D.H. GINSBURG, Circuit Judge:

The Town of Norwood seeks review of a Federal Energy Regulatory Commission order approving the New England Power Company's adoption of a marginal cost approach to setting wholesale electric rates. This is the first marginal cost rate design that the FERC has approved since we instructed the agency on the need for record support in Electricity Consumers Resource Council v. FERC, 747 F.2d 1511 (D.C. Cir. 1984) ( ELCON). Because the FERC has this time given adequate reasons for its decision to approve the proposed rate design, based upon evidence in the record, we deny the petition for review.

I. BACKGROUND

The rate design before us, like most wholesale electric rates, consists of separate monthly demand and energy charges. The demand component is calculated to recover NEPCO's fixed (or capacity-related) costs, such as construction and debt service, which it incurs regardless of how much electricity it produces. The energy charge is designed to recover the company's variable costs, which it incurs only in the course of actually producing electricity; fuel is a prime example.

In NEPCO's proposed rate design, the demand and the energy charge are each divided into two rates, one applicable to the customer's so-called "initial block" and the other to its "tail block." Each customer's initial block is an amount equal to 80% of its average monthly maximum demand and energy use during a base year. The rate for the initial block is based upon NEPCO's average cost, derived as usual from its historical or "embedded" costs. Demand and energy use in excess of that level is to be billed at the tail block rate, calculated on the basis of NEPCO's estimated long-run marginal cost (LRMC) for future capacity and energy.

In addition, NEPCO proposed to change the way it determines each customer's demand charge: whereas the demand charge had previously been calculated on the basis of each customer's actual peak usage during the month, it would now be calculated on the basis of the customer's usage at the time that total demand on NEPCO's system peaks each month. This switch from non-coincident to coincident peak billing was designed to impose upon each customer the cost of its contribution to NEPCO's need for increased capacity, which must be added in order to supply any increase in peak demand.

The ALJ approved these features, as well as other aspects of NEPCO's proposal that are no longer in controversy. See 49 FERC ¶ 63,007 (1989). The Commission affirmed, approving and to some degree supplementing the opinion of the ALJ, 52 FERC ¶ 61,090 (1990), and denied the petitioner's request for rehearing, see 54 FERC ¶ 61,055 (1991).

If as expected NEPCO's marginal costs exceed its historical costs, then the proposed use of block billing will increase the price of electricity to customers whose demand is increasing. The Town of Norwood is foremost among these; from 1985 through 1988 (when NEPCO filed the proposed rate design), Norwood's consumption rose almost 30% while that of all other NEPCO customers grew by 18.75%. In fact, the petitioner stands alone among NEPCO's seven unaffiliated wholesale customers in experiencing a rate increase by reason of the change in rate structure.

In ELCON, we reviewed the Commission's approval of a rate design that proceeded from the principle of marginal cost pricing but incorporated "substantial adjustments" made necessary in order to reconcile that principle with the utility's revenue constraint. The result was a complicated affair — the utility "would collect the entire marginal cost of energy and would consider any residual revenue collectable under the revenue constraint to be the demand charge," 747 F.2d at 1513 (emphasis in original) — that might have been entirely justifiable under the circumstances but was not recognizable as marginal cost pricing. We remanded the matter for the FERC's reconsideration, noting that "mere invocation of [marginal cost pricing] theory is an insufficient substitute for substantial evidence and reasoned explanations," particularly "where the theory has been severely compromised by the revenue constraint." Id. at 1517. We pointed out "unequivocally," however, that we were not "expressing our opposition" to the principle of marginal cost rate design; we even allowed that "[a] modified version of marginal cost pricing theory may indeed be perfectly acceptable, but not without record support and reasoned decision-making by FERC that address the particular nuances of the modified theory." Id. at 1518 (emphasis in original).

The present case brings once more before the court the issue of marginal cost pricing in a utility's rate design — this time without the Rube Goldberg-style modifications that the utility had cobbled together and the agency had approved in ELCON. Here the petitioner argues straightforwardly that (1) the proposed LRMC rate design is not just and reasonable because it is too volatile (at least relative to the stability of average cost rates), (2) marginal cost ratemaking is impermissibly retroactive, (3) the theory of marginal cost pricing supports only the use of short run marginal cost (SRMC) and not LRMC, and (4) the proposed rate structure involves an unlawful rate tilt. In addition, the petitioner argues that coincident peak demand billing is not just and reasonable and that the FERC impermissibly shifted the burden of proof from NEPCO to the petitioner on these issues.

II. ANALYSIS

Issues of rate design are fairly technical and, insofar as they are not technical, involve policy judgments that lie at the core of the regulatory mission. Not surprisingly, therefore, our review is deferential. In determining whether the FERC's decision that a rate design is "just and reasonable," 16 U.S.C. § 824d(a), we require only that the agency have made a reasoned decision based upon substantial evidence in the record.

A. The LRMC Rate Design

We find adequate justification in the record for the Commission's approval of NEPCO's proposal to move from an embedded cost to a marginal cost rate design. The ALJ correctly stated and clearly understood the conventional economic case for marginal cost pricing:

[NEPCO's marginal cost rate design applies] to public utility ratemaking . . . one of the best-established precepts of classical economics: social welfare is maximized when the marginal cost of purchasing any commodity is equivalent to the marginal cost of producing it. . . . [T]he rate design proposed in this proceeding by NEPCO would approach that objective more closely than continued adherence to a rate design based upon embedded costs.

49 FERC at 65,032. The Commission fully embraced this rationale, remarking that "customers must face prices that reflect their supplier's incremental costs in order for them to make efficient investment decisions and efficient choices when seeking alternative supply sources." 52 FERC at 61,335.

We note the agreement as well of the leading authorities in the field of public utility regulation. See James C. Bonbright, Albert L. Danielsen, David R. Kamerschen, Principles of Public Utility Rates 410 (2d ed. 1988) ("[M]arginal cost must play a major and even a dominant role in the elaboration of any scheme of rates or prices that seriously pretends to have as a major motive the efficient utilization of available resources and facilities") (quoting William Vickrey, Some Implications of Marginal Cost Pricing for Public Utilities, 45 Am.Econ.Rev., Papers and Proceedings 605 (1955)); Robert G. Uhler, Electric Utility Pricing Issues, in Current Issues in Public-Utility Economics 77, 81-82 (Albert L. Danielsen and David R. Kamerschen eds. 1983) ("The issue for the 1980s is not to debate the merits of marginal-cost pricing [, an idea] articulated by Bonbright two decades ago and clearly understood by the industry experts of today. The issue for the 1980s is to reaffirm Bonbright's three primary criteria for sound rate-making — revenue adequacy, optimal use of service, and fairness — and doggedly pursue marginal-cost pricing as a means to achieve those time-honored ends"); Alfred E. Kahn, 1 The Economics of Regulation: Principles and Institutions 65 (1970) [hereinafter Economics of Regulation] ("If economic theory is to have any relevance to public utility pricing, [the equation of price and marginal cost] is the point at which the inquiry must begin"); Samuel Huntington, The Rapid Emergence of Marginal Cost Pricing in the Regulation of Electric Utility Rate Structures, 55 B.U.L.Rev. 689, 696-97 (1975) ("Marginal cost pricing . . . has long been accepted by economists. . . . [M]any regulators and other participants in the regulatory process are beginning to bridge the gap between economic theory and regulatory practice"); J. Robert Malko and Philip R. Swensen, Pricing and the Electric Utility Industry, in Public Utility Regulation 35, 56-69 (Kenneth Nowotny et al. eds., 1989) (recounting and applauding successful implementation of marginal cost pricing by public electric utilities in Wisconsin); Irston R. Barnes, The Economics of Public Utility Regulation 321 (1942) (demonstrating graphically that setting marginal cost to marginal revenue achieves the optimal level of output).

The ALJ specifically credited the testimony of the FERC staff's expert economist, who endorsed marginal cost pricing and noted that the alternative — embedded cost pricing — is inherently inferior because it is backward- rather than forward-looking. The Commission in turn adopted and relied upon the ALJ's factual finding that "[t]he proposed marginal cost rates track NEPCO's costs over the short and long term more accurately than alternative rates based on embedded costs and send more accurate price signals to customers." 52 FERC at 61,335.

The ALJ also noted that

the marginal cost rate design should in theory reward customers who are able to shift their usage from peak periods to off-peak periods with lower rates.

49 FERC at 65,028. And the Commission expressly endorsed his conclusion that

[c]ompelling [NEP] [sic] to adhere to an embedded cost rate design would signal wholesale customers . . . that it costs no more to serve additional load at peak times than it does at off-peak times.

52 FERC at 61,335.

See also Kahn, 1 Economics of Regulation, supra, at 95-96 (criticizing use of non-coincident peaks), 101-03 (explaining justification for burden on peak users).

The Commission also thought it relevant to fairness that the marginal cost component of NEPCO's rate design is confined to the demand and energy charges for usage in excess of 80% of each customer's average base-year usage, which minimizes any transitional hardship attendant to the introduction of the new rate structure. Thus, the ALJ noted that "NEPCO has made every reasonable effort to avoid undue disruptions resulting from introduction of the marginal cost rate design. . . . The proof of the pudding in this regard is that only one non-affiliated customer [ viz. Norwood] receives a rate increase resulting from the shift to a marginal cost rate design, and the size of that increase is rather small." 49 FERC at 65,033.

Despite the substantial record support for NEPCO's LRMC rate design, as noted above, the petitioner asserts four grounds for thinking it unreasonable. We hold that only two of those grounds are properly before us, and neither is persuasive.

1. Rate volatility. The petitioner complains that LRMC pricing is inherently more volatile than embedded cost pricing. The ALJ frankly acknowledged the point: "[C]hanges in rates that reflect changes in future cost levels are entirely appropriate to provide accurate price signals to customers." Id.

Marginal cost pricing, being prospective, is more volatile than embedded cost pricing because, as the ALJ noted, it "is based on constantly changing estimates of future costs." Id. With the cost outlook constantly in flux due to changing economic conditions, some degree of volatility is necessary if prices are to signal the market accurately — as accurately, that is, as current prices can anticipate future costs. Price volatility alone, therefore, cannot provide a ground for overturning a marginal cost rate structure. Cf. Alfred E. Kahn, Applications of Economics to Utility Rate Structures, Pub.Util.Fort., Jan. 19, 1978, at 13, 15 ("The only economic function of price is to influence behavior") (emphasis in original).

2. Retroactive ratemaking. The petitioner's attempt to characterize LRMC pricing as retroactive ratemaking is badly misguided. The petitioner decries the use of past consumption decisions in current prices under the LRMC rate design: "customers could never know from year to year whether or not their future costs would be modified by their past purchasing practices." The past pattern of consumption is a major factor, however, in predicting a utility's need to incur the costs of adding and operating capacity in the future, and it is the presently projected cost of such future additions to which NEPCO's current prices respond.

The petitioner cites Associated Gas Distrib. v. FERC, 893 F.2d 349 (D.C. Cir. 1990), for the proposition that a past consumption pattern may not be used to determine a future price. That case involved a demand surcharge imposed upon past purchases. See id. at 355 ("the Commission now even forces past customers who no longer purchase any gas from Tennessee to pay their share of the take-or-pay liability") (emphasis in original). In that case, a customer had no way of knowing, when it made its decision to buy gas, that it would later incur a surcharge on that purchase. Nothing in the AGD decision prevents a utility from basing prices for future sales upon the past pattern of consumption. Cf. id. at 354 ("`predictability' [is] the fundamental policy underlying" the prohibition of retroactive ratemaking). Thus, NEPCO's new rate structure is not retroactive in any meaningful sense.

3. The SRMC alternative. The petitioner also argues that because there is uncontradicted testimony in the record to the effect that SRMC pricing is superior to LRMC, the FERC should have required NEPCO to adopt that approach to marginal cost pricing. This issue was not properly raised, however, and thus need not concern us at the present time.

The staff's economic expert, who ultimately endorsed NEPCO's rate design, did suggest along the way that SRMC pricing is in principle superior to LRMC pricing, although he thought the difference in result would not be significant in this case. Neither the petitioner nor anyone else had proposed using SRMC before the staff witness raised the theoretical possibility. Nonetheless, in its post-hearing brief the petitioner seized upon the SRMC possibility and argued vigorously for it.

The ALJ ruled that "[t]he [post-hearing] brief writing stage is too late in the day . . . to propose an alternative to the rate design filed by the utility. The time to put forward alternatives is when testimony is filed." 49 FERC at 65,036. The Commission agreed. See 52 FERC at 61,336 ("there is no live dispute on the subject, and the potential controversy can be pretermitted until another day") (quoting ALJ). So do we. Cf. Tennessee Gas Pipeline Co. v. FERC, 871 F.2d 1099, 1112 (D.C. Cir. 1989) (upholding agency's rejection of new arguments raised in a second petition for rehearing because the petitioner could have raised them earlier and there was no reasonable ground for not having done so). Like the agency, we of course express no view on the merits of the LRMC versus SRMC controversy.

4. Rate tilt. In a footnote to its opening brief in this court, the petitioner makes the following reference to its claim before the agency that the NEPCO rate design involves an impermissible cross-subsidy, or rate tilt: "As Norwood established in its application for rehearing at 35-40, it is unrebutted that NEPCO's rates in the instant case are also unlawfully `tilted.'" Mere reference to an argument presented elsewhere, however, is not sufficient to raise it here; because the petitioner did not preserve the argument in its opening brief, we reject as untimely the petitioner's attempt to raise the point in its reply brief. See McBride v. Merrell Dow and Pharmaceuticals, Inc., 800 F.2d 1208, 1210 (D.C. Cir. 1986) ("We generally will not entertain arguments omitted from an appellant's opening brief and raised initially in his reply brief"). Again, because the issue is not properly before us, we express no view of it.

B. Coincident Peak Demand Billing

The CP demand billing feature of the new rate structure allocates to each customer a share of the cost of future additions to capacity based upon that customer's proportionate use of existing capacity at the time of peak system demand. This ensures that the cost of new capacity is allocated to those who contribute to the need for adding it — an eminently sensible allocation, and one that we have endorsed before. See Union Elec. Co. v. FERC, 890 F.2d 1193, 1198 (D.C. Cir. 1989) ("[Costs] are assessed to the peak-period users because it is peak demand that determines how much a utility will invest in capacity"); see also Economics of Regulation, supra, at 89 ("if the same type of capacity serves all users, capacity costs as such should be levied only on utilization at the peak") (emphasis deleted).

Norwood's major criticism of the CP feature is that it cannot "know in advance exactly what the charges will be, at what times of day" — as it would with time-of-day rates that are higher at the system's historical peak than at off-peak times. We doubt that Norwood will long find it difficult to anticipate the daily CP. Even if the Town cannot do so "exactly," however, the Commission's prediction that "the use of coincident peak demand billing should facilitate the rate design goal of setting prices to more accurately track costs" is hardly unreasonable, 52 FERC at 61,337 — quite the contrary. That use of the CP is not the most convenient arrangement for a particular customer may be relevant but can hardly be dispositive.

In denying rehearing, the Commission elaborated upon its approval of CP demand billing, noting both that the utility is most vulnerable to reliability problems at the system peak and "that CP demand billing encourages conservation by discouraging consumption at the time of system peaks . . . thereby forestalling the need to install new capacity." 54 FERC at 61,206. Reliability problems and additions to capacity are real costs. Gearing price to responsibility for imposing those costs is a complete justification for using CP demand billing.

C. The Alleged Procedural Infirmity

The petitioner also claims that the Commission erred by shifting the burden of proof concerning the justness and reasonableness of the rate design from the utility proposing it to the customer challenging it. See 16 U.S.C. § 824d(e) ("At any hearing involving a rate or charge sought to be increased, the burden of proof to show that the increased rate or charge is just and reasonable shall be upon the public utility"). The ALJ was indeed somewhat careless in remarking that "it is not clear that NEPCO's use of long run marginal costs in the context of a wholesale rate filing was unreasonable." 49 FERC at 65,035. The FERC examined this statement in context however, and found "no basis for Norwood's charge that the presiding judge unlawfully reversed the burden of proof." 52 FERC at 61,336. We agree.

The ALJ clearly held that the "rate design filed by NEPCO is just and reasonable. It is not unduly discriminatory, nor is it unlawful in any other way." 49 FERC at 65,036. The Commission similarly "conclude[d] that the methods used by NEP were reasonable and that the resulting rates are just and reasonable." 52 FERC at 61,336. After stating that conclusion the Commission added, again somewhat carelessly, that "given NEP's case, Norwood has not provided credible counter-evidence to support its contention that NEP's [proposed] rate is unjust and unreasonable." Id. As we read this observation, however, it is a rather perfunctory restatement of the Commission's finding that the rate design is "just and reasonable," and not an indication that the FERC improperly shifted the burden of proof to the petitioner.

III. CONCLUSION

The petitioner makes other more fragmentary arguments, really no more than variations on the themes discussed above, that do not warrant separate treatment in this opinion. Without further ado, therefore, the petition is

Denied.