With its “employer mandate”—i.e., the requirement that applicable large employers make an offer of group health coverage to substantially all full-time employees or face the prospect of a penalty—the Affordable Care Act (ACA) opened a fault line in the previously monolithic market for group health insurance. There is large cohort of American workers who, before the ACA, were not offered major medical coverage under an employer-sponsored group health plan. These employees are sometimes referred to as the “contingent” workforce. They include part-time, seasonal and temporary employees, as well as employees whose work schedules are generally irregular or intermittent. Found predominantly though not exclusively in industries such as staffing, restaurants, media and advertising, transportation and hospitality, among others, these workers tend to be on the lower end of the pay scale. They also often have significant “deferred” health issues (a euphemism for undiagnosed conditions owing to lack of previous access to health care). The ACA provided “applicable large employers” (those with 50 or more full-time and full-time equivalent employees) with an incentive to cover these workers.

While the ACA requires mainstream health insurance carriers to issue policies to employers who cover large cohorts of contingent workers under the guaranteed issue and renewability rules, carriers remain free to set rates, which are in many cases unaffordable. This market is what the insurance industry refers to as a “hard” market. i.e., one with higher insurance premiums and more stringent underwriting criteria that together results in less competition and carriers writing fewer policies.

Large employers (over, say, 500 to 1,000 employees) with substantial populations of contingent workers can self-fund coverage for the purpose of complying with the ACA employer mandate. (For a description of the mechanics and benefits of self-funding of employer-sponsored group health plans, please see the white paper published by Spring Consulting Group, LLC entitled, Self-funding and the Management of Risk.) These employers benefit from the “law of large numbers” that holds that, in the group health plan context, the larger the number of lives independently exposed to loss, the greater the probability that actual loss experience will equal expected loss experience. In other words, the credibility of data increases with the size of the data pool, thereby leading to predictability. Similarly situated mid-sized employers (between 50 and 500 or so full-time and full-time equivalent employees) are unable to rely on this axiom, so their options for complying with the ACA are limited. They could, for example, offer a plan that required no underwriting such as a “MEC” that covers preventive services only. This approach gets the mid-sized employer off the hook for the more serious of the two levels of penalties (i.e., the excise tax imposed under Internal Revenue Code section 4980H(a)), but it leaves open the possibly of exposure to the other penalty (under Internal Revenue Code section 4980H(b)). To escape the latter penalty the employer must generally make an offer of major medical coverage, or in the parlance of the ACA, coverage that provides “minimum value.”

Association-style plans are also an option for mid-sized employers, but self-funding is generally not an option, since most states view self-funded association plans as unlicensed insurance companies. There is another, less conventional option: though not a new concept, these mid-sized employers can self-fund, but interpose a separate layer of stop-loss coverage under a “group captive” arrangement that includes other, similarly situated employers. In a group captive, each participating employer maintains its own self-funded health group health plan. Stop-loss coverage is purchased from a commercial medical stop-loss carrier that cedes a portion of the risk to a group captive insurer. The result is a transfer of responsibility for payment of claims that fall within a specified corridor immediately above the employer’s retained risk up to a preset amount. The reinsurance agreement also limits the total amount of claims the captive is responsible for paying, thereby limiting the captive’s aggregate exposure.

In our experience, these group captive programs have become, if not ubiquitous, at least commonplace. Moreover, they are being deployed by mid-sized employers of all stripes, not just those seeking to cover contingent workers. This post explains the basic features of these group captive arrangements, reviews and assesses the legal authority on which they rely, and provides compliance recommendations.

Background and Scope

The term “captive” insurer traditionally referred to “single-parent” captive, which is a subsidiary of a parent company that insures the risks of its parent. Single-parent captives offer certain tax and risk management advantages. Historically, single-parent captives insured property and casualty risks and workers’ compensation, but they have increasingly been pressed into service to cover health risks. In this latter case, issues arise under the rules governing “prohibited transactions” under the Employee Retirement Income Security Act (ERISA).

While a full-blown discussion of the ERISA prohibited transaction rules as they apply to captives is beyond the scope of this post, the use of group captives can implicate these rules. Briefly, ERISA prohibits the “transfer to, or use by or for the benefit of, a party in interest, of any assets of the plan.” (Emphasis added). Employer-sponsors of plans are parties in interest pursuant to ERISA, as are entities of which they own more than 50 percent. Thus single-parent captives, which are usually 100 percent-owned by the employer must navigate the ERISA prohibited transaction rules. And while not entirely clear, it is a safe bet that the same result would apply to a captive cell where the employer is entitled to more than 50 percent of the profits.

Key to the use of group captive arrangements is the concept of what constitutes “plan assets” for purposes of the prohibited transaction rules. ERISA specifically provides that, where an insurer issues a guaranteed benefit policy to a plan, such as a group term life policy, the assets of the plan include the policy, but not any assets of the insurer. Thus, if a self-funded plan purchases stop-loss coverage, the premium ceases to be a plan asset once it has been received by the insurer. But this is not the case of reinsurance using captives. This issue surfaced in connection with a 1979 prohibited transaction class exemption (Prohibited Transaction Class Exemption 1979-41 (August 7, 1979)). This class exemption permitted insurance companies to underwrite their own employee benefits provided that the employee benefits business did not exceed 50 percent of the insurance company’s business. Notably, the Department of Labor was unwilling to extend this class exemption to reach reinsurance provided by wholly-owned captives. According to the Department [44 Fed. Reg. 46365, 46368]:

[I]t is the Department’s view that if a plan purchases an insurance contract from a company that is unrelated to the employer pursuant to an agreement, arrangement or understanding, written or oral, under which it is expected that the unrelated company will subsequently reinsure all or part of the risk related to such insurance with an insurance company which is a party in interest of the plan, the purchase of the insurance contract would be a prohibited transaction.

In other words, any amounts deposited with a captive retain their status as “plan assets” that are subject to the ERISA prohibited transaction rules. As we explain below, group captives generally endeavor to side-step the ERISA prohibited transaction rules by purchasing stop-loss coverage and making group captive contributions with employer assets rather than plan assets.

In contrast to a single-parent captive, a group captive—also referred to as an “association captive”—is a legal entity jointly owned by a group of unrelated companies, and it is formed primarily to insure the risk of its member-owners. Group captives can be further subdivided into heterogeneous and homogeneous, the former covering dissimilar industries and the latter covering similar industries. In each case, the goal is to permit mid-sized employers to replicate the risk profile of a single large employer, and both types of group captives are commonplace.

Captives have caught the attention of both the Federal and state regulators on three fronts unrelated to this post.

Low attachment points

A stop-loss insurer might offer insurance policies with attachment points set so low that the insurer assumes nearly all the employer’s claims’ risk. For example, the attachment point could be set at $5,000 per employee, or $100,000 for a small group. While a plan might purport to be self-funded under these circumstances, the arrangement looks much more like a fully-insured, high-deductible health plan. The National Association of Insurance Commissioners (NAIC) has promulgated a model law that prohibits the sale of stop-loss insurance with a specific annual attachment point below $20,000. For groups of 50 employees or fewer, the aggregate annual attachment point must be at least the greater of (i) $4,000 times the number of group members, (ii) 120 percent of expected claims or (iii) $20,000. For groups of 51 employees or more, the model law prohibits an annual aggregate attachment point that is lower than 110 percent of expected claims. The group captive arrangements that we have encountered uniformly satisfy these standards.

Micro-captives/abusive tax shelters

A micro-captive arrangement is one in which a taxpayer endeavors to reduce his or her aggregate taxable income using a combination of an insurance contract and a captive insurance company. Each party claims deductions for insurance premiums, and the captive insurance company elects to be taxed only on investment income, thereby excluding payments it directly or indirectly received under the contracts from its taxable income. The IRS in Notice 2016-66 said the manner in which the contracts are interpreted, administered and applied is inconsistent with arm’s-length transactions and sound business practices. The group captives that we encounter do not take this approach.

Fully-insured arrangements that include reinsurance (Rev. Rul. 2014-15)

IRS Revenue Ruling 2014-15 describes and sanctions use of a captive to reinsure fully-insured health benefits. The ruling describes an arrangement in which an employer makes contributions to a funded welfare trust to provide health benefits to certain retirees and their dependents. The trust then purchases insurance from a commercial carrier, which cedes a portion of the risk to a captive 100 percent owned by the employer. Thus, this arrangement is similar to the group captive arrangements, except that the arrangement is fully insured. It is also worth noting that the employer was required, as a one of the conditions for approval of the arrangement, to obtain a prohibited transaction exemption from the Department of Labor.

Group Captive Structure and Administration

As we explained above, each participating member of a group captive establishes and maintains its own, self-funded group health plan. As a result, each employer can dictate its own plan rules, including levels of coverage such as copays and deductibles. Participating self-funded employers can also choose their own third-party administrators and provider networks, although in our experience, this rarely happens. Each employer also purchases commercial medical stop-loss coverage.

Stop-loss coverage reimburses claims when they exceed a specified amount, referred to as the “attachment point.” Coverage may be “specific,” i.e., covering individual claims, or “aggregate,” i.e., covering total annual claims. An employer may, for example, purchase an aggregate stop-loss policy that covers claims over 125 percent of total anticipated annual claims. In contrast, specific stop-loss coverage may reimburse any individual claim in excess of $50,000.

In a group captive arrangement, the employer enters into a reinsurance agreement with a commercial medical stop-loss carrier, under which responsibility for payment of claims in excess of the level of retained risk separately agreed to by each employer is transferred to the carrier. The medical stop-loss carrier then transfers a portion of the risk to the employer/sponsor’s captive pursuant to an “insurance treaty.” The arrangement is analogous to the securitization of mortgage securities in that the excess risk (that is the risk over and above the anticipated claims and any additional retained risk) is carved up in tranches. Under the reinsurance treaty, the captive is responsible for the tranche immediately above the employer/sponsor’s retained risk up to a pre-set amount. For example, in the case of specific coverage, the captive’s exposure might start at $50,000 per claim and go to, say, $500,000. The commercial medical stop-loss carrier would them be responsible for a claim that exceeds $500,000. In the case of aggregate coverage, the captive’s tranche might pay between 125 percent and 500 percent of total annual claims. Amounts above 500 percent would be the responsibility of the commercial medical stop-loss carrier.

While there are likely many ways to structure group captives, we routinely encounter two. Under the first approach, the group captive consists of a series of “fronted” captive cells that are sponsored by the commercial carrier from who the stop-loss coverage is purchased. Under the second approach, the group captive is separately maintained and subscribed to under an enabling state law. Both approaches accomplish the pooling of stop-loss coverage at the tranche of risk underwritten by the captive.

In one common model, each employer/member selects the level of its retained risk, which informs the level of the premium paid for reinsurance. (The greater the retained risk, the lower the cost of reinsurance coverage.) The employer/member then pays a premium, a portion of which is applied to the captive’s shared risk pool. If there are underwriting gains in the shared risk pools that exceed claims in a year, the excess is returned to the employer/members pro rata based on premiums and without regard to the employer/member’s individual experience. Going forward, each employer/member’s experience informs the following year’s premiums.

The MEWA Question

ERISA defines the term “multiple employer welfare arrangement,” or “MEWA,” in relevant part, as follows:

(A) The term “multiple employer welfare arrangement” means an employee welfare benefit plan, or any other arrangement (other than an employee welfare benefit plan) which is established or maintained for the purpose of offering or providing any benefit described in paragraph (1) [welfare plan benefits] to the employees of two or more [unrelated] employers . . ..”

(For a comprehensive explanation of MEWAs, please see the paper published by the U.S. Department of Labor entitled, Multiple Employer Welfare Arrangements under the Employee Retirement Income Security Act (ERISA): A Guide to Federal and State Regulation.) To be a “welfare plan” generally requires the plan to be established by an employer or a union. As explained in a recently issued Department of Labor advisory opinion, a plan can be a MEWA without being an employee welfare plan. This would occur, for example, if a commercial promoter established and maintained a group health plan covering a group of unrelated employers

Association health plan distinguished

An “association health plan,” i.e., a plan that covers entities in the same industry, is an example of a MEWA, since the plan covers employees of two or more unrelated employers. In order for a group or association to constitute an “employer,” there must be a bona fide group or association of employers acting in the interest of its employer-members to provide benefits for their employees. This requires a “commonality of interest” among the employers. Most association plans can satisfy these criteria.

For a MEWA to be subject to ERISA requires that the MEWA be an ERISA-covered welfare plan and that the arrangement cover employees of two or more unrelated employers. Only MEWAs that satisfy the Department of Labor “commonality of interest” standards can satisfy the first criteria—i.e., qualify as an ERISA-covered welfare plan. This is important because it affects the extent to which these plans are subject to state law. As a result of a 1983 amendment to ERISA, states are free to regulate self-funded association health plans. In contrast, a state’s ability to regulate fully-insured MEWAs is significantly constrained. Oversimplifying a bit, states may regulate the carrier that provides coverage to the full-insured MEWA but not the MEWA itself. As a result, absent special state legislation on the subject, a self-funded association health plan is an unlicensed insurance company.

An association health plan that fails the commonality of interest requirement is subject to regulation under state law irrespective of whether it is fully-insured or self-funded.

Treatment of group captives

For a group captive of the sort described in this post to work as advertised, it must be treated as a series or collection of individual, single-employer group health plans. It must not be a MEWA. Promoters and sponsors of group captive arrangements therefore take the position that:

• Each employer/sponsor of a group captive maintains its own, self-funded, single employer group health plan;

• Stop-loss insurance is in the nature of property and casualty insurance that insures the employer/sponsor and not health insurance that insures the employees and their beneficiaries; the former is not subject to ERISA, only the latter is; and

• Because the pooling of risk does not occur in connection with the providing of health insurance, there is no plan that covers employees of two or more unrelated employees—i.e., there is no MEWA.

The claim that stop-loss insurance is property and casualty insurance covering the employer, and not health insurance covering employees, is of central importance. There is, however, scant authority in support, or in derogation, of this claim. There is a 2014 technical release, in which the Department of Labor concedes that “Stop-loss insurance generally is not treated as health insurance under State law.” The Department made this statement in the context of opining on the power of the states to regulate stop-loss polices, so it should not be read as an endorsement of group captives.

Stop-loss Coverage and the ERISA “Plan Asset” Rule

In general, if a plan’s stop-loss policy is purchased by the plan, rather than by the employer, then it is considered a “plan asset” for ERISA purposes. But if the stop-loss policy is purchased by the employer and is intended to reimburse the employer, rather than the plan, it is not considered a plan asset. For a garden-variety self-funded plan (i.e., not involving a group captive), whether the stop-loss policy is a plan asset is important only for reporting purposes. If the stop-loss policy is plan asset, then it must be reported as such on a schedule to the plan’s annual report (Form 5500). But in the case of a group captive, the stakes are much higher, implicating both the ERISA prohibited transaction rules and the treatment of MEWAs under state law.

In a 1992 advisory opinion, the Department of Labor found that a stop-loss insurance policy purchased by an employer sponsoring a self-insured welfare benefit plan to which employees did not contribute would not be an asset of the plan if the following conditions are satisfied:

• The insurance proceeds from the policies are payable only to the plan sponsor, who is the named insured under the policy;

• The plan sponsor has all rights of ownership under the policy, and the policy is subject to the claims of the creditors of the plan sponsor;

• Neither the plans nor any participant or beneficiary of the plan has any preferential claim against the policy or any beneficial interest in the policy;

• No representations are made to any participant or beneficiary of the plan that the policy will be used to pay benefits under the plan or that the policy in any way represent security for the payment of benefits; and

• The benefits associated with the plans are not limited or governed in any way by the amount of stop-loss insurance proceeds received by the plan sponsor.

Some 13 years later, in a 2015 advisory opinion, the Department addressed contributory plans. Specifically, the Department opined that a stop-loss policy purchase by a plan that included participant contributions would not be a plan asset if the following conditions are satisfied:

• Except for the use of participant contributions to partly fund medical benefit under the plan, the facts surrounding the purchase of the stop-loss policies satisfy the requirements of the 1992 ruling;

• With respect to the use of participant contributions to fund in part the benefits under the plan, the employer must put in place an accounting system that ensures that the payment of premiums for the stop-loss policy includes no employee contributions;

• The purchase of stop-loss insurance must not relieve the plan of its obligation to pay benefits to plan participants, and the stop-loss insurer has no obligation to pay claims of participants; and

• The policies reimburse the plan sponsors only if the plan sponsors pay claims under the plans from their own assets so that the plan sponsors will never receive any reimbursement from the insurer for claim amounts paid with participant contributions.

The Department elaborated on the accounting system in the second bullet point above, saying:

Specifically, participant contributions are paid into the general account of [the employer] and recorded in a balance sheet. All health claims and other Plan expenses are paid from this [employer] general account. The plan sponsors will pay premiums for the policies, or any other stop-loss insurance, exclusively from a general account of [the employer].

It is the last requirement that is the most problematic in most group captives, since it would be unusual for plan sponsors to pay all claims in full and await reimbursement from the captive or the stop-loss carrier, as the case may be.

For group captives, the portion of tranche of stop-loss coverage provided by the captive must not be a plan asset for two reasons:

Regulation as an unlicensed insurance company under state law

What makes the group captive concept work is the ability to pool stop-loss risk. But if this coverage is a plan asset, then the pooled benefit is available to employees of two or more unrelated employers—i.e., a self-funded MEWA.

ERISA prohibited transactions and fiduciary self-dealing

If the stop-loss coverage is a plan asset, transactions with the captive will in all likelihood run afoul of the ERISA prohibited transaction rules described above. Employer-sponsors of group health plans are parties in interest, as are more than 50-percent-owned entities (or those entitled to more than 50 percent of the profits). This includes captive cells where the employer is entitled to more than 50 percent of the profits. The payment of premiums to a captive would trigger a prohibited transaction. In addition, it is possible to envision a violation of the ERISA self-dealing rules as well, since the employer/sponsor is also the fiduciary of its own plan, and arranges for that plan to enter into an arrangement that ultimately benefits the employer through its ownership interest in the captive. In either case, the penalties are severe.

Recommendations and Closing Thoughts

Group captives rest on a modestly reliable regulatory foundation. To be ideally positioned to withstand challenge, group captives would need to follow to the letter the steps outlined in the two Department of Labor advisor opinions described above. This would include the requirement to pay all claims then subsequently seek reimbursement from the captive or the commercial carrier. The Department’s rationale for this requirement is to ensure that “Plan Sponsors will never receive any reimbursement from the insurer for claim amounts paid with participant contributions.” But having the stop-loss pay the excess claims directly is not only commonplace, it also leaves the parties in the same financial position as the Department’s approach. Nevertheless, if the Department wanted a way to challenge group captives that don’t comply with this requirement, it could claim that the stop-loss policies including the captive are plan assets. We are unaware of the Department taking this approach on audit of examination.

There is another dynamic at work here. Earlier this year, the House Education and Workforce Committee approved a measure amending ERISA to exclude stop-loss coverage for self-insured health plans from the definition of “health insurance coverage.” The measure appears to be aimed at preventing the states from regulating stop-loss coverage for small groups. While the measure has not advanced, it does herald a change in the underlying politics. The impact on group captives is unclear, however, since the Committee at the same time reported out a bill intended to facilitate the formation of multistate small business association health plans.

It seems that the existing solutions for this particular hard insurance market—the one relating to contingent workers—need some help. While group captives appear to represent a viable market solution, they do not fit neatly into existing legal and regulatory structures. For now, the best that that can be said is they fit well enough to be considered a mainstream solution.