dissenting.
I am obliged to dissent from the majority’s judgment in favor of Dr. Lasser.
(1) The District Court failed to adhere to our established standards of review and our precedents. See Lasser v. Reliance Standard Life Ins. Co., 130 E.Supp.2d 616, 630 (D.N.J.2001) (“Lasser I”) (“th[is] Court anticipates adopting the procedure set forth in Kearney [v. Standard Ins. Co., 176 F.3d 1084 (9th Circuit 1999) (en banc) ] and making plenary findings of fact pursuant to Federal Rule of Civil Procedure 52, even though no new evidence may be received on that branch of the case.”)1 By so doing, the District Court erred by giving only lip service, but no required true deference, to the Administrator Walsh’s determinations even after it had held that the Administrator was neither conflicted nor biased. See Pinto v. Reliance Standard Life Ins. Co., 214 F.3d 377 (3d Cir.2000).
(2) This is a summary judgment case. The District Court, when called upon to decide summary judgment, cannot hold a trial on the merits a la Kearney. See infra note 3. In this case, the District Court did just that.
(3) The District Court, at its truncated April 10, 2001 hearing, while purporting to resolve only the Administrator’s status as to whether he was conflicted or biased (it held that he was neither), cannot proceed to ignore our prescribed standard of review which requires deference and then proceed to substitute its own judgment on the merits.
*393(4) The policy written by Reliance requires that for disability benefits, Dr. Las-ser had to be disabled from performing the material duties of his regular occupation. The burden of proving materiality rested on Dr. Lasser. No evidence of materiality appears in the record.
(5) No deference was shown by the District Court to Mr. Walsh’s, the Administrator’s, determinations that: (a) materiality was not demonstrated; (b) Dr. Lasser returned to, and was performing, his full-time occupation; (c) “emergency” and “on-call” services were not established as material; yet, these were the only services Dr. Lasser was told not to perform; and (d) neither stress nor future disabilities could establish Dr. Lasser’s 'present disability.
(6) The actions taken by Reliance and the evidence relied upon by Reliance, as well as the determinations made by Walsh, the Administrator, cannot be determined to be arbitrary nor capricious under our standards and precedents.
I.
This appeal comes to us after both parties had moved for summary judgment. It has remained a summary judgment case in which the District Court could look only to the developed administrative record.
I begin by briefly outlining the pertinent procedural history. In its February 8, 2001 opinion, the District Court denied the parties’ cross-motions for summary judgment. It then stated that it would hold a hearing to determine if Walsh, the Administrator, was biased or suffered a conflict because he was an employee of Reliance. Lasser I, 180 F.Supp.2d at 625-26, 680 (referring to Pinto’s, sliding scale of arbitrary and capricious review).2
The District Court then conducted a bench trial on April 10, 2001 where it heard only Walsh’s testimony relevant to the issue of how heightened our standard of review should be on Pinto’s sliding scale. Dr. Lasser did not testify, nor did any other witness. I reiterate-the only purpose of the hearing was to address whether Walsh was conflicted or biased because he was Reliance’s employee. Walsh had the duty of determining a claimant’s eligibility for benefits as well as of paying those benefits out of the funds from which Reliance reaps its profits. If biased, then under Pinto the deference to be accorded his determination would be less deferential, and the court must “calib-rare] the intensity of [the] review to the intensity of the conflict.” Pinto, 214 F.3d at 393. If not biased, then under Pinto, the standard of deference to be employed would be “deferential, but not absolutely deferential.” Id. But, in either event, deference had to be accorded.
Thereafter, the District Court determined that it would give a “moderate degree of deference” to Walsh’s determinations as it found no evidence of bias or conflict exhibited by Walsh. Lasser II, 146 F.Supp.2d at 623.
At this stage, the District Court was required to end its inquiries and hearing and review the existing administrative record-and no more. But, the District Court did not do that. It proceeded further, despite the summary judgment context of the instant case, and it made its own findings of fact on the merits. Id. at 620, 642.
Under our • precedents, district courts cannot conduct so-called “merits-trials” *394when determining the appropriate conflict-of-interest review standard.3 As I have stated, the District Court should have limited its findings to a determination of what the appropriate standard of review would be, and then should have continued on to resolve the summary judgment question pursuant to that standard.
Accordingly, I would have returned this case to the District Court with the direction that it expunge all its merits findings and discussion and then proceed to perform its prescribed function of ruling, on the administrative record only, whether Walsh’s denial of benefits was arbitrary or capricious.
II.
Although the District Court purported to subscribe to the correct standard of review, a fair reading of the District Court’s opinions indicates that it did not; it merely gave lipservice to the deference to be accorded to the Administrator’s determinations.
Pursuant to the plain language of Reliance’s policy, whether Dr. Lasser is “totally disabled” hinges on whether he can perform the “material” duties of his occupation. The relevant part of the policy provides:
“Totally Disabled” and “Total Disability” mean ... that as a result of an Injury or Sickness, during the Elimination Period and thereafter an Insured cannot perform the material duties of his/her regular occupation....
App. at 139 (emphasis added). Critical to this analysis, is that the burden to prove materiality is on Dr. Lasser because the burden of proving coverage is always on the claimant. See Lasser II, 146 F.Supp.2d at 639 (citing Pinto, 214 F.3d at 394 n. 8). Dr. Lasser, however, submitted no evidence, no proofs, no affidavits, no witnesses, and no depositions on the issue of materiality. Indeed, the record as to this issue is barren.4
Despite the silence of the record regarding materiality, the District Court nevertheless held that Dr. Lasser had made out his burden of proof on materiality. To do só, the District Court relied principally on the diminution of Dr. Lasser’s income when he discontinued emergency and on-call duties. Id. at 639-40. The majority, too, finds this diminution in income to be controlling. Yet, it must be remembered *395that during the period of time (from September 1996 to December 1997) that Dr. Lasser was “totally disabled,” he was attending to all his office functions and tasks and he was performing scheduled (elective) surgery. At the same time, he was collecting total disability benefits from Reliance.
The District Court also relied on the various medical reports from cardiologists, although they merely stated that Dr. Las-ser should not perform “emergency” and “on-call” work-not that those functions were “material” to his occupation of orthopedic surgeon. Id.
Reliance, on the other hand, presented evidence on the issue of materiality. Reliance looked to the duties of a general surgeon as defined in the Dictionary of Occupational Titles (“DOT”) from the United States Department of Labor. “Emergency” and/or “on-call” work are not included in material duties. Dr. Las-ser has presented no evidence to the contrary.
Furthermore, the DOT is a recognized authority in the vocational industry. See, e.g., Gallagher v. Reliance Standard Life Ins. Co., 305 F.3d 264, 272-73 (4th Cir.2002) (DOT is an objectively reasonable tool for determining the material duties of an occupation; a “general job description of DOT, to be applicable, must involve comparable duties but not necessarily every duty.”) (emphasis added).
Similarly, the District Court gave no deference whatsoever to Reliance in its use of a labor market survey. This study provided direct evidence that “on-call” and “emergency” duties-the only restrictions which Dr. Lasser’s physicians proscribed-were not material. Dr. Lasser provided no vocational studies nor, as I stated earlier, any other evidence which contradicted Reliance’s evidence on materiality.
Moreover, the use of such vocational studies in eligibility for disability benefits is an accepted practice in this Circuit. See, e.g., Russell v. Paul Revere Life Ins. Co., 288 F.3d 78, 79, 82 (3d Cir.2002) (labor market study which identified jobs within the claimant’s occupation that did not require travel, contravened plaintiffs contention that he remained disabled because he could not work extended hours and travel, which his current position required).
All of this evidence stands unrebutted by Dr. Lasser. Again, the appropriate standard of review is whether or not Reliance was arbitrary and capricious in determining that Dr. Lasser failed to meet his burden of materiality, that is: that Dr. Lasser could not perform the material duties of his regular occupation. Nowhere in this record is there evidence that “emergency” and “on-call” tasks-the only work which Dr. Lasser was told he could not perform-are material to Dr. Lasser’s every day occupation. Although I am sympathetic to Dr. Lasser’s financial and medical difficulties-as I am sure the members of the majority are too-such sympathy cannot substitute as evidence that “emergency” and “on-call” work are material to the occupation of an orthopedic surgeon.
Furthermore, I find it troubling that Dr. Lasser presented no vocational study, nor any evidence from experts in the field, or even evidence from the members of his own office or himself, to support his contention that these tasks were in fact material. In particular, there has been no evidence as to whether his three partners could handle and would respond to the practice’s on-call and emergency duties, although such circumstances could lead to a diminution of income for Dr. Lasser. There was also no evidence in the record that Dr. Lasser could not perform scheduled elective surgery, as well as all other office procedures and tasks.
*396III.
The majority, I sense, has fallen into the same pattern of analysis employed by the District Court. Instead of addressing and assessing the basis, the rationale and the evidence which led to the Administrator’s determination denying benefits, and then giving appropriate deference to that determination, the majority discusses the evidence as if it was subjected to de novo review after trial. In doing so, it draws its own conclusions and substitutes its own judgment as to who and what is to be believed.
Instead of looking to the evidence of materiality-(there is none)-or the evidence of whether “emergency” and “on-call” procedures are material-(the evidence is uncontradicted that they are not)-the majority opinion reviews the District Court record as if it arose from a non-jury fact-finding trial. That, I suggest, was the error the District Court fell into and it is compounded by the majority of this panel.
I also find that the majority’s preoccupation with the definition of “regular occupation” misses the mark. Here, we are concerned only with the 'policy’s definition of “total disability” which concentrates on material aspects of Dr. Lasser’s practice. The majority relies on the Second Circuit’s decision in Kinstler v. First Reliance Standard Life Ins. Co., 181 F.3d 243 (2d Cir.1999), even though that case is distinguishable inasmuch as the standard of review there, unlike here, was de novo. The majority further strains to justify the term “regular occupation” by relying on a Fourth Circuit unpublished opinion, O’Bryhim v. Reliance Standard Life Ins. Co., 188 F.3d 502 (Table), 1999 WL 617891 (4th Cir.1999) (unpublished per curiam). See maj. op. at 386. A Fourth Circuit unpublished opinion under no circumstances can be considered precedential by this Court. In doing so, the majority opinion ignores the claim evaluation framework established by the very language of the Reliance policy that it quotes: First, the claimant’s condition is evaluated. Next, the claimant’s regular occupation is evaluated and the material duties thereof are defined. Finally, the claimant’s condition is compared against the material duties determined to be required in his regular occupation. It appears that Dr. Lasser, the District Court, and now this Court would end the inquiry after just the first step.
IV.
Another example of the District Court’s failure to defer to Walsh’s determinations revolves around Dr. Lasser’s continuing to work while receiving disability payments of about $6,000 a month, as well as Dr. Lasser’s return to work after his benefits were discontinued.
Reliance argues that Dr. Lasser cannot be totally disabled because in December 1997 he returned to “full-time” unrestricted work, including “emergency” and “on-call” duties, without consequence to his health. See App. at 263. Reliance further contends that it was reasonable because none of the doctors stated that he would be incapable of performing a full-time schedule and Drs. Burke and Field specifically evaluated Dr. Lasser as being capable of full-time work.
The District Court concluded, and the majority here agrees, that because Dr. Lasser risked future heart attacks due to the stress from on-call and emergency duties-which on this record cannot be characterized as material-his continuing to work after the termination of his insurance benefits nevertheless is evidence of his “total disability.” In particular, the District Court focused on the fact that he risked his health, against the express rec*397ommendations of his doctors, by returning to those duties out of economic necessity. Lasser II, 146 F.Supp.2d at 626 n. 5, 627, 630. But, without satisfying the policy’s condition of inability to perform tasks material to his occupation, I cannot agree that Dr. Lasser, who was working in his office and performing elective surgery, was “totally disabled.”
The Eighth Circuit has addressed this issue of relating economic necessity to total disability in a context similar to Dr. Lasser’s. In Galman v. Prudential Ins. Co. of America, 254 F.3d 768 (8th Cir.2001), the Eighth Circuit held that a claimant cannot assert total disability when he has returned to work, regardless of whether he returned to work out of “economic necessity” after an initial denial of benefits while awaiting the results of his administrative appeal. Galman, 254 F.3d at 771-72. Galman was a trial attorney who suffered from coronary artery heart disease, with a history of, inter alia, two prior heart attacks. He based his disability claim, in part, on the risk of aggravating his heart condition due to the stress of work. After the claim was denied, Galman returned to work full time asserting that the denial “forced him to return to work for financial reasons.” Id. at 771. The court rejected his claim for benefits. Id. at 772.
The majority implies that Dr. Lasser was unable to work a forty-hour week, and thus was disabled. But the majority misapprehends our jurisprudence. An insurer is not arbitrary and capricious merely because it relies on other medical opinions that result in a denial of coverage; it is only arbitrary if such decision was “clearly not supported by the evidence in the record or the administrator has failed to comply with the procedures required by the plan.” Abnathya v. Hoffmann-La Roche, Inc., 2 F.3d 40, 41 (3d Cir.1993); see also, e.g., Orvosh v. Program of Group Ins. for Salaried Employees, 222 F.3d 123, 129 (3d Cir.2000); Hightshue v. AIG Life Ins. Co., 135 F.3d 1144, 1148-49 (7th Cir.1998) (a plan’s decision, if reasonable, must be affirmed even if the claimant presents an opposite position). Here, Reliance has fully complied with all established procedures and the Administrator’s record is replete with evidence supporting the Administrator’s decision.
V.
Moreover, the District Court’s and the majority’s reliance on the issue of “stress” does not do credit to their analyses.
Once again, I consider the evidence from the administrative record through the lens of deference. Under this deferential standard, it is clear that the District Court, and then the majority, erred.
First, it is undisputed that none of the doctors who supported the disability claim ever stated with any degree of medical certainty the increased risk to Dr. Lasser of returning to full-time employment. Reliance’s policy does not insure against future disability-only present disability. A doctor or other expert who determines that a patient is disabled based on the possible future effects of stress should have to quantify that risk so that the ERISA decision maker can determine if it constitutes a present disability. Thus, an opinion that Dr. Lasser’s venous graft may not survive another five years does not satisfy the condition of Reliance’s disability policy-a policy which requires that Dr. Lasser be unable to perform the material duties of his occupation at this time. To illustrate: I, as a judge, may go blind in five years-but if I am not blind now, that possibility cannot be deemed to render me presently disabled.
*398I think we all recognize that stress, while it may affect cardiac patients, does not necessarily incapacitate them, or prevent them from successfully returning to, and performing, stressful jobs, i.e., Vice President Dick Cheney. See, e.g., Coker v. Metropolitan Life Ins. Co., 281 F.3d 793, 796, 798-99 (8th Cir.2002) (plan not arbitrary and capricious in denying benefits, even where the claimant’s doctor declared him disabled based on the exacerbating factor of stress); Gooden v. Provident Life & Accident Ins. Co., 250 F.3d 329, 335 n. 8 (5th Cir.2001) (dismissing claimant’s complaint that his insurance company failed to account for his doctor’s concern of the stress associated with his job).
Based on the administrative record, it was not arbitrary and capricious for Reliance to determine that Dr. Lasser was not “disabled” based on any risk of relapse from stress. Indeed, stress tests subsequent to his heart attack in July 1996 did not reveal any chest pain or EKG abnormalities during cardiac testing, only a “very small myocardial infarction” which was “very small and probably clinically insignificant.” App. at 208. Three cardiologists rated Dr. Lasser at a Functional Class I level for physical capacity on the New York Heart Association Functional Classification System, which level designates “[pjatients with cardiac disease but with no limitation of physical activity.” Lasser II, 146 F.Supp.2d at 626. Dr. Aldrich’s March 25, 1998 report states: “I do not disagree [with Dr. Burke] that by all objective criteria Dr. Lasser is doing very well at this point in time” and also reports that Dr. Lasser “does not currently have any symptoms, his examination is negative with respect to the cardiovascular system, and his stress test from April of 1997, as well as the stress test performed by Dr. Burke showed good exercise capacity....” App. at 222. Dr. Raska’s August 5, 1998 report confirmed Dr. Lasser’s cardiovascular fitness as “excellent,” his cardiac condition as stable and further stated that patients with a New York Heart Association Functional Class I level typically have “an overall low risk of cardiovascular morbidity or mortality.” App. at 229.
VI.
I believe that the District Court should have determined, on the established administrative record, whether the Administrator abused his discretion and was arbitrary and capricious in determining at least the following matters:
(a) Did Dr. Lasser meet his burden of establishing the material tasks and aspects of his practice?
(b) Are emergency surgeries and on-call duties material parts of Dr. Lasser’s functions as an orthopedic surgeon?
(c) Could and did Dr. Lasser perform elective (scheduled, not emergency) surgery as material procedures of his occupation?
(d) Could Reliance’s policy as drawn take into account future disabilities that Dr. Lasser may suffer or was it limited to only the current or present disability which Dr. Lasser may or may not have?
It is these matters with which the District Court had to concern itself in determining whether the Administrator’s denial of benefits was arbitrary and capricious. By failing to do so, I believe the District Court erred.
One last word on this subject. I am troubled by the District Court’s reliance on the Ninth Circuit’s de novo standard in Kearney. See supra note 3. While Kear-ney may have its advocates, and I cannot help but note that the Ninth Circuit en banc court was severely divided, we have yet to depart from established review standards in ERISA and summary judgment *399cases. To my mind, the District Court’s abrogation of both these standards in favor of a Kearney review, rings a warning bell that cannot be ignored.
Had the District Court conducted its hearing as I suggested it should, that is, confining itself only to the resolution of whether Walsh was conflicted or biased, and had the District Court then resorted to the traditional summary judgment resolution based only on the record below, I would not have raised this issue. But this is not the way this case played out, as I have noted. If we permit the review that took place in this case, I fear it will signal to district courts in the future that they are free, in similar cases, to hold merits hearings and thus to deviate from the summary judgment and Pinto standards.
Thus, in sum, I would reverse the District Court’s judgment in favor of Dr. Las-ser and remand for the District Court to properly consider this case under the summary judgment standard, giving Pinto deference to Walsh’s determinations. As the Supreme Court just recently noted, a court must actually apply the correct standard; mere lipservice and mere citation to a standard of review will not suffice. See Price v. Vincent, — U.S. -, -, 123 S.Ct. 1848, 1852, 155 L.Ed.2d 877 (2003) (holding that the Sixth Circuit recited the correct standard of review under 28 U.S.C. § 2254(d)(1) but then did not apply it when it reviewed the double jeopardy question at issue.)
Thus, I respectfully dissent.
. The District Court filed two opinions in this case. On February 8, 2001, it filed Lasser I, 130 F.Supp.2d 616, and on June 13, 2001, it filed Lasser v. Reliance Standard Life Ins. Co., 146 F.Supp.2d 619 (D.N.J.2001) ("Lasser II").
. Pinto held that, "when an insurance company both funds and administers benefits, it is generally acting under a conflict that warrants a heightened form of the arbitrary and capricious standard of review." Pinto, 214 F.3d at 378.
. The District Court relied on a Ninth Circuit divided en banc case which, contrary to Third Circuit jurisprudence, required de novo, not deferential, review. See Kearney, 175 F.3d 1084. In Kearney, a lawyer with a heart condition who played several sets of tennis every weekend, car-raced up to 120 miles per hour and had medical opinions that he could return to work, contested the cessation of his disability policy payments. Id. at 1086. The District Court accorded de novo review and ruled in favor of the insurance company. Id. at 1086-87. A fractured en banc court reversed, approving de novo review but finding a material disputed fact. Id. at 1090, 1093.
. The majority opinion seeks to answer its own question, which it has posed as, "When he [Dr. Lasser] no longer handled night call and emergency surgeries, were they material? The District Court answered yes." Maj. op. at 387. But, as I have pointed out, it was not the function of the District Court to make such a finding. Nor is it the function of this Court to uphold such a finding when it appears in a summary judgment context.
Rather, the majority should have examined the record, as I have, for evidence of materiality as to night calls and emergencies. Had the majority done so, it would have learned, as I did, that no such evidence exists-and it is that evidence that the Reliance policy requires.
Moreover, the burden of proof as to materiality rests with Dr. Lasser and the majority opinion's purported clarification respecting the burden of proof does not change that prescription. See maj. op. at 391.