REVISED DECEMBER 10, 2010
           IN THE UNITED STATES COURT OF APPEALS
                    FOR THE FIFTH CIRCUIT   United States Court of Appeals
                                                                                       Fifth Circuit

                                                                                  FILED
                                                                              November 23, 2010
                                         No. 09-51079
                                                                                Lyle W. Cayce
                                                                                     Clerk
SIERRA CLUB, INC.; PUBLIC CITIZEN, INC.,

                                                     Plaintiffs-Appellants,
v.

SANDY CREEK ENERGY ASSOCIATES, L.P.,

                                                     Defendant-Appellee.




                      Appeal from the United States District Court
                           for the Western District of Texas


Before GARZA and BENAVIDES, Circuit Judges, and LYNN*, District Judge.
FORTUNATO P. BENAVIDES, Circuit Judge:
       The present case requires the Court to determine whether Sandy Creek’s
current and ongoing construction of a coal-fired power plant, for which no MACT
determination has ever been made, violates the Clean Air Act § 112(g).1 Because
we conclude that § 112(g)(2)(B) prohibits the act of construction, and not merely
the commencement thereof, we find that Sandy Creek’s current and ongoing


       *
           District Judge of the Northern District of Texas, sitting by designation.
       1
          42 U.S.C. § 7412(g). This opinion will routinely refer to statutory provisions by their
Clean Air Act section numbers (i.e. “section 112”), rather than from the U.S. Code codification
(i.e. “42 U.S.C. § 7412”).
                                        No. 09-51079

construction of a “major source” without a final MACT determination violates
the plain language of the statute. Accordingly, we REVERSE the judgment, and
remand to the district court for further proceedings not inconsistent with this
opinion.
               FACTS AND REGULATORY/PROCEDURAL BACKGROUND
      Defendant Sandy Creek Energy Associates, L.P. (“Sandy Creek”) is
currently constructing a coal-fired power plant in Riesel, Texas. The Texas
Commission on Environmental Quality (“TCEQ”) would ordinarily perform a
routine     case-by-case      “MACT       determination”         prior   to   the   company’s
commencement of construction on a coal-fired power plant, in order to comply
with federal Clean Air Act (“CAA”) requirements. See 42 U.S.C. § 7412(g)(2)(B)
(“No person may construct or reconstruct any major source of hazardous air
pollutants, unless the Administrator (or the State) determines that the
maximum achievable control technology emission limitation under this section
for new sources will be met. Such determination shall be made on a case-by-case
basis where no applicable emission limitations have been established by the
Administrator.”); see also 30 Tex. Admin. Code § 116.711(5) & (11)
(implementing § 112(g)’s “case-by-case MACT determination” by incorporating
it into the State’s pre-construction permitting process). Because Sandy Creek’s
Riesel plant will emit more than ten tons per year of hydrogen chloride–a listed
hazardous pollutant under the Act–the Riesel plant falls under § 112(g)’s MACT
construction proscription.2

      2
          Section 112(a)(1) defines “major source” as follows:

      The term “major source” means any stationary source or group of stationary
      sources located within a contiguous area and under common control that emits
      or has the potential to emit considering controls, in the aggregate, 10 tons per

                                               2
                                       No. 09-51079

       “MACT” refers to “Maximum Achievable Control Technology”–an emission
limitation standard for the listed “hazardous” pollutants Congress ordered the
EPA to regulate in § 112. Section 112 requires major sources, like the Riesel
plant, to “comply with technology-based emission standards requiring the
maximum degree of reduction in emissions EPA deems achievable, often referred
to as ‘maximum achievable control technology’ or MACT standards.” Nat’l Min.
Ass’n v. EPA, 

, 1353 (D.C. Cir. 1995) (quoting 42 U.S.C.
§ 7412(d)(1)-(2)). Congress noted that § 112’s MACT emissions standards would
“be more stringent” than the standards imposed by other provisions of the
Act–such as New Source Review or Prevention of Significant Deterioration
(“PSD”)–and explained this as necessary since the MACT “program is for the
control of extremely harmful air pollutants.” S. Rep. No. 101-228, at 140 (1989).
       Thus, in ordinary circumstances, we would not question whether
§ 112(g)(2)(B)’s requirement of a case-by-case MACT determination applied to
Sandy Creek’s Riesel Plant.3 The present case, however, presents us with an
unusual circumstance. In March of 2005, the EPA issued a rule removing coal
and oil-fired electric utility steam generating units (“EGUs”) from the list of
sources whose emissions are regulated under § 112.4 Thus, although Sandy


       year or more of any hazardous air pollutant or 25 tons per year or more of any
       combination of hazardous air pollutants.

The question of whether Sandy Creek’s Riesel plant qualifies as a “major source” under § 112
is not before this Court since both parties agree the plant will emit enough tons of mercury to
qualify the plant as a “major source” under the Act.
       3
        In its brief, Sandy Creek conceded that before March of 2005, § 112(g)’s MACT
requirement applied to its Riesel plant.
       4
        This agency action became known as the “Delisting Rule,” and will be discussed in
greater detail below.

                                              3
                                        No. 09-51079

Creek submitted an application for a MACT determination to TCEQ, TCEQ
concluded that, as a result of EPA’s Delisting Rule, no such determination was
legally required. On May 25, 2006, TCEQ stated that “[n]o case-by-case MACT
determination for the PC boiler is needed because the type of steam generating
unit (PC boiler) that Sandy Creek is proposing is not subject to MACT
regulation.” Sandy Creek then commenced construction on its coal-fired power
plant in Riesel, Texas, on January 7, 2008.
       One month later, on February, 8, 2008, the D.C. Circuit vacated EPA’s
March 2005 Delisting Rule, declaring that EPA’s decision to remove EGUs from
the list of § 112’s regulated sources violated “the plain text and structure of
section 112.”5      New Jersey v. EPA, 

, 583 (D.C. Cir. 2008).
Consequently, the D.C. Circuit held that EGUs “remain listed under section
112.” On March 14, 2008, the mandate issued and the Delisting Rule was
officially vacated. As a result of the D.C. Circuit’s decision in New Jersey,
§ 112(g)’s construction prohibition on “major sources” with no MACT
determination once again became applicable to all coal-fired power plants.6


       5
         The D.C. Circuit recognized that the plain language in § 112(c) meant that “the only
way EPA could remove EGUs from the section 112(c)(1) list was by satisfying section
112(c)(9)’s requirements.” New 

. The EPA, however, conceded “that
it never made the findings section 112(c)(9) would require in order to delist EGUs.” 

New Jersey Court held that “EPA’s purported removal of EGUs from the
section 112(c)(1) list therefore violated the CAA’s plain text and must be rejected under step
one of Chevron.” 

of the Delisting Rule means that EGUs are still legally
obligated to comply with § 112(g)(2)(B)’s MACT requirement for any hazardous pollutant they
emit. 

the CAA’s Citizen Suit Provision, see 42 U.S.C. § 7604(a), (f), gives this Court
jurisdiction over suits asserting that specific “major emitting facilities” violate the
requirements of the CAA, Congress specifically reserved jurisdiction over challenges to EPA’s
administrative rules and regulations implementing the statute as a whole to the D.C. Circuit
Court of Appeals. See 

Congress’s decision to designate the D.C. Circuit as the

                                                4
                                        No. 09-51079

       Soon thereafter, on August 8, 2008, Plaintiffs Public Citizen, Inc. and
Sierra Club, Inc. (collectively, “Sierra Club”) filed their complaint in federal
district court. In the complaint, the Plaintiffs alleged that Sandy Creek’s
construction of its coal-fired plant in Riesel was in violation of CAA
§ 112(g)(2)(B) because Sandy Creek had never obtained a MACT determination
for the plant.
       Sierra Club filed a motion for summary judgment on March 4, 2009, and
then on April 2, 2009, Sandy Creek filed its cross-motion for summary judgment,
arguing that the district court should abstain from deciding Sierra Club’s
§ 112(g)(2)(B) challenge pursuant to Burford v. Sun Oil Co., 

(1943).
On September 28, 2009, the district court issued its Order. Although the district
court declined to abstain pursuant to Burford, the district court determined that
§ 112(g)’s requirement for a MACT determination no longer applied to Sandy
Creek, and consequently, the district court granted summary judgment in Sandy
Creek’s favor and denied Sierra Club’s motion for summary judgment. Sierra
Club timely appealed, asserting that the district court erred when it found that
§ 112(g)(2)(B)’s requirement for a final MACT determination no longer applies
to Sandy Creek. On appeal, Sandy Creek asserts that the district court abused
its discretion when it declined to abstain under Burford.
       We will consider both claims in turn.



Court with exclusive jurisdiction over challenges to rules such as the Delisting Rule and other
“standard-setting actions by the Administrator was to ensure uniformity in decisions
concerning issues of more than purely local or regional impact.” Natural Res. Def. Council, Inc.
v. EPA, 

, 1357 (D.C. Cir. 1975). Consequently, it is clear, and no party contests,
that the D.C. Circuit’s vacatur of the Delisting Rule applies unequivocally throughout all of the
Federal Circuits.


                                               5
                                  No. 09-51079

                             STANDARD OF REVIEW
      “We review a grant of summary judgment de novo, applying the same legal
standard as the district court.” Croft v. Governor of Tex., 

, 742 (5th
Cir. 2009) (internal quotations omitted). Summary judgment should be rendered
if the record demonstrates that “there is no genuine issue as to any material fact
and that the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P.
56(c). “An issue is material if its resolution could affect the outcome of the
action.” Daniels v. City of Arlington, Tex., 

, 502 (5th Cir. 2001). “In
deciding whether a fact issue has been created, the court must view the facts and
the inferences to be drawn therefrom in the light most favorable to the
nonmoving party.” 

parties moved for summary judgment
before the district court, “[t]he mere fact that both appellants and appellee
moved for summary judgment does not warrant the grant of either motion if the
record reflects a genuine issue of fact.” Hindes v. United States, 

,
152 (5th Cir. 1964).
                                    ANALYSIS
      The question before this Court is how to interpret and apply § 112(g) to a
coal-fired power plant still under construction, when the construction on that
plant commenced before the D.C. Circuit’s decision in New Jersey–that is, where
the construction of the plant began during a time period in which a MACT
determination was not required under the EPA’s unlawful Delisting Rule–and
TCEQ declined to make one. Sandy Creek makes two arguments in support of
its position that it should not be required to obtain a MACT determination now.
First, Sandy Creek argues that when TCEQ issued Sandy Creek its pre-
construction permit on May 25, 2006, it actually made a MACT determination.


                                        6
                                 No. 09-51079

In the alternative, Sandy Creek also argues that because § 112(g) did not apply
to it when it commenced construction, § 112(g)’s requirement for a MACT
determination does not apply to its ongoing construction. Neither the record, nor
the plain language of the statute, support these two positions.
      I.    TCEQ HAS NOT MADE A MACT DETERMINATION
      Sandy Creek asserts that its ongoing construction of a coal-fired plant is
not in violation of § 112(g)(2)(B) because TCEQ did make a MACT
determination. In this regard, Sandy Creek conflates a decision not to make a
MACT determination with an actual final MACT determination that the Act
requires. Because TCEQ declared in its May 2006 Final Order that no MACT
determination was required, and Sandy Creek’s application for a MACT
determination and TCEQ’s preliminary determination were deficient as they did
not contain substantive evaluations of MACT limits or MACT floors for any
hazardous air pollutant, this Court finds that TCEQ did not make a proper
MACT determination in its May 2006 Final Order.
      The district court’s factual findings are ambiguous as to whether or not
TCEQ made a final MACT determination for Sandy Creek’s Riesel plant. The
evidence in the record, however, shows that TCEQ did not make a MACT
determination for Sandy Creek’s Riesel plant. First, the May 2006 Final Order’s
provision that the MACT determination requirement is inapplicable strongly
indicates that TCEQ did not make a MACT determination. The TCEQ provided,
“[n]o case-by-case MACT determination for the PC boiler is needed because the
type of steam generating unit (PC boiler) that Sandy Creek is proposing is not




                                       7
                                       No. 09-51079

subject to MACT regulation.”7 This May 2006 Final Order thus did not contain
a case-by-case evaluation of MACT limits or set a MACT floor for Sandy Creek.
       Nothing in the record demonstrates that TCEQ made a MACT
determination. Even though we accept as true Sandy Creek’s contention that
it prepared and submitted an application to TCEQ under the assumption that
a MACT determination would be required, that does not constitute evidence that
TCEQ actually made the determination.                 Moreover, Sandy Creek’s initial
§ 112(g) submissions were deficient, in that the application’s proposed MACT
limits did not include a MACT floor. A BACT limit cannot substitute for a
MACT limit because BACT limits do not include a floor as the MACT limits do.
Compare 42 U.S.C. § 7412(d)(3) (establishing MACT floor) with 42 U.S.C.
§ 7479(3) (defining BACT limits as taking economic costs into account).8


       7
        Most notably, Sandy Creek’s own engineer told TCEQ at the administrative hearing
that with regard to Sandy Creek’s Riesel plant, the “MACT regulation . . . is no longer
relevant.” R. at 140-41.
       8
         We note that the record does support Sandy Creek’s assertion that TCEQ conducted
a BACT (“Best Available Control Technology”) review for the Riesel plant’s emission of
hazardous air pollutants, including mercury. See R. at 693 (“[TCEQ] conducted a best
available control technology (BACT) and impacts review.”). We do not, however, agree with
Sandy Creek’s assertion that we should “consider this BACT review as equivalent to MACT
review.” 

§ 112’s language is clear in its command that MACT–and not BACT–be
applied to a coal-fired plant’s emissions of hazardous pollutants like mercury. Yet even if the
language of the statute could be construed as ambiguous, that ambiguity would fall in face of
the legislative history reflecting Congress’s intent that the more stringent MACT be applied
to hazardous pollutants–and not the less stringent BACT. See S. Rep. No. 101-228, at 140
(1989) (explaining that previously proposed legislation had “proposed to add a BACT concept
to section 112,” but Congress determined that “the case-by-case BACT determinations . . . were
not sufficiently stringent and in many cases provided no more control than would have been
required through application of new sources performance standards.”). Consequently,
Congress created an “emissions limitation based on section 112(d) [that] will, in most cases,
be more stringent than a new source performance standard for the same category of sources
or pollutants.” 

MACT, Congress further noted that the “emissions limitation
imposed here, and the standard for control which is its foundation, shall generally be referred

                                              8
                                     No. 09-51079

Moreover, a MACT determination requires assessment of every hazardous air
pollutant produced by a plant, including for emissions of arsenic, hydrochloric
acid, selenium, and cadmium, whereas the alleged MACT determination in
Sandy Creek’s permit is limited to mercury. See 42 U.S.C. § 7412(d)(1).
      In its May 2006 Final Order TCEQ did not acknowledge the deficiencies
in its preliminary decisions, and instead relied on the Delisting Rule to conclude
that no case-by-case MACT determination was necessary for the pulverized coal
boiler. R. at 680. A letter from an employee deeming the MACT determination
requirement satisfied is clearly insufficient to make up for these inadequacies
in the alleged MACT determination. Furthermore, Sandy Creek’s argument that
TCEQ’s decision not to make a MACT determination constitutes a MACT
determination is without merit. As such, we conclude that TCEQ never made
a MACT determination for Sandy Creek.
      II. DEFENDANT’S ONGOING CONSTRUCTION VIOLATES § 112(G)(2)(B)
      Furthermore, we find that Sandy Creek’s current, ongoing construction of
a coal-fired plant–for which Sandy Creek has received no final MACT
determination–violates the plain language of § 112(g)(2)(B). We reject Sandy
Creek’s argument that because it commenced construction when EPA’s unlawful
Delisting Rule was still in effect in January of 2008, the Delisting Rule should
continue to preclude the application of § 112(g)(2)(B) to its Riesel plant now.
      A plain reading of the statute does not support Sandy Creek’s position.
Instead, a plain reading of § 112(g)’s statutory language leads to the conclusion



to as the ‘maximum achievable control technology’ or MACT to distinguish it from other
requirements in the Act.”). As a result, we find that for purposes of § 112, MACT and BACT
are not equivalent, and a State Agency’s BACT review does not satisfy § 112’s mandate that
“major facilities” receive a MACT determination.

                                            9
                                        No. 09-51079

that the date of commencement of construction on a “major source” does not
alter, shift, or eradicate the application of § 112(g)’s prohibition on the
construction of that “major source” until, or unless, a MACT determination has
been made. Notably, § 112(g)(2)(B) states that “no person may construct or
reconstruct any major source of hazardous air pollutants, unless the
Administrator (or the State) determines that the maximum achievable control
technology emission limitation under this section for new sources will be met.”
See also 40 C.F.R. § 63.43(l)(2) (interpreting § 112(g)(2)(B) to mean that
violations are not limited to commencement of construction, but rather, occur
“for whatever period the owner or operator is determined to be in violation of
[MACT] requirements”).
       That is, § 112(g)(2)(B) simply states “no person may construct.” It does not
state “no person may begin construction” or “start to construct.”                      Section
112(g)(2)(B) does not state that prior to beginning construction, the
Administrator (or the State) must determine that the MACT emission limitation
has been met.9 Instead, § 112(g)(2)(B) simply renders the act of constructing


       9
         In this regard, Sandy Creek also argues that § 112(g)(2)(B) does not require that a
MACT determination be made now, in 2010, since “both TCEQ’s and EPA’s rules make the
§ 112(g) case-by-case MACT determination part of the preconstruction permitting process,
[and] not an ongoing, freestanding requirement.” Essentially, Sandy Creek argues that
because TCEQ and EPA have elected to administratively implement the requirements of
§ 112(g) in regulations that procedurally place MACT determinations in the preconstruction
permitting process, any violation of § 112(g) that occurs outside of this regulatory process is
no longer a violation of the statute itself. This is an incorrect reading of the interplay of
administrative law and statutory interpretation. Even if EPA and TCEQ intended to, they
cannot issue regulations that alter § 112(g)(2)(B)’s explicit requirement that no person
construct a major source without a MACT determination. It is a fundamental precept of
administrative law that an agency action, rule, or regulation “cannot overcome the plain text
enacted by Congress.” New 

; see also Chevron v. Natural Res. Def.
Council, Inc., 

, 842-43 (1984) (“[T]he court, as well as the agency, must give effect
to the unambiguously expressed intent of Congress.”). Section 112(g) says nothing about a

                                              10
                                        No. 09-51079

itself unlawful–unless and until a MACT determination has been made. The day
Sandy Creek actually commenced construction is, therefore, irrelevant to
§ 112(g)(2)(B)’s current application to Sandy Creek’s concurrent and ongoing
construction, since § 112(g)(2)(B) prohibits the act of construction itself– and not
the commencement thereof.10
       Thus, the question really is not whether Sandy Creek must comply with
§ 112, but rather, the question is when and how. Because Sandy Creek is
currently constructing a “major source,” we find that § 112(g)’s MACT
requirement for new sources constitutes the most appropriate application of
§ 112 to the Riesel plant. In so finding, we find it important to note that any
construction Sandy Creek undertook prior to March 14, 2008,11 should not be
considered in violation of § 112(g)(2)(B). We agree with the EPA’s position
“[that] consideration [must be given] to the effect of prior construction,
undertaken in reasonable reliance on now-vacated rules.”12                     However, the


preconstruction permitting process. It simply prohibits the act of constructing a major source
for which no MACT determination has been made.
       10
          Judge Ebel, of the Tenth Circuit Court of Appeals, recently came to the same
conclusion when, sitting on the District Court by designation, he reasoned that while ordinarily
the statute and its corresponding regulations “requir[e] an operator to obtain a MACT
determination before ‘construct[ing] or reconstruct[ing]’ a major source, nothing in the text
indicates that the operator is relieved of the responsibility of complying with the statute after
construction begins. ‘[C]onstruct or reconstruct’ are active verbs that have force after the
permit is issued and after construction or reconstruction has begun.” Wildearth Guardians v.
Lamar Utilities Bd., No. 1:09-CV-02974-DME-BNB, 

, at *5 (D. Colo. Aug.
13, 2010).
       11
            March 14, 2008, is the date the D.C. Circuit issued the mandate in New Jersey v.
EPA.
       12
         Notably, on January 7, 2009, in a letter from Robert J. Meyers, Principal Deputy
Assistant Administrator, the Agency stated that:


                                              11
                                      No. 09-51079

Delisting Rule has since been vacated, and consequently, Sandy Creek must now
come into compliance with § 112(g)(2)(B). Accordingly, we conclude that Sandy
Creek’s ongoing construction of a coal-fired power plant–for which no MACT
determination has been made–is in violation of § 112(g)(2)(B).
      III. HARPER DOES NOT PRECLUDE THE APPLICATION OF § 112(g)
      In coming to the conclusion that § 112(g) does not presently apply to Sandy
Creek’s ongoing construction of a “major source,” the district court erroneously
relied on the Supreme Court’s decision in Harper v. Virginia Dept. of Taxation,

(1993). In Harper, the Supreme Court considered whether a judicial
decision might have retroactive effect, reasoning “that the nature of judicial
review strips [the Court] of the quintessentially legislative prerogative to make
rules of law retroactive or prospective as we see fit . . . .” 

The Supreme
Court held that judicial decisions will only have retroactive effect on those “cases



      [a]lthough these EGUs may have relied in good faith on rules that EPA issued
      that were subsequently vacated, the Agency believes that these EGUs are
      legally obligated to come into compliance with the requirements of Section
      112(g). . . . We therefore request that the appropriate State or local permitting
      authority commence a process under Section 112(g) to make a new-source
      MACT determination in each of these cases. . . . Section 112(g) proceedings
      ordinarily are concluded before the commencement of any construction activity,
      so it is reasonable for the permitting authority– under these unique and
      compelling circumstances, and within the bounds of its discretion under Clean
      Air Act Section 112(g) and EPA’s section 112(g) regulations– to give
      consideration to the effect of prior construction, undertaken in reasonable
      reliance on now-vacated rules, in making the case-by-case determination of
      applicable MACT requirements. . . .

Thus, EPA’s interpretation protects the conduct of those who undertook construction in
reliance on what is now a vacated rule, while simultaneously effectuating the congressional
purpose behind § 112(g)–that is, ensuring that these new major sources of congressionally
listed hazardous pollutants will be constructed with “the maximum achievable control
technology emission limitation.” 42 U.S.C. § 7412.

                                             12
                                        No. 09-51079

still open on direct review.” 

The Harper Court, however, did not
consider–nor did it create–any guidelines for lower courts to apply when
considering the retroactive effect of a restored administrative rule or regulation.
Moreover, the district court seemed to assume that to find a violation of
§ 112(g)(2)(B), it would have to retroactively apply New Jersey, but that is not
the case. Because § 112(g)(2)(B) applies throughout the construction process, the
issue of whether New Jersey can be applied retroactively to make illegal the
construction that began before the D.C. Circuit decided New Jersey is not
determinative. We interpret the plain language of § 112(g)(2)(B) to bar Sandy
Creek from constructing a coal-fired plant without a final MACT determination.
The Supreme Court’s decision in Harper does nothing to absolve Sandy Creek
of its present duty to abide by the requirements in § 112(g)(2)(B).
       Furthermore, in its consideration of Harper, the district court also
concluded that because a Texas state district court affirmed TCEQ’s May 2006
Final Order for the Riesel plant’s PSD permit, the “direct review of TCEQ’s

       13
           We note that our discussion of Harper is limited to our recognition that the Supreme
Court’s holding in Harper is inapposite to the present case before us. In this regard, we
recognize that the Supreme Court’s jurisprudence on the retroactive application of judicial
decisions is more complex than our analysis herein necessitates. See, e.g., Danforth v.
Minnesota, 

, 271 (2008) (discussing the “confused and confusing ‘retroactivity’
cases decided in the years between 1965 and 1987.”); Whorton v. Bockting, 

, 416
(2007) (explaining that in the context of criminal appeals, “[a] new rule applies retroactively
in a collateral proceeding only if (1) the rule is substantive or (2) the rule is a watershed rule
of criminal procedure implicating the fundamental fairness and accuracy of the criminal
proceeding.”) (internal quotation marks and brackets omitted); Rivers v. Roadway Exp., Inc.,

, 312 (1994) (“‘Judicial decisions have had retrospective operation for near a
thousand years.’”) (quoting Kuhn v. Fairmont Coal Co., 

, 372 (1910) (Holmes, J.,
dissenting)). Consequently, it is not our intention to interpret Harper as reciting an exhaustive
list of judicial decisions that enjoy retroactive application and an exhaustive list of those that
do not–and any such reading of our discussion of Harper would be unnecessarily broad and
mistaken.


                                               13
                                        No. 09-51079

MACT-related findings and conclusions in the final permit was closed on the
date that the Texas district court rendered its decision on March 29, 2007.” This
particular application of Harper’s “direct review” rule is erroneous.
       That is, for purposes of its “direct review” analysis, the district court has
conflated TCEQ’s granting of a PSD permit, see 42 U.S.C. §§ 7470-7492, with the
issuance of a final MACT determination. See 42 U.S.C. § 7412(g). On March 29,
2007, the Texas state district court approved TCEQ’s granting of a PSD permit.
Of course, the state district court’s decision to affirm TCEQ’s granting of Sandy
Creek’s PSD permit closed the door to any future challenges to the legal
legitimacy of that particular PSD permit.                The Texas state district court,
however, did not affirm TCEQ’s MACT determination–mainly because TCEQ
never made one. Accordingly, since Sierra Club is challenging Sandy Creek’s
failure to obtain a MACT determination pursuant to § 112(g)(2)(B)–and because
Sierra Club is not challenging the legal legitimacy of Sandy Creek’s PSD
permit–what the Texas state district court affirmed on March 29, 2007,
regarding Sandy Creek’s PSD permit, does not influence our consideration of
Sandy Creek’s statutory obligation to obtain a MACT determination. This is
because the affirmation of this PSD permit said nothing about the legality (or
existence of) a MACT determination pursuant to § 112(g).14

       14
          In contrast to the permits that implement National Ambient Air Quality Standards,
Congress created § 112’s technology-based requirement of a case-by-case MACT determination
“to accelerate the regulation of hazardous air pollutants.” S. Rep. No. 101-228 at 133 (1989)
(emphasis added). That is, Congress determined that creating a more stringent program for
certain listed hazardous pollutants in § 112 was “appropriate” since, in contrast to PSD
permits, section 112(g) “is for the control of extremely harmful air pollutants.” 

Consequently, TCEQ’s determination that Sandy Creek was in compliance with the applicable
PSD emission limitations in its permit says nothing about Sandy Creek’s compliance with
§ 112(g)’s MACT requirement. See § 112(b)(6) (“Prevention of significant deterioration: The
provisions of part C of this subchapter (prevention of significant deterioration) shall not apply

                                              14
                                       No. 09-51079

       Accordingly, a plain reading of the statute demonstrates that securing a
PSD permit does not obviate the need to comply with § 112, and here, where
Sandy Creek has not yet obtained a final MACT determination, Sandy Creek’s
continued construction of a major source constitutes a violation of
§ 112(g)(2)(B).
        IV. THE DISTRICT COURT WAS CORRECT NOT TO ABSTAIN
       Finally, Sandy Creek asserts that the district court should have abstained,
pursuant to Burford, 

. This Court “generally review[s] abstention
decisions under an abuse of discretion standard.” Sierra Club v. City of San
Antonio, 

, 793 (5th Cir. 1997). Given this standard of review, and
the applicable law concerning Burford abstention, we would be hard pressed to
conclude that the district court abused its discretion in refusing to abstain in
this case.
       “We start with the command that the federal courts have a ‘virtually
unflagging obligation to exercise the jurisdiction given them.’” Wilson v. Valley
Elec. Membership Corp., 

, 313 (5th Cir. 1993) (quoting Colorado River
Water Conservation Dist. v. United States, 

, 817 (1976)). As a result,
“[a]bstention from the exercise of federal jurisdiction is the exception, not the
rule.” Colorado 

. This Court has recognized five factors for
a court to weigh when considering whether to abstain under Burford. In Wilson,
the Court listed these five Burford factors as follows:
       (1) whether the cause of action arises under federal or state law
       (finding abstention inappropriate where the case did not involve a
       state-law claim); (2) whether the case requires inquiry into
       unsettled issues of state law, or into local facts; (3) the importance


to pollutants listed under this section [§ 112].”) (emphasis added).

                                              15
                                         No. 09-51079

       of the state interest involved; (4) the state’s need for a coherent
       policy in that area; and (5) the presence of a special state forum for
       judicial review.

(internal quotation marks and citations omitted).
       In the present case, consideration of these five factors support the district
court’s decision not to abstain. The first factor overwhelmingly affirms the
district court’s decision, since no state cause of action is involved in a federal
CAA citizen suit. See New Orleans Public Serv., Inc. v. Council of City of New
Orleans, 

, 361 (1989) (holding that Burford abstention is not
appropriate where the plaintiff’s claim “does not involve a state-law claim,” and
rejecting the Fifth Circuit’s declaration that “‘the absence of a state law claim
[is] not fatal’” to the application of Burford abstention) (quoting and overturning
New Orleans Public Serv., Inc. v. City of New Orleans, 

, 861-62 (5th
Cir. 1986)). Whereas Burford consisted of a federal constitutional challenge to
a state-created agency action, see 

, the present challenge
raises no federal constitutional concerns about any state-created regulatory
body–but instead utilizes a federal congressionally-created cause of action to
challenge a particular entity’s failure to comply with a federally created
regulatory scheme.15
       Furthermore, the remaining factors weigh in favor of the district court’s
decision not to abstain. Although it is true that Congress created in the CAA “a
comprehensive national program that made the States and the Federal


       15
         Notably, Congress designed this regulatory scheme to specifically grant federal courts
subject matter jurisdiction over suits like the one presently before us. The CAA’s Citizen Suit
provision, see 42 U.S.C. § 7604(a), (f), gives this Court jurisdiction over suits such as this one,
where citizens challenge the construction of a “modified major emitting facility” 

for
which “any requirement under section 7411 or 7412 of this title” has not been met. 

                                               16
                                     No. 09-51079

Government partners in the struggle against air pollution,” GM Corp. v. United
States, 

, 532 (1990), it is also true that Congress gave EPA the
power to revoke a State’s regulatory authority if that State does not abide by the
CAA’s federal statutory requirements. See 42 U.S.C. § 7410(c)(1)(B) (granting
EPA the discretion to revoke a State’s air permitting authority if the EPA
“disapproves a State implementation plan submission in whole or in part”).
Recently, the EPA formally announced it has disapproved Texas’s State
Implementation Plan, and consequently, TCEQ’s continued authority to issue
air permits under the CAA will soon (perhaps temporarily) expire.16
Consequently, the present case would make for an odd application of Burford,
given that the federal district court would be abstaining from ruling when the
state-regulatory body may soon lose its federally-granted authority to regulate.
      In contrast to Burford and this Court’s decision in Sierra Club v. City of
San Antonio, the regulatory framework at issue here was created by the United
States Congress–not a state legislative body.            Further, the legal question
involves the application of federal law–not state law. And finally, any need for
a “coherent policy” is best attributed to the federal government, not the state of


      16
         On June 30, 2010, the EPA announced its formal disapproval of Texas’s state
permitting program run by TCEQ:

      Today, EPA announced final disapproval of the flexible permit program that the
      Texas Commission on Environmental Quality (TCEQ) had submitted for
      inclusion in its clean-air implementation plan. EPA has determined that this
      program does not meet several national Clean Air Act requirements that help
      to assure the protection of health and the environment.

EPA Disapproves Texas Flexible Air Permit Program, News Releases From Region 6, United
States Department of Environmental Protection Agency (Jun. 30, 2010),
http://yosemite.epa.gov/opa/admpress.nsf/e8f4ff7f7970934e8525735900400c2e/1d91bf2747c
5682b8525775200626aa6!OpenDocument (last visited Sept. 09, 2010).

                                           17
                                  No. 09-51079

Texas. One reason Congress decided to amend § 112 in 1990 and create a
federally enforceable MACT requirement was to remedy the states’ widely
divergent approaches to regulating hazardous pollutants nationwide. See H.R.
Rep. No. 101-490, pt. 1, at 330 (1990) (noting that “the approaches taken by
States to control air toxics vary considerably” and this has “produced a
patchwork of differing standards”).
      Thus, for the reasons described above, we conclude that the district court
did not abuse its discretion when it refused to abstain under Burford.
                                  CONCLUSION
      For the foregoing reasons, we find that the district court did not err in its
decision not to abstain under Burford, and because we find that Sandy Creek’s
ongoing construction of a coal-fired power plant with no final MACT
determination violates CAA § 112(g)(2)(B), we REVERSE the judgment of the
district court and remand for further proceedings not inconsistent with this
opinion.




                                        18