[108]*108Justice Breyer
delivered the opinion of the Court.
The Employee Retirement Income Security Act of 1974 (ERISA) permits a person denied benefits under an employee benefit plan to challenge that denial in federal court. 88 Stat. 829, as amended, 29 U. S. C. § 1001 et seq.; see § 1132(a)(1)(B). Often the entity that administers the plan, such as an employer or an insurance company, both determines whether an employee is eligible for benefits and pays benefits out of its own pocket. We here decide that this dual role creates a conflict of interest; that a reviewing court should consider that conflict as a factor in determining whether the plan administrator has abused its discretion in denying benefits; and that the significance of the factor will depend upon the circumstances of the particular case. See Firestone Tire & Rubber Co. v. Bruch, 489 U. S. 101, 115 (1989).
I
Petitioner Metropolitan Life Insurance Company (Met-Life) serves as both an administrator and the insurer of Sears, Roebuck & Company’s long-term disability insurance plan, an ERISA-governed employee benefit plan. See App. 182a-183a; 29 U. S. C. § 1003. The plan grants MetLife (as administrator) discretionary authority to determine whether an employee’s claim for benefits is valid; it simultaneously provides that MetLife (as insurer) will itself pay valid benefit claims. App. 181a-182a.
[109]*109Respondent Wanda Glenn, a Sears employee, was diagnosed with severe dilated cardiomyopathy, a heart condition whose symptoms include fatigue and shortness of breath. She applied for plan disability benefits in June 2000, and Met-Life concluded that she met the plan’s standard for an initial 24 months of benefits, namely, that she could not “perform the material duties of [her] own job.” Id., at 159a-160a. MetLife also directed Glenn to a law firm that would assist her in applying for federal Social Security disability benefits (some of which MetLife itself would be entitled to receive as an offset to the more generous plan benefits). In April 2002, an Administrative Law Judge found that Glenn’s illness prevented her not only from performing her own job but also “from performing any jobs [for which she could qualify] existing in significant numbers in the national economy.” App. to Pet. for Cert. 49a; see also 20 CFR § 404.1520(g) (2007). The Social Security Administration consequently granted Glenn permanent disability payments retroactive to April 2000. Glenn herself kept none of the backdated benefits: Three-quarters went to MetLife, and the rest (plus some additional money) went to the lawyers.
To continue receiving Sears plan disability benefits after 24 months, Glenn had to meet a stricter, Social-Security-type standard, namely, that her medical condition rendered her incapable of performing not only her own job but of performing “the material duties of any gainful occupation for which” she was “reasonably qualified.” App. 160a. MetLife denied Glenn this extended benefit because it found that she was “capable of performing full time sedentary work.” Id., at 31a.
After exhausting her administrative remedies, Glenn brought this federal lawsuit, seeking judicial review of Met-Life’s denial of benefits. See 29 U. S. C. § 1132(a)(1)(B); 461 F. 3d 660, 665 (CA6 2006). The District Court denied relief. Glenn appealed to the Court of Appeals for the Sixth Circuit. Because the plan granted MetLife “discretionary authority [110]*110to . . . determine benefits,” the Court of Appeals reviewed the administrative record under a deferential standard. Id., at 666. In doing so, it treated “as a relevant factor” a “conflict of interest” arising out of the fact that MetLife was “authorized both to decide whether an employee is eligible for benefits and to pay those benefits.” Ibid.
The Court of Appeals ultimately set aside MetLife’s denial of benefits in light of a combination of several circumstances: (1) the conflict of interest; (2) MetLife’s failure to reconcile its own conclusion that Glenn could work in other jobs with the Social Security Administration’s conclusion that she could not; (3) MetLife’s focus upon one treating physician report suggesting that Glenn could work in other jobs at the expense of other, more detailed treating physician reports indicating that she could not; (4) MetLife’s failure to provide all of the treating physician reports to its own hired experts; and (5) MetLife’s failure to take account of evidence indicating that stress aggravated Glenn’s condition. See id., at 674.
MetLife sought certiorari, asking us to determine whether a plan administrator that both evaluates and pays claims operates under a conflict of interest in making discretionary benefit determinations. The Solicitor General suggested that we also consider “ ‘how’ ” any such conflict should “ ‘be taken into account on judicial review of a discretionary benefit determination.’ ” Brief for United States as Amicus Curiae on Pet. for Cert. 22. We agreed to consider both questions. See 552 U. S. 1161 (2008).
II
In Firestone Tire & Rubber Co. v. Bruch, 489 U. S. 101, this Court addressed “the appropriate standard of judicial review of benefit determinations by fiduciaries or plan administrators under” § 1132(a)(1)(B), the ERISA provision at issue here. Id., at 105; see also id., at 108. Firestone set forth four principles of review relevant here.
[111]*111(1) In “determining the appropriate standard of review,” a court should be “guided by principles of trust law”; in doing so, it should analogize a plan administrator to the trustee of a common-law trust; and it should consider a benefit determination to be a fiduciary act (i. e., an act in which the administrator owes a special duty of loyalty to the plan beneficiaries). Id., at 111-113. See also Aetna Health Inc. v. Davila, 542 U. S. 200, 218 (2004); Central States, Southeast & Southwest Areas Pension Fund v. Central Transport, Inc., 472 U. S. 559, 570 (1985).
(2) Principles of trust law require courts to review a denial of plan benefits “under a de novo standard” unless the plan provides to the contrary. Firestone, 489 U. S., at 115; see also id., at 112 (citing, inter alia, 3 A. Scott & W. Fratcher, Law of Trusts § 201, p. 221 (4th ed. 1988); G. Bogert & G. Bogert, Law of Trusts and Trustees § 559, pp. 162-168 (rev. 2d ed. 1980) (hereinafter Bogert); 1 Restatement (Second) of Trusts § 201, Comment 6 (1957) (hereinafter Restatement)).
(3) Where the plan provides to the contrary by granting “the administrator or fiduciary discretionary authority to determine eligibility for benefits,” Firestone, 489 U. S., at 115 (emphasis added), “[t]rust principles make a deferential standard of review appropriate,” id., at 111 (citing Restatement § 187 (abuse-of-discretion standard); Bogert § 560, at 193-208; emphasis added).
(4) If “a benefit plan gives discretion to an administrator or fiduciary who is operating under a conflict of interest, that conflict must be weighed as a factor in determining whether there is an abuse of discretion.’ ” Firestone, supra, at 115 (quoting Restatement § 187, Comment d; emphasis added; alteration omitted).
The questions before us, while implicating the first three principles, directly focus upon the application and the meaning of the fourth.
[112]*112III
The first question asks whether the fact that a plan administrator both evaluates claims for benefits and pays benefits claims creates the kind of “conflict of interest” to which Firestone’s fourth principle refers. In our view, it does.
That answer is clear where it is the employer that both funds the plan and evaluates the claims. In such a circumstance, “every dollar provided in benefits is a dollar spent by . . . the employer; and every dollar saved ... is a dollar in [the employer’s] pocket.” Bruch v. Firestone Tire & Rubber Co., 828 F. 2d 134, 144 (CA3 1987). The employer’s fiduciary interest may counsel in favor of granting a borderline claim while its immediate financial interest counsels to the contrary. Thus, the employer has an “interest . . . conflicting with that of the beneficiaries,” the type of conflict that judges must take into account when they review the discretionary acts of a trustee of a common-law trust. Restatement § 187, Comment d; see also Firestone, supra, at 115 (citing that Restatement comment); cf. Black’s Law Dictionary 319 (8th ed. 2004) (“[C]onflict of interest” is a “real or seeming incompatibility between one’s private interests and one’s public or fiduciary duties”).
Indeed, Firestone itself involved an employer who administered an ERISA benefit plan and who both evaluated claims and paid for benefits. See 489 U. S., at 105. And thus that circumstance quite possibly was what the Court had in mind when it mentioned conflicted administrators. See id., at 115. The Firestone parties, while disagreeing about other matters, agreed that the dual role created a conflict of interest of some kind in the employer. See Brief for Petitioners 6-7, 27-29, Brief for Respondent 9, 26, and Brief for United States as Amicus Curiae in Firestone Tire & Rubber Co. v. Bruch, O. T. 1988, No. 87-1054, p. 22.
MetLife points out that an employer who creates a plan that it will both fund and administer foresees, and implic[113]*113itly approves, the resulting conflict. But that fact cannot change our conclusion. At trust law, the fact that a settlor (the person establishing the trust) approves a trustee’s conflict does not change the legal need for a judge later to take account of that conflict in reviewing the trustee’s discretionary decisionmaking. See Restatement §107, Comment / (discretionary acts of trustee with settlor-approved conflict subject to “careful scrutiny”); id., § 107, Comment/, Illustration 1 (conflict is “a factor to be considered by the court in determining later whether” there has been an “abuse of discretion”); id., § 187, Comment d (same); 3 A. Scott, W. Fratcher, & M. Ascher, Scott and Ascher on Trusts § 18.2, pp. 1342-1343 (5th ed. 2007) (hereinafter Scott) (same). See also, e. g., Bogert § 543, at 264 (rev. 2d ed. 1993) (settlor approval simply permits conflicted individual to act as a trustee); id., § 543(U), at 422-431 (same); Scott § 17.2.11, at 1136-1139 (same).
MetLife also points out that we need not follow trust law principles where trust law is “inconsistent with the language of the statute, its structure, or its purposes.” Hughes Aircraft Co. v. Jacobson, 525 U. S. 432, 447 (1999) (internal quotation marks omitted). MetLife adds that to find a conflict here is inconsistent (1) with ERISA’s efforts to avoid complex review proceedings, see Varity Corp. v. Howe, 516 U. S. 489, 497 (1996); (2) with Congress’ efforts not to deter employers from setting up benefit plans, see ibid.; and (3) with an ERISA provision specifically allowing employers to administer their own plans, see 29 U. S. C. § 1108(c)(3).
But we cannot find in these considerations any significant inconsistency. As to the first, we note that trust law functions well with a similar standard. As to the second, we have no reason, empirical or otherwise, to believe that our decision will seriously discourage the creation of benefit plans. As to the third, we have just explained why approval of a conflicted trustee differs from review of that trustee’s [114]*114conflicted decisionmaking. As to all three taken together, we believe them outweighed by “Congress’ desire to offer employees enhanced protection for their benefits.” Varity, supra, at 497 (discussing “competing congressional purposes” in enacting ERISA).
The answer to the conflict question is less clear where (as here) the plan administrator is not the employer itself but rather a professional insurance company. Such a company, MetLife would argue, likely has a much greater incentive than a self-insuring employer to provide accurate claims processing. That is because the insurance company typically charges a fee that attempts to account for the cost of claims payouts, with the result that paying an individual claim does not come to the same extent from the company’s own pocket. It is also because the marketplace (and regulators) may well punish an insurance company when its products, or ingredients of its products, fall below par. And claims processing, an ingredient of the insurance company’s product, falls below par when it seeks a biased result, rather than an accurate one. Why, MetLife might ask, should one consider an insurance company inherently more conflicted than any other market participant, say, a manufacturer who might earn more money in the short run by producing a product with poor quality steel or a lawyer with an incentive to work more slowly than necessary, thereby accumulating more billable hours?
Conceding these differences, we nonetheless continue to believe that for ERISA purposes a conflict exists. For one thing, the employer’s own conflict may extend to its selection of an insurance company to administer its plan. An employer choosing an administrator in effect buys insurance for others and consequently (when compared to the marketplace customer who buys for himself) may be more interested in an insurance company with low rates than in one with accurate claims processing. Cf. Langbein, Trust Law as Regulatory Law, 101 Nw. U. L. Rev. 1315, 1323-1324 (2007) (observing that employees are rarely involved in plan negotiations).
[115]*115For another, ERISA imposes higher-than-marketplace quality standards on insurers. It sets forth a special standard of care upon a plan administrator, namely, that the administrator “discharge [its] duties” in respect to discretionary claims processing “solely in the interests of the participants and beneficiaries” of the plan, § 1104(a)(1); it simultaneously underscores the particular importance of accurate claims processing by insisting that administrators “provide a ‘full and fair review’ of claim denials,” Firestone, 489 U. S., at 113 (quoting § 1133(2)); and it supplements marketplace and regulatory controls with judicial review of individual claim denials, see § 1132(a)(1)(B).
Finally, a legal rule that treats insurance company administrators and employers alike in respect to the existence of a conflict can nonetheless take account of the circumstances to which MetLife points so far as it treats those, or similar, circumstances as diminishing the significance or severity of the conflict in individual cases. See Part IV, infra.
IV
We turn to the question of “how” the conflict we have just identified should “be taken into account on judicial review of a discretionary benefit determination.” 552 U. S. 1161. In doing so, we elucidate what this Court set forth in Firestone, namely, that a conflict should “be weighed as a ‘factor in determining whether there is an abuse of discretion.’ ” 489 U. S., at 115 (quoting Restatement § 187, Comment d; alteration omitted).
We do not believe that Firestone’s statement implies a change in the standard of review, say, from deferential to de novo review. Trust law continues to apply a deferential standard of review to the discretionary decisionmaking of a conflicted trustee, while at the same time requiring the reviewing judge to take account of the conflict when determining whether the trustee, substantively or procedurally, has abused his discretion. See Restatement § 187, Comments d-j; id., § 107, Comment f; Scott § 18.2, at 1342-1344. [116]*116We see no reason to forsake Firestone’s reliance upon trust law in this respect. See 489 U. S., at 111-115.
Nor would we overturn Firestone by adopting a rule that in practice could bring about near universal review by judges de novo — i. e., without deference — of the lion’s share of ERISA plan claims denials. See Brief for America’s Health Insurance Plans et al. as Amici Curiae 3-4 (many ERISA plans grant discretionary authority to administrators that combine evaluation and payment functions). Had Congress intended such a system of review, we believe it would not have left to the courts the development of review standards but would have said more on the subject. See Firestone, supra, at 109 (“ERISA does not set out the appropriate standard of review for actions under § 1132(a)(1)(B)”); compare, e. g., C. Gresenz et al., A Flood of Litigation? 8 (1999), http://www.rand.org/pubs/issue_papers/2006/IP184.pdf (all Internet materials as visited June 9, 2008, and available in Clerk of Court’s case file) (estimating that 1.9 million beneficiaries of ERISA plans have health care claims denied each year), with Caseload of Federal Courts Remains Steady Overall (Mar. 11, 2008), http://www.uscourts.gov/Press_ Releases/2008/caseload.cfm (257,507 total civil filings in federal court in 2007); cf. Whitman v. American Trucking Assns., Inc., 531 U. S. 457, 468 (2001) (Congress does not “hide elephants in mouseholes”).
Neither do we believe it necessary or desirable for courts to create special burden-of-proof rules, or other special procedural or evidentiary rules, focused narrowly upon the evaluator/payor conflict. In principle, as we have said, conflicts are but one factor among many that a reviewing judge must take into account. Benefits decisions arise in too many contexts, concern too many circumstances, and can relate in too many different ways to conflicts — which themselves vary in kind and in degree of seriousness — for us to come up with a one-size-fits-all procedural system that is likely to promote fair and accurate review. Indeed, special procedural rules [117]*117would create further complexity, adding time and expense to a process that may already be too costly for many of those who seek redress.
We believe that Firestone means what the word “factor” implies, namely, that when judges review the lawfulness of benefit denials, they will often take account of several different considerations of which a conflict of interest is one. This kind of review is no stranger to the judicial system. Not only trust law, but also administrative law, can ask judges to determine lawfulness by taking account of several different, often case-specific, factors, reaching a result by weighing all together. See Restatement § 187, Comment d; cf., e. g., Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U. S. 402, 415-417 (1971) (review of governmental decision for abuse of discretion); Universal Camera Corp. v. NLRB, 340 U. S. 474 (1951) (review of agency factfinding).
In such instances, any one factor will act as a tiebreaker when the other factors are closely balanced, the degree of closeness necessary depending upon the tiebreaking factor’s inherent or case-specific importance. The conflict of interest at issue here, for example, should prove more important (perhaps of great importance) where circumstances suggest a higher likelihood that it affected the benefits decision, including, but not limited to, eases where an insurance company administrator has a history of biased claims administration. See Langbein, supra, at 1317-1321 (detailing such a history for one large insurer). It should prove less important (perhaps to the vanishing point) where the administrator has taken active steps to reduce potential bias and to promote accuracy, for example, by walling off claims administrators from those interested in firm finances, or by imposing management checks that penalize inaccurate decisionmaking irrespective of whom the inaccuracy benefits. See Herzel & Colling, The Chinese Wall and Conflict of Interest in Banks, 34 Bus. Law 73, 114 (1978) (recommending interdepartmental information walls to reduce bank conflicts); Brief for Blue [118]*118Cross and Blue Shield Association as Amicus Curiae 15 (suggesting that insurers have incentives to reward claims processors for their accuracy); cf. generally J. Mashaw, Bureaucratic Justice (1983) (discussing internal controls as a sound method of producing administrative accuracy).
The Court of Appeals’ opinion in the present case illustrates the combination-of-factors method of review. The record says little about MetLife’s efforts to ensure accurate claims assessment. The Court of Appeals gave the conflict weight to some degree; its opinion suggests that, in context, the court would not have found the conflict alone determinative. See 461 F. 3d, at 666, 674. The eourt instead focused more heavily on other factors. In particular, the court found questionable the fact that MetLife had encouraged Glenn to argue to the Social Security Administration that she could do no work, received the bulk of the benefits of her success in doing so (the remainder going to the lawyers it recommended), and then ignored the agency’s finding in concluding that Glenn could in fact do sedentary work. See id., at 666-669. This course of events was not only an important factor in its own right (because it suggested procedural unreasonableness), but also would have justified the court in giving more weight to the conflict (because MetLife’s seemingly inconsistent positions were both financially advantageous). And the court furthermore observed that MetLife had emphasized a certain medical report that favored a denial of benefits, had deemphasized certain other reports that suggested a contrary conclusion, and had failed to provide its independent vocational and medical experts with all of the relevant evidence. See id., at 669-674. All these serious concerns, taken together with some degree of conflicting interests on MetLife’s part, led the court to set aside MetLife’s discretionary decision. See id., at 674-675. We can find nothing improper in the way in which the court conducted its review.
[119]*119Finally, we note that our elucidation of Firestone’s standard does not consist of a detailed set of instructions. In this respect, we find pertinent this Court’s comments made in a somewhat different context, the context of court review of agency factfinding. See Universal Camera Corp., supra. In explaining how a reviewing court should take account of the agency’s reversal of its own examiner’s factual findings, this Court did not lay down a detailed set of instructions. It simply held that the reviewing judge should take account of that circumstance as a factor in determining the ultimate adequacy of the record’s support for the agency’s own factual conclusion. Id., at 492-497. In so holding, the Court noted that it had not enunciated a precise standard. See, e. g., id., at 493. But it warned against creating formulas that will “falsif[y] the actual process of judging” or serve as “instrument[s] of futile casuistry.” Id., at 489. The Court added that there “are no talismanic words that can avoid the process of judgment.” Ibid. It concluded then, as we do now, that the “[w]ant of certainty” in judicial standards “partly reflects the intractability of any formula to furnish definiteness of content for all the impalpable factors involved in judicial review.” Id., at 477.
We affirm the judgment of the Court of Appeals.
It is so ordered.