The Complex Landscape of Net Metering Reform in California: Why an Installed Capacity Charge?


Rooftop solar in California has grown from an infant industry two decades ago to a 10-gigawatt resource that contributes significantly to customer and electric system needs today. The state is blessed with ample sunshine in many regions, and its urgency on this and other clean-energy innovations was born out of the energy crisis in 2000 and 2001, as well as the need to address climate change and improve public health. But a proposed range of reforms to net metering for residential rooftop solar has prompted debate about the future of that important market segment, as well as the broader trajectory of state energy policy.

Full retail rate net energy metering with monthly netting (“traditional NEM”) — the method initially used in nearly every U.S. state, including California — is easy to implement and understand. Traditional NEM, which opted for simplicity over precision and was intended to kick-start an infant industry, might not be ideal from the perspectives of efficiency or fairness, but the extent of those problems depends on the level and design of retail rates, as well as the overall resource mix, load patterns, and customer solar adoption levels.

The maturation of the solar industry and the modernization of the electric grid make an evolution from traditional NEM to more efficient and sophisticated rate designs both possible and desirable. The exact speed of that transition, and exactly which direction to head, is not necessarily obvious and involves tradeoffs that policymakers and stakeholders should understand and debate fully. RAP recently released a report for the Michigan Public Service Commission that seeks to help Michigan regulators and policymakers understand these tradeoffs as they consider improving their rate designs for distributed energy resources (DERs), as well as the broader tariff specifications and DER program structures.

In 2016, California took a substantial step away from traditional NEM to “NEM 2.0.” That step required solar customers to be on a time-of-use (TOU) rate along with other reforms. All else equal, those reforms were designed to reduce the level of compensation for a new solar customer. Residential solar installations dropped modestly in 2017 but the pace of residential installations has grown steadily since then. On Dec. 13, the California Public Utility Commission’s proposed decision on NEM reform signaled a new stage in this debate. The proposed decision found there has been a significant cost shift from new residential solar customers to non-participating customers under NEM 2.0 and includes a wide range of additional reforms to rate design for new solar customers.

These proposed reforms are well intended, and many have substantial merit — including more differentiated time-of-use rates, value-based export credits, encouragement of solar plus storage, and a range of equity measures. More controversially, the proposed decision included a new “grid participation charge,” which is structured as a $/kW monthly fee on installed capacity for new residential solar customers. A new “market transition credit,” structured as a $/kW monthly credit on installed capacity, would partially offset the grid participation charge temporarily for customers of two of the three utilities, at a level designed to achieve a 10-year payback period for new solar installations.

The basic concept of the grid participation charge is not new. It has been debated before and can be generically described as an installed capacity charge. An installed capacity charge is not tied to any reasonable metric of the size of the customer or their impact on or usage of the grid, but is primarily a way to spread certain categories of costs. The direct incentive provided to customers by an installed capacity charge, all else equal, is to install fewer kW of the resources covered by such a charge. In this case, the market transition credit is attempting to partially counteract this effect by ensuring a reasonable payback period for new customers. In that context, it is important to consider whether such a spreadsheet analysis is reasonably accurate and whether there are more qualitative considerations, such as the complexity of the newly proposed rates, that could further hinder adoption.

New York is implementing a similar rate structure for new residential rooftop customers starting in January 2022, which has been labeled a “customer benefit contribution” charge. This charge is designed to cover a smaller set of program costs, namely energy efficiency and clean energy programs as well as low-income discounts. New York’s situation is different: Without its advanced metering infrastructure fully deployed, more sophisticated rates are difficult to implement. The New York Department of Public Service previously estimated that this new charge will be between $0.69 and $1.09 per kW (direct current) of installed capacity, depending on the utility. Final rates, going into effect next month, were recently filed and range from $0.72/kW to $1.33/kW. This means that the size of California’s proposed $8/kW charge for installed capacity is unprecedented, even if it is effectively reduced to $4 or $6 per kW by the market transition credit in the first year of this new structure for two of the utilities.

Jumping into such unexplored territory comes with risks. History shows that rate designs like this spur customers and vendors to find innovative workarounds. For example, customers may try to avoid this new charge by fully disconnecting from the grid. While unlikely for many customers today, it may become a more popular option if costs for storage continue to fall dramatically. Alternatively, customers may be able to set up their solar and storage systems to avoid exporting to the grid and thus avoid any need to notify the utility and be exposed to the grid participation charge. Solar installers have experience with both of these options in Hawaii. Such behavior is likely suboptimal from the societal perspective, and would likely cause significant cost-shifting to other customers. The CPUC could try to prevent these reactions, but this could just push such behavior further underground. Other unintended consequences will also likely arise.

Of course, the proposal for the grid participation charge cannot be evaluated in a vacuum — without a comparison to the relevant alternatives. Regulators should be guided by this principle for efficient rate design (which takes on increasing importance as customers have more options to invest in generation, storage, and load controls): Rate design should make the choices a customer makes to optimize their own bill consistent with the choices that would minimize system costs. Continued reforms to residential time-varying rates is an important option we will explore more fully in a follow-up blog. For now, we note several other rate design options that can address cost shifts from solar distributed generation in a manner similar to the grid participation charge. These should be considered instead of an installed capacity charge, or in combination with a small installed capacity charge:

  • Reasonably sized customer charges ($5 for low-income customers and multifamily building residents, $10 for everyone else);
  • A distribution flow charge on both inflows and outflows;
  • A demand-based or connected load charge to cover line transformer and other site infrastructure costs (approximately $1-2/kW); or
  • A higher minimum bill.

RAP has long advised against major reliance on large customer charges and demand charges for efficiency and equity reasons. In many states and for many customer classes, these charges are far too high. But within proper limits, these options all have a stronger cost causation argument than an installed capacity charge. The distribution flow charge, defined as a cents-per-kWh rate on both imports from and exports to the grid, has not been implemented by any state utility commission, but is a concept that RAP put forward in our report for the Michigan PSC as well as a 2013 report on distributed generation tariff design. Demand charges can also be more difficult for customers to understand and manage than other types of rates, although this can be mitigated with education efforts and data provision.

Ideally, these options would be considered as part of broader reforms to rate design for residential customers. Not all residential customers need to be exposed to more complex rates, however. Another idea discussed in the Michigan report is segmenting the residential class into an “advanced” category and a “basic” category. That way, a broad swath of the residential class could be moved onto more sophisticated and efficient rates without risking adverse impacts to low-income and low-usage customers. Broader residential rate design reform was not a focus of the current California proceeding, but it should be considered as the full CPUC takes up the proposed decision.

The grid participation charge, and other alternatives to address similar issues, is not the only important public policy concern as California considers reforms to its DER program. RAP will follow up with additional blogs in this area in January, discussing issues such as the structure of time-of-use rates, the newly proposed value-based export credit system, community solar and locational value.