Commissioner Richard Glick Concurring in Part and Dissenting in Part Regarding Public Utility ROE Methodology (Opinion No. 569)

Commissioner Richard Glick Statement
May 21, 2020
Docket No. EL14-12-004
Order: E-1 
Concurring in Part and Dissenting in Part Regarding Public Utility ROE Methodology (Opinion No. 569)

Today’s order is yet another twist in the Commission’s decade-long effort to adapt its methodology for setting public utilities’ return on equity (ROE) to the low-interest rate conditions that have prevailed since the late 2000s.  In that time, the Commission has proposed multiple different ways of dealing with the fact that its long-standing ROE methodology produces cost-of-equity estimates well below the ROEs it generally permitted public utilities to collect in the years before the Great Recession.  The Commission’s most recent attempt to address this issue, Opinion No. 569,[1] was far from perfect.  Nevertheless, I supported it because it represented a reasonable compromise that I hoped would bring some much-needed certainty and predictability to the Commission’s approach to setting public utilities’ ROEs. 

So much for that.  Today, we are once again changing course and revamping our ROE methodology.  And, in so doing, we are sacrificing whatever certainty Opinion No. 569 might have provided. 

In addition, I am particularly troubled that the Commission is portraying its change of heart as a dispassionate assessment of various technical questions—the comparative merits of one financial model, the right source of data for another, or the appropriate application of various assumptions.  It is hard for me to believe that anyone buys that this latest twist is a genuine reassessment of those technical minutiae or that those details are what led Chairman Chatterjee to express his eagerness to consider rehearing requests at the December 2019 Open Meeting, before those requests were even filed.  Instead, it appears that the Commission again has chosen a path directed by the results, in this case the perceived need to award a higher ROE, rather than the law and the facts. 

In fairness, it may be that the methodology established in Opinion No. 569 would yield ROEs that are too low.  And it may also be that the ROE established in this proceeding—10.02 percent—is a just and reasonable number.  But, even so, the Commission must be transparent about the factors driving its decisionmaking process.  If we think the ROEs set by the Commission’s methodology are too low—or, for that matter, too high—we ought to say so and explain our reasoning, rather than pretending to be concerned only with the technical details of our models, data, and assumptions.  Accordingly, I dissent in part because I do not believe that today’s order adequately justifies several of the changes it adopts, even if the end result is an appropriate number.

Finally, today’s order affirms the one aspect of Opinion No. 569 that merited a grant of rehearing.  Opinion No. 569 declined to order refunds for a period in which everyone agrees customers paid an unjust and unreasonable rate.  I continue to believe that decision was an abdication of our responsibility to protect consumers.  As a result, I also dissent from the portion of today’s order that affirms that decision.

*        *        *

The Commission Must Stop the Endless Fiddling with Its ROE Methodology

Between 2011 and 2015, various entities representing customers’ interests filed a series of complaints under section 206 of the Federal Power Act[2] (FPA) arguing that the base ROE available to transmission owners in ISO New England, Inc. and the Midcontinent Independent System Operator (MISO) was unjust and unreasonable.  In Opinion No. 531, the Commission addressed the first of those complaints, with its most significant findings being that “anomalous capital market conditions” required the Commission to consider a variety of financial models and that those models supported an elevated ROE.[3]  The Commission subsequently applied that approach to a similar complaint involving the MISO Transmission Owners.[4]  Shortly thereafter, however, the D.C. Circuit vacated Opinion No. 531, sending it back to the Commission and the Commission back to the drawing board.[5]  Following that remand, the Commission proposed to expressly rely on the four financial models considered in Opinion No. 531.[6]  A year later, in Opinion No. 569, we narrowed it to two models, while making a number of changes to how we implemented those models.[7]  Today, we’re back up to three models, with another round of tweaks to those models.[8] 

With the exception of the Commission’s finding of anomalous market conditions, which at least hinted at its real concern, the Commission’s various orders in this saga have suggested that each new iteration of its ROE methodology is a largely technical affair that turns on the Commission’s evaluation of discrete issues with the various financial models.  In so doing, the Commission has added new models,[9] removed some of those models,[10] tweaked some of those models,[11] introduced new inputs,[12] modified existing inputs,[13] introduced new screens,[14] modified existing screens,[15] and even altered how the Commission places the ROE within the zone of reasonableness.[16]  But, with each new twist, it becomes harder to buy that the Commission is genuinely reassessing the mechanics of each model rather than disagreeing with the ROE numbers those models produce.[17]

Today’s order is the culmination of all that.  Not long after completing a year-long process to re-evaluate our approach to setting ROEs following the D.C. Circuit’s decision in Emera Maine, the Commission is now once again charting a major change of course.  In so doing, the Commission is again portraying its change of heart as a technical matter based on its reassessment of a handful of discrete issues rather than what it is:  A determination that the old number was too low and now we need a higher one.

To be fair, I am sympathetic to the impulse to consider subjective factors.  The Commission’s approach to setting a just and reasonable ROE will often implicate broader policy considerations, equity, and other factors that cannot be captured in, for example, a discussion of dividend yields or the appropriate sources of growth rate calculations.  But while ROE policy will always be as much art as science, that is no excuse to pretend that art is science. 

If broader considerations, including policy goals, are preventing the Commission from settling on or consistently applying an ROE methodology, then we must acknowledge those goals and give the interested entities the chance to weigh in on them just as they do for the intricacies of dividend yields, growth rates, and the like.  All approaches to setting ROEs have their shortcomings, but the worst result by far is to continually fiddle with those approaches, undermining the certainty and predictability that help transmission owners make long-term investments.  If the Commission is going to purport to rely entirely on financial models to evaluate and set ROEs, it has to take those models at face value without second-guessing them when it does not like the results.    

In addition, today’s order illustrates the problems with disguising subjective policy considerations as technical determinations.  In a number of instances, the Commission is reversing determinations made in Opinion No. 569 using rationales that are far less convincing than those that supported the opposite outcome in Opinion No. 569.  Shifting from such strong arguments to such suspect ones underscores the extent to which subjective factors seem to be operating in the background while also opening the Commission up to considerable risk on judicial review, creating even more of the uncertainty we ought to be trying to minimize.

Take the example of the risk premium model.  Although Opinion No. 569 declined to utilize the risk premium model based on a long list of shortcomings, today’s order reverses course, adding it to the DCF and CAPM on which the Commission previously relied.  The record before us does not support that choice.

As an initial matter, and as explained in Opinion No. 569, the risk premium model does essentially the same thing as the CAPM by attempting to calculate the “premium that investors require over the risk-free rate of return.”[18]  Opinion No. 569 rightly pointed out that nothing in the record supports having two thirds of the Commission’s ROE methodology composed of such analytically redundant approaches.[19]  Today’s order tersely responds to that concern by asserting that the two models are “sufficiently distinct” since they use different inputs.[20]  But that ignores the point in Opinion No. 569 that the problem with relying on both models is that they replicate the same basic methodology, irrespective of their inputs.[21] 

Opinion No. 569 also explained how the risk premium model is, in most respects, just an inferior version of the CAPM in so far as it does not consider market-based cost-of-equity estimates[22] and introduces significant circularity concerns by relying on past judgments, which may not reflect the appropriate risk premium under current conditions.[23]  The Commission responds to those circularity concerns by contending that they are “mitigate[d]” by the fact that the Commission will average the results of the risk premium with the DCF and the CAPM, which do not present the same concerns.[24]  But observing that the Commission will also use models without significant circularity concerns is not a reasoned response to the argument that you should not use circular models in the first place.

In addition, the Commission convincingly explained in Opinion No. 569 how “the record contains insufficient evidence to conclude that investors rely on risk premium analyses utilizing historic Commission ROE determinations or settlement approvals to determine the cost of capital and make investment decisions.”[25]  The Commission noted that, while allowed ROEs are certainly important to investors’ decisionmaking, that does not suggest that investors’ perform anything remotely close to the analysis contemplated by the risk premium model—i.e., a backward looking comparison between riskless assets and allowed ROEs—when making their investment decisions.[26]  Today’s order now takes the opposite position, observing only that investors in regulated utilities expect to earn a return above a risk-free asset (which is obviously true) and that “investors . . . observe regulatory ROEs and how changes in authorized ROE levels could affect utility earnings” (which is equally obvious).[27]  It should go without saying that investors pay attention to ROEs earned by public utilities and expect them to be higher than debt backed by the U.S. government.  But neither of those self-evident statements provides any reason—much less substantial evidence—to believe that investors perform a risk premium analysis comparing past differences between risk free assets and Commission-allowed ROEs when evaluating whether to invest in Commission-regulated public utilities.   

And, finally, the risk premium model does not at all fit with the Commission’s new approach for evaluating whether an existing ROE is just and reasonable.  Opinion No. 569 established a framework for evaluating whether an existing ROE is just and reasonable based on ranges of presumptively just and reasonable results derived from the financial models used by the Commission.[28]  Unlike every other financial model used, or even considered by the Commission in Opinion No. 569,[29] the risk premium model produces a single point estimate of the just and reasonable ROE, not a zone of reasonableness.[30] 

Recognizing this “serious concern,” but nevertheless determined to fit a square peg into a round hole, today’s order resolves to “impute” the average width of the zone of reasonableness created by the DCF and CAPM methodologies to the risk premium model.[31]  For example, if the DCF and CAPM produce an average zone of 200 basis points, it seems that the Commission will just assume that the risk premium model does too.  Today’s order, however, does not point to any evidence suggesting that such imputation is appropriate or that any investors or financial experts have sanctioned the Commission’s method.  Presumably that is because the record lacks any evidence supporting such an odd repurposing of the risk premium model.[32]  After all, the Commission’s approach to using the risk premium in evaluating whether an existing ROE is just and reasonable is the equivalent of making someone a “custom” suit based on their siblings’ measurements:  Maybe it will fit, but there is no reason to believe that it will and, in any case, it misses the point of the exercise. 

In addition, today’s order adopts a series of equally unreasoned modifications to Opinion No. 569’s framework for conducting the first step of the section 206 inquiry.  As noted, Opinion No. 569 established a practice of dividing the zone of reasonableness into ranges of presumptively just and reasonable ROEs within the broader zone of reasonableness.[33]  In particular, the Commission created risk-adjusted “quartiles” of the zone of reasonableness centered on the three points that the Commission uses as the starting point for setting ROEs for utilities of different risk profiles[34]—the midpoint of the entire zone of reasonableness for average-risk utilities, the midpoint of the lower half of the zone of reasonableness for below-average risk utilities, and the midpoint of the upper half of the zone of reasonableness for above-average risk utilities.[35] 

The Commission justified the end points of each quartile by explaining that “[l]ogic dictates that the end points of those ranges should not be closer to the starting points for the ranges of utilities with different risk profiles than they are to their own starting point.”[36]  In other words, it would not make sense to presume that an existing ROE is just and reasonable if it was closer to the starting point used to set the ROE for a utility of a different risk profile than the starting point for a utility of the same risk profile.  The Commission’s quartile-based approach made sense given the emphasis that the Commission has historically placed on relative risk profiles when placing ROEs within the zone of reasonableness[37] and it ensured that the ranges of presumptively just and reasonable results were not just arbitrary sub-sections of the zone of reasonableness.

Today’s order abandons that well-reasoned approach and arbitrarily divides the entire zone of reasonableness into thirds, with each third providing a presumptively just and reasonable range of ROEs for certain utilities.  The Commission appears to suggest[38] that this maneuver is necessary to comply with the D.C. Circuit’s statement in Emera Maine that “the zone of reasonableness creates a broad range of potentially lawful ROEs.”[39]  But Emera Maine requires nothing of the sort.  Read in context, the quoted language stands only for the proposition that the Commission cannot prove that an existing rate is unjust and unreasonable simply by showing that its ROE methodology would produce a different number using current data.[40]  The court certainly did not suggest that every point within the zone of reasonableness must be presumptively just and reasonable for some utility, which is how today’s order appears to understand that language.  In any case, the quartile-based approach in Opinion No. 569 easily complied with even the Commission’s reading of the language in Emera Maine.  Because the ranges only represented presumptive findings, a public utility could still argue that an ROE outside those ranges was nevertheless just and reasonable based on other considerations,[41] making every ROE within the zone of reasonableness at least “potentially” just and reasonable.

And that’s just the start of it.  Recognizing that the decision to divide the zone of reasonableness into thirds obliterates the rationale for the ranges outlined in Opinion No. 569,[42] the Commission announces, without any explanation, that it will change the starting points it uses when setting the ROE for below- and above-average risk utilities to the midpoint of the lower third of the zone of reasonableness and the midpoint of the upper third of the zone of reasonableness, respectively.[43]  Now the tail is truly wagging the dog.  In Opinion No. 569, the Commission justified the ranges of presumptively just and reasonable ROEs based on the Commission’s longstanding approach to handling companies’ relative risk profiles, namely the use of the upper and lower midpoints for utilities of above- and below-average risk, respectively.[44]  In today’s order, the Commission uproots that longstanding approach, selecting entirely new starting points for placing ROEs within the zone of reasonableness in order to support its new ranges of presumptively just and reasonable results.  That gets it entirely backwards; the ranges of presumptively just and reasonable results should reflect how we set ROEs, not the other way around.  In any case, at no point in today’s order does the Commission explain why the new starting points themselves are an appropriate place to begin the process of placing the ROE for an above- or below-average risk utility within the zone of reasonableness.[45]  

Suffice it to say, the Commission has not justified its change of course with respect to either the risk premium model or its approach to step one of the section 206 inquiry.  Nevertheless, while I believe that Opinion No. 569 was a superior approach to setting ROEs, I also recognize that the roughly 10 percent ROE established in today’s order may well be a just and reasonable end result.[46]  In addition, for the reasons explained above, I firmly believe that the Commission must finally bring some certainty and predictability to how its sets transmission owner ROEs. 

The Commission Should Order Refunds for Unjust and Unreasonable Rates Paid by Consumers

I continue to disagree with the Commission’s refusal to order refunds for the fifteen-month refund period established pursuant to the Second Complaint.[47]  Throughout that period, customers within MISO paid an unjust and unreasonable ROE.  Nevertheless, the Commission refuses to order refunds on the specious basis that the FPA requires it to act as if the 10.02 percent ROE set in today’s order was in effect throughout that fifteen-month period.  In reality, however, customers actually paid a 12.38 percent ROE—a difference worth tens of millions of dollars—and nothing in the law requires us to pretend otherwise. 

The facts relevant to the issue of refunds are straightforward.  On November 12, 2013, multiple parties filed a complaint (First Complaint) alleging that the MISO Transmission Owners’ 12.38 percent ROE was unjust and unreasonable.[48]  The Commission set the matter for hearing and established a refund effective date of November 12, 2013 (the date the First Complaint was filed),[49] meaning that the 15-month refund period for the First Complaint lasted until February 12, 2015.[50]  On February 12, 2015, a different set of parties filed another complaint (Second Complaint) against the MISO Transmission Owners’ ROE.  The Commission again set the matter for hearing and established a refund effective date of February 12, 2015,[51] meaning that the 15-month refund period for the Second Complaint lasted until May 12, 2016.  Both proceedings were fully litigated and produced initial decisions by Administrative Law Judges.[52]  And, in both cases, the Commission did not get around to issuing orders on the initial decisions until well after both refund periods expired, meaning that customers paid rates reflecting a 12.38 percent ROE throughout both refund periods.[53] 

In today’s order, the Commission affirms its conclusion in Opinion No. 569 that the 12.38 percent ROE was unjust and unreasonable and it establishes a new just and reasonable ROE of 10.02 percent.  That is sufficient to order refunds for the refund periods established pursuant to both the First and Second Complaints.  To see why, let’s start with the text of section 206(b), which provides that

At the conclusion of any proceeding under this section [i.e., section 206], the Commission may order refunds of any amounts paid, for the period subsequent to the refund effective date through a date fifteen months after such refund effective date, in excess of those which would have been paid under the just and reasonable rate, charge, classification, rule, regulation, practice, or contract which the Commission orders to be thereafter observed and in force.[54]

All that text requires is that the Commission find that customers paid an unjust and unreasonable rate during the refund period and that the Commission have set a just and reasonable replacement rate, so that it can calculate refunds equal to the difference between those two rates.  Both conditions are satisfied here:  Customers paid 12.38 percent through the Second Complaint refund period and the Commission has determined that they should have paid 10.02 percent.  That is sufficient to order refunds pursuant to section 206(b).

Contrary to the suggestion in today’s order,[55] the text of section 206(b) does not limit the Commission’s refund authority to only those individual proceedings in which it sets a new rate.  Instead, it provides the Commission with the authority to order refunds “[a]t the conclusion of any proceeding under this section”—i.e., section 206.”[56]  Congress surely understood that not every section 206 proceeding would be resolved against the public utility and, had it so desired, it could have conditioned the Commission’s refund authority accordingly.  But by pairing the word “conclusion”—which would seem to contemplate proceedings in which the public utility prevailed as well as those in which it did not—with the phrase “any proceeding”—which is equally unlimited—Congress rejected such a narrow interpretation of the Commission’s refund authority.  Instead, as noted, the plain text of section 206 indicates that the Commission’s refund authority turns on the presence of a difference between the unjust and unreasonable rate that customers paid during the refund period and the just and reasonable rate that they should have paid, not whether the Commission set a new rate in every complaint it resolves.   

Recognizing that Congress did not explicitly limit the Commission’s refund authority, the Commission responds that it did so implicitly when it inserted the phrase “thereafter observed and in force” in section 206(b).[57]  The idea, as I understand it, is that “thereafter observed and in force” is supposed to reflect Congress’ understanding that the Commission would be setting a new rate in each complaint prior to ordering any refunds.[58]  Thus, the argument appears to go, the Commission cannot order refunds unless it sets a new rate in the complaint corresponding to each individual refund period. 

As an initial matter, that would be a remarkably convoluted way of limiting the Commission’s refund authority under section 206.  It envisions that, instead of limiting the Commission’s refund authority in the statutory text that establishes the proceedings in which the Commission can order refunds, Congress elected to do so through an opaque reference in the discussion of how the Commission should calculate any refunds that it may order.  That is a bizarre—and overly complicated—way to read an otherwise straightforward statute.[59] 

In any case, the “thereafter observed and in force” language is better read as a reference to the identical language in section 206(a).[60]  Under that reading, all that “thereafter observed and in force” does is clarify that the ceiling on the Commission’s refund authority under section 206(b) is the difference between the rate in effect during the refund period and the just and reasonable rate that the Commission established pursuant to subsection 206(a).[61]  In other words, that language specifies how the Commission should calculate any refunds it orders, not when it may order refunds.  As noted, my reading makes far more sense given the fact that the “thereafter observed and in force” language appears in the portion of 206(b) that defines how the Commission should calculate refunds, not when it should order them.  I see no reason to abandon that straightforward reading of the statute, which protects customers from paying unjust and unreasonable rates, in favor of a convoluted one that does not.[62] 

The Commission’s next argument is even more of a head scratcher.  The Louisiana Public Service Commission argues that it is irrational to use the ROE set in Opinion No. 569 as the baseline for evaluating whether to order refunds for the Second Complaint refund period because that ROE was never “demanded, observed, charged, or collected,” as section 206 requires.  The Commission responds with what might charitably be called a regulatory fiction.  It argues that Opinion No. 569 made the new just and reasonable ROE set in the First Complaint proceeding effective as of the beginning of the First Complaint refund period, which, the Commission argues, means that we must pretend that that lower ROE was in effect throughout the refund period for the Second Complaint as well.  The Commission seems to be suggesting that it must pretend that the 10.02 ROE established today was “demanded, observed, charged, or collected” during the second refund period.[63] 

But that interpretation is both demonstrably false and squarely foreclosed by section 206.  First and foremost, the ROE that the MISO Transmission Owners collected during the refund period for the Second Complaint was 12.38 percent, no ifs, ands, or buts.  In addition, the FPA flatly prohibited the MISO Transmission Owners from collecting any other ROE during that period.  As noted, section 206 is forward looking in that it gives the Commission the ability to set a new just and reasonable rate as of the date on which the Commission makes the findings required by section 206.[64]  The only exception to that rule is for the refund period, during which time the Commission is permitted to act as if the new rate were in effect when ordering refunds.[65]  The refund period for the Second Complaint, however, fell after the conclusion of the refund period for the First Complaint and before the date on which the Commission issued Opinion No. 569.  Suffice it to say, it is arbitrary and capricious for the Commission to assume that it did that which it is legally prohibited from doing.

The Commission’s next argument is that ordering refunds for the Second Complaint would represent an end-run around the 15-month limitation on refunds enshrined in section 206(b).[66]  That argument appears to have both a legal dimension and a policy dimension.  Beginning with the former, the Commission seems to be taking the position that ordering refunds in the Second Complaint period would effectively extend the refund period established for the First Complaint.  But the Commission has repeatedly held that the FPA permits such successive or “pancaked” complaints, which are “‘entirely new proceeding[s]’” and not “‘duplicative proceeding[s] intended solely to expand the amount of refund protection beyond 15 months,’”[67] provided that they raise new facts or arguments,[68] which the Commission held that the Second Complaint did.[69]  Accordingly, rather than extending the refund period for the First Complaint, ordering refunds pursuant to the Second Complaint would simply reflect the fact that the MISO Transmission Owners collected an unjust and unreasonable ROE during a period when all parties were on notice that the Commission might order refunds of such excessive rates.[70] 

From the perspective of public policy, I recognize that permitting pancaked complaints with multiple refund periods may be sub-optimal.  After all, pancaked complaints can create significant uncertainty in an area where certainty is especially important as transmission owners decide whether and how to invest in transmission infrastructure.  But the desirability of pancaked complaints is something for Congress to consider, not a reason for us to twist the text of the FPA.  So long as the FPA and the Commission’s precedents permit pancaked complaints, then the Commission should not let its antipathy toward such complaints prevent customers from receiving the refunds to which they are entitled.

For these reasons, I respectfully concur in part and dissent in part.