Early-Termination Assumption of Split-Dollar Arrangement a Bane for Taxpayer in Morrissette

“Intent to Terminate” Split-Dollar Arrangement Wins the Day for the IRS in Morrissette v. Comm’r

Recently, in Morrissette v. Comm’r (TCM 2021-60), the Tax Court finally arrived at a first, long-awaited, decision on the tax and valuation treatment of economic benefit split-dollar receivables. When such receivables are transferred, or transfers through an estate, two questions arise:

1. Does the transaction “work” – i.e., is there some reason in the tax code or regulations (Sections 2036 or 2703, for example) that causes the arrangement to be disregarded to the extent that the taxable fair market value is based on the premium paid, rather than the fair market value of the receivable itself, as a stand-alone instrument.
2. Assuming that the transaction form itself is respected, how should the receivable be valued? In other words, what do we determine the main inputs, including mortality rates, termination dates, and discount rates?

The answers, in Morrissette, based on this particular set of facts, are “Yes, it works” and “Very unfavorably (for the estate)”. I will in this piece ignore the first question – which may not have been resolved for all future in Morrissette in any case – and focus on the valuation. It should be noted, though, that some of the facts in Morrissette were unusual and highly adverse to the taxpayer.

The estate of Clara Morrissette held a receivable which gave it the right to the death benefit from a policy, a policy purchased some __ years earlier for an upfront premium of $32 million. The estate had valued the policy at a little over $7 million on the estate tax return. By the time of the trial, the value range between taxpayer and government had narrowed, slightly, but was still quite wide: $10 million to $28 million.

The four main inputs to this valuation, as presented in the Court’s opinion, in rough order of importance to the result are:

1. The projected death benefits (based on illustrations).
2. The mortality rate assumptions – which mortality table to use.
3. The discount rates and, most importantly, what research and set of empirical data should be applied in determining the discount rates.
4. Assumptions regarding early termination of the arrangement, i.e., termination of the arrangement before the maturity of the life insurance policy.

The Morrissette Court’s decisions on the first two items were a mixed bag on balance between the interests of taxpayer and the IRS, and I will not spend significant time reviewing these inputs here.

The truly salient inputs with the most impact on the value, both in Morrissette and quite possibly in future cases as well, are the discount rates and termination assumptions. In extreme cases, assuming very early assumption, the discount rate itself also becomes less important (as we shall see). However, it makes sense to start here by reviewing the discount rates first.

Discount Rates

Discount rates have an inverse relationship with the value of an instrument (discounting being the inverse of compounding interest). In the standard discounted cash flow model, each future cash flow is discounted to the present by a fraction determined by the discount rate and the length of time between the present and each expected future cash amount. Discount rates should always, to the extent possible, be determined by analogies or comparisons between the instrument being valued and similar or comparable instruments. For example, when valuing real estate, an appraiser can look at other buildings having sold, and derive a discount rate implicit in that sale. Enough data on comparable sales yields a very strong foundation to make a decision regarding the subject property.

The discount rates used by taxpayer (two separate appraisals) and government were far apart, indeed. And, with a longer assumed time horizon to future cash flows, those differences in discount rates might have had a major impact on the value. However, as we shall see, the early-termination assumption is the main driver of the Court’s value conclusion in Morrissette. The discount rates used by the government ranged from 6.4% to 8.85%. The taxpayer’s experts used discount rates ranging from 15% to 18%.

The government’s expert applied two different discount rates for the two policies that formed the basis for the receivables. Each of the discount rate conclusions were based on the debt yields for the insurance company underwriting each policy, plus a small liquidity premium. Comparing a bond that has a fixed repayment schedule and maturity date with a life insurance receivable, payable quite far into the future, and at a completely unpredictable point in time, might seem inappropriate. Certainly, the risk characteristics of the two instruments could not be further apart. The Court recognizes that the “split-dollar agreements involve a degree of uncertainty in the timing of the payout that is not present with corporate debt”. However, the Court does not give this difference significant weight in its analysis because mortality “statistics lessen the degree of uncertainty.”

The estate’s experts generally relied on life settlement yields when setting the discount rate. The Court notes:

Factors related to life settlements can explain the wide range of yields including the varying sizes of the underlying policies, the financial strength of the insurance companies, the insureds’ medical histories, mortality assumptions, and continued obligations to pay premiums.

Unfortunately, the Court points out, “we have no way of evaluating these factors … because such information is not provided.” The following are factors noted by the Court as making life settlements riskier than the receivables valued in Morrissette:

• The obligation to pay additional premiums
• The poor health of the insured
• The financial stability of the insurers
• The return on the investment portfolios of the insurers
• The insured’s life expectancy.

The Court concludes: “The differences between the life settlements and the split-dollar agreements make the comparison unreliable especially in the light of the lack of transparency associated with life settlement yields.”

Termination Assumption

Normally, when valuing a split-dollar receivable, the assumption is that the receivable will mature (and pay out) at the time the policy matures. Not so in Morrissette, which is key to the decision. The final maturity date, in the government’s analysis (adopted by the Court) was December 31, 2013, about 4.25 years out from the valuation date, and a small fraction of the expected maturity date of the policies. In the standard discounted cash flow approached used by all experts in this case, such a difference in timing makes a huge difference in values. So how did the Court arrive at this early-termination assumption? These are the main reasons:

• The family had decided “to purchase policies with high premiums and modest death benefits”
• Emails between family and the advisors structuring the split-dollar arrangement discussed “the possibility of canceling certain policies”
• One of the advisors “insisted that the policies not be canceled until the three-year period of limitations on the estate return had expired”

However, in the Court’s analysis, “the key factor in setting the … maturity date is the brothers’ complete control over the split-dollar agreements.”

Underpayment Penalties

The IRS had applied a 40% penalty for a gross valuation misstatement when auditing the tax return. There were some procedural reasons why penalties might not apply in this case, but those were rejected by the Court. The key remaining question, then, was whether the taxpayer could demonstrate that it acted with reasonable cause and in good faith.

The Court had already ruled in a prior order in the matter that “petitioners waived reliance on legal advice as a reasonable cause defense on the basis of the claim of attorney-client and work product privileges.” The only remaining possible defense being whether the estate’s reliance on the appraisal provided a reasonable cause defense. The Court held that it did not, for the following reasons:

• First and foremost, the appraised value, less than $7.5 million, was “not reasonable, and the brothers should have known that. The brothers had the … trust pay $30 million and turned it into $7.5 million for estate tax reporting purposes. They should have known the claimed value was unreasonable and not supported by the facts.”
• The taxpayer “also knew that (the advisors) were marketing the agreements as an estate tax saving strategy.” The advisors made it clear “the tax benefits of the split-dollar agreements would be obtained through the undervaluation of the” receivable.
• The taxpayer knew “that any estate tax saving depended on valuing the split-dollar rights at a substantial discount from the premiums” paid.
• Before the return filing, the advisor “warned Don that the IRS would likely see problems with the values” planned reported.
• “Knowing that any estate tax saving would be from the undervaluing of the split-dollar rights, the brothers engaged an appraiser that (the advisor) recommended.” The advisor reviewed the draft appraisal and asked the appraiser “to make changes that reduced his opined values.”

Summary

Split-dollar structures vary widely. Thus, the most important part of the Court’s decision in Morrissette may not carry over to future cases. The application of an early-termination assumption here may also have been influenced by evidence presented that this arrangement was promoted and marketed, perhaps primarily, as a tax avoidance scheme. May that also have colored the Court’s decision on penalties?

From the summary presented in the memorandum decision, the taxpayer’s appraisal does not sound all that unreasonable. This question, also, depends on your view of the early-termination assumption. Was it so completely unreasonable – prior to the Court’s decision on this point – to disregard early termination of the arrangement as an outcome? Note that, but for the early termination date, both the taxpayer’s and the government’s experts would have arrived at substantial discounts. Thus the presence of a discount cannot in and of itself be unreasonable – unless the taxpayer (perhaps not sharing all necessary information with the appraiser) by procuring a valuation that disregarded early termination as an option, intentionally stacked the deck and the result was an appraisal that was inherently unreliable. A harsh result, certainly.