Assessing California’s Climate Policies—Electricity Generation

Assessing California’s Climate Policies—Electricity Generation GABRIEL PETEK LEGISLATIVE ANALYST JANUARY 2020 analysis full gutter AN LAO REPORT LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT Executive Summary Chapter 135 of 2017 (AB 398, E. Garcia) requires our office to report annually on the economic impacts and benefits of the state’s greenhouse gas (GHG) limits. In this report, we assess the effects of some of the state’s major policies intended to reduce GHG emissions from the generation of electricity. Electricity Sector Primary Driver of GHG Emission Reductions. Over the last decade, the electricity sector has been the primary driver of statewide GHG emission reductions. Annual emissions from the electricity sector have declined by about 40 million metric tons (40 percent) over this period. Reductions have mostly been due to a change in the mix of resources used to generate electricity—primarily large increases in renewables (solar and wind) and, to a lesser extent, reductions in the amount of coal. State Policies Likely Key Factors in Reductions, but Magnitude of Effects Uncertain. In total, state policies were likely substantial drivers of changes to the generation mix that lowered annual emissions. However, a wide variety of other factors likely influenced emissions over the same period, including declines in natural gas prices, declines in prices for renewable generation, and federal policies. We did not identify any academic studies that comprehensively evaluated the overall effects attributed to state GHG reduction policies. RPS Likely a Substantial Driver of Emission Reductions at Moderate Cost Per Ton. Based on some “back-of-the-envelope” calculations, we estimate that the Renewable Portfolio Standard (RPS) program (1) reduced annual emissions by up to the low tens of millions of tons in 2018 and (2) costs about $60 to $70 per ton reduced in energy procurement costs. A variety of other costs—such as transmission and integration costs—are difficult to quantify, but could increase costs by tens of dollars per ton. Although the program likely generated other benefits— such as reducing local air pollutants and contributing to a global decline in solar prices—the magnitude of these effects appears to have been relatively small. Importantly, future costs to increase renewable generation are likely to be much different than past costs. This is because procurement costs for renewable energy are likely to be much lower in the future due to declining renewable prices, but this could be at least partially offset by higher integration costs. Rooftop Solar Policies Generally More Costly. State policies—such as the California Solar Initiative (CSI) and net energy metering (NEM)—likely had a significant impact on the amount of electricity generated from rooftop solar, which has reduced annual emissions by several million tons. However, these policies generally were a more expensive method for reducing emissions than policies focused on utility-scale renewables. Costs of electricity from distributed solar are at least a couple of times higher than utility-scale solar. Furthermore, estimated costs of emission reductions under CSI were about $150 to $200 per ton. The overall effects of NEM are not clear, but the policy has likely resulted in a substantial financial cost-shift from solar customers to nonsolar customers. The magnitude of other potential advantages of rooftop solar—including knowledge “spillovers” from learning-by-doing and reduced distribution system costs—are less clear, but appear to be relatively small. Little Known About Effects of SB 1368 and Cap-and-Trade on Emissions. Although a 2006 state law that prohibited new long-term contracts with coal power plants (Chapter 598, www.lao.ca.gov 1 analysis full gutter AN LAO REPORT SB 1368 [Perata]) likely reduced emissions from coal generation, we did not identify any empirical research assessing the magnitude of the effects. For cap-and-trade, the level of costs is clearer than the level of emission reductions. Market prices for allowances suggest the marginal costs for emission reductions encouraged by the program have been less than $20 per ton, but the overall amount of emission reductions from electricity generation attributable to the program is unclear. The cap-and-trade program has had significant distributional effects in the electricity sector. Specifically, the overall financial benefit to residential electricity customers from utilities selling allowances and using the proceeds to benefit ratepayers has exceeded the compliance costs that have been passed on to customers. Resource Shuffling Potentially Offsets Some of the Emission Reductions. Resource shuffling occurs when the mix of existing electricity supplies changes so that more low-carbon electricity is sent to California while more high-carbon electricity is sent to other states. Several different prospective analyses showed that there was potential for significant resource shuffling from imports. There has been limited retrospective empirical research estimating resource shuffling, but some preliminary work suggests it could be a significant factor. Key Issues for Legislative Consideration. We identify some key issues for the Legislature to consider as it modifies and adopts policies to achieve its GHG goals. • Comprehensive Policy Evaluations Lacking. Although the amount of information varies by program, we found a lack of rigorous retrospective evaluations for some programs and effects. The Legislature might want to consider directing agencies to identify opportunities to facilitate retrospective evaluation by ensuring data is available to researchers and, potentially, designing programs in ways that allow for more robust evaluations. In addition, the Legislature could consider additional reporting requirements and/or funding for research efforts in key areas, including resource shuffling and the effect of distributed solar on distribution system costs. • Mix of Policies Likely Not Most Cost-Effective Way to Reduce GHGs. There has been substantial differences in the costs of reducing emissions between cap-and-trade (marginal cost that are currently less than $20 per ton), RPS (average costs of about $60 to $70 per ton or more), and policies promoting distributed solar (average costs of roughly $150 to $200 per ton). In the future, the Legislature might want to rely more heavily on the most cost-effective programs, such as cap-and-trade. In certain limited instances, the Legislature could consider adopting policies that are a somewhat more costly way to reduce GHGs if those policies result in substantial benefits in other ways, such as reducing local air pollution and creating knowledge spillovers. • High Electricity Prices Could Be a Barrier to GHG Reductions. Retail electricity rates are substantially higher than the marginal social costs of providing electricity. This is due to a variety of factors including (1) utilities recovering fixed costs through volumetric rates, (2) declining electricity consumption (which means fixed costs are spread over a smaller base), and (3) costs for various state-mandated programs. High electricity rates discourage adoption of some technologies—such as electric vehicles and electric appliances—that could be used to substantially reduce statewide GHGs. As a result, the Legislature might want to consider actions that more closely align retail electricity rates with the marginal costs of providing the electricity. For example, the Legislature could direct regulators to exclude at least some of the fixed costs and certain state policy costs from utilities’ volumetric electricity rates, and potentially fund them in other ways. 2 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT INTRODUCTION Chapter 488 of 2006 (AB 32, Núñez/ However, the rate of reductions needed to reach Pavley) established the goal of limiting the SB 32 target are much greater. greenhouse gas (GHG) emissions statewide to Chapter 135 of 2017 (AB 398, E. Garcia) 1990 levels—431 million metric tons (MMT) of requires our office to report annually on the carbon dioxide equivalent (CO2e)—by 2020. In economic impacts and benefits of the state’s GHG 2016, Chapter 249 (SB 32, Pavley) extended the limits. In 2018, we issued two reports in fulfillment limit to 40 percent below 1990 levels—259 MMT of of this requirement. First, we released Assessing CO2e—by 2030. As shown in Figure 1, emissions California’s Climate Policies—An Overview, which have decreased since AB 32 was enacted and provided the analytical framework we are using were already below the 2020 target in 2017. to assess the economic impacts and benefits of Figure 1 State Met 2020 Goal Early, but 2030 Goal More Ambitious Million Metric Tons of Greenhouse Gases 500 450 2020 400 AB 32 Target 350 300 2030 250 Actual Emissions SB 32 Target 200 150 100 50 2006 2008 2010 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 www.lao.ca.gov 3 analysis full gutter AN LAO REPORT climate policies. Second, we released Assessing intended to reduce emissions from the generation California’s Climate Policies—Transportation, of electricity—hereafter referred to as electricity which applied that framework to the various state generation or electricity supply. (We do not assess programs designed to reduce GHG emissions the effects of programs primarily intended to reduce from the transportation sector. In this report, we electricity consumption, such as energy efficiency assess the effects of the state’s major policies programs, in this report.) OVERVIEW OF ELECTRICITY SECTOR EMISSIONS Electricity Sector Background Utilities—both public and private—own and operate distribution lines. Overview of Electric Grid. A wide variety of entities—both public and private—play a role in Load Serving Entities (LSEs) Procure providing electricity to California households and Electricity and Deliver it to Customers. Load businesses. In general, there are three main parts serving entities provide electricity to end users. of the electric grid: They are responsible for generating or purchasing electricity and ensuring it is delivered to households • Generation. Electricity frequently is and businesses. Historically, investor-owned generated at large power plants (such as utilities (IOUs) and publicly owned utilities natural gas, coal, or nuclear power plants) (POUs) have been the primary LSEs. IOUs are or large renewable generation sites (such as private companies regulated by the California wind farms or solar fields). This large-scale Public Utilities Commission (CPUC). POUs are generation is also known as utility-scale public agencies governed by locally elected or generation. These power plants typically are appointed officials. More recently, other types owned by private companies (including some of nonutility LSEs are providing an increasing utilities). Some generation occurs at a smaller share of electricity to customers. These include scale, such as solar installed at residences, community choice aggregators (CCAs), which are businesses, or other smaller-scale community local government-run entities that buy electricity for locations. This smaller-scale generation is customers but use IOU distribution to deliver the known as distributed generation and usually is electricity, and electric service providers (ESPs), owned by the property owner or a third-party which are private entities that sell electricity directly company that installs and owns the generation to commercial customers in IOU territories. In source. 2018, IOUs provided about 55 percent of electricity • Transmission. Electricity generated to California customers, POUs provided about at utility-scale is transported through 25 percent, and CCAs and ESPs provided about high-voltage power lines known as 10 percent each. transmission lines. These lines typically are Electricity Generated From a Wide Variety of owned by utilities. In some cases, electricity Sources. Figure 2 shows the different generation is sent directly from transmission lines to sources used to generate electricity that is end customers, such as large manufacturing consumed in California. Natural gas is by far the facilities. largest source of generation. Wind, solar, large • Distribution. Generally, electricity is hydroelectric, unspecified imports, and nuclear transferred from high-voltage transmission contribute a significant share as well. (Unspecified lines to low-voltage distribution lines before imports are imported electricity where it is not it is delivered to customers. For example, possible to identify the specific generation sources distribution lines are often on wooden poles used to produce the electricity.) Roughly 70 percent that run through cities and neighborhoods. of electricity consumed in California is generated 4 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT in-state and the remaining Figure 2 30 percent is generated out of Electricity Generated From a Wide Variety of Sources state but imported into California Percent of Total Generation, 2018 through transmission lines. Major Policies to Reduce Natural gas Electricity Sector Wind Solar Total = 314,955 Gigawatt Hours Emissions Large hydro The electricity sector accounts Unspecified for 16 percent of statewide GHG Nuclear emissions, according to the Distributed solar PV California Air Resources Board Geothermal (CARB) statewide GHG inventory. In-State Coal In addition to the statewide GHG Imports Biomass goals discussed above, in recent Small hydro years the state has established Other GHG goals that are specific to the 5 10 15 20 25 30 35% electricity sector. This includes Chapter 547 of 2015 (SB 350, de León), which requires CARB PV = photovoltaic. to establish 2030 GHG targets for the electricity sector (set at a range of 30 MMT to 53 MMT). In the last couple of decades. Chapter 516 of 2002 addition, Chapter 312 of 2018 (SB 100, de León) (SB 1078, Sher) established a 20 percent RPS by establishes a state policy of 100 percent zero 2017, and Chapter 464 of 2006 (SB 107, Simitian) carbon electricity by 2045. Over the past couple accelerated the 20 percent requirement to 2010. of decades, the state has implemented a variety of Subsequently, Chapter 1 of 2011 (SBX1 2, Simitian) policies intended to reduce GHG emissions from established a 33 percent requirement by 2020. In electricity generation. Figure 3, on the next page, 2015, SB 350 established a 50 percent requirement summarizes some of the major policies, which we by 2030—a target that SB 100 increased to describe in more detail below. 60 percent a few years later. State law and Renewable Portfolio Standard (RPS). State regulations also establish interim RPS requirements law requires LSEs (with a few exceptions) to and targets. Figure 4 on page 7, shows the RPS provide a minimum percent of retail electricity requirements under current law and regulation. sales from qualifying renewable generation. The CPUC oversees IOU, CCA, and ESP Qualifying renewables include solar, wind, biomass, compliance. The California Energy Commission geothermal, and small hydroelectric. Notably, under (CEC) oversees POU compliance. An LSE complies current law, some generation sources that do not by “retiring” enough renewable energy credits directly emit GHGs, such as large hydroelectric (RECs) to cover its required RPS percentage of and nuclear, do not qualify under RPS. Distributed retail sales. A REC is a certificate demonstrating generation, such as rooftop solar PV (photovoltaic), that one unit of electricity was generated and technically can qualify. However, in practice, very delivered from an eligible renewable resource. little of it is used to comply in part because certain State law establishes other requirements about administrative actions needed to certify RPS what types of RECs may be used to comply (such eligibility can be expensive for smaller PV units. as a maximum percent of RECs from renewable The Legislature has increased or extended energy that was generated in other states, but not the RPS requirements a few different times over delivered to California). www.lao.ca.gov 5 analysis full gutter AN LAO REPORT California Solar Initiative (CSI). In 2006, rebates—based on per kilowatt of generation Chapter 132 of 2006 (SB 1, Murray) provided capacity—to offset the upfront cost of the solar state agencies the authority to establish unit) for businesses and existing homes installing several programs aimed at providing incentives rooftop solar. The program had a declining for distributed solar—an effort known as Go incentives structure. The incentives started high Solar California. The overall goal was to install and then automatically decreased over time as 3,000 megawatts of distributed solar and transition each IOU hit certain thresholds for the total amount the solar industry to a point where it could be of solar installed in its jurisdiction. The incentives self-sustaining. The biggest program used to reduced the cost of installing a residential solar achieve this goal was the CSI, which provided unit by about 25 percent in the early years of the financial incentives to install rooftop solar on program and by about 5 percent to 10 percent businesses and existing homes in IOU territories. in the final years. This design was intended to Other programs included the New Solar Home gradually reduce customer reliance on subsidies Partnership Program, which provided financial as the solar industry matured and market prices incentives for solar on newly constructed homes, declined. The General Market Program stopped and a wide variety of solar programs offered accepting applications for incentives in 2016. through POUs. The statewide budget for these Net Energy Metering (NEM). The vast majority programs was $3.3 billion over a ten year period— of rooftop solar customers are enrolled in NEM, from 2006 to 2016—with about $2.7 billion going which supports onsite solar installations. Some to the CSI. The programs were primarily funded version of NEM has been in place since 1996, but through a surcharge on electricity bills. has been modified several times since then. Under The CSI included several different subprograms NEM, the utility effectively pays solar customers that provided customer incentives for distributed (through a bill credit) for the excess electricity solar. The largest subprogram—called the General they generate that is exported back to the grid. Market Program—provided a total of about Under NEM, the customer receives the retail rate $2 billion in upfront financial incentives (primarily for electricity, which includes costs associated Figure 3 Summary of Major Policies to Reduce Emissions From Electricity Generation Policy Year Implemented Description Renewable Portfolio Standard 2003 Requires LSEs to generate a minimum percent of retail electricity from qualifying renewable sources. Percentages increase over time from 20 percent in 2010 to 60 percent in 2030. California Solar Initiative 2007 Provided $2.7 billion over a ten-year period for financial incentives to reduce the cost of installing distributed solar, such as rooftop solar PV. Net Energy Metering 1996 Encourages customers to install distributed solar generation by paying them a retail electricity rate for the electricity they generate. Emissions Performance Standard 2007 Effectively prohibits LSEs from signing or extending (SB 1368)a long-term contracts with coal power plants. Cap-and-trade 2013 Requires electricity generators and importers to obtain an allowance or offset to cover each ton of GHG emitted. Program includes other emitters outside of the electricity sector, and entities can buy and sell allowances. a Chapter 598 of 2006 (SB 1368, Perata). LSE = load serving entity; PV = photovoltaic; and GHG = greenhouse gas. 6 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT with generation, transmission, Figure 4 and distribution. For example, if RPS Requirements Increase Over Time a customer consumes 100 kwh Renewable Generation as a Percent of Retail Sales of electricity from the grid, but exports 70 kwh of electricity from 70% their solar panels back to the grid, 60 then the customer would pay the retail rate for 30 kwh of electricity. 50 In response to state legislation, 40 CPUC made some changes to 30 the NEM program in 2016. Much 20 of the basic structure described above remains in place. Some 10 of the key changes included charging new NEM customers 2006 2010 2014 2018 2022 2026 2030 a one-time interconnection fee RPS = Renewable Portfolio Standard. and a requirement that new NEM residential customers use time-of-use (TOU) rates. Time generally purchase allowances at regular state of use is a rate plan in which rates vary according auctions or from other entities subject to the to the time of day and season. Higher rates are cap-and-trade regulations. In addition, some charged during typical high demand hours and allowances are given away for free. For example, lower rates are charged during low demand hours. the state allocates utilities additional allowances for Emissions Performance Standard. free, but they must be used to benefit ratepayers. Chapter 598 of 2006 (SB 1368, Perata) established In most cases, the utilities sell these additional the emissions performance standard for California allowances to other carbon emitters and use the LSEs. The standard prohibited LSEs from building revenue to provide bill credits to customers. This new generation or signing new long-term contracts is meant to offset the higher costs to consumers with generation sources that emit more than associated with cap-and-trade, but in such a 1,100 pounds of carbon dioxide per megawatt way as to not reduce their incentive to reduce hour. This effectively prohibited LSEs from signing electricity consumption. A small portion of the or extending long-term contracts with coal power revenue generated is used for other things, plants. such as renewable energy or energy efficiency programs. (For more information about the state’s Cap-and-Trade. Under the state’s cap-and-trade program, see our report The cap-and-trade program, in-state electricity 2017-18 Budget: Cap-and-Trade.) generators and electricity importers must obtain a compliance instrument—usually through the Other Programs. The state has a variety of purchase of “allowances” (or offsets)—to cover their other programs that are intended to facilitate GHG emissions. This adds costs to higher-carbon GHG reductions from electricity generation. sources of electricity (such as coal or natural These include the Self Generation Incentive gas) which, consequently, increases demand for Program, the New Solar Homes Partnership low-carbon sources of electricity (such as wind Program, and a mandate that utilities purchase and solar). The increased costs for higher-carbon a certain amount of electricity storage to help electricity are also intended to provide an incentive integrate larger percentages of intermittent for customers to reduce their consumption. renewables onto the grid. (Wind and solar are examples of intermittent resources—meaning The number of allowances issued each year they are only generated during certain days and declines over time as the state’s GHG targets hours.) These programs are not the primary decline. Electricity generators and importers www.lao.ca.gov 7 analysis full gutter AN LAO REPORT focus of this report because, based on our initial that are simply due to an accounting change. review, their effects are likely smaller than the This is especially relevant when discussing other policies identified above. changes in imports and coal generation. Electricity emissions reduced by 39 MMT Electricity Sector Is Primary Source (38 percent) from 2009 to 2017. During this same of State Emission Reductions period, there was a net increase of more than 5 MMT from other sources of emissions. Annual Emissions From Electricity Sector Have Decreased by About 40 Percent Since Overall Generation Relatively Steady, but 2006. Figure 5 summarizes the annual level of GHG Intensity Has Declined Substantially. Total emissions from various sources from 2000 to emissions from electricity depend on two basic 2017. As shown in the figure, the electricity factors: (1) total amount of electricity generated sector has been the major source of absolute (megawatt hours, for example) and (2) the emission reductions over the last decade. From emission intensity (tons of CO2e per megawatt 2006 to 2017, electricity sector emissions hour). As shown in Figure 6, total electricity have declined by 42 MMT (40 percent). It is generation has been relatively steady over the important to note that, in 2009, CARB changed last decade, but emission intensity has declined the methods it used to estimate emissions by about 40 percent. Thus, it is a change in from imported electricity, and CEC changed its the mix of generation resources used to supply methods for identifying the source of electricity electricity that has been the primary driver of imports. The accounting change was one factor absolute emission reductions. contributing to the observed decline in estimated Most Declines Have Come From Imported state emissions between 2008 and 2009. As a Electricity. As shown in Figure 7, on page 10, result, in this report, we often focus on changes emissions from both in-state generation and in emissions and generation sources that occur imports have declined since 2009, but imports after 2009 in order to avoid capturing changes have been the largest contributor to emission reductions. Overall generation from both in-state generation and Figure 5 imports has been relatively steady. Electricity Sector Emissions Have Declined Substantially Most of the changes have been Million Metric Tons of Carbon Dioxide Equivalent due to a reduction in emission intensity from imports, which has 200 decreased by half since 2009. 180 Renewables Increasing, While Transportation Coal and Nuclear Decreasing. 160 Figure 8, on page 10, illustrates 140 changes in the total mix of 120 generation sources since Electricity 2009 based on data from the 100 CEC. The substantial increase Industrial 80 in utility-scale renewables is the most notable change over this 60 Commercial and Residential period—the large majority of 40 which was solar PV and wind. 20 Agriculture High Global Warming Potential Distributed solar also increased substantially, even though the Recycling and Waste overall amount of generation is still 2000 2002 2004 2006 2008 2010 2012 2014 2016 relatively small. Generation from nuclear—a non-GHG emitting 8 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT source—and coal—a high source of emissions— time period. In the following section, we discuss the have decreased in the last several years. extent to which state policies have contributed to Generation from natural gas, large hydroelectric, these changes. and unspecified sources have all varied over this LAO ASSESSMENT OF POLICIES As discussed in our 2018 report, Assessing and discuss key issues for future legislative California’s Climate Policies—An Overview, state consideration. policies have a wide variety of potential effects. Policies Likely Substantial Drivers These include: of Emission Reductions, but Actual • Benefits. Includes GHG reductions, Magnitude Unclear reductions in criteria and toxic air pollutants, and promotion of activities—such as research Mix of Resources Used to Generate Electricity and development—that create knowledge Has Lower Emissions. The changing generation “spillovers” that have social benefits. mix described above has lowered emissions. • Costs. Includes changes that increase the For example, our simple “back-of-the-envelope” net cost of delivering electricity, including estimate reveals that from 2009 to 2018, the generation, transmission, and/or distribution. increase in renewable generation reduced annual The increase in costs could also discourage emissions by about 30 MMT of carbon dioxide— businesses and households from undertaking about 6 MMT from the increase in rooftop solar valuable economic activities. • Distributional Effects. Figure 6 Includes instances when Electricity Generation Has Been Relatively Steady, revenues or costs are shifted but Emissions Intensity Has Declined Substantially from certain households Percent Change Relative to 2006 or businesses to others, without any net changes in 20% economic costs or benefits. In this section, we summarize 10 our understanding of the major Total Generation (MWh) effects of California’s policies to reduce electricity emissions. Our assessment is based on a review -10 of academic studies and various reports; our own analysis of data -20 Emissions Intensity (Tons of CO2e per MWh) from government agencies and researchers; and conversations -30 with various stakeholders, agencies, and researchers. The primary focus of our assessment -40 is on the past effects of major policies, rather than projecting -50 future effects of policies. In the 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 following section, we identify MWh = megawatt hour and CO2e = carbon dioxide equivalent. www.lao.ca.gov 9 analysis full gutter AN LAO REPORT and 24 MMT from the increase in utility-scale • Decline in Natural Gas Prices. Natural gas renewables. In addition, the decline in coal fuel prices decreased significantly over the generation reduced annual emissions by about 8 last decade. Notably, prices declined by more MMT. These estimates are a rough proxy that do than 50 percent between 2008 and 2009. As not take into account a wide variety of complicating a result, natural gas generation might have factors, such as how a large addition of renewable generation Figure 7 capacity might affect the mix of Emissions From Imports other generation that was built Declined More Than In-State Emissions and used. However, they provide Millions of Metric Tons of Carbon Dioxide Equivalent a rough sense of the amount of GHG reductions associated with 60 changes in the generation mix In-State Generation over the last few years. 50 State Policies Likely a 40 Substantial Driver of Reductions . . . In total, state Imports 30 policies were likely substantial drivers of changes to the 20 generation mix that lowered emissions. RPS and rooftop 10 solar policies were almost certainly major factors in the 2009 2010 2011 2012 2013 2014 2015 2016 2017 significant expansion in renewable generation, particularly in early years when prices for this generation were much higher. (We discuss the price declines in Figure 8 more detail below.) In addition, Renewable Generation Increased SB 1368 was likely one factor While Coal and Nuclear Decreased that contributed to utilities Percent of Total Generation divesting from coal power plants. 45% Finally, by making high-carbon electricity more expensive 40 relative to low-carbon electricity, Natural Gas 35 cap-and-trade likely reduced 30 the GHG intensity of electricity purchased by California LSEs. 25 Renewables . . . But a Wide Variety of 20 Unspecified and Other Other Factors Also Likely 15 Influencing Emissions. Although Large Hydro state policies were likely 10 significant drivers of changes in 5 Coal Nuclear electricity sector emissions, a Distributed Solar PV wide variety of other factors may 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 have also increased or decreased electricity emissions over the last PV = photovoltaic. several years. These include: 10 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT replaced at least some of the coal generation including imports and natural gas generation. even without state policy. Based on other research that has been done • Decline in Prices for Renewable and our own estimates, the closure increased Generation. As shown in Figure 9, the costs annual emissions by about 7 MMT to 8 MMT of unsubsidized renewable generation— annually. particularly wind and solar—have declined • Annual Changes in Hydroelectric substantially over the last decade. For Generation. Hydroelectric generation is not example, global costs for utility-scale solar PV likely a factor contributing to the long-term declined by nearly 90 percent from 2009 to trend in declining emissions because total 2019. In recent years, LSEs likely would hydroelectric capacity has not changed much have purchased some renewable generation over this period. However, hydroelectric because it is less costly than other sources generation varies from year to year based of generation, even if the state did not have largely on the amount of rainfall in preceding an RPS policy. (As we discuss below, a small years and, therefore, can be a significant portion of the price decline might be driven factor affecting short-term differences in by California policies, but much of the cost emissions. For example, assuming natural declines are likely driven by other global gas as default, differences in hydroelectric factors.) generation over the last several years have • Federal Policies. The federal government changed annual emissions by about 10 MMT. offers tax credits for wind and solar that likely One working paper estimates that the drought were factors contributing to the increase in several years ago increased emissions by renewables. For example, in 2006, the federal about 8 MMT annually. government implemented a solar investment • Voluntary Purchases of “Green” Electricity. tax credit offering a tax deduction of up to Some households and businesses voluntarily 30 percent of the cost of the solar system. The tax Figure 9 credit has been modified Costs of Wind and Solar Have and extended a couple of Declined Substantially In Recent Years times. Notably, there was Dollars Per Megawatt Hour originally a $2,000 cap on residential deductions that $400 was eliminated in 2008. 350 Certain federal environmental regulations that limit pollution 300 from coal plants could also be factors affecting decisions 250 to retire coal plants. • San Onofre Nuclear Plant 200 Solar PV Stopped Operating. In 150 2012, one of the state’s two nuclear power plants Combined Cycle Natural Gas 100 stopped generating electricity due to safety Onshore Wind 50 concerns. This increased overall emissions since much 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 of the zero-carbon electricity had to be replaced by other PV = photovoltaic. sources of generation, www.lao.ca.gov 11 analysis full gutter AN LAO REPORT choose to purchase low- or zero-carbon overall benefits, costs, and distributional effects electricity even if it is more expensive than attributable to state climate change policies in the alternatives. This can be done through Green electricity sector. Given the many different factors Tariff programs offered through utilities, as well that affect California’s electricity emissions, it is as by businesses that sign direct contracts difficult to attribute changes in emissions to any with renewable electricity providers. This has particular set of state policies. Such an analysis likely driven some of the increase in renewable would likely require complex statistical modeling to generation. estimate the effect of state policies. Furthermore, • Economic Recession. In 2008, California as we discussed in previous reports, there are and the rest of the world suffered its worst significant interactions between state policies. For economic recession in several decades. example, state policies that reduce emissions from A reduction in economic activity tends to sources that are covered by the cap-and-trade reduce emissions. In the electricity sector, program—such as electricity sector policies that we would expect a recession to largely affect reduce emissions from in-state generators and overall electricity generation, rather than the importers—might simply free up allowances for mix of resources used to generate electricity. other sources to emit more. The net effect would This is because electricity consumption be to simply change the source of emissions but (and generation) is likely more closely tied to not reduce the overall amount that would have changes in economic activity. Although the been reduced if only cap-and-trade were in place. recession likely had an effect on emissions Even with the most sophisticated modeling tools by reducing overall generation below what it available, it is unclear whether it would be possible would have otherwise been, it likely is not a to precisely estimate the total effect of California significant driver of the decreases in electricity policies. emission intensity. The economic growth over While there are significant challenges associated the last several years may have contributed to with evaluating the overall effects of state policies some growth in generation. However, similarly, in the electricity sector, there is some information it likely did not have much effect on the available about the effects of specific policies. change in emission intensity. Some policies have been evaluated by researchers • Resource Shuffling. Emissions leakage using complex statistical techniques, while others is when emission reductions that occur have had almost no retrospective evaluation. in California are offset by an increase in Although there is not complete information on all emissions in other states and countries. the relevant effects of any particular policy, in some Resource shuffling is a specific type of cases, the available data and research can provide leakage that occurs when—in response to valuable information about some of the major state policies—more electricity generated from effects. In the next sections, we review the effects low-carbon sources is sent to California, but of the RPS, rooftop solar policies, SB 1368, and more high-carbon electricity is sent to other cap-and-trade. states. As a result, on paper, the electricity RPS Likely a Significant Driver of being used in California has lower emissions. Reductions At Moderate Costs However, the overall generation mix and total emissions throughout the western United Per Ton States does not change. We discuss the RPS Reducing State Emissions by Up to potential for resource shuffling in more detail the Low Tens of Millions of Tons Annually. below. In general, LSEs have met or exceeded RPS No Rigorous Analysis of Overall Effects of requirements so far. According to CEC estimates, State Policies. To our knowledge, there are no 34 percent of statewide retail sales in 2018 were studies that have comprehensively evaluated the met by RPS eligible resources. However, to our knowledge, there is no analysis of California GHG 12 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT emission reductions directly attributable to RPS. accounting could change the estimates by several As discussed above, our back-of-the-envelope million tons annually. calculations suggest large-scale renewables Direct IOU Compliance Costs Likely Over reduced emissions by roughly 24 MMT annually $1 Billion Annually—Roughly 5 Percent of Total between 2009 and 2018. Many of these reductions Costs. State law requires CPUC to report annually are likely attributable to the RPS. However, for on IOU RPS procurement and generation costs, many of the reasons discussed above, it is difficult increases in total utility costs from meeting RPS to isolate the effect of the RPS on the generation requirements, and avoided costs as a result of mix and emissions. Notably, some renewable meeting RPS. In total, CPUC estimates 2018 RPS generation would have been implemented even procurement expenditures for the three large IOUs without the RPS as we have seen in other areas were $1.1 billion higher than alternative sources of of the country. Figure 10 compares the growth electricity generation (the cost of a combined cycle in non-hydroelectric renewable generation in natural gas power plant). Although this estimate is different regions of the United States to the imperfect, it provides a rough sense of the RPS’s minimum growth required by state RPS policies higher generation costs. For context, the large IOUs in each region. In some regions—such as the collect about $24 billion in annual revenue from Northeast and Mid-Atlantic—growth in renewable “bundled” customers—or customers for whom the generation is largely consistent with minimum IOU procures the energy, as well as the distribution RPS requirements in those states. On the other and transmission. The $1.1 billion in costs reflects hand, growth in renewable generation for other an almost 5 percent increase in overall retail regions—such as Texas, the Midwest—far exceeds rates for bundled IOU customers. This increase is RPS requirements. The renewable generation in generally consistent with national studies that have these states has been driven by such things as found increased rates from RPS of about 3 percent lower unsubsidized costs for renewable generation to 8 percent. (especially wind) and federal tax credits. Properly accounting for these other factors would likely reduce Figure 10 the estimated emission reductions Growth in Renewable Energy in Some that are attributable to the RPS Other States Exceeds Minimum RPS Requirements program. As a result, in our view, Terawatt Hours the estimate of 24 MMT of annual reductions is likely toward the 120 Actual Growth in Non-Hydro Renewables high end of the range of likely Minimum Growth Required for RPS emission reductions attributable 100 to the RPS. It is also worth noting that we relied on total system 80 generation data from the CEC to estimate renewable generation. 60 However, there are key differences in accounting for renewable 40 generation and GHGs between CARB, CEC, and CPUC. These 20 add to the uncertainty of the estimate. We do not think these 0 accounting differences would Northeast Mid-Atlantic West Texas Midwest Southeast dramatically affect the magnitude of the estimates, but differences in RPS = Renewable Portfolio Standard. www.lao.ca.gov 13 analysis full gutter AN LAO REPORT As we discuss below, current RPS costs largely upgraded transmission lines. Some national reflect long-term renewable contracts that were research has found that building transmission signed several years ago when renewable prices— to wind and solar accounts for about particularly for solar and wind—were much higher 3 percent to 30 percent of overall costs of a than they are today. These costs do not necessarily project. However, the degree to which policies reflect future programmatic costs. have affected transmission costs in California Back-of-the-Envelope Calculations Suggest is not clear. Moderate Direct RPS Costs Per Ton of • Integration Costs. Regulators and grid Reductions. We are not aware of any retrospective operators must ensure there are enough other evaluations of the cost per ton of reducing GHGs resources available to supply energy (and through the RPS. There are many challenges other related grid services) when demand associated with making such an estimate. increases or the supply of intermittent However, below we provide a back-of-the-envelope renewables decreases. This could include calculation to provide a rough sense of the costs payments to natural gas generators to make per ton—focusing on only the estimated differences sure they are available in case they are in procurement costs. needed to balance the grid and the costs Assuming RPS implementation by the large IOUs of paying for electricity storage so it can be is responsible for about 60 percent to 70 percent used during other times when the renewable of the reductions from 2006 to 2018, then RPS sources are not producing as much energy. To emission reductions from IOUs are about 17 MMT our knowledge, there has been no complete to 18 MMT in 2018. If direct procurement costs retrospective evaluation of these integration are about $1.1 billion higher, then the program costs in California. Some national research is reducing emissions at a cost of roughly $60 to has estimated that integration costs are $70 per ton. We note that this is a rough calculation similar to transmission costs described that excludes many factors, such as transmission above—about 3 percent to 30 percent of and integration costs. This estimate also attributes total project costs. These costs are likely all of the increase in renewable generation to the lower in earlier years with low percentages RPS, rather than other factors. As a result, we of intermittent renewables, but can increase think this estimate reflects the low end of the range substantially as the percentage of renewables of costs per ton. Actual costs related to the RPS grows. could be tens of dollars higher per ton. • Reduced Local Air Pollution. Replacing For context, cap-and-trade allowance prices are coal and natural gas generation with currently about $17 per ton. In theory, this suggests certain renewables—such as wind and that this program is encouraging emissions solar—reduces other local air pollutants. reductions that cost up to $17 per ton. In contrast, (It is less clear how an increase in other as we found in our 2018 report, Assessing renewables, such as biomass, affect total California’s Climate Policies—Transportation, nitrogen oxide [NOx] emissions.) Using a estimated costs for some of the transportation back-of-the-envelope calculation, we estimate programs are much higher—hundreds of dollars per the increase in wind and solar generation ton or more. reduced about 8,000 tons of NOx and 1,000 tons of particulate matter (PM2.5) in RPS Has a Wide Variety of Other Costs and 2018. For context, this is a relatively small Benefits. The RPS likely has other effects—both reduction compared to total statewide annual positive and negative—that are not captured in emissions—less than 2 percent for NOx and the estimates described above, which only reflect less than 1 percent for PM2.5. Plus, some energy procurement costs. These include: of these reductions come from generators • Additional Transmission Costs. Utility-scale in other states so do not affect air quality in renewable generation often requires new or California. It is also worth noting that some 14 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT studies have found that the RPS benefits from capacity. That said, even a small contribution local air pollution reductions meet or exceed to the substantial decline in prices could have the benefits of GHG reductions in other significant global benefits. states. However, the largest local air pollution Future Program Costs Could Be Much reductions occur in states that rely heavily Different Than Past Costs. As discussed above, on coal. The local air pollution benefits of an global costs to install renewable energy have RPS are generally lower in California because decreased substantially in recent years. Figure 11 the mix of existing generation that is being shows how prices for new IOU RPS contracts— displaced (mostly natural gas) generates much particularly wind and solar—have declined over less local air pollution. time. The costs for certain types of renewable • Contribution to Decline in Solar Prices. energy—particularly solar PV—under these more As discussed above, costs for renewables— recent contracts will be lower than they have and solar PV in particular—have declined been in the past. Since 2007, the only increase substantially over the last decade. The decline in overall renewable contract prices was in 2016 in costs of installing solar PV largely have which was primarily due to an increase in contracts been driven by a decline in the costs for for biomass electricity (one type of bioenergy) the solar panels (also known as modules). in response to a legislative mandate to procure California policies likely partially contributed a certain amount of biomass capacity. These to the declining price of solar panels. contracts were more expensive than recent wind For example, as a result of the additional and solar contracts. production of solar panels driven by California While procurement costs for renewables is policies, manufacturers could have learned likely to decline, as the percentage of intermittent how to produce the panels more efficiently— renewables used for generation grows, integration also known as “learning-by-doing” (LBD). costs are likely to increase. The net effect of these Other firms might then learn these techniques changes depends on the future trends in renewable for reducing production costs and, as a result, Figure 11 the market price for solar panels declines. However, Prices for New IOU Renewable Contracts it is important to note that Declined Substantially solar panels are produced Cents Per Kilowatt Hour and sold in a global market 20 and the degree to which California policies—relative 18 to other global factors— 16 affected those prices is 14 unclear. Given the scale and timing of some of 12 Bioenergy California’s activities relative 10 to those that occurred in 8 Overall Renewables other jurisdictions—such as Germany, China, and 6 Small Hydro and Geothermal Japan—California policies Solar and Wind 4 are likely a relatively minor 2 driver of these reductions. For example, solar PV 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 capacity in California is only a few percent of global PV IOU = investor-owned utility. www.lao.ca.gov 15 analysis full gutter AN LAO REPORT costs and the costs of different strategies to research related to other programs that offer manage intermittency (storage costs, for example). rebates for new technologies—such as hybrid electric vehicles—that finds a high proportion of Rooftop Solar Policies Generally More rebates go to free-riders. Costly Rooftop Solar Much More Costly Than Utility-Scale Solar. There has been a significant The state has implemented several different decline in the cost of installing distributed solar PV. policies aimed at increasing adoption of distributed However, as shown in Figure 12, recent estimates solar—primarily rooftop PV—as a way to reduce of the costs of generating electricity from different GHGs. In this section, we focus on two key policies solar PV sources shows that distributed solar PV that have been used to increase adoption—the is much more expensive than utility-scale solar PV. CSI and NEM—as well as some of the effects of For example, rooftop residential solar PV is about rooftop solar more generally. Relative to the other five times more costly than utility-scale solar PV. climate policies that we reviewed in this report and Commercial and industrial rooftop solar is about in previous reports, there has been a significant two to three times more expensive than utility amount of retrospective evaluation of the effects scale. Although these are global cost estimates of some of the state’s rooftop solar policies. In that are subject to a wide variety of limitations particular, there is a robust literature on the effects and uncertainties, they suggest that there is a of the CSI. We summarize the key findings about large difference in the costs of installing and CSI and NEM below. generating energy from distributed solar compared CSI Increased Adoption of Rooftop Solar, to utility-scale renewables. The difference in cost but Significant Portion of Rebates Went to “Free-Riders.” As discussed earlier, total distributed solar PV Figure 12 generation reduced annual Energy From Distributed Solar PV emissions by up to 6 MMT More Costly Than Utility-Scale Solar PV in 2018. Academic studies Range of Global Costs in Dollars Per Megawatt Hour, 2019 consistently find that the CSI rebates offered for solar installations were mostly or fully passed through to consumers in Rooftop residential the form of lower prices for the solar installations, rather than increasing profit for businesses selling the units. Furthermore, Rooftop commercial and industrial studies consistently find that the CSI increased rooftop solar adoption relative to a scenario where no CSI rebates were Community offered. However, these studies also found that a large portion— sometimes 50 percent or more— of the households that installed Utility scale solar would have purchased rooftop solar without the CSI 50 100 150 200 250 $300 rebate. Such consumers are sometimes known as free-riders. PV = photovoltaic. This finding is consistent with 16 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT could be due to a variety of factors, including largest IOUs in 2016. This study also found that higher installation costs per unit for rooftop solar the total amount of the cost shift for each utility due to economies of scale and greater ability to was a few hundred million dollars annually. This install utility-scale solar in locations that have the amount grows over time as the amount of rooftop most sunlight to maximize generation. solar grows and if utilities incur additional costs for CSI Costs Likely Significantly Higher Than distribution and transmission. Importantly, CPUC RPS. In addition to the differences in costs recently modified NEM and required that new NEM described above, a few studies estimated the customers enroll in TOU pricing. This change—at costs of CSI specifically. For example, two studies the time it was adopted in late 2016 and early found that the CSI rebates reduced GHG emissions 2017—was expected to reduce a solar owners’ at a program cost of about $150 to $200 per overall financial benefit of generation during ton. These estimates are higher than the $60 to “off-peak” hours—such as hours in the middle $70 per ton estimates for RPS described above. of the afternoon when rooftop solar generation It is worth noting that—similar to the RPS—these is relatively high but retail prices are lower under estimates generally do not include any program TOU rates. However, this is partially offset by an benefits or costs related to knowledge spillovers, increase in the benefit during some of the peak improvements in local air pollution, or other effects hours—such as in the late afternoon—when solar on the grid. PV is still generating and TOU rates are higher. On net, these changes are likely to reduce the overall Overall Effects of NEM Less Clear but financial benefit to customers from NEM. These Results in Substantial Cost Shift to Nonsolar changes also reduce the amount of the cost-shift to Customers. To our knowledge, there have been nonsolar customers. no retrospective evaluations of the overall GHG reductions and/or economic costs from NEM. Recent Study Finds Small LBD Benefits. However, one aspect of the NEM program that has One common rationale for California’s rooftop been evaluated is the degree to which the program solar policies—including CSI and NEM—is LBD, shifts costs from solar customers to nonsolar whereby the cost of a technology declines with customers. The key mechanism by which NEM more cumulative experience with the technology. provides financial incentives for customers to install If there are learning spillovers—where, for distributed generation is through shifting fixed example, one firm learns how to install solar more costs from solar customers to nonsolar customers. efficiently but other firms also learn from that This occurs because—for each unit of rooftop experience—then there could be an economic solar generation—solar customers no longer pay rationale for government policies that encourage the retail rate for utility-generated electricity that greater deployment of new technologies. (This is includes fixed costs for the transmission and similar to the justification for governments funding distribution systems. When solar customers no research and development to create knowledge longer pay for these fixed costs, these costs are that is publicly available.) Most of the potential generally built into the electricity rates paid for LBD benefits for rooftop solar programs are likely by other (nonsolar) customers. It is important to to occur for what are known as balance of system note that changing who pays for fixed costs that (BOS) costs—or costs related to the installation have already been incurred is not considered of the solar panels, rather than the costs of the a net economic cost, but can have significant equipment. At least a few different studies have distributional implications. estimated the degree to which CSI has led to learning-by-doing for solar BOS costs—one as the One rough estimate by an economist at the program was beginning and two after the program University of California found that the additional was implemented. costs borne by nonsolar customers is about $65 per customer annually. Another evaluation • 2008 Prospective Study Found That found that the benefit to the solar customer of the Primary Benefit From CSI Was LBD . . . cost shift was about $1,200 annually for two of the One 2008 study found that LBD benefits www.lao.ca.gov 17 analysis full gutter AN LAO REPORT were roughly ten times greater than the The research on the effects of distributed solar direct environmental benefits associated with on distribution costs is somewhat limited but shows the CSI. The study found that, without LBD mixed results. One study found that the net costs benefits, environmental benefits did not justify depend on various factors including how much CSI subsidies. However, assuming a certain other local distributed solar PV exists, as well as level of LBD benefits, the level of CSI rebates certain other characteristics of the distribution grid were close to optimal. at the specific location. The same study found • . . . But Retrospective Studies Find Small that there was very little benefit associated with LBD Effects. More recent research has reducing “congestion” on most distribution circuits, found very weak evidence of LBD benefits but there was substantial value on 1 percent from 2002 to 2012, and the magnitude of circuits. The value is especially significant in of effect was relatively small. During the areas where circuits are very close to needing a study period, BOS costs declined by less capacity upgrade. Another working paper (which than $1 per watt ($3 per watt to a little focused only on the costs to modify the existing more than $2 per watt). Over a similar time network but not potential avoided or delayed costs) period, hardware costs declined from over found that (1) the vast majority of a 100 percent $7 per watt to less than $3.5 per watt. Only increase in average residential distribution network 15 percent ($0.12) of the decline in BOS prices between 2003 and 2017 can be explained costs were found to be attributable to LBD. by the increase in distributed solar generation Further, there was evidence of only very small and (2) larger amounts of rooftop solar that is learning spillovers, at least in the short run. more concentrated geographically predict higher Another study found that LBD contributed to a distribution network costs. 5 percent decrease in solar prices, which is a Rooftop Solar Has Other Advantages Over clear benefit but relatively small compared to Utility-Scale, but Magnitude of Benefits Unclear. the 33 percent decrease in solar prices over There are some other areas where rooftop solar the entire period of the study. has clear advantages over utility-scale solar. For example, one common rationale for encouraging Little Evidence of a Substantial Reduction in rooftop solar is because it has fewer land use Distribution Costs, Except in Certain Locations. impacts than utility-scale solar installations that Some stakeholders argue that rooftop solar require acres of land, which sometimes require reduces a utility’s costs associated with building the conversion of natural and working lands. We out its distribution network. In theory, this could did not identify any research that quantified the occur because it reduces the demand on the magnitude of this potential benefit, but it could be system during peak hours of electricity demand, significant. Another potentially substantial benefit thereby reducing or delaying the need for the utility is that distributed solar could provide enhanced to add potentially expensive distribution capacity. electric reliability during electric power shutoffs Other potential benefits include potentially reducing that are being implemented to reduce the risk the amount of “line-loss,” or the electricity lost as of wildfires, particularly when distributed solar is it travels through the grid system, because the accompanied by battery storage. While it is clear electricity does not have to travel as far. However, such advantages might exist, the magnitude of an increase in rooftop solar also has the potential these benefits—and the degree to which state to add distribution costs. This could occur because policies might be needed to help promote these adding distributed solar sometimes can require actions—are unclear. modifications to the existing distribution network to accommodate the new generation sources being Little Known About Effects of SB 1368 connected to the grid. In total, therefore, the overall and Cap-and-Trade magnitude—and even the direction—of the effect of adding distributed solar on distribution costs is not No Empirical Research on the Effects of obvious. SB 1368. Since 2009, coal generation for California 18 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT has declined by over 60 percent and now makes example, the cost of purchasing allowances— up only 3 percent of the state’s electricity supply also known as compliance costs—is generally mix. We did not identify any retrospective empirical not considered a net economic cost. Instead, the research that identified the degree to which purchase of allowances results in a transfer of SB 1368 contributed to this decline. Based on money from the entity who ultimately bears the conversations with various stakeholders, it is likely cost of purchasing the allowances to those who that the policy was a significant factor contributing get the revenue from selling the allowances. In to the decline in coal generation. However, other the electricity sector, electricity generators and factors—such as declines in natural gas prices and importers are directly responsible for purchasing cap-and-trade—were also likely important. allowances, and those costs are generally passed Level of Cap-and-Trade Costs More Clear on to utility consumers in the form of higher Than Level of Emission Reductions. Allowance electricity rates. prices are an indicator of the marginal costs While ratepayers bear costs associated with for emission reductions encouraged by the cap-and-trade for the electricity they use, the cap-and-trade program. Since the cap-and-trade state provides free allowances to IOUs that they program began in 2013, electricity generators and sell and use the revenue to benefit ratepayers. importers have had to pay a carbon price of roughly The large majority of this revenue is used to $10 to $17 per ton. In theory, if lower-carbon provide a semiannual bill credit to residential electricity can be provided for a net difference customers. As shown in Figure 13, on next page, in costs of less than $17 per ton, then the compliance costs for IOU ratepayers are far less lower-carbon electricity will be generated in or sent than the amount of IOU proceeds from the sale of to California. allowances going to IOU ratepayers. It is important The emission reductions associated with the to note that the amount of allowance revenue that program are more difficult to estimate. Carbon goes to benefit ratepayers is also used to provide prices have been incorporated into wholesale bill credits for CCA and ESP customers, but the electricity market bids, which at times has likely compliance costs do not include costs for CCA resulted in lower carbon mixes of electricity and ESP customers. However, even after adjusting supply being purchased in the market than would for this difference, this data suggests that electric otherwise have been the case. It is also possible ratepayers have, on average, benefited financially that expectations about future carbon prices have from the economic transfers that occur under affected LSE long-term procurement decisions. the program. It is also important to note that the However, to our knowledge, there has not been effect of cap-and-trade on consumers of other any empirical research estimating these effects. fuels—such as transportation fuels—is likely much Based on conversations with stakeholders and different because businesses and consumers researchers, the effect on electricity sector in those sectors do not receive as many free emissions is generally thought to have been allowances. relatively modest compared to other policies, Resource Shuffling Potentially such as RPS. As emissions targets become more Offsets Some of the Emissions ambitious in future years, cap-and-trade could result in significantly higher costs and emissions Reductions reductions associated with electricity generation. Resource shuffling occurs when, in response Allocating Free Allowances to IOUs Under to California climate policies, the mix of existing Cap-and-Trade Has Benefited Ratepayers. electricity supplies changes so that more As discussed in our 2018 report Assessing low-carbon electricity is sent to California while California’s Climate Policies—An Overview, some more high-carbon electricity is sent to other of the most visible effects of the cap-and-trade states. To the degree this occurs, the reduction program are not net economic costs, but what in the carbon intensity of California’s electricity are known in economic terms as transfers. For supply would not actually reflect a net reduction www.lao.ca.gov 19 analysis full gutter AN LAO REPORT in low-carbon generation. Understanding the long-term contracts with coal power plants, degree to which resource shuffling has occurred but these coal plants might continue to is an important factor in identifying the net effect operate and sell to entities in other states California policies have had on overall GHG instead, thereby not resulting in a decrease in emissions. Resource shuffling could be driven by total electricity generation from coal. a wide variety of policies in the electricity sector, • Regional Electricity Markets. In short-term but it is especially relevant for SB 1368 and wholesale markets where electricity is cap-and-trade. dispatched based on lowest cost, low-carbon Different Potential Mechanisms for energy is delivered to California to avoid the Resource Shuffling or Leakage. As outlined state’s carbon price while high-carbon energy in a 2018 report from the state’s Independent is sent to other states. Emissions Market Advisory Committee (IEMAC), Research Shows Potential for there are several different mechanisms through Significant Resource Shuffling, but Not which state policies could contribute to leakage or Much Retrospective Evaluation. As discussed resource shuffling in the electricity sector. These above, many of the emission reductions in the include: electricity sector have come through reduced • Bilateral Contract Shuffling. California emissions intensity of imports. Several different entities will no longer enter into bilateral prospective analyses showed that there was potential for significant resource Figure 13 shuffling. For example, as a result of SB 1368 and/or Electric IOU Ratepayers Get Net Financial Benefit cap-and-trade, coal power From Cap-and-Trade Allowance Allocation plants that no longer send (Total Amount in Millions) electricity to California could $1,200 provide electricity to other Costs Related to Purchasing Allowances states, while lower-carbon sources of electricity (such as Proceeds From Selling Allowances 1,000 hydroelectric) would be sent to California. Some of these studies estimated the magnitude of resource shuffling could be 800 at least several million tons annually. To our knowledge, however, 600 there has been very little retrospective empirical research estimating resource shuffling. 400 As a result, the degree to which resource shuffling has actually occurred is highly uncertain. 200 The only empirical research we are aware of is a working paper that found emissions decreased by 12 million tons annually in 2014 2015 2016 2017 2018 California and increased by about 8.5 million tons in other IOU = investor-owned utility. parts of the western United 20 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT States after cap-and-trade was implemented. Summary of Findings This suggests about 70 percent of emission Figure 14 summarizes the key findings from our reductions leaked out of state. review of the effects of the state’s major climate policies affecting the mix of electricity generation. Figure 14 Summary of Findings Policies likely substantial driver of emission reductions, but actual magnitude is unclear. • Changing mix of resources used to generate electricity—largely higher levels of renewables—has substantially lowered emissions. • State policies likely a substantial driver of reductions, but a wide variety of other factors also likely influence emissions. • No rigorous analysis of overall effects of state policies. Renewable Portfolio Standard (RPS) likely a significant driver of emission reductions at relatively moderate costs per ton. • RPS reducing California emissions by up to low tens of millions of tons annually. • Direct compliance costs for investor-owned utilities (IOUs) likely over $1 billion annually, or 5 percent of total electricity costs. • “Back-of-the-envelope” calculations suggests program reducing emissions at moderate direct cost (about $60 to $70 per ton). • A variety of other costs, such as transmission costs and integration costs, are more difficult to quantify. These could increase costs by tens of dollars per ton. • Some benefits from reduced local air pollution, but likely a relatively small impact on overall California air pollution. • Likely a contributing factor to the global decline in solar prices. Magnitude of effect is unclear, but likely relatively small compared to other factors and actions in other jurisdictions. • Future program costs could be much different than past costs. Decline in prices for wind and solar will lower energy costs, but other costs related to integrating intermittent renewables could grow. Rooftop solar policies generally more costly emission reduction strategy. • California Solar Initiative (CSI) increased rooftop solar adoption, but a significant portion of rebates went to “free-riders.” • Distributed solar much more costly way to generate electricity than utility-scale solar. • Costs of emission reductions under CSI (roughly $150 to $200 per ton) likely significantly higher than RPS. • Overall effects of Net Energy Metering unclear, but substantial cost-shift to nonsolar customers (roughly $65 higher costs annually) from solar customers (roughly $1,200 lower costs annually). • Recent study finds only small “learning-by-doing” benefits from policies. • Little evidence of a substantial reduction in distribution system costs, except in certain locations. • Rooftop solar has benefits related to reducing land use impacts and increasing electricity reliability during outages, but magnitude of benefits unclear. Little known about overall effects of SB 1368 and Cap-and-Trade. • No empirical research on effects of Chapter 598 of 2006 (SB 1368, Perata). • Level of marginal cap-and-trade costs (roughly $17 per ton) relatively clear, but level of emission reductions unclear. • Allocating free cap-and-trade allowances to IOUs under cap-and-trade has benefited electric ratepayers. Resource shuffling potentially offsets some of the observed emission reductions. • Prospective studies show potential for significant resource shuffling. • Limited retrospective evaluation, but some evidence of shuffling. www.lao.ca.gov 21 analysis full gutter AN LAO REPORT KEY ISSUES FOR LEGISLATIVE CONSIDERATION The prior section summarizes key findings methods that could better identify which solar from our review of the effects of policies that installations were attributable to the CSI versus have been implemented so far. In this section, we other factors. discuss some of the key issues for the Legislature Similar to our comments in previous reports, to consider going forward as the state modifies the Legislature might want to consider directing and adopts policies to achieve its GHG goals. agencies to identify opportunities to help facilitate Specifically, we identify considerations related to better retrospective evaluation before programs are facilitating future policy evaluations, promoting adopted or modified. This planning process could cost-effectiveness, and reducing barriers to include requiring implementing agencies to develop long-term electrification. a research plan for the program that would identify, for example, what data would be collected and how Comprehensive Policy Evaluations the program could be designed to help facilitate Lacking retrospective evaluation. The Legislature also might want to consider directing state agencies to consult As described above—and similar to findings in with academic researchers during this process. our 2018 report on transportation policies—we found a lack of rigorous retrospective evaluations RPS Reports Might Be Guide for Other of the major effects of some of the state’s climate Climate Programs With Potential for Some policies related to electricity generation. Below, Improvements. Much of the information on we provide options that the Legislature might want the effects of the RPS was based on multiple to consider to help ensure more robust evaluation reports that were required from CPUC on an of state climate policies in the future. Findings annual basis. One of these reports focuses on from these evaluations could help inform the progress in complying with the RPS requirements. Legislature’s future policy and budget decisions, Another focuses on the costs associated with as well as provide valuable information for other RPS procurement. These reports provide helpful jurisdictions considering adopting similar policies information on the past and current effects of the intended to reduce GHG emissions. program. Most notably, the cost report includes an estimate of what energy expenditures would Consider Directing Agencies to Identify have been without the RPS. Although imperfect, Opportunities to Facilitate Retrospective the estimate provides helpful information on the Evaluation. Agencies can help facilitate costs associated with the program compared to a retrospective evaluation by ensuring data are scenario where the RPS did not exist. We are not available to researchers and, potentially, designing aware of similar state reporting requirements for programs in ways that allow for more robust many of the other major climate policies (including evaluations. The CSI is an example of a program some of the state’s major transportation policies, that had both of these features and, as a result, which we discussed in our report last year). The there is a significant amount of information Legislature might want to consider whether similar about the program’s effects. First, the program requirements could be implemented for other collected—and made publicly available—a lot of climate programs. data about the amount of solar generation that was installed under the program. Second, some Although CPUC’s RPS reports provide valuable specific features of the program were structured in information, there might be opportunities to a way that facilitated more robust evaluation. For improve them. For example, there are additional example, rebates varied across time and location RPS costs—including transmission costs and depending on the amount of solar that had already integration costs—that are not included in the been installed in a utility’s service territory. This report. The Legislature might want to require more variation allowed researchers to utilize research reporting on RPS related to these costs. Although 22 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT more difficult to estimate, these costs—particularly want to consider directing CPUC to evaluate the integration costs—could be a substantial part these effects in more detail. of the overall costs of implementing RPS in future years. Additional information on these costs could Mix of Policies Likely Not Most help inform future legislative decisions about Cost-Effective Way to Reduce GHGs potential changes to the program. In addition, the current RPS reporting requirements do not require As the state’s GHG goals become more CPUC to estimate the GHG emission reductions stringent, the overall costs to reduce emissions is associated with the program. The Legislature might likely to grow. The higher costs are likely driven by want to require CPUC to report on the estimated (1) greater annual emission reductions needed to GHG benefits of the program since emission meet the target are higher because the reduction reductions is the primary goal of the program. targets are more aggressive and (2) the cost per ton of reducing emissions increases as the low-cost Other Future Reporting and Research emission reduction actions—or the “low hanging Priorities. Some of the key areas where the fruit”—have already been taken. As a result, Legislature might want to consider additional cost-effectiveness becomes increasingly important. reporting requirements and/or funding for research efforts include: Based on our review of the available information, there has been substantial differences in the costs • Resource Shuffling. Evaluation of resource of reducing emissions between cap-and-trade shuffling is particularly important since many (marginal cost that are currently less than $20 per of the emission reductions have come from ton), RPS (average costs of about $60 to $70 per imports, and there is a body of research ton or more), and distributed solar policies (average that suggests resource shuffling could be a costs of roughly $150 to $200 per ton). In the significant factor. The degree to which it has future, the Legislature might want to consider occurred is still unclear though. The 2018 relying more heavily on the most cost-effective Annual Report of the IEMAC noted some programs, such as cap-and-trade. key challenges associated with accurately There might be instances where there is a estimating resource shuffling, and it provided strong rationale for supporting policies that are recommendations intended to improve not the most cost-effective in the short term, monitoring and mitigation. These included but that provide other important benefits. For such things as better harmonization of example, some policies, such as those that data between CARB, CEC, and CPUC. The promote innovation or LBD by supporting new Legislature might want to consider requiring technologies, might have significant long-run some of those changes in order to facilitate benefits by creating knowledge spillovers. Some greater evaluation. targeted state policies focused on the development • Effect of Distributed Solar on Distribution of technologies that could help achieve those goals Costs. As discussed earlier, rooftop solar might be warranted. The focus of such efforts could can have costs and benefits related to the include (1) technologies that are in earlier stages of distribution grid. Some of these effects have development but that might end up being valuable been studied, but additional evaluation of to help meet long-term GHG goals, (2) technologies these effects could be particularly valuable to where increase in deployment is more likely to inform future decisions related to distributed result in LBD spillovers, and (3) technologies that solar. For example, to our knowledge, very are more likely to be adopted in other jurisdictions. little is known about the degree to which—or Another example, is policies that result in significant where—distributed solar has reduced costs reductions in local air pollutants. In some cases, by delaying the need to make distribution the Legislature might want to consider adopting infrastructure upgrades. The Legislature might policies that are a somewhat more costly way to reduce GHGs if those policies result in substantially www.lao.ca.gov 23 analysis full gutter AN LAO REPORT greater reductions in local air pollution. Ultimately, purchased from the utility) and improvements the Legislature will have to balance the higher costs in energy efficiency. Since fixed costs are against some of these other benefits, which can be largely recovered in volumetric rates, then difficult to quantify in some cases. declining electricity consumption can increase electricity rates. With fewer retail sales, higher High Electricity Prices Could Be electricity rates are needed to raise the same Barrier to Future Emission Reductions amount of revenue to cover fixed costs. • State Program Costs. A wide variety of state As the state’s GHG targets become more policies and programs also increase electricity ambitious and new technologies are deployed more rates for an average California customer. widely, it will become increasingly important for This includes policies discussed earlier in this the state to consider relationships across different report—such as RPS, cap-and-trade, and sectors. In other words, consider how policies in policies that promote distributed generation— one sector—such as electricity—affect emissions as well as a wide variety of other policies in other sectors—such as transportation fuels and to promote energy efficiency, fund electric fuels for home appliances. One important example vehicle infrastructure, and provide subsidized of this relationship is how electricity rates affect rates for low income customers. incentives to electrify other parts of the economy. We discuss this issue in more detail below. High Retail Rates Could Make It More Difficult California Rates Are Significantly Higher to Achieve Long-Term GHG Goals. There is Than Most Other States. Retail electricity rates nothing inherently wrong with having rates that in California are generally much higher than are higher than other states. For example, if the many other areas of the country. For example, rates reflect the true social costs of providing an the average rate in California in 2017 was about extra unit of electricity (including environmental 16 cents per kwh, or about 50 percent higher than damages), then the prices might be appropriate. the national average of roughly 10 cents per kwh. However, based on findings from a recent working Rates vary among LSEs. For example, rates for the paper, electricity rates in California are more three largest IOUs range from 16 cents to 24 cents than twice as much as the marginal social costs per kwh. of providing electricity in California, even after Wide Variety of Factors Contributing to accounting for environmental damages. High Rates. Some of the factors that contribute Rates that are much higher than the social to California’s comparatively high retail electricity marginal costs have adverse economic effects prices include: because they discourage valuable economic activities that might have otherwise occurred. For • How Fixed Costs Are Recovered. One example, high rates might make it more expensive key factor affecting electricity rates is how a for a business to produce valuable goods and utility collects revenue to cover fixed costs services in California. Similarly, households might for transmission and distribution. Generally, avoid electricity consumption that is valuable to California IOUs recover their fixed costs them, such as setting the home thermostat at a through volumetric rates, thereby increasing more comfortable temperature. per kwh rates paid by customers. Utilities in Furthermore, high electricity rates could present other parts of the country often collect more a barrier to long-term emission reductions. of their fixed costs through monthly fixed Although high electricity rates might encourage charges. some emission reduction in the electricity sector • Declining Consumption. All three large through reduced consumption and greater IOUs have had declining retail electricity efficiency, they serve as a barrier to GHG sales in recent years, at least partly driven reductions in other sectors. For example, one by increases in distributed solar generation strategy for substantially reducing statewide GHGs (which decreases the amount of electricity 24 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT is electrification—or using low-carbon electricity Reduced Revenue From Lower Rates Could to power vehicles and provide heat in buildings. Be Made up in Other Ways. If at least some fixed This includes using electric vehicles instead of grid costs or state policy costs are removed from gasoline-powered cars. It could also include electricity rates, the state would have to determine using electric appliances—such as electric heat who pays for those costs and how. A couple of pumps and water heaters—in place of appliances potential solutions are: powered by natural gas. Decisions by households • Fixed Charges. The state could direct utilities and businesses about whether or not to adopt to increase the use of fixed monthly charges these alternative technologies depend, in part, for electricity customers to pay for fixed grid on electricity rates. Higher electricity rates could costs. One of the primary concerns with this discourage some adoption of these lower-carbon approach is that it might shift a greater portion technologies. The relative weight given to energy of the fixed costs to households that tend to efficiency compared to electrification of other consume less electricity, including low-income sectors might depend, in part, on whether the households. One potential solution to this Legislature’s primary focus is on incremental concern would be to assess lower fixed near- to medium-term reductions, or whether charges for low-income customers. the primary goal is long-term decarbonization. • Other Funding Sources. The state could Although energy efficiency can potentially help consider using other funding sources to pay reduce emissions in the near- to medium-term, for costs for certain programs that are aimed the state cannot reach substantial economywide at achieving statewide policy goals, such as decarbonization with only energy efficiency. It deployment of electric vehicle infrastructure. must adopt other low- or zero-carbon sources of Since many of these programs are aimed energy for all sectors of the economy. Electrification at promoting statewide public policy goals, of a substantial portion of other sectors, along there is a rationale for using state budgetary with a decarbonized electricity grid, is one of the resources to cover the costs, including the strategies most often discussed for achieving those General Fund and Greenhouse Gas Reduction types of substantial GHG reductions. Fund (which gets funding from auctions of Consider Retail Rate Structures That More the state’s cap-and-trade allowances). The Closely Reflect Marginal Costs of Electricity. Legislature could also direct the utilities to use The Legislature might want to consider actions that revenue from the sale of their cap-and-trade encourage LSEs to adopt retail electricity rates that allowances to pay for some of these more closely reflect the marginal costs of providing programs. the electricity. For example, the Legislature could direct CPUC to exclude at least some of the fixed Each of these options involves its own set of costs and certain state policy costs from IOU implications and trade-offs that would need to volumetric electricity rates. From an economic be carefully examined. For example, using other perspective, this would improve efficiency by state funding sources would require a reduction in making rates more closely reflect the marginal funding for other state programs and/or additional social costs of providing the electricity. From a revenue collected from taxpayers. Each of these GHG perspective, the lower volumetric rates would has trade-offs that would need to be balanced reduce the incentive for energy conservation and against the potential ratepayer and emission energy efficiency, but would make it more financially reduction benefits related to lower electricity rates. attractive for households and businesses to purchase electric vehicles and appliances. www.lao.ca.gov 25 analysis full gutter AN LAO REPORT CONCLUSION The state’s climate policies in the electricity certain instances; (4) relatively little is known about sector are complex and have a wide range of the overall effects of SB 1368 and the benefits from effects—both positive and negative. Based on cap-and-trade; and (5) resource shuffling could our review, some of our main findings include: potentially offset some of the observed emission (1) state policies are likely a substantial driver of reductions. We also identify a variety of issues emission reductions, but the overall magnitude of for the Legislature to consider going forward as the effect is unclear; (2) the RPS program is likely a the state modifies and adopts policies to achieve significant driver of emission reductions at relatively its GHG goals. These issues include program moderate costs per ton, (3) rooftop solar policies evaluation, cost-effectiveness, and how high are generally a more costly emission reduction electricity rates could potentially serve as a barrier strategy, but could provide significant benefits in to long-term GHG reductions. 26 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT APPENDIX Selected References the Independent Emissions Market Advisory Committee.” Barbose, G (2019). “U.S. Renewable Portfolio Lazard (2019). “Lazard’s Levelized Cost of Standards: 2019 Annual Status Update.” Lawrence Energy Analysis—Version 13.0.” Berkeley National Laboratory. Lo Prete, C., A. Tyagi, and C. Hohl (2019). Bollinger, B. and K. Gillingham (2019). “California’s Cap-and-Trade Program and Emissions “Learning-by-Doing in Solar Photovoltaic Leakage in the Electricity Sector: An Empirical Installations.” Working paper, April 3 version. Analysis.” Manuscript in preparation. July 3rd Borenstein, S., and J. Bushnell (2018). “Do Two version. Electricity Pricing Wrongs Make a Right? Cost Mastrandrea, M.D., M. Inman, and D. Cullenward Recovery, Externalities, and Efficiency.” Natural (2019). “Assessing California’s progress toward its Bureau of Economic Research. Working Paper 2020 greenhouse gas emissions limit.” Pre-print. No. 24756. Rogers E. and S. Sexton (2015). “Effectiveness Bushnell, J., Y. Chen, and M. Zaragoza-Watkins. of Capacity-Dependent Rooftop Solar Subsidies: (2014). “Downstream regulation of CO2 emissions Lessons from California.” American Economic in California’s electricity sector.” Energy Policy 64: Association’s Annual Meeting: Allied Social Science 313-23. Associations Program. Presentation on January 3rd. California Public Utilities Commission (2019). Sergici, S., Y. Yang, M. Castaner and A. Faruqui “2019 Padilla Report: Costs and Savings for the (2019). “Quantifying net energy metering subsidies.” RPS Program.” The Electricity Journal (32) 106632. California Energy Commission (2018). “Tracking Van Bentham, A., K. Gillingham, and R. Sweeney Progress: Renewable Energy.” Updated December (2008). “Learning-by-Doing and the Optimal Solar 2018. Policy in California.” The Energy Journal 29(3): Cohen, M.A., B.A. Kauzmann, and D.S. Callaway 131-151. (2016). “Effects of distributed PV on California’s Wiser, R., G. Barbose, J. Heeter, T. Mai, L. Bird, distribution system, part 2: Economic analysis.” M. Bolinger, A. Carpenter, G. Heath, D. Keyser, Solar Energy 128: 139-152. J. Macknick, A. Mills, and D. Millstein (2016). “A Cullenward, D. (2014). “Leakage in California’s Retrospective Analysis of the Benefits and Impacts carbon market.” Electricity Journal 27(9): 36-48. of U.S. Renewable Portfolio Standards.” Lawrence Davis, L. “Why Am I Paying $65/year for Your Berkeley National Laboratory and National Solar Panels?” Energy Institute Blog, UC Berkeley, Renewable Energy Laboratory. March 26, 2018. NREL/TP-6A20-65005. Gorman, W., A. Mills, and R. Wiser (2019). Wolak, F. (2018). “The Evidence from California “Improving estimates of transmission capital costs on the Economic Impact of Inefficient Distribution for utility-scale wind and solar projects to inform Network Pricing.” National Bureau of Economic renewable energy policy.” Energy Policy. Pre-print. Research. Working Paper 25087. Hughes, J.E and M. Podolefsky. 2015. “Getting Green with Solar Subsidies: Evidence from the California Solar Initiative.” Journal of the Association of Environmental and Resource Economists. Volume 2, Number 2: 235-275. Independent Emissions Market Advisory Committee (2018). “2018 Annual Report of www.lao.ca.gov 27 analysis full gutter AN LAO REPORT 28 LEGISLATIVE ANALYST’S OFFICE analysis full gutter AN LAO REPORT www.lao.ca.gov 29 analysis full gutter AN LAO REPORT LAO PUBLICATIONS This report was prepared by Ross Brown and reviewed by Brian Brown and Anthony Simbol. The Legislative Analyst’s Office (LAO) is a nonpartisan office that provides fiscal and policy information and advice to the Legislature. To request publications call (916) 445-4656. This report and others, as well as an e-mail subscription service, are available on the LAO’s website at www.lao.ca.gov. The LAO is located at 925 L Street, Suite 1000, Sacramento, CA 95814. 30 LEGISLATIVE ANALYST’S OFFICE