Preferred Physicians Mutual Risk Retention Group (“Preferred”), plaintiff-appellee, is a risk retention group (“RRG”) chartered under Missouri law, which offers insurance to doctors practicing in the State of New York.1 Preferred is not licensed by the State of New York’s Commissioner of Insurance. It brought this action to challenge the lawfulness of New York’s Excess Insurance Law (“EIL”),2 because it gives a commercial advantage to insurers who are licensed in New York. Plaintiff contends principally that the federal Liability Risk Retention Act (“LRRA”), 15 U.S.C. § 3901-04, by exempting RRGs from most regulation by non-domieiliary states and barring discrimination against them, protects it from the injurious effects of the EIL. Judge Leisure granted summary judgment in plaintiff’s favor. For the reasons set forth below, we vacate and remand for further proceedings.
Background
RRGs are groups of risk-bearers, in this case anesthesiologists, who have banded together in an insurance cooperative to share liability. See 15 U.S.C. § 3901(a)(4). RRGs function pursuant to a federal statute, the 1986 LRRA. 15 U.S.C. § 3901-04. We have had occasion to review the LRRA’s history once before. See Insurance Company of the State of Pennsylvania v. Corcoran, 850 F.2d 88, 89-90 (2d Cir.1988). In brief, the LRRA was an expansion of the 1981 Product Liability Risk Retention Act (“PLRRA”), Pub.L. No. 97-45, 95 Stat. 949 (1981), which allowed RRGs to insure industry against product liability injuries only. Congress’s intention in passing the LRRA was to decrease insurance rates and increase the availability of coverage by promoting greater competition within the insurance industry. See H.R.Rep. No. 99-865, 1986 U.S.C.C.A.N. 5303, 5304-06 (report accompanying Risk Retention Amendments of 1986, Pub.L. No. 99-563, 100 Stat. 3170) [hereinafter “House Report”].
Under the MeCarran-Ferguson Act, 15 U.S.C. § 1012, the business of insurance is regulated by the states, rather than the federal government. Most states did not allow RRGs prior to the LRRA’s passage. However, the LRRA partially preempts state regulation of RRGs and prohibits states from excluding RRGs chartered under the laws of another state. See 15 U.S.C. § 3902(a)-(b)(exemptions from state law); Corcoran, 850 F.2d at 89-90 (describing history and functioning of LRRA); National Home Ins. Co. v. State Corp. Comm’n, 838 F.Supp. 1104, 1110 (E.D.Va.1993)(same). An RRG’s state of domicile may “regulate the formation and operation” of the group. 15 U.S.C. § 3902(a)(1). Non-domiciliary states retain some regulatory authority under the LRRA, but it is highly limited. National Home, 838 F.Supp. at 1111.
This ease requires that we determine whether New York’s EIL regulates or discriminates against Preferred in a manner prohibited by the LRRA. The EIL provides a substantial benefit to certain medical practitioners who obtain primary insurance coverage from a New York-licensed insurer. Under the EIL, physicians and dentists with hospital admitting privileges are provided— at no expense to the practitioners — a free layer of excess insurance coverage.3 Practitioners qualify for the insurance only if their primary coverage is with a New York-licensed insurer and the primary policy provides at least $1 million of coverage for each individual claimant, with a total protection of $3 million for all claims. See N.Y. Comp. *915Codes R. & Regs. tit. 10, § 91.1(a).4 Participating physicians receive excess coverage sufficient to provide them with total coverage of $2 million per claimant, and $6 million for all claims.
Because Preferred is not licensed in New York, physicians insured by it — or any other out-of-state RRG or insurer — are not eligible to receive the free excess insurance coverage provided by the EIL. Preferred brought this action in the district court against various New York State defendants. When the defendants’ moved for summary judgment, Preferred cross-moved for partial summary judgment. Preferred argued that the EIL violates the provisions of the LRRA which make RRGs “exempt” from any State law that either: (1) “regulate[s], directly or indirectly, the operation of a risk retention group,” 15 U.S.C. § 3902(a)(1),5 or (2) “otherwise diseriminate[s] against a risk retention group,” id. § 3902(a)(4).6 Preferred also claimed that the EIL impermissibly burdens interstate commerce. The district court found there to be no disputed issue of material fact, and granted Preferred’s motion for partial summary judgment. Subsequently, the trial judge ordered injunctive relief requiring, inter alia, that physicians insured by out-of-state RRGs be allowed to participate in the EIL program as if covered by state-licensed insurers. New York appeals.
Discussion
I. Indirect Regulation Under 15 U.S.C. § 3902(a)(1)
Preferred argues that the EIL constitutes indirect regulation of RRGs, and as such is prohibited by the LRRA. See 15 U.S.C. § 3902(a)(l)(exempting RRGs from any state law that would “regulate, directly or indirectly,” their operation) (emphasis added). It contends that, by granting a competitive advantage to regulated competitors, New York’s EIL effectively compels Preferred to submit to regulation by New York, and thus indirectly regulates it. The district court agreed, and granted Preferred’s motion for summary judgment. Summary judgment is, of course, appropriate only where there is no genuinely disputed issue of material fact, and the moving party is entitled to judgment as a matter of law. See Fed.R.Civ.P. 56(c). Without prejudice to any findings that the district court may make on remand, we think the district court’s determination was based on a misunderstanding of the showing required of Preferred in order to make out a claim of indirect regulation under the LRRA.
We agree with the district court that the LRRA’s preemption language is expansive. Both direct and indirect regulation of RRGs by non-domiciliary states are forbidden. 15 U.S.C. § 3902(a)(1). This expansiveness is emphasized by the structure of the Act, which sets forth a broad preemption followed by certain limited exceptions.7 See 15 U.S.C. § 3902(a)(1)(A-I); Corcoran, 850 F.2d at 91 (predecessor statute, PLRRA, “carved out specific exceptions to ... broad preemption”); State of Florida v. National Amusement Purchasing Group, Inc., 905 F.2d 361, 363 (11th Cir.1990)(same).
Furthermore, the legislative history of the Act makes clear that Congress intended to *916exempt RRGs broadly from state law “requirements that make it difficult for risk retention groups to form or to operate on a multi-state basis.” House Report, at 5305; see New York Conf. of Blue Cross & Blue Shield Plans v. Travelers Ins. Co., — U.S. -, -, 115 S.Ct. 1671, 1677, 131 L.Ed.2d 695 (1995)(consideration of text, structure, and legislative history of federal statute proper in determining whether state law is preempted). The House Committee Report explains that the Act is a response to the policies of many states that had “prevent[ed] insurance purchasing groups from achieving the advantages in rates and terms derived from the economic efficiency of collective purchasing that could contribute to resolving affordability and availability problems [in the insurance industry].” House Report, at 5305. Hence, the Congress “exempt[ed] risk retention ... groups from State law, in the respects specified in the [Act], in order to achieve [these] beneficial effects.” Id. at 5305-06.
Preferred urges us to consider the economic impact of the EIL on its ability to compete in the New York malpractice insurance market a form of indirect regulation forbidden under the LRRA. There is no disagreement between the parties that the EIL constitutes, in effect, a subsidy by New York to insurers licensed in that state who sell insurance to those EIL-eligible medical practitioners who are also in the market for excess coverage. Preferred argues, and the district court found, that this subsidy constitutes a “powerful incentive, bordering on compulsion, to out-of-state RRGs to receive New York accreditation.”
Were the evidence of economic impact well-developed, we would be required to determine whether the district court was correct in finding the effect of the subsidy on Preferred’s ability to compete in the New York medical malpractice market was sufficiently substantial to constitute — in essence — indirect regulation. But that question is not before us in answerable form. On the record as it now stands, the compulsive effect of the EIL on Preferred and other RRGs seeking to do business in New York is ambiguous.
To begin with, it is not clear from the record how much, if at all, the compulsive impact of this subsidy is counterbalanced by certain competitive advantages that New York argues RRGs possess because of their exemption under the LRRA from the bulk of New York’s insurance regulations. See 15 U.S.C. § 3902(a)(1).
Furthermore, the value of the subsidy received by New York licensed insurers under the EIL is unclear because it is dependent upon the importance to New York health care practitioners of the excess coverage provided. As to practitioners who would purchase excess insurance in any event, the subsidy is functionally the equivalent of the market value of the excess coverage policy. On the other hand, with regard to those practitioners who value the excess policy at less than its market value, and would therefore have otherwise not purchased excess coverage, the subsidy benefit received by the New York licensed insurer would be smaller. As to these customers, the value of the subsidy for the insurer would be the below-market price at which the insureds would have purchased excess insurance. And of course, if there are physicians who have no interest in receiving the excess coverage, and who therefore would not have purchased it at any price, as to them the New York licensed insurer receives no subsidy benefit at all.8
New York asserts on appeal that roughly two-thirds of all physicians licensed in New York do not participate in the EIL program. This fact, if true,9 may tend to support New *917York’s argument that physician RRGs covering medical malpractice risks are not coerced by the EIL into subjecting themselves to the state’s regulation as a condition of doing business in New York. The fact that Preferred has operated in New York in the face of the EIL without accepting regulation further tends to support New York’s contention. We are advised that no RRG has submitted to New York regulation. (The parties’ submitted papers do not, however, show the extent of operations in New York by unregulated RRGs.) On the other hand, if the EIL effectively forecloses RRGs from competing in a substantial submarket of New York’s physicians, that might constitute indirect regulation of RRGs.
Because the record is unclear, and the district court made no findings on these issues, we cannot determine whether the EIL has sufficient impact on RRGs to constitute indirect regulation. The mere existence of a competitive advantage to regulated insurers conferred by a New York State statute does not necessarily amount to indirect regulation. Whether unlicensed RRGs are either compelled to accept regulation in order to compete in New York or effectively foreclosed from a substantial submarket depends on facts not shown in the record. On this unclear record, we prefer not to venture to interpret the precise reach of the statute. We leave such interpretations to be made in the first instance by the district court in the light of its factual findings. We conclude that summary judgment should not have been granted in favor of Preferred.
New York argues in addition that its motion for summary judgment should have been granted. We disagree. On New York’s motion, Preferred was entitled to have the evidence viewed in the light most favorable to it, and to have the motion denied unless the evidence, seen in that light, left no issue of material fact and compelled judgment in New York’s favor. See Goenaga v. March of Dimes Birth Defects Found., 51 F.3d 14, 18 (2d Cir.1995); Olin Corp. v. Consolidated Aluminum Corp., 5 F.3d 10, 14 (2d Cir.1993). Preferred’s demonstration of the competitive advantage conferred on regulated insurers by the EIL sufficiently raises the issue of compulsion to accept regulation as to make summary judgment in New York’s favor inappropriate.
In sum, Preferred has put forward enough support for its claim of coercion to withstand New York’s motion for summary judgment. Equally, summary judgment in favor of Preferred is inappropriate on this record since New York has raised a material factual dispute as to the eoerciveness of the EIL. See, e.g., Cadbury Beverages, Inc. v. Cott Corp., 73 F.3d 474, 483-84 (2d Cir.1996)(denying on appellate review motion and cross-motion for summary judgment). We therefore remand for further consideration of Preferred’s claim under § 3902(a)(1).
II. Discrimination Under 15 U.S.C. § 3902(a) (b)
As an alternative ground for granting Preferred’s motion for summary judgement, the district court found that the EIL violated the LRRA’s general prohibition on discrimination against RRGs. See 15 U.S.C. § 3902(a)(4) (prohibiting any state from “otherwise discriminat[ing] against a risk retention group”). The question as to what constitutes discrimination against RRGs under the LRRA is one of first impression in this circuit. The district court ruled that because the EIL provides .advantages to licensed insurers that it does not provide to any unlicensed insurer, including to RRGs, it discriminates against RRGs for purposes of § 3902(a)(4). We think the question is more complex.
The word “discrimination” is of course inherently “ambiguous.” See Guardians Assoc. v. Civil Serv. Comm’n, 463 U.S. 582, 592, 103 S.Ct. 3221, 3227, 77 L.Ed.2d 866 (1983)(opinion of White, J.). The text and legislative history of the LRRA do little to illuminate Congress’s intent. However, we think — both as a matter of ordinary meaning and by analogy to similar uses of the term “discriminate” in other statutes — that dis*918crimination against a particular class is not proven merely by a showing that the class is part of a larger group that suffers some competitive disadvantage.
The concept of discrimination requires more. In its narrowest meaning it requires a showing that the contested act was done with the intent to discriminate against the protected class. The word is also sometimes employed, in particularly protective legislation, in a broader sense that is satisfied by a showing of “disparate impact” — that is to say, a harmful effect of a generally applicable policy that is disproportionately visited upon a protected class, regardless whether inspired by an intention to harm that class. Cf. Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 985-87, 108 S.Ct. 2777, 2783-85, 101 L.Ed.2d 827 (1988)(describing difference between Title VII “disparate treatment” and “disparate impact” analysis). In no case, however, is the concept of discrimination satisfied by a mere showing that the rule in question disadvantages a protected class, absent a further showing that the injurious practice either disparately effects the protected class, or was intended to injure it. Because the district court’s finding of discrimination in violation of § 3902(a)(4) was premised solely on the conclusion that RRGs (among others) are injured by the EIL, that finding must be vacated.
If Preferred were to demonstrate that the EIL was enacted with discriminatory intent, that would no doubt suffice to make out a claim of discrimination under the LRRA. Cf. Mears Transp. Group v. Florida, 34 F.3d 1013, 1018 (11th Cir.1994) (suggesting showing of discriminatory intent would be sufficient to make out discrimination claim under § 3902(a)(4)), cert. denied, — U.S. -, 115 S.Ct. 1960, 131 L.Ed.2d 852 (1995). Summary judgment is not appropriate on this point, however, as there is a material factual dispute as to whether the EIL was enacted with discriminatory intent. New York notes that the EIL was enacted in its initial form in 1985, prior to the passage of the LRRA. The State therefore suggests that the EIL could not possibly have been intended to diseriminate against RRGs, which had not existed in New York prior to the LRRA’s passage. Preferred rejoins that the EIL expires of its own terms each year, and so has been renewed many times since the passage of the LRRA with full knowledge of its effect on RRGs and with an intent to discriminate against them. These are matters for the consideration of the district court on remand.
Turning to the possibility that Preferred might prevail on a disparate impact theory, this claim raises questions which the parties did not address in their first round of litigation.
As a matter of law, we note that it is unclear whether the LRRA forbids only intended targeting of RRGs, or whether it also prohibits legislation with an unintended disparate impact.10 The Supreme Court has indicated that the application of disparate impact theory to a given statute’s anti-discrimination provisions is a troublesome question, requiring careful assessment of congressional purpose. See, e.g., Alexander v. Choate, 469 U.S. 287, 292-99, 105 S.Ct. 712, 715-19, 83 L.Ed.2d 661 (1985)(considering whether to apply disparate impact theory to § 504 of the Rehabilitation Act); cf. Washington v. Davis, 426 U.S. 229, 238-42, 96 S.Ct. 2040, 2047-49, 48 L.Ed.2d 597 (1976)(finding of discrimination under Equal Protection Clause generally requires showing of invidious intent).
We do not undertake to answer this question here, because the parties have failed to brief it adequately in their presentation to us, and because the issue may well be mooted by the district court’s findings on remand. If the district court finds on remand, after exploring the factual issues presented in the first part of this opinion, that the EIL indirectly regulates RRGs, the disparate impact question will be moot. Similarly, if Preferred proves able to demonstrate intentional discrimination, or fails to show the existence of facts that would satisfy a disparate impact standard, these events too would render it unnecessary for the district court to reach the question.
*919We note the existence of a factual dispute material to the question of whether Preferred, presuming the applicability of disparate impact analysis, can prevail on such a claim. To demonstrate that the EIL has a disparate impact on RRGs, Preferred would have to offer evidence as to the limited number of unlicensed non-RRG insurers in the New York market affected by the EIL, in order to show that the effect of the New York law falls disparately on RRGs rather than more generally on a wide group of unlicensed insurers doing business in New York.
We inquired about this matter at oral argument. In post-argument letter briefs submitted to us, the parties have advanced very different views. New York contends that the “mail order” exception to licensing freely allows unlicensed foreign insurers to write policies on New York risks, so long as they do not solicit in New York or enter the state to complete the transaction. N.Y. Ins. Law § 1101(b)(2). New York suggests that numerous foreign unlicensed entities other than RRGs are actively engaged in providing malpractice insurance to New York doctors and dentists in this fashion, and that the burden of the EIL meaningfully falls on them to the same extent as on RRGs, dispelling the contention that it disparately injures RRGs. Preferred’s letter brief suggests that the barriers imposed by New York’s regulatory scheme are so substantial that, among unlicensed issuers, only RRGs acting under the protection of the LRRA can effectively compete for New York business.
Because the EIL on its face does not burden RRGs any more severely than other foreign unlicensed insurers competing for New York business, Preferred will need to develop a factual record supporting its contention that no meaningful business is done by other unlicensed insurers if it is to prevail on a disparate impact claim, presuming that the statute allows for one at all.11
Conclusion
The judgment of the district court is vacated. The ease is remanded for further findings.
. By stipulation of the parties, U.S. Physicians Mutual Risk Retention Group has withdrawn from this case.
. The EIL was originally enacted in a slightly different form in 1985. See 1985 N.Y. Laws 294. Although the statute expires by its own terms each year, it has been repeatedly reenacted— with amendments not relevant to this dispute. See e.g., Act of July 1, 1994, 1994 N.Y. Laws 223; see also N.Y. Comp.Codes R. & Regs. tit. 10, Part 91 (1995).
.This coverage is funded, through a complex mechanism, by general hospitals and third-parly payers. See N.Y. Comp.Codes R. & Regs. tit. 10, Part 91. The details are not relevant to this appeal.
. A practitioner not covered by an individual primary policy may also participate in the program if insured through an approved hospital policy. See N.Y. Comp.Codes R. & Regs. tit. 10, § 91.1(a).
. This prohibition does not apply to an RRG’s state of domicile. See 15 U.S.C. § 3902(a)(1).
. The LRRA expressly exempts RRGs from:
any State law, rule, regulation, or order to the extent that [it] ... would—
(1) make unlawful, or regulate, directly or indirectly, the operation of a risk retention group except that the jurisdiction in which it is chartered may regulate the formation and operation of such a group and any State may require such a group to [comply with certain specified requirements]____
(4) otherwise discriminate against a risk retention group or any of its members, except that nothing in this section shall be construed to affect the applicability of State laws generally applicable to persons or corporations.
15 U.S.C. § 3902(a).
.For example, RRGs may be required by a non-domiciliary state to "comply with unfair claim settlement practices law,” 15 U.S.C. § 3902(a)(1)(A), “pay, on a nondiscriminatory basis, applicable premium and other taxes which are levied on admitted insurers,” id. at § 3902(a)(1)(B), and “register with and designate the State insurance commissioner as its agent” for service of process, id. at § 3902(a)(1)(D).
. The New York licensed insurer also receives no subsidy benefit with regard to those physicians who are ineligible for excess coverage under the EIL for reasons unrelated to their primary coverage, for example because they lack hospital admitting privileges.
. These materials are not formally in the record, and we note them purely for purposes of illustration. We recognize that Preferred may be able to rebut the significance of this figure. New York’s numbers might include physicians who are retired, not practicing in New York, or do not possess hospital admitting privileges. Furthermore, even if only some portion of physicians (perhaps only those practicing in relatively high-risk areas of specialty), find New York licensed insurers preferable because of the EIL coverage *917that results, this may, as noted below, still constitute indirect regulation of a sub-market within the medical malpractice insurance industry. These are matters for the district court to consider on remand.
. We recognize that the Eleventh Circuit’s ruling in Mears may be read to suggest the inapplicability of disparate impact analysis to § 3902(a)(4) claims. See Mears, 34 F.3d at 1018.
. The district court also ruled that summary judgement in favor of Preferred was appropriate because the EIL impermissibly burdens interstate commerce. The district court expressly based its ruling on the determination that the EIL violates the LRRA. Because we have determined that this finding was inappropriate, we remand for further consideration of Preferred's Commerce Clause claim by the district court in the first instance.