delivered the opinion of the Court.

The respondents, 18 owners of independent pharmacies in San Antonio, Tex., brought an antitrust action in a Federal District Court against the petitioners, Group Life and Health Insurance Co., known as Blue Shield of Texas (Blue Shield), and three pharmacies also doing business in San Antonio. The complaint alleged that the petitioners had violated § 1 of the Sherman Act, 15 U. S. C. § 1, by entering agreements to fix the retail prices of drugs and pharmaceuticals, and that the activities of the petitioners had caused Blue Shield’s policyholders not to deal with certain of the respondents, thereby constituting an unlawful group boycott. The trial court granted summary judgment to the petitioners on the ground that the challenged agreements are exempt from the antitrust laws under § 2 (b) of the McCarran-Ferguson Act, 59 Stat. 34, as amended, 61 Stat. 448, 15 U. S. C. § 1012 (b), because the agreements are the “business of insurance,” are “regulated by [Texas] law,” and are not “boycotts” within the meaning of § 3 (b) of the Act, 59 Stat. 34, 15 U. S. C. *208§ 1013(b).1 415 F. Supp. 343 (WD Tex.). The Court of Appeals for the Fifth Circuit reversed the judgment. Holding that the agreements in question are not the “business of insurance” within the meaning of §2 (b), the appellate court did not reach the other questions decided by the trial court. 556 F. 2d 1375. We granted certiorari because of intercircuit conflicts as to the meaning of the phrase “business of insurance” in § 2 (b) of the Act.2 435 U. S. 903.

*209I

Blue Shield offers insurance policies which entitle the policyholders to obtain prescription drugs. If the pharmacy selected by the insured has entered into a “Pharmacy Agreement” with Blue Shield, and is therefore a participating pharmacy, the insured is required to pay only $2 for every prescription drug. The remainder of the cost is paid directly by Blue Shield to the participating pharmacy. If, on the other hand, the insured selects a pharmacy which has not entered into a Pharmacy Agreement, and is therefore a nonparticipating pharmacy, he is required to pay the full price charged by the pharmacy. The insured may then obtain reimbursement from Blue Shield for 75% of the difference between that price and $2.

Blue Shield offered to enter into a Pharmacy Agreement with each licensed pharmacy in Texas. Under the Agreement, a participating pharmacy agrees to furnish prescription drugs to Blue Shield’s policyholders at $2 for each prescription, and Blue Shield agrees to reimburse the pharmacy for the pharmacy’s cost of acquiring the amount of the drug prescribed. Thus, only pharmacies that can afford to distribute prescription drugs for less than this $2 markup can profitably participate in the plan.3

*210The only issue before us is whether the Court of Appeals was correct in concluding that these Pharmacy Agreements are not the “business of insurance” within the meaning of § 2 (b) of the McCarran-Ferguson Act. If that conclusion is correct, then the Agreements are not exempt from examination under the antitrust laws.4 Whether the Agreements are illegal under the antitrust laws is an entirely separate question, not now before us.5

II

A

As the Court stated last Term in St. Paul Fire & Marine Ins. Co. v. Barry, 438 U. S. 531, 541,6 the starting point in a case involving construction of the McCarran-Ferguson Act, like the starting point in any case involving the meaning of a statute, is the language of the statute itself. See also Blue Chip Stamps v. Manor Drug Stores, 421 U. S. 723, 756 (Powell, J., concurring). It is important, therefore, to observe at the outset that the statutory language in question *211here does not exempt the business of insurance companies from the scope of the antitrust laws. The exemption is for the “business of insurance,” not the “business of insurers”:

“The statute did not purport to make the States supreme in regulating all the activities of insurance companies; its language refers not to the persons or companies who are subject to state regulation, but to laws 'regulating the business of insurance.’ Insurance companies may do many things which are subject to paramount federal regulation; only when they are engaged in the ‘business of insurance’ does the statute apply.” SEC v. National Securities, Inc., 393 U. S. 453, 459-460. (Emphasis in original.)

Since the law does not define the “business of insurance,” the question for decision is whether the Pharmacy Agreements fall within the ordinary understanding of that phrase, illumined by any light to be found in the structure of the Act and its legislative history. Cf. Ernst & Ernst v. Hochfelder, 425 U.S. 185,199, and n. 19.

B

The primary elements of an insurance contract are the spreading and underwriting of a policyholder’s risk. “It is characteristic of insurance that a number of risks are accepted, some of which involve losses, and that such losses are spread over all the risks so as to enable the insurer to accept each risk at a slight fraction of the possible liability upon it.” 1 G. Couch, Cyclopedia of Insurance Law § 1:3 (2d ed. 1959). See also R. Keeton, Insurance Law § 1.2 (a) (1971) (“Insurance is an arrangement for transferring and distributing risk”); 1 G. Richards, The Law of Insurance § 2 (W. Freedman 5th ed. 1952).7

*212The significance of underwriting or spreading of risk as an indispensable characteristic of insurance was recognized by this Court in SEC v. Variable Annuity Lije Ins, Co., 359 U. S. 65. That case involved several corporations, representing themselves as “life insurance” companies, that offered variable annuity contracts for sale in interstate commerce. The companies were regulated by the insurance commissioners of several States. Purchasers of the contracts were not entitled to any fixed return, but only to a pro rata participation in the investment portfolios of the companies. Thus a policyholder could receive substantial sums if investment decisions were successful, but very little if they were not. One of the questions presented was whether these variable annuity contracts were the “business of insurance” under § 2 (b) of the McCarran-Ferguson Act.8 The Court held that the annuity contracts were not insurance, even though they were regulated as such under state law and involved actuarial prognostications of mortality. Central to the Court's holding was the premise that “the concept of 'insurance' involves some investment risk-taking on the part of the company.” 359 U. S., at 71. Since the variable annuity contracts offered no guarantee of fixed income, they placed all the investment risk on the annuitant and none on the company. Ibid. The Court concluded, therefore, that the annuities involved “no true underwriting of risks, the one earmark of insurance as it has commonly been conceived of in popular understanding and usage.” Id., at 73 (footnote omitted). Cf. German Alliance Ins, Co. v. Lewis, 233 U. S. 389, 412 (“The effect of insurance — indeed *213it has been said to be its fundamental object — is to distribute the loss over as wide an area as possible”).

The petitioners do not really dispute that the underwriting or spreading of risk is a critical determinant in identifying insurance. Rather they argue that the Pharmacy Agreements do involve the underwriting of risks. As they state in their brief:

“In Securities and Exchange Commission v. Variable Annuity Lije Insurance Co., 359 U. S. 65, 73 (1959), the 'earmark’ of insurance was described as the 'underwriting of risks’ in exchange for a premium. Here the risk insured against is the possibility that, during the term of the policy, the insured may suffer a financial loss arising from the purchase of prescription drugs, or that he may be financially unable to purchase such drugs. In consideration of the premium, Blue Shield assumes this risk by agreeing with its insureds to contract with Participating Pharmacies to furnish the needed drugs and to reimburse the Pharmacies for each prescription filled for the insured. In short, each of the fundamental elements of insurance is present here — the payment of a premium in exchange for a promise to indemnify the insured against losses upon the happening of a specified contingency.”

The fallacy of the petitioners’ position is that they confuse the obligations of Blue Shield under its insurance policies, which insure against the risk that policyholders will be unable to pay for prescription drugs during the period of coverage, and the agreements between Blue Shield and the participating pharmacies, which serve only to minimize the costs Blue Shield incurs in fulfilling its underwriting obligations.9 The *214benefit promised to Blue Shield policyholders is that their premiums will cover the cost of prescription drugs except for a $2 charge for each prescription.10 So long as that promise is kept, policyholders are basically unconcerned with arrangements made between Blue Shield and participating pharmacies.11

The Pharmacy Agreements thus do not involve any underwriting or spreading of risk, but are merely arrangements for the purchase of goods and services by Blue Shield. By agreeing with pharmacies on the maximum prices it will pay for drugs, Blue Shield effectively reduces the total amount it must pay to its policyholders. The Agreements thus enable Blue Shield to minimize costs and maximize profits. Such cost-savings arrangements may well be sound business practice, and may well inure ultimately to the benefit of policyholders in the form of lower premiums, but they are not the “business of insurance.”12

*215The Pharmacy Agreements are thus 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. Suppose, for example, that an insurance company entered into a contract with a large retail drug chain whereby its policyholders could obtain drugs under their policies only from stores operated by this chain. The justification for such an agreement would be administrative and bulk-purchase savings resulting from obtaining all of the company's drug needs from a single dealer. Even though these cost savings might ultimately be reflected in lower premiums to policyholders, would such a contract be the “business of insurance”? Or suppose that the insurance company should decide to acquire the chain of drug stores in order to lower still further its costs of meeting its obligations to its policyholders. Such an acquisition would surely not be the “business of insurance.” SEC v. National Securities, Inc., 393 U. S. 453.13

C

Another commonly understood aspect of the business of insurance relates to the contract between the insurer and the insured. In enacting the McCarran-Ferguson Act Congress was concerned with:

“The relationship between insurer and insured, the type of policy which could be issued, its reliability, interpreta*216tion, and enforcement — these were the core of the ‘business of insurance.’ Undoubtedly, other activities of insurance companies relate so closely to their status as reliable insurers that they too must be placed in the same class. But whatever the exact scope of the statutory term, it is clear where the focus was — it was on the relationship between the insurance company and the policyholder.” SEC v. National Securities, Inc., supra, at 460.

The Pharmacy Agreements are not “between insurer and insured.” They are separate contractual arrangements between Blue Shield and pharmacies engaged in the sale and distribution of goods and services other than insurance.

The petitioners argue that nonetheless the Pharmacy Agreements so closely affect the “reliability, interpretation, and enforcement” of the insurance contract and “relate so closely to their status as reliable insurers” as to fall within the exempted area.14 This argument, however, proves too much.

At the most, the petitioners have demonstrated that the Pharmacy Agreements result in cost savings to Blue Shield which may be reflected in lower premiums if the cost savings are passed on to policyholders. But, in that sense, every business decision made by an insurance company has some impact on its reliability, its ratemaking, and its status as a *217reliable insurer. The manager of an insurance company is no different from the manager of any enterprise with the responsibility to minimize costs and maximize profits. If terms such as “reliability” and “status as a reliable insurer” were to be interpreted in the broad sense urged by the petitioners, almost every business decision of an insurance company could be included in the “business of insurance.” Such a result would be plainly contrary to the statutory language, which exempts the “business of insurance” and not the “business of insurance companies.”

Ill

A

The conclusion that the Pharmacy Agreements are not the “business of insurance” is fully confirmed by the legislative history of the McCarran-Ferguson Act. The law was enacted in 1945 in response to this Court’s decision in United States v. South-Eastern Underwriters Assn., 322 U. S. 533. The indictment in that case charged that the defendants had conspired to fix insurance rates and commissions, and had conspired to boycott and coerce noncooperating insurers, agents, and insureds. In the District Court the defendants had successfully demurred to the indictment on the ground that the insurance industry was not a part of interstate commerce subject to regulation under the Commerce Clause.15 On direct appeal, this Court reversed the judgment, holding that the business of insurance is interstate commerce, and that the Congress which enacted the Sherman Act had not intended to exempt the insurance industry from its coverage.

B

The primary concern of Congress in the wake of that decision was in enacting legislation that would ensure that *218the States would continue to have the ability to tax and regulate the business of insurance.16 This concern is reflected in §§ 1 and 2 (a) of the Act,17 neither of which is involved in this case. A secondary concern was the applicability of the antitrust laws to the insurance industry.18 Months before *219this Court’s decision in South-Eastern Underwriters was announced, proposed legislation to totally exempt the insurance industry from the Sherman and Clayton Acts had been introduced in Congress.19 Less than three weeks after the actual decision, the House of Representatives passed a bill which would also have provided the insurance industry with a blanket exemption from the antitrust laws, thus restoring the state of law that had existed before the decision in SouthEastern Underwriters.20

Congress, however, rejected this approach.21 Instead of a total exemption, Congress provided in § 2 (b) that the antitrust laws “shall be applicable” unless the activities of insurance companies are the business of insurance and regulated by state law. Moreover, under § 3 (b) the Sherman Act was made applicable in any event to acts of boycott, coercion, or intimidation. To allow the States time to adjust to the applicability of the antitrust laws to the insurance industry, *220Congress imposed a 3-year moratorium.22 After the expiration of the moratorium on July 1, 1948, however, Congress clearly provided that the antitrust laws would be applicable to the business of insurance “to the extent that such business is not regulated by State law.” 23

By making the antitrust laws applicable to the insurance industry except as to conduct that is the business of insurance, regulated by state law, and not a boycott, Congress did not intend to and did not overrule the South-Eastern Underwriters case.24 While the power of the States to tax and regulate insurance companies was reaffirmed, the McCarran-Ferguson Act also established that the insurance industry would no longer have a blanket exemption from the antitrust laws. It is true that § 2 (b) of the Act does create a partial exemption from those laws. Perhaps more significantly, however, that section, and the Act as a whole, embody a legislative rejection of the concept that the insurance industry is outside the scope of the antitrust laws — a concept that had prevailed before the South-Eastern Underwriters decision.

C

References to the meaning of the “business of insurance” in the legislative history of the McCarran-Ferguson Act *221strongly suggest that Congress understood the business of insurance to be the underwriting and spreading of risk. Thus, one of the early House Reports 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 insuring companies in the sense that the producer or manufacturer may control cost factors.” H. R. Rep. No. 873, 78th Cong., 1st Sess., 8-9 (1943).25 See also S. Rep. No. 1112, 78th Cong., 2d Sess., 6 (1944); 90 Cong. Rec. 6526 (1944) (remarks of Rep. Hancock).

Because of the widespread view that it is very difficult to underwrite risks in an informed and responsible way without intra-industry cooperation, the primary concern of both representatives of the insurance industry and the Congress was that cooperative ratemaking efforts be exempt from the antitrust laws. The passage of the McCarran-Ferguson Act was preceded by the introduction in the Senate Committee of a report and a bill submitted by the National Association of Insurance Commissioners on November 16, 1944.26 The views of the NAIC are particularly significant, because the Act ultimately passed was based in large part on the NAIC bill.27 The report emphasized that the concern of the insurance commissioners was that smaller enterprises and insurers other than life insurance companies were unable to underwrite risks accurately, and it therefore concluded:

“For these and other reasons this subcommittee believes it would be a mistake to permit or require the unrestricted competition contemplated by the antitrust laws to apply to the insurance business. To prohibit oom-*222bined efforts for statistical and rate-making purposes would be a backward step in the development of a progressive business. We do not regard it as necessary to labor this point any further because Congress itself recently recognized the necessity for concert of action in the collection of statistical data and rate making when it enacted the District of Columbia Fire Insurance Rating Act.” Id., at A4405 (emphasis added).

The bill proposed by the NAIC enumerated seven specific practices to which the Sherman Act was not to apply.28 Each of the specific practices involved intra-industry cooperative or concerted activities. None involved contractual arrangements that insurance companies might make with providers of goods or services to reduce the costs to the companies of meeting their underwriting obligations to their policyholders.29

*223The floor debates also focused simply on whether cooperative ratemaking should be exempt. Thus, Senator Ferguson, in explaining the purpose of the bill, stated:

“This bill would permit — and I think it is fair to say that it is intended to permit — rating bureaus, because in the last session we passed a bill for the District of Columbia allowing rating. What we saw as wrong was the fixing of rates without statutory authority in the States; but we believe that State rights should permit a State to say that it believes in a rating bureau. I think the insurance companies have convinced many members of the legislature that we cannot have open competition in fixing rates on insurance. If we do, we shall have chaos. There will be failures, and failures always follow losses.” 91 Cong. Rec. 1481 (1945).

The consistent theme of the remarks of other Senators also indicated a primary concern that cooperative ratemaking would be protected from the antitrust laws. Id., at 1444 and 1485 (remarks of Sen. O’Mahoney); 485 (remarks of Sen. Taft).30 President Roosevelt, in signing the bill, also em*224phasized that the bill would allow cooperative rate regulation. He stated that “Congress did not intend to permit private rate fixing, which the Antitrust Act forbids, but was willing to permit actual regulation of rates by affirmative action of the States.” S. Rosenman, The Public Papers and Addresses of Franklin D. Roosevelt, 1944-1945 Vol., p. 587 (1950).31 There is not the slightest suggestion in the legislative history that Congress in any way contemplated that arrangements such as the Pharmacy Agreements in this case, which involve the mass purchase of goods and services from entities outside the insurance industry, are the “business of insurance.” 32

*225D

At the time of the enactment of the McCarran-Ferguson Act, corporations organized for the purpose of providing their *226members with medical services and hospitalization were not considered to be engaged in the insurance business at all, and thus were not subject to state insurance laws. E. g., Jordan v. Group Health Assn., 71 App. D. C. 38, 107 F. 2d 239 (1939); California Physicians’ Service v. Garrison, 155 P. 2d 885 (Cal. App. 1945). affd, 28 Cal. 2d 790, 172 P. 2d 4 (1946); Commissioner of Banking & Insurance v. Community Health Service, 129 N. J. L. 427, 30 A. 2d 44 (1943); State ex rel. Fishback v. Universal Service Agency, 87 Wash. 413, 151 P. 768 (1915).33 Similarly, States which regulated prepaid health-service plans at the time the Act was enacted either exempted them from the requirements of the state insurance code or provided that they “shall not be construed as being engaged in the business of insurance” under state law. Rorem, Enabling Legislation for Non-Profit Hospital Service Plans, 6 Law & Contemp. Prob. 528, 534 (1939).34 Since the legisla*227tive history makes clear that Congress certainly did not intend the definition of the “business of insurance” to be broader than its commonly understood meaning, the contemporary perception that health-care organizations were not engaged in providing insurance is highly significant in ascertaining congressional intent.

The Jordan v. Group Health hLssn. case, supra, is illustrative of the contemporary view of health-care plans. Group Health was organized as a nonprofit corporation to provide various medical services and supplies to members who paid a fixed annual premium. To implement the plan, Group Health contracted with physicians, hospitals, and others, to provide medical services. These groups were compensated exclusively by Group Health. By contracting with the various medical groups directly, Group Health was able to obtain *228services at a lower cost than if each member contracted separately. The plan, therefore, was somewhat similar to the Pharmacy Agreements in this case. The court in Group Health held that this type of arrangement was not insurance:

“Whether the contract is one of insurance or of indemnity there must be a risk of loss to which one party may be subjected by contingent or future events and an assumption of it by legally binding arrangement by another. Even the most loosely stated conceptions of insurance . . . require these elements. Hazard is essential and equally so a shifting of its incidence.

“Although Group Health’s activities may be considered in one aspect as creating security against loss from illness or accident, more truly they constitute the quantity purchase of well-rounded, continuous medical service by its members. Group Health is in fact and in function a consumer cooperative. The functions of such an organization are not identical with those of insurance or indemnity companies. The latter are concerned primarily, if not exclusively, with risk .... On the other hand, the cooperative is concerned principally with getting service rendered to its members and doing so at lower prices made possible by quantity purchasing and economies in operation.” 71 App. D. C., at 44, 46, 107 F. 2d, at 245, 247. (Emphasis supplied in part; footnotes omitted.) 35

*229Indeed, Blue Cross and Blue Shield organizations themselves have historically taken the position that they are not insurance companies in seeking to avoid state regulation and taxation.36 It is thus difficult to assume that contrary to this historical position and a majority of court decisions, Congress in 1945 understood that advance-payment medical-*230benefits plans are the “business of insurance.” 37 It is next to impossible to assume that Congress could have thought that agreements (even by insurance companies) which provide for the purchase of goods and services from third parties at a set price are within the meaning of that phrase.38

*231IV

It is well settled that exemptions from the antitrust laws are to be narrowly construed. E. g., Abbott Laboratories v. Portland Retail Druggists Assn., Inc., 425 U. S. 1; Connell Construction Co. v. Plumbers Steamfitters, 421 U. S. 616; FMC v. Seatrain Lines, Inc., 411 U. S. 726; United States v. McKesson & Robbins, Inc., 351 U. S. 305. This doctrine is not limited to implicit exemptions from the antitrust laws, but applies with equal force to express statutory exemptions. E. g., Abbott Laboratories v. Portland Retail Druggists Assn., Inc., supra, at 11-12 (the Nonprofit Institutions Act); FMC v. Seatrain Lines, Inc., supra, at 733 (§15 of the Shipping Act); United States v. McKesson Robbins, supra,'at 316 (the Miller-Tydings and McGuire Acts).

Application of this principle is particularly appropriate in this case because the Pharmacy Agreements involve parties wholly outside the insurance industry. In analogous contexts, the Court has held that an exempt entity forfeits antitrust exemption by acting in concert with nonexempt parties. The Court has held, for example, that an exempt agricultural cooperative under the Capper-Yolstead Act loses its exemption if it conspires with nonexempt parties. Case-Swayne Co. v. Sunkist Growers, Inc., 389 U. S. 384; United States v. Borden Co., 308 U. S. 188. Similarly, the Court has consistently stated that a union forfeits its exemption from the antitrust laws if it agrees with one set of employers to impose a wage scale on other bargaining units. Ramsey v. Mine Workers, *232401 U. S. 302, 313; Mine Workers v. Pennington, 381 U. S. 657, 665-666.39

If agreements between an insurer and retail pharmacists are the “business of insurance” because they reduce the insurer’s costs, then so are all other agreements insurers may make to keep their costs under control — whether with automobile body repair shops or landlords.40 Such agreements *233would be exempt from the antitrust laws if Congress had extended the coverage of the McCarran-Ferguson Act to the "business of insurance companies.”41 But that is precisely what Congress did not do.

For all these reasons, the judgment of the Court of Appeals is

Affirmed.

with whom The Chief Justice, Mr. Justice Marshall, and Mr. Justice Powell join, dissenting.

The McCarran-Ferguson Act, 59 Stat. 33, as amended, 15 IT. S. C. §§ 1011-1015, renders the federal antitrust laws inapplicable to the “business of insurance” to the extent such business is regulated by state law and is not subject to the “boycott” exception stated in § 1013 (b).1 The single question presented by this case is whether the “business of insurance” *234includes direct contractual arrangements (“provider agreements”) between petitioner Blue Shield and third parties to provide benefits owed to the insurer’s policyholders. The Court today holds that it does not.

I disagree: Since (a) there is no challenge to the status of Blue Shield’s drug-benefits policy as the “business of insurance,” I conclude (b) that some provider agreements negotiated to carry out the policy obligations of the insurer to the insured should be considered part of such business, and (c) that the specific Pharmacy Agreements at issue in this case should be included in such part. Before considering this analysis, however, it is necessary to set forth the background of the enactment of the McCarran-Ferguson Act.

I

SEC v. National Securities, Inc., 393 U. S. 453, 459 (1969), recognized that the legislative history of the McCarran-Ferguson Act sheds little light on the meaning of the words “business of insurance.” See S. Hep. No. 20, 79th Cong., 1st Sess. (1945); H. R. Rep. No. 143, 79th Cong., 1st Sess. (1945). But while the legislative history is largely silent on the matter,2 it does indicate that Congress deliberately chose *235to phrase the exemption broadly. Congress had draft bills before it which would have limited the “business of insurance” to a narrow range of specified insurance company practices, but chose instead the more general language which ultimately became law.3

*236The historical background of the statute’s enactment, developed by the Court in SEC v. National Securities, Inc., supra, provides the guide to congressional purpose:

“The McCarran-Ferguson Act was passed in reaction to this Court’s decision in United States v. South-Eastern Underwriters Asm., 322 U. S. 533 (1944). Prior to that decision, it had been assumed, in the language of the leading case, that ‘[iIssuing a policy of insurance is not a transaction of commerce.’ Paul v. Virginia, 8 Wall. 168, 183 (1869). Consequently, regulation of insurance transactions was thought to rest exclusively with the States. In South-Eastern Underwriters, this Court held that insurance transactions were subject to federal regulation under the Commerce Clause, and that the antitrust laws, in particular, were applicable to them. Congress reacted quickly . . . [, being] concerned about the inroads the Court’s decision might make on the tradition of state regulation of insurance. The McCarran-Ferguson Act was the product of this concern. Its purpose was stated quite clearly in its first section; Congress declared that The continued regulation and taxation by the several States of the business of insurance is in the public interest.’ 59 Stat. 33 (1945), 15 U. S. C. § 1011. As this Court said shortly afterward, ‘[ojbviously Congress’ purpose was broadly to give support to the existing and future state systems for regulating and taxing the business of insurance.’ Prudential Insurance Co. v. Benjamin, 328 U. S. 408, 429 (1946).

“The . . . Act was an attempt to turn back the clock, to assure that the activities of insurance companies in dealing with their policyholders would remain subject to state regulation.” 393 U. S., at 458-459.

See also St. Paul Fire & Marine Ins. Co. v. Barry, 438 U. S. 531, 538-539 (1978); 90 Cong. Rec. 6524 (1944) (Cong. Wal*237ter) (“[T]he legislation ... is designed to restore to the status quo the position the insurance business of this Nation occupied before the Supreme Court recently legislated [in SouthEastern Underwriters]”).

Since continuation of state regulation as it existed before South-Eastern was Congress’ goal,4 evidence of what States *238might reasonably have considered to be and regulated as insurance at the time the McCarran-Ferguson Act was passed in 1945 is clearly relevant to our decision. This does not mean that a transaction not viewed as insurance in 1945 cannot be so viewed today.

“We realize that . . . insurance is an evolving institution. Common knowledge tells us that the forms have greatly changed even in a generation. And we would not undertake to freeze the conceptt] of 'insurance’ . . . into the mold [it] fitted when these Federal Acts were passed.” SEC v. Variable Annuity Lije Ins. Co., 359 U. S. 65, 71 (1959).

It is thus logical to suppose that if elements common to the ordinary understanding of “insurance” are present, new forms of the business should constitute the “business of insurance” for purposes of the McCarran-Ferguson Act. The determination of the scope of the Act, therefore, involves both an analysis of the proximity between the challenged transactions and those well recognized as elements of “insurance,” and an examination of the historical setting of the Act. On both counts, Blue Shield’s Pharmacy Agreements constitute the “business of insurance.”

*239II

I start with common ground. Neither the Court, ante, at 230 n. 37, nor the parties challenge the fact that the drug-benefits policy offered by Blue Shield to its policyholders — as distinguished from the contract between Blue Shield and the pharmacies — is the “business of insurance.” Whatever the merits of scholastic argument over the technical definition of “insurance,” the policy both transfers and distributes risk. The policyholder pays a sum certain — the premium — against the risk of the uncertain contingency of illness, and if the company has calculated correctly, the premiums of those who do not fall ill pay the costs of benefits above the premiums of those who do. See R. Mehr & E. Cammack, Principles of Insurance 31-32 (6th ed. 1976). An important difference between Blue Shield’s policy and other forms of health insurance is that Blue Shield “pays” the policyholder in goods and services (drugs and their dispensation), rather than in cash. Since we will not “freeze the concep[t] of 'insurance’ . . . into the mold it fitted” when McCarran-Ferguson was passed, this difference cannot be a reason for holding that the drug-benefits policy falls outside the “business of insurance” even if our inquiry into the understandings of what constituted “insurance” in the 1930’s and 1940’s were to suggest that a contrary view prevailed at that time.5

Fortunately, logic and history yield the same result. It is true that the first health insurance policies provided only cash indemnities. However, although policies that specifically provided drug benefits were not available during the 1930’s and 1940’s, analogous policies providing hospital and medical services — rather than cash — were available.

The hospital service-benefit concept originated in Texas in *2401929; medical services were first offered in 1939. R. Eilers, Regulation of Blue Cross and Blue Shield Plans 10, 15 (1963) (hereinafter Eilers). In 1940, 4,500,000 people in 60 communities were covered by Blue Cross or related hospital-benefits plans. C. Rorem, Non-Profit Hospital Service Plans 1-2 (1940) (hereinafter Rorem I). During the 1940’s, health insurance became a subject of collective bargaining, with unions demanding the service-benefit approach of Blue Cross and Blue Shield. S. Law, Blue Cross 11 (1974) (hereinafter Law). By 1945, the year the McCarran-Ferguson Act was enacted, over 20 million people were enrolled in service-benefit programs, with service-benefit plans comprising 61% of the total hospitalization insurance market. See Hearings before the Senate Committee on Education and Labor, A National Health Program, 79th Cong., 2d Sess., pt. 1, p. 173 (1946); Eilers 19; Law 11.

Moreover, regulation of the service-benefit plans was a part of the system of state regulation of insurance that the McCarran-Ferguson Act was designed to preserve. Led by New York in 1934, 24 States passed enabling Acts by 1939 which, while relieving the plans of certain reserve requirements and tax obligations, specifically subjected service-benefit plans to the supervision and control of state departments of insurance.6 See Rorem, Enabling Legislation for Non-Profit Hospital Service Plans, 6 Law & Contemp. Prob. 528, 531, 534 (1939) (hereinafter Rorem II); N. Sinai, O. Anderson, & M. Dollar, Health Insurance in the United States *24148-49 (1946) (hereinafter Sinai); Comment, Group Health Plans: Some Legal and Economic Aspects, 53 Yale L. J. 162, 174 (1943). Another 16 States apparently limited the issuance of hospitalization insurance to stock and mutual insurance companies. Nine acted on the premise that the plans were not “insurance” and authorized operation under general corporation laws, exempt from reserve requirements. Rorem II, p. 532. By the time the MeCarran-Ferguson Act was passed, 35 States had enabling legislation.7 During this period, the National Association of Insurance Commissioners (NAIC), the organization of state insurance directors which played a major role in drafting the MeCarran-Ferguson Act,8 was also drafting model state enabling legislation to govern service-benefit health plans. Proceedings of the NAIC, 75th Sess., 226 (1944); id., 76th Sess., 250 (1945).9

*242Thus, when the McCarran-Ferguson Act became law, service-benefit plans similar to the Blue Shield plan at issue here were a widespread and well-recognized form of insurance, subject to regulation in most of the States. Congress itself treated these important programs as insurance. In 1939, Congress adopted an enabling Act incorporating a hospitalization-benefits plan in the District of Columbia, with supervi*243sory authority placed in the hands of the Superintendent of Insurance. See H. R. 6266, 76th Cong., 1st Sess. (1939); H. R. Rep. No. 1247,76th Cong., 1st Sess. (1939); 84 Cong. Rec. 11224 (1939). And in hearings held the year after passage of the McCarran-Ferguson Act, the same Congress that approved that Act debated Blue Shield-type programs as alternatives to national health insurance, with participating Congressmen frequently referring to them as “insurance.” Hearings before the Senate Committee on Education and Labor, A National Health Program, 79th Cong., 2d Sess., pt. I, pp. 55, 83, 108, 172, pt. 2, p. 558 (1946).10 The status of service-benefit policies as “insurance,” both logically and historically, is therefore sufficiently established to make that the first premise in an analysis of the status of the Pharmacy Agreements at issue in this case.

Ill

The next question is whether at least some contracts with third parties to procure delivery of benefits to Blue Shield's insureds would also constitute the “business of. insurance.” Such contracts, like those between Blue Shield and the druggists in this case, are known as “provider agreements.” The Court, adopting the view of the Solicitor General, today holds that no provider agreements can be considered part of the “business of insurance.” 11 It contends that the “underwriting or spreading of risk [is]' an indispensable characteristic of *244insurance,” ante, at 212,12 and that “[ajnother commonly understood aspect of the business of insurance relates to the contract between the insurer and the insured.” Ante, at 215. Because provider agreements neither themselves spread risk, nor involve transactions between insurers and insureds, the Court excludes them from the “business of insurance.”

The argument fails in light of this Court’s prior decisions and the legislative history of the Act. The Court has held, for example, FTC v. National Casualty Co., 357 U. S. 560 (1958), that the advertising of insurance, a unilateral act which does not involve underwriting, is within the scope of the McCarran-Ferguson Act. And the legislative history makes it abundantly clear that numerous horizontal agreements between insurance companies which do not technically involve the underwriting of risk were regarded by Congress as within the scope of the Act’s exemption for the “business of insurance.” For example, rate agreements among insurers, a conspicuous congressional illustration, see, e. g., 91 Cong. Rec. 1481, 1484 (1945) (remarks of Sens. Pepper and Ferguson), and the subject of the South-Eastern Underwriters case, see SEC v. National Securities, Inc., 393 U. S., at 460, do not themselves spread risk. Indeed, the Court apparently concedes that arrangements among insurance companies respecting premiums and benefits would constitute the “business of insurance,” despite their failure to fit within its formula. Ante, at 221 and 224-225, n. 32.

But the Court’s attempt to limit its concession to horizontal transactions still conflicts with the legislative history. Compelling evidence is the fact that Congress actually rejected a proposed bill to limit the exemption to agreements between *245insurance companies. S. 12, 79th Cong., 1st Sess. (1945). See n. 3, supra. Moreover, vertical relationships between insurance companies and independent sales agencies were a subject of the indictment in United States v. South-Eastern Underwriters Assn., 322 U. S. 533, 535 (1944), were the object of discussion in the House, 90 Cong. Ree. 6538 (1944) (remarks of Cong. Celler), and were expressly included as part of the “business of insurance” in an early draft of the Act, id., at A4406 (NAIC bill, §4 (b)(5)). Again, the Court concedes that such transactions, between insurers and agents, might fall within the “business of insurance,” despite the inconsistency with the Court’s own theory. Ante, at 224r-225, n. 32.13

The Court’s limitation also ignores the significance of pervasive state insurance regulation — prevailing when the Act was passed — of hospitalization-benefits plans whose “distinctive feature,” Rorem I, p. 64; Proceedings of the NAIC, 75th Sess., 228 (1944), was the provider contract with the participating hospital to provide service when needed. The year prior to adoption of the Act the NAIC emphasized the relationship between provider agreements and service-benefit policies:

“A hospital service plan is designed to provide service rather than to indemnify and this can only be guaranteed through contractual arrangements between plans and hospitals.” Ibid.

The Association also proposed, in the year MeCarran-Ferguson passed, a model state enabling Act requiring “full approval of . . . contracts with hospitals ... by the insurance commissioner.” Proceedings of the NAIC, 76th Sess., 250 *246(1945). That proposal reflected well the actual contents of existing state enabling Acts which armed insurance commissioners with considerable authority to regulate provider agreements.14 Congress itself authorized the service-benefit plan it incorporated in the District of Columbia “to enter into contracts with hospitals for the care and treatment of '[its subscribers].” H. it. 6266, 76th Cong., 1st Sess. (1939). In light of Congress’ objective through the McCarran-Ferguson Act to insure the continuation of existing state regulation, the conclusion that at least some provider agreements were intended to be within the “business of insurance” is inescapable.

Logic compels the same conclusion. Some kind of provider agreement becomes a necessity if a service-benefits insurer is to meet its obligations to the insureds. The policy before us in this case, for example, promises payment of benefits in drugs. Thus, some arrangement must be made to provide those drugs for subscribers.15 Such an arrangement obtains *247the very benefits promised in the policy; it does not simply relate to the general operation of the company. A provider contract in a service-benefit plan, therefore, is critical to “the type of policy which could be issued” as well as to its “reliability” and “enforcement.” It thus comes within the terms of SEC v. National Securities, Inc., 393 U. S., at 460. That case explained that the “business of insurance” involves not only the “relationship between insurer and insured,” but also “other activities of insurance companies [that] relate so closely to their status as reliable insurers that they too must be placed in the same class.” Thus, “ [s]tatutes aimed at protecting or regulating . . . [the insurer/insured] relationship, directly or indirectly, are laws regulating the 'business of insurance.’ ” Ibid, (emphasis added).

The Congress that passed McCarran-Ferguson was composed of neither insurance experts nor dictionary editors. Rather than use the technical term “underwriting” to express its meaning, Congress chose “the business of insurance,” a common-sense term connoting not only risk underwriting, but contracts closely related thereto.16 Since Congress knew of service-benefit policies, and viewed them as insurance, it would strain common sense to suppose Congress viewed con*248tracts necessary to effectuate those policies’ commitments as being outside the business it sought to exempt from the antitrust laws.

IV

The remaining question is whether the provider agreement in this case constitutes the "business of insurance.” Respondents contend that even if some contract between Blue Shield and the pharmacies is necessary, this one is not. Under the contract at issue, the druggist agrees to dispense drugs to Blue Shield’s insureds for a $2 payment, and Blue Shield agrees to reimburse the druggist for the acquisition cost of each drug so dispensed. The pharmacy is thus limited to a $2 "markup.” With support from the Court of Appeals, respondents argue that only the first half of the bargain is necessary for Blue Shield to fulfill its policy obligations. Those are fulfilled when Blue Shield binds the pharmacy to dispense the requested drug for $2. The second' half of the agreement, the amount Blue Shield reimburses the druggist, is assertedly irrelevant to the policyholder. As an alternative to the existing plan, the respondents and the Court of Appeals suggest that Blue Shield could simply pay the pharmacist his usual charge (minus the $2 paid by the policyholder). The present plan, which limits reimbursement to acquisition cost and freezes the markup at $2, is said to set a “fixed” price. From this premise respondents argue that such fixed-price plans are “anticompetitive,” and therefore not the “business of insurance.”

Respondents’ argument is directly contradicted by history. The service-benefit plans available when the McCarran-Ferguson Act was passed actually “fixed” more of the payment to their participating providers than does the plan here, which “fixes” only the markup. Those early plans usually paid established and equal amounts to their participating hospitals, rather than paying whatever each hospital charged. Rorem I, p. 64. Moreover, under the typical state enabling Act, those *249payments were subject to the approval of the state department of insurance.17 The 1937 Pennsylvania statute, for example, provided that “all rates of payments to hospitals made by such [service-benefit plan] corporations . . . and any and all contracts entered into by any such corporation with any hospital, shall, at all times, be subject to the prior approval of the Insurance Department.” 1937 Pa. Laws No. 378. Therefore, as insurer/provider fee agreements were part of the system of state regulation which the McCarran-Ferguson Act sought to preserve, there is no historical reason to exclude Blue Shield’s Pharmacy Agreements from the ambit of the exemption; there is instead a good historical reason for including them.

Nor does respondents’ claim that the Pharmacy Agreements are “anticompetitive” exclude them from constituting the “business of insurance.” The determination of whether Blue Shield’s Pharmacy Agreements actually involve antitrust violations or are otherwise anticompetitive has been held in abeyance, pending final decision as to whether the agreements fall within the scope of the McCarran-Ferguson Act. But even if the agreements were anticompetitive, that alone could not be the basis for excluding them from the “business of insurance.” An antitrust exemption by its very nature must protect some transactions that are anticompetitive; an exemption that is extinguished by a finding that challenged activity violates the antitrust laws is no exemption at all.

While this reason for excluding the Pharmacy Agreements from the circle of exempt provider agreements is unconvincing, there are substantial reasons, in addition to history, for including them within that circle. First, it is clear that the contractual arrangement utilized by Blue Shield affects its *250costs, and thus affects both the setting of rates and the insurer’s reliability. This is definitely a factor relevant to the determination of whether a transaction is within the “business of insurance.” See SEC v. National Securities, Inc., 393 U. S., at 460. See also Proctor v. State Farm Mutual Automobile Ins. Co., 182 U. S. App. D. C. 264, 561 F. 2d 262 (1977). True, that factor alone is not determinative, for as argued by the Court, innumerable agreements, including the lease on the insurance company’s offices, affect cost. This contract, however, has more than a mere incidental connection to the policy and premium. It is a direct arrangement to provide the very goods and services whose purchase is the risk assumed in the insurance policy. It is therefore integral to the insurer’s rate-setting process, as the correlation between rates and drug prices in a drug-benefits policy is necessarily high. Moreover, the ability of state insurance commissioners to regulate rates, an important concern of the Act, is measurably enhanced by their ability to control the formulas by which insurers reimburse providers.18 The same is true of state efforts to ensure that plans are financially reliable. See Travelers Ins. Co. v. Blue Cross of Western Pennsylvania, 481 F. 2d 80, 83 n. 9 (CA3 1973) (quoting the Pennsylvania Insurance Commissioner). This close nexus between the Pharmacy Agreements and both the rates and fiscal reliability of Blue Shield’s plan speaks strongly for their inclusion within the “business of insurance.” See generally Proctor v. State *251Farm Mutual Automobile Ins. Co., supra, at 271-272, 561 F. 2d, at 269-270.

Another reason, in addition to this nexus to basic insurance elements, also supports the conclusion that fixed-price provider agreements are the “business of insurance/’ Such agreements themselves perform an important insurance function. It may be true, as the Court contends, that conventional notions of insurance focus on the underwriting of risk. But they also include efforts to reduce the unpredictable aspects of the risks assumed. Traditional plans achieve this end by setting ceilings on cash payments or utilizing large deductibles. R. Mehr & E. Cammack, Principles of Insurance 222 (6th ed. 1976). Even if the insurer cannot know how often a policyholder might become ill, it can know the extent of its exposure in the event of illness. The actuarial uncertainty, therefore, is greatly reduced. A fixed-price provider agreement attempts to reach the same result by contracting in advance for a price, rather than agreeing to pay as the market fluctuates. The agreement on price at least minimizes the variance of the “payoff” variable, even if the probability of its occurrence remains an unknown. Indeed, if examined carefully, this function comes within the latter half of the definition of “underwriting” offered by the Solicitor General: “spread[ing] risk more widely or reducing] the role of chance events." See n. 12, supra. Of course, the Pharmacy Agreements in this case do not totally control “the role of chance” in drug prices since acquisition costs may fluctuate even if “markup” is fixed, but they are at least an attempt to reduce the role of chance to manageable proportions.19

Moreover, a service-benefit plan which “pay[s] the cost. . . whatever it might be,” as hypothesized by the Court of *252Appeals, 556 F. 2d, at 1381, would run grave risks of bankruptcy. Since it would expose the insurer to unknown liability, it would measurably increase the probability that an incorrect assessment of exposure would occur. This could lead to a failure to cover actual losses with premiums. Respondents argue that this fiscal-reliability problem could be solved by placing a dollar limit on benefits. But such a plan would be almost indistinguishable from a cash-indemnity policy. It would not be the full-service-regardless-of-price plan for which the policyholders bargained.20 The Pharmacy Agreements are thus “other activities of insurance companies relate [d] so closely to their status as reliable insurers that they too must be placed in the same class.” SEC v. National Securities, Inc., supra, at 460.

Y

The process of deciding what is and is not the “business of insurance” is inherently a case-by-case problem. It is true that the conclusion advocated here carries with it line-drawing problems. That is necessarily so once the provider-agreement line is crossed by holding some to be within the “business.” But that is a line which history and logic compel me to cross. I would hold that the concept of a provider agreement for benefits promised in the policy is within the “business of insurance” because some form of provider agreement is necessary to fulfill the obligations of a service-benefit policy. I would hold that these provider agreements, Blue Shield's Pharmacy Agreements, are protected because they (1) directly obtain the very benefits promised in the policy21 and therefore *253directly affect rates, cost, and insurer reliability, and (2) themselves constitute a critical element of risk “prediction.”22 The conclusion that these kinds of agreements are the “business of insurance” is that reached by every Court of Appeals except the Court of Appeals in this case.23

I would not suggest, however, that all provider agreements come within the McCarran-Ferguson Act proviso. Given the facts found by the District Court upon summary judgment, this is not a case where the petitioner pharmacies themselves conspired to exclude others from the market, and either pressured Blue Shield to go along, or were voluntarily joined by the insurer. See also Government Brief 13 n. 6. Such an agreement among pharmacies, itself neither necessary nor related to the insurer’s effort to satisfy its obligations to its policyholders, would be outside the “business of insurance.” An insurance company cannot immunize an illegal conspiracy by joining it. Cf. Parker v. Brown, 317 U. S. 341, 351-352 *254(1943). Moreover, since in this case the Blue Shield plan was offered to all San Antonio pharmacies and was in fact agreed to by at least 12, I am not called upon to decide whether an exclusive arrangement with a single provider would be so tenuously related to providing policyholder benefits as to be beyond the exemption’s protection. See generally Proctor v. State Farm Mutual Automobile Ins. Co., 182 U. S. App. D. C., at 270 n. 10, 561 F. 2d, at 268 n. 10.24

Finally, the conclusion that Blue Shield’s Pharmacy Agreements should be held within the “business of insurance” 25 *255does not alone establish whether the agreements enjoy an exemption from the antitrust laws. To be entitled to an exemption, petitioners still would have to demonstrate that the transactions are in fact truly regulated by the State, 15 U. S. C. § 1012 (b), and that they do not fall within the “boycott” exception of 15 U. S. C. § 1013 (b). The District Court held for petitioners on both issues. Neither issue was reached by the Court of Appeals, however, in light of its holding that the contracts were not the “business of insurance.” Accord*256ingly, I would reverse the judgment of the Court of Appeals and remand the case for further proceedings.26