EMILIO M. GARZA, Circuit Judge:
This appeal concerns the disposition of unclaimed funds from a class action settlement. The plaintiffs and defendants settled the underlying antitrust claims, and funds from the settlement were allocated to various identified members of the plaintiff class. The settlement administrator sent checks to the last known addresses of plaintiffs, but many were returned as undeliverable or were never cashed. The district court invoked the doctrine of
cy pres
and ordered that all unclaimed funds be distributed to the Center for Energy and Environmental Resources at the University of Texas (“Center”). The State of Texas (“Texas” or “State”) intervened seeking to enforce its right to custody of and investment income from unclaimed funds that had been allocated to plaintiffs with last known addresses in Texas. Texas and the Appellees filed cross-motions for summary judgment. The district court granted the Appellees’ motion for summary judgment, denied the State’s, and ordered that the funds allocated to Texas plaintiffs be returned to the fund subject to
cy pres
distribution. Texas appeals. For the reasons below, we REVERSE and VACATE the judgment of the district court and REMAND for further proceedings in accordance with this opinion.
I
Many of the details of this case can be found in
In re Lease Oil Antitrust Litigation,
570 F.3d 244 (5th Cir.2009)
(“Lease Oil”).
The plaintiffs (“Appellees”) filed a class action antitrust suit against various oil companies. The parties settled, the district court approved the settlements, and settlement orders were entered in 1999. Under the terms of the settlements, identified plaintiffs were entitled to payment of funds by the settlement administrator. The settlement administrator accordingly sent out thousands of checks to addresses of identified plaintiff class members. Many checks, however, were never cashed or were returned as undeliverable. While the administrator reissued checks and attempted to distribute the unclaimed funds, over ten million dollars eventually remained. Those remaining funds came from three sources: checks that were mailed, but never cashed; checks that were returned as undeliverable; and settlement awards below a
de minimis
amount. Of the over ten million dollars, more than four million had been allocated to plaintiffs whose last known addresses are in Texas. The settlement agreements themselves, however, did not resolve what the settlement administrator was to do with the remaining unclaimed funds. Rather, the agreements merely provided
that, if funds remained unclaimed, the settling parties would apply to the district court for directions regarding the disposition thereof.
Unable to distribute the funds to their rightful owners, the district court decided instead to distribute those funds by
cy pres
order to the Center. After the district com't approved the
cy pres
distribution, Texas filed a motion to intervene and a motion to reconsider. Texas asserted that, under the State’s unclaimed property laws, it had a right to custody of the funds allocated to individuals with last known addresses in Texas, as well as a property right to investment income from those funds. The district court denied the motions, concluding that Texas’s intervention was untimely. Anticipating an appeal, however, the district court ordered that the funds allocated to plaintiffs with last known addresses in Texas be placed in a separate account pending the resolution of Texas’s claims. Texas appealed the denial of its motion to intervene. We concluded that Texas’s motions were timely and that it had met the requirements for intervention as of right. We therefore reversed the denial of intervention and remanded for further proceedings.
Lease Oil,
570 F.3d at 252.
Upon remand, Texas sought custody of the disputed funds. Texas and the Appellees filed cross-motions for summary judgment. The district court granted judgment in favor of the Appellees, concluding that the court was permitted to dispose of the funds via
cy pres,
regardless of the terms of the State’s unclaimed property statutes. It granted summary judgment to the Appellees, denied summary judgment to Texas, and ordered that the funds the prior order had set aside be returned to the general unclaimed funds account. Texas now appeals.
II
Texas argues that the disposition of the unclaimed funds allocated to Texas plaintiffs should have been governed by the Texas Unclaimed Property Act (“Unclaimed Property Act” or “Act”), Tex. Prop. Code Ann. §§ 72.001-74.710. Although the Appellees ultimately contend that state law should not control the disposition of the funds, they argue, in the alternative, that the Unclaimed Property Act does not reach funds held by the settlement administrator in a class action case in federal court. Because Texas’s interest in the funds is premised on the applicability of the Act, we consider this issue first.
Texas argues that, under the Act, the disputed funds are “property that is presumed abandoned” and therefore the “holder” of the funds must provide them to the Texas comptroller (“Comptroller”). Tex. Prop.Code Ann. § 74.301. The Comptroller is then empowered to invest the unclaimed funds, with any resulting income to be allocated to the State. Tex. Prop.Code Ann. § 74.601(b)(4), (d). At any time, however, the rightful owner of the unclaimed funds may file a claim with the Comptroller, and the Comptroller is instructed to pay all valid claims. Tex. Prop. Code Ann. § 74.501(b). For the purposes of the Act, the holder is the person who is “(1) in possession of property that belongs to another; (2) a trustee; or (3) indebted to another on an obligation.” § 72.001(e). Texas argues that the holder of the funds it seeks is the settlement administrator. The Appellees seemingly concede that the settlement administrator fits the statutory definition provided for the holder of unclaimed funds, but argue that, because the settlement administrator is merely carrying out the orders of the district court, we should proceed as if the district court is the holder. The Appellees then argue that a federal district court cannot be a holder
under the Act, and therefore the Act does not apply.
The Appellees’ argument is without merit. The settlement administrator is plainly a holder, as that term is defined under the Act, because the settlement administrator is “in possession of property that belongs to another.”
Id.
The Appellees have, moreover, identified no exception on the face of the Act that is applicable to the settlement administrator. We will not engage in far-ranging discussion regarding the applicability of the Act to funds held directly by a district court, based merely on the speculation that the district court could, in the words of the Appellees, “order[ ] the funds to be deposited into the registry of the court tomorrow.” Appellees’ Br. at 41. The funds are in the possession of the settlement administrator, and the settlement administrator fits the Act’s definition of “holder.” The Act applies.
Cf. State v. Snell,
950 S.W.2d 108, 111-13 (Tex.App.—El Paso 1997, no writ) (enforcing Unclaimed Property Act to reverse distribution of unclaimed funds in state court class action).
Ill
The State argues that the Unclaimed Property Act is substantive state law that does not conflict with federal law, and therefore the settlement administrator should comply with the terms of the Act and the district court cannot order the funds to be otherwise distributed via the doctrine of
cy pres.
The Appellees counter that settlements binding members of a plaintiff class are governed by Rule 23(e) of the Federal Rules of Civil Procedure, and that Rule 23(e) grants the district court broad discretion to approve settlement terms. The Appellees note that the original settlement agreements in this case reserved the district court’s authority to approve the eventual method for distributing any unclaimed funds, and therefore argue that the eventual use of
cy pres
fell within the scope of its broad discretion of Rule 23(e).
When the original settlement agreements were approved in 1999, Rule 23(e) consisted of little more than a straightforward requirement that a class action settlement be made pursuant to judicial approval and notice to the class: “A class action shall not be dismissed or compromised without the approval of the court, and notice of the proposed dismissal or compromise shall be given to all members of the class in such manner as the court directs.” Fed.R.CivP. 23(e) (1999). When the district court first ordered
cy pres
distribution, Rule 23(e) had been substantially amended to “strengthen the process of reviewing proposed class-action settlements” in order “to assure adequate representation of class members who have not participated in shaping the settlement.” Fed.R.CivP. 23(e), 2003 advisory committee’s note. Both Texas and the Appellees have proceeded, before us, as if the amended version applies in this case. We will therefore assume, without deciding, that the posb-2003 version of the Rule is the relevant one for our analysis. In any event, although the new version of the Rule includes additional procedures, the fundamental role of the district court is largely the same: review and, if appropriate, approval of the settlement reached by
the parties.
See In re Katrina Canal Breaches Litig.,
628 F.3d 185, 194-95 (5th Cir.2010) (discussing role of district court in approving class action settlement).
The background principles governing whether courts should apply state or federal law can be found in
Erie Railroad Co. v. Tompkins,
304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938). When a party has alleged a direct conflict between the Federal Rules and state law, however, an additional step precedes the
Erie
analysis.
Hanna v. Plumer,
380 U.S. 460, 469-70, 85 S.Ct. 1136, 14 L.Ed.2d 8 (1965). “The initial step is to determine whether, when fairly construed, the scope of [the Rule] is ‘sufficiently broad’ to cause a ‘direct collision’ with the state law or, implicitly, to ‘control the issue’ before the court, thereby leaving no room for the operation of that law.”
Burlington N. R.R. Co. v.
Woods, 480 U.S. 1, 4-5, 107 S.Ct. 967, 94 L.Ed.2d 1 (1987) (quoting
Walker v. Armco Steel Corp.,
446 U.S. 740, 749-50 & n. 9, 100 S.Ct. 1978, 64 L.Ed.2d 659 (1980);
Hanna,
380 U.S. at 471-72, 85 S.Ct. 1136). In determining whether the Rule covers a particular issue, we look to the plain meaning of the Rule’s language.
Walker,
446 U.S. at 750 n. 9, 100 S.Ct. 1978. If we determine that such a “direct collision” does occur, we must apply the Federal Rule as long as that Rule is a valid exercise of Congress’s rulemaking authority.
Burlington N.,
480 U.S. at 4-5, 107 S.Ct. 967. Only if this initial inquiry is not determinative will we wade into the “murky waters” of
Erie
itself.
Shady Grove Orthopedic Assocs., P.A. v. Allstate Ins. Co.,
— U.S. -, 130 S.Ct. 1431, 1437, 176 L.Ed.2d 311 (2010).
The Supreme Court has never determined whether Rule 23(e) permits a district court to disregard state unclaimed property laws. The Court has, however, considered the effect of other provisions of Rule 23 on state laws. In
Shady Grove Orthopedic Associates, P.A. v. Allstate Insurance Co.,
the Court concluded that a New York law prohibiting class actions in suits seeking penalties or statutory minimum damages conflicted with Rule 23(a)-(b), which provides that a class action “may be maintained” if certain requirements — none of which bar penalties or statutory damages — are met. 130 S.Ct. at 1436 & nn. 1-2, 1438 (quoting Fed.R.Civ.P. 23(b)). The Court concluded that Rule 23 “[b]y its terms ... creates a categorical rule entitling a plaintiff whose suit meets the specified criteria to pursue his claim as a class action.”
Id.
at 1437. The Court stressed that a district court has a nondiscretionary duty to permit a class action that complies with Rule 23(a)-(b):
Allstate asserts that Rule 23 neither explicitly nor implicitly empowers a federal court “to certify a class in each and every case” where the Rule’s criteria are met. But that is
exactly
what Rule 23 does: It says that if the prescribed preconditions are satisfied “[a] class action
may be maintained”
(emphasis added) — not “a
class action may be permitted.”
Courts do not maintain actions; litigants do. The discretion suggested by Rule 23’s “may” is discretion residing in the plaintiff: He may bring his claim in a class action if he wishes.
Id.
at 1438 (citation omitted). Because Rule 23 required a district court to authorize suits that the New York statute forbade, the Court concluded that the two laws were in inescapable conflict.
Id.
at 1442.
Rule 23(e), in contrast, contains no categorical rule entitling plaintiffs to
cy pres
distribution — and, in fact, does not mention
cy pres
distribution at all. Appellees argue instead that the Unclaimed Property Act collides with Rule 23(e) by reducing the scope of a district court’s
discretion to approve a settlement. The Appellees rely heavily on
Burlington Northern Railroad Co. v. Woods,
in which the Supreme Court concluded that an Alabama statute imposing a 10% “mandatory affirmance penalty” on unsuccessful civil appellants who meet certain criteria was in conflict with Fed. R.App. P. 38’s grant of discretionary authority to impose damages on those who bring frivolous appeals. 480 U.S. at 3-4, 107 S.Ct. 967 (citing Ala.Code § 12-22-72 (1986)). The Court found that the mandatory nature of the Alabama statute and the discretionary nature of the Federal Rule gave rise to a conflict:
Rule 38 affords a court of appeals plenary discretion to assess “just damages” in order to penalize an appellant who takes a frivolous appeal and to compensate the injured appellee for the delay and added expense of defending the district court’s judgment. Thus, the Rule’s discretionary mode of operation unmistakably conflicts with the mandatory provision of Alabama’s affirmance penalty statute. Moreover, the purposes underlying the Rule are sufficiently coextensive with the asserted purposes of the Alabama statute to indicate that the Rule occupies the statute’s field of operation so as to preclude its application in federal diversity actions.
Id.
at 7, 107 S.Ct. 967.
The Appellees argue that “[precisely the same” analysis that prevailed in
Burlington Northern
dictates that we permit the district court to disregard the Unclaimed Property Act in this case. Appellees’ Br. at 12. We disagree.
Burlington Northern
involved an explicit grant of discretion on a specific issue: the award of costs and damages for a frivolous appeal. 480 U.S. at 4, 107 S.Ct. 967 (citing Fed.R.Civ.P. 38 & advisory committee’s notes). Rule 23(e), in contrast, merely empowers a district court to approve a settlement. Rule 23(e) does not mention the district court’s discretion — or even its authority— to extinguish the right of recovery of identified class members through a later
cy pres
order. Appellees read the judicial approval requirement as a blanket authorization for the district court to disregard state laws throughout the administration of an approved settlement, even long after the initial settlement orders are entered. Neither the language nor the purpose of the rule justifies so broad a reading. “The gravamen of an approvable proposed settlement is that it be fair, adequate, and reasonable and is not the product of collusion between the parties.”
Newby v. Enron Corp.,
394 F.3d 296, 301 (5th Cir.2004) (citation and internal quotation marks omitted). “ ‘The purpose of this salutary requirement is to protect the nonparty members of the class from unjust or unfair settlements affecting their rights’ as well as to minimize conflicts that ‘may arise between the attorney and the class, between the named plaintiffs and the absentees, and between various subclasses.’”
Strong v. BellSouth Telecomms., Inc.,
137 F.3d 844, 849 (5th Cir.1998) (quoting
Piambino v. Bailey,
610 F.2d 1306, 1327 (5th Cir.1980)). Specifically, we have identified six factors that guide- the court’s review:
(1) evidence that the settlement was obtained by fraud or collusion; (2) the complexity, expense, and likely duration of the litigation; (3) the stage of the litigation and available discovery; (4) the probability of plaintiffs’ prevailing on the merits; (5) the range of possible recovery and certainty of damages; and (6) the opinions of class counsel, class representatives, and absent class members.
In re Katrina Canal Breaches Litig.,
628 F.3d at 194-95 (quoting
Newby,
394 F.3d at 301 (5th Cir.2004)). Nothing about this crucial — but limited — inquiry implies an
authority to disregard state property laws.
In sum, Rule 23(e) only “collides” with the Act in this case if one construes the Rule to include a blanket authorization to disregard state property laws in the context of administering a settlement. Nothing in the Rule’s text or structure leads us to adopt such an aggressive construction. Nor does the text or structure suggest that the Rule implicitly occupies the field that the Act seeks to regulate. We conclude, therefore, that Rule 23(e) does not preclude the application of the Act to unclaimed funds allocated to identified class members in this case. The Act survives the initial
Hanna
analysis.
IV
We turn, then, to whether the Act applies under
Erie.
Texas argues that the district court was bound to apply the Act under
Erie;
the Appellees argue that it was not. The black letter rule of
Erie
is that federal courts “apply state substantive law and federal procedural law.”
Foradori v. Harris,
523 F.3d 477, 486 (5th Cir.2008) (citing
Gasperini v. Ctr. for Humanities, Inc.,
518 U.S. 415, 426-27, 116 S.Ct. 2211, 135 L.Ed.2d 659 (1996)). As courts have discovered, however, “[classification of a law as ‘substantive’ or ‘procedural’ for
Erie
purposes is sometimes a challenging endeavor.”
Gasperini,
518 U.S. at 427, 116 S.Ct. 2211. Accordingly, the Supreme Court has set forth a number of touchstones to consider.
See, e.g., Gasperini,
518 U.S. at 427, 116 S.Ct. 2211;
Hanna,
380 U.S. at 471, 85 S.Ct. 1136;
Byrd v. Blue Ridge Rural Elec. Coop., Inc.,
356 U.S. 525, 538-39, 78 S.Ct. 893, 2 L.Ed.2d 953 (1958);
Guaranty Trust Co. v. New York,
326 U.S. 99, 109, 65 S.Ct. 1464, 89 L.Ed. 2079 (1945). Appellees argue that we must confine our analysis to one particular touchstone, the “twin aims” analysis set forth in
Hanna.
Appellees posit that, under that analysis, the laws at issue are procedural. Texas argues that we should not rely on the “twin aims” analysis, because that analysis overemphasizes factors that are irrelevant when the party seeking to enforce state law is a post-judgment intervenor. Texas argues that we should instead conclude that the laws at issue are substantive, because any seemingly procedural aspects of those laws are bound up with substantive state property rights. In the alternative, Texas argues that it would also prevail under a pure “twin aims” analysis.
The relationship between the various doctrines that courts have developed under
Erie
makes most sense in the historical context of the developing case law. The “outcome determination” test, which has provided the foundation for most subsequent analyses, was set forth in
Guaranty Trust Co. v. New York:
The question is whether [the state law] concerns merely the manner and the means by which a right to recover, as recognized by the State, is enforced, or whether [it] is a matter of substance in the aspect that alone is relevant to our problem, namely,
does it significantly affect the result of a litigation for a federal court to disregard a law of a State that would be controlling in an action upon the same claim by the same parties in a State court?
326 U.S. at 109, 65 S.Ct. 1464 (emphasis added). Soon thereafter, however, the Court conceded that whether a rule might determine the outcome of particular litigation could not be the “only consideration” guiding courts.
Byrd,
356 U.S. at 537, 78 S.Ct. 893.
In Byrd v. Blue Ridge Rural Electric Cooperative, Inc.,
the Court suggested that courts might also consider
inter alia
whether a state rule is “bound up” with state-secured substantive rights and obligations. 356 U.S. at 535-38, 78 S.Ct. 893. In
Hanna,
the Supreme Court further refined its test, instructing courts to look to the “twin aims” of
Erie:
“discouragement of forum-shopping and avoidance of inequitable administration of the laws.”
Hanna,
380 U.S. at 468, 85 S.Ct. 1136;
see also Hall v. GE Plastic Pac. PTE Ltd.,
327 F.3d 391, 395 (5th Cir.2003). As the Supreme Court has made clear, the “twin aims” analysis does not supplant the “outcome determination” test, but rather is intended to qualify that test and guide its application, so that the “outcome determination” inquiry is not “applied mechanically to sweep in all manner of variations.”
Gasperini,
518 U.S. at 428, 116 S.Ct. 2211.
The Appellees argue that we should consider only the concerns explicitly acknowledged in the “twin aims” analysis and disregard any factors, such as the “bound up with” inquiry set forth in
Byrd,
that cannot be fit neatly into the Appellees’ narrow reading of those aims. As the district court acknowledged, however, a review of case law reveals that
Byrd
and the “twin aims” analysis can function, and have functioned, side by side. Certainly, we have not stopped citing or relying on
Byrd. See, e.g., Hall,
327 F.3d at 395 (citing
Byrd,
356 U.S. at 537, 78 S.Ct. 893);
Vaught v. Showa Denko,
107 F.3d 1137, 1146 (5th Cir.1997) (same);
Herbert v. Wal-Mart Stores, Inc.,
911 F.2d 1044, 1047-48 (5th Cir.1990) (per curiam) (same). While
Byrd
is no longer the last word on the substantive/procedural distinction, it is by no means irrelevant.
Texas is moreover correct that some aspects of the “twin aims” analysis — in particular, the focus on forum shopping— seem better suited to disputes between plaintiffs and defendants, rather than those involving post-judgment intervenors. Regardless, we agree with the Appellees that our precedents’ embrace of the “twin aims” analysis prevents us from disregarding that analysis altogether.
See Hall,
327 F.3d at 395 (“To determine whether an issue is substantive or procedural, this Court
must
consider the ‘twin aims’ of
Erie....
” (emphasis added)). The fact that we must consider the “twin aims,” however, does not preclude us from also considering the unique circumstances of this litigation. It is well-settled that “[t]he line between ‘substance’ and ‘procedure’ shifts as the legal context changes.”
Hanna,
380 U.S. at 471, 85 S.Ct. 1136. We therefore will not rotely apply the “twin aims” test without considering which aspects of that test are most salient in light of the posture of the case before us.
As to the first “twin aim,” Texas seemingly concedes that the availability of
cy pres
does not pose a significant threat of forum-shopping by plaintiffs. Lack of an apparent forum-shopping threat, however, is not fatal to Texas’s position if other considerations show that the purposes of
Erie
support applying the Act.
See Walker,
446 U.S. at 753, 100 S.Ct. 1978 (applying state law despite finding that “in this case failure to apply the state ... law might not create any problem of forum shopping”). Citing the second of the “twin aims,” Texas argues that permitting federal courts to disregard the Unclaimed Property Act in favor of
cy pres
distribution, while state courts follow the Act, would lead to the inequitable administration of justice.
See Hanna,
380 U.S. at 467-68, 85 S.Ct. 1136. We agree. Under the Act, the State has an enforceable property right in the income from unclaimed property.
Lease Oil,
570 F.3d at 251 (citing Tex. Prop.Code Ann. § 74.709). If the district court is permitted to disregard the Act, the result will be the elimination of that property right, based purely on the fact that this case happened to be settled in federal, rather than state, court. Moreover, identified class members who would have a right to recover their property from the State under the Act would instead lose that right of recovery forever — again, solely because the case was in federal court. Such a holding would, in our view, amount to inequitable administration of the laws.
Byrd
only bolsters that conclusion.
Cf. Herbert,
911 F.2d at 1047-48 (applying
Byrd
as part of the “inequitable administration” analysis). The aspects of the Act that are arguably procedural are plainly “bound up” with “state-created rights and obligations” — that is, the State’s underlying scheme of allocating property rights.
Byrd,
356 U.S. at 535, 78 S.Ct. 893. We cannot disregard the procedural aspects of the Act without also destroying the rights and obligations that the Act creates.
We therefore conclude that the question of who shall have a property right in the unclaimed funds is substantive, as that term was set forth in
Erie
and refined in subsequent cases including
Guaranty Oil, Byrd, Hanna,
and
Gasperini.
The district court therefore erred in concluding that it was free to disregard the Unclaimed Property Act and the rights secured thereunder in favor of distributing the funds to the Center via a
cy pres
order. Those funds, insofar as they were allocated to plaintiffs with a last known address in Texas, are governed by Texas law of unclaimed property.
V
Based on our review of the relevant law, we conclude that the unclaimed funds allocated to Texas plaintiffs are subject to the Unclaimed Property Act, that Rule 23(e) is not so broad as to preclude application of the Act, that the question of who has a right to the unclaimed funds is substantive in nature, and that therefore the Act controls. The judgment of the district court is REVERSED and VACATED, and the case REMANDED for proceedings in accordance with this opinion.