We have before us challenges to a decision by the Federal Energy Regulatory Commission concerning the reasonableness of rates for the transmission of electricity over facilities owned by utilities that belong to a Regional Transmission Organization (that is, a power pool) called PJM Interconnection. PJM Interconnection, L.L.C., 119 F.E.R.C. ¶ 61,063, 2007 WL 1748619 (2007), rehearing denied, 122 F.E.R.C. ¶ 61,082 (2008); see 16 U.S.C. § 824e; Atlantic City Electric Co. v. FERC, 295 F.3d 1, 10 (D.C.Cir. 2002). (“PJM” stands for “Pennsylvania-New Jersey-Maryland,” but the full name is not used any more.) “RTOs are voluntary associations in which each of the owners of transmission lines that comprise an integrated regional grid cedes to the RTO complete operational control over its transmission lines.” Richard J. Pierce, Jr., “Regional Transmission Organizations: Federal Limitations Needed for Tort Liability,” 23 Energy L.J. 63, 64 (2002); see also Regional Transmission Organizations, 65 Fed.Reg. 810-01, 2000 WL 4557 (FERC Jan. 6, 2000); Morgan Stanley Capital Group Inc. v. Public Utility District No. 1, - U.S. -, -, 128 S.Ct. 2733, 2741, 171 L.Ed.2d 607 (2008). PJM’s region stretches east and south from the Chicago area, primarily to western Michigan and eastern Indiana, Ohio, Pennsylvania, New Jersey, Delaware, Maryland, the District of Columbia, and Virginia. PJM Interconnection, L.L.C., supra, p. 3, 2007 WL 1748619, see FPL Energy Marcus Hook, L.P. v. FERC, 430 F.3d 441, 442-43 (D.C.Cir. 2005). The region is home to more than 50 million consumers of electricity.

Two issues are presented. The first, raised by American Electric Power Service Corporation and the Public Utilities Commission of Ohio (participation by state commissions in rate proceedings before FERC is authorized by 16 U.S.C. § 825g(a); see also § 825Z(a)), involves the pricing of electricity transmitted from the Midwest to the East through Ohio. PJM wants that transmission to be priced on the basis of the cost to American Electric of transmitting one more unit of electricity, that is, the marginal cost; and FERC agrees. Such a price excludes the cost that the company incurred when it built the transmission facilities. That cost— which American Electric wants to be permitted to reflect in its rates — is what economists call a “sunk” cost, that is, a cost that has already been incurred. So while its financial burden can be shifted (from American Electric to the eastern utilities), the cost itself cannot be shifted, and therefore shifting the financial burden created by the cost from one set of shoulders to another will have no direct effect on service or investment.

Had FERC decided that American Electric would not be permitted to charge a price that covered the cost of building a new transmission facility or upgrading an existing one, its decision would have affected the allocation of resources and not just of money. It would have deterred the building of new facilities that benefited customers outside American Electric’s ser*474vice area, because building them would become an unprofitable venture. FERC emphasizes, however, that the company’s existing facilities, which are all that are involved in this case, were built before 2001 when PJM became a Regional Transmission Organization, and were intended to serve American Electric’s customers only. So even if the facilities had not been fully paid for, there would be no economic basis for shifting any part of their costs to other members, because American Electric did not expect when it built the facilities that any part of their cost would be defrayed by anyone besides its customers. PJM and FERC have made clear that American Electric will be allowed to charge a price that covers its costs for transmission to other utilities over new or upgraded facilities.

American Electric points out that some of its existing facilities are not fully depreciated. But it can continue to depreciate them over their remaining useful life in order to create an accounting reserve or obtain a tax benefit. And when it builds a new facility it will be allowed, as we said, to recover the full costs of that facility in its prices.

The company may be trying to extract a monopoly price for the use of its facilities. It stands between western sellers of electricity and their eastern customers and would like to extract a toll for giving the former passage to the latter, a toll that has no relation to its costs of rendering that service. It charged its customers for the costs of building its existing facilities and recovered those costs fully and now wants to recover them all over again from another group of consumers. And it’s not as if American Electric were being required to provide transmission to the east at zero price. It is permitted to charge for the service — -just not to include in the charge its sunk costs.

The second issue relates to the financing of new transmission facilities. Here the Ohio commission joins its Illinois counterpart, representing the interests of the midwestern utilities in PJM’s region, in objecting to PJM’s proposed method, approved by FERC, for pricing new transmission facilities that have a capacity of 500 kilovolts or more. Heretofore all new facilities in PJM’s region have been financed by contributions from the region’s electrical utilities calculated on the basis of the benefits that each utility receives from the facilities. This will continue to be the rule for facilities with capacities of less than 500 kV. But for the higher-voltage facilities FERC has decided that all the utilities in PJM’s region should contribute pro rata; that is, their rates should be raised by a uniform amount sufficient to defray the facilities’ costs.

FERC’s stated reasons are that some of PJM’s members entered into similar pro rata sharing agreements with each other more than forty years ago and would like to follow that precedent, that figuring out who benefits from a new transmission facility and by how much is very difficult and so generates litigation, and that everyone benefits from high-capacity transmission facilities because they increase the reliability of the entire network. Despite the stakes in the dispute — the new policy might, for example, force Commonwealth Edison to contribute hundreds of millions of dollars to an above-500 kV eastern project called “Project Mountaineer,” when it would not have had to pay a dime under the benefits-based system applicable to lower-voltage transmission facilities — no data are referred to in FERC’s two opinions (the original opinion and the opinion on rehearing). No lawsuits are mentioned. No specifics concerning difficulties in assessing benefits are offered. No particulars are presented concerning the contri*475button that very high-voltage facilities are likely to make to the reliability of PJM’s network. Not even the roughest estimate of likely benefits to the objecting utilities is presented. The first sentence in this paragraph is an adequate summary of the Commission’s reasoning, minus recourse to metaphor, as in the Commission’s repeated references to very high-voltage facilities as the “backbone” of PJM’s network. The Commission’s insouciance about the basis for its ruling is mirrored by its lawyers: their brief devotes only five pages to the 500 kV pricing issue.

The objections to the Commission’s ruling pivot on an asymmetry between the eastern and western portions of PJM’s region. In the west the electrical generating plants usually are close to the customers — Chicago for example is ringed by power plants. As a result, relatively low-voltage transmission facilities — mainly 345 kV — are preferred. In the east, where the power plants generally are farther away from the customers, 500 kV and even higher-voltage transmission facilities are preferred, because high voltage is more efficient than low for transmitting electricity over long distances. So far as appears, few if any such facilities will be built in the objectors’ service areas, that is, in the Midwest, within the foreseeable future. FERC seems not to care whether any will ever be built, because the reasons it gave for approving PJM’s new pricing method are independent of where the facilities are located.

The first two reasons the Commission gave can be dispatched briefly. The fact that some of the same members of PJM who agreed to share the costs of such facilities with each other many years ago would like contributions from midwestern utilities carries no weight. The eastern utilities that created PJM refer to themselves revealingly as the “classic” PJM utilities, and the fact that these utilities thought it appropriate to share costs in 1967 says nothing about the advantages and disadvantages of such an arrangement in the larger, modern PJM network.

The Commission said that it would be inclined to defer to “regional consensus,” but acknowledged there was none; the midwestern utilities are members of PJM but did not agree to the eastern utilities’ cost-sharing proposal. As we shall see, the fact that one group of utilities desires to be subsidized by another is no reason in itself for giving them their way.

The second reason the Commission gave for approving PJM’s pricing scheme — the difficulty of measuring benefits and the resulting likelihood of litigation over them — fails because of the absence of any indication that the difficulty exceeds that of measuring the benefits to particular utilities of a smaller-capacity transmission line. Like the D.C. Circuit in Sithe/Independence Power Partners, L.P. v. FERC, 285 F.3d 1, 5 (D.C.Cir. 2002) (citation omitted), we acknowledge “that feasibility concerns play a role in approving rates, indicating that FERC is not bound to reject any rate mechanism that tracks the cost-causation principle less than perfectly.” But we also agree that “the Commission’s cursory response simply will not do. At no point did the Commission explain how these considerations [that the tariffs and refund mechanism produced ‘efficient price signals,’ and that petitioner’s requested refunds would somehow disrupt that price signaling, would be ‘infeasible,’ and a matter of ‘unending controversy’] applied. Why, we wonder, would a different method of refunds, based more closely on cost-causation principles, jeopardize desirable price signaling or be infeasible?” Id.

No doubt the more a transmission facility costs, and therefore the greater the stakes in a dispute between potential con*476tributors to that cost, the more litigation there is likely to be. But how much more (at least approximately) is the critical consideration and the Commission ignored it.

That leaves for consideration the benefits that the midwestern utilities might derive from the greater reliability that the larger-capacity transmission facilities might confer on the network as a whole. The reason for building such facilities is to satisfy the demand of eastern consumers for electricity, but the more transmission capacity there is, the less likely are blackouts or brownouts caused by surges of demand for electricity on hot summer days or by accidents that shut down a part of the electrical grid. Because the transmission lines in PJM’s service region are interconnected, a failure in one part of the region can affect the supply of electricity in other parts of the network. So utilities and their customers in the western part of the region could benefit from higher-voltage transmission lines in the east, but nothing in FERC’s opinions in this case enables even the roughest of ballpark estimates of those benefits.

At argument FERC’s counsel reluctantly conceded that if Commonwealth Edison would derive only $1 million in expected benefits from Project Mountaineer, for which it is being asked to chip in (by its estimate) $480 million, the disparity between benefit and cost would be unreasonable. The concession was prudent. Algonquin Gas Transmission Co. v. FERC, 948 F.2d 1305, 1313 (D.C.Cir. 1991); Pacific Gas & Electric Co. v. FERC, 373 F.3d 1315, 1320-21 (D.C.Cir. 2004). As FERC itself explained in Transcontinental Gas Pipe Line Corp., 112 F.E.R.C. ¶ 61,170, 61,924-61,925, 2005 WL 1848553 (2005), “a claim of generalized system benefits is not enough to justify requiring the existing shippers to subsidize the uncontested increase in electric costs caused by the Cherokee project.... The rehearing applicants suggest that the use of the Cherokee shippers’ transportation quantities in deriving the fuel retention percentages and their payment of such charges reduce the fuel costs borne by the existing shippers. However, they point to no evidence in the record that seeks to quantify this benefit, or even shows that such a benefit has occurred.... The Commission concludes that all these alleged benefits are simply too speculative and unsupported to be taken into account.”

FERC is not authorized to approve a pricing scheme that requires a group of utilities to pay for facilities from which its members derive no benefits, or benefits that are trivial in relation to the costs sought to be shifted to its members. “ ‘[A]ll approved rates [must] reflect to some degree the costs actually caused by the customer who must pay them.’ KN Energy, Inc. v. FERC, 968 F.2d 1295, 1300 (D.C.Cir. 1992); Transmission Access Policy Study Group v. FERC, 225 F.3d 667, 708 (D.C.Cir. 2000); Pacific Gas & Elec. Co. v. FERC, 373 F.3d 1315, 1320-21 (D.C.Cir. 2004). Not surprisingly, we evaluate compliance with this unremarkable principle by comparing the costs assessed against a party to the burdens imposed or benefits drawn by that party.” Midwest ISO Transmission Owners v. FERC, 373 F.3d 1361, 1368 (D.C.Cir. 2004); see also Alcoa Inc. v. FERC, 564 F.3d 1342, 1346-47 (D.C.Cir. 2009); Sithe/Independence Power Partners, L.P. v. FERC, supra, 285 F.3d at 4-5; Federal Power Act, 16 U.S.C. § 824d. To the extent that a utility benefits from the costs of new facilities, it may be said to have “caused” a part of those costs to be incurred, as without the expectation of its contributions the facilities might not have been built, or might have been delayed. But as far as one can tell from the Commission’s opinions in this *477case, the likely benefit to Commonwealth Edison from new 500 kV projects is zero. The opinion on rehearing attributes the need for new transmission capacity in PJM to the threat of “degraded reliability in Eastern PJM,” 122 F.E.R.C. ¶ 61,082, p. 13, 2008 WL 276596 (emphasis added), and nowhere do the Commission’s opinions suggest that degraded reliability is a danger in Midwestern PJM.

No doubt there will be some benefit to the midwestern utilities just because the network is a network, and there have been outages in the Midwest. But enough of a benefit to justify the costs that FERC wants shifted to those utilities? Nothing in the Commission’s opinions enables an answer to that question. Although the Commission did say that a 500 kV transmission line has twice the capacity of a 345 kV line, it added that “the reliability of 500 kV and above circuits in terms of momentary and sustained interruptions is 70 percent more reliable than 138 kV circuits and 60 percent more than 230 kV circuits on a per mile basis,” PJM Interconnection, L.L.C., supra, 119 F.E.R.C. ¶ 61,063, p. 23; 122 F.E.R.C. ¶ 61,082, p. 16 (emphasis added) — but did not compare the reliability of a 500 kV line to that of a 345 kV line, even though network reliability is the benefit that the Commission thinks the midwestern utilities will obtain from new 500 kV lines in the East.

Rather desperately FERC’s lawyer, and the lawyer for the eastern utilities that intervened in support of its ruling, reminded us at argument that the Commission has a great deal of experience with issues of reliability and network needs, and they asked us therefore (in effect) to take the soundness of its decision on faith. But we cannot do that because we are not authorized to uphold a regulatory decision that is not supported by substantial evidence on the record as a whole, or to supply reasons for the decision that did not occur to the regulators. E.g., 5 U.S.C. § 706; Bethany v. FERC, 276 F.3d 934, 940 (7th Cir. 2002); Central Illinois Public Service Co. v. FERC, 941 F.2d 622, 627 (7th Cir. 1991); Pacific Gas & Electric Co. v. FERC, supra, 373 F.3d at 1319. The reasons that did occur to FERC are inadequate.

We do not suggest that the Commission has to calculate benefits to the last penny, or for that matter to the last million or ten million or perhaps hundred million dollars. Midwest ISO Transmission Owners v. FERC, supra, 373 F.3d at 1369 (“we have never required a ratemaking agency to allocate costs with exacting precision”); Sithe/Independence Power Partners, L.P. v. FERC, supra, 285 F.3d at 5. If it cannot quantify the benefits to the midwestern utilities from new 500 kV lines in the East, even though it does so for 345 kV lines, but it has an articulable and plausible reason to believe that the benefits are at least roughly commensurate with those utilities’ share of total electricity sales in PJM’s region, then fine; the Commission can approve PJM’s proposed pricing scheme on that basis. For that matter it can presume that new transmission lines benefit the entire network by reducing the likelihood or severity of outages. E.g., Western Massachusetts Elec. Co. v. FERC, 165 F.3d 922, 927 (D.C.Cir. 1999). But it cannot use the presumption to avoid the duty of “comparing the costs assessed against a party to the burdens imposed or benefits drawn by that party.” Midwest ISO Transmission Owners v. FERC, supra, 373 F.3d at 1368. Nor did it in the Western Massachusetts case.

In Midwest ISO, where the objecting utilities contended that they were being asked to pay far more than their share of the benefits — which they said was a measly 5 percent — the court found that they were misrepresenting the record. *478373 F.3d at 1370. There is no comparable basis on which to affirm the Commission’s decision in this case. Our review of decisions by FERC is deferential, e.g., Town of Norwood v. FERC, 962 F.2d 20, 22 (D.C.Cir. 1992); “we require only that the agency have made a reasoned decision based upon substantial evidence in the record.” Id. But the Commission failed to do that, and so the case must be remanded for further proceedings; we intimate no view on their outcome.

To summarize, the petitions for review that concern the pricing of existing transmission facilities are denied, but the petitions concerning the pricing of new facilities that have a capacity of 500 kilovolts or more are granted.

concurring in part and dissenting in part.

I concur fully in the majority’s approval of FERC’s rate design for existing facilities’ transmission costs. I write separately to express my concerns over the majority’s disapproval of the proposed rate design for new transmission lines operating at voltages at or in excess of 500,000 volts.

The United States is now engaged in an urgent project to upgrade its electric transmission grid, which for years has been generally regarded as inadequate,1 and may become more deficient with the addition of major new anticipated loads.2 The existing transmission system originally served vertically integrated utilities that built their own generation relatively close to their customers. The system was not designed for long-distance power transfers between different parts of the country. The inadequacy of the present network and the urgency of the need for its improvement has only been exacerbated by the additional burdens imposed by deregulation (or restructuring), which “unbundled” generation and transmission and created a need to bring power from distant generators.3 Additional challenges have been posed by the demand for power from renewable generation sources (such as wind farms) that are often located in places remote from centers of electric consumption.4

Long-distance transmission, which inherently presents challenges to reliability, is accomplished most efficiently by the highest levels of voltage — 500 kV and above. According to FERC, “500 kV and above circuits ... [are] 70 percent more reliable than 138 kV circuits and 60 percent more than 230 kV circuits on a per mile basis.” PJM Interconnection LLC, 122 FERC ¶ 61,082, 2008 WL 276596, at *479*16 (Jan. 31, 2008) (order on rehearing). Further, because power transfer capability-increases with the square of voltage,5 extra-high voltage transmission also facilitates enormous transfers of power: “the maximum transfer capability at 500 kV and above is approximately 6 times greater than a similar transmission line operated at 230 kV and more than twice that at 345 kV....” Id. In light of its unique contributions to reliability and transfer capability, extra-high voltage transmission is especially fitted to be financed equally by all utilities that benefit from its role as the “backbone” of the system.6 Pro rata rates for extra-high voltage transmission, through their simplicity of application, also provide a strong incentive to build transmission undeterred by fruitless controversy over the allocation of costs.

It is significant that FERC’s conclusion that the costs of extra-high voltage transmission facilities should be shared is consistent with the proposals of fifteen of PJM’s seventeen members. In the course of this proceeding, various parties proposed voltages lower than 500 kV as the threshold above which proportional cost-sharing should apply. Although PJM’s members were unable to agree on a specific voltage cutoff, they were broadly in agreement that the rate structure should be designed to share the costs of facilities providing general systemic benefits. There was thus an effort by many parties to broaden the area of rate-simplification by enlarging the- set of new transmission facilities to be governed by cost-sharing, not to narrow or eliminate it. I think these efforts illustrate the value of simplification and the difficulties in the design of a transmission rate structure that attempts rigidly and in all circumstances to trace benefits to specific utilities.

However theoretically attractive may be the principle of “beneficiary pays,” an unbending devotion to this rule in every instance can only ignite controversy, sustain arguments and discourage construction while the nation suffers from inadequate and unreliable transmission. Unsurprisingly, it is not possible to realistically determine for each utility and with reference to each major project the likelihood that rate-simplification .will reduce litigation, or to calculate the precise value of not having to cover the costs of power failures and of not paying costs associated with congestion, and all this over the next forty to fifty years. Concerns about the real value to individual utilities of the stability and efficiency provided by improvements to the backbone grid are answered by their voluntary participation in the power pool and its collaborative “RTEP” (or regional transmission expansion planning) process. Rate-making based on cost causation is assured by this process, since universal cost-sharing is recommended only when developments are found to benefit the integrated system as a whole.7

*480Contrary to the majority’s suggestion, FERC did not violate principles of “cost causation” by failing to propose a number that would represent the specific monetary benefits to each utility of a more reliable network. Cost causation requires that “approved rates reflect to some degree the costs actually caused by the customer who must pay them.” Midwest ISO Transmission Owners v. FERC, 373 F.3d 1361, 1368 (D.C.Cir. 2004) (Roberts, J.) (quoting KN Energy, Inc. v. FERC, 968 F.2d 1295, 1300 (D.C.Cir. 1992)) (internal quotation marks omitted). However, until today, no court has found that cost causation requires FERC to monetize the benefits of reliability improvements in order to share the costs. Indeed, the cases the majority cites support the opposite conclusion. Most notably, in Midwest ISO, the panel was quite clear that utilities that draw benefits from being a part of a power pool should share the cost of having a power pool. Id. at 1371. As then-Judge Roberts explained, “upgrades designed to preserve the grid’s reliability constitute system enhancements that are presumed to benefit the entire system.” Id. at 1369 (internal quotation marks, citations and alterations omitted, and emphasis added); see also Entergy Servs., Inc. v. FERC, 319 F.3d 536, 543 (D.C.Cir. 2003); Western Massachusetts Elec. Co. v. FERC, 165 F.3d 922, 927 (D.C.Cir. 1999). Since there is a presumption that enhanced reliability benefits all of the systems members, Commonwealth Edison (ComEd) can be required to bear a proportional share of an improvement’s costs even where it is not possible to determine precisely how much it benefits. Put otherwise, the burden is on ComEd to show that it would not benefit from the newly planned transmission facilities; the burden is not on FERC to estimate how much ComEd would benefit from a more reliable grid.

Indeed, in Midwest ISO, the panel rejected the objecting utility’s argument that it could not be made to pay sixty to seventy percent of an investment’s costs because it would obtain only five percent of the benefits. 373 F.3d at 1370. As the majority notes, the panel found no record support for the utility’s claim that its benefits would be so low. (Maj. Op. at 477.) However, the panel also held that cost causation principles do not require the costs of a new facility to be apportioned based on the objecting utility’s actual use of that facility. To the contrary, the “benefits” of system enhancements must be understood more broadly than this. Again, then-Judge Roberts:

even if they are not in some sense using the ISO [roughly a term for a power pool], the MISO Owners still benefit from having an ISO. In this sense, MISO is somewhat like the federal court system. It costs a considerable amount to set up and maintain a court system, and these costs — the costs of having a court system — are borne by the taxpayers, even though the vast majority of them will have no contact with that system (will not use that system) in any given year ... The MISO Owners’ position is tantamount to saying that if *481they are not a litigant, they should not be made to pay for any of the costs of having a court system. Since the MISO Owners do, in fact, draw benefits from being a part of the MISO regional transmission system, FERC correctly determined that they should share the cost of having an ISO.

Id. at 1371. I fear that the majority has lost sight of this basic principle.8

Because the majority’s decision is based on an unusually narrow conception of cost-causation, its characterizations of FERC’s and the intervenor’s arguments as “insouciant” (Maj. Op. at 475) and “desperate” (Maj. Op. at 477) strike me as conspicuously misplaced. FERC responded to ComEd’s objections by indicating that the proposed projects would improve reliability and reduce congestion. See PJM Interconnection, 2008 WL 276596, at *16. It did not explain how PJM’s members benefit from a reliable network because no court had hitherto required it to do so. Until now, it went without saying that, network reliability benefits the network’s members. This is not insouciance; “[explanations come to an end somewhere.” Ludwig Wittgenstein, Philosophical Investigations § 1 (G.E.M. Anscombe trans., 1968).

The big picture here is that FERC’s proposal to spread the cost of very high voltage transmission on a uniform basis seems to me in the interest of efficient, high-capacity transfer capability and of the closely linked improvement of reliability, which affects the system generally.9 Deregulation created a demand for competitive sources of power, often at a distance. Because 500 kV and above lines satisfy these new systemic needs, their separate *482treatment for rate-making purposes is both sensible and innovative. While an effort to identify specific benefits to specific utilities is a traditional rate design approach and may be appropriate for most electric plant facilities, it may miss the forest and focus on the trees when applied to very high voltage “backbone” facilities having a generalized role in supporting reliability and high capacity power transfer. Perhaps as important in this picture is the urgency of the need to build transmission and the need for incentives to that end. Pro rata assignment of costs eliminates not only lawsuits but nitpicking controversies of every sort and delays standing in the path of action. From that point of view, I think FERC may be in a better position to implement a policy leading to prompt improvement in a deficient transmission grid than this court, focused as it is on the inevitable complaints of utilities demanding more for their money. I therefore respectfully dissent from the majority’s unfortunate rejection of FERC’s rate scheme for new transmission lines carrying 500 kV or higher.