Design Matters: Unpacking Xcel’s Current Rates Proposal

Across the electricity system, utilities face a reckoning between traditional rate designs and disruptive trends. As distributed solar raises concerns about cost shifts onto a shrinking customer base, and distributed energy resources (DERs) grow ever more capable, the electricity rates charged to customers are central to industry-wide efforts to rethink the utility business model. RMI has advocated more sophisticated rates to send more accurate price signals that reflect the underlying benefits and costs of customers’ grid use while recovering utility costs. Now, Xcel Energy in Colorado has put forth its vision for a sophisticated rate future, offering an opportunity to evaluate it against desired outcomes.

Xcel’s Phase II rate case before the Colorado Public Utilities Commission (PUC) describes the utility’s long-term intention to move from its existing two-part mass-market rate to a five-part rate. Xcel proposes three-part rates for most residential customers in the near term, including expanded use of fixed and demand charges. The proposal’s increase in rate sophistication is commendable, but it risks adverse impacts if customers aren’t equipped to respond to complicated price signals. The proposal will also affect customer-sited solar, which can reduce what is owed in proposed demand charges, but may provide lower bill savings than today’s rates due to lower kWh energy prices. Hence, the bottom-line impact is unclear.

Xcel puts rate design at the center of the modernized utility

It’s not Hawaii, but Colorado’s solar market is growing, supported by Xcel’s net energy metering (NEM) policy. Xcel, like utilities in Nevada and California, is voicing concerns about cost shifts: as NEM customers buy less grid energy, non-NEM customers’ share of grid costs grows. Xcel has concluded that residential solar contributed to an under-recovery of $4 million in fixed costs over 2014 and 2015. Such claims are controversial and depend on how one calculates the benefits and costs of solar photovoltaics (PV); in a similar case in Nevada, for instance, a recent study concludes that PV may provide net benefits of $7–14 million to all customers.

Xcel’s response to these concerns is “Our Energy Future,” a multipronged strategy of which rate design proposals are one piece. Xcel states that the plan would help reclaim “reasonable opportunity for the utility to recover its approved revenue requirement.”

In particular, Xcel seeks to recover a larger portion of its costs through a combination of increased fixed charges and new demand charges. Xcel believes these incremental steps will help realize a sweeping revision to conventional rate design—a mass-market rate with five key components:

1. A basic time-of-use (TOU) charge to recover fuel and purchased-energy costs
2. An ex ante coincident-peak demand charge (based on a predetermined peak period) to recover fixed generation and transmission (G&T) costs
3. A noncoincident demand charge to recover distribution costs
4. A flat energy charge to recover nonfuel variable energy costs
5. A customer charge to recover customer-specific costs

Such a rate requires advanced metering and communication technology, which Xcel has not built out everywhere. Also, customer experience could suffer from suddenly moving to complex pricing schemes. Accordingly, Xcel proposes intermediate rate options—to take effect in 2017—which Xcel argues will enhance cost recovery and provide a new level of choice. Two of the proposed rates for residential customers are outlined below.

A revised Schedule R would be the default rate for most mass-market customers, but changed from a two-part to a three-part tariff. This rate includes basic service and facilities charges (i.e., fixed charges) that cover customer-related metering and billing costs, a seasonally differentiated energy charge, and a “grid-use charge” to cover other distribution costs that have some relationship to total energy demand. The grid-use charge would be levied on the rolling average of customers’ 12-month net consumption. This is similar to a service charge levied by Burbank Water and Power.

Existing NEM customers on Schedule R, and those who sign up for NEM by year-end, can choose to stay on the old “Optional Energy Charges” (see Table 1) and avoid paying any grid-use charges.

A new Schedule RD-TOU would be a model for future Xcel rates. That proposal includes a fixed customer charge, a flat energy charge, a grid-use charge like that in schedule R, and a seasonally differentiated ex ante coincident-peak demand charge. A limited number of customers would volunteer to test the rate design, which would inform future rates. Despite its name, the RD-TOU rate doesn’t appear to include conventional volumetric time-based rates. In addition, the extraordinarily low energy charge would provide reduced incentive for energy efficiency.

Table 1
Source: RMI analysis of Xcel Energy filings; some numbers rounded

What can we expect if the Xcel proposals are adopted?

Complex proposals like Xcel’s, with sometimes confusing terminology, need a common language and—more importantly—for their likely impact to be carefully evaluated. RMI’s report, A Review of Alternative Rate Designs, released last month, helps do just that. The report provides a meta-analysis of dozens of empirical studies, expert reports, and specific utility rates, and identifies key design decisions for electricity rates of the future. Specifically, the report identifies a set of dimensions that can support regulatory review beyond cost recovery, including goals for peak reduction, energy conservation, customer acceptance, and overall system efficiency.

There are too many components within Xcel’s proposal to thoroughly evaluate them here. However, our research provides some insight on these proposals, and highlights important questions. (However we note at the outset that limited industry experience and a lack of robust empirical results makes assessments of demand charges uncertain; there is more experience with time-based rates, supporting wider adoption of those.)

Demand charges – The rates proposed by Xcel move toward new demand charges and higher fixed fees to recover grid costs. Although TOU rates would be part of the five-part rate, volumetric time-based rates aren’t a meaningful part of this proposal.

“Extensive demand charges” such as those in the RD-TOU proposal (which would include costs of generation and transmission infrastructure) are hotly debated. Proponents argue that each customer’s demand directly contribute to system infrastructure requirements. Opponents respond that any individual’s peak demand is far removed from bulk system costs, which can be recovered through volumetric rates just as effectively.

Little data exists showing that demand charges can reduce peak demand—a worthwhile objective for any rate proposal. We found time-based rates can effectively reduce customer peak and energy consumption, without negatively affecting customer acceptance. There is less evidence for the efficacy of demand charges to achieve any goal beyond cost recovery.

Measurement intervals – The proposed grid use charges are essentially a demand charge with an extremely long measurement interval. By pegging the current month’s charge to the average of the past 12 months, they would also serve as a ratchet mechanism and dampen customers’ incentive to reduce consumption, as any change in current energy use would have a limited impact on the next bill. However, a long measurement interval can reduce bill volatility and stabilize utility revenues. These mechanisms have been used for larger commercial and industrial customers but are relatively untested for mass-market customers.

Xcel’s more conventional demand charge under Schedule RD-TOU would be based on a customer’s 60-minute average demand over a predefined four-hour window. This is an ex ante peak-coincident demand charge, which can more effectively allow customers to adjust their usage to reduce their bill. There is limited experience with how this type of signal would affect residential customers’ responses or total energy use, but the rate is an improvement on noncoincident demand charges that have little relation to system costs.

Enabling Technology – Xcel suggests that enabling technology (hardware and software that can support customers to effectively respond to price signals) will enable response to these rates in the future. Enabling technology is essential to making alternative rates effective. However, specific technology, program design, and marketing choices will determine whether technology delivers on its promise. To make programs successful, utilities should incorporate enabling technology based on program designs elsewhere. For example, Black Hills Power in South Dakota actively markets its “demand controller” technology to customers on its demand charge program.

What might this proposal mean for solar customers in Colorado?

Last year’s Colorado PUC ruling determined NEM will remain in place. But the economics for new solar customers would be impacted by the rates proposed by Xcel—it’s just not clear exactly how.

Beginning in 2017, new NEM customers will have to decide between Schedule R (where they pay grid-use charges and lower energy charges) and Schedule RD-TOU (where they pay both grid-use and demand charges). In either case, the energy charge for net consumption will be 25–27% lower than under current rates, thereby reducing solar’s traditional bill-reduction benefit. But because the tiered grid-use charge and demand charges would be assessed against consumption net of solar, rooftop solar can help to reduce these charges.

Across Xcel’s proposal, these countervailing affects for solar and nonsolar customers call for an exhaustive analysis of bill impacts. Xcel claims that approximately two-thirds of customers would see bills decrease under the proposal. But, as the RMI rate design report demonstrates, the devil is in the details. Specific design choices will determine the effectiveness of these rate reforms, including the ability of customers to manage their bills, the effect on specific customer segments, including low-income customers, and impacts on the solar market.

The Xcel proposals are notable for their relative sophistication and the potential for a new relationship between customers and their electricity supply. But proposals like this point to the criticality of thoroughly reviewing their anticipated impacts. In that light, a more detailed analysis is needed to understand how the economics of solar and nonsolar customers would be impacted by these new rates, and the full costs and benefits of rooftop solar in Colorado.