concurring in the judgment:
I am loath to write separately from the majority’s well-penned opinion, particularly given the oft-befuddling pleadings and record in this case. But it is because of that confusion that I am writing separately. FERC has not adequately explained what it has wrought. Thus, like the majority, I believe a remand is in order, but I am not yet willing to say that FERC’s orders are irredeemable, or that at this time we need resolve the scope of § 10(f)’s preemption. Instead, I think we should remand and let FERC better explain itself before we decide anything more.
The deep problem is that there is a logical disconnect between the front and back ends of FERC’s orders. The reasoning underlying the two parts seem to be in conflict. In particular, FERC (1) set forth a lengthy preemption analysis, but (2) then explained that “even to the extent that [the District’s assessment system] is preempted by section 10(f), [FERC] ha[s] no authority over the District’s actions.” Order on Rehearing, 119 FERC ¶ 61,141 at P 55. If by this explanation FERC meant that it lacks authority to compel its licensees to follow the Federal Power Act, then that is obviously ridiculous. But if FERC meant something more subtle, which it very well might have, then it has not adequately explained itself.
As I read the record, I conclude FERC intended its remedies determination to be independent from its preemption determination. I simply do not know how else to read FERC’s Order on Rehearing. In it, FERC frankly acknowledged it found the “District’s assessment system” not “entirely compatible with section 10(f),” but nonetheless denied Albany any relief, even for that acknowledged incompatibility. Id. The majority remands, believing its holding that § 10(f) requires total preemption materially changes this case’s context, but that interpretation of FERC’s orders does not comfortably jibe with what FERC said: “[E]ven to the extent that [the District’s assessment system] is preempted by section 10(f), [FERC] ha[s] no authority over the District’s actions.”
I have struggled to understand what FERC meant by saying preemption does not matter, and this is what I believe FERC might have intended: § 10(f) requires headwater benefits assessments paid from downstream licensees to upstream licensees for “interest, mainte*1082nance, and depreciation,” to be equitable. However, without a headwater benefits determination or settlement, FERC does not know what fees are appropriate under § 10(f), as it does not know how much the downstream beneficiary is benefited. A headwater benefits investigation' — -and thus determination — only can occur if a licensee requests it. Without a request, there is no determination; without a determination, FERC has no authority to forbid an upstream licensee from charging fees on a downstream licensee, because it does not know whether fees repugnant to § 10(f) are being assessed, and consequently the question of preemption is premature until FERC has determined head-water benefits. Thus, FERC said it had “no authority to prevent a storage project from attempting to assess charges from downstream projects under color of state law and in the absence of a Commission headwater benefits determination.” Order, 117 FERC ¶ 61,321 at P 55. (emphasis added).
Laying aside for the moment the question of whether this proposed logic is actually FERC’s position (or if it is, whether it is reasonable), for this analysis to be internally credible, a couple of things must be true. First, it does not matter how an upstream licensee labels a headwater benefits charge, but only the amount. In other words, if the District sends a “head-water benefits” bill to Albany for only “operations” costs, but the District is entitled under § 10(f) to the exact same amount of money for “interest, maintenance, and depreciation,” then there would be no factual need to discuss preemption. It is only when a fee — assessed under color of state law — is greater in amount than what is required by § 10(f) that a question of preemption must be resolved.
Second, FERC need not conduct a head-water benefits investigation on its own accord. This clearly is FERC’s position, but it is a curious one under the statute.1 Nonetheless, no one in this case has challenged FERC’s narrow interpretation of its role, and it is likely that no licensee ever will. After all, if the expected benefit of a headwater benefits investigation is greater than the expected costs, a licensee will request one. But if not, no licensee will want one.
An interesting thing happens, however, when FERC does not conduct a headwater benefits investigation, but instead allows the licensees to negotiate a settlement, with FERC only conducting an investigation if one of the licensees requests it. A “windfall” is created, which must go to either the upstream or the downstream licensee. Who receives the windfall de*1083pends on which licensee is benefited by the status quo.
To illustrate, consider this hypothetical. Assume a river on which there is only one upstream licensee, and one downstream licensee, and they both have perfect information about each other, and about how much a headwater benefits investigation will cost.2 Assume further that a downstream licensee is, and will be in perpetuity, benefited $20,000 a year from the upstream licensee’s expenditures on “interest, maintenance, and depreciation” — costs clearly covered by § 10(f)— but that a headwater benefits investigation will cost the upstream and downstream licensees $250,000 each. For simplicity, assume under § 10(f) that the “equitable” rate of compensation between downstream and upstream licensees for headwater benefits is 100% (in fact, the percentage is lower, but that is of no consequence). Next assume that the interest rate is 10%, and that state law permits the upstream licensee to assess benefits from the downstream licensee.
If, in this hypothetical, the status quo favors the downstream beneficiary (i.e., by not allowing any charges until a headwater benefits investigation has happened or the licensees have settled), then a rational upstream licensee will not request a determination. By investing that $250,000, the upstream licensee will receive annual payments of $25,000, more than it would receive if it requested a determination. The downstream licensee will know this, and thus, in settlement conversations, will agree to pay nothing, because it will know the upstream licensee has no credible threat. The result is a windfall for the downstream licensee: it does not have to compensate the upstream licensee at all.
If the status quo is reversed (i.e., by allowing any charges until a headwater benefits investigation has happened or the licensees have settled), then the upstream licensee receives the windfall: it will request something less than $45,000 a year— say, $44,000 — from the downstream licensee (the $20,000 in headwater benefits, plus a share of the cost of a headwater benefits determination). If the request is greater than that, the downstream licensee will seek a determination. But as long as the amount charged is less than that, the upstream licensee will know that any threat by the downstream licensee to go to FERC for an investigation is empty. It would not be economically rational. Thus, the upstream licensee gains the windfall: the amount above the actual benefits.
Assuming that FERC need not undertake a headwater benefits investigation on its own accord, which again no party has contested, then which placement of the status quo is more reasonable under federal law? If a downstream licensee is favored by the status quo, then it would be “directly benefited by the construction work of another licensee,” but “the licensee so benefited” would not “reimburse the owner of such reservoir” for those benefits, contrary to § 10(f). If the status quo preference is reversed, the upstream licensee receives the windfall. But, importantly, that result is preferable for the downstream licensee as compared to the alternative possibility of universal headwa-ter benefits determinations. And because the upstream licensee is the creator of the benefit, it makes some intuitive sense that the windfall goes to it. Or at least that might be a reasonable reading of the statute. After all, if a downstream licensee *1084receives headwater benefits, then everyday the upstream licensee is not compensated is a violation of federal law.
So, back to the case at hand, what if FERC had responded to Albany’s complaint by saying something really simple? “Even if Albany is right that these charges may in fact be preempted, we will not get involved until Albany has requested a headwater benefits investigation, because we do not know for certain how much the District is entitled to under § 10(f). Without knowing how much the District is entitled to under § 10(f), we do not know whether there are even any charges beyond those the District is already entitled to, so, as a factual matter, there may be no need to discuss preemption at all.” On review, we would not rule on preemption. My understanding of the record is that FERC may have said essentially what I have just outlined, with the only difference being that instead of holding its tongue on preemption, FERC gave its view. Unfortunately, in giving its view on preemption in the front-end of its orders, FERC undermined the logic of its back-end analysis: that preemption is premature.
For instance, instead of saying that it is irrelevant how an upstream licensee labels its headwater benefit fees because it only matters whether the amount is greater than what is actually due under § 10(f), FERC expounded at length on preemption, treating operations costs as different than those for “interest, maintenance, and depreciation,” and explaining why operation costs are not preempted. In other words, FERC actually answered the question it need not have answered, and it did so in a way that contradicted the most plausible reason why an antecedent head-water benefits determination might have mattered. That is incoherence.
Next, instead of explaining why a head-water benefits determination was required before it would intervene on Albany’s behalf, FERC offered nothing but the most cursory of analysis. I have pieced together what I believe may have been FERC’s rationale, but I am not confident enough even to say “the agency’s path may reasonably be discerned.” Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. 281, 286, 95 S.Ct. 438, 42 L.Ed.2d 447 (1974). Instead, while I believe FERC meant something when it said that preemption did not matter to its decision to deny Albany relief, and while I may be able to hazard an educated guess, I do not feel comfortable in saying this is what FERC must have meant. Pablo Picasso purportedly said “If I spit, they will take my spit and frame it as great art.” Likewise here, I do not know if FERC has offered a plausible argument, or a Rorschach inkblot.
And it matters what FERC meant. If FERC intended to say that preemption was irrelevant without a headwater benefits determination because until there is that determination, no one knows whether the District has requested fees it is not entitled to under § 10(f), then there might be no need for this court to decide whether FERC erred in denying Albany’s requested relief because its preemption analysis went awry. Instead, it would be as if FERC said nothing at all about preemption. Thus, for purposes of judicial economy I would remand without deciding the scope of preemption, to let FERC explain itself anew, and better.
. Section 10(f) says “whenever any licensee ... is directly benefited by the construction work of another licensee ... the Commission shall require as a condition of the license that the licensee so benefited shall reimburse the owner of such reservoir ... for interest, maintenance, and depreciation ... as the Commission may deem equitable.” Then, “[t]he proportion of such charges ... shall be determined by the Commission. The licensees ... shall pay to the United States the cost of making such determination as fixed by the Commission.” Id.
Three times this short statute says “shall.” Thus, § 10(f) seems to requires that if a downstream licensee receives headwater benefits from an upstream licensee, FERC must, as a condition to licensing that downstream beneficiary, ensure that the upstream licensee is equitably compensated, with the licensees paying FERC for the costs of any requisite headwater benefits investigation. But FERC reads for itself a passive role. Unless one of the licensees requests a headwater benefits investigation, FERC does ... nothing, even when the licensees are in conflict as to the equitable assessment. See Order, 117 FERC ¶ 61,321 at P 55 (“[W]e do not undertake a headwater benefits determination for benefits from a non-federal storage project in the absence of a request to do so.”).
. Of course, in real life, there is imperfect information. But unless there is cause to believe that either the upstream or downstream licensee has a systematic advantage in obtaining accurate information, then this may not be much of a problem: if the same uncertainty is built into both sides of the equation, it cancels out.