Full opinion text
OPINION LAWSON, District Judge. Dow Chemical Company and its subsidiaries (Dow) have filed a complaint in this Court against the United States claiming that they are entitled to refunds for taxes paid for calendar/fiscal years 1989, 1990, and 1991, totaling $22,209,570, plus interest. The center of the dispute is the disal-lowance of deductions which Dow claimed for payment of interest on loans that were used to pay premiums on broad-based, corporate owned life insurance (COLI) policies purchased by Dow, and for administrative expenses associated with the purchase and maintenance of those insurance policies. In a very real sense, then, this case involves both death and taxes. Dow purchased COLI policies on the lives of 4,051 of its upper management employees from Great West Life Assurance Company in 1988. In 1991, Dow purchased a group COLI policy on the lives of 17,061 employees from Metropolitan Life Insurance Company (MetLife). The premiums on these policies were financed by means of an elaborate plan to borrow from the insurer to pay premiums in the first three and eighth years of the policies. The loans were secured by the cash value of the policies. Premiums in years four through seven were paid by means of partial withdrawals of the accumulated cash value of the policies. The policy acquisition plan thus drastically minimized the initial cash outlay for premium payments. The United States claims that because the payment of the premiums and the loans and withdrawals occurred simultaneously on the first day of each policy anniversary, and were accomplished virtually or literally simultaneously by means of netting transactions, the transactions never actually occurred and constitute factual shams. The United States also asserts that the COLI plans had no practical economic purpose apart from generating the tax deductions for the interest payments and therefore are shams in substance. The government also contends that the Great West COLI plan does not constitute “life insurance” under Michigan law because Dow did not have an insurable interest in all of the 4,051 employees insured under that plan. Consequently, the argument goes, the plan fails to comport with Internal Revenue Code § 7702(a) and therefore Dow is not entitled to any of the tax advantages afforded life insurance, particularly the deferral of tax on the “inside build-up,” and the interest paid on policy loans is not deductible under Internal Revenue Code § 161. Further, the government argues that the use of simultaneous netting transactions to finance the premiums by means of policy loans and withdrawals, which it believes are factual shams, causes the plan to fail the “four-of-seven” test set forth in Internal Revenue Code § 264(c)(1), and therefore, for that additional reason, interest deductions should be disallowed. Many of these issues were thoroughly litigated in three prior cases, American Electric Power, Inc. v. United States, 136 F.Supp.2d 762 (S.D.Ohio 2001), In re CM Holdings, Inc., 254 B.R. 578 (D.Del.2000), aff’d 301 F.3d 96 (3d Cir.2002), and Winn-Dixie Stores, Inc. v. Commissioner of Internal Revenue, 113 T.C. 254, 1999 WL 907566 (1999), in which the courts have found that the broad-based COLI plans constituted shams in substance. The constellation of these cases has formed a lodestar which has guided and shaped the parties’ presentation of evidence; the plaintiff has endeavored to demonstrate through testimony and exhibits that its COLI plans are substantially different from the plans condemned in the prior cases, while the defendant has attempted to show that the Great West and MetLife COLI plans are virtually identical, with nearly all of the offending features of the other plans in common. Trial began on January 8, 2002, and the proofs concluded on March 12, 2002. The Court heard the testimony of 26 witnesses and received 1,526 exhibits. The parties filed a stipulation of facts consisting of 137 separate paragraphs. Initial and amended proposed findings of fact were filed, along with post-trial briefs. The parties then presented their final arguments in open court on May 28, 2002. The following constitutes the Court’s findings of fact under Federal Rule of Civil Procedure 52, followed by its application of the governing law. I. Background and Facts of the Case A. Life Insurance General Terms and Features In its most basic form, a life insurance contract consists of an agreement by an insurance company to pay a sum of money, known as a death benefit, to the beneficiary named on the insurance policy upon the death of the insured life. In consideration of the payment of a premium, the insurance company assumes the risk that the insured will die during a fixed period of time, usually one year. The premium charged for that year is calculated based upon data which includes the age of the insured, life expectancies of individuals of that same age, the likelihood that individuals of that age will die during that year, and the amount of the death benefit. If there are no other features to the insurance contract, this kind of insurance coverage is known as term insurance or “pure insurance,” and the charge associated with assuming the risk of premature death of the insured during that period is called the cost of insurance (COI). Because the likelihood of death increases as age advances, the COI for renewable term insurance becomes increasingly expensive as an insured grows older. The insurance industry over the years has developed products that ameliorate the high cost of term insurance in the later years. ‘Whole life insurance” is a form of cash value insurance that is designed to provide coverage over the course of one’s entire life, which is typically calculated at 95 years. A level, annual premium in excess of the cost of insurance is charged. The excess premium is invested by the insurance company so that the insurance policy accumulates “cash value,” which consists of the accumulation of the excess premiums and earnings. The earnings are referred to as “inside build-up.” If the insured dies, the cash value is paid out as part of the death benefit. As the insured advances in age, the cash value becomes an increasingly larger component of the total death benefit, and the pure insurance element correspondingly decreases, thereby moderating the COL In a typical whole life insurance policy, the annual premium is comprised of the COI, an excess amount which is invested for the purpose of accumulating cash value, charges for expenses such as policy administration and commissions, and a profit for the insurance company. The COI element of the premium is based in part on calculations using complex actuarial formulae which endeavor to quantify the risk of mortality of an insured in relation to a given population. Information concerning rates of death generally comes from statistical studies and compilations by actuaries who assess the mortality experience of a given population. The results are assembled in mortality tables. In some circumstances, the actual mortality experience of an insurance company, that is, the frequency of the incidence of death among actual policy holders compared to the expected mortality of the comparator population, can determine the profitability of an insurance company. Another component of profitability comes from the performance of the insurance company’s investment of excess premium. However, an insurance company may choose to share favorable mortality, expense, and investment experience with its policy holders by paying dividends when this experience outperforms expectations. Insurance policies which have this feature are known as participating (par) policies. Those without this feature are known as non-participating (non-par) contracts. Generally, in whole life policies the expense and pricing components of the premiums and, in par policies, dividends formulae, are not revealed to the policy holder. In some cash value insurance policies, the policy holder has a contractual right to access the cash value. This may occur in the form of loans from the insurance company which are secured by the cash value of the policy. The insurance company charges interest, usually at a rate in excess of the rate of return paid on the investment principal of the contract. The rate of return on the investment portion of the insurance premium is known as the “credited rate.” The amount of interest charged on the policy loan is known as the “loan rate.” The difference between the loan rate and the credited rate is called the “spread,” which can be established or adjusted to serve a variety of payment and financing goals. If an insured dies, a portion of the death benefit is used to pay off the loan and outstanding interest charges. The policy holder may also access the cash value of the policy in certain insurance contracts by partial -withdrawals, or “partial surrender,” of the policy. Partial withdrawals need not be repaid, but there is a corresponding reduction in the death benefit of the policy. When the cash value of the policy reaches a point where the return on investment covers the COI required to satisfy the death benefit along with the accumulated cash value and the expense of administration, the policy is considered “paid up” and no further premiums are due. Some policies are designed to require premium payments throughout the insured’s “whole life,” while others, such as the policies in this case, may compress the payments by requiring a larger premium for a lesser number of years for the same death benefit. Some par policies include a feature that dividends declared are used to purchase “paid-up additions” which increase the amount of the death benefit. In the late 1970s, the insurance industry developed a “universal life” policy, which is a cash value policy in which all of the economic components are “unbundled,” or revealed, to the policy holder. The development of universal life policies and the unbundling of economic components stimulated the development of different life insurance products that could address varying investment goals and returns for individual policy holders. For example, a younger individual who wants to purchase a maximum death benefit for a finite period of years may elect to purchase term insurance. Someone who wants a very low premium but permanent protection may elect to purchase a whole life insurance policy with limited borrowing features. A person who intends to use an insurance policy as a savings vehicle may elect to purchase a universal life policy which includes a partial withdrawal feature. An individual who wants to use his insurance policy as a source of funds may look for an insurance contract which permits borrowing and partial withdrawals. The insurance industry has developed products over the years to accommodate all of these goals. Insurance contracts are highly regulated at the state level, and the standard policies, or “forms,” themselves are submitted to state regulators for approval. Policy forms may be approved as individual policies — i.e., contracts which insure a single life — or as group policies. Under a group policy form, a single policy can provide insurance on the lives of several individuals. Insurance provided to several employees as a benefit of employment frequently takes the form of a group contract, with the corporation owning the policy and the employees designating their respective beneficiaries. Some states, including Michigan, as will be explained later, limit the range of group insurance contracts in which corporations may be named as beneficiaries. B. Tax Treatment of Life Insurance and Changes in Tax Law It has been a long-standing feature of Congressional tax policy that the death benefit of insurance policies is not subject to income tax. See IRC § 72, 101(a). In addition, the inside build-up is tax deferred, and if paid out as a death benefit it is non-taxable. See id. Policy withdrawals are treated as coming first from basis and then from earnings, so withdrawals up to basis are likewise not taxable. See id. The proceeds from loans secured by the cash value of insurance are, of course, not taxable since they do not constitute “income.” See Internal Revenue Code (IRC) § 61 (not listing loan proceeds as gross income); United States v. Ivey, 414 F.2d 199, 202-03 (5th Cir.1969). Furthermore, interest paid on these loans in the past has been tax deductible, although Congress has curtailed and ultimately eliminated the interest deduction. In 1964, Congress amended IRC § 264 to limit interest deductions on loans used “to purchase or carry a life insurance ... contract ... pursuant to a plan of purchase which contemplates the systematic direct or indirect borrowing of part or all of the increases in the cash value of such contract.” Pub.L. 88-272 (1964); IRC § 264(a)(3). Deductions were still allowed where “no part of 4 of the annual premiums due during the 7-year period (beginning with the date the first premium on the contract to which such plan relates was paid) is paid .under such plan by means of indebtedness.” IRC § 264(c)(1). In 1986, Congress once again amended section 264 to allow interest deductions only on the first $50,000 borrowed and secured by the cash value of a policy. See Pub.L. 99-514, § 1003, 100 Stat.2085 (1986). Congress eventually eliminated the interest deduction altogether when it enacted the Health Insurance Portability and Accountability Act of 1996 (HIPA), Public Law No. 104-191, 110 Stat. 1936 (1996). C. COLI in General Traditionally, individuals purchased policies insuring their own fives naming other persons or entities as the beneficiaries designated to receive the death benefit upon the death of the insured policyholder. In earlier times, it was not uncommon for strangers to purchase insurance on the lives of prominent people, in effect wagering on the likelihood of their premature death. See Crossman v. Amer. Ins. Co. of Newark, N.J., 198 Mich. 304, 308, 164 N.W. 428, 429 (1917). The concept of “insurable interest” arose to curb this disturbing trend. Generally speaking, an “insurable interest” is “a reasonable ground, founded upon the relations of the parties to each other, either pecuniary or of blood or affinity, to expect some benefit or advantage from the continuance of the life of the assured.” Warnock v. Davis, 104 U.S. 775, 778-79, 26 L.Ed. 924 (1881) An individual is presumed to have an insurable interest in his or her own life. 3 Couch on Insurance, § 41.19 (1995). Likewise, it has long been recognized that corporations have an insurable interest in their important employees, supporting the development of so-called “key person” insurance in which a corporation purchases insurance on its employees’ fives naming the corporation as the beneficiary. 3 Couch on Insurance, § 43.13. The death benefit provides economic protection against the untimely death of employees important to the success of the business. In these plans, the insured neither names the beneficiary nor owns the policy. Rather, it is “corporate owned.” Because of the favorable tax treatment of death benefits and inside build-up, corporate-owned fife insurance on the fives of key employees was marketed and sold as an investment vehicle. Proceeds were commonly used to fund deferred compensation and other employee benefit plans. In addition, the industry developed plans to leverage the purchase of such insurance by borrowing from the insurance company, using the cash value of the policy as collateral, and using the loan proceeds to pay the annual premiums. Large cash value policies were marketed on the tax arbitrage opportunity based on the deductibility of policy loan interest. When Congress limited the interest deduction to policy loans of $50,000 or less, insurance entrepreneurs marketed COLI policies which insured a broader employee base, taking advantage of a national trend recognizing an employer’s insurable interest in lower-level employees. Contributing to the desirability of these plans was the concept of “aggregate funding,” i.e., using the cash generated by the policy or group of policies to fund entire programs, rather than simply tying the policy’s death benefit to the benefit costs of a specific insured individual. In many circumstances, the financial success of these broad-based COLI plans relied on the favorable tax treatment of cash value fife insurance — that is, tax-exempt death benefits, tax deferral of inside build-up, and the deductibility of policy loan interest — rather than the economic gain solely from premature mortality of the insured employees. In prior challenges to broad-based, highly leveraged COLI plans, the United States has not assailed the taxpayers’ reliance on and utilization of tax benefits afforded by the treatment of death benefits and inside build-up. The broad-based COLI plans that have been under attack are those in which the government has argued that the taxpayer will derive no economic benefit from the plan absent the tax deductions on policy loan interest. In fact, it is the absence of the likelihood of profit from mortality and the stripping of cash value which has made those plans suspect and susceptible to the economic sham argument, as is described in the AEP, CM, and Winn-Dixie cases. The government claims that Dow’s two COLI plans suffered the same infirmities. D. The COLI Plans in AEP, CM Holdings and Winn-Dixie The insurance policies involved in Winn-Dixie, CM Holdings, and AEP, were all characterized as highly leveraged, broad-based COLI plans, which combined large, front-loaded premiums and liberal access to quickly accumulating policy cash value. The cash from the policies accessed by a combination of loans and other distributions (partial withdrawals and dividends) was used to pay premiums and loan interest charges. The respective corporations owned the policies and were named as beneficiaries for the death benefits. 1. Winn-Dixie Stores, Inc., v. Commissioner of Internal Revenue In Winn-Dixie, the taxpayer purchased insurance on the lives of nearly all of its 36,000 full-time employees. Annual premiums were charged in the amount of $3,000 per insured payable in years one through fifteen, with the amount of death benefit varying depending on the age of the insured. For the first three years, approximately 93% of the premium was paid by means of loans secured by policy cash value. Winn-Dixie elected an option in which interest was charged at a variable rate equal to Moody’s high-risk bond average (Moody’s BAA) instead of a lower fixed-term rate that was offered, and the borrowed cash value was credited with earnings at 40 basis points below the loan interest rate. This amounted to a 10.66% return on borrowed funds in the first year, yet the crediting rate on unborrowed funds was 4%. Based on pre-purchase 60-year projections, the Court found that Winn-Dixie would pay premiums and loan interest primarily through partial withdrawals of cash value in years four through seven, and then continue borrowing to pay loan interest throughout the remaining policy years. In the first four years, the amount of the premium actually paid in cash roughly approximated the insurance company’s COI plus expense charges. The same projections demonstrated that the net pre-tax cash flow generated by the plan was a negative $682 million; however, when accounting for policy loan interest deductions and assuming a 38% tax bracket, Winn-Dixie would realize over $2 billion in positive cash flow over the life of the plan. The policies issued by AIG Life Insurance Company also contained a provision establishing a “claims stabilization reserve” which effectively limited the ability to profit from mortality charges. The policies allowed for partial withdrawals and borrowing up to net cash value, with the death benefit applied first to retire any outstanding policy loans. After the passage of HIPA which eliminated tax deductions for COLI interest payments, Winn-Dixie cancelled its policies with AIG. The Tax Court found that the loan and other premium-financing transactions “actually occurred,” and therefore confined its analysis to whether the plan constituted a sham in substance. 113 T.C. at 278, 1999 WL 907566. In examining the overall transaction, the Court concluded that since there was no reasonable basis for Winn-Dixie to expect to profit from death benefits due to the ameliorating effect of the claim stabilization reserve, and policy cash value was relatively small throughout the projected life of the plan, the deduction of policy loan interest payments was “clearly the dominant element” of the plan. Id. at 281, 1999 WL 907566. Without the benefit of interest deductions, the plan yielded substantial negative cash flow in each of the sixty years. The plan was economically viable only as long as Winn-Dixie’s “appetite for interest deductions remains large,” and tax considerations “permeated” the pre-purchase analyses. Id. at 288, 1999 WL 907566. Finding that consistent pre-tax negative cash flow “precludes any economic value, economic significance, economic substance, or commercial substance other than the tax benefit,” the Court concluded that the plan was a sham in substance. Id. at 290,1999 WL 907566. 2. In re CM Holdings, Inc. In CM Holdings, the COLI policies were “designed to be owned on a broad base of employees, to be financed through a highly leveraged transaction, and had to provide the policyholder with a positive cash flow in every year of the policy.” CM Holdings, 254 B.R. at 582-82. The COLI policies, known as the “COLI VIII Plan,” were issued by Mutual Benefit Life Insurance Company (MBL) to Camelot Music, Inc., a wholly owned subsidiary of CM Holdings, and carried a fixed annual premium of $10,000 payable in years one through nine for each of the policies purchased on 1,431 employees (one policy was later rescinded). The death benefits increased over time, varied based on the issue age of the policies on each individual insured, and the policies had several cash-access and transparency features of universal life policies. Camelot paid $1 million of the $14 million in cash for the first-year premium. The balance of the premium, or approximately 93%, was paid by policy loans. Camelot financed the second and third year premiums and accumulated loan interest charges in the same fashion. Camelot continued to take policy loans in the second and third years to pay approximately 90% of the annual premiums. Id. at 593. The other 10% of the annual premiums was paid in cash. Id. at 593. The policy loan interest in these years was paid by taking additional policy loans. Id. at 607. The premium payment and loans occurred in simultaneous netting transactions in which the amount of the policy loan was deducted from the gross premium, the payment of which was the basis for the cash value which secured the loan. The amount of the premium which Camelot elected was designed to create first-day, first-year cash value that earned interest at the crediting rate, which Camelot elected to be indexed to Moody’s BAA enhanced rate. This election in turn determined the policy loan interest rate, because Camelot chose a variable rate calculated at 100 basis points above the crediting rate for borrowed funds. Customarily in cash value policies, the annual premium less an expense charge and the COI is added to the policy’s cash value. The expense charge is a percentage of the premium set aside to cover commissions and other administrative costs, and may include a “margin” which is intended as a hedge against higher than anticipated costs. The expense charge is typically between 5% and 8% of the annual premium, including the margin which is intended to reasonably relate to the risk that higher than anticipated charges will materialize. In CM Holdings, the annual premiums and loan interest payments in years four through seven were financed by cash, partial withdrawals, and a device called a “loading dividend.” Id. at 593. By design, 95% of the gross premium was taken as an expense charge, known as the “loading charge.” Id. at 593. In actuality, the expenses were between 5% and 8% of the gross premium. Id. at 593. The difference between the “loading charge” and MBL’s actual expense charges generated excess funds in the policies which resulted in the payment of a “loading dividend,” which amounted to about 92% to 95% of the loading charge. Id. at 594. In simultaneous netting transactions, the loading dividend was used to pay the premium. Id. at 593. A partial withdrawal was taken in an amount equal to about 99% of the policy loan interest payment and used to make that payment. Id. at 593. The Camelot COLI VIII policies did not limit the amount of policy value that could be taken as a partial withdrawal. Id. at 594. Camelot paid the balance of premiums (about 5%) and loan interest payment (about 1%) in cash. Id. at 593. Although Camelot originally planned to take policy loans in years eight and nine, those loans were never taken. Id. at 592 n. 16. During the first eight years, $31.3 million in policy loan interest accrued. Id. at 607. Camelot paid $12 million, or about 40% of the policy loan interest, in cash. Id. at 607. Cash value was stripped from the Camelot COLI VIII policies by means of a highly efficient computer program designed to achieve zero net equity on the last day of each policy year, id. at 595, such that the net equity of the policies “would not exceed one penny.” Id. at 631. Moreover, the Camelot COLI VIII plans were designed to be “mortality neutral.” This meant that the cumulative COI charge paid by Camelot was anticipated to equal the cumulative death benefit that would be distributed to Camelot, with the exception of a profit margin to the insurance company of 20% of the COI charge in the first plan year, 10% in the second plan year, and 2% thereafter. Id. at 632-33 (footnote omitted). Although so designed, the Camelot COLI VIII plan did not operate in a mortality neutral way. Id. at 633. Over the first seven years, Camelot received death benefits and mortality dividends $1.3 million higher than its COL Id. at 633-34. This was the result of pooled dividends in the first three policy years and a combination of pooled and experience-rated dividends in the fifth policy year; the experience-rated dividend was totaled only $293, all received in the fifth policy year. Id. at 634-35. In that fifth policy year, the Camelot COLI VIII plan moved to an experience-rated approach. MBL explained that [t]he new mortality mechanism will result in a much closer match between the expected cash flow from projected death benefits and claims that are actually paid. This will minimize any volatility or variation in the cash flows and corporate earnings which are expected in each year of the plan. Id. at 635. With the passage of HIPA in 1996, Camelot stopped paying premiums and allowed the policies to function as paid-up policies for a reduced amount of death benefit coverage. Id. at 641. This resulted in a $30 million reduction in death benefits causing a $30 million partial withdrawal which forced all of Camelot’s taxable policy gain out of the policies. Id. at 641. Albeit cleverly designed, the COLI VIII plan ran afoul of the governing case law concerning so-called “shams in fact” and “shams in substance,” according to the district court. Id. at 598. Although the policy loans used in the first three years and the interest accrued on them were found to possess factual substance, the loading dividends used in years four through seven of COLI VIII did not fare so well. In order to calculate the dividend amounts, the policies charged administrative fees considerably out of proportion to the actual costs incurred. Furthermore, instead of distributing the dividend from accumulated surplus, the policies “sourced” their dividends from excessive loading charges which were instantaneously offset against payment of the premium. The lack of any contributions to the dividends from investment yields was also suspicious. The district court was also troubled by the payment of the dividends at the beginning of the policy year, rather than the industry-standard practice of paying at the end of the year. The legitimacy of the dividends was further compromised by the fact that they were guaranteed, not contingent. Topping it off, the designers of the COLI plan did not treat the dividends as a liability on its balance sheets. The combination of all these factors provided “overwhelming evidence” to the district court that the loading dividends were factual shams. Id. at 617-20. The district court further concluded that the COLI plan, as a whole, was a sham in substance that lacked any rational economic purpose other than the creation of tax savings through interest-payment deductions. The court’s determination centered on two factors: “the objective economic substance of the transactions and the subjective business motivation behind them.” Id. at 621 (citation and internal quotation marks omitted). From the objective standpoint, all of the pre-purchase illustrations projected Camelot’s cash flows to be negative absent the interest deductions in all years of the plan and in the aggregate. In addition, projections considered by the court, when discounted to present value, made it clear that the plans generated negative cash flow absent the interest deductions and positive cash flow only when those deductions were considered. This conclusion was bolstered by the economic neutrality of the plans, which were designed to preclude any cash build-up in the policies. Rather, any value received was immediately stripped from the policies in the form of loans or dividends. Similarly, the policies’ focus on “mortality neutrality” meant that there was no risk involved for either party with respect to the disbursement of death benefits, unlike traditional insurance contracts. Id. at 637. Camelot’s subjective intent in adopting the COLI plan did not salvage the transactions. Although the court saw no reason to discredit Camelot’s explanation that it purchased the COLI policies to offset the cost of medical benefits, that alone was not sufficient to confer economic substance on what was otherwise an “economically empty transaction.” Id. at 638. Camelot’s position was also undermined by evidence indicating that the interest deductions played a crucial role in convincing Camelot to purchase the policies, and by management’s awareness prior to purchase that the policies could not be profitable on a pre-tax basis. Id. at 639-41. The district court further concluded that Camelot could not save the deductions taken in years one through three via the four-of-seven safe harbor provided by I.R.C. § 264(c)(1), which permits deductions on policy loans taken from life insurance policies as long as “no part of 4 of the annual premiums due during the 7-year period (beginning with the date the first premium ... was paid) is paid under such plan by means of indebtedness.” 26 U.S.C. § 264(c)(1). The court concluded without contest by the policyholder that the statute implied that the premiums paid must be level throughout the seven-year period. Id. at 645. Once the court deducted the portions of those premiums that were “paid” by loading dividends, the remaining amounts actually paid were substantially lower that those paid in previous years with the aid of the policy loans (which were found not to be factual shams). The result was that the effective amount of the premiums paid was not truly level throughout the first seven years of the policy, and the taxpayer could not take advantage of the safe harbor of Section 264(a)(2), which otherwise precludes deductions for premium financing arising from a corporate life insurance plan “which contemplates the systematic direct or indirect borrowing of part or all of the increases in the cash value” of the policy. Having violated the implicit requirement that premiums remain level, the interest deductions taken for the first three years of the policies were also invalid. Id. at 646-47. 3. American Electric Power, Inc. v. United States The COLI plan at issue in AEP was almost identical to the one considered in CM Holdings. See AEP, 136 F.Supp.2d at 768-69. In that case, the plaintiff, American Electric Power Company, purchased COLI policies on its employees’ lives intending to offset the effect on its earnings reports of new accounting requirements for expensing post-retirement medical benefit obligations. Although the death benefit of each policy varied according to the age of the employee, the other terms were identical. Id. at 774. Each policy in the plan provided fixed, annual premiums of $16,667 that were designed to permit the accumulation of $50,000 in cash value within three years. Over the first three years, AEP would pay approximately ninety percent of the premiums through policy loans “in simultaneous netting transactions in which the loans were offset against the premiums.” Id. at 776. This structure allowed AEP to pay approximately $23.5 million in cash, including administration fees, for premiums costing $330 million. That $23.5 million, in turn, would be offset by deductions for interest paid on the policy loans and the receipt of tax-free disbursements in the amount of $10.8 million. AEP thus ended up with $3.5 million of positive cash flow in the plan’s first year, rising to $3.9 million in the second year, $10.5 million in the third year, and more than $35 million after eight years. Id. After the third policy year, borrowings to pay premiums approached the $50,000 policy loan limit of I.R.C. § 264, so premium payments and accrued loan interest were paid using dividends and partial withdrawals in simultaneous netting transactions once per year. Ninety-five percent of each premium was considered an expense charge by the insurer, and the remainder was returned to AEP as a “loading dividend.” Id. No premiums were due from that point forward, although the policy loans remained outstanding and generated more than $100 million of interest expense each year, offset by cash withdrawals which stripped excess cash from the policies, resulting in zero net equity. Through year 20, positive cash flows of $35 million to $39 million were projected, after the tax deductions were taken into account. In addition, the court found that these policies were “mortality neutral,” since there was no risk to either party that the amount of death benefit paid would vary from the COI. This was achieved by means of an annual dividend that would refund excess monies after any year that morality “prove[d] more favorable to the [insurance] company than expected.” AEP, 136 F.Supp.2d at 777. The net effect of these structures was to create a scheme in which policy payouts would exactly offset the premiums paid less a fixed profit for the insurer. The AEP Court agreed with the CM Holdings Court that this design was problematic. First, although the Court agreed that the loading dividends in years four through seven could be workable, the mortality neutral design of the COLI scheme eliminated any “reasonable relationship to the insurance company’s risk of incurring higher than expected expenses.” Id. at 782. Second, the AEP court agreed with CM Holdings that the Code’s “four-of-seven” rule required the premiums to be level in all events. Third, the AEP COLI plan as a whole was an economic sham because its neutrality on mortality and absence of excess cash in the policy gave it no “practicable economic effect other than the creation of income tax losses.” Id. at 785 (quoting Rose v. Comm’r, 868 F.2d 851, 853 (6th Cir.1989)). Finally, although the court determined that simultaneously borrowing against the policy and paying that year’s premiums with those loan proceeds was not a factual sham, the first-year loans were suspiciously backdated in a manner inconsistent with industry practice. When AEP decided to participate in the Mutual Benefit Life Insurance (MBL) COLI VIII plan, it signed prepayment agreements on February 16, 1990 that allowed AEP to purchase COLI on its employees between that date and the date the policies were finally issued. The policies were issued on March 23, 1990, and MBL originally calculated March 23, 1990 as the inception date for the policies. Because using the March date cost it almost $2 million in deductions, AEP complained, and Integrated Administration Services (IAS), which was administering the program for MBL, responded by backdating the inception date to February 16, 1990, when the prepayment agreements were signed. The Court found this to create a sham in fact with respect to interest “accrued” between February 16 and March 23, 1990 because loans were never issued during the prepayment period, and AEP paid no interest during that time. Furthermore, since the policies were considered to have been issued to a grantor’s trust on March 21, 1990, the trust became liable for debt that predated its existence. Id. at 781-82. The court brushed off, with little explanation, AEP’s contention that these sort of “bridge” agreements were common in the life insurance industry for applicants who wanted to be protected while the policy was being considered and processed: While there was evidence presented showing that it is common in the life insurance industry to issue a conditional receipt at the time of taking an application for life insurance and to provide death benefit coverage from the date of the receipt — after it has been determined that the insured meets the company’s underwriting requirements— there was no evidence presented of a custom or practice to backdate policy loans in the manner in which it was done in the instant case. The court concludes that the manner in which IAS originally intended to calculate the first-year policy loan interest is more likely the industry norm for a transaction of this kind. Id. at 782. Thus, in these three cases, the courts generally determined that the COLI plans constituted economic shams, functioning only as interest-deduction engines that drove no legitimate financial vehicles. The courts pointed to artificially high loan interest rates which had no practical adverse effect on the borrower because the fixed spread correspondingly drove up the supercharged credited rate on borrowed funds; relatively small interest rates on unborrowed funds to discourage leaving cash in the policies; the elimination of mortality risk by means of fully retrospective, annual equalization of COI and death benefit payments; the use of unconventional “loading dividends” as a means of paying premiums in four of the first seven years; an exquisitely efficient computer program which stripped virtually all equity from the policies year after year; and prepurchase illustrations which showed only negative cash flows without the tax deductions in each of the policy years through the duration of the programs. These courts concluded that there could be no profit on premature mortality because of the retrospective adjustment which eliminated mortality risk in advance, and no return on inside build-up because all the cash was stripped from the policies throughout the program. As will be explained in greater detail below, the COLI plans purchased by Dow were similar, although not identical to, the COLI plans in the cases described above. They all involved cash value policies that compressed the premium payments into a relatively short period, they called for a highly-leveraged premium financing strategy that avoided policy loans in years four through seven, they all carefully monitored the relationship between COI and the actual death benefit payout, and they all were used to fund in the aggregate future corporate benefit obligations. There are critical differences in Dow’s plans, however, which preclude finding that the plans are factual shams. The Court concludes that there was an economic benefit that potentially could be derived from the plans without relying solely on the tax deductions for policy loan interest. As will be explained, Dow articulated a legitimate purpose for embarking on the programs: providing a source of cash to cover unfunded future medical obligations for its retirees. Dow had turned to similar devices in the past, albeit on a much smaller scale, to cover contingent liabilities. E. Dow’s 1983 and 1985 COLI Policies In 1983, Dow purchased 89 COLI policies from Connecticut General Life Insurance Company (CIGNA). Stip., ¶ 28. Purchased on senior executive-level employees, including Dow’s President, its Chief Executive Officer, its Executive Vice Presidents, and a number of Senior Directors, the policies, dated December 1983, were intended to provide a funding source for deferred compensation obligations Dow had to those employees. Stip., ¶ 28. These policies were purchased to protect Dow from financial damage in the event of loss of its key employees’ unique knowledge and expertise and to offset unfunded liabilities that Dow had to the insured employees. Pierce, Tr. at 4276, 4328. The premiums and death benefit varied for each policy according to the salary level of the insured employee. Pierce, Tr. at 4278-80; Ex. D970. Dow paid the annual premiums for the first, fifth, sixth, and seventh policy years in cash; annual premiums for the second, third, and fourth policy years were paid using policy loans. Pierce, Tr. at 4281; Ex. D970. Neither first-day, first-year policy loans nor partial withdrawals were taken with these policies. Pierce, Tr. at 4281; Ex. D970. These policies were acquired and subsequently monitored by Dow’s Human Resources department. Falla, Tr. (1/9 a.m.), at 32-33. In 1988, Clark/Bardes, Inc. (Clark/Bardes), a broker, took over administrative support for the CIGNA policies. Stip., ¶ 29. In 1986, Dow purchased 52 COLI policies from Great West Life Assurance Company (Greah-West). Stip., ¶ 30. These policies, dated November 10, 1985, were purchased on senior executive-level employees to offset Dow’s unfunded deferred compensation liabilities to those employees. Stip., ¶ 30; Pierce, Tr. at 4334; White, Tr. (1/10 a.m.), at 12. These policies were individual, traditional, participating, whole-life, paid-up-at-age-95 base policies with an attached rider. Ex. P91. Level annual premiums were payable at issue and continued until age 95. Ibid. Like the 1983 COLI program, the annual premiums and death benefit on the 1985 COLI policies varied according to the salary level of the insured employee. Pierce, Tr. at 4285-87; Ex. J1191. Dow paid the annual premiums for four out of the first seven policy years in cash; annual premiums for the other three policy years were paid using policy loans. Pierce, Tr. at 4281. Neither first-day, first-year policy loans nor partial withdrawals were taken with these policies. Id. at 4292. The purchase of these policies was again handled by Human Resources; however, Clark/ Bardes provided the administrative support for the management of the policies. Stip., ¶ 31. Both the 1983 CIGNA and apparently the 1985 Greatr-West COLI policies provided for a variable loan interest rate based on Moody’s Corporate Average. Ex. Jll; Ex. P91. Moody’s Corporate Average is a long-term index which approximates the rate of return earned on assets invested by insurance companies. By charging interest at the Moody’s Corporate Average rate, an insurer avoids discrimination between its borrowing and non-borrowing policyholders and protects the insurance company from the risk of disintermediation, that is, the risk that a policyholder will borrow cash from the insurance policy at below market rates when returns on investments are considerably higher than the policy loan rate. Todd, Tr. at 636-37; Puglisi, Tr. at 5276-77, 5278; DesRochers, Tr. at 35970-98, 3599; Hoag, Tr. at 6205-06. Also, by using Moody’s Corporate Average as a borrowing rate for policy loans, an insurance company decreases the sensitivity of the insurance company’s financial statements to policyholder borrowing. Plotkin, Tr. at 4085-86; DesRochers, Tr. at 3598. The cash flows generated by the 1983 and 1985 COLI programs were not segregated as specific pools of funds for the purpose of offsetting Dow’s unfunded liabilities to the employees insured; rather, the cash flows generated were fungible and could have been used for any corporate purpose. Pierce, Tr. at 4328, 4334. Dow’s interest deductions relating to the 1983 CIGNA and 1985 Great-West COLI policies were not challenged by the IRS. Pierce, Tr. at 4335. F. Dow’s Purchase of the Great West COLI Policies 1. Dow’s Business Purpose: Mounting Employee Retirement Medical Expenses The 1983 and 1985 COLI programs were purchased to fund deferred compensation obligations. In 1982, Dow retained an outside consulting firm, Watson Wyatt Worldwide (Wyatt), to calculate the present value costs of its future liabilities for retiree benefits and the annual’ contribution required to fund those benefits. Ex. J7; J9; Falla, Tr. (1/8 a.m.) at 121. For each dollar Dow pays an average employee in salary, it pays that employee an additional 22 to 25 cents in benefits. Falla, Tr. (1/18 a.m.) at 115. The estimated present value of Dow’s accrued post-retirement life and medical plan liabilities in 1983 was approximately $500 million. Ex. J9; Pierce, Tr. at 4336. The estimated liabilities rose to over $1 billion by 1987. Ex. J73. The increase was the result of the escalating cost of benefits and the significant increase in the utilization of medical, health, and life insurance benefits. Falla, Tr. (1/8 a.m.) at 117, 123; Lake, Tr. (1/10 p.m.) at 85. Dow executives worried that Dow might no longer be able to offer post-retirement benefits to its employees unless it found a way to fund these costs. White, Tr. (1/10 a.m.) at 65. In 1983, Wyatt assigned Gary Lake as Dow’s actuarial consultant. Lake, a member of the American Academy of Actuaries since 1977, started at Wyatt in 1973 and became its national resource on COLI-related matters, advising approximately fifty clients. As Dow’s actuarial consultant, Lake assisted Dow in evaluating whether COLI was a viable means for funding the employee benefit obligations identified, educated Dow as to insurance products and risks inherent in COLI, and helped develop specifications to solicit and compare bidder’s proposals. Lake, Tr. (1/10 p.m.) at 73-74, 78. Lake advised Dow on its purchase of all its COLI programs: the 1983 CIGNA and 1985 Great-West, as well as the two COLI programs in this case. 2. Changes in Accounting Requirements for Retiree Medical Expenses Prior to 1989, Dow accounted for its retiree medical benefits on a pay-as-you-go basis. Falla, Tr. (1/8 a.m.) at 116-17; Brink, Tr. at 2398-99. In February 1989, the Financial Accounting Standards Board (FASB) issued Exposure Draft 105, Employer’s Accounting for Post Retirement Benefits Other than Pensions. If implemented, this document would have required employers to accrue current liabilities for retiree medical and life insurance benefits on a current basis for the purpose of their financial statements. Stip., ¶ 32; Lake, Tr. (1/10 p.m.) at 81. Exposure Draft 105 would have required Dow to accrue $1.6 billion as the present value of its accrued retiree medical liabilities in 1989. Stip., ¶ 33. In December 1990, the FASB issued FASB Statement No. 106 (FAS 106), Employer’s Accounting for Post-Retirement Benefits Other than Pensions. FAS 106 adopted the requirements of Exposure Draft 105, effective for fiscal years beginning after December 15, 1992. Stip., ¶ 34. A 1991 calculation placed Dow’s accrued retiree medical liabilities at $1.34 billion. Ex. J411. Dow adopted FAS 106 effective January 1, 1992. Stip., ¶ 35. In making the transition to this accounting standard, Dow took an after-tax charge of $994 million against its 1992 income for unfunded retiree obligations with a net present value of $1.45 billion accrued to date. Ex. J641. The liability continued to grow. Ex. J722 ($1.54 billion in 1993); Ex. J774 ($1.62 billion on January 1,1994). 3. Requests for Proposals (RFP) Dow officials decided to explore COLI as a means of funding these liabilities and, as before, turned to Gary Lake for advice. The “kick-off’ date for the exploratory project, as Lake referred to it, was April 1, 1987. Because the Human Resources Department acquired Dow’s other COLI policies, Loren Pierce, Dow’s Manager of Executive and International Benefits at the time, was approached by Clark/Bardes about a new version of corporate-owned life insurance on the market, known as “COLI III,” to fund various employee benefit liabilities. Ex. J12; Ex. J14; Pierce, Tr. at 4303-04. In April 1987, Dow, assisted by Lake, initiated a review of COLI as a potential funding vehicle for Dow’s post-retirement liabilities. Ex. J58; Lake, Tr. (1/14 a.m.) at 5. On May 28, 1987, Pierce wrote a memo inviting various Dow employees to a presentation on COLI III, noting that “COLI III is a significantly different product than we have previously seen and is becoming more of an investment than benefit funding mechanism; thus we thought more financial people should hear about it.” Ex. P5; Falla, Tr. (1/8 a,m.) at 137-40. The proposed transactions involved large and complex financial issues, but also addressed liabilities with which Human Resources was familiar. Nevertheless, because COLI was described as more of an investment than a benefit funding mechanism, Pierce contacted representatives from Dow’s Treasury Department. Ex. P5; Falla, Tr. (1/8 a.m.) at 140; Pierce, Tr. at 4337-39. That spring, Lake conducted a two-step evaluation of COLI for Dow. In step one, he evaluated Dow’s accrued post-retirement benefit liabilities; in step two, he helped Dow evaluate COLI as a potential funding vehicle for the liabilities. Ex. J55; Ex. J60; Lake, Tr. (1/10 p.m.) at 92. Lake made a side-by-side comparison of the net present value of the benefit liabilities and the projected after-tax cash flows of a COLI program, even though this was not necessary because the benefit liabilities far exceeded the projected earnings. Lake, Tr. (1/10 p.m.) at 93; Lake, Tr. (1/14 a.m.) at 15-16. Between June 12 and August 25, 1987, Dow employees heard three presentations on COLI III. The June 12, 1987 presentation by Clark/Bardes detailed the then-recent amendment to I.R.C. § 264 which limited interest deductions to $50,000 of policy loans per employee. Ex. J65. The participants at the meeting also discussed future tax law changes, likely insurance carriers for Dow to consider, and issues to consider in the decision-making process. Ex. J65. On August 4, 1987, a representative for Management Compensation Group (MGG), a broker, gave a presentation regarding a COLI product offered by Mutual Benefit Life. Ex. J71. After this presentation, Dow representatives made two decisions: 1) Proceed with COLI for [approximately] 1200 employees in the MIP [Management Incentive Program], This will likely be done through Clark Bardes with Conn. Mutual since forms that were signed by employees contemplate this. 2) The new COLI concept with MCG will be investigated for up to 4 or 5,000 additional employees strictly as an investment vehicle to take advantage of the tax arbitrage. Ex. J71, at A011224; Stip., ¶ 69. On August 25, 1987, representatives from Clark/Bardes made a presentation regarding a proposal for a Connecticut Mutual COLI plan covering approximately 1,200 Dow employees. The Connecticut Mutual COLI product would offer a guaranteed dividend formula, guaranteed mortality, and a 75-basis-point spread with the “[g]ain to Dow com[ing] from tax leverage.” Ex. J75, Bates A011212; Bur-dett, Tr. at 1492-93. As noted above, the “spread” is the difference between the interest rate earned by Dow on the money in the policy (i.e., the credited rate) and the interest rate charged to Dow on loans taken (i.e., the loan rate). The next day, the COLI Task Force was formed by Enrique Falla, Dow’s Chief Financial Officer, and Jay Hornsby, head of Human Resources. Falla, Tr. (1/8 a.m.) at 142. The COLI Task Force was chaired by Glenn White, director of taxes and ex-officio member of the finance committee. Id. at 142-43. The other members were Anita Jenkins, an attorney who worked in the tax department on benefits issues; William Wales and Janet VanAl-sten, attorneys assigned to work with the human resources department; Pierce; Bill Schmidt, assistant comptroller; and Howard Burdett, assistant treasurer. Stip., ¶ 70. Falla provided the Task Force members a “road map” of issues to be addressed before proceeding with a transaction, consisting of: (1) tax issues, including concerns with the then-current law and with prospective legislative changes to that law; (2) legal issues, including the question of insurable interest; and (3) financial issues. Ex. J76; Ex. J77. The COLI Task Force issued a report on September 8, 1987. The report indicated that “[COLI] is a program that the corporation may find suitable for providing funding for employee benefit plans or perhaps for other general revenue purposes.” Ex. J1218. According to Falla, the “general revenue purposes” language was used to give Dow a “little latitude in the event [it] want[s] to modify a program.” Falla, Tr. (1/8 a.m.) at 157. a. Dow’s pre-purchase analysis of tax issues In examining the tax issues, the tax force recognized that there are three tax benefits generally available through the corporate use of leveraged cash value life insurance: (1) the tax-free build-up of the cash value; (2) the tax-free payout of death benefits; and (3) the deductibility of interest up to a $50,000 loan cap. The defendant’s experts acknowledged that corporations like Dow should take full advantage of the tax laws to minimize their tax liability and maximize shareholder wealth. Puglisi, Tr. at 5311, 5312; Hoag, Tr. at 6409-10, 6411, 6414, 6356. Dow’s Tax Department, with the assistance of Lake, researched and resolved several tax issues prior to Dow’s purchases of its 1988 and 1991 COLI, including evaluating Dow’s options for unwind scenarios if the tax law changed. White, Tr. (1/10 a.m.) at 13, 20-21; Jenkins, Tr. at 462021; J67. On August 10, 1987, Jenkins authored a memorandum to White identifying potential tax issues concerning COLI, including: (1) the qualification of the policy as real life insurance under Code Section 7702, and (2) the use of financing mechanisms to pay premiums in years four through seven as potentially violative of Section 264. Ex. J72 at H00229-H00231; Jenkins, Tr. at 4535-37. White followed up on Jenkins’s memorandum with a memorandum to Enrique Falla on August 25, 1987. He identified for Falla the primary tax exposure risks of COLI as outlined in Jenkins’s August 10 memorandum. White also noted that Dow should incorporate a suitable cancellation provision in the contract in case Congress enacted legislation curtailing the tax benefits of COLI. Ex. J76; White, Tr. (1/10 a.m.) at 1516. On October 22,1987, Jenkins again identified the Tax Department’s concerns regarding the design of COLI policies, advising that if Dow were to proceed with the purchase of COLI, then several conditions had to be met, including: (1) satisfaction of Code Section 264 by avoiding characterization as a single premium policy and complying with the so-called “four-out-of-seven” safe harbor; and (2) satisfaction of the requirements of Code Section 7702, including qualification as insurance under state law. Ex. J124; Jenkins, Tr. 4531-35; White, Tr. (1/10 a.m.) at 31. Gary Lake likewise identified as risks (1) the use of partial withdrawals or loading dividends to finance premiums in years four through seven under Section 264; (2) the definition of life insurance under Section 7702; and (3) the risks of sham transaction characterizations resulting from aggressive policy features such as policy loan rates above Moody’s Corporate Average and the use of arbitrary dividend payments to minimize cash flow. Ex. J60; Ex. J67 at A011245. Ultimately, White, Lake, and the COLI Task Force were satisfied that they had resolved all of their issues favorably. White, Tr. (1/10 a.m.) at 70-71, 73-77. Dow planned to use partial withdrawals instead of loading dividends to pay premiums in years four through seven. Lake, Tr. (1/14 a.m.) at 20-21, 92-93. Dow representatives were never comfortable with the use of unconventional “loading dividends” for that purpose. With respect to Section 7702, Lake provided actuarial verification that the COLI programs Dow considered satisfied the alternative mathematical tests of that section. Lake, Tr. (1/14 a.m.) at 14. Dow decided that the Great-West policies were not subject to Section 7702A. DesRochers, Tr. at 3564-65. As for financing premiums with policy loans, Dow rejected the use of enhanced policy loan interest rates and instead selected Moody’s Corporate Average as an appropriate policy loan interest rate. According to Lake, Dow’s selection of Moody’s Corporate Average reflected Dow’s desire to be on the “conservative side” of the policy loan feature. Lake, Tr. (1/14 a.m.) at 19, 59-60; White, (1/10 a.m.) Tr. at 36-38. On mortality issues, Dow and Lake determined that a COLI product that trued up COI charges based on actual mortality at the end of each policy year (i.e., a product that was 100% experience-rated) would violate the requirement that an insurance policy transfer risk to the insurer. Lake, Tr. (1/14 a.m.) at 124, 135. Although Dow and Lake wanted to reflect Dow’s favorable mortality experience in the mortality charges, Dow wanted to avoid a 100% experience-rated policy. Dow also was mindful of the shifting landscape in tax legislation impacting COLI; the Task Force considered the possibility of legislative action that would diminish various tax advantages of COLI, particularly the tax-free character of the inside build-up and the deductibility of policy loan interest. J77 at A011205; Falla, Tr. (1/8 a.m.) at 149. White, as tax director, monitored pending tax legislation through Dow’s Washington office, as did Paul Brink, White’s successor. b. Dow’s pre-purchase analysis of legal issues Among the legal issues explored was whether Dow had an insurable interest in all of the proposed insured employees. Jenkins raised the issue of insurable interest in an August 10, 1987 memo to White, concluding that a “corporation has an insurable interest in the lives of its officers, directors or managers (key men); however, the mere existence of the relationship of employer and employee is not sufficient to give the employer an insurable interest in an employee.” Ex. J72. On October 5, 1987, Kirkland & Ellis, an outside law firm, advised Dow that it could structure a COLI program “so that the policies issued will not be void as a matter of public policy due to a lack of insurable interest.” To do so, Kirkland & Ellis recommended that Dow (1) require all insureds to consent to insurance coverage, and (2) purchase an aggregate amount of insurance that was proportionate to the perceived potential liabilities to employees under all benefit plans. Ex. J107; White, Tr. (1/10 a.m.) at 48. MetLife also brought the insurable interest issue to Dow’s attention in its response to Dow’s November 1987 RFP. There, MetLife attached a memo on insurable interest in large employee populations from Roy Albertalli, a member of Met-Life’s Law Department. Albertalli identified the same issues that had been raised in Jenkins’s and Kirkland & Ellis’s memo-randa. Ex. J136 at A03870; Ex. J797; Lake, Tr. (1/14 a.m.) at 42; Ryan, Tr. at 1725:7-24. Since the 1960s, Dow had used a “Hay Point” system to measure the importance of its employees to the corporation. The system evaluates and quantifies a job’s content and value on three dimensions: know-how, problem-solving, and accountability. Falla, Tr. (1/8 p.m.) at 18; Pie