Primer: ERISA Preemption

Executive Summary

  • The Employee Retirement Income Security Act of 1974 (ERISA) limits the ability of states to regulate many employer-sponsored health insurance plans for the purpose of avoiding duplicative or conflicting (and thus costly) regulations that would make it more difficult for employers to offer health insurance to employees.
  • Several Supreme Court cases, including a recent 2020 case, have struggled to provide clarity around the limits of ERISA’s preemption, leading to routine challenges to the law by states and countersuits by employers over the past several decades.
  • To maintain the statute’s protection of ERISA plans, Congress should clarify the extent to which ERISA preempts state laws, which would provide more certainty for employers and beneficiaries.


The Employee Retirement Income Security Act of 1974 (ERISA) was created as the result of several scandals surrounding employee pension funds in the 1970s.[1] ERISA was meant to provide uniform national standards for reporting and accountability for employee benefit plans and to protect them from potential loss or abuse. In addition to pensions, Congress also placed regulation of self-insured employer-sponsored insurance (ESI) plans under ERISA and preempted them from state control in many instances. This preemption was designed to encourage employers to offer health benefits to employees without the difficulty of dealing with numerous, and potentially conflicting, state regulations on health insurance, thereby reducing the cost of offering health benefits.

The ERISA preemption is neither absolute nor sharply defined in statute. Several Supreme Court cases over the past several decades have failed to provide a consistent and clear test on the limits of ERISA preemption. This statutory ambiguity has resulted in numerous attempts by states to legislate and regulate ERISA plans, ultimately eroding ERISA preemption and increasing costs for these plans and their beneficiaries. This primer explains how ERISA preemption works, the importance of preemption to the U.S. health care system, and how the courts have interpreted ERISA preemption. Finally, it examines the need for congressional action to clarify the extent to which ERISA preempts state laws, which would provide more certainty for employers and beneficiaries by preventing further state attempts to erode ERISA.


As critical as the ERISA preemption clause is, it takes up only a few lines of statute. The clause, 29 USC 1144(a), reads: “…the provisions of this subchapter and subchapter III shall supersede any and all State laws insofar as they may now or hereafter relate to any employee benefit plan described in…this title.” The statute goes on to clarify that states are prevented from defining self-funded health insurance plans offered by employers or unions as “insurance plans” under their jurisdiction. All other state laws regulating insurance plans are not preempted.

Essentially, ERISA exempts applicable insurance plans from state insurance laws and specifies that those applicable plans include all self-insured private-sector employer and union plans, with some exceptions (notably for multiple employer welfare plans, governmental plans, and church plans). Self-insured health insurance plans are defined as those in which the sponsoring organization takes on the full risk of medical costs for beneficiaries (and are thus “established” and/or “maintained” by the organization). In contrast, fully insured plans are those plans where a third-party insurance company takes on the risk of beneficiary costs and are therefore not covered by ERISA. An organization may contract administration of their self-insured plans to third-party insurance companies, but the establishing organization of that self-insured plan is still on the hook for all monetary risk.

ERISA incentivizes employers to provide benefits packages by avoiding numerous regulations in one or multiple states, thereby reducing costs to both the employer and the beneficiary. This has led to ERISA plan-provided coverage for over 133 million people in 2020, according to the Department of Labor, and coverage of 65 percent of workers with ESI.[2],[3] These plans are expensive for organizations, and are only sustainable for employers provided the incentives the plans provide in attracting and retaining employees and members outweigh the costs of providing the benefits. Health insurance is an important part of both recruitment and retention: One survey reported that 51 percent of employees stay at jobs based on their health insurance benefits, while another found that 96 percent of employees believed it was important for employers to offer health insurance and that 52 percent would not accept a job that did not offer health insurance.[4],[5]

Despite ERISA preemption’s importance to the U.S. health care system, the preemption clause is built of ambiguous language: “all State laws insofar as they may now or hereafter relate to any employee benefit plan.” The ambiguity of the phrase “relate to” has resulted in numerous Supreme Court cases over four decades. Indeed, the Supreme Court has struggled to come up with a clear and consistent test of when a state law “relate[s] to” an employee benefit plan, in large part because the Court is hemmed in by the non-specific statutory language. The justices have attempted to interpret and reinterpret Congress’ broad language over the years to define how laws can and cannot relate to ERISA plans. As the cases below demonstrate, the courts (by one justice’s admission) add little in the way of clarity and, ultimately, Congress will need to act to clarify the “relate to” language in statute.

Court Decisions

Shaw v. Delta Air Lines, Inc. (1983)[6]

New York’s Human Rights Law and its Disability Benefits Law lay at the center of this case involving a requirement for pregnancy coverage prior to the passage of the federal Pregnancy Discrimination Act of 1978. The Court held that, under the plain language and history of ERISA, the Human Rights Law “relate[s] to any employee benefit plan,” but is preempted under ERISA only if the state law prohibits practices considered legal under federal law. The Disability Benefits Law was ruled to not be preempted. Of note, the Court ruled that states could not mandate specific types of benefits in ERISA plans (although they could mandate separate plans such as a disability plan). Additionally, the majority opinion stated that a “law ‘relates to’ an employee benefit plan…if it has a connection with or reference to such a plan.” While Shaw affirmed the idea that plan design is generally off-limits for state laws, it also contained the first of the Court’s many attempts to define “relate to” – in this case, with the equally vague “connection with or reference to” phrase (both of which arguably mean the same thing as “relate to”) that provided fodder for future cases.

Ingersoll-Rand Co. v. McClendon (1990)[7]

In McClendon, a former employee sued his employer for wrongful termination and claimed it was an attempt to prevent him from attaining his employee benefits. Texas courts sided with the plaintiff, but the Supreme Court ultimately ruled that the existence of a plan was essential to establishing liability under Texas law and therefore the law “relate[d] to” an ERISA plan and was thus preempted. Important for our purposes, the Supreme Court emphasized in the McClendon ruling that Congress’ goal in enacting ERISA was “to ensure that plans and plan sponsors would be subject to a uniform body of benefits law” and thus “minimize the administrative and financial burden of complying with conflicting directives.” The Court laid out that it will base rulings not just on the direct language of the law but also its understanding of Congress’ intent in passing ERISA (“ensure…a uniform body of benefits law”). The Court then added that intent includes minimizing a “financial burden.” What constitutes a “financial burden” was left undefined, and this led to further points of contention between states and plan sponsors.

New York State Conference of Blue Cross & Blue Shield Plans v. Travelers Insurance Co. (1995)[8]

A New York statute required hospitals to charge more (up to 13 percent) for patients not covered by a Blue Cross/Bule Shield plan, as well as place additional surcharges on health maintenance organizations. The Supreme Court ruled that this “indirect economic influence” did not “bind plan administrators to any particular choice,” and thus did not create an impermissible connection between ERISA plans and the statute in question. Essentially, the ruling stated that imposed costs did not inherently interfere with the goal of ERISA which, according to the Court, is to enable uniform plan design and administration but not to ensure cost uniformity. The Court stated that the “relate to” language in ERISA was not intended to nullify “the starting presumption that Congress does not intend to supplant state law.” State laws and regulations that only increase costs (to a degree) or alter incentives for ERISA plans, but do not force them into specific coverage arrangements, are not preempted. The Court clarified, however, that it “does not hold that ERISA pre-empts [sic] only direct regulation of ERISA plans. It is possible that a state law might produce such acute, albeit indirect, economic effects as to force an ERISA plan to adopt a certain scheme of coverage or effectively restrict its choice of insurers.” The Court said the “relate[s] to” clause wasn’t intended to replace state law, but it did not provide a clear line of when ERISA preemption does or does not supplant state law. Increasing the confusion, the Court seemed to say on the one hand that imposed costs do not trigger ERISA preemption, but then on the other hand it somewhat-contradictorily stated the economic effects of imposed costs may trigger ERISA preemption. What’s more, it gave no clarity in its ruling as to when the line of “economic effects” is crossed, nor when a plan design becomes “forced.” Despite the clear “financial burden” (see McClendon) this New York law placed on ERISA plans, it was apparently not enough to produce an “acute…economic effect.”  The Travelers ruling has been one of the most cited ERISA cases and had a major impact on the Rutledge ruling – to be mentioned later – but is also a primary example of the multiple tests the Court adopted for preemption that lack precise definitions and boundaries.

California Division of Labor Standards Enforcement v. Dillingham Constr., N.A., Inc. (1997)[9]

A California public works contractor had an unapproved apprenticeship program and was paying sub-prevailing wage to the participants (which would have been allowed had the program been approved). The Supreme Court held that, because the law applied to all apprenticeship program sponsors and only incentivized participation, nothing in the apprenticeship program’s law or regulations related to ERISA plans in a meaningful way. While not health care related, this case’s importance stems from the explicit statement by the Court that it would look “both to ERISA’s objectives as a guide to the scope of state law that Congress understood would survive…and to the nature of the law’s effect on ERISA plans.” But because of the vagueness of the statute, the Court attempted to divine congressional objectives with little guidance. Additionally, the Court did not define the parameters of “the nature of a law’s effect” in a meaningful way – yet another ambiguity. The Dillingham ruling also held that a state law makes “reference to” (see Shaw) ERISA plans when the law “acts immediately and exclusively upon ERISA plans,” and is thus preempted by ERISA statute. The Court held that ERISA may also preempt state laws if there is a “connection with” (see Shaw) ERISA plans in that law. Essentially, this creates another two-fold test for preemption: State laws need to directly target ERISA plans or have a “connection with” (as laid out in Travelers) ERISA plans. This test is inherently confusing: There isn’t an unambiguous binary inflection (e.g. if it ruled that a law must directly single out ERISA plans), but instead the test contains elasticities (“connection with”/”reference to”) which provide little clarity as to its actual application.

Egelhoff v. Egelhoff (2001)[10]

A beneficiary who divorced his wife (previously the beneficiary of his life insurance policy and pension plan) died without a will. His children from a previous marriage filed suit that they should, under Washington state law, be the recipients of the assets. State law provided an automatic de-designation of a spouse as beneficiary to assets, including life insurance policies and employee welfare benefit plans, upon divorce. The Court held that because the statute directly affected the administration of a benefit plan – in this case, over how and to whom benefits are paid out – the statute was preempted by ERISA. Critically, the Court noted that the statute interfered with “nationally uniform plan administration,” which it called “one of the principal goals of ERISA.” In Egelhoff, the Court affirmed McClendon in that it again looked toward its understanding of congressional intent to divine the principal goals of ERISA. While the Court’s understanding of the goal of ERISA is not in dispute, the exact meaning of “nationally uniform plan administration” is. Many plan sponsors argue that costs imposed by state laws inherently affect nationally uniform plan administration, as those costs determine the boundaries of plan design (an inherent part of plan administration) and potentially limit options for plan design. Again, in this case, the Court struggled to produce definite terms in its rulings.

None of this is to say the Court was incorrect in any of these rulings. The Court’s rulings are constrained by vague statutory language, and the Court is forced to both attempt to divine congressional intent from ambiguous statute language and to try to provide a consistent test that lower courts may easily follow. These tests center around the meaning of “relate to” with terms that mean functionally the same thing – “connection with,” “reference to,” etc. – and require their own interpretations and litmus tests. As the numerous different and conflicting lower court rulings on each case demonstrate, none of these litmus tests has provided any consistency.

The Rutledge v. Pharmaceutical Care Management Association Ruling and Its Implications


In 2015, Arkansas passed a law, Act 900, that required pharmacy benefit managers (PBMs) to reimburse pharmacies at or above the pharmacies’ acquisition cost for prescription drugs. The Pharmaceutical Care Management Association (PCMA), the trade industry group for PBMs, sued on the grounds that such price-setting is preempted by ERISA. In December 2020, the Supreme Court issued an 8-0 verdict in Rutledge holding that Act 900 was not preempted by ERISA.[11] The Court relied heavily on (and reaffirmed) the Travelers ruling, but the previous cases mentioned above all played a role. How the Court laid out its reasoning will inform future cases and provides important points of consideration for Congress as it decides on potential reforms for ERISA.

The Court’s ruling emphasized that “not every state law that affects an ERISA plan…has an impermissible connection with an ERISA plan. That is especially so if a law merely affects costs.”[12] Citing Travelers, the Court noted that the New York law relevant in that case was more intrusive than Arkansas’ Act 900. The New York law at issue in Travelers clearly benefited certain plans, while Act 900 applied to all PBMs and pharmacies in the state. While the Court acknowledged that any cost increases would likely be passed on to plans and thus increase the amount ERISA plans pay for prescriptions relative to other states, it cites Travelers in noting that “cost uniformity was almost certainly not an object of preemption.”[13]

The Court also reaffirmed Dillingham, noting that Act 900 “does not ‘refer to’ ERISA.” Act 900 therefore passed the two-part test the Court established for states to avoid preemption – it neither specifically targets ERISA plans nor does it have an impermissible connection to them. The Court noted that Act 900 does not directly regulate health plans, just PBMs, and makes no specific or exclusive reference to ERISA plans. This last point is important – the Court distinguishes a third-party administrator (TPA), in this case the PBM, from the plan itself. This logic has the potential to then open all TPAs to regulation, even if they are an integral part of plan design and administration.

It is worth reviewing Justice Clarence Thomas’ concurring opinion in Rutledge, as it laid out the problems the Court has had in attempting to clarify the preemption statute. Justice Thomas criticized the Court’s attempts to define “relate to” through the lens of the congressional objectives of ERISA – arguing this has created even more confusion. He argued that the impermissible connection test, where a law has a connection if it governs a central matter of plan administration or interferes with nationally uniform plan administration, “has proven just as difficult to apply consistently.”[14] Justice Thomas did not fault the decisions of lower courts that the Supreme Court overturned, as they are merely “apply[ing] the amorphous test that we gave them.”[15] While Justice Thomas was primarily concerned with the Court’s interpretations, his concurring opinion speaks to the fundamentally problematic ambiguity in ERISA’s preemption statute as written by Congress.


Rutledge, while not earth-shattering, reaffirms the idea that state laws and regulations that impose additional, indirect costs on ERISA plans (as long as those costs are not exclusive to ERISA plans) are not preempted by ERISA. While the Court did state in Travelers that state laws “might produce such acute, albeit indirect, economic effects as to force an ERISA plan to adopt a certain scheme of coverage or effectively restrict its choice of insurers,” it is clear that the Court has a very high bar for what meets this standard. The New York law in Travelers imposed a 13 percent fee on hospital bills for non-Blue Cross Blue Shield plans, but if such a large fee is not considered “acute,” very few realistic economic effects on plans would be. Unlike in Travelers, however, Act 900 did not directly dictate a cost. Instead, Act 900 functionally forced PBMs to change the way they calculate payments to pharmacies to a method that would likely increase costs for PBMs (and thus plans and beneficiaries). Payment calculations are part and parcel of plan design and administration – but the Court has now ruled that laws dictating how those calculations must be done are functionally cost impositions and not relevant to plan governance or a nationally uniform plan design and administration.

Additionally, the Court’s claim that “ERISA plans are likewise not essential to Act 900’s operation,” as Act 900 only concerns PBMs, creates a separation between TPAs and the plans they to which administer. Yet self-funded plans, which are secured by organizations who are not insurers, must rely on TPAs to function. The average company is simply not capable of negotiating its drug prices and provider networks, ensuring payments are made and rebates are collected, and a wide variety of other tasks that come with administering a health plan. As such, TPAs are essential to a plan’s operation, and the regulation of a TPA seems to functionally regulate plan design. It is, after all, the TPA that handles the benefit design and plan administration, while the employer is the one that takes the financial risk.

None of which is to say that the Court got the Rutledge ruling wrong based on current statute. Rather, it is to highlight ERISA’s statutory ambiguity that limits the Court’s ability to interpret the law’s preemption clause in clear and consistent ways. Nothing in the current statute specifically identifies TPAs as a critical part of a plan’s administration. Indeed, TPAs such as PBMs were not used by plans to any large extent until over a decade after the law was written.

While Rutledge generally served as a reaffirmation of previous cases, its largest impact may be that it has created a perception of weakness in the ERISA preemption statute by implicitly suggesting that TPAs’ benefit design and administration processes are not inherent to the plan itself. Proof of this perception can be seen in the wide variety of new legislation states have introduced and passed over the last several years impacting ERISA preemptions through regulations on TPAs. States clearly believe that Rutledge potentially provides a broader window for them to regulate and control ERISA plans indirectly, creating additional requirements, and therefore costs, for ERISA plans.

Rutledge will affect future cases before the Supreme Court, specifically the potential PCMA v. Mulready case, which has not yet been granted certiorari by the Court.[16] In Mulready, PCMA is suing Oklahoma over a 2019 law that required PBMs to cover the vast majority of pharmacies in a given coverage area, not restrict a patient’s choice of an in-network pharmacy, not deny from preferred status any in-network pharmacies who are willing to accept the same conditions as other preferred providers, and prevented PBMs from denying, limiting, or terminating a provider’s contract based on probation status. Various lower court rulings have been split, with the most recent decision from the appellate court ruling in favor of PCMA and rejecting Oklahoma’s argument that the law regulates PBMs, not insurance plans, and applies to all PBMs regardless of whether they administer to ERISA plans. The Oklahoma law imposes requirements on plan design and administration, but with Rutledge having implied that PBMs are not part of the plan, there is the potential for the Court to rule in Oklahoma’s favor given that these regulations apply only to PBMs regardless of whether the plans they administer to are self-insured.

The Need for Congressional Action

These Supreme Court rulings indicate the need for Congress to clarify and update the ERISA statute to remove ambiguities surrounding the “relate to” clause. The courts have routinely failed to provide a clear, consistent test for when state laws relate to self-funded health insurance plans, as evidenced by the Supreme Court’s many rulings on the issue. While Congress wrote the ERISA preemption statute to incentivize employers to offer health insurance by means of reducing the costs associated with state regulation, previous Supreme Court rulings have muddied this feature of the law and created a perception of weakness in the statute that states have begun to probe through increased regulation of TPAs. Congressional action is needed to clarify ERISA’s preemption clause and prevent unnecessary and costly regulatory burdens that threaten health insurance offerings by private companies.


ERISA’s preemption clause is a crucial part of our nation’s private health care system. It creates a statutory environment that allows employers to provide health insurance to employees in a more efficient and ultimately more economically feasible manner. Yet vague statutory language in ERISA has led to decades of court cases, prompting attempts by the courts to reinterpret ERISA with little congressional guidance. The Rutledge ruling opens the door for more state intervention in ERISA plans by redefining what counts as plan administration and design. Congressional action to clarify ERISA is needed to avoid potential end-problems with increased state regulation, namely more expensive and less attractive plans for beneficiaries. A forthcoming paper will further explore these problems, as well as potential solutions for Congress to consider.






[6] Shaw v. Delta Air Lines, Inc., 463 U.S. 85 (1983),

[7] Ingersoll-Rand Co. v. McClendon, 498 U.S. 133 (1990),

[8] New York State Conference of Blue Cross & Blue Shield Plans v. Travelers Ins. Co., 514 U.S. 645 (1995),

[9] California Div. of Labor Standards Enforcement v. Dillingham Constr., N. A., Inc., 519 U.S. 316 (1997),

[10] Egelhoff v. Egelhoff, 532 U.S. 141 (2001),

[11] Rutledge v. Pharmaceutical Care Management Association, 592 U.S. ___ (2020),

[12] Ibid.

[13] New York State Conference of Blue Cross & Blue Shield Plans v. Travelers Ins. Co., 514 U.S. 645 (1995),

[14] Ibid.

[15] Ibid.