Daniel Demer v. IBM Corp Ltd Plan

BYBEE, Circuit Judge,

dissenting from Part II.B, but concurring in the judgment:

An ERISA plan administrator has a structural conflict of interest where it “both funds the plan and evaluates the claims.” Metropolitan Life Ins. Co. v. Glenn, 554 U.S. 105, 112 , 128 S.Ct. 2343 , 171 L.Ed.2d 299 (2008); see also Abatie v. Alta Health & Life Ins. Co., 458 F.3d 955, 965 (9th Cir. 2006) (en banc). The federal courts have offered at least two ways that such conflicts “should prove less important (perhaps to the vanishing point).” Metro. Life, 554 U.S. at 116 , 128 S.Ct. 2343 . First, administrators may “wall[] off claims administrators from those interested in firm finances.” Id.; Abatie, 458 F.3d at 969 & n.7 (an administrator may show that “any conflict did not influence its decisionmak-ing process” by showing that “its employees do not have incentives to deny claims”); Davis v. Unum Life Ins. Co. of Am., 444 F.3d 569, 575-76 (7th Cir. 2006) (holding that absent evidence of “any specific incentive [for the in-house doctors] to derail [a] claim,” such as giving the doctors “some specific stake in the outcome of [a] case,” the theoretical argument that “in-house doctors have an inherent conflict in every case” is insufficient to change the standard of review). Second, plan administrators may refer medical evaluations to outside experts, such as doctors, who also have no interest in firm finances. Abatie, 458 F.3d at 969 & n.7 (“[T]he administrator might demonstrate that it used truly independent medical examiners or a neutral, independent review process”).

MetLife listened very carefully to what we said. It employed both of these methods: First, it walled off its claims administrators from its financial offices. And then, second, the claims office sought medical evaluations from outside, independent physicians who have no interest in MetLife’s finances. For its trouble, the majority is going to give MetLife additional scrutiny — -the majority is “skeptical” of MetLife precisely because it did what we told it to. Maj. Op. at 903-04. The majority’s new skepticism has been rejected by every other circuit to have considered it. When companies structure their operations in response to our opinions and then we penalize them for doing exactly as we have suggested, we sow uncertainty into both law and business. I dissent from Section II.B of the majority’s opinion. Because that section is not otherwise necessary to the majority’s opinion, I concur in the judgment.

I

The problems with the majority’s new concept of skepticism are at least three fold. First, nothing in our decisions supports the majority’s new concept of skepticism. Our relevant cases stand for two propositions: (1) a claims administrator has an internal conflict of interest when it stands to profit from the claims it denies; and, (2) administrators can cleanse this *909 conflict by taking steps such as creating ethical walls, making sure that claim denials are not rewarded, and hiring outside physician reviewers. The majority takes these rules about internal conflicts of interest for which “some skepticism” is warranted, and discovers an external conflict in outside reviewers that is deserving of “modest but ... some, but not substantial” skepticism — something between “no skepticism” and “enhanced skepticism.” Maj. Op. at 903-04. The majority has taken a mechanism we gave administrators for cleansing their conflicts — hiring independent reviewers — and turned it into a new source of conflicts. Second, there is- no basis for inferring a conflict of interest in the outside reviewers simply because they reviewed multiple flies for MetLife and were compensated for their work. Finally, I have great reservations about the use of “skepticism” as a standard of review.

A

In Abatie , we considered the problem with “an insurer that acts as both the plan administrator and the funding source for benefits,” and explained that the insurer thus “operates under what may be termed a structural conflict of interest.” Abatie, 458 F.3d at 965 . We discussed this sort of structural conflict only, and then commented that it could be remedied by, for example, using “independent medical examiners.” Id. at 969 n.7 (emphasis added).

MetLife has an internal conflict of interest in this case. It both pays the benefits and evaluates the claims. No one disputes it. And no one disputes that MetLife has addressed this internal conflict by walling off its financial department from its claims department. See Maj. Op. 901 (assuming that “there is no residual structural conflict ... because of affirmative steps taken by MetLife to insulate its claims department”). But the Majority then finds that an external conflict exists because MetLife refers files to outside physician reviewers.

The majority says that we are to view this relationship with skepticism, but I can find no basis in our decisions for this conclusion. Our cases simply hold that a structural conflict of interest may warrant skepticism, nothing more. See, e.g., Glenn, 554 U.S. at 112 , 128 S.Ct. 2343 (holding that an ERISA conflict emerges where the same administrator “both funds the plan and evaluates the claims,” because “every dollar provided in benefits is a dollar spent by ... the [administrator]; and every dollar saved ... is a dollar in [the administrator’s] pocket”) (quotation omitted); Abatie, 458 F.3d at 965 . The majority claims that in Montour v. Hartford Life & Accident Insurance Co., we “assumed that a relevant piece of extrinsic evidence would be ‘how frequently [the insurance company] contracts with the file reviewers it employed in this case.’ ” 588 F.3d 623, 634 (alteration in Montour). With respect, the majority overstates Montour. In Montour, we found that the Hartford Insurance Company, like MetLife, had a structural conflict of interest. But in Montour we also cited to extensive district court findings that caused us to conclude that “Hartford’s bias infiltrated the entire administrative decisionmaking process, which leads us .to accord significant weight to the conflict.” Id. We then observed that Hartford had failed to show any efforts that it had made “to ‘assure accurate claims assessment^]’ such as utilizing procedures to help ensure a neutral review process.” Id. (quoting Metro. Life, 554 U.S. at 117 , 128 S.Ct. 2343 ). Notice what we said next: “To the contrary, in fact, Hartford’s nurse case manager took an advocacy position in her letters to Montour’s physicians soliciting their agreement with her disability conclusion.” Id. (emphasis added). Not only'had Hartford failed to present any evidence to *910 show its neutrality, the evidence suggested that Hartford was telling Montour’s doctor what it wanted to hear. We then faulted the claimant, Montour — and here is the passage the majority relies on — for “not submitting] any extrinsic evidence of bias, such as statistics regarding Hartford’s rate of claims denials or how frequently it contracts with the file reviewers it employed in this case.” Id. Far from demonstrating that we should be suspicious of outside reviewers, Montour stands for nothing more remarkable than the proposition that it would be relevant to have known in that case where Hartford “took an advocacy position” in its letters to Montour’s doctor what the relationship, if any, was between Hartford’s rate of claims denials and its reliance on outside reviewers. Here, without the predicate for our comment in Montour, the majority simply assumes that the evidence that MetLife used and paid outsider reviewers is sufficient to show bias. 1 The majority calls this a “financial conflict! ],” Maj. Op. at 901, but we have no explanation why an outside reviewer has a conflict of interest by virtue of being compensated for her time. 2

I don’t see the conflict at all. The Supreme Court in Firestone found that “ERISA abounds with the language and terminology of trust law,” and that we should be “guided by principles of trust law.” Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 111, 112 , 109 S.Ct. 948 , 103 L.Ed.2d 80 (1989). The trust analogy is an important one, because a trustee is under a dual obligation: It must ensure that trust funds are paid to those who are eligible; at the same time, it must ensure that trust funds are not paid to those who are not eligible. A trustee-administrator as surely violates its obligation by violating the one charge as the other. MetLife knows that, as trustee, it has an internal conflict of interest and must compensate for that conflict in some way, and perhaps in several ways. Going to outside consulting physicians — literally, getting a second opinion — is one way of compensating for the conflict. But where is the conflict in the outside consulting physicians? Reliability, not bias, is the incentive for the outside reviewer. In the long run, it doesn’t do a company like MetLife any good to get bad medical advice because just as it has a duty to pay deserving claims, it has a duty to avoid paying undeserving claims. 3

I recognize that there could be a cognizable conflict of interest where an indepen *911 dent physician is so dependent on an administrator that it effectively becomes an employee of the administrator, see McDonald v. Hartford Life Grp. Ins. Co., 361 Fed.Appx. 599 , 609 (5th Cir. 2010), but this would simply bring the reviewer within the administrator’s internal umbrella. And in that case, the outside reviewer would be subject to the same conflict of interest as the administrator’s own employees; the outside reviewer cannot be more conflicted than the administrator’s own claim processors (assuming the administrator has not given the reviewer some additional incentive tied to results). See Armstrong v. Aetna Life Ins. Co., 128 F.3d 1263 , 1265 (8th Cir. 1997) (insurer provided incentives and bonuses to claims reviewers for “claims savings”). What is so odd about the majority’s analysis is that if the outside consultants were MetLife’s own internal employees, we would find that they were not conflicted — or at the least that any conflict had been neutralized by MetLife’s claims-handling practices. But for the majority, because the reviewers are independent, suddenly they are untrustworthy. Yet there is not one hint in the record that these outside reviewers are given financial incentives based on their results or in any other way biased.

The principle the majority adopts has profound implications for other areas of the law — notably Social Security claims. The SSA, and its state partners, frequently rely on outside medical sources to review a claimant’s file and offer a second opinion. The views of these reviewing physicians are given significant weight under SSA regulations and our decisions. See 20 C.F.R. § 404 .1519a-q (SSA rules governing the hiring and use of physician reviewers); Reed v. Massanari, 270 F.3d 838 , 842 n.1 (9th Cir. 2001) (noting the “important role played by independent medical specialists” in SSA cases, and overturning the ALJ’s decision not to order an outside physician review); Kish v. Colvin, 552 Fed.Appx. 650, 651 (9th Cir. 2014) (overturning the ALJ’s determination because he failed to order an outside physician review); see also Standards For Consultative Examinations and Existing Medical Evidence, 56 Fed. Reg. 36932 , 36949 (Aug. 1, 1991) (“[The SSA] spend[s] considerable sums annually to obtain consultative examinations.”). Sometimes the SSA uses outside physician reviewers, and sometimes it uses physicians employed full time by the SSA’s state-partner agencies. See Wilson v. Comm’r of Soc. Sec., 280 Fed.Appx. 456, 462 (6th Cir. 2008) (relying in part on evaluation conducted by psychologist employed by state agency); Patty v. Barnhart, 189 Fed.Appx. 517, 519 (7th Cir. 2006) (relying on both state-employed and outside physician reviewers); see also 20 C.F.R. § 404 .1519g (stating the SSA may purchase an outside examination from any qualified provider, including the claimant’s own physicians or “another source”). So far as I know, we have never questioned the bona fides of reviewing physicians’ views on the grounds that SSA or state agencies are sending them lots of business and paying them well, or, even worse, employing them full time. It will turn our cases upside down if we start down that road.

B

Even if we considered Demer’s evidence on the outside reviewers in this case, there *912 is “no there there.” The majority’s new skepticism is based on two facts: First, the majority thinks that MetLife’s outside reviewers are doing a lot of work for Met-Life and, second, the majority thinks the outside reviewers are getting paid a lot of money for their work. Maj. Op. at 902-03 (referring to the “magnitudes of these numbers”). Neither of these reasons will bear scrutiny.

Let’s start with the idea that the two doctors in question here — both of whom were board-certified in internal medicine— are doing a lot of work for MetLife. The record discloses that Dr. Del Valle reviewed some 250 files per year; Dr. Gor-dan, some 200-300 files. We have no basis for judging the significance of these numbers. The majority has just decided that these numbers are big. Let’s indulge the assumption: so what? Even if MetLife decided to stop paying these two doctors as outside consultants and brought them in-house, it wouldn’t implicate a conflict. Met-Life has followed our directions and walled off its claim processors — including any medical personnel — from its financial people. Its claim processors have not been given any incentives to deny claims. That is precisely what we said would cleanse MetLife’s conflicts. Abatie, 458 F.3d at 969 n.7. So whether the doctors are doing a little bit of work for MetLife or a lot of work is irrelevant because there is no smidgen of evidence in the record that MetLife will compensate the doctors based on the results of their evaluations. The majority has no basis whatsoever for impugning the reputations of these medical professionals.

The majority’s second theory- — that the doctors are well compensated — doesn’t hold any more water than the first. Dr. Gordan was compensated $175,000/year for his work; Dr. Del Valle, $125,000. Based on nothing more than its own ipse dixit, the majority declares this to be “significant compensation.” Maj. Op. at 900. We have nothing in the record to tell us how doctors are compensated, whether as practicing physicians or as outside consultants. 4 I could opine that this doesn’t seem like a lot of money for an experienced medical professional, but I have no more basis than the majority for saying so. And perhaps there is some professional jealousy at work here, but if these doctors were lawyers, they aren’t even making first-year associate wages. 5

Would we feel better if MetLife found physicians who were willing to be poorly compensated? Does it get better advice? Does it avoid the conflict? The doctors’ compensation tells us nothing about how the doctors make their medical judgments. And if we are going to indict the medical profession for understanding the potential interests of its clients, do we get to be skeptical of Demer’s personal physician because Demer’s insurance is paying for his opinions? What if the doctor knows that Demer will go get a second opinion (and perhaps leave his personal physician and give him a poor review on Yelp) unless the doctor finds that Demer is disabled? Does Demer’s family doctor have a finan *913 cial incentive to reach a particular medical judgment? What if Demer’s doctor refers him to a second physician for an opinion? Does the second doctor have a conflict of interest because he fears that if he doesn’t corroborate the referring doctor’s opinion he won’t get referrals in the future? There is no end to the gamesmanship that can be played here. •

C

That brings me to the question of “skepticism.” The majority, of course, should not be faulted for relying on our prior decisions and this “skepticism” scheme. We first used the term in Abatie, 458 F.3d at 968 (referring to the “level of skepticism” with which we view a “structural conflict of interest”). In Montour, we picked up on the “skepticism” comment and amplified it, finding that courts should “adjust the level of skepticism” “in accordance with the facts and circumstances of the case” and that where “the conflict may have tainted the entire administrative decisionmaking process,” the courts should view the administrator’s decision with “enhanced skepticism.” 588 F.3d at 631 . We have continued to solidify and elevate Mont-our’s scheme. See Stephan v. Unum Life Ins. Co. of America, 697 F.3d 917, 934 (9th Cir. 2012); Salomaa v. Honda Long Term Disability Plan, 642 F.3d 666, 675 (9th Cir. 2011); Nolan, 551 F.3d at 1155. Even as we acknowledge that it is not “easy to decide how many metaphorical grams should go on the metaphorical scale,” we have added additional qualifiers to “skepticism” to give us a range of skepticism, including “special skepticism,” “additional skepticism,” and a “higher degree of skepticism.” Salomaa, 642 F.3d at 675-76 .

I question the usefulness of “skepticism” as a standard of review. Standards of review are, necessarily, not “rules” and are subject to the vagaries of language. In Justice Frankfurter’s memorable phrase, standards of review capture a “mood,” but a “mood [that] must be respected, even though it can only serve as a standard for judgment and not as a body of rigid rules assuring sameness of applications.” Universal Camera Corp. v. NLRB, 340 U.S. 474, 487 , 71 S.Ct. 456 , 95 L.Ed. 456 (1951). We are burdened with a variety of “word formulas,” FTC v. Sun Oil Co., 371 U.S. 505, 527 , 83 S.Ct. 358 , 9 L.Ed.2d 466 (1963), to describe amorphous concepts such as evidence, error, and discretion. And we (and Congress) have developed lists of adjectives to qualify those concepts. Thus, we have at least six different kinds of “evidence”: a “scintilla of evidence,” “some evidence,” “substantial evidence,” a “preponderance of evidence,” “clear and convincing evidence,” and “evidence beyond a reasonable doubt.” It isn’t enough to commit an “error,” we must know whether it was “harmless error,” “clear error,” “plain error,” or “invited error.” When a court or agency exercises discretion, we need to know whether it was “an abuse of discretion” or merely “arbitrary and capricious.” And I won’t even go down the road of constitutional “scrutiny.”

The problem with “skepticism” is not that it is not descriptive in some useful way, but that we don’t need another noun modified by a raft of adjectives to capture our mood when we review decisions under ERISA. We know that the standard of review for ERISA plans in which some discretion is conferred on the plan trustee is the well-established “abuse of discretion.” Firestone, 489 U.S. at 115 , 109 S.Ct. 948 ; Abatie, 458 F.3d at 962-63 . That standard is hard enough to apply without adding a layer of “skepticism” on top of it. We are trained in the law, and we are, either by nature or by avocation, skeptical. It’s what we do. We are skeptical of what defendants, witnesses, and even lawyers *914 tell us. We question their memories, reliability, and motives. And we have developed our own ranges of skepticism, from ordinary eyebrow-raising skepticism to the more ambiguous harrumphing skepticism to full-blown, exasperated snorting skepticism. In terms of expressing a mood, we are much better off with the facial expressions and the occasional snort than with trying to define the difference between “some skepticism,” “special skepticism,” “enhanced skepticism,” and “high-degree skepticism.” I would abandon “skepticism” as a separate standard of review in ERISA cases and try to deal with run-of-the-mill “abuse of discretion.”

II

These are problems of our own making. Every circuit that has considered the use that MetLife made of outside reviewers here has approved the practice, and they have managed to do so without invoking the standard of “skepticism.” The courts not only don’t apply even a modicum of skepticism, they often reject the claim that there is any conflict in the first place. For example, in Davis v. Unum Life Insurance Company of America, the Seventh Circuit noted the absurdity of turning “an administrator’s decision to seek independent expert advice [which] is evidence of a thorough investigation” into grounds for criticizing the administrator. 444 F.3d 569, 575-76 (7th Cir. 2006) (quotation omitted). The court explained that “[playing for a legitimate and valuable service in order to evaluate a claim thoroughly does not create a review-altering conflict.” Id. at 575 . This was true even where the reviewers’ entire salary came from a single administrator. 6 The Fifth, Sixth, and Tenth Circuits have adopted similar rules. See Hagen v. Aetna Ins. Co., 808 F.3d 1022, 1029 (5th Cir. 2015) (holding that the relationship between the administrator and reviewer was not enough, on its own, to show bias; there must be some evidence of specific stake in the claim); Hunt v. Metro. Life Ins. Co., 587 Fed.Appx. 860, 862 (6th Cir. 2014) (requiring more than mere evidence of payments between the administrator and reviewer); Benson v. Hartford Life &. Acc. Ins. Co., 511 Fed.Appx. 680, 685 (10th Cir. 2013) (rejecting claim of reviewer conflict because “[n]othing in the record suggested] a specific bias on the part of the ... reviewing physicians.”); Christoff v. Ohio N. Univ. Employee Ben. Plan, 495 Fed.Appx. 710, 712 (6th Cir.. 2012) (declining to find conflict of interest because the administrator used “independent reviewers” to deny claims).

Where the plan administrator is operating under a structural conflict of interest, courts have often concluded that the conflict is mitigated by the administrator’s decision to “reduce its inherent bias by hiring ... independent physicians.” Holcomb v. Unum Life Ins. Co. of Am., 578 F.3d 1187, 1193 (10th Cir. 2009); see also Fite v. Bayer Corp., 554 Fed.Appx. 712, 717 (10th Cir. 2014) (“Bayer took active steps to reduce any potential bias and to promote accuracy: it sought an independent review of Ms. Fite’s medical records by a different psychiatrist ... and it obtained an independent psychiatric evaluation of Ms. Fite from a fourth psychiatrist *915 before reaching its final decision. We therefore give the conflict-of-interest factor limited weight in determining whether Bayer abused its discretion.”); Menge v. AT & T, Inc., 595 Fed.Appx. 811, 814 (10th Cir. 2014) (holding that an administrator had mitigated its “inherent conflict of interest” by “rel[ying] on the medical opinions of independent physician advisors”); Keith v. Prudential Ins. Co. of Am., 347 Fed.Appx. 548, 552 (11th Cir. 2009) (holding that an administrator’s use of three independent physician reviewers showed the administrator was not “influenced by [its] [structural] conflict”). Indeed, it is more likely that courts will hold it against an administrator who doesn’t use independent reviewers. As the First Circuit explained, it is “difficult to fault a plan administrator for seeking expert assistance (indeed, it probably would be easier to fault a plan administrator for not seeking such assistance).... [C]ommon sense dictates that retaining outside physicians to assist in evaluating disability claims, without more, does not constitute a conflict of interest.” Leahy v. Raytheon Co., 315 F.3d 11, 16 (1st Cir. 2002); see also Loan v. Prudential Ins. Co. of Am., 370 Fed.Appx. 592, 598 (6th Cir. 2010) (“Prudential did not seek the opinion of an outside expert who was not operating under Prudential’s conflict-of-interest.... Prudential’s, in-house doctor had less incentive than an independent outside doctor to conduct a thorough and accurate review of the record and address the arguments Plaintiffs raised on appeal.” (emphasis added)).

Ill

The majority takes the remedies we offered to administrators for cleansing conflicts — walling off its claim processors and hiring independent reviewers — and turns it into a sword to punish administrators with skepticism. After today’s decision, we cannot fault administrators for their confusion over what they can rely on in our decisions. And we can predict with near certainty how they will respond: at least in our circuit, administrators will stop using outside, independent reviewers; instead, they will try to bring them in house where, they hope, we will still respect the administrator’s efforts to wall them off. Today’s decision injects confusion and change for no reason. We are not likely to end up with better decisions in ERISA claims.

Count me skeptical. I respectfully dissent from Section II.B.

. Montour cited to Nolan v. Heald College, 551 F.3d 1148 (9th Cir. 2009), where, in a footnote, we suggested that there might be some relationship between an outside reviewer’s regular contract with the insurance company and the reviewer’s bias. Id. at 1152 n.3. We offered no explanation why this was a problem. We later commented that “This is not to say that [the insurance company] was not entitled to rely on the opinions of its independent physicians,” but we assumed that it was evidence of bias that the district court might have viewed “with skepticism.” Id. at 1155 .

. The majority contends that I have "mistakenly equate[d] outside experts with independent experts.'” Maj. Op. at 903. Not so. I have simply equated better-paid outside experts with slightly-lesser-paid outside experts. The question in this case is not whether there is any circumstance in which an outside expert could lose her independence — of course there is'. The question here is whether paying a board-certified physician an amount that strikes the majority as a lot, on its own, makes that physician no longer "independent.” And to me, the answer must be no.

.It is far from clear from the cases (including the record in this case) what files we think the outside physicians are reviewing. Are the reviewers looking at all files, including files the claims department is inclined to pay? If so, why would the reviewing physicians offer anything other than their best view?

But what if the claims department is only sending outside reviewers the files of claims it is inclined to deny? In that case the outside reviewers still have little incentive to shade their views in favor of the claims department. *911 This is not advocacy, it is medical opinion, and the claims department, after all, is asking the reviewers for a second opinion before denying the claim.

If there were evidence an administrator was signaling to its outside reviewers what result it wanted them to reach, that would be a serious problem. In such a case, the signaling would be direct evidence that the administrator had not addressed its structural conflict of interest.

. The majority's approach to the burdens here makes little sense. If the claimant bears the burden, as Montour suggests, 588 F.3d. at 634 , then gaps in the record — as we have here-— should weigh against Demer. The majority fails to explain how pointing to two numbers (the number of files received and the sum of payments received) establishes that the outside reviewers were biased.

. See David Lat, Breaking: NY to $180KH! Cravath Raises Associate Base Salaries!!!, June 6, 2016, http://abovethelaw.corn/2016/06/ breaking-ny-to-180k-cravath-raises-associate-base-salaries/.

. The majority suggests that its approach to analyzing outside experts is unremarkable, and points out that other circuits do not hold "that outside experts are immune from judicial scrutiny for possible bias.” Maj. Op. at 904. We can (and should) scrutinize outside experts when there are facts that demonstrate that their judgment is not independent. But that is not where I and our sister circuits depart from the majority. Our disagreement is with the majority's contention that the bare fact of compensating an outside expert for her opinion makes her biased.