PUBLISHED


UNITED STATES COURT OF APPEALS
             FOR THE FOURTH CIRCUIT


NATIONAL ELECTRICAL                  
MANUFACTURERS ASSOCIATION,
                       Petitioner,
               v.
UNITED STATES DEPARTMENT OF
                                     
ENERGY; UNITED STATES OF
AMERICA,                                   No. 10-1533
                    Respondents.


NATURAL RESOURCES DEFENSE
COUNCIL,
  Amicus Supporting Respondents.
                                     
          On Petition for Review of an Order of
               the Department of Energy.
              (EERE-2007-BT-STD-0007)

                  Argued: May 12, 2011

                Decided: August 16, 2011

   Before KING, SHEDD, and WYNN, Circuit Judges.



Petition for review denied by published opinion. Judge King
wrote the majority opinion, in which Judge Wynn joined.
Judge Shedd wrote a dissenting opinion.
2                      NEMA v. DOE
                        COUNSEL

ARGUED: John Andrews Hodges, WILEY REIN, LLP,
Washington, D.C., for Petitioner. H. Thomas Byron, III,
UNITED STATES DEPARTMENT OF JUSTICE, Washing-
ton, D.C., for Respondents. ON BRIEF: Clark R. Silcox,
NATIONAL ELECTRICAL MANUFACTURERS ASSOCI-
ATION, Rosslyn, Virginia; Eric Andreas, WILEY REIN,
LLP, Washington, D.C., for Petitioner. Scott Blake Harris,
General Counsel, Daniel Cohen, Assistant General Counsel,
Michael Kido, Bettina Mumme, UNITED STATES
DEPARTMENT OF ENERGY, Washington, D.C.; Tony
West, Assistant Attorney General, Michael S. Raab, UNITED
STATES DEPARTMENT OF JUSTICE, Washington, D.C.,
for Respondents. Timothy D. Ballo, EARTHJUSTICE, Wash-
ington, D.C.; Katherine Kennedy, Christine Chang, NATU-
RAL RESOURCES DEFENSE COUNCIL, New York, New
York, for Amicus Supporting Respondents.


                         OPINION

KING, Circuit Judge:

   The National Electrical Manufacturers Association
("NEMA") petitions for review of a final rule promulgated by
the United States Department of Energy (the "DOE") setting
forth energy conservation standards for electric induction
motors ranging in power output from .25 to 3 horsepower. See
Energy Conservation Standards for Small Electric Motors:
Final Rule, 75 Fed. Reg. 10874 (Mar. 9, 2010) (codified at 10
C.F.R. pt. 431) (the "Final Rule"). In promulgating the Final
Rule, the DOE invoked its authority to establish energy con-
servation standards for "small electric motor[s]," a term
defined by the Energy Policy and Conservation Act (the
"EPCA"). See 42 U.S.C. § 6311(13)(G). NEMA contends that
the relevant statutory definition unambiguously excludes all
                             NEMA v. DOE                                   3
such motors exceeding 1 horsepower, as well as certain
motors rated at and less than 1 horsepower, from being regu-
lated as small electric motors. As explained below, because
the Final Rule embodies a permissible interpretation of the
statutory definition, we deny the petition for review.

                                     I.

   A "small electric motor" is a type of electric induction
motor, which is a machine that converts electricity into rota-
tional mechanical power. The more efficient electric induction
motors create more mechanical power using less electricity.
The Final Rule is the DOE’s response to the EPCA’s directive
that the DOE promulgate energy conservation (or efficiency)
standards to govern certain small electric motors. The motors
covered by the Final Rule are manufactured in two-, four-,
and six-pole configurations and range from .25 to 3 horsepower.1
The covered motors have a wide variety of industrial and
commercial applications, the largest of which are used in
pumping and in heating, ventilating, and air-conditioning
("HVAC") systems. The DOE estimates that the standards
pronounced in the Final Rule "will save approximately 2.2
quads . . . of energy over 30 years," which "is equivalent to
about 2.2 [percent] of total annual U.S. energy consumption"
and will "eliminate the need for approximately eight new 250-
megawatt . . . power plants" as well as "result in cumulative
greenhouse gas emission reductions of . . . an amount equal
to that produced by approximately 25 million new cars in a
year." Final Rule, 75 Fed. Reg. 10874, 10876 (Mar. 9, 2010)
(codified at 10 C.F.R. pt. 431).

   NEMA’s petition for review challenges whether certain of
the electric induction motors covered by the Final Rule fall
  1
    "Pole configuration" refers to the number of magnetic poles in an elec-
tric induction motor. The number of poles is directly related to the theoret-
ical maximum rotational speed at which a motor can operate, as measured
in revolutions per minute (rpm).
4                            NEMA v. DOE
within the statutory definition of the term "small electric
motor." Resolving this challenge requires us to examine a
highly technical regulatory framework that governs a complex
market and is characterized by fine technological distinctions.
Accordingly, we begin by sketching the regulatory architec-
ture of which the definition is a part, as well as the relevant
provisions of NEMA Standards Publication MG1-1987, the
industry publication to which the definition refers. We then
describe the process culminating in the promulgation of the
Final Rule and NEMA’s challenge thereto.

                                    A.

   The story underlying this litigation begins with the enact-
ment of the Energy Policy Act of 1992, Pub. L. No. 102-486,
106 Stat. 2776 (the "Energy Policy Act"). The Energy Policy
Act amended the EPCA, directing the DOE to promulgate
energy conservation standards for various products and equip-
ment. Among other things, the EPCA requires the DOE to
"prescribe, by rule, energy conservation standards for those
small electric motors for which" the DOE determines that
such standards are "technologically feasible and economically
justified, and would result in significant energy savings." 42
U.S.C. § 6317(b)(1)-(2).2 Such standards must satisfy the
criteria specified in 42 U.S.C. § 6295(o). See 

    2
     The energy conservation standards promulgated under the EPCA do
not apply to "any small electric motor which is a component of" certain
other covered products and equipment, such as residential air conditioners
and heat pumps, clothes washers and dryers, and commercial-packaged air
conditioning and heating equipment. 42 U.S.C. § 6317(b)(3).
   3
     Under the EPCA, any new or amended standard for small electric
motors must "be designed to achieve the maximum improvement in
energy efficiency . . . which the [DOE] determines is technologically fea-
sible and economically justified." 42 U.S.C. § 6295(o)(2)(A). The DOE
may not promulgate a standard that "is not technologically feasible or eco-
nomically justified," or that would not result in the "significant conserva-
tion of energy." 

The DOE is obliged to weigh several
benefits and costs to determine whether a standard is economically justi-
fied. See 

                             NEMA v. DOE                                    5
   The resolution of this challenge turns on the meaning of the
term "small electric motor," which, as provided by the EPCA,
"means a NEMA general purpose alternating current single-
speed induction motor, built in a two-digit frame number
series in accordance with NEMA Standards Publication MG1-
1987." 42 U.S.C. § 6311(13)(G). MG1-1987 is an industry
publication that sets forth specifications for electric motors
and generators, which collectively are called "machines." The
publication classifies machines in various ways, such as by
size, application, electrical type, and variability of speed. See
J.A. 13-22.4 Two of those classifications are relevant here:
size and application.

   With respect to classification by size, a machine may be a
"Small (Fractional) Machine," a "Medium (Integral)
Machine," or a "Large Machine." J.A. 13-14. Paragraph 1.02
of MG1-1987 provides that a motor is a "small machine" if it
"is either (1) a machine built in a two-digit frame number
series in accordance with [¶] 11.01.1 or (2) a machine built
in a frame smaller than that frame of a medium machine . . .
which has a continuous, open-construction rating at 1700-
1800 rpm of 1 horsepower." 

(emphases added).5 By
contrast, under ¶ 1.03.1, "[a]n alternating-current medium
machine is a machine (1) built in a three- or four-digit frame
number series in accordance with [¶] 11.01.2 . . . and (2) hav-
ing a continuous, open-construction rating up to and includ-
ing" 500 horsepower. 

-14 (emphasis added).

   MG1-1987 also classifies machines by application. Rele-
  4
     Citations herein to "J.A. __" refer to the contents of the Joint Appendix
filed by the parties in this appeal.
   5
     Subparagraph 11.01.1 of MG1-1987 governs the numbering of frames
for small machines, and predicates such numbering on physical size. That
subparagraph provides that "[t]he frame number for small machines shall
be the D dimension in inches multiplied by 16," J.A. 52, where the D
dimension is the distance from the "[c]enterline of shaft to bottom of feet"
of the motor frame, 

6                           NEMA v. DOE
vant here is ¶ 1.05, which defines the term "general-purpose
alternating-current motor" as

        an induction motor, rated 200 horsepower and less,
        which incorporates all of the following:

               1.   Open construction.

               2.   Rated continuous duty.

               3.   Service factor in accordance with
                    MG1-12.47.

               4.   Class A insulation system with a
                    temperature rise as specified in
                    MG1-12.42 for small motors or
                    Class B insulation system with a
                    temperature rise as specified in
                    MG1-12.43 for medium motors.

           It is designed in standard ratings with standard
        operating characteristics and mechanical construc-
        tion for use under usual service conditions without
        restriction to a particular application or type of appli-
        cation.

J.A. 14. Paragraph 1.05 of MG1-1987 thus limits general pur-
pose motors to a maximum of 200 horsepower and requires
that they be built in an open (rather than enclosed) construction.6
A subsequent version of MG1 expanded ¶ 1.05 to provide that
such motors may also be built in an enclosed construction.
Paragraphs 12.42 and 12.43, referenced in ¶ 1.05, make no
mention of horsepower rating.
    6
   "Open" or "enclosed" construction, as used in ¶ 1.05, refers to the type
of housing utilized in an electric induction motor. An open housing allows
outside air to pass through and cool the motor during operation, while an
enclosed housing does not circulate outside air.
                             NEMA v. DOE                                   7
   Two other aspects of the definition in ¶ 1.05 are important
for our inquiry. First, it references ¶ 12.47, which sets forth,
in Table 12-2, the service factor specifications for small and
medium general-purpose alternating-current electric induction
motors. See J.A. 72.7 Table 12-2, in turn, delineates small and
medium motors at certain horsepower ratings. Specifically,
for two-pole motors operating at 3600 rpm, 1 horsepower
motors are "small" while 1.5 horsepower motors are "me-
dium." For four-pole motors operating at 1800 rpm, a .75
horsepower motor is small while a 1 horsepower motor is
medium; for six-pole motors operating at 1200 rpm, a .5
horsepower motor is small while a .75 horsepower motor is
medium. Second, ¶ 1.05 requires a general purpose motor to
be designed in "standard ratings with standard operating char-
acteristics," which means that such motors must meet pre-
scribed performance characteristics, one example of which is
locked rotor torque as specified in ¶ 12.32. Subparagraphs
12.32.2 and .3 reflect the same delineations between small
and medium motors as found in Table 12-2.8

                                     B.

   In 2006, the DOE took the first step toward establishing
energy conservation standards for small electric motors when
it determined that such standards appeared to be technologi-
cally feasible and economically justified, and would result in
significant energy savings. See Determination Concerning the
Potential for Energy Conservation Standards for Small Elec-
tric Motors, 71 Fed. Reg. 38799 (published July 10, 2006)
(the "Determination"). In the Determination, the DOE identi-
  7
     The "service factor" of an electric induction motor "is a measure of the
overload capacity at which a motor can operate without thermal damage,
while operating normally within the correct voltage tolerances." Final
Rule, 75 Fed. Reg. at 10885.
   8
     Other performance characteristics are also specified in Parts 10 and 12
of MG1-1987, and those provisions do not specify ratings for small motors
exceeding 1 horsepower. See J.A. at 45-46, 64-72.
8                       NEMA v. DOE
fied one "key issue" as "the definition of a ‘small electric
motor’ and precisely which motors are covered by this rule-
making." 

The DOE consulted MG1-1987 to
address this question, observing that "the two-digit frame
series" specified in the statutory definition "encompasses
NEMA frame sizes 42, 48, and 56, and motors with horse-
power ratings ranging from [.25] to 3 horsepower." 

Determination, the DOE also recognized that the
statutory definition requires small electric motors to be
"NEMA general purpose . . . motor[s]," prompting the DOE
to state that "[t]he EPCA definition of a small motor is tied
to the NEMA Standards Publication MG1-1987 performance
requirements that NEMA has established for general purpose
motors, such as the minimum levels for breakdown and
locked rotor torque for small electric motors presented in
MG1-1987 paragraph 12.32." Determination, 71 Fed. Reg. at
38800. In light of those requirements, the DOE found "that of
the motors that satisfy the frame-size requirements of the
small-motors definition, only a subset satisfies the other per-
formance requirements of the definition." 

   In 2007, the DOE released a document announcing the
scope of coverage for its rulemaking concerning this matter.
This document identified relevant portions of MG1-1987,
including ¶ 1.05. In particular, the document addressed the
horsepower ratings of electric induction motors subject to the
rulemaking. It observed first that "[t]he MG-1 standards dis-
cussed in the [section of the document dealing with horse-
power ratings] are not explicitly referenced by the definitions
of" the other terms in MG1-1987. J.A. 82. The DOE then
remarked that, "[a]s seen in [Tables 10-1 and 10-2], MG1-
1987 identifies small induction motors as motors with horse-
power ratings from 1 millihorsepower up to 1 horsepower."

after public input and the DOE’s further study,
DOE issued a Notice of Proposed Rulemaking. See Energy
                         NEMA v. DOE                            9
Conservation Standards for Small Electric Motors: Proposed
Rule, 74 Fed. Reg. 61410 (proposed Nov. 24, 2009) (the "No-
tice"). The Notice addressed the import of MG1-1987 regard-
ing two aspects of the electric induction motors subject to the
rulemaking relevant here: horsepower rating and open or
enclosed construction. As to horsepower ratings, the Notice
observed first that the statutory definition "does not explicitly
limit the scope of coverage to certain horsepower ratings." 

The Notice stated further that "the small electric
motor industry generally considers 3 hp as the upper limit for
rated capacity of such motors." 

also addressed the comments of Earthjustice, an
environmental advocacy group, regarding the regulation of
electric induction motors built in an enclosed construction.
Earthjustice contended that DOE could regulate such motors,
even though ¶ 1.05 of MG1-1987 requires a general purpose
motor to be built in an open construction. Earthjustice based
its view on subsequent versions of MG1 having expanded the
definition of "general-purpose . . . motor" to include motors
built in both open and enclosed constructions. Earthjustice’s
view took as one premise that the statutory definition’s refer-
ence to MG1-1987 applied only to the frame size requirement
found in the definition’s second clause, and not to the term
"NEMA general purpose . . . motor" in the definition’s first
clause. See Notice, 74 Fed. Reg. at 61421. According to
Earthjustice, therefore, nothing prevented DOE from treating
enclosed motors as "NEMA general purpose . . . motor[s]."
NEMA disagreed with Earthjustice, and the DOE declared
that it "agrees with NEMA that the reference MG1-1987
applies to all facets of the statutory definition of a small elec-
tric motor." 

of time following issuance of the Notice
involved a comment period and an additional public meeting.
At the public meeting and in written comment, NEMA
objected on the ground that some of the electric induction
motors identified in the Notice exceeded what NEMA
10                     NEMA v. DOE
asserted were MG1-1987’s horsepower limitations for small
motors, and therefore fell outside the statutory definition.
Earthjustice also reiterated its view that enclosed motors
could be regulated in the rulemaking because the reference to
MG1-1987 does not unambiguously apply to the first clause
of the statutory definition. Other commenters echoed Earth-
justice’s position, urging the DOE to announce a broader rule.
See, e.g., J.A. 351-52, 355, 359-60.

  On March 9, 2010, the DOE promulgated the Final Rule.
The Final Rule set forth energy conservation standards for
certain electric induction motors ranging in power from .25 to
3 horsepower:

        (a) Each small electric motor manufactured . . .
     after March 9, 2015, shall have an average full load
     efficiency of not less than the following:
                       NEMA v. DOE                          11




10 C.F.R. § 431.446.

   The preamble to the Final Rule responded to NEMA’s
objections that some of the motors are too powerful to be reg-
ulated as small electric motors. Specifically, it stated:

       DOE understands that NEMA MG1-1987 does not
    provide ratings for small motors of the identified
    higher horsepower ratings. However, DOE does not
    believe this precludes certain higher horsepower rat-
    ings built in a two-digit NEMA frame consistent
    with NEMA MG1-1987 from coverage. In addition,
    upon review of NEMA manufacturer catalogs, DOE
    noted that two-digit frame size motors of higher
    horsepower ratings are commonly marketed as gen-
    eral purpose. DOE also observed from NEMA ship-
    ment data provided to DOE for the determination
    analysis that when NEMA surveyed its members and
    requested shipments of general purpose motors built
    in a two-digit frame number series, responding man-
    ufacturers provided shipments data in horsepower
12                      NEMA v. DOE
     ratings exceeding those listed in [NEMA’s] com-
     ments above. Although NEMA argued that these
     motors do not fall within this rulemaking, NEMA
     did not deny that these motors are considered general
     purpose motors. Thus, DOE believes that even
     though NEMA MG1-1987 does not provide standard
     ratings for higher horsepower small electric motors,
     many of these motors are considered NEMA general
     purpose motors that could be considered for cover-
     age by DOE.

Final Rule, 75 Fed. Reg. at 10883.

   As recognized in the Final Rule’s preamble, the DOE deter-
mined which electric induction motors were within the statu-
tory definition by consulting materials from motor
manufacturers, including catalog entries that advertise "gen-
eral purpose" motors built in a two-digit frame series with
power outputs up to 3 horsepower. J.A. 420-30. The preamble
explained further that the Final Rule "does not codify a defini-
tion for ‘NEMA general purpose motor,’" although the "DOE
will consider proposing a definition for this term" in a future
rulemaking. Final Rule, 75 Fed. Reg. at 10886. The DOE also
reiterated that MG1-1987 informs the meaning of the term
"NEMA general purpose . . . motor" in the statutory defini-
tion, such that motors within the definition must satisfy MG1-
1987’s performance requirements. See J.A. 533-34.

   The preamble also addressed the remarks of Earthjustice,
which contended that the DOE could regulate electric induc-
tion motors with both open and enclosed construction by
interpreting "the phrase MG1-1987" in the statutory definition
of "small electric motor" to apply "only to the two digit frame
number series requirement" in the definition’s second clause.
Final Rule, 75 Fed. Reg. at 10882. DOE responded by reiter-
ating its "belie[f] that its scope of coverage in this final rule
is appropriate," 

MG1-1987 applies to "all
facets" of the definition, J.A. 537.
                            NEMA v. DOE                                13
                                   C.

   NEMA pursued its challenge to the Final Rule by filing this
petition for review. NEMA contends that the statutory defini-
tion of a small electric motor unambiguously excludes all
electric induction motors exceeding 1 horsepower, as well as
certain motors rated at and less than 1 horsepower. Specifi-
cally, NEMA’s objections track the delineations between
small and medium motors set forth in Parts 10 and 12 of
MG1-1987. See supra at 7 & note 8. Thus, NEMA maintains
that six-pole motors are not within the definition if they
exceed .5 horsepower; that four-pole motors are not if they
exceed .75 horsepower; and that two-pole motors are not if
they exceed 1 horsepower. NEMA further asserts that,
because these more-powerful motors fall outside the defini-
tion, we must vacate the Final Rule in its entirety, as those
motors were included in the calculations supporting DOE’s
determination that the Final Rule was warranted.

                                   II.

   The EPCA provides that "[a]ny person who will be
adversely affected by a rule prescribed under [42 U.S.C.
§ 6295] may . . . file a petition with the United States court
of appeals for the circuit in which such person resides or has
his principal place of business, for judicial review of such
rule." 42 U.S.C. § 6306(b)(1).9 We possess "jurisdiction to
review the rule in accordance with [5 U.S.C. §§ 701-706], and
to grant appropriate relief as provided in" those provisions. 

We are thus tasked with "decid[ing] all relevant
questions of law" and "interpret[ing] . . . statutory provisions"
at issue. 5 U.S.C. § 706. We are bound to "hold unlawful and
set aside" the Final Rule if we determine that it is "(A) arbi-
trary, capricious, an abuse of discretion, or otherwise not in
accordance with law; . . . (C) in excess of statutory jurisdic-
  9
   NEMA’s principal place of business is located in this Circuit, that is,
in Rosslyn, Virginia.
14                           NEMA v. DOE
tion, authority, or limitations, or short of statutory right; [or]
(D) without observance of procedure required by law." 

                                    III.

   The sole issue before us in this proceeding is whether
DOE’s interpretation of "small electric motor" is within the
range of permissible interpretations thereof. NEMA argues
that Parts 10 and 12 of MG1-1987 restrict the horsepower of
small electric induction motors, and that the statutory defini-
tion unambiguously incorporates this limitation.10

   In its papers before this Court, the DOE agrees with NEMA
that MG1-1987 applies to both clauses of the statutory defini-
tion, but does not agree that such an upper horsepower limita-
tion is so incorporated. The DOE contends that the
definition’s second clause, which expressly refers to MG1-
1987, requires merely that a small electric motor is one built
in an appropriate frame series. The DOE has interpreted the
term "NEMA general purpose . . . motor" in the definition’s
first clause to incorporate the requirements of MG1-1987
¶ 1.05. Paragraph 1.05, in turn, specifies (among other things)
that a general purpose motor must possess "standard ratings"
and "standard operating characteristics" found in Parts 10 and
12 of MG1-1987. The DOE points out, however, that the defi-
nition does not unambiguously incorporate the delineations
between "small" and "medium" motors found in Parts 10 and
  10
    The DOE does not dispute, and we agree, that if the more-powerful
motors do not fall within the statutory definition, we must vacate the Final
Rule in its entirety, and cannot simply strike down the standards for those
motors and leave in place the standards for the motors that NEMA con-
cedes fall within the definition. This is because DOE’s finding that energy
conservation standards are warranted is predicated on calculations includ-
ing the more-powerful motors. Moreover, there is no doubt that DOE
intended the Final Rule as "one, integral action," all the components of
which "must stand or fall together." North Carolina v. EPA, 

,
929 (D.C. Cir. 2008).
                        NEMA v. DOE                           15
12, and has declined to interpret the definition in such a way.
Rather, the DOE explains, a 3-horsepower motor built in a
two-digit frame series is a "small electric motor" for purposes
of the statutory definition, so long as it is general purpose and
possesses standard ratings and operating characteristics enu-
merated in Parts 10 and 12 — even if those ratings and oper-
ating characteristics correspond with those of a "medium"
motor.

  To evaluate the parties’ contentions, we must employ the
familiar framework of Chevron U.S.A. Inc. v. Natural
Resources Defense Council, Inc., 

(1984). We
will first describe the relevant precepts of the Chevron frame-
work, and then apply those principles to DOE’s interpretation
of the statutory definition.

                               A.

   We have explained that "Chevron deference is a tool of
statutory construction whereby courts are instructed to defer
to the reasonable interpretations of expert agencies charged by
Congress to fill any gap left, implicitly or explicitly, in the
statutes they administer." Am. Online, Inc. v. AT&T Corp.,

, 817 (4th Cir. 2001) (internal quotations and
emphasis omitted). Under Chevron, we first ask "whether
Congress has directly spoken to the precise question at issue,"
and "[i]f the intent of Congress is clear, that is the end of the
matter; for the court, as well as the agency, must give effect
to the unambiguously expressed intent of 

. If, however, "the statute is silent or ambiguous with
respect to the specific issue, the question for the court is
whether the agency’s answer is based on a permissible con-
struction of the statute." 

   Under the first step of Chevron, a reviewing court is to
"employ[ ] traditional tools of statutory construction" to deter-
mine whether Congress addressed "the precise question at

, 843 n.9. In this inquiry, we begin with
16                      NEMA v. DOE
the text and structure of the statute. See Cabell Huntington
Hosp., Inc. v. Shalala, 

, 986 (4th Cir. 1996). At
step one, we also bring to bear principles of grammatical
usage. See Barnhart v. Thomas, 

, 26 (2003).
While we ordinarily wade into a statute’s legislative history
only after deeming the statute ambiguous, see United States
v. Hatcher, 

, 226 (4th Cir. 2009), we have
described legislative history as one of the traditional tools of
interpretation to be consulted at Chevron’s step one. See, e.g.,
Elm Grove Coal Co. v. Dir., O.W.C.P., 

, 293-94
(4th Cir. 2007).

   Importantly, though, in consulting legislative history at step
one of Chevron, we have utilized such history only for limited
purposes, and only after exhausting more reliable tools of
construction. For example, in Elm Grove Coal, we consulted
legislative history solely to bolster our conclusion that the
agency’s interpretation was permissible — a conclusion
formed after a thorough examination of the statutory language
at issue and the applicable canons of construction. 

. NEMA has not identified, nor are we aware
of, any decision in which we relied on legislative history,
standing alone, to reject an agency’s interpretation of a statute
it administers. In this regard, we have consistently empha-
sized that "the plain language of the statute is . . . the most
reliable indicator of [c]ongressional intent." Soliman v. Gon-
zales, 

, 281-82 (4th Cir. 2005). Accordingly, fed-
eral courts are obliged to "avoid the pitfalls that plague too
quick a turn to the more controversial realm of legislative his-
tory." Lamie v. U.S. Tr., 

, 536 (2004).

   Where "Congress has [not] directly spoken to the precise
question at issue," 

, we "must . . .
proceed to the second step of the Chevron analysis," Capitol
Mortg. Bankers, Inc. v. Cuomo, 

, 155 (4th Cir.
2000). In this respect, the Supreme Court has instructed that
"[f]ew phrases in a complex scheme of regulation are so clear
as to be beyond the need for interpretation when applied in a
                        NEMA v. DOE                           17
real context." Nat’l R.R. Passenger Corp. v. Boston & Maine
Corp., 

, 418 (1992). We have proceeded to
Chevron’s second step where the statutory language "neither
plainly compel[led] nor clearly preclude[d] [an] interpreta-
tion," because in such circumstances the "precise import" of
the language "is ambiguous and certainly not free from
doubt." United Seniors Ass’n, Inc. v. Social Security Admin.,

, 403 (4th Cir. 2005) (internal quotation marks
omitted). Similarly, we have reached Chevron’s second step
after describing statutory language as "susceptible to more
precise definition and open to varying constructions." Md.
Dep’t of Health & Mental Hygiene v. Centers for Medicare
and Medicaid Servs., 

, 434 (4th Cir. 2008) (inter-
nal quotation marks omitted).

   If we determine that "the statute is ambiguous on the" pre-
cise question at issue, "we defer at [Chevron’s] step two to the
agency’s interpretation so long as the construction is a reason-
able policy choice for the agency to make." Nat’l Cable &
Telecomms. Ass’n v. Brand X Internet Servs., 

,
986 (2005) (internal quotation marks omitted). We afford
"controlling weight" to an agency’s reasonable interpretation
even where we would have, if writing on a clean slate,
adopted a different interpretation. Regions Hosp. v. Shalala,

, 457 (1998). We have deferred to an agency’s
interpretation of an ambiguous statute where the interpretation
"is grammatically correct and . . . accommodates the purpose
of the [enactment]." Gatlin Oil Co., Inc. v. United States, 

, 211 (4th Cir. 1999). The degree of deference
afforded the agency "is particularly great where . . . the issues
involve a high level of technical expertise in an area of rap-
idly changing technological and competitive circumstances."
Verizon Tel. Cos. v. FCC, 

, 909 (D.C. Cir. 2002)
(internal quotation marks omitted).

                               B.

  We now turn to whether the statutory definition unambigu-
ously forecloses the DOE’s interpretation. We must, of
18                      NEMA v. DOE
course, evaluate the permissibility of the DOE’s interpretation
by reference to the entirety of the definition, and not by read-
ing any portion thereof in isolation. See Elm Grove 

("[W]e must not be guided by a single sentence
or member of a sentence, but instead must look to the provi-
sions of the whole law, and to its object and policy." (internal
quotation marks omitted)). In this regard, we observe that the
statutory definition consists of two clauses — the first refer-
ring to a "NEMA general purpose . . . motor" (a term not
defined in either the statute or regulations), the second refer-
ring to the frame series in which a small electric motor must
be built. 42 U.S.C. § 6311(13)(G).

   For NEMA to prevail, the statutory definition must unam-
biguously incorporate the horsepower limitation that NEMA
contends is found in MG1-1987. In light of the language and
structure of the definition, as well as ordinary principles of
grammar, we conclude that the definition does not unambigu-
ously incorporate such a limitation. To begin with, we are not
free to disregard Congress’s decision to break the definition
into two clauses. Second, the express reference to MG1-1987
in the definition’s second clause does not unambiguously
incorporate all of that publication’s provisions, but rather only
its specifications for frame size. Third, the term "NEMA gen-
eral purpose . . . motor" in the definition’s first clause like-
wise does not unambiguously impose NEMA’s proposed
horsepower limitation. Finally, the available legislative his-
tory does not justify NEMA’s reading.

                               1.

  We begin by examining the definition as a whole. In this
regard, Congress’s deliberate choice to break the statutory
definition into two clauses is revealing. The definition first
requires a small electric motor to be a "NEMA general pur-
pose . . . motor," while the comma and following language
impose the additional requirement that such a motor is one
built in specified frame numbers. The addition of the second
                        NEMA v. DOE                           19
clause therefore differentiates small electric motors from oth-
ers on the basis of but one of MG1-1987’s specifications —
those provisions relating to frame size — and not its specifi-
cations relating to power output.

   Importantly, Congress had at its disposal many straightfor-
ward ways, had it so wished, to incorporate the horsepower
limitations that NEMA contends are found in MG1-1987 —
none of which it adopted. Cf. 

(refus-
ing to substitute "clear language" of statute "with . . . unenac-
ted legislative intent" where Congress "quite eas[ily]" could
have "includ[ed] limiting language" but did not do so (internal
quotation marks omitted)). Congress could have set forth an
express horsepower range within which small electric motors
must fall. Indeed, Congress took this very course elsewhere in
the same enactment, establishing statutory energy conserva-
tion standards for electric motors built in a three-digit frame
series ranging from 1 to 200 horsepower. See 42 U.S.C.
§ 6313(b)(1); J.A. 547.

   In the alternative, Congress could have simply omitted the
comma and reference to frame size from the statutory defini-
tion — much in the way that NEMA excerpts the definition
in its papers to this Court. See Petr.’s Br. 25 ("Congress
defined a small electric motor as a ‘NEMA general-purpose
[sic] alternating-current [sic] induction motor . . . built in
accordance with NEMA Standards Publication MG1-1987.’")
(omission in original; emphasis omitted); see also 

(same). Or, Congress could have opted for simpler language
and structure adopting the delineations between small and
medium motors found in MG1-1987 Parts 10 and 12, provid-
ing, say, that a small electric motor must comport with the rat-
ings and operating characteristics specified for small motors
in MG1-1987.

   Suffice it to say, Congress imposed no express horsepower
limitation; thought it important enough to add the comma and
reference to frame size to the definition; and did not see fit to
20                           NEMA v. DOE
adopt a simpler definition that incorporated MG1-1987’s pro-
visions wholesale. As to the absence of an express horse-
power limitation, other provisions of the same enactment
make clear that Congress knew how to impose such a limita-
tion when it so desired. And as to the presence of the comma
and reference to frame size, Congress went out of its way to
identify but one specification enumerated in MG1-1987
(frame size) while avoiding any reference to the publication’s
other specifications. NEMA thus undermines its reading of
the definition when it purports to quote the definition in its
papers. NEMA’s reading would, quite literally, excise the def-
inition’s specific reference to frame series and alter it to con-
tain only a single clause. In so doing, NEMA runs headlong
into the cardinal principle that "we have a duty, where possi-
ble, to give effect to all operative portions of the enacted lan-
guage, including its every clause and word." Shipbuilders
Council of Am. v. U.S. Coast Guard, 

, 244-45
(4th Cir. 2009) (internal quotation marks omitted). NEMA’s
reading would simultaneously render superfluous the first
clause’s use of the term "NEMA": if the language "NEMA
Standards Publication MG1-1987" in the second clause had
the sweeping reach that NEMA gives it, there would be no
need for the generic reference to "NEMA" in the first clause.

                                    2.

   Turning now to the express reference to MG1-1987 in the
statutory definition’s second clause, we conclude that its only
unambiguous import is to require small electric motors to be
built in an appropriate frame.11 First, the language of the sec-
  11
     Our dissenting colleague asserts that the DOE has, in essence, hood-
winked us into accepting a post hoc litigation position on the import of the
second clause’s express reference to MG1-1987. To the contrary, we
explain in considerable detail infra at 31-35 why the DOE’s arguments on
review are consistent with its underlying rulemaking approach. Left with-
out explanation, however, is the dissent’s insistence that NEMA’s inter-
pretation of MG1-1987, adopted for the express purpose of challenging
                           NEMA v. DOE                                21
ond clause is nearly identical to MG1-1987’s definition of a
small machine, and both distinguish small and medium
motors based on frame series (and thus physical size) rather
than power output. Compare 42 U.S.C. § 6311(13)(G)
(requiring, inter alia, that small electric motors must be "built
in a two-digit frame number series in accordance with . . .
MG1-1987"), with J.A. 13 (MG1-1987 ¶ 1.02) (providing that
one type of "small machine" is one "built in a two-digit frame
number series in accordance with [¶] 11.01.1"). Subparagraph
11.01.1, in turn, is a clear reference to physical size. See supra
at 5 & note 5. Paragraph 1.02 nowhere references the delinea-
tions between small and medium motors found in MG1-1987
Parts 10 and 12. Furthermore, ¶ 1.02 makes no reference to
maximum horsepower, while ¶ 1.03 limits medium motors to
500 horsepower.

   It follows that Congress did not aimlessly hurl the reference
to MG1-1987 into the definition, but rather inserted it with
care — first segmenting the definition into two clauses, only
the second of which references MG1-1987, and then choosing
language in the latter clause that mirrors the portion of MG1-
1987 relating solely to physical size. Moreover, the presence
of a horsepower limitation in the "medium motor" definition
— and lack of a corresponding limitation in the "small motor"
definition — indicates that MG1-1987 supplies horsepower
limitations expressly. Accordingly, the reference to MG1-
1987’s specifications for frame size cannot be said to unam-
biguously incorporate all provisions of the publication into the
definition, or to impose a 1-horsepower maximum.

the DOE’s rulemaking, is entitled to special weight on the ground that
NEMA authored the publication. See post at 43. The dissent’s skewed
point of view neglects the fundamental proposition that we are obliged to
defer to the reasonable legal interpretations of expert and politically
accountable executive agencies, affording no especial weight to the self-
interested constructs advanced by regulated parties — even those who
have drafted the particular language under examination.
22                      NEMA v. DOE
   Second, the familiar grammatical principle known as the
last-antecedent rule strongly counsels against reading the
express reference to MG1-1987 to apply unambiguously
beyond the second clause. The last-antecedent rule is "[a]n
elementary principle of statutory construction . . . which holds
that ordinarily a clause modifies only its nearest antecedent."
Commonwealth of Va. v. Browner, 

, 877 (4th Cir.
1996). In Barnhart v. Thomas, the Supreme Court applied the
last-antecedent rule to evaluate an agency’s interpretation of
a provision of the Social Security Act with a structure strik-
ingly similar to the definition at issue here. The relevant stat-
ute in Barnhart provided:

        "An individual shall be determined to be under a
     disability only if his physical or mental impairment
     or impairments are of such severity that he is not
     only unable to do his previous work but cannot, con-
     sidering his age, education, and work experience,
     engage in any other kind of substantial gainful work
     which exists in the national economy . . . ." [42
     U.S.C.] § 

(omission in original; emphases altered). The
Social Security Administration, as the agency charged with
administering the statute, interpreted the qualifier "which
exists in the national economy" to apply only to the "any
other kind of substantial gainful work" requirement, and not
to the "previous work" requirement.

   The Supreme Court’s Barnhart decision recognized that the
agency’s interpretation "is at least a reasonable construction
of the text and must therefore be given effect" under Chevron,
reversing the court of appeals’ ruling that the qualifier unam-
biguously applied to both 

. In so
ruling, the Court reasoned that applying the qualifier to both
of the requirements "disregards — indeed, is precisely con-
trary to — the grammatical ‘rule of the last antecedent.’" 

further explained that "[w]hile this rule is not an
                        NEMA v. DOE                           23
absolute and can assuredly be overcome by other indicia of
meaning, . . . construing a statute in accord with the rule is
quite sensible as a matter of grammar." 

marks omitted). It follows in this case that the reference to
MG1-1987 in the statutory definition’s second clause requires
merely that a small electric motor be built in the appropriate
frame size, and that the reference — standing alone — does
not unambiguously incorporate MG1-1987 Parts 10 and 12.

                               3.

   As the preceding discussion demonstrates, the reference to
MG1-1987 in the statutory definition’s second clause is too
flimsy to bear the weight that NEMA places on it. Conse-
quently, we turn to the import of the phrase "NEMA general
purpose alternating current single-speed induction motor,"
found in the definition’s first clause. Here, too, we conclude
that the DOE’s interpretation is not unambiguously foreclosed
at Chevron’s first step. We begin by reiterating that Congress
had much less roundabout ways of imposing a horsepower
limitation. Instead, by breaking the definition into two clauses
and referencing MG1-1987’s specifications for frame size
only in the second clause, Congress went out of its way to
incorporate but one specification in that publication — a spec-
ification that relates to physical size rather than power output.

    We also observe that, although Congress adopted MG1-
1987’s language relating to physical size in the second clause,
it used wholly different language — "NEMA general purpose
. . . motor" — in the first clause. Congress’s choice of terms
is instructive in at least two respects. First, the choice demon-
strates that Congress knew how to reference MG1-1987 when
it so desired, and it did not do so in the first clause. Second,
Congress’s use of a different and seemingly broader term in
the first clause suggests that, at minimum, the DOE could rea-
sonably read that term to be informed by MG1-1987 but not
to incorporate the delineations between small and medium
motors in Parts 10 and 12 of that publication. See Soliman,
24                          NEMA v. 

("Where Congress has utilized distinct terms
within the same statute, the applicable canons of statutory
construction require that we endeavor to give different mean-
ings to those different terms . . . .").

   Moreover, not only does the definition lack an express
horsepower limitation, but also ¶ 1.05 of MG1-1987 — on
which NEMA so heavily relies — limits only the maximum,
and not the minimum, horsepower of "general-purpose
alternating-current motor[s]." See J.A. 14 (stating that such
motors are "rated 200 horsepower and less"). NEMA must
therefore seek refuge in another requirement of ¶ 1.05 — that
such motors must be "designed in standard ratings with stan-
dard operating characteristics." 

directs us to
the ratings and characteristics of small and medium motors
found in Parts 10 and 12 of MG1-1987, which, according to
NEMA, limit the horsepower of small motors.

   NEMA’s contention is thus unconvincing. First, ¶ 1.05 on
its own terms provides only that general-purpose alternating-
current motors must possess some standard ratings and operat-
ing characteristics; it does not specify which ratings and oper-
ating characteristics those must be. And, having already set an
express 200-horsepower limitation for both small and medium
general purpose motors in ¶ 1.05 — and a 500-horsepower
limitation for medium motors in ¶ 1.03 — we would expect
that MG1-1987 would be more explicit about any horsepower
limitation applicable to small motors than the internal delinea-
tions in Parts 10 and 12 of the publication.12 Second, the very
presence of the second clause in the statutory definition sug-
gests that Congress delineated between "small" and not-
  12
     The dissent selectively quotes ¶ 1.05, see post at 38-39, excising its
only express reference to horsepower — that general purpose motors are
"rated 200 horsepower and less." Instead, the dissent, like NEMA, delves
into the internal delineations of MG1-1987 Parts 10 and 12. As explained
above, the express horsepower limitation in ¶ 1.05, with which the Final
Rule complies, is the superior guide to the import of MG1-1987 on the
statutory language.
                        NEMA v. DOE                          25
"small" motors on the basis of frame size as described in ¶¶
1.02 and 11.01.1 of MG1-1987, and not necessarily on the
basis of the delineations found in Parts 10 and 12. As such,
the statutory definition does not unambiguously incorporate
the distinctions between small and medium motors found in
Parts 10 and 12 of MG1-1987.

   We do not reach or decide the broader question of whether
the term "NEMA general purpose . . . motor" unambiguously
incorporates the 1987 version of MG1. First, because the
DOE has in fact interpreted the first clause to incorporate
some aspects of MG1-1987, whether the DOE was obliged to
interpret that clause to incorporate any aspect of MG1-1987
is not before us. Rather, the question we must decide is
whether "NEMA general purpose . . . motor" unambiguously
incorporates all aspects of MG1-1987. It does not.

   Second, Chevron’s mandate to accord government agencies
the flexibility to respond to changing conditions compels us
to reserve judgment as to MG1-1987’s ultimate import with
respect to the first clause. 

("An initial
agency interpretation is not instantly carved in stone. On the
contrary, the agency . . . must consider varying interpretations
and the wisdom of its policy on a continuing basis."). The
broader question of whether the term "NEMA general pur-
pose . . . motor" unambiguously incorporates MG1-1987
could be presented later if, for example, the DOE seeks to
regulate enclosed motors by adopting Earthjustice’s interpre-
tation of the statutory definition. See supra at 9. Because
enclosed motors are "general purpose" motors only under
more recent versions of MG1, declaring the first clause to
unambiguously and exclusively incorporate the 1987 version
could deprive the DOE of the flexibility to regulate enclosed
motors, which some groups identify as a significant compo-
nent of the relevant market. See, e.g., J.A. 351. Worse, such
a declaration could also deprive the DOE of the requisite flex-
ibility without providing the agency the opportunity to defend
26                      NEMA v. DOE
what it portrays as a "reasonable alternative interpretation."
Resp’t.’s Br. 38.

                               4.

   NEMA has also scoured the legislative history seeking sup-
port for its view of the statutory definition, but to no avail.
NEMA holds up a single sentence of a committee report
accompanying the House of Representatives version of the
legislation as proof that the definition unambiguously
excludes all electric induction motors exceeding one horse-
power. See H.R. Rep. No. 102-474, pt. 1, at 175 (1992),
reprinted in 1992 U.S.C.C.A.N. 1953, 1998. That report states
that the provision now codified at 42 U.S.C. § 6317, pursuant
to which the DOE promulgated the Final Rule, "requires DOE
to prescribe energy conservation standards for . . . electric
motors of less than one horsepower that it determines would
result in significant energy savings." 

unconvinced by NEMA’s contention in this regard.
Put simply, accepting NEMA’s contention would require us
to (1) rely on a negative inference from (2) a single sentence
of unenacted legislative history that (3) the statute unambigu-
ously incorporates the horsepower limitation NEMA urges,
even as (4) that very legislative history does not reflect the
same horsepower limitation, and (5) several more reliable
indicia of meaning strongly support the DOE’s interpretation.

   To begin with, the horsepower limitation NEMA identifies
in the report is not even an express one — while the report
contemplates the DOE’s regulation of sub-1 horsepower elec-
tric induction motors, it is silent with respect to other motors.
In its submissions here, NEMA candidly asks that we infer
the limitation based on the report’s "less than one horse-
power" language. Petr.’s Reply Br. 19. But even where we are
dealing with statutory language and not mere snippets of leg-
islative history, we will draw such a negative inference only
where it appears that Congress meant to exclude the unmen-
                         NEMA v. DOE                             27
tioned item. See INS v. Fed. Labor Relations Auth., 

, 272 (4th Cir. 1993) (declining to draw negative inference
where "statute displays not an effort at legislative micro-
management . . . , but rather a positive intention to leave com-
plexities and trivialities to" agency). Additionally, in those
Chevron cases where we have consulted legislative history,
we have demanded only that the agency’s interpretation be
"consistent" with the legislative history. Elm Grove 

. Here, DOE’s interpretation is consistent because
sub-1 horsepower motors, the motors discussed in the com-
mittee report, are among those regulated by the Final Rule.

   Finally, we observe that the legislative history itself belies
its unreliability. If we were to take seriously its statement that
all small electric motors must be "less than" 1 horsepower,
then certain motors that even NEMA concedes are small
under MG1-1987 Parts 10 and 12 (namely, 1-horsepower
motors in a two-pole configuration) would fall outside the
statutory definition. See Petr.’s Br. 20. On the other side of
the coin, NEMA does challenge the Final Rule’s inclusion of
certain motors (namely, .75-horsepower motors in a six-pole
configuration) that, because they are rated less than 1 horse-
power, meet the condition supposedly described in the legisla-
tive history. See 

of the incongruities between
MG1-1987’s delineations and the committee report, we do not
accept the report as reflecting a considered view of the intrica-
cies of regulation — many of which Congress expressly dele-
gated to the DOE in the Energy Policy Act — or a reliable
means to ascertaining the import of the language in the statu-
tory definition. Rather, the "uncertainties [in the report] illus-
trate the difficulty of relying on legislative history . . . and the
advantage of . . . rest[ing] our holding on the statutory text."

(declining to rely on legislative
history where available material "creates more confusion than
clarity about the congressional intent").

  Of whatever usefulness legislative history may be in other
Chevron cases, for the reasons discussed above, the available
28                       NEMA v. DOE
history is not helpful here. It is conceivable, for instance, that
where the application of textual canons and grammatical prin-
ciples leaves a handful of competing and similarly plausible
interpretations, strong legislative history against one interpre-
tation would "restrict the range of choices that the bare text
of the provision might otherwise seem to leave the adminis-
tering agency." Caleb Nelson, Statutory Interpretation 766
(2011) (discussing purposes for which reviewing courts
employ legislative history in Chevron cases). In General
Dynamics Land Systems, Inc. v. Cline, by way of example, the
Supreme Court acknowledged that the statute at issue was
susceptible to the agency’s interpretation "[i]n the abstract,"
yet refused to defer to the agency because the purpose of the
statute, "30 years of judicial interpretation" and legislative
inaction, as well as the legislative history, demonstrated "be-
yond a reasonable doubt" that the agency was "clearly
wrong." 

, 586, 590, 599-600 (2004).

   Alternatively, perhaps strong legislative history could sug-
gest an interpretation we would not glean from other tools of
construction, such that the legislative history "rehabilitate[s]
a reading that the courts would otherwise put off limits" to the
agency. 

; cf. Bellum v. PCE Construc-
tors, Inc., 

, 739 (5th Cir. 2005) (affording Chev-
ron deference to agency interpretation of statute that "is not
inherently ambiguous," but where contrary interpretation
"would frustrate the [statute’s] unmistakable purpose," as evi-
denced in part by legislative history). But NEMA would have
us invoke the legislative history not merely to choose between
competing interpretations not resolved by the text or canons
of construction, or to enable the agency to adopt an interpreta-
tion that fulfills the statutory purpose, but instead for the
novel and unprecedented purpose of overriding an interpreta-
tion that is strongly supported by more reliable interpretive
tools. We are leery of such an approach as a general proposi-
tion, and are more skeptical still of invoking the thin record
here to conclude that DOE’s interpretation is unambiguously
foreclosed.
                       NEMA v. DOE                        29
                             C.

   The second step of the Chevron analysis requires us to
determine whether the interpretation reflected in the Final
Rule "is a reasonable policy choice for the agency to make."
Brand 

(internal quotation marks omitted).
Because the DOE logically looked to market realities in for-
mulating its interpretation, we conclude that DOE’s interpre-
tation — that both clauses of the definition are informed by
MG1-1987, but that they do not incorporate the delineations
between small and medium motors found in MG1-1987 Parts
10 and 12 — is a reasonable policy choice.

                             1.

   In reaching its interpretation, the DOE recognized (and
agreed with NEMA) that the import of MG1-1987 informs the
meaning of the term "NEMA general purpose . . . motor" in
the statutory definition’s first clause. The DOE went on to
recognize, however, that Congress chose to break the defini-
tion into two clauses, the second of which singles out MG1-
1987’s specification relating to the physical size of "small
machine[s]." J.A. 13. Faced with the structure of the defini-
tion and the intricacies of MG1-1987, DOE commendably
looked to industry practice and market realities to guide its
understanding of the definition. See Determination, 71 Fed.
Reg. 38799, 38800 (published July 10, 2006) (explaining that
DOE "began the analysis for this determination by collecting
information from manufacturers of small motors and others"
and that, "[s]ubsequently, the Department received data and
information, including that provided by . . . [NEMA]").

  DOE’s investigation of the market revealed that electric
induction motor manufacturers, including some NEMA mem-
bers, regularly interpret the "general-purpose alternating-
current motor" definition in MG1-1987 ¶ 1.05 as DOE does.
That is to say, they manufacture and market 1-plus horse-
power motors in a two-digit frame number series (as specified
30                      NEMA v. DOE
in the MG1-1987 definition of small machine and the EPCA
definition of small electric motor), and conform those motors
to the standard ratings and operating characteristics applicable
to what MG1-1987 Parts 10 and 12 would call medium
motors. See J.A. 420-30 (manufacturer catalog excerpts).
Notably, NEMA itself made this point in its comments to
DOE. See J.A. 121 ("NEMA Design Types A, B, C, and D are
not defined for polyphase small electric motors. However, a
search of manufacturers’ catalogs will indicate that it is com-
mon for a polyphase 1 hp small electric motor to be designed
to the same performance requirements as a polyphase 1 hp
medium electric motor. This is expected, if not demanded, by
the motor users.").

   In light of the information gleaned from manufacturers and
NEMA itself, the DOE quite sensibly decided that the electric
induction motors subject to the rulemaking ought to include
motors actually manufactured, marketed, and sold as general
purpose motors and built in a two-digit frame number series,
even if the power output of those motors is not always identi-
fied as corresponding to the output of "small" machines as the
term is used in certain tables of MG1-1987. As the DOE
explains, "[t]he absence of ratings in some tables in MG1-
1987 does not keep such motors off the market; nor should it
keep such motors from complying with energy conservation
standards." Resp’t.’s Br. 33.

                               2.

    Finally, NEMA contends that the DOE’s position in this
proceeding is not entitled to deference because it is a post hoc
litigation position that contradicts the view it set forth in the
rulemaking process. See Bowen v. Georgetown Univ. Hosp.,

, 212-13 (1988). NEMA seizes on two sources
to contend that the DOE has contradicted itself in defending
the Final Rule in this challenge. First, NEMA points to lan-
guage in the 2007 scoping document, which states that, "[a]s
seen in [Tables 10-1 and 10-2], MG1-1987 identifies small
                        NEMA v. DOE                           31
induction motors as motors with horsepower ratings from 1
millihorsepower up to 1 horsepower." J.A. 82. Second,
NEMA relies on DOE’s response to Earthjustice in the Notice
and Final Rule that MG1-1987 applies to "all facets" of the
statutory definition. Notice, 74 Fed. Reg. at 61421; J.A. 537.

   We are unable to find NEMA’s contention persuasive. To
begin with, we are confident that the DOE’s arguments to this
Court are not at odds with the interpretation reflected in the
Final Rule. Rather, the DOE’s papers simply elaborate on the
legal theory underpinning the Final Rule’s interpretation of
the statutory definition. We are mindful that "an agency’s
action may not be upheld on grounds other than those relied
on by the agency" in the actual course of its decisionmaking.
Boston & 

(citing SEC v. Chenery
Corp., 

, 88 (1943)). But deference is appropriate,
and we need not remand, when the agency’s litigation papers
merely set forth an interpretation that was a "necessary pre-
supposition" of its underlying action. 

Supreme
Court has explained, "the fact that the [agency] did not in so
many words articulate its interpretation" in the underlying
action "does not mean that we may not defer to that interpre-
tation, since the only reasonable reading of" the explanation
offered by the agency in the underlying action, "and the only
plausible explanation of the issues that the [agency] addressed
after considering the factual submissions by all of the parties,
is that the [agency’s action] was based on" the interpretation
it proffered in subsequent litigation. 

the DOE advances here simply does not devi-
ate from the view it articulated in the course of the rulemak-
ing — that certain electric induction motors with outputs
between 1 and 3 horsepower, if built in a two-digit frame
series, are small electric motors. Cf. Coeur Ala., Inc. v. South-
east Ala. Conservation Council, 

, 2474-77
(2009) (rejecting argument that agency memorandum inter-
preting agency’s regulations contradicted agency’s prior
view). True enough, the DOE maintains that it is not required
32                      NEMA v. DOE
to interpret the term "NEMA general purpose . . . motor" to
incorporate all provisions of MG1-1987. See Resp’t.’s Br. 20-
21 ("The bare statutory reference to MG1-1987 at the end of
[the statutory definition] was not intended to incorporate all
elements of that industry standards publication into the stat-
ute. . . . Instead, the statute refers to MG1-1987 only in a lim-
ited context: those industry standards define the ‘two digit
frame number series’ in which a small electric motor is
built."); 

   Nevertheless, it does not follow that DOE’s position con-
tradicts the explanation it offered in the Final Rule. After all,
the DOE — both in this Court and in the Final Rule — has
interpreted both clauses of the statutory definition to refer to
and incorporate MG1-1987. See Resp’t.’s Br. 29 ("DOE has
interpreted the earlier part of the phrase — ‘NEMA general
purpose . . . motor’ — to refer to MG1-1.05, but not to incor-
porate the distinctions between small and medium machines
in Parts 10 and 12."); Final Rule, 75 Fed. Reg. at 10883.
Although the DOE has not interpreted the definition to incor-
porate the delineations between small and medium electric
induction motors in MG1-1987 Parts 10 and 12, the DOE has
never taken that position and, as explained in Part III.B, the
statutory language does not require it to do so.

   In support of its proposition in this respect, NEMA makes
much ado about language in the 2007 scoping document and
the DOE’s statements that MG1-1987 applies to "all facets"
of the statutory definition, but these are in the nature of red
herrings. Regarding the 2007 document, we observe first that
the language upon which NEMA seizes merely describes cer-
tain tables found in MG1-1987, and is not a conclusive (or
even preliminary) pronouncement on the import of those
tables for the meaning of the statutory definition. This is
borne out by the 2007 document’s disclaimer, found on the
same page, that "[t]he MG-1 standards discussed in [the sec-
tion dealing with horsepower ratings] are not explicitly refer-
enced by the definitions of any of the [MG-1] terms listed
                         NEMA v. DOE                            33
above." J.A. 82. Nor does the "all facets" language support
NEMA’s contention. That language was included to address
a separate question raised by Earthjustice — the regulation of
enclosed as well as open motors — and not the horsepower
ratings of motors subject to the rulemaking. No observer
could have mistaken these statements to refer to horsepower
ratings, as the statements were made in the same breath that
DOE interpreted the more-powerful motors to fall within the
statutory definition. Moreover, the "all facets" language is not
in tension with interpreting the definition to include more-
powerful motors, as the DOE in fact interpreted MG1-1987 to
apply (to some extent) to the first clause of the definition.

   To the extent that DOE did not state its interpretation in
exacting detail in the rulemaking — and elaborated on its
interpretation in arguments to this Court — NEMA’s argu-
ment yet flounders. The DOE was under no obligation to go
to such lengths to defend its interpretation at the time it pro-
mulgated the Final Rule; it is enough that the Final Rule and
the accompanying explanation set forth a coherent interpreta-
tion and a plausible rationale underpinning it. See Madison
Gas & Elec. Co. v. EPA, 

, 529 (7th Cir. 1994)
(articulating the "undemanding standard" that, under Chevron,
"[a]lthough the [agency] is free to choose any reasonable
interpretation for an undefined term in the statutes it adminis-
ters, it must exhibit the reasons for its choice and those rea-
sons must be at least plausible" (internal quotation marks and
citations omitted)). Indeed, the Supreme Court has instructed
that "[e]ven when an agency explains its decision with less
than ideal clarity, a reviewing court will not upset the decision
on that account if the agency’s path may reasonably be dis-
cerned." Ala. Dep’t of Envtl. Conservation v. EPA, 

, 497 (2004) (internal quotation marks omitted).

   We will not hold against the DOE the more sophisticated
legal arguments it sets forth in its submissions to this Court.
The DOE is tasked, in the first instance, with fulfilling its stat-
utory mandate. Of course, this responsibility includes provid-
34                      NEMA v. DOE
ing reasons for its actions sufficient to permit assessment by
a reviewing court. See Pension Benefit Guar. Corp. v. LTV
Corp., 

, 654; 

. Most
assuredly, however, this responsibility does not oblige the
agency to provide exhaustive, contemporaneous legal argu-
ments to preemptively defend its action. Similarly, when (and
if) its action is challenged, the DOE is not hamstrung to limit
its legal arguments to the four corners of the administrative
record. In promulgating the Final Rule, the DOE set forth a
coherent interpretation of the statutory definition and gave a
sufficient explanation thereof. And in response to NEMA’s
challenge, it was not only expected but also a duty that the
DOE would further explain and elucidate its interpretation
and how it fits within the Chevron framework. Neither of
these actions undermine the weight given to DOE’s interpre-
tation.

   In contending that the interpretation set forth by the DOE
in its papers warrants no deference, NEMA relies heavily on
Bowen v. Georgetown University Hospital. Such reliance is
misplaced. In Bowen, the Supreme Court refused to afford
Chevron deference to an agency counsel’s interpretation of a
statute "where the agency itself has articulated no position on
the question" — in other words, where the agency’s litigating
position was "wholly unsupported by regulations, rulings, or
administrative 

. The Court further
explained that even if, as a general principle, it would defer
to an agency’s litigating position in the absence of such sup-
porting authority, it would not do so there because the agen-
cy’s litigating position did not reflect a "reasoned and
consistent view of the" statutory provision, but in fact was
"contrary to the . . . view of that provision advanced in past
cases." 

   The present circumstances scarcely could be more distinct
from Bowen. Most obviously, the DOE’s position as advanced
in its papers is hardly "wholly unsupported by" the Final
Rule. See 

. Indeed, quite the contrary
                             NEMA v. DOE                                 35
— as explained above, the DOE’s pronouncements in and
prior to the Final Rule sufficiently set forth and justify its
interpretation, and the DOE’s position here is fully consistent
with that interpretation. As a result, it cannot be said that the
DOE’s papers constitute "nothing more than an agency’s con-
venient litigating position." See 

                                    IV.

  Pursuant to the foregoing, we deny NEMA’s petition for
review.

                               PETITION FOR REVIEW DENIED

SHEDD, Circuit Judge, dissenting:

   When viewed in context, the statute we are reviewing is
clearly unambiguous, and the Department of Energy’s
("DOE") final rule is contrary to the statute and Congress’s
clear direction. Therefore, I would grant the petition, vacate
the final rule, and remand the matter to DOE for action con-
sistent with the law.

   Nineteen years ago, Congress enacted the Energy Policy
Act of 1992 (the "Act"), part of which directed DOE to pre-
scribe energy conservation standards for small electric
motors. See 42 U.S.C. § 6317(b). The Act defines "small elec-
   13
      We need not and do not rule today that the interpretation adopted by
the DOE is the only permissible interpretation of the statutory definition.
Rather, "[i]t suffices to conclude, as we do," that the statutory definition
"does not unambiguously require a different interpretation, and that
[DOE]’s regulation is an entirely reasonable interpretation of the text."

. In the course of the rulemaking, the DOE
received comments regarding the scope of the rulemaking not only from
NEMA but also from environmental groups and other interested parties.
Faced with an ambiguous definition and plausible arguments that the defi-
nition could have an even broader meaning, we glean nothing unreason-
able from the DOE’s decision to chart a course between extremes.
36                           NEMA v. DOE
tric motor" as "a NEMA1 general purpose alternating current
single-speed induction motor, built in a two-digit frame num-
ber series in accordance with NEMA Standards Publication
MG1-1987." 42 U.S.C. § 6311(13)(G)). In the House Report
accompanying the Act, Congress explained the Act "requires
DOE to prescribe energy conservation standards for . . . elec-
tric motors of less than one horsepower." H.R. REP. 102-
474(I), at 208 (1992), reprinted in 1992 U.S.C.C.A.N. 1953,
1998 (emphasis added). The dispositive question before us is
straightforward: as used in the Act, what does "small electric
motor" mean?

   The first step in answering that question is deciding
whether "in accordance with NEMA Standards Publication
MG1-1987" modifies the entire definition or only "two-digit
frame number series." Under the facts in this case, there
would appear to be no doubt that it applies to the whole defi-
nition because both NEMA and DOE have consistently
agreed throughout the rulemaking process that the reference
to NEMA Standards Publication MG1-1987 applies to all fac-
ets of the statutory definition of small electric motors. In other
words, the entire definition in 42 U.S.C. § 6311(13)(G)
should be read in accordance with MG1-1987.

  However, the majority has adopted DOE’s newly
announced litigation posture that MG1-1987 applies only to
"two-digit frame number series" in 42 U.S.C. § 6311(13)(G).2
  1
     The petitioner in this appeal, National Electrical Manufacturers Associ-
ation ("NEMA"), is a not-for-profit trade association that represents a vari-
ety of electrical manufacturers. Among other things, NEMA develops and
publishes standards relating to electrical products, including the 1987 edi-
tion of the Standards Publication MG1 at issue here.
   2
     In this petition for review, DOE takes the position — for the first time
— that the "in accordance with" language modifies only the second
"clause" of the definition. This is completely contrary to their stated posi-
tion throughout the rulemaking process. Compare 74 Fed. Reg. 61421
("[DOE agrees] with NEMA that the reference MG1-1987 applies to all
facets of the statutory definition of a small electric motor. The language
                             NEMA v. DOE                                   37
In reaching its decision, the majority focuses primarily on the
placement of the comma3 and selective canons of statutory
construction to determine that the phrase "in accordance with
NEMA Standards Publication MG1-1987" does not apply to
"a NEMA general purpose alternating current single-speed
induction motor." Accordingly, the majority finds that "small
electric motor" is statutorily limited by a motor’s frame size,
but not by the amount of horsepower a motor has.

   But the obvious reading of the definition is that "in accor-
dance with NEMA Standards Publication MG1-1987" modi-
fies the entire statutory definition of "small electric motor."
The dual references to "NEMA" unmistakably link the entire
definition together as a whole, see U.S. v. Atlantic Research
Corp., 

, 135 (2007) ("Statutes must ‘be read as
a whole.’"), and nothing suggests that the "in accordance
with" language should not be read as applying to the entire
definition, see Porto Rico Ry., Light & Power Co. v. Mor, 

, 348 (1920) ("When several words are followed by

of the statute specifies that the requirements of MG1-1987 apply in deter-
mining what constitutes a small electric motor. . . . The statutory definition
of a small electric motor is bound to the definition of a general-purpose
alternating-current motor as defined in NEMA MG1-1987.") (emphasis
added), with Resp. Br. at 29 ("As explained above, the reference to MG1-
1987 comes later, and expressly modifies only the phrase ‘built in a two-
digit frame number series.’"). This switch in position appears to be an
acknowledgement by DOE that its final rulemaking could not otherwise
withstand the challenge by NEMA. Such a post-hoc litigation position is
entitled to no weight or deference. See, e.g., Wheeler v. Newport News
Shipbuilding & Dry Dock Co., 

, 290 (4th Cir. 2011) (showing
no deference to the agency’s interpretation where it was merely a litigating
position); Village of Barrington v. Surface Transp. Bd., 

, 660
(D.C. Cir. 2011) (giving no deference to an agency’s "‘litigating positions’
raised for the first time on judicial review").
   3
     While the location of the comma is part of the majority’s analysis, its
removal would not in any material manner affect its analysis or its ulti-
mate result. See the majority’s discussion of the last antecedent rule, ante
at 22-23.
38                           NEMA v. DOE
a clause which is applicable as much to the first and other
words as to the last, the natural construction of the language
demands that the clause be read as applicable to all."). This
reading is bolstered by the fact that the phrase "general pur-
pose alternating current" motor, used in the first line of the
statutory definition, is lifted verbatim from Section 1.05 of
MG1-1987 (which section DOE and the majority would at
least partially apply, see ante, at 31-32).4 Thus, in my opin-
ion, § 6311(13)(G) unambiguously incorporates all of MG1-
1987, just as if MG1-1987 had been written into the statute.
See, e.g., United States v. Myers, 

, 331 (4th Cir.
2009) ("The general rule is that when one statute adopts a pro-
vision of another statute by specific reference, it is as if the
adopting statute had itself spelled out the terms of the adopted
provision . . . .").

   Therefore, we turn to MG1-1987 to determine what "small
electric motor" means, and it is clear that MG1-1987 does not
include standards for small motors greater than one horse-
power.5 A general purpose alternating-current motor is
defined in Section 1.05 of MG1-1987 (emphasis added):

        A general-purpose alternating-current motor is an
  4
     In its new litigation position, DOE seeks to minimize the relevance of
MG1-1987. But DOE’s position underscores its inconsistency. Despite
arguing that MG1-1987 (which DOE describes as "decades-old industry
standards") applies only to the "second clause" of § 6311(13)(G), through-
out its briefs DOE repeatedly interprets the "first clause" in the definition
in light of selective provisions and standards in MG1-1987. See, e.g.,
Resp. Br. at 29 ("DOE has interpreted the earlier part of the phrase —
‘NEMA general purpose . . . motor’ — to refer to MG1-1.05 . . . ."); cf.
J.A. 146, 538 (acknowledging that "the 1987 version is the only applicable
version of NEMA MG1.").
   5
     DOE concedes this point, acknowledging in the Final Rule "that
NEMA MG1-1987 does not provide ratings for small motors of the identi-
fied higher horsepower ratings" (i.e. motors of greater than one horse-
power which DOE is seeking to regulate under this rulemaking). 75 Fed.
Reg. at 10883.
                             NEMA v. DOE                                  39
     induction motor . . . which incorporates . . . the fol-
     lowing:

     ....

     (3) Service factor in accordance with MG1-12.47.

     ....

     It is designed in standard ratings with standard
     operating characteristics . . . .6

These factors, ratings, and operating characteristics are
defined more fully in various sections of MG1-1987. First,
Section 1.05 directs us to the service factors found in Section
12.47 (and illustrated in Table 12-2); those factors are delin-
eated in terms of horsepower and speed ratings and clearly
distinguish between small and medium motors; small motors
are one horsepower or less. Second, the "standard ratings" ref-
erenced in Section 1.05 are found in MG1-1987 Part 10.
There, Section 10.32 and Table 10-1, both entitled "Horse-
power and Speed Ratings," illustrate the horsepower and
speed ratings for small and medium induction motors —
small motors do not exceed one horsepower. Finally, Section
1.05’s "standard operating characteristics" requirement is
found in Section 12.32.2. This section provides the standard
   6
     The majority states that I "selectively" quoted Section 1.05, ante at 24
n.12, with the apparent implication that I have purposely avoided relevant
statutory provisions. Such an implication is unfounded. I have expressly
noted only the provisions of Section 1.05 that are relevant to the question
presented in this case, which is the maximum horsepower limitation for
small electric motors. The 200 horsepower ceiling, which includes
medium motors, in no way sets the maximum horsepower for small elec-
tric motors. Neither the 200 horsepower ceiling nor other requirements
found in Section 1.05 but which I have not expressly noted — such as
"open construction," "continuous duty," or "insulation systems" — address
the question presented in this case. Not even DOE argues that small horse-
power motors here exceed 3 horsepower.
40                         NEMA v. DOE
locked-rotor torque operating characteristics for small and
medium motors — small motors include only motors up to
one horsepower.

   Thus, Section 1.05 of MG1-1987, by incorporating Parts 10
and 12, provides the full and complete definition of a general
purpose alternating current motor. And that complete defini-
tion in Section 1.05 provides a clear demarcation between
"small" and "medium" general purpose alternating current
motors. As noted therein, small motors do not exceed one
horsepower.

   Although the majority concludes that the definition of small
electric motor adopts MG1-1987’s standards only as it applies
to frame size, it endorses DOE’s fallback position that MG1-
1987 does not clearly contain a horsepower limitation even if
it were to be applied in its entirety. Although acknowledging
that Section 1.05 of MG1-1987 and Parts 10 and 12 do pro-
vide standards and ratings applicable to small electric motors,
the majority and DOE inexplicably conclude the statutory def-
inition does not incorporate the distinctions between small
and medium motors found in those specified standards and
ratings. Ante, at 24. In fact, under this interpretation, the
majority effectively excises the word "small" from the very
term — small electric motor — the statute is defining. Thus,
under the majority’s view, the MG1-1987 definition of a
small motor includes motors deemed by MG1-1987 to be
medium motors.

   Congress certainly intended no such result. By referring to
MG1-1987 in the statute, Congress incorporated the publica-
tion in its entirety, including the ratings and standards which
defined small motors. Thus, Congress intended to grant DOE
authority only to regulate small electric motors, which in 1992
Congress and NEMA understood to be motors of one horse-
power or less.7 In fact, the congressionally incorporated publi-
  7
   MG1-1987 was incorporated as it existed in 1992 when Congress refer-
enced it in the Act. The fact that DOE has "commendably looked to indus-
                             NEMA v. DOE                                   41
cation MG1-1987 contains no reference — directly or
indirectly — to "small" motors greater than one horsepower.

   Lest there be any doubt about this point, Congress’s clear
understanding is expressed explicitly in the language of the
House Report, which states that the Act "requires DOE to pre-
scribe energy conservation standards for . . . electric motors
of less than one horsepower." H.R. REP. 102-474(I), at 208
(1992), reprinted in 1992 U.S.C.C.A.N. 1953, 1998 (discuss-
ing 42 U.S.C. § 63178) (emphasis added). The House Report
language is important because it explains what the statutory
language at issue meant to the members of Congress who had
expertise over these matters; namely, the membership of the
House Energy and Commerce Committee. Other members of
Congress look to such Report language to understand how the
bill works, what it means, and how it is to be applied. See,
e.g., Garcia v. United States, 

, 76 (1984) ("In sur-
veying legislative history we have repeatedly stated that the
authoritative source for finding the Legislature’s intent lies in
the Committee Reports on the bill, which ‘represen[t] the con-
sidered and collective understanding of those Congressmen
involved in drafting and studying proposed legislation.’"
(quoting Zuber v. Allen, 

, 186 (1969))). Second
only to the words used in the statute itself, a committee report
is perhaps the most definitive expression of congressional
intent. Here, the impact of these words is heightened because
we have more than a general statement by Congress from
which we would need to draw an inference. This is a specific
congressional expression on the very issue before us. Indeed,
congressional intent on this specific issue is clear: DOE is

try practice" and "logically looked to market realities," ante at 29, is of no
moment. That the industry may have developed motor technology not con-
templated in 1992 does not authorize DOE to extend its authority beyond
what the statute authorizes.
   8
     DOE acknowledges it promulgated the final rule at issue pursuant to
the authority delegated to it in 42 U.S.C. § 6317.
42                          NEMA v. DOE
given authority "to prescribe energy conservation standards
for . . . electric motors of less than one horsepower" — noth-
ing more.9 DOE and the majority cavalierly dismiss this legis-
lative language or read it in a fashion that turns the
congressional directive on its head.

   The majority first dismisses the House Report language
because it believes Congress did not speak directly enough.
According to the majority, although the House Report con-
templates regulation of small motors "less than one horse-
power," it is silent as to other horsepower sizes, and to infer
such a horsepower limitation is to draw an impermissible
"negative inference." But reading the House Report language
as limited to motors of one horsepower or less is not based on
a negative inference; it is based on the positive statement of
what Congress intended by this specific provision of the Act
— limiting DOE’s regulatory authority to motors of one
horsepower or less. See, e.g., Am. Petroleum Inst. v. U.S.
Envtl. Prot. Agency, 

, 278 (D.C. Cir. 2000) ("[I]f
Congress makes an explicit provision for apples, oranges and
bananas, it is most unlikely to have meant grapefruit.").

   Then, to the extent the majority would consider this con-
gressional language at all, it adopts the overreaching view
asserted by DOE that even if Congress indicated it wanted to
regulate motors of one horsepower or less, DOE has acted
consistently with this requirement by regulating motors of
even larger horsepower as long as the one horsepower motors
are included in such a regulation.10 This approach turns Con-
  9
    Although Congress stated the limitation as "less than one horsepower,"
that imprecise wording has no real significance here. DOE affirmatively
asserts it makes no difference, and the majority’s analysis is not based on
this particular wording. DOE understands the question before us concerns
motors of one horsepower or less. See J.A. 82 (stating DOE’s acknowledg-
ment in the scope of coverage for the proposed rulemaking that "MG1-
1987 identifies small induction motors as motors with horsepower ratings
from 1 millihorsepower up to 1 horsepower").
   10
      See Resp. Br. at 31-33 (arguing that by regulating electric motors up
to three horsepower, DOE also regulated electric motors of less than one
horsepower and was therefore acting consistently with the express intent
of Congress).
                             NEMA v. DOE                                   43
gress’s pronouncement on its head: it would mean if Congress
authorizes a specific, limited action in a particular regulatory
area, it has authorized all regulatory action in that area. The
suggestion that a partial grant of governmental authority
equals a complete grant of authority is astonishing — only in
"government speak" can "less than one horsepower" actually
mean "up to three horsepower."

   In the final analysis, DOE asks us to: (1) ignore the fact
that Congress statutorily incorporated the entirety of NEMA
Standards Publications MG1-1987; (2) ignore the word
"small" in our effort to understand what Congress meant by
"small electric motor;" (3) ignore the fact that for purposes of
this litigation DOE has completely reversed its longstanding
interpretation of the statute; (4) ignore NEMA’s considered
view of a NEMA publication which Congress intentionally
incorporated into the statute; and (5) ignore specific, relevant
House Report language or in the alternative accept DOE’s
reading which turns that language on its head. We should
reject all of these premises.11

   We should also reject DOE’s new position because it alters
the fundamental relationship between Congress and the
bureaucracy. Instead of giving a real world assessment of
Congress’s statements, the majority’s analysis shifts control
of this entire regulatory area from Congress to the bureau-
cracy, absent congressional authorization. See Brown & Wil-
   11
      Contrary to the majority’s suggestion, I do not believe the majority
has been "hoodwinked." Ante, at 20 n.11. This is a description the majority
applies to itself. I merely point out that the majority has accepted DOE’s
new position on how to read the statutory definition, a position which
DOE indisputably takes for the first time in this litigation. Further, perhaps
to deemphasize DOE’s changed position, the majority asserts that NEMA
has "adopted [its current view] for the express purpose of challenging the
DOE’s rulemaking." 

has the power to reject NEMA’s
view of a NEMA document, but there is no basis in the record to assert
that NEMA, like DOE, has recently changed its position on how the defi-
nition is to be read.
44                        NEMA v. DOE
liamson Tobacco Corp. v. Food & Drug Admin., 

, 161 (4th Cir. 1998) ("[A]gency power is not the power
to make law. Rather, it is the power to adopt regulations to
carry into effect the will of Congress as expressed by the stat-
ute." (quoting Ernst & Ernst v. Hochfelder, 

,
213-14 (1976) (internal quotation marks omitted))). What
DOE proposes is not gap-filling; it is misreading congressio-
nal intent to justify an agency claiming more authority. No
Supreme Court or Fourth Circuit case requires us to take such
a dramatic step with this statute, and I would decline to do so.

   At bottom, this is more than an academic exercise: we are
deciding what Congress meant by "small electric motor" in a
statute that has actual impact in the real world. When we
review a congressional enactment, we should be about the
business of ascertaining and following congressional intent,
which "is of particular importance where, as here, an agency
is attempting to expand the scope of its jurisdiction." Brown
& 

) ("[T]he more intense scrutiny
that is appropriate when the agency interprets its own author-
ity may be grounded in the unspoken premise that govern-
ment agencies have a tendency to swell, not shrink, and are
likely to have an expansive view of their mission." (quoting
Hi-Craft Clothing Co. v. NLRB, 

, 916 (3d Cir.
1981))). We simply should not apply selective canons of con-
struction or caselaw to grant authority to DOE which it can
only justify by creating a litigation position at odds with its
longstanding regulatory position. The statute, confirmed by
the House Report language, is clear and unambiguous that
small electric motors are motors of one horsepower or less.

     Accordingly, I respectfully dissent.