The Grantor ILIT – What Should You Look For?

Comparing assumptions to reality when examining the income tax status of an irrevocable life insurance trust

Executive Summary:  There are many paths to achieve grantor trust status.  It is often assumed that an irrevocable life insurance trust (ILIT) is a 100% grantor trust, and many life insurance planning techniques are based on this assumption.  However, one must be cautious when determining the income tax status of an ILIT. 

Introduction

Irrevocable life insurance trusts, or ILITs, are a staple of estate planning.  While these trusts create management and protection for life insurance death benefits, they often have a primary purpose of saving transfer taxes.  But, one must approach these types of trusts with caution. 

While life insurance death benefits are often excluded from gross income for income tax purposes, the same does not hold true from a transfer tax perspective.  IRC Section 2042 generally includes life insurance death benefits in the gross estate of an insured if the insured held any incidents of ownership at the time of death.  IRC Section 2035 creates a look-back, where the transfer of incidents of ownership (other than through a bona fide sale for full and adequate consideration) within 3 years of the insured’s death will be disregarded for purposes of IRC Section 2042.  And, one must be conscious of the fact that a transfer of incidents of ownership for an existing policy, or the payment of premiums on a policy not owned by the insured, may be gifts for gift tax purposes.

Even where transfer taxes are avoided, one must be cautious with how an ILIT is approached from an income tax perspective.  An ILIT must be respected as a separate taxpayer in order for transfer tax savings to be achieved, meaning that the grantor cannot maintain incidents of ownership through retained control or benefit over the trust using principles similar to (and perhaps more onerous than) IRC Sections 2036 and 2038.  All of these exceptions recognize a bona fide sale exception, which avoids gift (and sometimes estate) tax so long as adequate consideration is received by a transferor who may be subject to these taxes.  Nonetheless, the life insurance transfer-for-value rule may be invoked where any consideration is received by the insured.

What is the transfer-for-value rule?  This rule, codified under IRC Section 101, generally removes the income tax exclusion for life insurance death benefits when a policy is converted into an investment.  In other words, the purchase of a policy outside of narrow exceptions causes the purchaser to be taxed on the death benefit as ordinary income, minus investment basis equal to the purchase price plus premiums paid by the purchaser. 

(There are some other issues that come into play with split-dollar life insurance, which we will not examine in this discussion, but these issues are especially prescient given recent cases such as Morrissette and Levine). 

This issue with grantor trusts especially rears its ugly head where the transfer-for-value rule is invoked.  Why?  Because one of the exceptions to the transfer-for-value rule is a sale to the insured, meaning that the insured’s purchase of a policy from another owner (even if the sale itself is a taxable exchange) never causes the death benefit to become taxable when paid out.  However, a sale of a policy by an insured grantor to a non-grantor trust could potentially invoke the transfer-for-value rule, while a sale of the policy back to the grantor would not (due to the sale-to-the-insured exception). 

Diving deeper, since the assets of a grantor trusts are treated as if they are the assets of the deemed owner (whether the grantor or a beneficiary) for income tax purposes under the grantor trust rules, this means that a grantor ILIT will generally create greater income tax flexibility for sales or exchanges of policies between the insured and the trust.  Unlike a non-grantor trust, sales of policies both ways between a grantor and a grantor trust would avoid the transfer-for-value rule so long as the trust is a fully-grantor trust.  In determining whether an ILIT is a fully-grantor trust, the mechanism for obtaining grantor trust status may matter more than we think. 

IRC Section 677(a)(3)

There are a variety of mechanisms used to obtain grantor trust status, but the mechanisms are often blindly viewed with a binary outcome.  In other words, it is assumed that a trust is fully grantor, or fully non-grantor.  This is where practitioners should pump the breaks, because this assumption can be faulty. 

IRC Section 677(a)(3) is often thought of as the mechanism to attain grantor trust status for an ILIT.  This Code Section provides that grantor shall be the owner of “any portion” of a trust for which income is, or may be, applied for the payment of life insurance premiums on the life of the grantor or the grantor’s spouse.  Reading the statute, there are two standards for examining the use of income – “is” and “may be.”  These are sometimes misinterpreted, but each has its own separate deployment. 

The ”is” standard applies where income is so applied for payment of premiums, unless coupled with the approval or consent of an adverse party.  In other words, if income is actually applied to premiums, this can create a fully- or partially-grantor trust unless a beneficiary consented to such use of income.  If there was consent of an adverse party (with some added nuance we will not get into here), you do not have a grantor trust in this situation unless the next standard (“may be”) is satisfied.

The “may be” standard looks to the discretionary use of trust income for insurance premiums, by either the grantor or a nonadverse party, or both.  In this situation, a trustee’s discretion to use income for the payment of premiums by itself can create a grantor trust.  This is the standard often relied upon when drafting a grantor ILIT.  If you read an ILIT, there is often an express provision permitting the trustee to use income and/or principal for the payment of life insurance premiums on the life of the grantor (we will refer to this power in this article as a “677(a)(3) power”).  This is customarily accepted by many drafters as the mechanism to attain fully-grantor trust status for an ILIT.  On the other hand, a non-grantor ILIT is achieved by expressly stating that life insurance premiums can only be paid from principal, and not from income.  (This latter point has special application, for example, in a charitable remainder trust, which cannot be a grantor trust). 

As noted above, these two tests are often misinterpreted and misapplied.  There are two key items of authority we must consider when examining if either test creates a fully grantor trust. 

First, there is a companion section – IRC 677(a)(2) – which usually converts an inter vivos irrevocable trust (such as a spousal lifetime access trust, or SLAT) into a fully-grantor trust.  This principle also has the same two-pronged test, but in the context of whether income is or may be distributed to the grantor or the grantor’s spouse.  There is an added layer, however, which often gives us the boost to assume that a 677(a)(2) power creates a fully-grantor trust – the application where income is “accumulated for future distribution” to the grantor or the grantor’s spouse.  What this means is that, in a SLAT, income that is added to principal (such as capital gains, or accumulated income that is not distributed) can create a grantor trust if principal is or may be distributed to the grantor or the grantor’s spouse.  There are some nuances in the regs, but this 677(a)(2) power generally creates a fully-grantor trust without question.

However, IRC Section 677(a)(3) does not consider whether income may be accumulated for future payment of premiums from principal.  This, at the very least, prevents an ILIT from being a grantor trust in any given tax year where premiums are, or may be, paid from principal. 

Second, many practitioners often cite Revenue Ruling 66-313 as authority for the conclusion that only income that is actually applied to the payment of premiums of life insurance on the grantor or the grantor’s spouse is taxed to the grantor.  This reading may be erroneous, mainly because (as noted above) a plain reading of the “may be” standard can override the “actually applied” standard to at least cause the income of the ILIT to be taxed to the grantor in any given year where a non-adverse trustee has a 677(a)(3) power.   

We will dive into this Ruling a bit more as follows, but for the time being I want to leave our analysis as:

·         It is often assumed that an ILIT is a fully-grantor trust, especially when there is a 677(a)(3) power.

·         However, there are multiple 677(a)(3) powers – one which examines whether income was actually applied to premiums, without an analysis of discretionary distribution powers, but subject to the qualifier of the consent of adverse parties, and the other of which examines whether income may be applied to premiums in the discretion of the non-adverse trustee, or the grantor, or both.

·         It may be safe to assume that a trust with the “may be” standard is a fully-grantor trust, regardless of whether income is actually applied to the payment of premiums, but this is dependent on the trust terms – the trustee must have authority to pay premiums from the income, and not just the principal, of the trust, and accumulation of income in principal is not enough to get away with claiming grantor trust status where premiums must be sourced from principal only. 

Revenue Ruling 66-313

This Ruling, as introduced above, is fundamental to the understanding of this grantor ILIT issue.  It is often relied upon for the premise that only the income of a trust actually applied to the payment of life insurance premiums is taxed to the grantor under both prongs of Code Section 677(a)(3).  However, as you may have guessed, this Ruling applies only the “is” standard, and not the “may be” standard. 

In this Ruling, there were two trusts (Trust A and Trust B), each with the grantor’s three children as the income beneficiaries.  Trust B was to own life insurance.  The Ruling involved the income tax treatment of Trust A, which did not own life insurance.  The proposed transaction involved the application of the income of Trust A to pay the life insurance premiums incurred by Trust B.  The outcome?  Only the income of Trust A actually applied to the payment of premiums incurred by Trust B, a different trust, would be taxed to grantor. 

Ironically, there is a flaw in this Ruling.  Of note is the fact that the Ruling only generically references IRC Section 677, without actually analyzing how Code Section applies.  If there were an analysis, surely the analysis would have taken into account one key representation of the grantor:

“The grantor represented that each beneficiary of trust A will consent in writing (revocable at will) to have his share of the income from trust A applied toward the payment of the premiums on the life insurance policies in trust B.”

Remember above, how we stated that the “is” standard does not apply when you have the consent of an adverse party?  Here, we have all adverse parties, the children, revocably consenting to the application of Trust A income to Trust B premiums.  Whether or not this affected the Service’s analysis is not known, as the Ruling’s text is only two paragraphs.  But, if the consent were not revocable, one could argue that the “is” standard of IRC Section 677(a)(3) should not have applied at all.  Alternatively, one could argue that this revocability is what led to the conclusion that only the income actually applied to premiums was the income that would be taxable. 

Beyond these limitations, perhaps the greatest value is in the universe of application.  Discretionary distribution standards were not addressed, and if they were, they likely would only have applied in isolation to each of the two trusts.  It is rare that a trustee would have the ability to use the income of one trust to pay life insurance premiums incurred by a separate trust.  Few drafters even think to consider this possibility, as it carries separate gift tax issues (which were not addressed here).

Given its limitations, the following are some key takeaways from this Ruling:

·         The scope of the Ruling was limited to the “is” standard, and may even have misapplied this standard as stated within IRC Section 677(a)(3). 

·         This Ruling’s facts may not apply in a one-trust situation, which is the norm for ILITs which are being examined for grantor trust status.

A Case for Fully-Grantor ILITs

I will address the gold standard for a fully-grantor trust – the swap or substitution power – in the next section, but for the time being I want to dive deeper into this question of whether an ILIT is a fully-grantor trust under IRC Section 677(a)(3). 

In keeping with this introduction, keep in mind that a 677(a)(3) power could be superseded in scope by another grantor trust power.  This would remove any argument that you have a fully-grantor ILIT.  For example, in a single-life ILIT with a spousal beneficiary, this trust would likely be a fully-grantor trust if the spousal beneficiary can receive distributions of income or principal without the consent of an adverse party under IRC Section 677(a)(2).  In addition, a substitution power could supersede the limited scope of a 677(a)(3) power. 

So, the analysis here applies in a situation where the only string of a grantor’s retained control or benefit is found under IRC Section 677(a)(3).  Unfortunately, much like the Revenue Ruling cited above, we have no direct, definitive analyses which hold that a 677(a)(3) power would create a wholly-grantor trust.  Instead, we have some assumptions and conclusory statements which are founded on this principle. 

Take, for example, Estate of Petter v. Commissioner, T.C. Memo 2009-280.  This is a popular case for gift tax formula clauses, but also contains a loose conclusion that a trust with a 677(a)(3) power would be a fully-grantor trust with respect to the grantor-insured. 

Many other examples can be found in Private Letter Rulings, which are not binding authority but nonetheless reach the same conclusion as Petter.  For example, PLRs  8014078, 8007080, 8103074, and 8118051 each contained identical language providing that a trust subject to a 677(a)(3) power would be a fully-grantor trust.  Similarly, PLR 8852003 treated a trust subject to a 677(a)(3) power as a fully-grantor trust for purposes of qualifying the trust as an S corporation shareholder.  Finally, PLR 200228019 treated a trust subject to a 677(a)(3) power as a fully-grantor trust for purposes of qualifying for the exception to the transfer-for-value rule (cited above) allowing a “deemed” transfer to the insured by means of a transfer to a grantor trust where the insured was the deemed owner of trust assets. 

On this latter point, the Service appears to have taken the blanket assumption (without elaborating about the mechanisms to get there) that ILITs are fully grantor trusts under Rev. Proc. 2007-13.  In this Ruling, the Service concluded that the transfer-for-value rule would not apply to a sale of a policy to a trust for which the insured was the deemed owner under the grantor trust rules. 

And, interestingly, PLR 8839008 applied the “is” standard of IRC Section 677(a)(3) to a nongrantor trust in accord with the conclusion reached in Rev. Rul. 66-313 – that only the income of a nongrantor trust actually applied to the payment of life insurance premiums on the life of the grantor would be included in the grantor-insured’s gross income for the year of payment. 

Swap Powers: An Effective Backup?

The gold standard for intentionally created a defective grantor trust has often been the power of substitution, or swap power, described in IRC Section 675(4).  This power, when reserved by the grantor, allows the grantor to reacquire trust corpus by substituting other property of an equivalent value.  Such an exchange is not taxable, due to the fact that the trust is disregarded for income tax purposes with respect to the grantor, as further discussed in Rev. Rul. 85-13.  In fact, it is this very principle upon which Rev. Rul 2007-13 (above) avoids the application of the transfer-for-value rule since the insured is technically selling the policy to, or purchasing the policy from, themselves for income tax purposes.

But, since the goal of an ILIT is to avoid the retention, or transfer (within 3 years of death), of incidents of ownership, one must question whether a swap power runs afoul of this goal.  Arguably, so long as the exchange of assets with the ILIT is for equivalent value, there should be no issue under IRC Sections 2035-2038.  This was clarified in Rev. Rul. 2004-64, which clarified that IRC Section 2036 would generally not apply to a swap power unless there was a mandatory tax reimbursement clause and, as a bonus, that no gift would result from the grantor’s payment of income tax on behalf of the trust and its beneficiaries. 

I would be remiss, however, if I did not bring up Rev. Rul. 2008-22.  This ruling held that a swap power does not trigger IRC Sections 2036 or 2038, but only if a trustee (other than the grantor) has a duty to ensure that both the assets exchanged are of equivalent value, and that the exchange does not have the effect of shifting benefits among the trust beneficiaries. 

However, what many practitioners do not recognize is that this Ruling has a companion – Rev. Rul. 2011-28 – which basically extends the holding of Rev. Rul. 2008-22 to the question of whether a swap power constitutes the retention of incidents of ownership to the grantor-insured under IRC Section 2042. 

This begs the question of whether it is good practice to also include a swap or substitution power when drafting an ILIT if the grantor wants to be assured of wholly-grantor trust status.  Such a backup may be a good idea to supplement a 677(a)(3) power, especially if the grantor plans to have the ILIT purchase an existing policy in reliance on Rev. Rul. 2007-13.  Similarly, such a power may be helpful where there is a 677(a)(2) power to prevent an inadvertent termination of grantor trust status if the grantor’s spouse predeceases the grantor. 

But, the issue comes in not when examining the swap power itself, but examining how it is actually applied.  If, for example, a policy is sold to a grantor ILIT (whether or not there is a swap power) for less than the fair market value of the policy, there is at least a partial gift, and it is unclear whether this would cause part or all of the policy’s death benefit to be included in the gross estate if the insured dies within 3 years of the sale under IRC Section 2043.

It is also important to keep in mind that a swap power is not an absolute safe harbor, even when there is an exchange of equivalent value.  This is especially true in the case of an ILIT, since the payout may be substantially delayed.  It is possible that the exchange of liquid assets in a trust for a life insurance held by the grantor-insured, while for adequate consideration, could have the effect of shifting beneficial interests by favoring the remainder beneficiaries (who will receive death benefits after the grantor’s death) over the current income beneficiaries (who could otherwise have benefitted from the assets transferred to the grantor in exchange for the policy).  For this reason, one may question whether it is a good idea to have the ILIT purchase the grantor’s policy in exchange for a promissory note.  But, this is a rabbit hole we will not dive down for now. 

Takeaways

It is easy to take for granted the question of whether an ILIT is a fully-grantor trust or partially-grantor trust.  While the rules on determining what “portion” of a trust’s income (and possibly principal) is deemed to be owned by a grantor or beneficiary are beyond the scope of this discussion, the tides seem to have turned towards the assumption that a trust from which the income “may be” applied for payment of life insurance premiums is a fully-grantor trust, notwithstanding the question of whether income was actually applied toward the payment of premiums. 

While reliance on the “may be” standard appears to be the custom from a drafting perspective for ILITs, it is important to weigh this standard against the possibility of other grantor trust rules being invoked.  And, as always, one must weigh the effect of grantor trust powers against the application of IRC Sections 2035 and 2042. 

In the meantime, if you enjoy this subject matter, I am working on a comprehensive course on grantor trusts and their principles.  If you would like to receive updates on this course, and provide input on what nuances of the grantor trust rules are most confusing for you, please click here: https://forms.gle/eK63jy6s28cEn2Vr6