It's been five years since the Internal Revenue Service issued Notice 2002-8, the pronouncement that was supposed to clarify treatment of those split-dollar plans that had been established before the Service issued the final regulations. But many of the most fundamental questions about the tax implications of even the simplest steps taken to address these plans remain unanswered. Despite the uncertainties, though, one thing is clear: The tax and economic issues underlying many of these plans, particularly those involving irrevocable life insurance trusts (ILITs), are not going to be any easier to address as insureds get older and the amounts potentially subject to tax get bigger. It may be time to stop hoping that the IRS will provide the necessary guidance, start drawing some conclusions ourselves, and figure out what to do with these plans.

The type of plan we'll focus on here is typical: a collateral assignment equity split-dollar plan on the life of an employee (or joint lives of an employee and spouse) established before Sept. 17, 2003, the date the final regulations became effective. More particularly, we're focusing on plans that were in effect as of Jan. 27, 2002, the key date for special safe harbors afforded by Notice 2002-8. We'll refer to these pre-Jan. 28, 2002, plans as “existing plans.”

First, we'll give some background on how these plans were typically structured, the tax and economic expectations that people had when they implemented them, and the most important changes in the ground rules for taxation of these plans. Then we'll provide questions to consider before settling on a strategy for coping with these plans in light of the current guidance. Next, we'll explore four possible strategies: (1) leave the split-dollar arrangement intact and keep your options open; (2) roll it out; (3) recast it as a loan; and (4) roll in and terminate the plan or maintain it under an endorsement plan.

In a typical plan, the ILIT owns the policy and the employer pays the premium. The employer's premium advances are secured by a collateral assignment of the policy. Upon termination of the arrangement during the insured's lifetime, the employer is repaid either the premiums it advanced or the lesser of the premiums it advanced or the policy's cash value. Once repaid, the employer releases the assignment. We refer to the employer's release of the assignment as the “roll out.” At roll out, any cash value greater than the employer's share is the equity, which belongs to the ILIT at the time the plan is terminated.

When collateral assignment equity split-dollar plans were established, their success from a tax standpoint depended on two essential elements: One was the continued availability of very low carrier term rates, used as the measure of the economic benefit for income and gift tax purposes. The other and more significant element was the belief that the equity would never be subject to income or gift tax. In reality, the success of the plan depended on several additional factors, such as policy performance and preparation for accomplishing the roll out at the earliest possible time.


In a split-dollar plan, the insured has an imputed economic benefit because of the insurance protection he receives. From the mid-1960s to the end of 2001, the measure of this economic benefit was the lower of the so-called “PS 58 rate” or the insurer's published, generally available one-year term rate, based on the insured's attained age. The economic benefit for a survivorship policy in a split-dollar plan while both insureds were alive was measured by joint PS 58 rates that reflected the actuarial probability of both insureds dying in the same year. The joint-life rate is much lower than the single-life rate, thereby enhancing the tax leverage of the survivorship split-dollar plan. When one of the insureds dies, the economic benefit increases to the single-life rates, to reflect that payment of the death benefit is now conditioned on only one life.

The insured's imputed income from the plan is reduced by any premium contribution he makes to the policy. Because an ILIT owns the policy, the annual economic benefit is a gift by the insured to the ILIT. In a contributory plan, the insured usually makes gifts of the economic benefit to the ILIT, which contributes that amount to the plan.

The economic benefit attributable to the coverage is imputed (or must be contributed) each year as long as the plan is in force — even if the company is no longer advancing premiums to the insurer. Therefore, most plans were designed and funded so they'd terminate before the insured's age caused the annual term cost to become oppressive for income tax purposes and, when an ILIT was involved, for gift tax purposes. Plans usually called for employers to pay premiums for 10 to 15 years. The premiums were set high enough to build enough cash value in the policy to enable the ILIT to tap the cash value (by loan or withdrawal) in the 16th year for funds to repay the employer (do a roll out) and support a targeted death benefit with no further cash premiums from the employer or the insured.

Then, in January 2001, the IRS issued Notice 2001-10. The portion of that notice that remains relevant today is its declaration that the PS 58 Table would not be applicable after 2001. A new, interim table — Table 2001 — was provided for split-dollar arrangements and is still in effect. The Table 2001 rates are substantially lower than the PS 58 rates, but substantially higher than a typical insurer's qualifying one-year term rates. Taxpayers could continue to use the insurer's lower, one-year term rates until the end of 2003. After 2003, the insurer's qualifying one-year rate could be used only if the insurer's rate met several stringent requirements.

A year later, the IRS issued Notice 2002-8 offering safe harbors for the economic benefit in existing plans. Essentially, plans established before Jan. 28, 2002, were permitted to use the insurer's qualifying one-year term rates as the measure of the economic benefit beyond 2003. In practice today, some carriers use “old” rates; others do not. While carriers using old term rates might tell planners they're confident those rates qualify under Notice 2002-8, they generally don't provide a guarantee. Some carriers let the planner decide which rate, old or new, to use. So, to a certain extent, planners have to make their own determinations about which rate is appropriate. But if the IRS challenges an old rate, then presumably any income and gifts (and associated tax) would have to be recalculated in accordance with the higher Table 2001.


With respect to the equity in existing plans, in Technical Advice Memorandum (TAM) 9604001, the IRS considered the income and gift tax implications of a collateral assignment equity split-dollar arrangement when an ILIT owns the policy. The Service's conclusion was that growth in policy equity (that is to say, the annual increase in the amount by which cash value exceeds the amount due back to the employer) is taxable income to the insured and constitutes a gift from the insured to the trust. The IRS applied Internal Revenue Code Section 83, which governs when property is transferred to an executive in connection with performance of services. Under that section, the insured is taxable on such property once the property is transferable or free of substantial risk of forfeiture. Because the employer's creditors could reach only the cash value equal to the amount of premiums paid by the employer, any excess cash value or equity belongs to the insured without risk of forfeiture. After issuing this TAM, the IRS fell silent on the Section 83 issue — until it released Notices 2001-10 and 2002-8.

In Notice 2001-10, the IRS indicated that it would tax the equity in these arrangements, but was not specific as to how or when it would do so. The successor to Notice 2001-10 was Notice 2002-8. We already noted the impact of Notice 2002-8 on the measurement of the economic benefit. The impact of the notice on the taxation of the equity is more important — and more vexing. Essentially, for plans established before Jan. 28, 2002, the IRS will find no transfer (of equity) under IRC Section 83 and therefore will not tax the equity as long as the plan remains in place. Notice 2002-8 gives safe harbors for avoiding taxation of the equity upon termination. The equity would not be taxed if the plan was either terminated (rolled out) before Jan. 1, 2004 — or recast as a loan under IRC Section 7872 or the applicable provisions of the tax law. Presumably, if the plan was neither terminated nor recast by Jan. 1, 2004, the IRS would attempt to tax the equity upon termination. Bear in mind that if an ILIT owns the policy, taxable equity is both income to the employee and a gift to ILIT.

The vexing portion of the notice is language stating that, except for the standards for valuing current life insurance protection, “no inference shall be drawn from Notice 2002-8 regarding the appropriate federal income, employment, and gift tax treatment of split-dollar life insurance arrangements entered into prior to the date of the publication of the final regulations.” There are different schools of thought about what the so-called “no inference” language means in real terms to real practitioners who must advise real clients and even sign real tax returns. One school of thought says the IRS essentially has grandfathered existing deals and won't pursue the equity for income or gift tax purposes. The other school, which we and most corporate clients seem to attend, believes that the IRS was merely saying that it reserves the right to make its arguments in favor of taxation.

Collateral assignment plans established after Jan. 28, 2002, also were given some relief. The IRS won't tax the equity so long as the plan remains in force, but there's no safe harbor for terminating or recasting those plans as loans by Jan. 1, 2004.

Meanwhile, in June 2002, the Sarbanes-Oxley Act (SOX) became law. Most advisors/securities lawyers interpreted SOX to classify collateral assignment plans for directors or covered executives as prohibited personal loans under IRC Section 402 of SOX. The act did grandfather in premium advances/loans as of June 30, 2002. Importantly, many advisors/securities lawyers interpreted SOX as allowing endorsement split-dollar arrangements and non-equity (economic benefit) collateral assignment arrangements. In the typical endorsement plan, the employer/company owns the policy and endorses a portion of the death benefit out to the insured's designated beneficiary.

In July 2002, the IRS issued proposed regulations that established the new playing field for split-dollar plans: economic benefit regime with traditional annual taxation of that economic benefit pursuant to the insurer's one-year term cost or Table 2001 (endorsement plans and some, but not all, non-equity collateral assignment plans) and loan regime (collateral assignment plans). These regulations were to apply to plans implemented after the regulations became final in September 2003. The final regulations apply to split-dollar arrangements implemented after Sept. 17, 2003 or — importantly — to pre-existing arrangements that are thereafter “materially modified” (a largely undefined term.) Therefore, Notice 2002-8 remains the operative guidance for the taxation of pre-final regulation split-dollar arrangements.

Planners trying to figure out what to do with existing plans are likely to find key provisions of the final regulations as vexing as the “no inference” language — if not more so. For all practical purposes, we have very little insight into what constitutes a “material modification” that would transform taxation of the existing plan from a traditional economic benefit regime to a loan regime. Clearly, the most common question is whether a policy exchange is a material modification. But there are many other concerns, such as whether a change of the premium-paying entity (perhaps after a merger/acquisition occurring after 2003), or a change from one policy-owning ILIT to another, or a decrease in the death benefit necessitated by a loss of premium funding due to SOX, are material modifications that would force tax regime change.

In some cases, indeed many, the long-term economic advantages of exchanging an old policy for a new one, transferring a policy from one ILIT to another, or doing both, are likely to trump concerns about risking a material modification. Even when these advantages are not present, a change to the loan regime might be in the best long-term interest of the insured or, as it happens in so many cases, of the surviving insured in a survivorship split-dollar case.


Many, if not most people who had these collateral assignment split-dollar arrangements did nothing about them in reaction to Notice 2002-8. Either they didn't know about the changes or didn't have the equity to worry about; or SOX precluded recasting as a loan; or once they started to look into it and considered the questions we'll pose in a moment, the strategies for dealing with the notice were simply too difficult from an accounting, tax or policy perspective.

Some companies did terminate plans because, putting the notice and new regulations aside, theirs weren't working as they'd hoped. This was especially true with arrangements involving variable universal life insurance hit by a bear market that made it look like roll outs from those policies would not be feasible for many years, if ever.

But everyone attached to a pre-Jan. 28, 2002 collateral assignment split-dollar arrangement must take action — even if it's simply to make the conscious and fully informed decision to do nothing for the time being. And before the parties can explore their options, we think the company, the insured, the ILIT trustee, the agent and other advisors have to jointly and severally address these 14 threshold questions:

  1. Does the insured still want the insurance?

    We've found several situations in which executives who “needed” the insurance when it involved no cash outlay on their part and no taxation of the equity decided that, based on the economic and tax implications of the new rules or SOX or whatever, their kids had enough money already. Plan terminated and case closed!

  2. Is the employer a public reporting company?

    If so, is the insured covered by SOX? Note that in the five years since SOX was introduced, executives whose plans were put on hold because of SOX may have retired. If so, the options for dealing with the plan are considerably greater than if the insured is still in an executive position. Often, executives who retired remain on the company's board, in which case SOX still applies. But if the insurance is an important enough component of that insured's planning, he might consider resigning as a director so that the company can resume payments under the plan.

  3. What is the date of the plan? Is there a written agreement? Was an assignment filed with the carrier?

  4. Is the policy in a gain position? If not, when will it be in a gain position? What about surrender charges?

  5. How much death benefit is the plan intended to provide to the ILIT, for how long and under what assumptions?

  6. What was the intended timing and strategy for a roll out? Who bears the risk that the policy will perform well enough to actually enable a roll out? The company or the insured?

  7. Does the company expect and require full repayment of its premium advances? If the company is a public reporting company, is it willing to accept the book and tax implications of releasing its assignment for less than its due?

  8. Would the company be willing to support the policy by increasing the insured's compensation?

    In many cases, particularly those that count on a roll out from policy values, it will be counterproductive (if not simply impermissible) to continue advancing premiums under the split-dollar arrangement. Doing so will only add to the aggregate obligation to be repaid at roll out, thereby putting that much more pressure on the policy to perform. The agent should be able to show how a change to a bonus approach, when the premiums are paid as additional compensation to the insured (and as full measure of gift to the ILIT) can be a better solution than trying to bring the policy financed roll out to fruition.

  9. If an ILIT is involved, is the insured (really) aware of the gift tax implications of the escalating annual economic benefit? What is the insured's tolerance for making large gifts as part of the safe harbor strategies available under Notice 2002-8? What about when these issues involve trusts with generation-skipping tax provisions?

  10. Is the ILIT a grantor trust for income tax purposes?

    While this question may seem to be tangential to the “real” split-dollar questions, it's actually crucial when large premiums (and gifts) are involved. There will be more flexibility to consider loan-based strategies involving the insured as the lender if the ILIT is a grantor trust, than if it's not. Strategies to accelerate the roll out by complementing policy values with trust assets gifted or sold to the ILIT to create a “side fund” to help with the roll out are certainly more feasible if the ILIT is a grantor trust. In cases with a certain constellation of facts, it may be that if the roll out is projected to be many years away, the wiser course will be to recast the plan as a loan, transfer the policy to a new ILIT designed as a grantor trust, and perhaps, exchange the policy for a more flexible, more efficient product.

  11. Here's a reprise of the first question, but now with emphasis: If the plan was terminated and the insured had to pay the full premium with after-tax dollars, how much insurance would he want?

  12. What kind of policy is it? How much flexibility is there to reconfigure the policy if the arrangement is changed?

    When the whole life illustration projects that a policy-financed roll out won't be feasible for another 15 to 20 years, it's likely worthwhile to look at exchanging the existing policy for a competitively designed flexible premium policy (while remembering that such an exchange might constitute a material modification).

  13. When and with how much more premium will the policy be able to support a roll out?

  14. What possible issues or concerns will the trustee have to contend with when he explores strategies that reduce the amount of insurance, moves the policy to another trust, etc.?


So, what's the action plan for existing collateral-assignment plans? Here are the four choices:

  1. Leave the plan intact and the options open.

  2. Roll out the policy and terminate the plan.

  3. Recast the split-dollar arrangement as a loan.

  4. “Roll in” (transfer the policy to the company) and terminate the plan, or maintain the arrangement under an endorsement plan.

Let's examine these one at a time:

  1. Leave the plan intact and the options open.

    Notice 2002-8 tells us that as long as the insured is taxed on (or the ILIT contributes to) the annual economic benefit, the equity will not be taxed. Therefore, the arrangement can be kept in place, using either the insurer's qualifying one-year term rates or Table 2001. This option can be a permanent solution, subject to several caveats. Or, it can be a strategy of wait-and-see (who else gets audited), meaning that if the IRS indicates it won't pursue taxation of the equity or the Service loses in the courts, the parties can terminate the plan at the appropriate time without undue consequences.

    This option might be a permanent solution when, for example, the in-force illustrations project that the policy will never build enough cash value to roll out of the plan within the insured's life expectancy. The insured will have an ongoing taxable economic benefit for both income and gift tax, but (presumably) there will never be an issue of equity taxation if the plan continues until the insured dies.

    But there are caveats for the employer and the insured. The company would have to reflect that the value of its receivable must be discounted for the executive's life expectancy. (Note that recent developments in accounting rules may require this change anyway as of 2008.) If the terms of the split-dollar agreement call for the arrangement to be terminated upon the retirement of the insured, does maintaining that arrangement post-retirement constitute a material modification that will impose loan regime treatment?

    If there is no equity in the policy and no issue with SOX, a comparison of ongoing one-year term costs and the costs associated with variations on a loan approach might suggest recasting the arrangement as a loan anyway. Also, depending upon whether the plan is maintained “as is” or recast as a loan, the employer may be able to reduce the premiums it's paying on the policy. After all, if the company is willing to agree that it will defer its recovery until the executive dies, the cash value will not have to be tapped to repay the company upon the executive's retirement and would be available to support the death benefit for the long term.

  2. Roll out and terminate the plan.

    The plan is terminated by the employer's releasing the collateral assignment for full, partial or no consideration. The insured includes in income any shortfall between what the company is due and the amount paid for the release of the assignment. The shortfall is also a gift to the ILIT. Meanwhile, the company will have accounting concerns and may have IRC Section 162(m) limitations on the amount of its deduction.

    If the policy is in equity when the plan is terminated, the company is going to have to determine if it will report the equity as compensation to the insured. The company will study the “thread” of taxation of the equity, from Revenue Ruling 66-110, to the TAM, to Notice 2002-8, to the provisions for taxable transfers of policy value in the (admittedly inapplicable) final regulations. The insured and her advisors should not be surprised to find that the company will conclude that, in light of the penalties associated with an employer's underreporting income and its essential neutrality on the issue, the prudent course is to report the equity as compensation upon termination of the plan. The insured presumably then reports the equity as a gift to the ILIT.

    In many cases, even if there is equity in the policy at the time of the roll out, the policy still will need more premiums before it becomes self-sustaining. The executive would continue to make any additional premium payments on the policy. The funds for such payments can come from the company as compensation to the executive or from the executive's own funds. If the plan is indeed terminated, the company still might want to enable the executive to maintain a permanent life insurance program. So, the agent would explore what can be done with the existing policy to reduce the premium needed to support a given amount of death benefit. For example, there would be no further need to have an increasing death benefit design because the company no longer has an interest in the policy. Also, the policy would no longer have to support the withdrawal of the company's premiums, so there wouldn't be the need for robust cash value accumulation.

    The measure of the executive's gift to the ILIT would now be the full premium, not just the one-year term cost. Therefore, the executive is likely to be very interested in any policy restructuring/replacement that can reduce the amount of her outlay, the duration of that outlay or both, and the possible use of a private split-dollar arrangement to reduce the gift tax cost of funding ongoing premiums.

  3. Recast the split-dollar arrangement as a loan before equity (possible only when there are no SOX issues).

    The employer's existing and subsequent premium advances are considered a loan to the insured. Depending on a number of factors, the loan can bear interest at the applicable federal rate (AFR) or it can be imputed under IRC Section 7872. If the loan bears interest, it can be designed as a demand or a term obligation. If it does not bear adequate interest, the loan can be designed as an obligation that is a demand, term or hybrid (a term loan treated as a demand loan.) Generally, if an ILIT is involved, the loan will bear interest, which either will be paid or accrued, as facts and circumstances allow. Meanwhile, we urge the insured to fund the ILIT with property that can be used, perhaps in concert with policy values, to repay the loan as soon as is practicable. Clearly, the insured now would prefer the ILIT to be a grantor trust to avoid treating the interest as income.

    A variation on this theme is for the company to recast as a loan only its existing advances and pay the remaining premiums as bonus compensation to the insured. Here, again, the premium outlay and/or product design might be revisited in light of the fact that the obligation to the employer will have been frozen, so the roll out theoretically could occur sooner.

  4. Roll in and terminate the plan or maintain it under an endorsement plan.

While this strategy would usually be considered when SOX applies, it might be considered when the employer insists on owning the policy. The ILIT transfers the policy to the company and the company terminates the plan or changes it to a non-equity endorsement plan. If the latter, the insured will have imputed income for the annual economic benefit until the plan is terminated, which may not be until the insured's death. The potential downsides to this strategy are that economic benefit may have to be based on the less-advantageous Table 2001, the trustee must be comfortable relinquishing the policy, and assignment of the death benefit to an ILIT (perhaps the same ILIT) will likely be subject to the three-year rule of IRC Section 2035.


Any difficulties dealing with Notice 2002-8 were compounded for collateral assignment plans for executives of those public companies that interpreted SOX as prohibiting the company from paying premiums after June 30, 2002.

Responses to the SOX/Notice 2002-8 dilemmas were as varied as responses to the threshold questions we've articulated. Some plans were simply terminated, with their policies surrendered or rolled in to the company. In many cases, companies suspended their premium payments under the split-dollar plans but the executives continued to contribute (at least) the one-year term costs to their ILITs to maintain coverage for the foreseeable future. Not infrequently, executives were only a few years from retirement, when SOX would no longer apply to them and the parties would have more latitude to deal with their plans. Of course, retirement was not going to remedy the problem for an executive who also served on the company's board and therefore would continue to be covered by SOX.

When SOX still applies (and will do so for the foreseeable future) and a change to a loan regime is out of the question, a company can continue to “freeze” the plan, meaning suspend its premium advances, and the executive can pay all or part of the remaining premiums with personal gifts or loans to the ILIT. Of course, the company might be able to increase the executive's compensation, essentially turning the split-dollar plan into an executive bonus. The company still would be due its premium advances under the frozen plan, but the executive's contributions should nullify any imputed income attributable to the company's outstanding premium advances. The parties would work with the agent to determine whether, based upon the insured's age, the type of policy and other factors, the originally planned roll out from cash values will ever be feasible. If not, they might be able to reduce the premium enough to maintain the plan (and enough death benefit to reimburse the company) until the insured's death.

An altogether different strategy would be for the executive to purchase the assignment from the company and recast the arrangement as a private split-dollar plan or a private premium financing with the ILIT. Alternatively, the ILIT can enter into a premium financing arrangement with a commercial lender to acquire the funds to repay the employer.

When circumstances won't allow for these strategies, the ILIT could transfer the policy to the company (roll in), which could establish a new endorsement plan under the economic benefit regime that it would maintain for the rest of the insured's life. A variation on this theme would be for the parties to change the nature of the plan from an “old” collateral assignment split-dollar to a “new” non-equity collateral assignment plan under which the company would be deemed to be the owner of the policy and the measure of the economic benefit would be either the insurer's one-year term rate or the Table 2001. But there are caveats: In particular, the company's counsel must be comfortable that SOX would not apply to a non-equity endorsement arrangement.


As any planner who has attempted to deal with a pre-Jan. 28, 2002 collateral assignment plan can tell you, the ambiguity of the guidance is clear. But time waits for no one, and unless all parties are comfortable with a “wait-and-see” strategy, it's probably time to begin to map out the strategies, make the operative assumptions and reckon with these plans before they are too expensive to reckon with.

The views expressed in this article are those of the authors and do not necessarily reflect the views of Ernst & Young LLP or Bryan Cave LLP. This document should not be construed as legal, tax, accounting or any other professional advice or service. No one should act upon the information contained herein without appropriate professional advice after a thorough examination of the facts of a particular situation.