dissenting:
Ten years ago appellants filed suit claiming violations of the federal antitrust laws.1 Appellee insurance companies 2 had allegedly combined and conspired among themselves to fix the hourly labor rate to be paid automobile body repair shops such as the appellants. In addition, the insurance companies had allegedly implemented this “horizontal” (intra-industry) agreement by entering into “vertical” (inter-industry) agreements with certain preferred repair shops that agreed to work at the rates prescribed by appellees.
In 1975 the District Court granted summary judgment to the insurance companies after concluding that their activities were exempt from antitrust laws since they constituted the “business of insurance” under the McCarran-Ferguson Act.3 Proctor v. State Farm Mutual Auto. Ins. Co., 406 F.Supp. 27 (D.D.C.1975). This court affirmed, 561 F.2d 262 (D.C.Cir.1977), over my dissent, id. at 276. In 1979 the Supreme Court vacated the judgment of this court and remanded the case for further consideration in light of Group Life & Health Ins. Co. v. Royal Drug Co., 440 U.S. 205, 99 S.Ct. 1067, 59 L.Ed.2d 261 (1979) (hereafter Royal Drug ).4
On remand the District Court granted another summary judgment for the insurance companies after finding once again that the alleged horizontal and vertical agreements were exempt from antitrust scrutiny.5 In the opinion of the District Court, Royal Drug had left “untouched and intact” 6 the vacated Court of Appeals judgment. The majority affirms the District Court’s grant of summary judgment, although it relies in part on different grounds.
At the outset, the majority agrees with the lower court that, even after Royal *340Drug, the alleged horizontal price-fixing agreement is the “business of insurance.” At 318-325. Insurance companies can therefore legally conspire among themselves to fix labor rates in the automobile repair industry, and summary judgment is justified as a matter of law. In the alternative, the majority affirms the grant of summary judgment on different grounds. It conducts “an independent review”7 of the entire record and finds no evidentiary support for the alleged horizontal agreement, id. at 325-326, even though this issue was never addressed by the District Court.8 As for the vertical agreements between the insurers and certain repair shops, the majority properly rejects the District Court’s finding that they constitute the “business of insurance.” Id. at 336-337. Nonetheless, the majority finds the agreements legal under a “rule of reason” analysis, id. at 337-338, although it does not decide whether they would be lawful if used to implement an illegal horizontal agreement, id. at 338.
I respectfully dissent on three main grounds. First, in my view an agreement among insurers to fix labor rates of automobile repairers is not the “business of insurance.” Under Royal Drug the alleged horizontal agreement simply does not constitute the business of insurance. See Part I infra. Second, the majority’s attempt, in the alternative, to grant summary judgment on the horizontal agreement issue based on its own review of the record is, in my judgment, misguided. See Part II infra. Finally, since the vertical agreements were designed to implement an unlawful horizontal agreement, the vertical agreements themselves were illegal. See Part III infra.
I. THE HORIZONTAL AGREEMENT IS NOT THE “BUSINESS OF INSURANCE”
Appellants have alleged that the insurance companies exchanged data relating to, and also agreed upon, a “prevailing” labor rate at which to pay repair costs of automobile damage claims.9 .According to the majority, such actions are the “business of insurance” even after Royal Drug. In my view, these activities do not constitute the business of insurance under the McCarran Act. Analysis of Royal Drug and the legislative history of the antitrust exemption clearly warrants this conclusion.
A. The “Business of Insurance"
In Royal Drug the Supreme Court explained in some detail the meaning of the term “business of insurance.” The Court emphasized that “spreading of risk” — and not mere risk reduction — was an “indispensable” characteristic of insurance. 440 U.S. at 212, 214-215, 99 S.Ct. at 1073, 1074-75. Since agreements with third-party providers to reduce an insurer’s expenses merely cut costs and did not spread risk, the Court concluded that the agreements did not constitute the “business of insurance” even though they were the “business of insurance companies.” Id. at 217, 232-233, 99 S.Ct. at 1076, 1083-84.
Royal Drug indisputably narrowed the range of activities exempt from antitrust scrutiny. As the Second Circuit recently held, the exemption for the business of insurance after Royal Drug “is to be strictly limited to only the quintessential insurance functions.” Pireno v. New York State Chiropractic Ass’n, 650 F.2d 387, 392 (2d Cir. 1981), cert. granted, - U.S. -, 102 S.Ct. 595, 70 L.Ed.2d 587 (1981). Even the majority concedes that a literal reading of Royal Drug might exclude “almost every practice outside of the basic insurance contract itself * * At 319. This reading seems compelled by the Court’s analysis. See Note, 80 Colum.L. Rev. 1473, 1479 (1980); Sullivan & Wiley, Recent Antitrust Developments: Defining *341the Scope of Exemptions, Expanding Coverage, and Refining the Rule of Reason, 27 UCLA L.Rev. 265, 282-283 (1979).
The Supreme Court found strong confirmation for its narrow construction of the exemption in the legislative history of the McCarran Act. According to the Court, Congress understood the business of insurance to be the spreading of risk, and Congress’ primary concern was to create an exemption for “ ‘combined efforts for statistical and rate-making purposes * * *.’ ” Royal Drug, 440 U.S. at 221-222, 99 S.Ct. at 1078 (emphasis by the Court; quoting Report of National Association of Insurance Commissioners, 90 Cong.Rec. A4405 (1944)). Because the legislative history bears directly on the practices alleged on appeal, it deserves special emphasis.10
In 1944 the National Association of Insurance Commissioners (NAIC) issued a report and proposed legislation concerning regulation of the insurance business. See 90 Cong.Rec. A4403-A4408 (1944). The views of the NAIC were deemed “particularly significant” by the Court in Royal Drug since the McCarran Act was based “in large part” on the NAIC proposal. 440 U.S. at 221, 99 S.Ct. at 1078. The foremost concern of the NAIC report was that insurance companies in fields other than life insurance11 had great difficulty underwriting risks accurately.12 In such fields there was “no guarantee that the contingency insured against will occur at all.” 90 Cong.Rec. at A4405. Rates could only be estimated “with a lesser degree of certainty.” Id. It followed that:
Since rates in these other fields are based upon the law of averages it is manifest that the broader the statistical base the more accurate the average. The experience of individual companies is seldom a reliable guide for rate-making purposes. * * *
Id.
The NAIC report identified a secondary concern of the insurance commissioners. Other than in the life insurance field, where to sell below a proper rate was to “invite insolvency” given the certainty that people would die and claims accrue, there existed the “temptation upon the part of some underwriters to assume that the contingency insured against will not occur.” Id. This raised the spectre of “inadequate rates and eventual insolvency or sharp claim practices.” Id.
Based on its concerns, the NAIC report found a need for “cooperation in obtaining statistical data and in the promulgation of rates based thereon.” Id. Since this result could be obtained “only through concert of action,” id., Congress needed to enact an exemption from the antitrust laws.
Congressional sources expressed many of the same thoughts as the NAIC report. For example, an early House Report stated:
The theory of insurance is the distribution of risk according to hazard, experience, and the laws of averages. These factors are not within the control of the insuring companies in the sense that the producer or manufacturer may control cost factors. Obviously, these factors, in their bearing upon rates and forms of policies[,] can best be determined through cooperative action * * *. * * *
H.R.Rep.No.873, 78th Cong., 1st Sess. 8-9 (1943) (emphasis added).13 See also S.Rep. No.1112, 78th Cong., 2d Sess. 6 (1944) (ways of safeguarding against the “untoward manifestations of nature and other vicissitudes of life” were not subject to free com*342petition) (quoting Osborn v. Ozlin, 310 U.S. 53, 65, 60 S.Ct. 758, 762, 84 L.Ed. 1074 (1940)).
Floor debates echoed similar themes. For instance, Representative Hancock stressed that the essence of insurance is “distribution of risk according to hazard,” and that, while “insurance companies cannot control the law of averages, * * * they must reckon with them.” 90 Cong.Rec. 6526 (1944). Also, members from both the Senate and the House emphasized that the new Act would exempt the activities of rating bureaus, whereby insurers pooled their loss experience in order to formulate rates. 91 Cong.Rec. 1481 (1945) (remarks of Senator Ferguson); 90 Cong.Rec. 6527 (1944) (remarks of Representative Miller of Connecticut).
B. The Alleged Horizontal Agreement
The Supreme Court has deemed it “well settled” that express statutory exemptions from the antitrust laws are to be “narrowly construed.” Royal Drug, 440 U.S. at 231, 99 S.Ct. at 1083 (citing prior cases). We should therefore only exempt those activities that clearly fall within the scope of the exemption. Here, Royal Drug and the legislative history of the exemption make clear that the challenged activities are not exempt from antitrust scrutiny.
To begin with, an agreement to pay repair costs using a “prevailing” labor rate manifestly does not involve spreading risk, but only risk reduction. Just like the pharmacy agreements in Royal Drug, the agreement as to the labor rate serves to cut the insurers’ costs and thereby reduces the magnitude of the risk the insurer faces. However, the risk of automobile damages is not in any way distributed by the horizontal agreement in this case. The majority never explains, and I fail to see, how the challenged practices meet the “indispensable characteristic”14 of risk-spreading. Instead, the majority argues that the horizontal agreement is the business of insurance because, in its judgment, it fits within Congress’ and Royal Drug’s understanding of statistical data-sharing and cooperative ratemaking. At 322. The majority reasons that since the costs of repair are directly related to the calculation of premiums, the challenged agreement is “virtually” a part of ratemaking and is therefore the business of insurance. Id. at 324.
I strongly disagree with the majority’s view15 that the alleged agreement falls within Congress’ concern about exempting the sharing of statistical data and cooperative ratemaking. With respect to data-sharing, Congress worried about the insurers’ ability to deal with factors (a) that were subject to hazard and the law of averages,16 (b) that were beyond the control of the insurance companies,17 and (c) that the insurers might be tempted to underestimate.18 For instance, data concerning the number of accidents involving drivers in a certain age cohort definitely fall within the scope of congressional concern.19 But data concerning the prevailing labor rate differ substantially from the statistical data that insurers were supposed to be able to share.
First, the labor rate in a given market is not a matter of random occurrence; while it is normally subject to the vagaries of the free market, it is not subject to hazard in the way that accidental events, are. Second, and most important, the labor rate is not a factor that is beyond the control of insurance companies. Indeed, appellants allege that the insurers have actually set a prevailing labor rate. By contrast, an insurance company has no direct control over *343the number of accidents its insureds will have in a given year20; therefore, companies can legally share data involving the frequency of accidents. Third, insurers have little incentive to underestimate the labor rate in a given market, for they continually pay claims that are based on the rate. An insurer would face immediate losses if it arbitrarily assumed a labor rate below that set by the market. (On the other hand, an insurer does not face immediate losses by assuming that future contingencies such as accidents will not occur. This poses the risk of insolvency, a concern addressed by the antitrust exemption.) Thus the data-sharing envisioned by Congress in enacting the McCarran exemption differs fundamentally from the exchange of data alleged in this case.
Nor does the horizontal agreement in this case fall within the meaning of cooperative ratemaking. Ratemaking involves the setting of prices within the insurance industry itself; it does not encompass fixing prices in markets outside the insurance industry. For example, no one would suggest that a state authority charged with insurance ratemaking could also set labor rates in provider markets. Rating bureaus certainly have no such authority. See, e.g., 7 D.C.Code §§ 35-1601 to 35-1609 (1981) (Fire Insurance Rating Bureau); 91 Cong. Rec. 1481 (1945) (remarks of Senator Ferguson) (McCarran Act exemption covered activities of D.C. rating bureau). Therefore, while the business of insurance encompasses the combined setting of rates based on pooled data involving a given set of variables, I do not see how it extends to the combined manipulation of those variables.21
The majority’s conception of the business of insurance has pernicious implications. As long as a factor is “directly related to the calculation of premiums,” (emphasis in original), insurers can conspire to affect it. Thus, under the majority’s standard, automobile insurers can conspire and combine to fix the prices of glass, metals, and rubber used in the auto repair process, since all of these are important inputs like the labor rate. Similarly, health insurers could agree among themselves to fix the prices charged by all types of physicians for their services, since the labor rate of these individuals is analogous to that of automobile repairers. And malpractice insurers could fix the hourly legal rates of attorneys who defend their insureds. In short, under the guise of the “business of insurance,” the majority would allow insurance companies to reach price-fixing agreements on an endless array of factors that affect the magnitude of the payments that insurers must make. Such an approach amounts to a stark departure from the narrow limits of the business of insurance exemption.22
II. THE MAJORITY’S ERRONEOUS DECISION TO GRANT SUMMARY JUDGMENT ON THE BASIS OF THE EVIDENCE CONCERNING THE HORIZONTAL AGREEMENT
Based on its examination of apparently all of the record evidence, see at 327 & n.40, the majority, to secure a back-up posi*344tion in the event its Royal Drug position proves unacceptable, decides to address an issue never reached by the District Court: whether appellees have demonstrated the absence of any genuine issue of fact material to appellants’ price-fixing claim. Since this issue was not discussed by the District Court, the issue was not properly presented for review. Thus it was not addressed directly by appellants. Instead, appellants’ argument focused oft whether appellees’ alleged activity constituted the “business of insurance” under the McCarran Act,23 although they naturally pointed to some of the factual support for their claims. Regardless, the majority decides to conduct its own summary judgment review of the record and thereby provide an alternative disposition of the case to which appellants have had no opportunity to respond.
A. The Use of Trials in Antitrust Cases
The majority’s efforts run counter to clear guidance from the Supreme Court concerning disposition of complex antitrust cases. While lower courts have long attempted to handle such actions without trials, these attempts have usually been rebuffed. For example, one federal judge sought to dispose of two antitrust suits by referring them to a master for resolution. The Supreme Court decisively rejected such an approach and stated:
[Mjost litigation in the antitrust field is complex. It does not follow that antitrust litigants are not entitled to a trial before a court. On the contrary, we believe that this is an impelling reason for trial before a regular, experienced trial jU(Jge * * * * * *
La Buy v. Holmes Leather Co., 352 U.S. 249, 259, 77 S.Ct. 309, 315, 1 L.Ed.2d 290 (1957).
Similarly, lower courts have often sought to dispose of complex antitrust cases through the use of summary judgment. In Poller v. CBS, Inc., 368 U.S. 464, 82 S.Ct. 486, 7 L.Ed.2d 458 (1962), the Court strongly disapproved of resort to summary disposition.
We believe that summary procedures should be used sparingly in complex antitrust litigation where motive and intent play leading roles, the proof is largely in the hands of the alleged conspirators, and hostile witnesses thicken the plot. It is only when the witnesses are present and subject to cross-examination that their credibility and the weight to be given their testimony can be appraised. Trial by affidavit is no substitute for trial by jury which so long has been the hallmark of “even handed justice.”
Id. at 473, 82 S.Ct. at 491 (footnote omitted).
Poller established a “rigorous standard” for summary judgment. 10 C. Wright & A. Miller, Federal Practice and Procedure § 2732 at 609 (1973). Where “motive and intent play leading roles,” 368 U.S. at 473, 82 S.Ct. at 491, courts should avoid summary dispositions. On the other hand, the Court has indicated that where the law is “so well developed that * * * the gist of the case turns on documentary evidence, the rule at times can be divined without a trial.” White Motor Co. v. United States, 372 U.S. 253, 259, 83 S.Ct. 696, 9 L.Ed.2d 738 (1963).
B. Summary Disposition in This Case
In this appeal we confront precisely the situation where “motive and intent” play leading roles. As the majority concedes, appellants have presented “some evidence of parallel behavior by appellees.” At 334. In particular, according to the majority, the evidence suggests that
upon occasion certain appellees used the same labor rate in writing estimates, that they had similar arrangements with repair shops that agreed to do volume work at the low rates used in their estimates, that they conducted surveys of repair shops to determine the average rate charged by shops in particular areas, and that they tended to resist price increases by repair shops. * * *
Id. at 334-335 (emphasis in original).
Therefore, as the majority itself recognizes, the crucial question in this case is *345whether these parallel activities were undertaken pursuant to agreement or whether they were undertaken independently in the “economic self-interest of each of the individual appellees.” Id. at 335. Indisputably, this issue involves “motive and intent” since it goes to the reason why insurance companies acted as they did. The case is thus in the grip of Poller, and summary disposition is ill-advised.
The inappropriateness of summary judgment is magnified by the fact that the District Court chose not to discuss the state of the record, even though motion papers before the court had dealt with the existence of the horizontal agreement. Instead, we deal here with a more egregious practice: summary disposition by a Court of Appeals based on its own review of several volumes of lifeless record. Appellate courts are hardly well-suited to this task.
As for the record itself, the majority would grant summary judgment because it finds no evidence that supports or creates an inference of a conspiracy or agreement. At 327. However, the record is hardly devoid of genuine factual issues. For example, the record amply demonstrates that one of the appellees visited several other major insurance companies on more than one occasion and discussed in extensive detail their procedures for handling automobile damage claims, including use of backup and preferred garages. See II Appendix for Appellants (App.) 282A-374A (Liberty Mutual memoranda). Other meetings among appellees on the subject of the costs of automobile damage repair also took place. See, e.g., Record, Vol. 9, Entry No. 235, Exhibit C. Moreover, one of the appellees (Travelers) opened up an experimental body repair shop in order to determine repair costs more accurately, and another appellee (Nationwide) subsequently secured data from Travelers about the hourly rate at which a profit could be made. Id., Exhibit B. The record also shows attempts by one company to induce collective action to resist an increase in the labor rate in one market. See App. 126A (State Farm document stating that no other insurers “would join us to this extent”). Similarly, a Liberty Mutual memorandum to its claims personnel instructed them to “check with competitors, as some of our offices have done,” if they had concerns about the number of back-up garages which would do work at given rates. App. 378A. Moreover, a survey of auto repair labor rates conducted by one company shows a surprising uniformity across several major markets. See Record, Vol. 9, Entry No. 235, Exhibit E (showing same hourly rate in Boston, Charlotte, Chicago, Dallas, New Haven, New Orleans, Philadelphia, San Diego, and Syracuse). These facts, and others spread throughout the record, create an inference of agreement among the insurers, especially when combined with the clear evidence of parallel behavior. The existence of a conspiracy is surely a matter of dispute for a jury to resolve.
The majority questions the significance of such evidence, e.g., at 329-330 n.46, and it refuses to infer any agreement despite evidence of contacts and parallel behavior, id. at 45. However, “if the evidence presented on the motion is subject to conflicting interpretations, or reasonable men might differ as to its significance, summary judgment is improper.” 10 C. Wright & A. Miller, supra, § 2725 at 515 (footnotes omitted). See, e.g., United States v. Perry, 431 F.2d 1020, 1022 (9th Cir. 1970) (“Summary judgment should not be granted where contradictory inferences may be drawn from undisputed evidentiary facts.”); Chenette v. Trustees of Iowa College, 431 F.2d 49, 53 (8th Cir. 1970) (“Evaluative judgment between two rationally possible conclusions from facts cannot be engaged in on summary judgment.”).
III. THE VERTICAL AGREEMENTS
The majority concedes that appellants have presented evidence to support the existence of vertical agreements between insurers and repair shops to do work at specified rates. E.g., at 335. The majority also rejects the District Court’s judgment that such agreements are the business of insurance. Id. at 336-337. The only *346issue in dispute is whether such agreements are illegal. However, because of its holdings concerning the horizontal agreement, the majority offers “no opinion on whether the alleged vertical arrangements would be unlawful if they were used to implement an unlawful horizontal agreement * * Id. at 338.
Since I have rejected the majority’s conclusions as to the legality and existence of the alleged horizontal agreement, I address the legality of a vertical agreement used to implement an illegal horizontal agreement. While vertical agreements are not illegal per se, they would be illegal if anticompetitive in purpose and effect. See Nat’l Society of Professional Engineers v. United States, 435 U.S. 679, 692, 98 S.Ct. 1355, 1365, 55 L.Ed.2d 637 (1978). Here, the vertical agreements were a necessary concomitance to the alleged price-fixing agreement. They advanced a horizontal agreement that was inherently anticompetitive and that was per se illegal. See id. Seen in this light, the vertical agreements themselves were necessarily anticompetitive. Accordingly, given the existence of the illegal horizontal agreement, a finding that the vertical agreements violated the law seems inescapable.
IV. CONCLUSION
Ten years after filing their complaint, appellants are once again deprived of a trial on the merits. In order to achieve this result, the majority has in my judgment badly misconstrued the meaning of Royal Drug and of the legislative history of the McCarran Act exemption. Moreover, despite contrary guidance from the Supreme Court, it has summarily disposed of the case after examining the record even though the District Court was unwilling to do so. As a result, it fails to address the legality of vertical agreements which implement unlawful horizontal agreements.
For all of these reasons, I respectfully dissent.
. In particular, § 1 et seq. of the Sherman Act, 15 U.S.C. § 1 et seq. (1976).
. The five appellees — State Farm Mutual Automobile Insurance Co., Liberty Mutual Insurance Co., Nationwide Mutual Insurance Co., Allstate Insurance Co., and Travelers Indemnity Co. — are among the largest automobile insurers in the United States.
. 15 U.S.C. § 1012(b)(1976) (hereafter referred to as the McCarran Act).
. Proctor v. State Farm Mutual Auto. Ins. Co., 440 U.S. 942, 99 S.Ct. 1417, 59 L.Ed.2d 631 (1979).
. Proctor v. State Farm Mutual Auto. Ins. Co., [1980-81] Trade Cases (CCH) H 63,591 (D.D.C. Oct. 12, 1980). In the alternative the court found the vertical agreements legal. Id at 77,-140.
. Id (discussing the horizontal agreement); see also id. (finding Royal Drug was not controlling as to the vertical agreements)'.
. Id. at 325 n.37.
. Id. at 325 (issue “not reached” by District Court).
. Brief for appellants at 5-6.
. This discussion draws heavily on sources cited in Royal Drug, 440 U.S. at 220-224, 99 S.Ct. at 1077-79.
. The NAIC report noted that life insurance rates were based in part on mortality tables, which in turn “are based upon the certainty that everyone must die * * 90 Cong.Rec. at A4405. This certainty was contrasted with the uncertainty in fields such as casualty insurance. Id.
. Royal Drug, 440 U.S. at 221, 99 S.Ct. at 1078.
. The House Report also stressed the importance of “preserving the solvency of insuring companies” through provision of adequate rates. H.R.Rep.No. 873, 87th Cong., 1st Sess. 9 (1944).
. Royal Drug, 440 U.S. at 212, 99 S.Ct. at 1073.
. At 322.
. See text at pp. 311-312 supra.
. See text at p. 312 supra.
. See text at pp. 311-312 & n.13 supra.
. The frequency of accidents is obviously sub-' ject to hazard and the law of averages, is beyond the direct control of the insurance companies, and is the type of data insurers might be tempted to underestimate. Historical data concerning the average extent of damage per accident are probably analogous.
. Nor does an insurance company exert direct control over the actual physical damage resulting from accidents involving its insureds.
. See Note, 80 Colum.L.Rev. 1475, 1479 (1980).
. The alleged practices in this case simply bear no relationship to the concerns underlying passage of the McCarran exemption. Those concerns centered on the inability of insurance companies, especially smaller ones, to underwrite risks accurately in the absence of cooperation. Royal Drug, 440 U.S. at 221, 99 S.Ct. at 1078. Yet we deal in this case with enormous insurance companies that did not even engage in the alleged price-fixing agreement until 1968 or 1969. Brief for appellants at 5. For over 20 years after passage of the exemption, insurance companies accurately underwrote risks without agreeing upon the labor rates of auto repair shops. Consequently, the challenged activities hardly amount to “quintessential” insurance functions. Like the agreements in Royal Drug, the activities here are “legally indistinguishable from countless other business arrangements that may be made by insurance companies to keep their costs low and thereby also keep low the level of premiums charged to their policyholders.” 440 U.S. at 215, 99 S.Ct. at 1075. While the horizontal agreements may be the business of insurance companies, they are definitely not the business of insurance.
. See brief for appellants at 10 (statement of argument).