United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued January 5, 2018 Decided April 6, 2018
No. 16-1284
UNITED STATES POSTAL SERVICE,
PETITIONER
v.
POSTAL REGULATORY COMMISSION,
RESPONDENT
NATIONAL POSTAL POLICY COUNCIL, ET AL.,
INTERVENORS
On Petition for Review of Orders
of the Postal Regulatory Commission
David C. Belt, Attorney, U.S. Postal Service, argued the
cause and filed the briefs for petitioner. Stephan J. Boardman,
Chief Counsel, U.S. Postal Service, entered an appearance.
Dana Kaersvang, Attorney, U.S. Department of Justice,
argued the cause for respondent. With her on the brief were
Michael S. Raab, Attorney, David A. Trissell, General Counsel,
Postal Regulatory Commission, Christopher J. Laver, and
Erica A. Barker, Attorneys.
Before: PILLARD, Circuit Judge, and EDWARDS and
WILLIAMS, Senior Circuit Judges.
2
Opinion for the Court filed by Senior Circuit Judge
WILLIAMS.
WILLIAMS, Senior Circuit Judge: Under the Postal
Accountability and Enhancement Act, Pub. L. No. 109–435,
120 Stat. 3198 (2006), the Postal Regulatory Commission is
authorized to set rate caps for the “market-dominant products”
of the United States Postal Service. (The act contrasts such
products, consisting most obviously of products where the
Postal Service enjoys a legal monopoly, such as first-class mail,
with the Postal Service’s “competitive” products.) The
Commission is to set “an annual limitation on the percentage
changes in rates” equal to the rate of inflation, 39 U.S.C.
§ 3622(d)(1)(A); the statute defines “rates” as “fees for postal
services,” 39 U.S.C. § 102(7).
In USPS v. Postal Regulatory Comm’n, 785 F.3d 740 (D.C.
Cir. 2015) (“USPS I”), we wrestled with the question of
whether, and if so under what circumstances, the Commission
could treat Postal Service changes in mail preparation
requirements as “changes in rates” subject to the cap.
We rejected the Postal Service’s theory that the statute
encompassed “only changes to the official posted prices of each
product,” id. at 751, saying that “the Commission may have the
authority under the price cap statute and regulations to consider
mail preparation requirement changes of the kind at issue in this
case as changes in rates,” id. at 755. But, mystified by the
Commission’s efforts to explain how it would decide when a
mailing requirement actually was a rate change, we found its
action arbitrary and capricious and remanded to the
Commission for it to “enunciate an intelligible standard and
then reconsider its decision in light of that standard.” Id. at 756.
In due course the Commission produced the order now
before us, Order No. 3047, Order Resolving Issues on Remand
3
(Jan. 22, 2016). The order rules that a mail preparation change
constitutes a change in rates if it results “in the deletion of a rate
cell” or “in the redefinition of a rate cell if the mail preparation
change causes a significant change to a basic characteristic of a
mailing.” Id. at 15.
Applying this standard, the Commission reaffirmed its
earlier decision that the proposed mail preparation change
constituted a change in rates. The Postal Service moved for
reconsideration, which the Commission denied. Order No.
3441, Order Resolving Motion for Reconsideration of
Commission Order No. 3047 (July 7, 2016). The Postal Service
again petitions for review.
We now find that the Commission’s new analysis adds no
discernible clarity to the reasoning it supplied on the last round
and that it rests on an unreasonable interpretation of “changes
in rates” that “goes beyond the meaning that the statute can
bear.” MCI Telecomm. Corp. v. AT&T Co., 512 U.S. 218, 229
(1994). “[U]nder the familiar standard of Chevron, . . . a
‘reasonable agency interpretation prevails’”; “[o]f course, ‘if
Congress has directly spoken to an issue then any agency
interpretation contradicting what Congress has said would be
unreasonable.’” Loan Syndications & Trading Ass’n v. SEC,
882 F.3d 220, 222 (D.C. Cir. 2018) (quoting Entergy Corp. v.
Riverkeeper, Inc., 556 U.S. 208, 218 n.4 (2009)). “Even under
Chevron, after all, agencies only ‘possess whatever degree of
discretion [an] ambiguity allows.’” Id. at 224 (quoting City of
Arlington v. FCC, 569 U.S. 290, 307 (2013)). We grant the
petition and vacate the orders.
* * *
Our 2015 decision recites the relevant background. See
USPS I, 785 F.3d at 744–50. We cover the same ground only
as necessary.
4
The Commission does not apply the rate cap by limiting
the rate for each product in isolation. Rather, it allows the
Postal Service to trade off above-inflation increases in the rate
of one product with below-inflation increases in the rate of
another product within the same class (or, of course, constant-
dollar decreases in rates). As a result, apart from market-driven
changes in volume, the Postal Service’s aggregate revenue for
each class of market-dominant products should increase no
faster than inflation. We explained in USPS I:
Thus, for example, the Commission’s rules ensure that the
Postal Service may not generate extra revenue beyond the
price cap by taking advantage of the different volume
levels of different products within a class to raise rates
unevenly while technically complying with the class-level
price cap. To achieve this, the Commission has
promulgated regulations specifying that the calculation of
a “change in rates” in a class should be weighted by the
mail volume of any given rate cell in a class. 39 C.F.R.
§ 3010.23(b). So, if the Postal Service has two rate cells
in a given class but one of them accounts for the lion’s
share of the mail volume, any increase in the rate for that
rate cell will be weighted according to volume when
determining its contribution toward the classwide rate
change cap.
Id. at 745. (The Postal Service does not question this aspect of
the Commission’s work.) Thanks to this interrelation of rate
cells, the Commission might have tried to integrate mail
preparation requirements into its authority over “changes in
rates” with the following argument: Where an increase in mail
preparation requirements for one cell will drive mailers to use
a higher-priced cell, the resulting increase in volume in the
latter should count against the rate cap. This is emphatically
not the road taken by the Commission. We identify this
approach not in order to offer any final judgment on it but to
5
indicate how treating a change in mail preparation requirements
as a rate change might, as a matter of arithmetic, be integrated
with the Commission’s system of volumetric assessment.
We now return to the history that has brought us here. As
of 2009, the Postal Service had three rates for bulk mail (listed
in descending order of price and ascending order of the related
mailing requirement’s stringency): (1) one for mail sent
without a barcode (the “nonautomation” rate), (2) one for mail
sent with either a POSTNET or Basic Intelligent Mail barcode
(the “standard automation” rate), and (3) one for mail sent with
a Full-Service Intelligent Mail barcode (the “discounted
automation” rate).
In January 2013 the Postal Service disallowed use of the
POSTNET barcode as a basis for obtaining the standard
automation rate. POSTNET Barcode Discontinuation, 77 Fed.
Reg. 26,185 (May 3, 2012). The Commission did not treat this
change as a change in rates.
In April 2013 the Postal Service sought to make a further
change: To qualify for the standard automation rate, mailers
would have to upgrade their barcodes to the Full-Service
Intelligent Mail variety, which would necessarily qualify them
for the “discounted automation” rate as well. Implementation
of Full-Service Intelligent Mail Requirements for Automation
Prices, 78 Fed. Reg. 23,137 (Apr. 18, 2013). The Commission
held this redefinition of requirements for the standard
automation rate to be a change in rates. Order No. 1890, Order
on Price Adjustments for Market Dominant Products and
Related Mail Classification Changes (Nov. 21, 2013). Its
explanation was that the barcode changes “require mailers to
alter a basic characteristic of a mailing in order for the mailing
to qualify for the same rate category for which it was eligible
before the change in requirements.” Id. at 18. It ruled that this
change in mailing requirements, coupled with changes in
6
nominal rates that the Postal Service had earlier proposed,
would in the aggregate violate the price cap. Id. at 35–37.
That ruling led to our decision in USPS I, to the
Commission’s new articulation of its theory and application of
that theory against the Postal Service, and thus to the petition
now before us.
* * *
Our 2015 decision laid down a marker for what might
qualify as rates and “changes in rates.” Time and again we tied
“rates” to payments by mailers to the Postal Service, and
“changes in rates” to changes in those payments. We started
by pointing to the controlling statute’s definition of “rates” as
“fees for postal services.” 39 U.S.C. § 102(7); USPS I, 785
F.3d at 751. We saw the central issue as being identification of
the circumstances where “the Service changed mail preparation
requirements that would have the likely effect of changing
rates paid by certain mailers for sending the same mailpieces
that they sent in the prior year.” USPS I, 785 F.3d at 746
(emphasis added). We said that under the barcode changes
mailers “who did not upgrade their systems to comply with the
full-service requirements would have to pay the higher,
undiscounted rates.” Id. at 747. We framed the basic statutory
issue as whether “changes in rates” encompassed not only
changes in posted prices but “also changes to the prices
actually applied to particular mailpieces, as the Commission
argues.” Id. at 751 (emphasis added). We saw the Commission
as seeking to regulate “changes in rates paid by mailers.” Id. at
752. And, since a change in posted prices and a change in mail
preparation requirements “can cause a change in the rates paid
by mailers,” id. at 753 (emphasis added), we accepted the
Commission’s reading where the Postal Service failed to show
anything in the statute or regulations requiring a distinction
between those two types of changes, see id.
7
We note that USPS I’s link of rates to amounts paid to the
provider is anything but unique. “The standard dictionary
definition of the term ‘rate’ (as used with reference to prices) is
‘[a]n amount paid or charged for a good or service.’” FERC v.
Elec. Power Supply Ass’n, 136 S. Ct. 760, 777 (2016) (quoting
Black’s Law Dictionary 1452 (10th ed. 2014)). It is “the
amount of money a consumer will hand over in exchange for”
a product or service. Id.
In 2015 we were baffled by the Commission’s contrasting
treatment of changes in bundling requirements for flat-shaped
mailpieces. We noted its acknowledgement that the
requirement “‘may result in some mailers paying higher prices’
because those mailers who did not change their shipping
methods would be forced into a higher rate cell,” but quoted the
Commission as saying these changes “do not count as changes
in rates because the requirements ‘do not change the basic
characteristics of a mailing.’” USPS I, 785 F.3d at 754. We
observed, “This is hard to fathom.” Id.
In its present order the Commission seems just as devoted
to its previous murky notion of changes in “basic
characteristics”: Its new formula for finding whether a
redefinition of a rate cell produces a change in rate turns on
whether a change in mailing requirements “causes a significant
change to a basic characteristic of a mailing.” The Commission
now defines a “basic characteristic of a mailing” as “what the
mailer sends” and “how the mailer sends it,” while measuring
significance “by assessing the operational adjustments and/or
costs required by the mailer for compliance with the new mail
preparation requirement.” Order No. 3047 at 15–17. This new
verbiage, which the Commission offers with no effort to link
the words to any clear concept of “changes in rates,” can hardly
be said to advance matters from its earlier reliance on “basic
characteristics” simpliciter.
8
We need not devote more effort to evaluating the
Commission’s diminutive increase in clarity because its
explanation suffers from a more fatal flaw: It seeks to expand
the Commission’s regulatory domain beyond any permissible
meaning of “rates” under § 102(7) and “changes in rates” under
§ 3622. In USPS I we found the latter section “ambiguous” on
the question whether “changes in rates” could mean something
more than changes to posted rates. 785 F.3d at 751–52. But of
course that ambiguity doesn’t give the Commission carte
blanche to treat anything as a change in rates. “[W]hile
ambiguity in a statute may imply a ‘delegat[ion] to the agency
[of] the power to fill those gaps,’ the agency must still stay
within the bounds of the delegation in promulgating regulations
under the statute.” Aid Ass’n for Lutherans v. USPS, 321 F.3d
1166, 1177 (D.C. Cir. 2003) (quoting Cty. of Los Angeles v.
Shalala, 192 F.3d 1005, 1016 (D.C. Cir. 1999)); see also City
of Arlington v. FCC, 569 U.S. 290, 307 (2013); Village of
Barrington v. Surface Transp. Bd., 636 F.3d 650, 659–60 (D.C.
Cir. 2011).
We have already reviewed at length USPS I’s equation of
rates with sums paid for a mailing. At no point did we suggest
that a change in mailing requirements might qualify as a change
in rates simply because it might change a mailer’s aggregate
cost for sending an item. The Commission offers no case
holding “rate” or “change in rate” to encompass such costs or
changes. No aspect of its new formulation even attempts to
assess whether a mail preparation change, “significant” or not,
will cause a mailer to pay more to the Postal Service.
Unlike the Commission, the Postal Service identifies
circumstances where a change in mailing preparation costs
would yield “a change in the rates paid by mailers,” USPS I,
785 F.3d at 753:
9
This is not to say that compliance costs are inherently
irrelevant to the issue of prices paid. . . . If a mailer
believes that a new mail-preparation requirement requires
it to incur compliance costs that are higher than the costs
associated with simply paying a higher rate, then a rational
customer will not pay those costs and instead will pay the
higher rate (or perhaps leave the mail entirely). If, on the
other hand, the customer believes that the compliance costs
are lower than the cost of paying higher rates, then the
mailer will comply with the requirement and its mail
volume will accordingly not shift from one rate category
to another.
Reply Comments of the United States Postal Service, at 9 (Aug.
31, 2015); see also Motion for Reconsideration of Order No.
3047, at 11 (Feb. 22, 2016) (“The Commission, however,
makes no attempt to explain why the magnitude of a mail
preparation change is the proper basis for determining whether
mailers will shift to paying higher rates.”).
Curiously, the Commission recognized in briefing USPS I
that a mail preparation requirement change would present
mailers with a choice between compliance and non-
compliance. See Resp. Br. at 16, USPS I, 785 F.3d 740 (D.C.
Cir. 2015) (No. 13-1308) (“And those mailers who previously
had used basic barcoding now had to choose between paying
higher prices, or making processing changes that the Postal
Service itself acknowledged would be ‘significant.’”). Before
us, however, it disclaims any obligation to take account of such
choices.
Because the Commission blocked the Postal Service’s
proposed barcode change, we are able to examine mailer
behavior under the rate structure existing at the time. It
demonstrates the Commission’s drastic implicit underestimate
of mailer readiness to adopt the Postal Service’s most costly
10
(for mailers) barcodes. At the time of its original proposal in
2013, 64% of all eligible bulk mail used the Full-Service
Intelligent Mail barcode and therefore qualified for the
discounted automation rate. The Commission calculated that
the entire remaining 36%—the mail using only the Basic
Intelligent Mail barcode—would necessarily pay higher rates if
it had accepted the proposed change in standard automation
eligibility requirements (requiring Full-Service Intelligent Mail
barcodes). But just two years later (without the proposed
changes taking effect and so far as appears with no other
alteration of mailers’ incentives), Full-Service barcode
utilization jumped to over 88%. Put another way, barcoded
mail not enjoying the most discounted rate fell from 36% to
12%. Had the Postal Service proceeded in the face of the
Commission’s ruling, it would have been docked over $855
million in price cap authority under the assumption that it
would receive that amount in increased prices. The Service
would never have seen any of it.
Because the rate cell “redefinition” aspect of the
Commission’s new theory makes no effort to single out mail
preparation changes that induce mailers to shift to a higher-
priced service, it goes beyond “the bounds of the delegation”
identified in USPS I. Aid Ass’n for Lutherans, 321 F.3d at
1177.
The Commission’s alternative—asking if a rate cell has
been deleted—fares no better. First, there is no impermeable
line between redefinitions and rate cell deletions. Nearly all
redefinitions can be alternatively characterized as deletions.
Take the facts of this case. Before the proposed barcode
change, there were three rates. The Postal Service’s new
insistence on use of the Full-Service Intelligent Mail barcode
for the standard automation rate could be seen as either
redefining or deleting that rate. It redefined the rate in the sense
of requiring the Full-Service barcode for mailings to qualify for
11
it. And it deleted the rate in the sense that no mail would
henceforth be charged the standard automation rate: mailers
would pay either the nonautomation rate or the discounted
automation rate. We do not see how the semantics of labeling
in any way saves the Commission’s interpretation or ties it any
more closely to the regulatory authority as defined by § 3622.
Further, the Commission’s deletion theory eschews any
measure of significance altogether. The redefinition concept
attempted a measure of significance, but it used an absolute
metric that focused solely on the magnitude of costs and was
therefore of no value in predicting possible mailer migration to
higher-priced products. The deletion concept goes further,
making no attempt to measure significance at all.
In defense of both its redefinition and deletion formulae,
the Commission offers two principal responses, neither of
which alters our conclusion that its order has adopted a notion
beyond the bounds of the relevant statutory phrase: “changes in
rates.”
The Commission first points to one of its regulations, 39
C.F.R. § 3010.23, which describes how the Commission will
calculate the magnitude of a change in rate. That section,
discussed at the outset of the opinion, weights any change in
rates by the volume of mail for the product in question, and uses
an assumption that volume is constant, deriving the data “from
the most recent available 12 months.” Id. § 3010.23(d)(1).
Modifications to the volume inputs are permitted in some
circumstances: “The Postal Service shall make reasonable
adjustments to the [volume levels] to account for the effects of
classification changes such as the introduction, deletion, or
redefinition of rate cells.” Id. § 3010.23(d)(2). “Whenever
possible, adjustments shall be based on known mail
characteristics or historical volume data, as opposed to
forecasts of mailer behavior.” Id. § 3010.23(d)(3).
12
The Commission effectively argues that if its order
complies with its regulation, that is the end of the matter. The
Commission reminds us that § 3010.23 speaks of deletions and
redefinitions, and privileges “historical volume data” over
“forecasts of mailer behavior” whenever “possible.” Because
the Commission asserts that the disputed barcode change is
properly characterized as either a redefinition or a deletion, and
because the Commission has adopted a test that avoids mailer
forecasts, the Commission contends that it need not show
anything further to establish that the barcode rule would effect
a rate change.
The Commission’s reliance on its regulation plainly cannot
justify its giving the statutory “changes in rates” a meaning
outside the range of genuine ambiguity. The Commission
recognized as much in USPS I, observing, “[W]hether or not
the Commission correctly interpreted its own regulations has
no bearing on the separate question of whether the Commission
correctly interpreted 39 U.S.C. § 3622(d).” Resp. Br. at 25,
USPS I, 785 F.3d 740 (No. 13-1308). Further, the regulation
does not seem relevant. Section 3010.23, titled “Calculation of
percentage change in rates,” appears designed simply to govern
that process. The statutory question is how to identify when a
mail preparation change can be deemed to effect a change in
rates; the regulation answers a separate and subsequent
question about how to calculate the magnitude of such a
change.
It is true that we were called on to interpret § 3010.23(d)
in our 2015 decision. USPS I, 785 F.3d at 752–53. The
question we confronted was whether the regulation itself
unambiguously barred the Commission’s interpretation that
“classification changes” could refer to mail preparation
changes. Id. at 752. We held that it did not. But we did not
say that the Commission could treat as a rate change every mail
preparation change that could be somehow shoehorned into the
13
text of § 3010.23(d)(2). To the extent that in 2015 we affirmed
the historical volume rule (i.e., the assumption, for purposes of
calculating the extent of a rate change, that volume is constant
based on the prior year’s data), we did so in the context of the
question of how to calculate a rate change—not in asking
whether a rate change has occurred in the first instance. Id. at
756; Pet. Br. at 49–54, USPS I, 785 F.3d 740 (No. 13-1308);
Resp. Br. at 33–40, USPS I, 785 F.3d 740 (No. 13-1308).
The Commission also contends that it would be infeasible
to compare the costs of compliance with the rate increase a
mailer would face if it failed to comply with a new requirement.
We offer two responses. First, even if these practical concerns
were valid, they couldn’t trump a statutory limit. The
Commission must first confine its oversight to “changes in
rates” and then, within that realm, craft a workable standard.
Looking solely to mailer costs (as the Commission would
prefer), without comparing those costs to the additional
payment a mailer would avoid by making the mail preparation
change, is no basis for predicting that mailers will pay a higher
rate. Second, it is unclear whether the calculation the
Commission embraces (securing cost estimates from the
mailers and pegging them as “significant” or not) is necessarily
less onerous than comparing mailers’ compliance costs with the
offsetting rate benefit. If there is an advantage, it seems to lie
mainly in the Commission’s extravagant claim of discretion in
pinning the “significant” tag on a number (here it relies on
estimates of the one-time costs for “larger mailers” ranging
from $400,000 to $3.25 million, without even saying which
number makes the grade). The comparison required in order to
ascertain mailer movement to a higher rate might be more
complex, but mainly because the costs (however estimated)
would have to be compared with a benchmark—the rate
increment faced by mailers—that would be quite precise.
14
* * *
Our 2015 decision found that the Commission may have
authority to treat some Postal Service changes in mail
preparation requirements as changes in rates. But that potential
authority depends on its articulating and applying a test
consistent with the statute. Its present orders have failed to do
so.
We grant the petition for review and vacate the
Commission’s orders.
So ordered.