undervalues other attributes such as sustainability and resilience. These prices do not reflect the full incremental cost of adding new resources that deliver additional benefits during non-peak periods such as green energy, nor the true opportunity cost that is exercised when a generator is interconnected rather than during later operations. Texas has overbuilt its fossil-fueled generation thanks to this paradigm. It needs an external market based on long-run incremental costs to achieve the necessary environmental goals.

What is driving California’s high electricity prices?

This report by Next10 and the University of California Energy Institute was prepared for the CPUC’s en banc hearing February 24. The report compares average electricity rates against other states, and against an estimate of “marginal costs”. (The latter estimate is too low but appears to rely mostly on the E3 Avoided Cost Calculator.) It shows those rates to be multiples of the marginal costs. (PG&E’s General Rate Case workpapers calculates that its rates are about double the marginal costs estimated in that proceeding.) The study attempts to list the reasons why the authors think these rates are too high, but it misses the real drivers on these rate increases. It also uses an incorrect method for calculating the market value of acquisitions and deferred investments, using the current market value instead of the value at the time that the decisions were made.

We can explore the reasons why PG&E’s rates are so high, much of which is applicable to the other two utilities as well. Starting with generation costs, PG&E’s portfolio mismanagement is not explained away with a simple assertion that the utility bought when prices were higher. In fact, PG&E failed in several ways.

First, PG&E knew about the risk of customer exit as early as 2010 as revealed during the PCIA rulemaking hearings in 2018. PG&E continued to procure as though it would be serving its entire service area instead of planning for the rise of CCAs. Further PG&E also was told as early as 2010 (in my GRC testimony) that it was consistently forecasting too high, but it didn’t bother to correct thee error. Instead, service area load is basically at the save level that it was a decade ago.

Second, PG&E could have procured in stages rather than in two large rounds of request for offers (RFOs) which it finished by 2013. By 2011 PG&E should have realized that solar costs were dropping quickly (if they had read the CEC Cost of Generation Report that I managed) and that it should have rolled out the RFOs in a manner to take advantage of that improvement. Further, they could have signed PPAs for the minimum period under state law of 10 years rather than the industry standard 30 years. PG&E was managing its portfolio in the standard practice manner which was foolish in the face of what was occurring.

Third, PG&E failed to offer part of its portfolio for sale to CCAs as they departed until 2018. Instead, PG&E could have unloaded its expensive portfolio in stages starting in 2010. The ease of the recent RPS sales illustrates that PG&E’s claims about creditworthiness and other problems had no foundation.

I calculated the what the cost of PG&E’s mismanagement has been here. While SCE and SDG&E have not faced the same degree of exit by CCAs, the same basic problems exist in their portfolios.

Another factor for PG&E is the fact that ratepayers have paid twice for Diablo Canyon. I explain here how PG&E fully recovered its initial investment costs by 1998, but as part of restructuring got to roll most of its costs back into rates. Fortunately these units retire by 2025 and rates will go down substantially as a result.

In distribution costs, both PG&E and SCE requested over $2 billion for “new growth” in each of its GRCs since 2009, despite my testimony showing that growth was not going to materialize, and did not materialize. If the growth was arising from the addition of new developments, the developers and new customers should have been paying for those additions through the line extension rules that assign that cost responsibility. The utilities’ distribution planning process is opaque. When asked for the workpapers underlying the planning process, both PG&E and SCE responded that the entirety were contained in the Word tables in each of their testimonies. The growth projections had not been reconciled with the system load forecasts until this latest GRC, so the totals of the individual planning units exceeded the projected total system growth (which was too high as well when compared to both other internal growth projections and realized growth). The result is a gross overinvestment in distribution infrastructure with substantial overcapacity in many places.

For transmission, the true incremental cost has not been fully reported which means that other cost-effective solutions, including smaller and closer renewables, have been ignored. Transmission rates have more than doubled over the last decade as a result.

The Next10 report does not appear to reflect the full value of public purpose program spending on energy efficiency, in large part because it uses a short-run estimate of marginal costs. The report similarly underestimates the value of behind-the-meter solar rooftops as well. The correct method for both is to use the market value of deferred resources–generation, transmission and distribution–when those resources were added. So for example, a solar rooftop installed in 2013 was displacing utility scale renewables that cost more than $100 per megawatt-hour. These should not be compared to the current market value of less than $60 per megawatt-hour because that investment was not made on a speculative basis–it was a contract based on embedded utility costs.

Drawing too many conclusions about electric vehicles from an obsolete data set

The Energy Institute at Haas at the University of California published a study allegedly showing that electric vehicles are driven about only one-third of the average standard car in California. I responded with a response on the blog.

Catherine Wolfram writes, “But, we do not see any detectable changes in our results from 2014 to 2017, and some of the same factors were at play over this time period. This makes us think that newer data might not be dramatically different, but we don’t know.“

A recent study likely is delivering a biased estimate of future EV use. The timing of this study reminds me of trying to analyze cell phone use in the mid-2000s. Now household land lines are largely obsolete, and we use phones even more than we did then. The period used for the analysis was during a dramatically changing period more akin to solar panel evolution just before and after 2010, before panels were ubiquitous. We can see this evolution here for example. Comparing the Nissan Leaf, we can see that the range has increased 50% between the 2018 and 2021 models.

The primary reason why this data set is seeing such low mileage is because is almost certain that the vast majority of the households in the survey also have a standard ICE vehicle that they use for their extended trips. There were few or no remote fast charge stations during that time and even Tesla’s had limited range in comparison. In addition, it’s almost certain that EV households were concentrated in urban households that have a comparatively low VMT. (Otherwise, why do studies show that these same neighborhoods have low GHG emissions on average?) Only about one-third of VMT is associated with commuting, another third with errands and tasks and a third with travel. There were few if any SUV EVs that would be more likely to be used for errands, and EVs have been smaller vehicles until recently.

As for copurchased solar panel installation, these earlier studies found that 40% or more of EV owners have solar panels, and solar rooftop penetration has grown faster than EV adoption since these were done.

I’m also not sure that the paper has captured fully workplace and parking structure charging. The logistical challenges of gaining LCFS credits could be substantial enough for employers and municipalities to not bother. This assumption requires a closer analysis of which entities are actually claiming these credits.

A necessary refinement is to compare this data to the typical VMT for these types of households, and to compare the mileage for model types. Smaller commuter models average less annual VMT according to the California Energy Commission’s vehicle VMT data set derived from the DMV registration file and the Air Resources Board’s EMFAC model. The Energy Institute analysis arrives at the same findings that EV studies in the mid 1990s found with less robust technology. That should be a flag that something is amiss in the results.

How to increase renewables? Change the PCIA

California is pushing for an increase in renewable generation to power its electrification of buildings and the transportation sector. Yet the state maintains a policy that will impede reaching that goal–the power cost indifference adjustment (PCIA) rate discourages the rapidly growing community choice aggregators (CCAs) from investing directly in new renewable generation.

As I wrote recently, California’s PCIA rate charged as an exit fee on departed customers is distorting the electricity markets in a way that increases the risk of another energy crisis similar to the debacle in 2000 to 2001. An analysis of the California Independent System Operator markets shows that market manipulations similar to those that created that crisis likely led to the rolling blackouts last August. Unfortunately, the state’s energy agencies have chosen to look elsewhere for causes.

The even bigger problem of reaching clean energy goals is created by the current structure of the PCIA. The PCIA varies inversely with the market prices in the market–as market prices rise, the PCIA charged to CCAs and direct access (DA) customers decreases. For these customers, their overall retail rate is largely hedged against variation and risk through this inverse relationship.

The portfolios of the incumbent utilities, i.e., Pacific Gas and Electric, Southern California Edison and San Diego Gas and Electric, are dominated by long-term contracts with renewables and capital-intensive utility-owned generation. For example, PG&E is paying a risk premium of nearly 2 cents per kilowatt-hour for its investment in these resources. These portfolios are largely impervious to market price swings now, but at a significant cost. The PCIA passes along this hedge through the PCIA to CCAs and DA customers which discourages those latter customers from making their own long term investments. (I wrote earlier about how this mechanism discouraged investment in new capacity for reliability purposes to provide resource adequacy.)

The legacy utilities are not in a position to acquire new renewables–they are forecasting falling loads and decreasing customers as CCAs grow. So the state cannot look to those utilities to meet California’s ambitious goals–it must incentivize CCAs with that task. The CCAs are already game, with many of them offering much more aggressive “green power” options to their customers than PG&E, SCE or SDG&E.

But CCAs place themselves at greater financial risk under the current rules if they sign more long-term contracts. If market prices fall, they must bear the risk of overpaying for both the legacy utility’s portfolio and their own.

The best solution is to offer CCAs the opportunity to make a fixed or lump sum exit fee payment based on the market value of the legacy utility’s portfolio at the moment of departure. This would untie the PCIA from variations in the future market prices and CCAs would then be constructing a portfolio that hedges their own risks rather than relying on the implicit hedge embedded in the legacy utility’s portfolio. The legacy utilities also would have to manage their bundled customers’ portfolio without relying on the cross subsidy from departed customers to mitigate that risk.