The following is a contributed article by Fredrich Kahrl, managing partner at 3rdRail.
In a nailbiter, California narrowly escaped rotating outages on Sept. 6, the third CAISO stage 3 emergency in two years. There will certainly be much blaming of climate change for the stress on California’s electrical grid last week, but California’s resource adequacy, or RA, troubles are of human origin. At root, they are the result of a broken system that needs more radical organizational reform: shifting responsibility for RA from state agencies to the California Independent System Operator.
California’s current RA program was conceived in 2004 as a bandage to address the lack of adequate forward contracting during the 2000-2001 electricity crisis. Nearly 20 years later, and despite significant changes in California’s electricity industry, the original RA program design has remained stubbornly intact. The California Public Utilities Commission, which oversees the program, initiated proceedings in 2017 to restructure it, with a focus on addressing procurement challenges resulting from load migration. However, the main sources of California’s recent RA troubles are in planning and incentives and, underlying these, organizational responsibilities.
A robust RA program rests on six components: (1) reasonably accurate load forecasting; (2) a planning reserve margin that accurately accounts for generation and transmission outages, load forecast uncertainty, and in the future perhaps some solar and wind forecast uncertainty; (3) reasonably accurate capacity crediting for different resources; (4) penalties for non-compliance with RA procurement obligations and/or “backstop” procurement on behalf of load serving entities; (5) performance requirements to ensure that resources are available and can perform in real-time; and (6) market designs and operational processes that enable system operators to effectively commit and dispatch available generation when it is needed.
If a system experiences resource inadequacy, one of these components must have gone awry: load might have been under-forecasted; the planning reserve margin might have been wrong or applied in the wrong hours; capacity credits for different resources might have been inaccurate relative to real-time performance; load serving entities or the entity doing backstop procurement may not have procured adequate resources; resources might have been unavailable or non-dispatchable in real-time; or the system operator might not have been able to deploy resources in time even though adequate generation was available and dispatchable.
Though buried in a technical appendix (Appendix B), CAISO’s root cause analysis of the 2020 blackouts provides insight into what went wrong in 2020: the planning reserve margin was too low and was applied in the wrong hours, a significant amount of generation (especially gas and solar) was not available when needed due to capacity crediting and performance issues, programmatic demand response did not respond as expected, load was under-scheduled in the day-ahead market, and CAISO software issues allowed export scheduling even as the CAISO system was stressed.
The root cause analysis suggests that a principal reason why California’s RA program keeps coming up short is due to failures in planning and performance incentives, but these failures rest on deeper organizational flaws. California is unique among states with organized wholesale electricity markets in that the first three (or 3.5) RA components are managed by a state agency rather than by an ISO/RTO or by utilities. Keeping most of the RA program within the CPUC’s fence was a way to avoid it becoming FERC jurisdictional, but this has led to some problematic outcomes.
For one, the program’s organizational coordination has become increasingly complex over time, with overlapping and duplicative responsibilities between state agencies (the CPUC and California Energy Commission) and the system operator (CAISO) that obfuscate planning and blur incentives. Additionally, the CPUC, with limited resources and expertise, is now being tasked with upgrading an RA program that has grown more technically complex and that will likely need to be adjusted regularly over the next decade as new RA planning and implementation challenges arise due to continued changes in resource mix, electrification-driven changes in load, and a changing climate.
The most obvious fix to coordination and expertise challenges would be to move the RA program so that it falls under CAISO’s domain, aligning forward RA planning with short-term operational reliability and taking advantage of CAISO’s system data access, technical expertise, and its ability to regularly diagnose problems and problem solve. This move would also be consistent with CAISO’s potential longer-term evolution into a Western RTO and the regionalization of RA in the West. The CPUC dismissed this possibility as a solution to procurement issues, but the CPUC and state lawmakers should consider it as a means to simplify the RA program, address coordination challenges, and shore up expertise. Claims that a CAISO-run, and FERC overseen, RA program would hinder state energy policy lack imagination and appear to set up a false tradeoff between reliability and emission reductions.
Absent moving RA to the CAISO, meaningfully bolstering California’s RA program would require the CPUC to significantly up its game, investing in the in-house technical expertise (vis-à-vis external consultants) needed to be able to proactively identify and find solutions to emerging RA challenges. While there are valid concerns over the time needed for CAISO to absorb and adapt the state’s RA program, this upgrading of the CPUC’s capacity and the CPUC’s own improvements to the RA program would also take time.
Continuing to muddle through with periodic CAISO stage 3 emergencies, in addition to ongoing utility wildfire shutoffs, sends a signal to consumers that they should take electric reliability into their own hands. Whether and to what extent this is a desired result is an open question, but decentralization of reliability to customers should be an intentional outcome rather than an accident. A combination of customers voting with their feet and multi-year, centralized RA procurement by utilities, beginning in 2023, could lead to a third wave of stranded costs for the state — think ISO4 PURPA contracts plus nuclear cost overruns and the competition transition charge of the 1980s and 1990s, and high renewable power purchase agreement costs and the power charge indifference adjustment of the 2000s and 2010s.
California’s continued RA woes have lent credence to the notion that unreliable electricity systems are the new normal, that the transition to non-fossil fuel energy sources and climate change will inexorably lower historical levels of bulk system reliability. But this is simply not true. Any electricity system, even those transitioning to energy-limited resources and new temperature regimes, can be as reliable as forward planning, organizational coordination and expertise, performance incentives, and customer spending will support. The real questions are what level of bulk reliability we want and whether we are willing to invest in the institutional change needed to support it.