Trust-Owned Life Insurance and IRC 2042 on a Non-Grantor’s Life: A Workaround?

Examining the effects of entity ownership of life insurance in the context of a trust

To start, this may be an idea that has been covered elsewhere.  The views here are my own, based on limited research which has excluded past IRS rulings and written determinations.  That being said, the idea is so relatively straightforward (to me) in its application, from an estate tax perspective, that I find it surprising that I have not found better guidance. 

The problem we are attempting to solve is simple.  Can a trust purchase life insurance on a non-grantor beneficiary, without forcing the beneficiary to change their relationship to the trust? 

Setting the Stage

In estate tax planning, a major concern is the creation of irrevocable trusts and the question of whether the trust will be respected for estate tax purposes with respect to the grantor.  We see a number of ways this can be achieved, such as IRC Sections 2035-2038 which kick in when a transferor retains interests in previously-transferred assets for which there was not a bona fide sale. 

However, for a beneficiary of a trust, IRC Sections 2035-2038 do not apply (absent one exception below).  Usually, there are only two Code Sections under which the assets of a trust can be included in a beneficiary’s gross estate.  One is found in IRC Section 2041, to the extent of trust assets over which a beneficiary either (1) held a general power of appointment at the time of death, (2) released a general power of appointment during life while also retaining an interest described in IRC Sections 2035-2038, or (3) exercised a limited power of appointment in a manner which extends the perpetuities period applicable to the trust.  The other is found in IRC Section 2044, which includes the assets of a trust or portion of a trust in a surviving spouse’s gross estate to the extent a QTIP election was made by the personal representative of the grantor spouse’s gross estate. 

These are blanket rules which apply to all trust assets.  There is, however, one lurking issue for beneficiaries of a trust – the estate tax rules for life insurance.  IRC Section 2042(2) includes life insurance proceeds in the gross estate of the insured if the insured held, at the time of death, incidents of ownership in a life insurance policy which could be exercised alone or in conjunction with others.  Further, a transfer (which is not a bona fide sale) of incidents of ownership within three years of the insured’s death is disregarded under IRC Section 2035, triggering the application of IRC Section 2042. 

Many practitioners are familiar with these rules with respect to the creation and funding of irrevocable life insurance trusts, or ILITs.  ILITs are designed to hold life insurance for which the grantor is the insured, so as to avoid the application of IRC Section 2042.  The outcome, when the ILIT is deployed correctly, is an avoidance of estate tax on the life insurance death benefit. 

However, an issue that often flies under the radar is the treatment of incidents of ownership where a beneficiary, and not the grantor, is the insured.  The technical letter of IRC Section 2042 could extend to a beneficiary’s interest in a trust, and it is unclear whether a trustee’s discretion could avoid this application.  This is perhaps the only Code Section that applies two ways – to both a grantor and a beneficiary (mainly because the grantor’s retention of a power of appointment under IRC Section 2041 results in an incomplete gift, and this no use of gift tax applicable credit). 

More often, where we see this issue pop up is where a beneficiary is also a trustee of a trust holding life insurance on their life, or where a beneficiary holds a nonfiduciary power of appointment.  In fact, many modern trusts contain limiting language which prevents an insured beneficiary from exercising incidents of ownership over a policy through fiduciary or nonfiduciary powers. 

If we dive into the Regulations under IRC Section 2042, we find in Treas. Reg. 20.2042-1 the following conclusion regarding trusts:

A decedent is considered to have an “incident of ownership” in an insurance policy on his life held in trust if, under the terms of the policy, the decedent (either alone or in conjunction with another person or persons) has the power (as trustee or otherwise) to change the beneficial ownership in the policy or its proceeds, or the time or manner of enjoyment thereof, even though the decedent has no beneficial interest in the trust.

Based on this reading, any power of appointment in a fiduciary or non-fiduciary capacity would be an incident of ownership.  Again, the last sentence dealing with a beneficial interest is not clear on whether the interest itself is sufficient to be an incident of ownership, but it is certainly possible that a distribution right that is enforceable under state law (even if limited by an ascertainable standard) could rise to the level of an incident of ownership.  Even where a trustee has discretionary authority to select which assets might be distributed to satisfy an enforceable distribution right, the IRS could argue that the distribution right is exercisable by the beneficiary in conjunction “with another person” which is enough to attribute incidents of ownership under IRC Section 2042. 

So, does this mean that a beneficiary can never have their trust own life insurance?  I don’t know.  At the very least, it means that a beneficiary cannot serve as trustee of a trust owning life insurance on the beneficiary’s life, nor can the beneficiary retain any nonfiduciary power of appointment.  How, then, can we attain some certainty on the application of IRC Section 2042?  It seems inequitable that in the absence of a general power of appointment or QTIP election, no other trust asset but life insurance could be included in a beneficiary’s gross estate where a beneficiary holds any power of appointment. 

That being said, could we convert the life insurance into another form of asset at the trust ownership level, so as to avoid the application of this rule?  And, in so doing, could we allow a beneficiary to retain fiduciary control of the trust or even control of the alternate form of asset?  Possibly, but to get there, we would have to prevent attribution of incidents of ownership through the entity to the beneficiary as trustee or powerholder. 

Entity Ownership

The most obvious solution here is to have the trust create an entity, which will be respected for estate tax purposes, to own and be the beneficiary of the life insurance policy.  The entity could then be funded using income and principal of the trust, which could then be used to purchase life insurance on the (trust) beneficiary’s life.  The question, however, becomes the optimal form of entity.  There are income tax considerations, but IRC Section 2042 must also be considered. 

Hidden within Treas. Reg. 20.2042-1 is a possible safe harbor workaround for corporations.  In Treas. Reg. 20.2042-1(c)(6), we find the following helpful nugget:

In the case of economic benefits of a life insurance policy on the decedent's life that are reserved to a corporation of which the decedent is the sole or controlling stockholders, the corporations' incidents of ownership will not be attributed to the decedent through his stock ownership to the extent the proceeds of the policy are payable to the corporation.

Read on its face, this Regulation seems to imply that any form of corporation could be owned 100% by the trust, without any incidents of ownership being attributable to the trustee or insured beneficiary, so long as the corporation is the beneficiary of the policy.  But, this also means that the corporation must be the owner of the policy in order for the incidents of ownership to be constrained to the corporation. 

Diving deeper, this safe harbor speaks to a situation where the insured (directly or through a trust) owns a controlling interest in the stock.  The Regulation goes on to define this as more than 50% ownership.  So, could it then be possible for multiple trusts to aggregate their stock interests, so that no trust owns more than 50% of the stock?  This could be a valuable deployment if, for example, multiple siblings in one generation wanted to pool resources for life insurance. 

Let’s say, however, that the trust had to be the sole or controlling shareholder.  There has to be a catch, and indeed there is – two sentences later this Regulation (20.2042-1(c)(6)) clarifies:

See § 20.2031-2(f) for a rule providing that the proceeds of certain life insurance policies shall be considered in determining the value of the decedent's stock. Except as hereinafter provided with respect to a group-term life insurance policy, if any part of the proceeds of the policy are not payable to or for the benefit of the corporation, and thus are not taken into account in valuing the decedent's stock holdings in the corporation for purposes of section 2031, any incidents of ownership held by the corporation as to that part of the proceeds will be attributed to the decedent through his stock ownership where the decedent is the sole or controlling stockholder.

Simply put, the ticket to entry here is that the life insurance ends up being taxable, just in a different form.  We turn to Code Section 2031, which applies to increase the value of the corporate stock in the gross estate of an insured stockholder by the life insurance paid.

But, there are two catches, each of which could be its own safe harbor. 

One, this rule only applies if the trust in our proposed structure is the sole or controlling shareholder. 

Two, even if the trust is the sole or controlling shareholder, IRC Section 2031 does not apply in isolation to include assets in the gross estate of a non-grantor beneficiary of a trust.  Instead, in this context, it is a valuation rule that is only invoked by another Code Section.  And, if neither IRC Sections 2041 or 2044 apply to include corporate stock in the gross estate of the beneficiary, we are left simply with the question of whether IRC Section 2042 would.  The answer, based on my reading, is no, because 2042 does not apply to the corporate stock itself but instead to the life insurance which enhances the value of the corporate stock.  Since the stock itself is not included in the insured beneficiary’s gross estate, the increase in value of the stock attributable to the life insurance would not matter. 

It seems too simple, and there are certainly income tax catches to note below, but for the time being it appears that the safe harbor language of Treas. Reg. 20.2042-1(c)(6) could in theory allow a beneficiary to retain either fiduciary or non-fiduciary powers of appointment over a trust owning life insurance, so long as the life insurance is purchased within a “blocker” corporation of sorts.

Could we use a single-member LLC, or a partnership, instead?

Possibly.  But, these forms of entities are not expressly addressed in the Regulations.  Remember, the key here is whether entity ownership can result in attribution of incidents of ownership from the entity to the entity holder, or (in the case of a trust) any insured beneficiary holding a fiduciary or non-fiduciary power of appointment.  The only safe harbor we have is for corporations, but one could argue that the Regulations including life insurance in the value of corporate stock would also extend to any other entity owning life insurance. 

Still, however, we run into the issue that this value-inclusion rule must apply in order for the safe harbor to apply.  The value inclusion rule should not move the needle in the context of life insurance on a beneficiary, but one must avoid the disqualification of this value inclusion rule.  In light of cases such as Connelly v. U.S. (E.D. Mo. 2021), it is common to include buy-sell language in entity documents which includes a statement excluding the value of life insurance from the value of the entity’s equity interests in determining the purchase price for any purchase option.  Especially where this purchase price (net of life insurance) could be respected under IRC Section 2703, this could ironically disqualify the entity from the safe harbor under Treas. Reg. 20.2042-1(c)(6).  So, in drafting such an entity, care must be taken to remove this language from any forms used. 

What about entity control at the officer, manager, or general partner level?

A natural question is whether the insured beneficiary could also have managerial control of the entity owning life insurance, notwithstanding the stock ownership.  Could this give the IRS room to disregard the safe harbor under Treas. Reg. 20.2042-1(c)(6)? 

The answer is maybe.  Rulings in other contexts have clarified this, but the overly-conservative path would be to have someone other than the insured (directly or indirectly) exercise any managerial powers over the life insurance policy.  In fact, many trusts and entities contain stock language clarifying that an insured trustee or manager cannot exercise any incidents of ownership over a life insurance policy owned by the entity (I will address this language below). 

But, this is an area where we may be able to rely on a fiduciary duty exception.  While this is a narrow reading of the case, U.S. v. Byrum (1972) supports the proposition that a corporate director or officer has fiduciary duties to the shareholders which could prevent the application of IRC Section 2036.  Such fiduciary duties have been discounted in family ownership situations for partnerships not conducting an active trade or business, but not so much for corporations. 

Since life insurance is involved, however, is it possible that all of the dicta surrounding entities and retained interests must be discarded since it invokes different Code Sections?  Possibly.  But, in those Code Sections, there was no authority as clear as the safe harbor under Treas. Reg. 20.2042-1(c)(6). 

We can argue this from the opposite perspective, invoking some of the policies of Rev. Rul. 95-58.  In this Ruling, the powers of a trustee were attributed to a grantor who had the unqualified right to remove and replace a trustee.  The Service reasoned that this power could be used to remove a trustee, and allow the grantor to name themselves as trustee. 

What does this have to do with our safe harbor?  Well, for one, the safe harbor allows an insured to the sole or controlling stockholder.  Arguably, the sole or controlling stockholder would have the same power – to remove any corporate director, and name themselves as director.  It stands to reason that if a director’s or officer’s incidents of ownership could be attributed to a stockholder, this would have been addressed in the Regulations. 

In fact, such attribution was addressed in Treas. Reg. 20.2036-1(b)(3) (“But, for example, if the decedent reserved the unrestricted power to remove or discharge a trustee at any time and appoint himself as trustee, the decedent is considered as having the powers of the trustee”) and Treas. Reg. 20.2041-1(b)(1) (“A power in a donee to remove or discharge a trustee and appoint himself may be a power of appointment”).  While it may be a stretch to assume that such attribution language would have been included in Treas. Reg. 20.2042-1(c)(6) if it was intended, it is difficult to argue that this was an oversight if similar Regulations include this language. 

If limiting language is in the entity and the trust, why go through this exercise?

Just having the language in a trust or entity’s governing documents is not enough.  Who will actually exercise the authority to purchase life insurance, and exercise incidents of ownership?  At the very least, you would have to follow a procedure under the terms of the trust or under state law to identify and appoint an independent trustee meeting the requirements of IRC Section 672(c).  In so doing, you would have to also (again under the terms of the trust or under state law) sever a separate share of the trust that would own the life insurance policy, with the insured beneficiary releasing any power of appointment (and possibly any non-discretionary or enforceable distribution right) over the trust. 

Given this procedure, it may be easier to simply form a corporation to hold the life insurance policy.  And, as noted above, it could be possible for the insured beneficiary to be the director and chief officer of the corporation due to the Regulation’s omission of language recognizing the sole or controlling shareholder’s ability to name themselves as director and/or chief officer.  Even if this is not a safe enough exception, it may be easier to have an independent corporate director/officer than it would be to have an independent trustee.  However, separate share treatment may still be required as we will see below if an S corporation election is made. 

What about income tax?

This is tricky.  The safe harbor applies to a “corporation,” but we can infer that this reference within estate tax regulations would look to state law classification more than income tax classification.  So, you would need a state law corporation. 

The question then becomes one of tax status.  Unfortunately, you cannot use the check-the-box regulations to elect disregarded entity or partnership treatment (for income tax purposes) for a state law corporation.  Instead, you are left with two choices – C corporation, or S corporation. 

While life insurance proceeds are generally received tax-free under IRC Section 101 (except where the transfer-for-value rule or employer-owned life insurance exception applies), life insurance proceeds are included in a corporation’s earnings and profits under Treas. Reg. 1.312-1(b)(1).  The outcome, in a case of a C corporation, is that a distribution of the life insurance proceeds to the trust would be a taxable dividend under IRC Section 301.  So, the proceeds essentially become locked in the C corporation, and the use of the proceeds must be carried out through the C corporation structure in order to maintain tax-free status.  So, for example, if the proceeds were used to pay life insurance, this would have to be booked as a loan to the estate, which may still be treated as a constructive dividend to the trust followed by a loan from the trust to the estate.  The alternative would be the use of the proceeds to purchase assets from the estate or other trusts, which could trigger gain if the assets did not receive a stepped-up basis at the insured’s death. 

One also cannot assume that the life insurance proceeds would be tax-exempt – the effects of IRC Section 101(j) on employer-owned life insurance must be considered.  While the corporation arguably would not be an employer, protective compliance with the notice and consent requirements of IRC Section 101(j)(4) (accomplished through the annual filing of IRS Form 8925) could help avoid this outcome. 

So, what about an S corporation?  This seems to be the classification of choice given the dividend treatment for life insurance proceeds above, along with the application of IRC Section 311 to property within the C corporation (which prevents distribution of appreciated property from a C corporation without recognition of gain).  But, we must consider the effect of the life insurance proceeds on the basis of the S corporation stock held by the trust.  Luckily, the outcome is not one of doom and gloom – IRC Sections 1366(a)(1) and 1367(a)(1)(A) clarify that tax-exempt income increases a shareholder’s basis in its S corporation stock.  The effect?  The receipt of life insurance proceeds would increase the trust’s basis in its stock by the amount of the proceeds, allowing the S corporation to then distribute the insurance proceeds to the trust without triggering gain, as the distribution would not reduce the trust’s basis in the S corporation shares below zero. 

However, there are some headaches with S corporations. 

S corporation qualification

The S corporation itself will likely have no income (depending on the method of funding premiums discussed below), but will still be required to file IRS Form 2553 (to make the S corporation election) and IRS Form 1120-S annually (even if no income is earned). 

In addition, trust ownership of the S corporation stock will require either a QSST, or ESBT, election if the trust is not a grantor trust.  It is rare that a trust would be a grantor trust with respect to the insured beneficiary (creating other issues with the transfer-for-value rule below), unless it is structured as a BDOT (a beneficiary deemed-owner trust) under which the insured beneficiary has retained the right to withdraw all taxable income at least annually.  But, ironically, the effect of this right of withdrawal under IRC Section 2042 could be fixed by the S corporation structure proposed herein. 

As long as the S corporation has no income, and as long as the life insurance proceeds are received by the S corporation tax-free, the choice of a QSST or ESBT should not matter.  From an accounting perspective, however, the effect of the QSST election at the death of the insured beneficiary must be considered, along with the accounting treatment of life insurance proceeds distributed from the S corporation if the QSST election(s) continues with respect to one or more successor income beneficiaries under the trust or subtrusts formed at the death of the insured beneficiary.  As we will see below, the election may also affect the funding of premiums and the application of the transfer-for-value rule, tipping the scales in favor of a QSST election during the insured beneficiary’s life. 

And, either election will require separate share treatment for the S and non-S portions of the trust. 

Funding premiums – a huge headache?

I would be remiss if I did not address the payment of premiums.  In an ILIT context, the trust usually must either rely on annual payments of premiums (which are treated as a gift by an insured grantor to the beneficiaries whether contributed to the ILIT or directly paid), or the funding of the ILIT with assets producing enough income to pay premiums (the transfer of which may itself use gift tax applicable credit).  This is not, however, the same as an ILIT.  The appeal of life insurance on a beneficiary of a trust, instead of a grantor, is the avoidance of these hoops since the assets to pay premiums are already there. 

The problem, though, is the income tax treatment of contributions to the S corporation by the trust.  We can mimic the annual contribution structure by having the trust contribute sufficient assets to the S corporation to pay premiums, annually.  But, non-cash contributions may run afoul of IRC Section 351, resulting in gain recognition on the exchange of appreciated assets for stock, if the S corporation is treated as an investment company under IRC Section 351(e).  Since the sole asset of the S corporation is an insurance contract, along with assets used to pay premiums, the S corporation could be an investment company if the life insurance contract is treated as a “readily marketable stock or security.”  Term policies arguably would not rise to this level, but it is possible that permanent policies (especially indexed insurance products) could rise to this level. 

So, if annual cash contributions are not desirable, any income-producing assets contributed to the S corporation must prevent the crossing of a threshold where more than 80% of the noncash assets of the S corporation consist of marketable securities. 

Split-dollar arrangements between the trust and the S corporation could also be helpful, but care must be taken to structure the arrangement in a way that does not violate the safe harbor of Treas. Reg. 20.2042-1(c)(6).  Recall that this safe harbor requires the S corporation to be respected as both the owner and beneficiary of the life insurance at all times, and deemed ownership rules under the split-dollar regulations could affect compliance with this safe harbor.

So, in the grand scheme of things, contributions of cash for the payment of premiums would likely be optimal.  But, there is one wild card we have not yet considered – the effect of the separate share treatment of S and non-S shares of the trust.  A QSST can hold non-S corporation assets, but an ESBT cannot.  If an ESBT election is made, could the non-S trust make contributions directly to the S corporation?  These issues may make a BDOT or QSST a better structure, but this is not always an option.  If an ESBT is required (which would be the case if the insured is not a U.S. citizen), the S corporation would have to initially be funded with enough assets or income-producing assets to pay premiums for perpetuity, without reliance on subsequent contributions from the trust.  Alternatively, split-dollar would have to be used where an ESBT election is involved. 

Transfer-for-value rule

Where the transfer-for-value rule of IRC Section 101(a)(2) applies, a life insurance death benefit becomes taxable to a beneficiary to the extent it exceeds the beneficiary’s investment in the life insurance contract.  This could affect the contribution of an existing policy from the trust to the S corporation, a distribution of the policy from the S corporation to the trust, or a sale or any other non-gift transfer of the policy. 

There are two notable exceptions to this rule, which come into play in the S corporation arena.  The first exception applies to a transfer of a policy to a corporation in which the insured is a shareholder or officer.  The second exception is a transfer directly to the insured. 

As noted above, the trust would be the shareholder, so the initial impression is that the shareholder exception would not apply to transfers of the policy between the S corporation and the trust unless the trust is a grantor trust with respect to the insured.  As noted above, the trust likely would not be a grantor trust in isolation unless it is structured as a BDOT.  However, the QSST or ESBT election would have the effect of treating the trust as a grantor trust for S corporation stock with respect to its primary income beneficiary (in the case of a QSST) or with respect to each potential current beneficiary (in the case of an ESBT).  But, remember in the case of a QSST, the insured beneficiary would have to be a U.S. citizen. 

However, remember that the look-through rules of this deemed grantor trust treatment apply only to the S corporation stock, and not to property distributed from the S corporation.  Thus, a distribution of the policy to the trust, or a contribution of the policy from the trust, may not be a deemed transfer to the insured under these rules.  This leaves a situation where the only way to bail out the policy may be a sale to the insured.  Even if the insured is an S corporation shareholder with the trust (which would cause the life insurance to be taxable in the shareholder’s estate to the extent of the increase in value of their stock from the proceeds under IRC Section 2031), the prohibition on multiple classes of stock would prevent a disproportionate distribution of the policy to the shareholder at the expense of the trust. 

Conclusion

There are many angles from which this transaction can be analyzed, but corporate ownership of life insurance on the life of a trust beneficiary may prevent the application of IRC Section 2042, without the need to change the insured’s relationship to the trust as trustee, beneficiary, or holder of a power of appointment.  It may also be possible to allow the beneficiary to be a director or officer of the corporation itself. 

While the safe harbor of Treas. Reg. 20.2042-1(c)(6) seems to be clear, the ancillary income tax rules point to other key transaction points:

·         An S corporation may be the optimal tax election for the corporation;

·         If the trust is not a grantor trust with respect to the insured beneficiary, the QSST election may be the optimal tax election for the trust as an S corporation shareholder, at least during the life of the insured beneficiary;

·         Care must be taken to maintain the S corporation election through the filing of IRS Forms 2553 initially and 1120-S annually, even if the S corporation has no income;

·         While the insurance likely would not be employer-owned life insurance, the annual filing of IRS Form 8925 may be helpful;

·         Funding with annual contributions of cash may be optimal under the investment company rules and split-dollar regulations;

·         Given the issues with the transfer-for-value rule, the S corporation’s purchase of a new policy may be the safest avenue to avoid this rule given the myriad issues with income tax treatment of trust beneficiaries under the QSST and ESBT election; and

·         If the insured is not a U.S. citizen, you may be constrained to a BDOT structure or an ESBT election, and if the ESBT election is your only option, you may have to fully pre-fund the S corporation with enough assets to pay premiums or consider a split-dollar arrangement between the non-S share and the S corporation; and

·         If split-dollar is required or selected, be wary of the deemed ownership treatment under the split-dollar regulations for income, gift and estate tax purposes.