OPINION
CUDAHY, Circuit Judge.
Stephanie Elliot, a terminally ill former paralegal, sued Fortis Benefits Insurance Company, claiming benefits and other damages under a long-term disability insurance policy she had through her employer. She prevailed on her Employee Retirement Income Security Act (ERISA) claims, and recovered policy benefits and attorney’s fees and costs, but lost on For-tis’s motion for judgment on the pleadings on her state law claims, under which she sought non-ERISA compensatory and punitive damages. Although recent case law requires us to reconsider whether ERISA preempts such state law claims, we affirm the judgment of the district court.
I.
Stephanie Elliot was first diagnosed with cancer in 1995. She successfully underwent a right modified mastectomy, chemotherapy, stem-cell transplant and other treatment, and her breast cancer was found to be in complete remission. As of July 9,1997, her oncologists noted that she had “no measurable disease,” and regular examinations deemed her cancer-free as late as September 16, 1999. Nonetheless, to lessen the risk that she would develop cancer in the future, Elliot was placed on Tamoxifen
therapy beginning in July 1996. According to her doctor, the Tamoxifen was not prescribed for the treatment of any active cancer or other disease.
Elliot started working as a paralegal at the Crowley Law Firm on September 27, 1999. As part of her compensation, she was enrolled in Crowley’s long-term disability insurance policy with Fortis Benefits Insurance Company. This coverage was effective as of her first day of work. This policy, however, contained restrictions on coverage for pre-existing conditions, which were defined as follows:
A “pre-existing condition” means an injury, sickness, or pregnancy or any related injury, sickness, or pregnancy for which you:
• consulted with or received advice from a licensed medical or dental practitioner; or
• received medical or dental care, treatment, or services, including taking drugs, medicine, insulin, or similar substances during the 3 months that end on the day before you became insured under the long term disability insurance policy.
Defs Statement of Uncontroverted Facts Ex. A at 20, Record at 22. The policy excludes benefits “for any disability caused by a pre-existing condition” until three consecutive months without treatment have passed or the beneficiary has been continuously insured under the policy for twelve consecutive months.
The first sign that Elliot’s health was deteriorating came in December 1999, when her doctor noted elevated levels of Cancer Antigen 27.29, an early portent of cancer. Although her Cancer Antigen 27.29 levels subsequently declined, she suffered from growing headaches and pain in Summer 2000. On August 19, 2000, Elliot was diagnosed with brain and bone cancer. She was taken off Tamoxifen and did not
return to work. At the time of this appeal, she was totally disabled and terminally ill.
For the first three months of her disability, Elliot was paid full salary under Crowley’s short-term disability policy. Elliot then applied to Fortis for long-term disability benefits.
On October 30, 2000, Fortis sent Elliot a letter explaining the pre-existing condition limitation of her policy. The letter, however, in excerpting the policy provision dealing with pre-existing conditions, recited a requirement, not contained in the actual policy, that insureds be “at active work for a full day following” the three consecutive months of nontreatment or twelve consecutive months of coverage required under the policy. According to the new language, it appeared that Elliot would not qualify for coverage unless she returned to active work. The letter explained that Fortis would be conducting “a routine preexisting review” because Elliot’s date of disability, August 19, was within twelve months of the start of her coverage, September 27 of the preceeding year.
After a review of Elliot’s medical history, Fortis denied disability benefits. For-tis reasoned that because Elliot was taking Tamoxifen and undergoing periodic checkups, her earlier breast cancer constituted a pre-existing condition. Because Elliot had a pre-existing condition, she was covered only if she could show three consecutive months of nontreatment or twelve consecutive months of coverage. Fortis found that she had seen doctors and had prescriptions dispensed regularly, with no three-month gap, and that her disability began within twelve months of the start of her coverage. To reach these conclusions, Fortis relied in part on the opinion of Dr. Patrick Cobb, Elliot’s oncologist, who reported to Fortis that Elliot’s current cancer was a metastasis from her earlier breast cancer and that Tamoxifen was prescribed to prevent the recurrence of breast cancer and its metastasis.
Elliot filed suit in the District of Montana against Fortis, making two claims for relief. First, Elliot sought policy benefits under ERISA. Second, she asserted violations of Montana’s Unfair Trade Practices Act (UTPA),
MontCode Ann. § 33-18-201, and sought non-ERISA compensatory and punitive damages. She prevailed on the first count, and that is not before us. On the state law claim, the district court granted Fortis’s motion for judgment on the pleadings, Fed.R.Civ.P. 12(c), concluding that the state law claim is preempted by ERISA. Elliot appeals this ruling.
II.
We review
de novo
a Fed.R.Civ.P. 12(c) judgment on the pleadings.
McGann v. Ernst & Young,
102 F.3d 390, 392 (9th Cir.1996). ERISA preemption is a question of law, which we also review
de novo. Waks v. Empire Blue Cross/Blue Shield,
263 F.3d 872, 874 (9th Cir.2001).
The district court based its finding of preemption on two separate ERISA provisions. First, the district court held that the express preemption of ERISA § 514,
29 U.S.C. § 1144, which contains both a preemption and a saving clause, defeated Elliot’s state law claim. Second, the district court noted that even if it were to find the substantive provisions of the UTPA not preempted, ERISA § 502(a), 29 U.S.C. § 1132(a), would preempt the enforcement provision of the UTPA which allows her to enforce the state law rights. We consider in some detail both lines of analysis.
A.
We begin by noting a recent change in the law involving § 1144. Section 1144(a) expressly preempts “any and all State laws insofar as they may now or hereafter relate to any employee benefit plan” in favor of federal regulation under ERISA. However, § 1144(b)(2)(A) saves from preemption “any law of any State which regulates insurance, banking, or securities.” The parties agree that Montana’s UTPA, as applied here against Fortis, would fall under § 1144(a). How do we determine whether the UTPA is included in the class of laws regulating insurance that are saved by § 1144(b)(2)(A) from ERISA preemption?
Until April of the current year, courts determining the reach of the ERISA saving clause (including the district court in this case) used a two-step test set forth in
Metropolitan Life Ins. Co. v. Massachusetts,
471 U.S. 724, 740, 105 S.Ct. 2380, 85 L.Ed.2d 728 (1985). The
Metropolitan Life
test instructed courts to ask first whether, from a “common-sense view,” the state law regulates insurance. The second step was to determine whether the state law regulates “the business of insurance.” This is the telling phrase that was bor
rowed
from the McCarran-Ferguson Act.
See
15 U.S.C. § 1011 (“Congress hereby declares that the continued regulation and taxation by the several States of
the business of insurance
is in the public interest, and that silence on the part of the Congress shall not be construed to impose any barrier to the regulation or taxation of such business by the several States.”) (emphasis added). In order to determine whether a given state law regulated “the business' of insurance,” courts were to consider three elements laid out in
Union Labor Life Ins. Co. v. Pireno,
458 U.S. 119, 129, 102 S.Ct. 3002, 73 L.Ed.2d 647 (1982) (emphasis omitted): “first, whether the practice has the effect of transferring or spreading a policyholder’s risk; second, whether the practice is an integral part of the policy relationship between the insurer and the insured; and third, whether the practice is limited to entities within the insurance industry.”
But in
Kentucky Ass’n of Health Plans, Inc. v. Miller,
— U.S. —, —, 123 S.Ct. 1471, 1479, 155 L.Ed.2d 468 (2003), the Supreme Court “ma[de] a clean break from the McCarran-Ferguson factors” for determining the reach of the ERISA saving clause. “[F]or a state law to be deemed a ‘law ... which regulates insurance’ under § 1144(b)(2)(A), it must [now] satisfy two requirements. First, the state law must be specifically directed toward entities engaged in insurance.... Second, ... the state law must substantially affect the risk pooling arrangement between the insurer and the insured.”
Id.
(citations omitted). Thus, to determine the reach of § 1144 preemption, we
must ask whether the UTPA is “specifically directed toward entities engaged in insurance” and whether it “substantially affect[s] the risk pooling arrangement between the insurer and the insured.” As is immediately apparent, both these questions are actually remnants of the
Metropolitan Life
test, which asked as part of its analysis “whether the practice is limited to entities within the insurance industry,” and “whether the practice has the effect of transferring or spreading a policyholder’s risk.”
B.
Because of the similarities between the
Metropolitan Life
and
Kentucky Association
approaches, it is well worth considering the substantial body of case law applying the older test. Of these cases, perhaps the most relevant Supreme Court precedent is
Pilot Life Ins. Co. v. Dedeaux,
481 U.S. 41, 107 S.Ct. 1549, 95 L.Ed.2d 39 (1987).
In that case, a policyholder had sued Pilot Life for failure to pay benefits under a permanent disability policy that he had obtained through his employment. The Supreme Court held that a Mississippi state cause of action alternatively called “tortious breach of contract” and “the law of bad faith” was preempted by ERISA.
Id.
at 48, 57, 107 S.Ct. 1549. The Court held that the Mississippi law fell under the ERISA preemption clause, but was not saved by the ERISA saving clause, because it failed the
Metropolitan Life
test. Of this analysis, two findings made by the
Pilot Life
court remain relevant. The Court first found that Mississippi’s bad faith law was not directed specifically at insurance companies because while “the Mississippi Supreme Court has identified its law of bad faith with the insurance industry, the roots of this law are firmly planted in the general principles of Mississippi tort and contract law.”
Pilot Life,
481 U.S. at 50, 107 S.Ct. 1549. “Any breach of contract, and not merely breach of an insurance contract, may lead to liability for punitive damages under Mississippi law.”
Id.
Second, the Court found that the “common law of bad faith does not effect a spreading of policyholder risk.”
Id.
In addition to this straightforward application of the
Metropolitan Life
test,
Pilot Life
also introduced the concept of preemption by ERISA § 502(a), 29 U.S.C. § 1132(a). Accepting an argument made by the Solicitor General, the
Pilot Life
court stated that § 1132 (the private enforcement provision),
is “the exclusive vehicle for actions by ERISA-plan participants and beneficiaries asserting improper processing of a claim for benefits.”
Pilot Life,
481 U.S. at 52, 107 S.Ct. 1549. The Court compared the broad preemption of the ERISA enforcement provision to that of the Labor-Management Relations Act of 1947,
id.
at 51-54, 107 S.Ct. 1549, and concluded that, the
Metropolitan Life
test aside, state laws such as the Mississippi bad faith law may be preempted simply because they provide individuals with a private cause of action that is not available to them under ERISA.
We have previously followed
Pilot Life.
In
Greany v. Western Farm Bureau Life Ins. Co.,
973 F.2d 812, 819 (9th Cir.1992), we applied the § 1132 preemption of
Pilot Life
specifically to the Montana UTPA, holding that the UTPA “do[es] not ‘regulate[ ] insurance’ ... but [is a] civil enforcement provisionf ]” preempted by ERISA. Other Ninth Circuit cases have also found similar state laws preempted, seemingly relying on § 1132 preemption as
well as on § 1144 preemption and its
Metropolitan Life
test to dismiss state law insurance claims.
See Bast v. Prudential Ins. Co. of Am.,
150 F.3d 1003, 1008 (9th Cir.1998) (noting that “we have held that insurance bad faith claims are preempted by ERISA”);
Tingey v. Pixley-Richards West, Inc.,
953 F.2d 1124, 1133 (9th Cir.1992) (holding that an Arizona tort for breach of duty of good faith and fair dealing was preempted by ERISA);
Kanne v. Conn. Gen. Life Ins. Co.,
867 F.2d 489, 493-94 (9th Cir.1988) (holding that a California law prohibiting unfair insurance practices was preempted by 29 U.S.C. § 1132).
It would be relatively straightforward to apply the reasoning of
Pilot Life
and
Greany
to the present case. For one, we could look to § 1144 and the
Kentucky Association
test: Is the UTPA “specifically directed toward entities engaged in insurance” and does it “substantially affect the risk pooling arrangement between the insurer and the insured?” According to
Pilot Life,
the fact that Montana recognizes a tort of bad faith in non-insurance contexts,
see Story v. City of Bozeman,
242 Mont. 436, 791 P.2d 767, 776 (1990) (“The [bad faith] tort remedy may also be available in contracts ... not otherwise controlled by specific statutory provisions.”); Fortis Br. at 30-31, may defeat the fact that the UTPA is, superficially at least, specifically directed at insurance,
see
Mont.Code Ann. § 33-18-101 (“The purpose of this chapter is to regulate trade practices in the business of insurance.”). However, because the Montana law is statutory, unlike the Mississippi common law, our conclusion may be that the provision
is,
by its own terms, “specifically directed” at insurance companies. Second, the
Pilot Life
court remarked that the Mississippi bad faith law did not effect a spreading of risk. This conclusion would probably apply equally to the Montana law, resulting in the UTPA’s failing the second prong of the
Kentucky Association
test and not being saved by the ERISA saving clause.
Alternatively, we could look to
Pilot Life’s
§ 1132 preemption. Under that provision of
Pilot Life,
the UTPA could be preempted because UTPA’s civil enforcement provision, Mont.Code Ann. § 33 — 18— 242, violates the implicit requirement that § 1132 is “the exclusive vehicle for actions by ERISA-plan participants and beneficiaries asserting improper processing of a claim for benefits.”
Pilot Life,
481 U.S. at 52, 107 S.Ct. 1549.
In summary, there are three aspects of
Pilot Life
that still indicate preemption here: 1) claims processing laws are not “specifically directed” at insurance if bad faith claims can be made in other settings, 2) claims processing laws do not affect the risk-pooling arrangement and 3) all state private causes of action for claim processing are preempted by ERISA’s § 1132 enforcement provision. If any one of these
Pilot Life
conclusions is supported by subsequent case law, we must rely on the holding
of Greany
and affirm the district court. But we find that subsequent ease law puts the validity of all three of these
Pilot Life
conclusions into some doubt, making the present case considerably more difficult.
C.
The principal case challenging some of
Pilot Life’s
conclusions is
UNUM Life Ins. Co. of Am. v. Ward,
526 U.S. 358, 119 S.Ct. 1380, 143 L.Ed.2d 462 (1999). In
UNUM,
the Supreme Court considered whether a California law that validated claims filed after a filing deadline, provided that the delay did not cause prejudice to the insurance company, a so-called “notice-prejudice” rule, was preempted by ERISA. A unanimous Supreme Court concluded, upholding the Ninth Circuit, that the notice-prejudice rule is saved by the ERISA sav
ing clause.
Id.
at 373, 119 S.Ct. 1380. In doing so,
UNUM
cast doubt on all three of the
Pilot Life
principles that are potentially controlling here.
First,
UNUM
took a much more generous approach than did
Pilot Life
in analyzing whether a law was specifically directed at insurance. The Court found that the California law “by its very terms, ‘is directed specifically at the insurance industry and is applicable only to insurance contracts,”’ and rejected UNUM’s argument that the general disposition of California law against forfeiture, including cases applying something akin to a notice-prejudice rule in non-insurance contexts, somehow defeated the circumstance that the California notice-prejudice rule itself was literally directed at insurance companies.
UNUM,
526 U.S. at 368-73, 119 S.Ct. 1380. Applying this view, the UTPA probably meets the first requirement of the
Kentucky Association
test, since the UTPA “by its very terms” is directed at insurance and contains provisions applicable only to insurance companies.
See, e.g.,
Mont.Code Ann. § 33-18-101 (“The purpose of this chapter is to regulate trade practices in the business of insurance -”);
id.
§ 33-18-201 (forbidding certain claim settlement procedures).
UNUM
also calls into question
Pilot Life’s
conclusion that a claim processing law does not affect risk allocation. Although this was not a question the
UNUM
court reached, the
UNUM
court noted without disapproval the federal government’s amicus brief argument that “[ijnso-far as the notice-prejudice rule shifts the risk of late notice and stale evidence from the insured to the insurance company in some instances, it has the effect of raising premiums and spreading risk among policyholders.”
UNUM,
526 U.S. at 374, 119 S.Ct. 1380. The
Kentucky Association
court confirmed that the notice-prejudice rule “certainly qualifies as [having] a substantial effect on the risk pooling arrangement between the insurer and insured.”
Kentucky Association,
— U.S. at - n. 3, 123 S.Ct. at 1477 n. 3.
UNUM
and
Kentucky Association
thus leave open the possibility that risk spreading might be found in a much wider variety of statutes than
Pilot Life
suggested. Under these potentially broader concepts of risk-pooling, it is entirely plausible to find that the UTPA allocates risk “by obligating insurers to make advance payments before the duty to indemnify under the policy is triggered” and “by requiring insurers to pay excess judgments against their insureds” in the case of a UTPA violation.
See
Elliot’s Opening Br. at 22 (citing
Ridley v. Guaranty Nat’l Ins. Co.,
286 Mont. 325, 951 P.2d 987, 992 (1997), which held that, under the UTPA, insurance companies must make payments as soon as liability becomes reasonably clear).
But see Howard v. Coventry Health Care, of Iowa, Inc.,
293 F.3d 442, 447 (8th Cir.2002) (reiterating,
post-UNUM,
that bad faith laws do not spread risk).
Most significantly, perhaps,
UNUM
questioned the preemptive force of ERISA’s civil enforcement provision, 29 U.S.C. § 1132(a). UNUM Life Insurance had argued that § 1132 barred any state cause of action involving claim processing, a broad, literal reading of
Pilot Life.
The
UNUM
court found this argument irrelevant because the
UNUM
plaintiff had not brought the action based on a state cause of action; instead, the California statute became relevant only when cited to meet the insurer’s defense that the plaintiffs claim was untimely. However, the Court specifically noted in its footnote seven the Solicitor General’s argument that the ERISA saving cause, 29 U.S.C. § 1144(b)(2)(A), also saves state causes of action that regulate insurance, i.e., that as long as the state law being enforced is
saved as a law regulating insurance under § 1144(b)(2)(A), its enforcement provision is also saved from § 1132 preemption.
See UNUM,
526 U.S. at 376 n. 7, 119 S.Ct. 1380. With this footnote (bringing potential § 1132 preemption within the preclusive coverage of § 1144(b)(2)(A)),
UNUM
appeared to narrow significantly
Pilot Life’s
language on § 1132’s independent preemptive power.
D.
Thus, none of
Pilot Life’s
three principles noted above escaped unquestioned by
UNUM.
Following the latter case, we might well find the UTPA to be specifically directed toward entities engaged in insurance, and it is certainly possible that we would find the UTPA to substantially affect the risk pooling arrangement between the insurer and the insured.
See, e.g., Rosenbaum v. Unum Life Ins. Co. of Am.,
No. Civ. 01-6758, 2002 WL 1769899 (E.D.Pa. July 29, 2002) (finding, under a post-UNCTM application of the
Metropolitan Life
test, a state bad faith law not to be preempted by ERISA). As things stand, however, we do not have to reach these issues because
Pilot Life’s
theory of § 1132 preemption has been narrowly but sufficiently re-affirmed by the Supreme Court subsequent to
UNUM’s
footnote seven and the questions it might raise.
The level of preemptive force of § 1132 was clarified in
Rush Prudential HMO, Inc. v. Moran,
536 U.S. 355, 122 S.Ct. 2151, 153 L.Ed.2d 375 (2002). In considering an Illinois law that provided a right to independent medical review of certain denials of benefits, the Supreme Court concluded that, while the independent review may resemble a separate arbitration proceeding going beyond ERISA’s enforcement provisions, the independent review was not precluded by the preemption principle of
Pilot Life. Rush Prudential,
536 U.S. at 380-83, 122 S.Ct. 2151. In reaching this conclusion,
Rush Prudential
removed much of the precedential force of this aspect of
Pilot Life,
calling
Pilot Life’s
discussion of preemption by § 1132
“dictum.” Id.
at 377, 122 S.Ct. 2151. The
Rusk Prudential
court explained the genesis of
Pilot Life’s
invocation of § 1132, and recharacterized
Pilot Life’s
holding to mean only that “ERISA would not tolerate a diversity action seeking monetary damages for breach generally and for consequential emotional distress, neither of which Congress had authorized in § 1132(a).”
Id.
at 378, 122 S.Ct. 2151. The
Rush Prudential
court noted that only once since
Pilot Life,
in
Ingersoll-Rand Co. v. McClendon,
498 U.S. 133, 111 S.Ct. 478, 112 L.Ed.2d 474 (1990), had it re-affirmed the
Pilot Life
conclusion that § 1132 might preempt a state remedy.
Rush Prudential,
536 U.S. at 379, 122 S.Ct. 2151. According to the Court in
Rush Prudential,
the problem in
Inger-solL-Rand
was that a “state law duplicated the elements of a claim available under ERISA” while converting the remedy “from an equitable one under § 1132(a)(3) .. i into a legal one for money damages.”
Id.
The Court explained, thus narrowly reaffirming
Pilot Life,
that preemption by § 1132 constitutes “a limited exception from the saving clause” for causes of action and remedies alternative to ERISA.
Id.
at 381, 122 S.Ct. 2151. While the Court described this exception to the saving clause as quite limited, the exception does include state causes of action that would “significantly expand[ ] the potential scope of ultimate liability imposed upon employers” by ERISA, and such claims are preempted.
Id.
at 379, 122 S.Ct. 2151. In other words, § 1132’s preemptive effect depends on the nature of the state remedy, including the availability of non-ERISA compensatory and punitive damages. Thus, the availability of such non-ERISA damages is a strong indicator that a state cause of action may be preempted.
Applying
Rush Prudential
as controlling here would dictate preemption. But there are two possible distinctions between
Rush Prudential
and the instant case. First, Montana’s UTPA arguably does not fall squarely under
Rush Prudential’s
characterization of
Ingersoll-Rand.
According to
Rush Prudential,
the problem with the state law cause of action in
Ingersoll-Rand
was that it “duplicated the elements of a claim available under ERISA, [but] converted the remedy.”
Rush Prudential,
536 U.S. at 379, 122 S.Ct. 2151. In contrast, here the UTPA may instead provide rights in addition to those provided by ERISA by, for example “obligating insurers to make advance payments before the duty to indemnify under the policy is triggered.”
Ridley,
951 P.2d at 992. We do not, however, read
Rush Prudential’s
re-affirmation of
Pilot Life
as limited narrowly to the precise circumstances of
Ingersoll-Rand. See Rush Prudential,
536 U.S. at 380, 122 S.Ct. 2151 (citing
Pilot Life, Ingersoll-Rand
and
Mass. Mut. Life Ins. Co. v. Russell,
473 U.S. 134, 105 S.Ct. 3085, 87 L.Ed.2d 96 (1985), as examples of the types of cases invoking § 1132 preemption). The instant action, which seeks non-ERISA damages for what are essentially claim processing causes of action, clearly falls under the § 1132 preemption exemplified by
Pilot Life.
Also, Elliot asks us to distinguish
Rush Prudential
on the ground that its main concern was the potential liability imposed upon
employers,
whereas the present action is against an insurance com-party rather than an employer. While this may be true, the nature of the defendant alone does not suffice to save a claim which conflicts with ERISA’s enforcement scheme.
Cf. Bast,
150 F.3d at 1008 (noting that for questions of ERISA preemption, “[t]he key issue is whether the parties’ relationships are ERISA-governed relationships” and not the precise roles played by the parties).
Section 1132, as construed by
Rush Prudential,
thus preempts Elliot’s claims. Unlike the notice-prejudice rule involved in
UNUM,
which only provided a “relevant rule of decision” for a suit brought under § 1132,
UNUM,
526 U.S. at 377, 119 S.Ct. 1380, Elliot’s claim relies in the first instance on Montana’s UTPA’s civil enforcement provision, Mont.Code Ann. § 33-18-242. Unlike the
not
preempted independent review mechanism of
Rush Prudential,
Mont.Code Ann. § 33-18-242 provides damages above and beyond those provided in ERISA, including punitive damages. Because the present case “involve[s] the sort of additional claim or remedy exemplified in
Pilot Life,” Rush Prudential,
536 U.S. at 380, it falls within § 1132 preemption.
Accord Fink v. Dakotacare,
324 F.3d 685, 689 (8th Cir.2003) (applying
Rush Prudential
and declaring a state claims processing law preempted);
Conover v. Aetna U.S. Health Care, Inc.,
320 F.3d 1076, 1079-80 (10th Cir.2003) (same);
Caffey v. UNUM Life Ins. Co.,
302 F.3d 576, 582 (6th Cir.2002) (same);
Emil v. UNUM Life Ins. Co. of Am.,
No.
3:CV02-2019, 2003 WL 256781 (M.D.Pa. Feb.5, 2003) (same).
III.
We review a district court’s decision to award attorney’s fees under ERISA for abuse of discretion.
Williams v. Caterpillar, Inc.,
944 F.2d 658, 668 (9th Cir.1991). ERISA provides that a “court in its discretion may allow a reasonable attorney’s fee and costs of action to either party.” 29 U.S.C. § 1132(g)(1). In order to determine the appropriateness of an award of fees, the court should generally consider five factors:
(1) the degree of the opposing parties’ culpability or bad faith; (2) the ability of the opposing party to satisfy an award of fees; (3) whether an award of fees ... would deter others from acting under similar circumstances; (4) whether the parties requesting fees sought to benefit all participants and beneficiaries of an ERISA plan or to resolve a significant legal question regarding ERISA; and (5) the relative merits of the parties’ positions.
Hummell v. S.E. Rykoff & Co.,
634 F.2d 446, 453 (9th Cir.1980);
see also Landwehr v. DuPree,
72 F.3d 726, 739 n. 5 (9th Cir.1995).
Fortis argues that much of the fees Elliot incurred were not compensable under § 1132(g)(1) because those fees were incurred pursuing non-ERISA claims. However, we have held that the
Hummell
factors should be “liberally construed in favor of protecting participants in employee benefits plans.”
McElwaine v. U.S. West, Inc.,
176 F.3d 1167, 1172 (9th Cir.1999). Successful plaintiffs in ERISA suits should ordinarily recover fees unless special circumstances would render such an award unjust.
Id.
The district court’s grant of fees was well within its discretion.
IV.
For the foregoing reasons, we AFFIRM the decision of the district court.