In 1978, Congress enacted the Public Utility Regulatory Policies Act ("PURPA"). PURPA made several changes to energy regulation, particularly to how utilities would interact with small independent energy producers. PURPA charges the Federal Energy Regulatory Commission ("FERC") with enacting implementing regulations. FERC's regulations, in turn, allow state regulatory agencies to determine exactly how they will comply with PURPA and FERC's regulations. The relevant state agency here is the California Public Utilities Commission ("CPUC").

*932Californians for Renewable Energy ("CARE") and two of its members, Michael E. Boyd and Robert Sarvey, are small-scale solar producers. They allege that CPUC's programs do not comply with PURPA. Specifically, they argue that CPUC has incorrectly defined the amount that PURPA requires utilities to pay qualifying facilities ("QFs"). CARE argues that PURPA also allows equitable damages and attorney fees.

The district court dismissed CARE's claims for equitable damages and attorney fees and entered summary judgment for CPUC on CARE's PURPA challenges. We affirm in part and reverse in part.

I. FACTUAL AND PROCEDURAL BACKGROUND

A. Statutory Background

Congress enacted PURPA "to encourage the development of cogeneration and small power production facilities, and thus to reduce American dependence on fossil fuels by promoting increased energy efficiency." Indep. Energy Producers Ass'n, Inc. v. Cal. Pub. Utils. Comm'n ("IEP "), 36 F.3d 848, 850 (9th Cir. 1994).

To achieve this objective, Congress sought to eliminate two significant barriers to the development of alternative energy sources: (1) the reluctance of traditional electric utilities to purchase power from and sell power to non-traditional facilities, and (2) the financial burdens imposed upon alternative energy sources by state and federal utility authorities.

Id.

PURPA created a new category of energy producers: qualifying facilities. QFs can be either "small power production facilit[ies] or "cogeneration facilit[ies]." 18 CFR §§ 292.201 & 292.203. FERC has authority to define the requirements for being a QF. 16 U.S.C. §§ 796(17)(C) & (18)(B).

To address the barriers facing QFs, PURPA required utilities to purchase electricity from QFs, i.e. the mandatory purchase requirement, 16 U.S.C. § 824a-3(a), and to pay QFs rates that "shall be just and reasonable to the electric consumers of the electric utility and in the public interest." 16 U.S.C. § 824a-3(b). Utilities must compensate QFs at a rate equal to the utility's "avoided cost." 18 CFR § 292.304(d). "Avoided cost" is "the incremental cost[ ] to an electric utility of electric energy or capacity or both which, but for the purchase from the qualifying facility or qualifying facilities, such utility would generate itself or purchase from another source." 18 C.F.R. § 292.101(6).

State regulatory agencies have the responsibility of calculating avoided cost, but FERC has set forth factors that states should consider. 18 C.F.R. § 292.304(e). Those factors are:

(1) the utility's system cost data;

(2) the terms of any contract including the duration of the obligation;

(3) the availability of capacity or energy from a QF during the system daily and seasonal peak periods;

(4) the relationship of the availability of energy or capacity from the QF to the ability of the electric utility to avoid costs; and

(5) the costs or savings resulting from variations in line losses from those that would have existed in the absence of purchases from the QF.

Cal. Pub. Util. Comm'n ("CPUC "), 133 FERC ¶ 61,059, 61,265, 2010 WL 4144227 (2010). "Avoided cost rates may also 'differentiate among qualifying facilities using various technologies on the basis of the supply characteristics of the different technologies.' " Id. at ¶ 61,265-66 (quoting 18 C.F.R. § 292.304(c)(3)(ii) ). Avoided cost can also include the capacity costs that the *933utility avoids by purchasing electricity from QFs. CPUC , at ¶ 26.

Congress changed this statutory scheme in 2005 with the Energy Policy Act ("EPAct"). With EPAct, Congress acknowledged that QFs no longer faced the same barriers that prompted PURPA. EPAct thus eliminated the must-purchase obligations for any QF that FERC determined had "nondiscriminatory access to" particular markets as specified in 16 U.S.C. § 824a-3(m). In 2011, FERC released California utilities from PURPA's mandatory purchase obligations for QFs over 20 MW. Pac. Gas and Elec. Co. , 135 FERC ¶ 61234, 62305 (2011). FERC established a presumption that the mandatory purchase obligation would apply for QFs 20 MW or smaller unless the utility showed that "each small QF ..., in fact, has nondiscriminatory access to the market." New PURPA Section 210(m) Regulations Available to Small Power Production and Cogeneration Facilities ("Order 668"), 71 Fed. Reg. 64342, 64363 (Oct. 20, 2006). The facilities that CARE represents produce less than 20 MW of energy.

In addition to mandatory purchase requirements, PURPA requires utilities to connect QFs to the power grid. The interconnection requirement goes hand-in-hand with the mandatory purchase requirement for "[n]o purchase or sale can be completed without an interconnection between the buyer and seller." Am. Paper Institute, Inc. v. Am. Elec. Power Serv. Corp ., 461 U.S. 402, 418, 103 S.Ct. 1921, 76 L.Ed.2d 22 (1983). Using its authority under PURPA, FERC promulgated a rule requiring that "any electric utility shall make such interconnection with any qualifying facility as may be necessary to accomplish purchases or sales under [PURPA]." 18 C.F.R. § 292.303(c)(1). FERC's rule also specifies that "[e]ach qualifying facility shall be obligated to pay any interconnection costs which the State regulatory authority ... may assess against the qualifying facility on a nondiscriminatory basis with respect to other customers with similar load characteristics." 18 C.F.R. § 292.306(a).

B. The Challenged CPUC Programs

In the 1980s, CPUC required utilities to offer one of four standard contracts if a QF requested one. These contracts "differ[ed] primarily in the length of the contract, the availability of capacity and energy from a QF, and the avoided cost rate payments corresponding to such availability." IEP , 36 F.3d at 852. This program was successful but did not "accurately reflect[ ] the avoided cost of ... utilities." Solutions for Utilities, Inc. v. Cal. Pub. Utilities Comm. , CV 11-04975 SJO (JCGx), 2016 WL 7613906, at *5 (C.D. Cal. Dec. 28, 2016). CPUC discontinued using these contracts in the mid-1980s because of "QF oversubscription." Id. The elimination of these contracts and the subsequent search for a better mechanism for compensating QFs sparked years of litigation. Rather than use long-term pricing, CPUC moved to using short-run pricing. State legislation in 1996 "set[ ] forth certain elements to be included in setting [short-term avoided cost ('SRAC') ]." Order Instituting Rulemaking to Promote Policy, Program Coordination and Integration in Electric Utility Resource Planning, No. D.07-09-040, 2007 WL 2872674, at *9 (Cal. P.U.C. Sept. 20, 2007). Disputes, however, continued.

This situation was finally resolved in 2010 with the Qualifying Facility and Combined Heat and Power ("CHP") Program Settlement ("QF Settlement"). Solutions for Utilities, Inc. , 2016 WL 7613906, at *6. Among other things, the QF Settlement established four standard contracts. Id. One of these standard contracts was designed specifically for QFs with capacity of 20 MW or less. Id. Any QF 20 MW or smaller may avail itself of this contract, regardless of where the QF sources its energy. This contract sets the price paid to *934QFs based on both capacity and energy. The price for capacity is a fixed rate while the price for energy is variable, based on the Short Run Avoided Cost ("SRAC").

"Energy costs are the variable costs associated with the production of electric energy (kilowatt-hours). They represent the cost of fuel, and some operating and maintenance expenses. Capacity costs are the costs associated with providing the capability to deliver energy; they consist primarily of the capital costs of facilities."

Small Power Production and Cogeneration Facilities; Regulations Implementing Section 210 of PURPA, ("Order 69") 45 Fed. Reg. 12,214, 12,216 (Feb. 25, 1980).

Separate from the QF Settlement, the California legislature, through Assembly Bill 1613, created the Combined Heat and Power Facilities Program on January 1, 2008. Solutions for Utilities, Inc. , 2016 WL 7613906, at *6. The CHP Program applies to CHP facilities with capacities under 20 MW. Id. Under this law, CPUC set up a different program for compensating CHPs based "on the Market Price Referent ('MPR'), which is defined as the cost to design, build, and operate a 500 MW Combined cycle natural gas turbine generator ('CCGT')." Id.

CPUC also operates the Feed-in-Tariff ("FiT") or Renewable Market Adjusting Tariff ("Re-MAT") program. This program applies to renewable generators with capacities of 3 MW or less. Id. at 7. Under this program, utilities must purchase electricity at the program-specified rates "until the [utility] meets its proportionate share of a statewide cap of 750 [MWs] cumulative rated generation capacity." Id. The Re-MAT price is calculated using three pricing values. First, the Re-MAT takes "the weighted average contract price of [three California utility's] highest priced executed contract resulting from the CPUC's auction held in November 2011 for three different product types." Id. Second, Re-MAT uses "a two-month price adjustment 'based on the market response.' " Id. Finally, the participating power producer receives "a 'time-of-delivery adjustment' based on the generator's actual energy delivery profile and the individual utility's time-of-delivery factors." Id. As CARE describes it, CPUC assumes that market bids take account of capacity costs.

The last CPUC program at issue is the Net Energy Metering ("NEM") Program. The NEM Program was established by state statute, Assembly Bill 920, and took effect in January 2011. Solutions for Utilities, Inc. , 2016 WL 7613906, at *7. This program is limited to consumers with capacity of 1 MW or less. Id. The NEM Program calculates how much electricity a consumer uses and how much electricity a consumer generates over a twelve-month period. If the consumer generates more electricity than it uses, then the excess electricity goes back into the electrical grid. Id. The utility pays the consumer for this electricity based on the default load aggregation point ("DLAP") price. DLAP is "an hourly day-ahead electricity market price," in other words, what "the utility is paying one day out in the marketplace." Id. DLAP does not include capacity costs, even as defined by CPUC.

California has also enacted a Renewables Portfolio Standard ("RPS"). The first RPS, enacted in 2002, required utilities to source 33% of their electricity from renewable sources by the end of 2020. Those standards have since been increased to require 50% of a utility's electricity to be from renewable sources by 2030. CPUC represents that "CPUC-regulated utilities have met their 2020 targets and are on track to reach their [2030] targets."1 Most *935of these goals have been met by purchasing energy from producers with capacity over 20 MW.

II. Procedural Background

A. CARE v. CPUC I

CARE and Solutions for Utilities Inc. ("SFUI") sued CPUC and Southern California Edison Company ("SCE") in 2011. That suit alleged violations of PURPA and violations of § 1983 based on allegations of suppressing SFUI's and CARE's First Amendment rights. The district court dismissed the § 1983 claims and CARE's PURPA violation claim but left SFUI's PURPA claim. The district court also entered summary judgment for CPUC and SCE, finding that SFUI did not have standing to bring its PURPA claim. CARE appealed. This Court affirmed dismissal of the § 1983 claims but reversed and remanded on CARE's PURPA claim, finding that the CARE Plaintiffs had met PURPA's administrative exhaustion requirement. Solutions for Utilities, Inc. , 2016 WL 7613906, at *2.

B. The Current Action

CARE moved for leave to file a fourth amended complaint on March 8, 2016. The district court denied CARE's motion for leave to file without prejudice. In that order, the district court found that CARE could not amend its complaint to assert a claim for equitable damages and attorney fees. CARE then filed an amended complaint on April 14, 2016. CPUC moved for summary judgment. On December 28, 2016, the district court granted summary judgment for CPUC on all claims. This appeal followed.

III. JURISDICTION AND STANDARD OF REVIEW

The district court denied CARE's Motion for Leave to File Fourth Amended Complaint. In that order, the district court found that damages and attorney fees were not available under PURPA. This Court reviews a "denial of a motion to amend a complaint ... for an abuse of discretion." Chodos v. West Publishing Co. , 292 F.3d 992, 1003 (9th Cir. 2002). A denial of leave to file is "strictly reviewed, in light of the strong policy permitting amendment." Moore v. Kayport Package Express, Inc. , 885 F.2d 531, 537-38 (9th Cir. 1989) (quoting Thomas-Lazear v. Federal Bureau of Investigation, 851 F.2d 1202, 1206 (9th Cir. 1988) ). The "district court does not err in denying leave to amend where the amendment would be futile, or where the amended complaint would be subject to dismissal." Saul v. United States , 928 F.2d 829, 843 (9th Cir. 1991) (citations omitted). If the district court is correct in making a finding that "there was no possibility of stating a cause of action .... the dismissal would not be an abuse of discretion." Shermoen v. United States , 982 F.2d 1312 (9th Cir. 1992).

The district court next granted summary judgment for CPUC on CARE's PURPA challenges. This Court reviews summary judgment orders de novo . Sonner v. Schwabe North America, Inc. , 911 F.3d 989 (9th Cir. 2018). This Court "[v]iewing the evidence in the light most favorable to the nonmoving party ... must determine whether there are any genuine issues of material fact and whether the district court correctly applied the relevant substantive law." Devereaux v. Abbey , 263 F.3d 1070, 1074 (9th Cir. 2001) (en banc) (citing Lopez v. Smith , 203 F.3d 1122, 1131 (9th Cir. 2000) (en banc)). On summary *936judgment, "it is not our task ... to scour the record in search of a genuine issue of triable fact." Keenan v. Allan , 91 F.3d 1275, 1279 (9th Cir. 1996) (quoting Richards v. Combined Ins. Co. , 55 F.3d 247, 251 (7th Cir. 1995) ). Rather, "[w]e rely on the nonmoving party to identify with reasonable particularity the evidence that precludes summary judgment." Id.

We recognize that FERC intended to leave states with discretion in implementing its regulations under PURPA. Order 69 , 45 Fed. Reg. at 12226 (stating that a state's implementation of avoided cost is satisfactory if it "reasonably accounts for the utility's avoided costs" and encourages "small power production."). But a state's broad authority in determining how to implement PURPA, IEP , 36 F.3d at 856, and the corresponding deference due state utility regulators, does not mean that we abdicate our responsibility to ensure that the state program complies with PURPA. See, e.g. , Exelon Wind 1, L.L.C. v. Nelson , 766 F.3d 380, 394 (5th Cir. 2014) (explaining that a state is owed deference in PURPA implementation); Allco Renewable Energy Limited v. Massachusetts Electric Company , 208 F.Supp.3d 390, 399 (D. Mass. 2016) (noting that a state cannot implement a program that conflicts with PURPA).

IV. ANALYSIS

CARE alleges that CPUC is not enforcing PURPA's requirement that utilities pay QFs the "full avoided cost" and that utilities must connect QFs to the power grid ("mandatory inter-connection"). CARE challenges several of CPUC's programs based on three theories. First, CARE argues that avoided cost cannot be based on the cost for multiple energy sources. Second, CARE argues that avoided cost must also include capacity costs. Third, CARE argues that the NEM Program violates PURPA's mandatory interconnection requirements. CARE also appeals the district court's dismissal of the equitable damages and attorney fees claims under PURPA.

A. Calculating full avoided cost based on a mix of energy sources

CARE argues that CPUC improperly calculates avoided cost based on multiple sources of electricity, rather than calculating the avoided cost for each type of electricity ("multi-tiered pricing"). CARE argues that if a utility purchases energy from natural gas producers, coal producers, and solar producers, the utility would be required to calculate an avoided cost for natural gas, an avoided cost for coal, and an avoided cost for solar; rather than calculating a single avoided cost based on all the energy sources. CARE argues that several CPUC programs impermissibly base avoided cost on the cost of a natural gas benchmark, rather than a renewables benchmark. CPUC argues that states have discretion in determining how they will comply with PURPA and that, thus, while FERC has said that multi-tiered pricing is permissible, it is not mandatory. While we do not think that PURPA requires utilities to always use multi-tiered pricing, we find that summary judgment was improperly granted here.

In 1995, FERC issued two orders that interpreted "avoided cost."2 In N. Little Rock , FERC stated that "avoided costs are determined ... by all alternatives *937available to the purchasing utility ... [and] include[s] all supply alternatives." N. Little Rock Cogeneration, L.P. and Power Sys., Ltd. v. Entergy Servs., Inc. ("N. Little Rock "), 72 FERC ¶ 61263, 62173, 1995 WL 556544 (Sept. 19, 1995). Similarly, in SoCal Edison , FERC stated that avoided cost must "reflect prices available from all sources able to sell to the utility whose avoided costs are being determined." Re Southern California Edison Co. (SoCal Edison ), 70 FERC ¶ 61215, 61676 (1995), reconsideration denied , 71 FERC ¶ 61269 (1995).

FERC issued an important qualification to this "all sources" requirement in CPUC , 133 FERC ¶ 61,059. In CPUC , FERC clarified that "if a state required a utility to purchase 10 percent of its energy needs from renewable resources, then a natural gas-fired unit, for example, would not be a source 'able to sell' to that utility for the specified renewable resources segment of the utility's energy needs, and thus would not be relevant to determining avoided costs for that segment of the utility's energy needs." Id. at ¶ 61267. California has an RPS. The district court dispensed with the argument that an RPS changes the avoided cost calculation, reading the language in CPUC as permissive rather than mandatory.

The district court erred in reading FERC's pronouncement in such a way. Although FERC initially stated in CPUC that a "state may take into account obligations imposed by the state that, for example, utilities purchase energy from particular sources of energy," CPUC , 133 FERC at ¶ 61266 (emphasis added), later in CPUC , FERC reiterated that when a state has a requirement that utilities source energy from a particular type of generator, "generators with those characteristics constitute the sources that are relevant to the determination of the utility's avoided cost for that procurement requirement." Id. at ¶ 61267. Thus, where a state has an RPS and the utility is using a QF's energy to meet the RPS, the utility cannot calculate avoided costs based on energy sources that would not also meet the RPS.

This reading of FERC's regulations is consistent with other FERC pronouncements. In FERC's final rule implementing Section 210 of PURPA ("Order 69"), FERC explained that if purchasing energy from a QF allowed a utility to forego energy purchases, then the cost of energy was to be included in the avoided cost. But "if a purchase from a qualifying facility permits the utility to avoid the addition of new capacity, then the avoided cost of the new capacity ... should be used." Order 69 , 45 Fed. Reg. at 12216. In other words, FERC interpreted PURPA to require an examination of the costs that a utility is actually avoiding . This comports with PURPA's goal to put QFs on an equal footing with other energy providers. Where a utility uses energy from a QF to meet the utility's RPS obligations, the relevant comparable energy sources are other renewable energy providers, not all energy sources that the utility might technically be capable of buying energy from.

The dissent misreads the majority opinion when it says we require pricing based on each type of energy source for all avoided cost calculations. We do not hold that the avoided cost must be calculated for each individual type of energy. We hold only that where a utility uses energy from a QF to meet a state RPS, the avoided cost must be based on the sources that the utility could rely upon to meet the RPS. If the CPUC chooses to calculate an avoided cost for each type of energy source, it may do so. But it may just as permissibly aggregate all sources that could satisfy its RPS obligations. And if a QF is not aiding a utility in meeting its RPS obligations, the avoided cost in that context need not be *938limited to RPS energy sources. Neither does this opinion hold that CPUC's programs are de facto impermissible under PURPA. Because we hold that the district court misinterpreted PURPA's requirements, we remand for the district court to make such a determination in the first instance.

Because the district court did not read CPUC as requiring an avoided cost based on renewable energy where energy from QFs was being used to meet RPS obligations, it did not consider whether utilities are fulfilling any of their RPS obligations through the challenged CPUC programs. We therefore remand the case to the district court for a determination in the first instance of whether CPUC's programs comply with this aspect of PURPA.

B. Excluding capacity costs from a full avoided cost calculation

CARE next contends that several CPUC programs violate PURPA because they do not include capacity costs as part of the full avoided cost. In granting summary judgment for CPUC, the district court reasoned that PURPA did not require state regulatory agencies to take into account capacity costs. Rather, the regulations required state utility regulators to consider capacity costs only "to the extent practicable." 18 C.F.R. § 292.304(e). The district court found no genuine dispute of material fact that NEM participants were not being paid avoided cost, nor were utilities required to include capacity costs because NEM customers did not provide capacity to the utility. Finally, the district court found that avoided cost did not require the use of long-run avoided cost ("LRAC") as opposed to SRAC.

It would go too far to say that state regulatory agencies are never required to include capacity costs in an avoided cost calculation. The FERC regulations set forth factors for states to consider in setting avoided cost but states that those factors, including capacity, "shall, to the extent practicable, be taken into account." 18 C.F.R. § 292.304(e). FERC has "made clear that an avoided cost rate need not include capacity costs (as distinct from energy costs) where a QF does not 'permit the purchasing utility to avoid the need to construct a generating unit, to build a smaller, less expensive plant, or to reduce firm power purchases from another utility.' " City of Ketchikan, Alaska , 94 FERC ¶ 61293, 2001 WL 275023, at *6 (2001) (quoting Order No. 69, FERC Stats. & Regs., Regs. Preambles 1977-1981 ¶ 30,128 at 30,865. FERC Order 69, however, clarifies that capacity costs are required in some circumstances. Specifically, FERC stated:

[i]f a qualifying facility offers energy of sufficient reliability and with sufficient legally enforceable guarantees of deliverability to permit the purchasing electric utility to avoid the need to construct a generating unit, to build a smaller, less expensive plant, or to reduce firm power purchases from another utility, then the rates for such a purchase will be based on the avoided capacity and energy costs.

Order 69 , 45 FERC at 12216.

Thus, a QF would not be entitled to capacity costs unless it actually displaced the utility's need for additional capacity. If a QF displaces the utility's need for additional capacity, however, the utility is required to include capacity costs as part of avoided costs.

1. The QF Settlement Contract price

CARE challenges the QF Settlement contract price because it does not include capital costs as part of capacity costs.3 As *939CARE acknowledges, the QF standard contract does include capacity costs. Although CARE argues that capital costs, as distinct from capacity costs, are required, CARE has not shown how capital costs differ from capacity costs except for a statement at oral argument that capacity costs are essentially a subset of capital costs. CARE presents no evidence as to why capacity costs, without capital costs, do not accurately reflect a utility's avoided cost. CARE has pointed to "mere conclusory allegations made in [CARE's] own affidavits." Keenan , 91 F.3d at 1279. This is not enough to raise a genuine issue of material fact. Thus, summary judgment was appropriate on this question.

2. The NEM Program

CARE next challenges the DLAP price used in the NEM Program because DLAP does not include capacity costs. CPUC acknowledges that NEM participants are not compensated for avoided capacity but argues that participants in the NEM program are not owed capacity costs because they do not provide any capacity for utilities. CPUC also asserts that net metering programs are not PURPA programs.4

NEM programs are not, as a general matter, state programs categorically exempt from PURPA. In the very CPUC decision implementing the NEM program, CPUC acknowledged that if customers are compensated in the form of a credit on their utility bill, PURPA does not apply. But if the utility is making a separate payment to customers, PURPA applies and the payment must be the full avoided cost.

CPUC is not required to take capacity costs into account in the NEM program. PURPA requires utilities to compensate QFs for capacity costs only when purchasing energy from the QF allows the utility to forgo spending its own money on capacity. FERC has explained that capacity costs are required when "a qualifying facility offers energy of sufficient reliability and with sufficient legally enforceable guarantees of deliverability to permit the purchasing electric utility" to forgo capital investments. Order 69 , 45 FERC at 12216 (emphasis added).

The energy that customers provide to utilities through the NEM Program does not have "sufficient legally enforceable guarantees of deliverability" because customers are not legally required to provide the utility with energy. If, at the end of twelve months, a customer has used more energy than it produced, the customer simply would not provide any energy to the utility. This scenario does not allow utilities to forgo spending on capacity elsewhere because the utility cannot know in advance how much surplus energy NEM participants will provide, and CARE has failed to make any showing that NEM decreases utilities' spending on capacity. Thus, this aspect of the NEM program does not violate PURPA.

3. The Re-MAT and CHP Programs

CARE has given perfunctory treatment to any possible challenge to the Re-MAT

*940and CHP programs, stating only that CPUC operates these programs and that "[a]ll of these programs have one thing in common. Plainly and simply, there is no component for actual avoided capacity costs." Given CARE's bare-bones assertion of the programs' deficiencies, we decline to speculate as to why CARE believes that these programs allow utilities to forgo capacity spending and will not address these programs on appeal. See Navajo Nation v. U.S. Forest Serv ., 535 F.3d 1058, 1079 n.26 (9th Cir. 2008) (en banc) ("It is well-established that a bare assertion in an appellate brief, with no supporting argument, is insufficient to preserve a claim on appeal."). To the extent, however, that CARE challenges either program for basing capacity costs on a new natural gas facility, rather than renewable energy facilities, the district court should consider such a challenge on remand, consistent with our holdings above regarding avoided cost and capacity cost in the context of an RPS.

4. Renewable Energy Credits ("RECs")

CARE next challenges whether CPUC can allow utilities to condition energy purchases from QFs on transfers of the QF's RECs to the utility. As CARE acknowledged in its brief, RECs are not covered under PURPA; rather, they are considered state programs and do not factor into the avoided cost determination. See American Ref-Fuel Co. , 105 FERC ¶ 61,004, 61,008 (2003) ; CGE Fulton, LLC , 70 FERC ¶ 61,290 (1995), reconsideration denied , 71 FERC ¶ 61,232 (1995) ; SoCal Edison , 71 FERC at ¶ 62,080. CARE argues, nonetheless, that RECs are valuable to utilities that do not comply with California's greenhouse gas emission standards (and could thus use the RECs to become compliant) and that allowing utilities to require that QFs give RECs to utilities reduces the cost that QFs receive to below full avoided cost. CPUC argues, and CARE appears to acknowledge, that QFs are compensated for RECs under the NEM program.

CARE cites no legal authority in support of its argument that the value of RECs should be considered as reducing the cost that utilities pay QFs. Given FERC's treatment of RECs as outside the purview of PURPA, however, utilities do not violate PURPA in not compensating QFs for RECs.

C. CPUC's NEM program and PURPA's "must purchase" requirements

CARE alleges that the NEM program violates the mandatory interconnection requirement of PURPA. PURPA requires that utilities "shall make such interconnection with any [QF] as may be necessary to accomplish purchases or sales under this subpart." 18 C.F.R. § 292.303(c). FERC regulations place the burden of paying the cost to connect to the power grid on the QF. 18 C.F.R. § 292.306 (a).

The NEM program does not violate PURPA's mandatory interconnection requirements. Participants in the NEM program are, by definition, connected to the utility's infrastructure. CARE objects to the NEM Program being "imposed unilaterally." While QFs can choose to be compensated based on energy pricing "at the time of delivery" or based on energy pricing at the time a contract is made, 18 CFR § 292.304(d)(2), the interconnection provisions of PURPA merely mandate that utilities connect QFs when needed to comply with PURPA. CARE challenges the imposition of fees, but the regulations specifically allow utilities to charge QFs for the connection fees. Thus, the NEM Program does not violate PURPA.

*941D. Equitable damages and attorney fees

The district court denied CARE's motion for leave to amend its complaint to add a request for equitable damages and attorney fees. The district court found that CARE had not shown that justice so required equitable damages and said that it would "likely conclude" that PURPA does not authorize damages. The district court concluded that suits against Commissioners in their official capacity can only seek "prospective injunctive relief" and that Commissioners had absolute immunity. The district court found attorney fees unavailable because PURPA does not have a fee-shifting provision. We affirm.

As this Court previously noted on appeal, "PURPA has a comprehensive remedial scheme." Solutions for Utilities, Inc. v. Cal. Pub. Utilities Comm'n , 596 F. App'x 571, 572 (9th Cir. 2015). PURPA allows for suits in federal courts and authorizes "such injunctive or other relief as may be appropriate." 16 U.S.C. § 824a-3(h)(2)(B). This Circuit has yet to rule on whether PURPA authorizes equitable damages. We find it unnecessary to reach that issue, however, because the Eleventh Amendment precludes such damages here.

We have previously held that CPUC is immune from suit "as an arm of the state" based on the Supreme Court's determination in Will v. Michigan Dep't of State Police , 491 U.S. 58, 109 S.Ct. 2304, 105 L.Ed.2d 45 (1989) that "Congress did not intend states to be subject to suit under Section 1983." Sable Commc'ns of Cal., Inc. v. Pac. Tel. & Tel. Co. , 890 F.2d 184, 191 (9th Cir. 1989). As an arm of the state, CPUC is protected by the Eleventh Amendment. Air Transport Ass'n of America v. Public Utilities Comm'n of Cal. , 833 F.2d 200, 204 (9th Cir. 1987). The Eleventh Amendment bars citizens from suing their own states in federal court. Edelman v. Jordan , 415 U.S. 651, 663, 94 S.Ct. 1347, 39 L.Ed.2d 662 (1974). A state need not be a "named party to the action." Id. Ordinarily, the Eleventh Amendment would bar suit against CPUC for any purposes.

The Supreme Court rejected a claim similar to CARE's claim for equitable damages in Edelman . There, the Court found that an award of "retroactive benefits," essentially what CARE seeks here, would be in essence "an award of damages against the State," Edelman , 415 U.S. at 668, 94 S.Ct. 1347, and therefore barred by the Eleventh Amendment. Id. at 677, 94 S.Ct. 1347. Thus, the Eleventh Amendment bars CARE's claim for equitable damages. CARE can, however, sue CPUC under the Ex Parte Young exception to the Eleventh Amendment, that allows for "prospective injunctive relief only." Edelman , 415 U.S. at 677, 94 S.Ct. 1347. CARE's reliance on Albemarle Paper Co. v. Moody , 422 U.S. 405, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975), is to no avail, as Albemarle was a suit against private employers, not a state or state agency. CPUC Commissioners in their individual capacity have absolute immunity for "acting in a legislative capacity." Lake Country Estates, Inc. v. Tahoe Regional Planning Agency , 440 U.S. 391, 405-06, 99 S.Ct. 1171, 59 L.Ed.2d 401 (1979).

CARE next argues that the lack of statutory authorization for attorney fees is no bar to their recovery. Attorney fees are not necessarily barred by the Eleventh Amendment. Hutto v. Finney , 437 U.S. 678, 690-93, 98 S.Ct. 2565, 57 L.Ed.2d 522 (1978). Hutto is distinguishable from CARE's claims because the district court in Hutto first found bad faith before imposing attorney fees, making such fees analogous to fines for civil contempt. Here, CARE alleges no bad faith. Hutto additionally examined the availability of attorney fees under 42 U.S.C. § 1988, *942finding that "Congress has plenary power to set aside the States' immunity from retroactive relief in order to enforce the Fourteenth Amendment." Id. at 693, 98 S.Ct. 2565. But unlike § 1988, PURPA creates no attorney fee remedy.

CARE argues that it is entitled to attorney fees under a private attorney general theory. CARE cannot claim attorney fees, however, under that theory. Under a private attorney general theory, a plaintiff could recover attorney fees if the plaintiff: (1) advanced "the interests of a significant class of persons by (2) effectuating a strong congressional policy." Brandenburger v. Thompson , 494 F.2d 885, 888 (9th Cir. 1974). CARE seeks to vindicate the interests of, at a minimum, other solar producers, if not all renewable energy producers. And PURPA evinces a strong policy of encouraging small energy producers. But the Supreme Court long ago foreclosed awarding attorney fees under the private attorney general theory without statutory authorization. See Alyeska Pipeline Serv. Co. v. Wilderness Soc'y , 421 U.S. 240, 269-70, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975). As the Supreme Court made clear in Alyeska Pipeline , Congress may authorize attorney fees in federal statutes. Without such statutory authorization, however, the judiciary would be determining which statutory objectives are important enough to merit shifting the burden of attorney fees. Id. at 263-64, 95 S.Ct. 1612. That is a policy question not suited for judicial resolution. Id. at 269-70, 95 S.Ct. 1612. Therefore, we cannot impose attorney fees under the private attorney general theory as PURPA makes no provision for such fees.

CARE relies on Hall v. Cole , 412 U.S. 1, 93 S.Ct. 1943, 36 L.Ed.2d 702 (1973), to argue for attorney fees under the "private attorney general" theory. Hall , however, concerned the "common benefit" theory of attorney fees rather than the private attorney general theory. The common benefit theory does not apply to CARE, as that theory requires a common fund from which to compensate plaintiffs. In other words, that theory operates to spread the cost of litigation among the beneficiaries of the litigation; it does not shift the fees from the plaintiff to the defendant. See Alyeska Pipeline Serv. Co. , 421 U.S. at 257-59, 95 S.Ct. 1612. Although CARE protests that it is left without a remedy, that is a complaint for Congress, not the courts.

CONCLUSION

The district court erred in not interpreting FERC's regulations to require state utility commissions to consider whether an RPS changed the calculation of avoided cost. This case is reversed and remanded on that issue. In all other respects, the decision below is affirmed.

AFFIRMED IN PART and REVERSED IN PART.

NGUYEN, Circuit Judge, dissenting in part:

Under the Public Utility Regulatory Policies Act of 1978 ("PURPA") and its implementing rules and regulations, states "play the primary role in calculating avoided costs," and are afforded "a great deal of flexibility" in doing so. Indep. Energy Producers Ass'n v. Cal. Pub. Utils. Comm'n , 36 F.3d 848, 856 (9th Cir. 1994) (quoting Administrative Determination of Full Avoided Costs, 4 FERC Statutes & Regs. ¶ 32,457, at 32,173 (proposed Mar. 16, 1988)). While "a state cannot implement a program that conflicts with PURPA," Maj. Op. at 16 (construing Allco Renewable Energy Ltd. v. Mass. Elec. Co. , 208 F.Supp.3d 390, 399 (D. Mass. 2016) ), the majority identifies no such conflict in any of the programs at issue here. Because the majority's misreading of the law substantially undercuts the discretion intended for the *943states and inflicts significant consequences upon their energy policy, I dissent.

I.

A.

Start with the statute itself. PURPA instructs the Federal Energy Regulatory Commission (the "FERC"), "after consultation with representatives of Federal and State regulatory agencies," to develop rules that "require electric utilities to offer to ... purchase electric energy from [qualifying small power production] facilities" ("QFs"). 16 U.S.C. § 824a-3(a). PURPA says little about the rates that utilities must pay for such energy other than that they "shall be just and reasonable to the electric consumers of the electric utility and in the public interest," "shall not discriminate against [QFs]," and cannot "exceed[ ] the incremental cost to the electric utility of alternative electric energy." Id. § 824a-3(b). As FERC interprets these directives, utilities must compensate QFs based on the utilities' "avoided costs," 18 C.F.R. § 292.304(d), which FERC defines as "the incremental costs to an electric utility of electric energy or capacity or both which, but for the purchase from the [QF] or [QFs], such utility would generate itself or purchase from another source." 18 C.F.R. § 292.101(b)(6).

The flexibility afforded to state regulatory authorities and utilities in determining avoided costs is evident in the regulation providing ratemaking guidance. It directs ratemakers to take certain factors into account "to the extent practicable."1 18 C.F.R. § 292.304(e). These factors are framed at an extremely high level of generality to allow states to exercise wide discretion in balancing them.

None of this statutory and regulatory language suggests that utilities must compensate individual QFs based on the costs that the utility would otherwise have incurred by purchasing the same type of energy. For example, a QF selling energy generated from photovoltaic cells is not entitled to receive a rate based on the utility's cost of procuring solar energy from another source. Indeed, the regulations suggest the opposite-that utilities can aggregate energy sources when determining avoided costs. See 18 C.F.R. § 292.101(b)(6) (looking to costs avoided by purchasing "from the [QF] or [QFs]"); see also id. § 292.304(e)(2)(vi) (directing ratemakers to consider "[t]he individual and aggregate value of energy and capacity from [QFs] on the electric utility's system").

B.

In concluding that a utility using energy from QFs to satisfy state-mandated renewable energy targets "cannot calculate avoided costs based on energy sources that would not also meet [those targets]," Maj. Op. at 937, the majority relies on a single sentence from a FERC order that it misinterprets. See Cal. Pub. Utils. Comm'n ("CPUC "), 133 FERC ¶ 61,059, 61,261 (2010). In CPUC , the question was not whether utilities must calculate avoided costs in that manner but whether they could do so consistently with PURPA and FERC regulations. Specifically, CPUC sought clarification that utilities setting *944avoided cost rates could consider factors other than those set forth in 18 C.F.R. § 292.304(e) (the "avoided cost factors") and that avoided costs "need not be the lowest possible avoided cost and can properly take into account real limitations on 'alternate' sources of energy imposed by state law." CPUC , 133 FERC at ¶ 61,262.

Then, as now, the ratemaking regulation required each electric utility to establish "standard rates" for energy purchases from QFs that are "consistent with" the avoided cost factors. 18 C.F.R. § 292.304(c)(3)(i). In addition, standard rates "[m]ay differentiate among qualifying facilities using various technologies on the basis of the supply characteristics of the different technologies." Id. § 292.304(c)(3)(ii) (emphasis added). However, the regulation is not clear whether supply characteristics can be considered only when determining standard rates or whether they can be considered in determining avoided costs generally. FERC explained that supply characteristics can be considered generally. See CPUC , 133 FERC at ¶¶ 61,265-66.

[I]n determining the avoided cost rate, just as a state may take into account the cost of the next marginal unit of generation, so as well the state may take into account obligations imposed by the state that, for example, utilities purchase energy from particular sources of energy or for a long duration. Therefore, the CPUC may take into account actual procurement requirements, and resulting costs, imposed on utilities in California.

Id. at ¶ 61,266 (emphases added).

FERC stressed that "states are allowed a wide degree of latitude in establishing an implementation plan for [determining avoided cost rates], as long as such plans are consistent with [FERC] regulations." Id. at ¶ 61,266 (quoting Am. REF-FUEL Co. of Hempstead , 47 FERC ¶ 61,161, 61,533 (1989) ). Because "the determinations that a state commission makes to implement [PURPA's] rate provisions ... are by their nature fact-specific and include consideration of many factors," FERC was "reluctant to second guess the state commission's determinations." Id.

The majority cherry picks a sentence from CPUC to reach its result. That sentence concerns a different decision "support[ing] the proposition that, where a state requires a utility to procure a certain percentage of energy from generators with certain characteristics, generators with those characteristics constitute the sources that are relevant to the determination of the utility's avoided cost for that procurement requirement." Id. at ¶ 61,267 (construing S. Cal. Edison Co. ("SoCal Edison "), 70 FERC ¶ 61,215 (1995) ).

The problem, CPUC explained, was that "there is language in the SoCal Edison proceeding that would seem to permit state commissions to base avoided costs on 'all sources able to sell to the utility ,' and other language that requires a state commission to take into account 'all sources'" without qualifying language. Id. CPUC clarified that avoided costs calculations do not have to take into account all alternative sources; rather FERC was "permitting states to set a utility's avoided costs based on all sources able to sell to that utility." Id. (emphasis added).

Nothing in CPUC implies that states are required to consider supply characteristics. To the contrary, both in CPUC and the regulations it interprets, the repeated use of terms such as "may," "permits," and "consistent with" all suggest that it is a matter of state discretion.

The majority's only other interpretive support is FERC's statement that "if a purchase from a [QF] permits the utility to avoid the addition of new capacity," i.e. , new generation facilities, "then the avoided cost of the new capacity and not the average embedded system cost of capacity *945should be used." Regulations Implementing PURPA Section 210, 45 Fed. Reg. 12,214, 12,216 (Feb. 25, 1980). But this has nothing to do with consideration of supply characteristics when determining avoided energy costs. Rather, it explains why avoided costs should be based on a utility's "incremental cost" of obtaining alternative energy, 16 U.S.C. § 824a-3(b), rather than the utility's average cost. "Under the principles of economic dispatch, utilities generally turn on last and turn off first their generating units with the highest running cost," so by purchasing energy from a QF, an economically efficient utility "can avoid operating its highest-cost units." Regulations Implementing PURPA Section 210, 45 Fed. Reg. at 12,216.

If anything, this discussion undermines the majority's position. It illustrates "[o]ne way of determining the avoided cost," id. , implying that there are others and, more generally, that states have discretion in their calculations. See id. at 12,226 ("[T]o the extent that a method of calculating the value of capacity from [QFs] reasonably accounts for the utility's avoided costs, and does not fail to provide the required encouragement of cogeneration and small power production, it will be considered as satisfactorily implementing [FERC] rules.").

"The question ... is what costs the electric utility is avoiding. Under [FERC] regulations, a state may determine that capacity is being avoided ... to determine the avoided cost rate." CPUC , 133 FERC at ¶ 61,266 (emphasis added). The majority usurps the state's prerogative.

II.

This is the wrong case to be deciding these issues in a published decision, which will inflict significant consequences on energy policy throughout our circuit. Plaintiffs' briefing, both here and in the district court, is impenetrable. For example, this is plaintiffs' summary of the argument that the majority finds meritorious:

[T]hey2 manipulate the "multi-tiered structure" for pricing, which refers to pegging avoided cost calculations between similar energy sources, which means both in terms of the energy production and, again, capital [capacity] costs. They push for multi-tiered pricing when it serves the utilities, when crafting different contracts for different energy producers; and not when it does not suit them, when renewable energy producers object to an avoided cost computation based on the cheapest source that the utilities can invoke. In either case, the governing rationale is the same: one purpose of PURPA is to expand total capacity and encourage new sources, with policy objectives that include avoidance of risks of shortages, and those objectives are not served by relegating all cost calculations to the cheapest available source which is likely to be existing, aged production facilities.

From that, the majority divines an argument "that CPUC improperly calculates avoided cost based on multiple sources of electricity, rather than calculating the avoided cost for each type of electricity ('multi-tiered pricing')." Maj. Op. at 936.

To the extent plaintiffs have an argument, they seem to be complaining that the CPUC is inconsistent about implementing multi-tiered pricing in a way that always benefits the utilities-not, as the majority seems to assume, that multi-tiered pricing is always required or, for that matter, desirable. Neither the majority nor plaintiffs explain which CPUC programs fail to calculate avoided costs by supply source, let alone how . The majority leaves it to the district court to make plaintiffs' argument for them in the first instance. I do not envy its task.

*946Even under the majority's interpretations, I see no obvious problem if plaintiffs' utility considers sources other than solar energy when calculating the costs it avoids by purchasing energy from solar QFs like plaintiffs. Plaintiffs participate in the Net Energy Metering ("NEM") program which, as the majority acknowledges, means that they have no contractual obligation to sell any amount of electricity to the utility. Maj. Op. at 939. This is a relevant consideration in determining a utility's avoided costs, see 18 C.F.R. § 292.304(e)(2), because it affects the QF's reliability as a source of solar energy. See Regulations Implementing PURPA Section 210, 45 Fed. Reg. at 12,226 ("[T]he value of the service from the [QF] to the electric utility may be affected by the degree to which the [QF] ensures by contract or other legally enforceable obligation that it will continue to provide power."). The CPUC could reasonably find that NEM participants' inherent unreliability in providing solar energy makes them unsuitable as capacity sources to meet a utility's state-mandated renewable energy requirements. While "the diversity of [solar QFs] may collectively comprise the equivalent of [solar] capacity," id. at 12,227 (emphasis added), nothing in the regulations compels such a finding.

The programs at issue here were forged in a hard-fought settlement to end a long-running dispute between QFs and the CPUC. See Maj. Op. at 933. In a stroke, the majority upends this settlement by calling all of these programs into question. There is no reason to create such regulatory uncertainty.

We should affirm the district court's judgment in its entirety. I respectfully dissent.