Bethesda Memorial Hospital, In Re:

HATCHETT, Chief Judge, dissenting:
Without the benefit of briefing or clear guidance from the
Supreme Court, a majority of this panel sua sponte abandons a
firmly rooted line of circuit authority holding that a district
court’s statutory-based remand order must be challenged by
Petition for Writ of Mandamus. In so doing, the majority (1)
contravenes the well-established rule that only the en banc court
or the Supreme Court may reverse prior panel decisions; (2)
alters the standard of review applicable in this case; and (3)
compounds the error of a decision that is otherwise wrong on the
merits. For these reasons, I respectfully dissent.
If the majority followed Congress’s clear command in 28
U.S.C. § 1447, we might save for another day our dispute
regarding circuit precedent, the proper method of appeal and the
standard of review. That is because section 1447(d) provides
that an “order remanding a case to the State court from which it
was removed is not reviewable on appeal or otherwise . . . .” 28
U.S.C. § 1447(d)(1994) (emphasis added). In other words, where
section 1447 is appropriately invoked, appellate review is
unavailable through mandamus or direct appeal.
Unfortunately, the majority refuses to acknowledge that
section 1447 operates to bar any form of appellate review in this
case. The majority refuses to do so because it believes that the
district court’s remand order was not based on (1) a lack of
subject matter jurisdiction, or (2) a timely motion to remand for
a defect in the removal procedure. I respectfully suggest that
this belief is untenable on the record before us and rooted in a
flawed reading of section 1447.
With respect to subject matter jurisdiction, the majority
says that “the remand order in this case is clearly based on a
defect in the removal process.” Majority Op. at 5. To buttress
its claim, the majority cites portions of the record strung
together with ellipses. See Majority Op. at 5-6 n.2.
The first two substantive sentences of the district court’s
order read as follows: “Section 1446(a), Title 28 of the United
States Code requires a party seeking removal to file a notice of
removal listing all grounds which support the federal court’s
jurisdiction. The removal procedures are strictly construed
because of this Court’s limited removal jurisdiction.” (Emphasis
added.) These two sentences alone raise the strong inference
that the district court’s concern in its remand order is
ultimately jurisdictional and not procedural in nature. Cf.
Ariail Drug Co., Inc. v. Recomm Int’l Display, Ltd., ___ F.3d
___, ___, No. 96-6570, slip op. at 3562-63 (11th Cir. Sept. 3,
1997) (indicating that removal jurisdiction is a species of
subject matter jurisdiction); see also BJT, Inc. v. Molson
Breweries USA, Inc., 848 F. Supp. 54, 56 (E.D.N.C. 1994) (“a
federal court’s removal jurisdiction is a form of subject matter
jurisdiction”); Emrich v. Touche Ross & Co., 846 F.2d 1190, 1194-
95 (9th Cir. 1988) (removal of a case from state court to federal
2
court pursuant to section 1441 is a question of federal subject
matter jurisdiction).1
1
The Ariail Drug panel characterizes removal
jurisdiction as a hybrid form of subject matter jurisdiction,
because a lack of removal jurisdiction is considered in some
contexts a procedural defect. __ F.3d at ___ n.5, slip op. at
3563 n.5. Compare In re Ocean Marine Mut. Protection & Indem.
Ass’n, Ltd., 3 F.3d 353, 356 (11th Cir. 1993) (“Failure to comply
with § 1446(a) and (b) constitutes a defect in removal procedure
within the meaning of § 1447(c).”), and In re Fed. Sav. & Loan
Ins. Co., 837 F.2d 432, 434-35 (11th Cir. 1988) (stating that the
“authority” of the Federal Deposit Insurance Corporation to bring
a case to federal court was, like a private litigant’s authority,
once subject to limitations of 28 U.S.C. §§ 1331 and 1441, and
stating that the express consent of all defendants is a
“condition precedent to removal” pursuant to 28 U.S.C. §
1441(a)”).
I am willing to accept, given our case law, that violations
of 28 U.S.C. § 1446 fall in the category of procedural defects
for purposes of determining reviewability under 28 U.S.C. §
1447(c). Indeed, from a purely textual point it makes perfect
sense to think that section 1447(c)’s use of the phrase “defect
in removal procedure” refers to the requirements of section 1446,
which is entitled “procedure for removal.”
It does not follow, however, that failures to comply with
the requirements of 28 U.S.C. § 1441, which is entitled “actions
removable generally,” are also defects in removal procedure for
purposes of section 1447(c). Section 1441 is more properly
viewed as a specialized subject matter jurisdiction statute for
defendants, akin to 28 U.S.C. § 1331, which allows a plaintiff to
bring a suit in federal court if the claims therein arise under
federal law. This is so because section 1441 confers a limited
statutory right on defendants to remove certain cases to federal
court, provided specific preconditions are met -- conditions such
as consent among all defendants under section 1441(a), and
diversity of citizenship under section 1441(b). Given the
jurisdictional nature of section 1441, I believe Congress
probably intended for removal that is defective under section
1441 to be cognizable under section 1447(c)’s provision relating
to subject matter jurisdiction. See Charles Everingham IV,
Removal, Waiver, and the Myth of Unreviewable Remand in the Fifth
Circuit, 45 Baylor L. Rev. 723, 753 (1993) (“Based on the text of
the removal statutes, the legislative history, and the
Congressional policies behind removal jurisdiction, Congress
probably meant for a defect in removal jurisdiction to deprive
the court of subject matter jurisdiction.”); cf. In re Fed. Sav.
& Loan Ins. Co., 837 F.2d at 435; Emrich, 846 F.2d at 1194 n.2.
3
The district court’s order follows the above-quoted
sentences with a citation reference to two prior cases that this
particular district court judge authored. The first case is
Senter v. Sears, Roebuck & Co., 712 F. Supp. 179 (S.D. Fla.
1989). In Senter, Sears & Roebuck filed a motion for
reconsideration after the district court remanded a case that
Sears & Roebuck had removed to federal court. Sears & Roebuck’s
initial petition for removal was technically defective in some
respects and the district court indicated that these defects
constituted “one of the reasons why this court found in its
previous order that it lacked jurisdiction over this matter.”
712 F. Supp. at 179. The Senter court went on to state:
When a defendant moves a federal district court to
grant a removal petition, the court must make an
initial inquiry whether it has subject matter
jurisdiction over the cause. In making this
determination, the court must consider the procedural
posture of the case at that time including the status
of all parties and the action of the state court in
disposing of the matter. For example, in this case,
the defendant’s act of including a third-party
complaint with the notice of removal gave the
appearance that there was an additional party to this
suit, but it was unclear if this party had been joined
in the action by the state court.
Even if the view I express is wrong, the critical issue in
this case remains whether the district court viewed removal
jurisdiction as a form of subject matter jurisdiction when it
ordered a remand. If it did, the remand order in this case is
unreviewable, even if also incorrect. See In re Decorator
Indus., Inc., 980 F.2d 1371, 1374 (11th Cir. 1992) (“We are
prohibited from reviewing a remand order for lack of jurisdiction
even when the district court’s determination is clearly
erroneous.”). For the reasons discussed in the main text, I am
firmly convinced that the district court viewed its remand order
as a subject matter-based jurisdictional order.
4
In granting the petition for removal, the federal court
must make this important initial decision with
certainty so as not to unduly interfere with the
jurisdiction of the state courts . . . .
712 F. Supp. at 179-80 (emphasis added). The second case cited
in the district court’s order, Cury v. Royal Palm Savings Ass’n,
713 F. Supp. 388, 389 (S.D. Fla. 1989), contains language to the
same effect.
In essence, these cases reflect a view that the district
court’s ability to make a proper determination of subject matter
jurisdiction is often inextricably linked to the formal contents
of the removing party’s petition for removal. The cited cases
also appear to endorse a categorical view that a district court
cannot determine subject matter jurisdiction with certainty where
a petition for removal is defective. Whether these views are
correct or not in every case is of no real moment for present
purposes, because reviewability under section 1447 turns on the
district court’s intent when making a remand determination, not
on the actual accuracy or wisdom of the determination. See In re
Decorator Indus., Inc., 980 F.2d at 1374; In re Amoco Petroleum,
964 F.2d at 713 (“A search for error is precisely what § 1447(d)
forbids.”). If the district court decided that it could not
ascertain subject matter jurisdiction in this case with certainty
because of a defective petition for removal, our tribunal has no
authority to review or second-guess that decision. 28 U.S.C. §
1447(d) (1994).
If any doubt remains that the district court intended to
make a jurisdictional-based remand order, the remaining
5
paragraphs of the district court order dispel that doubt. In the
paragraph immediately following the sentence regarding the
“[c]ourt’s limited removal jurisdiction,” the district court’s
order states that “all defendants must join the petition [for
removal] even if the basis for removal is a federal question.”
The district court then cites Darras v. Transworld Airlines, 617
F. Supp. 1068 (N.D. Ill. 1985), an opinion where the district
court sua sponte remanded a Warsaw Convention Treaty-based case
to a state court because one of the defendants did not join in
the removal petition, rendering the case “removed improvidently
and without jurisdiction.” 617 F. Supp. at 1069 (emphasis
added).
After then proceeding to describe several defects in the
instant petition for removal -- including the lack of an
allegation that all the defendants even consented to removal (as
required in section 1441(a)) -- the district court concluded its
order as follows:
Accordingly, having reviewed the amended notice of
removal, the record, and being otherwise duly advised,
it is hereby:
ORDERED AND ADJUDGED that this cause be REMANDED
to Circuit Court, Palm Beach County, Florida, and
DISMISSED from this Court’s Federal docket. It is
further hereby:
ORDERED AND ADJUDGED that because this Court did
not have jurisdiction over this cause, any and all
previously filed Orders of this Court are VACATED.
(Emphasis added.) The majority attributes no significance to
this language. Instead, the majority focuses on the district
6
court’s preceding discussion of various defects in the petition
for removal and confidently concludes that “the district court
remanded the case to state court for the failure of all
defendants to timely join in or consent to a petition for
removal.” See Majority Op. at 6 n.3.
If the district court intended to remand this case because
of the failure of all the defendants to timely join in a petition
for removal, one would expect the conclusion of the district
court’s order to say as much. We know, however, that the
district court said no such thing. Indeed, some of the words the
majority uses to describe the district court’s order never even
appear in the order’s text: not in the conclusion, and not in the
preceding discussion. The district court simply ends its order
with the statement that “this Court did not have jurisdiction
over this cause.” (Emphasis added.)
I would take this unambiguous statement for what it is -- an
unreviewable decision to remand for lack of subject matter
jurisdiction. That decision may have been predicated in part on
the deficient nature of the petition for removal in this case,
but that fact alone does not alter the jurisdictional nature of
the district court’s ultimate conclusion, especially in light of
the repeated references to jurisdictional concerns throughout the
entire district court order.
The majority also errs in establishing a hard and fast rule
that section 1447 does not bar appellate review and relief
whenever a district court remands a case sua sponte shortly after
7
the expiration of the thirty-day period within which the parties
must file motions for remands based on procedural defects. The
policy behind section 1447's thirty-day motions period, as
explained in the Fifth Circuit’s Loyd opinion, is to avoid
shuttling cases between state and federal courts based on purely
procedural defects when both state and federal courts have
jurisdiction over the relevant claims. FDIC v. Loyd, 955 F.2d
316, 322 (5th Cir. 1992). Section 1447’s thirty-day motions
period also operates to prevent the parties from engaging in
forum shopping once the litigation process has commenced in
earnest. See generally Charles Everingham IV, Removal, Waiver,
and the Myth of Unreviewable Remand in the Fifth Circuit, 45
Baylor L. Rev. 723 (1993) (discussing, inter alia, policy goals
Congress sought to advance in section 1447).
The rule section 1447(c) establishes is not a rigid and
absolute rule. The plain language of section 1447(c) only
requires the parties to file a motion noting a defect in the
removal process within thirty days. It does not require the
district court to rule on the motion within thirty days; nor does
it explicitly restrict the supervisory authority and discretion
of a district court to act sua sponte or grant equitable relief
in connection with section 1447(c); nor does it unequivocally
guarantee a federal forum to the parties once thirty days have
elapsed. Section 1447(c) is thus not a comprehensive measure
covering all contingencies, but a mechanism designed to prod the
parties to inform each other in a prompt way -- before litigation
8
commences in earnest -- if any of them has an objection to
litigation within the extant forum.
In this case, Edwards clearly placed Bethesda Hospital and
the other petitioners on notice that she objected to the forum(s)
of the litigation. She did so with motions for consolidation and
extension of time, which she filed on Friday, June 13, 1996 --
the twenty-seventh day following Bethesda’s notice of removal.
At that time, the parties had not commenced any litigation on the
merits of the claims at issue. The district court’s ruling a
week later (and only four days after the thirty-day time limit
expired) thus does not appear to have upset any truly settled
expectations the petitioners had of being in a particular forum -
- after all, the petitioners were not even all properly before
the same federal judge at the time. Indeed, if the district
court had granted Edwards’s motion for consolidation and invoked
the doctrine of equitable tolling to extend the thirty-day
motions period, cf. Roe v. O’Donohue, 38 F.3d 298, 302 (7th Cir.
1994) (assuming that the “30-day period [of section 1447(c)] is
subject to equitable tolling and estoppel”), the petitioners
might have quickly found themselves faced with a proper motion
from Edwards to remand the case to state court on procedural
and/or jurisdictional grounds. Given these facts, and the fact
that shuttling -- the principal evil section 1447 seeks to
eliminate -- will again occur here if review is permitted and
relief granted, I cannot agree that the district court’s decision
9
in this case is of a type we should find cognizable for review
and relief.
I believe that controlling Eleventh Circuit precedent
indicates that the proper standard of review for relief in a case
such as this is the standard applicable to petitions for a writ
of mandamus. See, e.g., New v. Sports & Recreation Inc., 114
F.3d 1092, 1094 n.3 (11th Cir. 1997) (“A writ of mandamus is the
proper means by which a party may challenge a remand order.”);
Loftin v. Rush, 767 F.2d 800, 802 n.3 (11th Cir. 1985) (“The
proper way to challenge a remand is by way of writ of mandamus,
not appeal.”).
Under this standard, an applicant for a writ of mandamus
must establish a “clear abuse of discretion or [conduct amounting
to an] usurpation of power” in order to establish a right to
relief. In re Lopez-Lukis, 113 F.3d 1187, 1187 (11th Cir. 1997);
In re Amoco Petroleum Additives Co., 964 F.2d 706, 713 (7th Cir.
1992). As we recently observed in In re Lopez-Lukis, 113 F.3d at
1187-88, “mandamus is an extraordinary remedy . . . [and] the
petitioners have the burden of showing that their right to
issuance of the writ is 'clear and indisputable.'” Moreover, as
our colleagues in the Seventh Circuit have noted, “mandamus is
not the appropriate means to resolve doubtful issues of procedure
or statutory construction.” In re Amoco Petroleum, 964 F.2d at
713. Since, in my view, the majority’s statutory interpretation
of section 1447(c)’s purview is not indisputably correct based on
the plain language of that statute, it is imprudent for us to
10
find a right of review and relief for the petitioners in this
case.
The majority avoids this issue altogether, claiming that the
Supreme Court overruled Loftin and its progeny in Quackenbush v.
Allstate Ins. Co., 116 S. Ct. 1712 (1996). According to the
majority, Quackenbush holds “that a district court’s order to
remand a case to state court is a final judgment that can be
reviewed on direct appeal.” Majority Op. at 2 (citing
Quackenbush, 116 S. Ct. at 1720.). I respectfully suggest that
this interpretation of Quackenbush overstates that opinion’s
holding. The Supreme Court’s opinion does not consider the
question of whether all remand orders can be reviewed on direct
appeal. To the contrary, the Supreme Court’s opinion determined
only that an abstention-based remand order is appealable under 28
U.S.C. § 1291. See Quackenbush, 116 S. Ct. at 1717 (“In this
case, we consider whether an abstention-based remand order is
appealable as a final order under 28 U.S.C. § 1291"); see also In
re United States Brass Corp., 110 F.3d 1261, 1267 (7th Cir. 1997)
(following Quackenbush, “[a]n order of abstention that takes the
form of a dismissal or a remand, rather than merely of a stay of
the proceedings before the district court is an appealable final
decision”). While it is true that some circuits have allowed
statutory-based remand order challenges to proceed through direct
appeal in the wake of Quackenbush, other circuits have continued
to adhere to their prior practice of reviewing statutory-based
remand orders through mandamus. Compare Gaming Corp. of Am. v.
11
Dorsey & Whitney, 88 F.3d 536, 542 (8th Cir. 1996) (direct
appeal), with In re Excel Corp., 106 F.3d 1197, 1200-01 (5th Cir.
1997) (mandamus).
Given the well-established rule in this circuit that only
the Supreme Court or the court of appeals sitting en banc may
overrule prior panel decisions -- as well as the fact that
neither party briefed or raised the issue -- I respectfully
believe the majority oversteps its bounds when it expands
Quackenbush beyond its abstention-based context to hold that
Loftin and its progeny are no longer good law. But see Ariail
Drug __ F.3d at __, slip op. at 3562 (mandate pending) (similarly
suggesting in obiter dictum that Quackenbush overrules Loftin).
Loftin is precisely on point and represents well-established law,
followed in this circuit as recently as this year. See New v.
Sports & Recreation, Inc., 114 F.3d at 1094 n.3. Quackenbush,
while controlling on the issue of abstention-based remand orders,
has been read narrowly in at least one other circuit, and no
party to this case even cites it in its brief. It is thus
neither necessary nor prudent for the majority to disregard
established precedent in this case to convert petitioners’s writ
of mandamus application into a direct appeal.
Because the majority makes an unwarranted break with
controlling circuit precedent, employs an erroneous standard of
review and reaches a result that appears wrong to me in every
regard, I respectfully dissent.
12