6. Are the features of the proposed tariffs the best way to design a critical peak pricing tariff for the future?
Although we choose not to implement a new default rate design for Summer 2005, we remain committed to modifying utility rates, through the normal rate design process, to send better signals to all bundled customers to accomplish our policy objectives. We are not convinced that in the long term voluntary rate programs will accomplish our objective of being able to reduce summer peak demand by 5% as set forth in the Energy Action Plan. Therefore, we take this opportunity to lay out the lessons we have learned from this proceeding and provide guidance for the rate design applications currently underway Application (A.) 04-06-024 (PG&E) and A.05-02-019 (SDG&E) and that will be filed shortly by SCE.
We learned in this proceeding that many large customers have facilities in multiple service territories, and often vest responsibility for development of energy management strategies statewide, rather than on a facility-specific basis. Therefore, the general rate design approach, event definition, and event triggers, should be as consistent as possible between service territories although the actual rate of each utility may vary based on its different cost structure. Statewide consistency in design will facilitate customer ability to provide demand response. In particular, we direct the ALJs in A.04-06-024 and A.05-02-019, to suspend the current schedule (if needed) and to require revised rate designs by the subject utilities to accomplish the objectives we set forth below. SCE should also prepare its next rate design application consistent with this approach.
6.1 Investment Signal to Customers
One of the shortfalls of the rates proposed in these applications was that they did not appear to be structured in a way that would motivate customers to reduce demand. For that reason, the ALJ asked all parties to provide their comments on their preference for a rate that includes a fourth higher priced time-of-use period every summer weekday (generally described as between 3:00 and 6:00 p.m.) as compared to a critical peak rate that is triggered the day-ahead for a limited number of days each summer. Most customer groups (although not all) indicated that a fourth fixed time-of-use period would send a stronger investment signal to customers to remove usage from hours targeted than a day-ahead on-call structure. For example, the CLECA witness testified that CLECA member companies would prefer refinement of the TOU pricing periods, potentially by narrowing the on-peak period or by adding a 4th super-peak period, over the implementation of an on-call CPP rate option. She noted that the predictability and regularity of pricing that is set in advance is most likely to permit customers to adapt their operations to new price signals. Others also suggested that instead of adding an additional TOU period, it might make sense to narrow the existing peak period to a shorter number of hours. However, several customers would prefer a rate structure where the pricing differs only when there is an actual emergency or reliability event, more like the structure explored in these applications.
A fourth TOU period allows customers to plan their investments more easily than a day-ahead on-call approach because with a fourth TOU period, the customer knows that if it makes an investment, the rate will be in effect every day and savings will occur every day. Thus the fourth TOU period rate differential leads to a sustained reduction in use from increased investment in efficiency improvements. With a day-ahead call, the customer is less likely to make significant investments in equipment to improve efficiency on a daily basis, because the likelihood of the program being called is unknown.40 Rather, the customer will evaluate whether or not to reduce load on a given day based on a comparison of the bill impacts of dropping load to the economic impact of a reduction in energy usage on core business. If the bill impacts are insufficient to outweigh the cost of disruption, then reduction in usage is unlikely. This balancing, of course, assumes that the customer is sufficiently educated about the rates in effect and usage patterns to be able to perform this calculation.
Price signals sent by a fourth TOU period result in an overall lowering of peak demand on all days, not just the most critical days, because the prices reflect average costs to provide energy during each time-of-use period, rather than actual market prices. Economic theory says that on non-critical peak days, this outcome is inefficient because it is only during a very narrow set of hours (many fewer than would be encompassed by a fourth TOU period) that there is actual shortage on the electric system and therefore we would be sending improper signals to customers that reducing their load everyday is important. Thus the Commission must decide, as it considers how to motivate customers to achieve demand response, whether its primary objective is a sustained lowering of peak demand or a temporary response to short-term shortage. Which of these objectives it places higher priority on should drive its decision of whether a fourth TOU period (or narrower peak period) or a day-ahead called program better meets the state's needs.
We believe that by modifying the approach to adopting revenue requirements for customers with loads greater than 200 kW as described below, we will be better able to accomplish both objectives. We are intrigued by the information provided by CCEA about the significant inertia that customers face when changing their rate schedule, even when a different schedule would provide a lower bill, often as a result of high switching costs. This inertia points to the need for aggressive education efforts to position customers to make the right decisions about whether to remain on a given tariff. SVMG suggests that we place customers on a critical peak pricing tariff as a default, with the ability to convert back to their current TOU tariff.41 We will adopt this approach, with the TOU rate modified slightly, as described below, with revenue requirements calculated as described below.
In addition, instead of establishing a fourth time-of-use period, we direct the utilities to explore narrowing the current peak period to cover the hours of 2:00 p.m. to 6:00 p.m. The information provided by SCE in Exhibit 16 and the CAISO's Opening Brief convince us that this narrower peak period will generally capture the peak system loads without significant risk of peak shifting. For example, "SCE notes that the average change in load between its peak hour of 3-4 p.m. and non-peak hour of 1-2 p.m. during the ten highest load days in each of several years was approximately 600 to 700 MW. Therefore, for the peak to shift to the 1-2 p.m., for example, the amount of load migration attributable to CPP would have to be greater than 600-700 MW." (CAISO Opening Brief, p. 6.) Narrowing the on-peak TOU period will likely increase the peak period price slightly as compared to the current peak period price, but decrease partial- and off-peak rates. This slight change is likely to better signal the price differential between the hours that represent the highest load, and other hours. By narrowing the peak period, the price differential between the peak and partial-peak TOU rates will increase, sending a stronger investment signal than adding a fourth TOU period. At the same time, the TOU rate provides a stable and predictable price for those who prefer certainty over the critical peak pricing rate.42
6.2 Establishing Revenue Requirement Upon Which to Design Rates
The key issue for establishing a day ahead critical peak pricing rate for customers 200 kW and above is identifying the correct revenue requirement to collect. In these applications we required the utilities to file rates that were revenue neutral based on existing revenue requirements. We required them to propose a "hedge premium" for customers wishing to remain on a traditional TOU rate. As a result of these requirements, there were structural winners and losers from a bill impact perspective. Because of the designs, customers who did have increased bills would not have seen significant reductions to those bills, even if they made significant reductions in their usage. Because the utilities could not count on reductions in usage from the rates, procurement costs were not reduced.
The revenue requirement used to set current TOU rates includes the costs to serve forecasted load which inherently includes both the expected costs to meet load during normal operating conditions and costs to meet load during higher peak periods. Establishing a revenue requirement that incorporates these higher, more critical period costs is the "hedging" the utility performs as a matter of course in order to ensure that it will recover its revenue requirement under expected load conditions. Therefore, we agree with parties that as long as the revenue requirement used to establish TOU rates includes the costs to meet load during these higher, critical peak periods, no additional hedging premium should be required if a customer chooses not to participate on the critical peak pricing tariff.
In order to send the correct pricing signal to customers under a critical peak pricing rate, the critical peak period costs need to be unbundled from the revenue requirement and recovered from customers only when a critical peak event is called. The Commission should calculate non-critical peak rates based on an adopted revenue requirement for all hours that reflects expected costs in a year with no critical peak events. Separately, the Commission should establish the rate for the critical peak period to reflect the utility's anticipated marginal cost to procure for power for those customers during critical peak periods.43
In creating new CPP rates in this manner, we shift procurement cost risk to customers who remain on the CPP rate, so we need to establish the base level of revenue requirement that would be recovered in rates assuming no CPP events are called. The no-call revenue requirement would not include certain generation costs, like high spot market prices induced by short term supply-demand imbalances. More stressed market conditions or system supply-demand imbalances would impose larger procurement costs per unit and would be collected in the CPP event price. Clearly identifying the generation related revenue requirements during "orderly" market conditions from incremental generation costs recoverable during a CPP event requires much more precise allocation of generation procurement costs than rate designs have used in the past, but it does not prejudge whether for rate design purposes, the revenue should be recovered via demand or energy charges - that is left to the parties to develop in the rate design phase of these proceedings.
This approach is similar to the approach taken by ORA in that it would recover generation marginal capacity costs in the energy charge during critical peak periods. Our approach differs from ORA's though by separately establishing a revenue requirement for non-critical peak hours assuming no critical peak events occur and setting rates to collect that revenue requirement. By doing so, cost recovery of the revenue requirement is not tied to the number of events that occur as it is under ORA's primary approach. ORA points out in comments on the proposed decision that its alternate rate design approach was not addressed and that it believes its alternate approach addresses many of these concerns.
By removing the costs to meet the higher critical peak loads from revenue requirement allocated to 200 kW and above customers, and charging for those costs only during the critical peak, customers receive a stronger price signal that usage during that period is costly than under a standard TOU rate and will have additional motivation to reduce load during those critical peaks. This approach would allow the utility to fully recover its necessary revenue requirement and avoid procurement costs on peak as customers modify their usage in response to the rates. By calculating rates in this manner, we do not need to establish any particular crediting mechanism for when an event is called, since the revenue requirement being collected from customers on the critical peak pricing rates during non-event hours has already excluded the costs associated with meeting the utility's critical peak needs. Because customers have the option to convert back to standard TOU rate without additional cost, there is no need to exclude any customer group from default tariff applicability.44
We do not expect that this approach will require different marginal cost studies or revenue allocation to classes than would normally be performed. Instead, how the rates are designed to recover the revenue allocated to that class, how to extract the critical peak costs to determine the "no-call" revenue requirement, and the proper critical peak rate will be the incremental work required to establish critical peak pricing tariffs.
6.3 Event Triggers
Several different event triggers were proposed for Summer 2005 proposed rates. The primary recommendation was to use an ISO Alert to trigger an event. SDG&E also proposed a combination of temperature and system load as well as local grid emergencies to trigger a CPP event. However, SDG&E's proposed temperature/load trigger was designed to accomplish 12 calls in the summer in order to limit revenue undercollections from its rate design.
The CAISO defines an "Alert Notice," based on situation following the close of the Day Ahead Market, which closes at 1:00 p.m., as:
A notice issued by the ISO when the operating requirements of the ISO Controlled Grid are marginal because of Demand exceeding forecast, loss of major generation or loss of transmission capacity that has curtailed imports into the ISO Control Area, or if the Day Ahead Market is short of scheduled Energy and Ancillary services for the ISO Control Area.
In establishing the revenue requirement for the critical peak pricing rate, it is clear that the Commission will need to determine at what level of load costs should be allocated to normal operation (non-CPP hours) vs. critical peak (CPP hours). Based on that assessment, revenue will be allocated to non-CPP hours and CPP hours and rates will be set accordingly. In theory, the load level relied on to perform this allocation should have a relationship to the demand level that the utility must procure reserves for as part of the Commission's resource adequacy requirements. Given the approach we have described for setting revenue requirement to calculate CPP rates, we agree with ORA that the event trigger should bear a relationship to the load levels assumed in rate design and for resource adequacy. We direct each utility in its proposed rate design to designate a specific MW load level for its system that will trigger a CPP event call, consistent with the load level used in its rate design and resource adequacy requirements. For example, this MW level could be set as a specific MW amount or as the difference between the long term and day-ahead forecast load. When the day-ahead load forecast reaches this level, the utility's CPP price will be triggered for the following day. As proposed in the instant applications, notification should be effected by 3:00 p.m. the day ahead.
6.4 Limit on Events
We are convinced by numerous parities that more than four hours is too long for calling a critical event. As described above for our approach to narrowing the TOU period to the 2:00 to 6:00 p.m. time frame, we are convinced that this four hour period will adequately cover the critical peak. We will not specify at this time the number of events that should be called each summer. Instead, in each rate design proceeding, the number of events should be determined based on the forecasts used to allocate revenue to the critical peak. If the forecasts show that there will be five events in the next summer, and revenue is allocated accordingly, then the limit on events should be five. If the forecast shows 12 events, then the event limit should likewise be 12.
6.5 Non-Firm Conversion to BIP
In the short term, we concur that eliminating existing non-firm and interruptible rates is inappropriate. However, given that the BIP reservation payment was designed to provide the same bill impact to customers as the non-firm rates, we see no practical reason from a customer standpoint that the customer would not participate in BIP but would participate on the rate program provided that the customer truly has load it can curtail on short notice. In each utility's rate design proceeding, we will review whether the reservation payment adopted for BIP provides, on average, a consistent bill impact as the non-firm rate discount. In the rate design proceedings, we will ensure that the reservation payment is at a level sufficient to make customers, on average, neutral to the change to a reservation payment program vs. rate discount.45
Contrary to what many commenters say, converting to BIP does not eliminate non-firm programs. BIP is a non-firm program designed to provide the same bill effect to customers as the current non-firm tariffs. The fact that less customers have enrolled on BIP than the tariffs appears to be a function of most non-firm load already being served on the tariffs rather than an inherent problem with the BIP design or how responsive customers are who sign up for BIP, given that it has the same triggering and penalty criteria as well as cost basis as non-firm rates. Several commenters assert that there is no basis for converting customers to BIP from the non-firm rate programs, but neglect to mention the fact that for many years now, including as far back as 1992, the Commission has expressed discomfort with the structure of the non-firm rate program, and indicated its intent to move customers to another non-firm program structure. (See for example, D.92-11-049.) Therefore, we make no change to our plan to convert the current non-firm rate programs to the BIP structure over the three year GRC cycles.
For first year CPP is available, we should retain the non-firm rate option rather than immediately migrating customers to either CPP or TOU rate. In second year of the GRC cycle, half of the non-firm rate discount should be removed from the rate and converted into BIP reservation payment. In the third year of the GRC cycle, customers would convert either to CPP or TOU rate with the entire non-firm discount provided through the BIP reservation payment.46
Participation in BIP while taking service under the critical peak pricing rate should be allowed because under BIP, the customer must commit to a particular firm load reduction level, whereas under the critical peak pricing rate, the customer is not required to reduce its load, although it has the incentive to do so. A BIP event may or may not coincide with a CPP event so there is value in the customer being positioned to respond to either. In fact, if a non-firm customer is on a CPP rate and receives the day-ahead call, they will be even better positioned to respond to a potential emergency the next day, having received a day-ahead notice that supplies were tight. That customer can begin to take steps to adjust its load during the following day in anticipation of the possibility of an emergency call. In any event, should the customer that is currently on a non-firm rate wish to not be exposed to the critical peak pricing rate, it will be able to select the traditional TOU rate with reservation payments for its non-firm load provided through the BIP rider.
6.6 Customer Education Efforts
Once rates are adopted in the rate design proceedings, the utilities will clearly need to provide educational materials to customers to inform them of the upcoming change to the default CPP, the revised TOU peak period and resulting rates, and their option to stay with CPP or switch back to a traditional TOU. In order to ensure that customers have sufficient information to determine whether to remain on the critical peak pricing rate or switch to the traditional TOU rate, once the revised rates are adopted, the utilities shall provide each customer with
load over 200 kW with comparative bills under each rate using that customer's prior summer load data at least two months prior to the start of the summer period.
