American Elec. Power v. United States, 136 F. Supp. 2d 762 (S.D. Ohio 2001)

US District Court for the Southern District of Ohio - 136 F. Supp. 2d 762 (S.D. Ohio 2001)
February 20, 2001

136 F. Supp. 2d 762 (2001)

AMERICAN ELECTRIC POWER, INC., et. al., Plaintiffs,
UNITED STATES of America, et al., Defendants.
American Electric Power, Inc., et al., Plaintiffs,
United States of America, et al., Defendants.

Nos. C2-99-724, C2-98-304.

United States District Court, S.D. Ohio, Eastern Division.

February 20, 2001.

*763 *764 *765 Randolph C. Wiseman, Bricker & Eckler, Columbus, OH, Arthur L. Bailey, J. Walker Johnson, Jame P. Holden, Susan H. Serling, Jean Male Baxley, Steptoe& Johnson, Washington, DC, Richard E. May, Hunton & Williams, Washington, DC, of counsel, for Plaintiffs.

Dennis M. Donahue, Sp. Litigation Counsel, Office of Civ. Litigation, Tax Div., Washington, DC, James D. Hill, Office of Chief Counsel, Cincinnati, OH, Alex A. Sadler, Charles M. Ruchelman, Joseph A. Sergi, Tria; Atty's., Dept. of Justice, Tax Div., Washington, DC, Sharon J. Zealey, U.S. Atty., Deborah F. Sanders, U.S. Atty's. Office, Columbus, OH, for Defendants.


GRAHAM, District Judge.


This is a tax refund suit initiated by American Electric Power Company, Inc. ("AEP") to recover an alleged overpayment of federal income tax with respect to its 1996 tax year. AEP is a public utility holding company organized under the laws of the State of New York. Its principal office is in Columbus, Ohio. AEP is one of the largest shareholder-owned electric utility companies in the United States, operating in Ohio, Indiana, Kentucky, Michigan, Tennessee, Virginia, and West Virginia. AEP owns various operating, service, and coal-producing subsidiary corporations. The amount of tax in dispute with respect to the 1996 tax year for all the AEP companies is $25,478,773. The resolution of the tax controversy in this case will affect other AEP tax years, including 1991 to 1995, 1997 and 1998.

This case arises out of AEP's participation in a highly leveraged broad-based *766 program of corporate-owned life insurance ("COLI"), wherein AEP purchased life insurance policies on the lives of over 20,000 employees constituting most of its work force. AEP was the beneficiary of these life insurance policies. A substantial amount of the premium cost for the COLI policies was financed by loans secured by the cash value of the policies. The COLI program was designed to maximize the tax benefits of the deductibility of interest on the policy loans, the deferral of taxation on the increase in cash value, and the ultimate non-taxability of the death benefits.

The Internal Revenue Service ("IRS" or "government") has disallowed AEP's deductions for interest paid on the COLI policy loans for 1996. The government claims that the policy loans, dividends, and partial withdrawals of policy value, which were used to pay premiums, lacked factual or economic substance and were sham transactions, i.e., that they were shams in fact. Alternatively, the government maintains that the AEP COLI plan, taken as a whole, had no practical economic consequences, other than the creation of income tax benefits, and therefore, was a sham in substance. The government also contends that when these transactions are viewed in their proper light, the relationship of AEP's borrowing to its net premiums, i.e. gross premiums minus dividends and partial withdrawals, the plan fails the "four of seven" test set forth in I.R.C. § 264(c) (1).[1] The Internal Revenue Code generally disallows deductions on amounts borrowed on an insurance policy to pay premiums. See I.R.C. § 264(a) (3). However, this exclusion does not apply if "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[.]" § 264(d) (1) (known as the "four out of seven safe harbor").

AEP contends that its COLI plan and its use of policy loans, policyholder dividends, and withdrawals to pay premiums and accrued policy loan interest were all within the intendment of the provisions of the Internal Revenue Code that governed the federal income tax treatment of life insurance and the deductibility of interest on policy loans during the years in question. AEP asserts that the policy loans, dividends and policy withdrawals were real transactions. AEP also claims that its COLI plan had objective economic substance and was undertaken for a valid business purpose, and that the interest on the policy loans was properly deductible under I.R.C. § 163(a).

General principles of life insurance

This case involves many of the unique attributes of life insurance. There are two basic kinds of life insurance: term insurance, which, as its name implies, provides a death benefit for a specified term of years; and whole life insurance, which provides a death benefit throughout the lifetime of the insured. Term insurance policies are priced on the basis of the actuarially-determined cost of providing the stipulated death benefit during the term of the policy. This is referred to in the insurance industry as the cost of insurance ("COI"). The premium for a term insurance policy includes a profit for the insurance company, but the contract provides no benefits to the insured other than the death benefit. Whole life policies, on the other hand, are customarily priced to permit the accumulation of value in addition to the cost of providing the death benefit. These policies provide a benefit known as cash value, which may be accessed by the policyholder during the insured's lifetime through policy loans or withdrawals of policy value. The accumulation of cash value within the policy may be likened to a savings account. The cash value of the policy represents not only the excess of premiums *767 over COI charges, but also additional credits added by the insurance company in the nature of interest. The source of these additional credits is the income generated by the insurance company's investments. This additional contribution to cash value is referred to as "inside buildup."

The owner of a whole life insurance policy which has accumulated cash value has a contractual right to obtain a loan from the insurance company, referred to as a policy loan, whereby the policyholder pledges the cash value of the insurance policy as collateral for the loan. The loan is deemed to have been made from the general assets of the insurance company, not as a withdrawal from the accumulated cash value of the policy. If the insured dies before the loan is repaid, the amount of the loan and accrued interest is deducted from the death benefit payable under the policy. Some whole life insurance policies permit the policyholder to access the cash value by making withdrawals therefrom. Withdrawals reduce the death benefit; however, the policyholder has no obligation to repay amounts withdrawn, nor is there any interest obligation on the amount withdrawn.

Many whole life insurance policies have an additional mechanism for distributing benefits to the policyholder, known as dividends. Such policies, which are called "participating" policies, provide the opportunity for the insurance company to distribute additional value to the policyholder in the form of dividends. Dividends may be paid in cash, credited against premiums due, or they may be accumulated and added to the cash value. Generally, dividends are paid when the insurance company's operating results have exceeded its profit expectations and there are additional funds available for distribution to policyholders. Policyholders do not have a right to receive dividends, they are payable in the discretion of the insurance company.

Today, there is a great deal of flexibility in the structuring of life insurance policies with respect to premiums, death benefits and other benefits. Premiums may range from a single payment to a lifetime of payments. They may be fixed or variable in amount. Death benefits may be fixed or variable and there are many permutations of other policy benefits, such as cash value, policy loans, and dividends. In the past, the traditional whole life insurance policy provided a fixed death benefit in return for fixed premiums payable for the lifetime of the insured, usually deemed to be ninety years. Such policies accumulated cash values in accordance with a schedule stipulated in the policy. The policyholder was not given any information about the insurance company's cost of providing the death benefit, the amount of premium allocated to cash value, or the rate at which additions would be credited to cash value. The investment or savings component of the life insurance contract was hidden. In the late 1970s, a new form of cash value life insurance, known as "universal life," was introduced to the market. In a universal life policy, all of the economic components of the policy are revealed. The introduction of universal life had a significant impact on the life insurance industry by providing a product that could be marketed with an emphasis on investment goals and returns. With the introduction of universal life, the flexibility of life insurance as an investment vehicle was enhanced and the variations in the kinds of policies multiplied.

The tax advantages of life insurance

Historically, Congress has deferred taxation of the inside buildup that occurs within a cash value life insurance policy. Additionally, death benefits have been exempt from taxation. Congress has, however, *768 regulated and limited these substantial tax benefits by imposing restrictions on the amount of the investment component which can be built into a life insurance policy and by regulating the extent to which policyholders are permitted to access cash values without incurring tax consequences.

Prior to the enactment of the Health Insurance Portability and Accountability Act of 1996 ("HIPA"), Pub.L. No. 104-191, 110 Stat.1936, 2090, corporate policyholders enjoyed another significant tax advantage, namely the full deductibility of interest on policy loans. The combination of deferral of taxation of inside buildup, the exemption of taxation of death benefits, and the deductibility of interest on policy loans created the opportunity for tax arbitrage by corporate policyholders.[2] Not surprisingly, insurance industry entrepreneurs developed products for the corporate market which exploited this opportunity. Similar, although less sophisticated, products had been developed and marketed to high income individuals before Congress eliminated the deductibility of ordinary interest expense for individuals in the 1986 Tax Reform Act.

Corporate-Owned life insurance (COLI)

Corporations have, for many years, purchased life insurance on the lives of their most valuable employees in order to protect the corporation against economic losses which would occur as a result of the untimely death of such an employee. This form of COLI is commonly known as key person life insurance. After it became widely accepted that corporations had an insurable interest in key employees, insurance companies began to market COLI as a funding vehicle for deferred compensation plans, where the lives of the beneficiaries of the deferred compensation plans, usually executive-level employees, were insured and the death benefits were used to fund the deferred compensation obligation. When these kinds of insurance programs began to exploit the tax arbitrage opportunities inherent in COLI, through the use of large policy loans, and the interest deductions which they generated, Congress reined them in by eliminating the interest deduction on policies in excess of $50,000. See I.R.C. § 264(a) (4) (B) (1994).

Insurance industry entrepreneurs soon found a way to circumvent this limitation by designing programs in which a corporation could purchase insurance policies on a large number or even all of its employees. These "broad-based" COLI plans could include any employee regardless of the length of employment or the value of the employee's services to the corporation. Thus, by increasing the number of policies issued, the opportunity for policy loans and loan interest deductions could be multiplied while remaining within the $50,000 per policy loan limit. These efforts to continue the exploitation of the tax arbitrage opportunities of COLI caused Congress to consider additional legislation to limit or eliminate the deductibility of interest on policy loans. It was in this context that AEP made its decision to purchase broad based COLI covering over 20,000 of its employees.

Related Litigation

This is the third in a series of cases in which the IRS has challenged the deductibility of interest on policy loans in highly-leveraged broad-based COLI programs. See Winn-Dixie, Inc. v. Comm'r, 113 T.C. 254, 1999 WL 907566 (1999), and In re C.M. Holdings, Inc., 254 B.R. 578 (D.Del. 2000). The C.M. Holdings case is particularly relevant because it involved the same *769 COLI plan purchased by AEP which was offered by the same insurer and sold by the same brokers.

In C.M. Holdings, Camelot Music, Inc., a wholly-owned subsidiary of C.M. Holdings, Inc., purchased life insurance policies on the lives of 1,430 of its employees. The policies were in all relevant respects identical to those purchased by AEP. The arguments advanced by the parties were essentially the same as those advanced in the instant case. The IRS was represented by the same team of litigation counsel and the parties called most of the same witnesses who testified here, including the expert witnesses. The trial in C.M. Holdings commenced in late March, 2000, and lasted for six weeks. On October 16, 2000, Senior U.S. District Judge Murray M. Schwartz, issued a 143-page opinion, which marshals all of the facts and competing theories which these two highly complex cases share. The trial of the instant case commenced on October 30, 2000, and it also lasted six weeks. After thoroughly studying the court's opinion in C.M. Holdings, it is apparent that this court has been presented with essentially the same evidence and the same legal arguments. This court has independently reached many of the same conclusions as the court in C.M. Holdings. In the interests of brevity and judicial economy, this court will make liberal use of the opinion in C.M. Holdings. AEP's purchase of broad-based COLI

Historically, AEP has provided medical benefits to its retired employees. Prior to 1993, Generally Accepted Accounting Principles permitted corporations to account for postretirement medical benefits on a "pay-as-you-go" basis. Thus, AEP reported the costs of medical benefits for retired employees as an expense at the time the benefits were paid. In the mid-1980s, the Financial Accounting Standards Board ("FASB") announced that it was considering a change that would accelerate the recognition of postretirement medical benefits expense. Under this new method, postretirement medical benefits expense would be accrued and reflected in financial statements during the employees' working years. The FASB indicated that it would also require the creation of a "transition liability" to recognize that no prior expense accruals had been made for postretirement medical benefits. This change in accounting rules, known as FAS 106, was issued in 1990 and became effective in 1993.

Preliminary analysis performed by AEP staff and consultants in 1987 and 1988 indicated that FAS 106 would increase AEP's annual expense for postretirement medical benefits from approximately $10 million on the pay-as-you-go basis, to $60-$80 million on the new accrual basis. This increase in expense would significantly reduce AEP's reported net earnings.

In early 1989, AEP formed an internal task force on postretirement benefits and charged it with the responsibility to analyze the effects of proposed FAS 106 and to develop solutions to the problems it would create. AEP also requested its actuarial consulting firm, Towers, Perrin, Foster & Crosby ("Towers, Perrin") to evaluate the impact of the anticipated changes.

As a public utility, AEP recovers the costs of providing services through rates charged to its customers, subject to the approval of federal and state regulators. It was unclear to AEP whether it would be permitted to pass on to consumers the incremental increase in benefits expense caused by FAS 106. AEP considered eliminating or severely reducing the medical benefits available to retirees but rejected this approach. AEP decided to attempt to recover its FAS 106 expenses through rate increases, and it sought ways to mitigate or offset its FAS 106 expense in order *770 to enhance the possibility of obtaining the approval of regulators for rate increases.

AEP's consideration of purchasing broad-based COLI appears to have originated with AEP's accountants, DeLoitte Haskins + Sells ("DeLoitte"). In October of 1989, Robert S. Gorab of DeLoitte mentioned the use of COLI for the purpose of offsetting anticipated FAS 106 expenses in a telephone conversation with Gerald P. Maloney, AEP's chief financial officer. Gorab followed up by sending Maloney materials he had received from the insurance brokerage firm of Alexander and Alexander. These materials described a COLI plan offered by New York Life Insurance Company and included a paper entitled "METHODS TO PRE-FUND POST-EMPLOYMENT MEDICAL BENEFITS."[3] J1273. Maloney had prior experience with COLI in connection with the funding of senior executive's deferred compensation plans in 1982 and 1986, in which life insurance was purchased on the lives of key executives in order to fund the deferred compensation. Maloney assembled a group of AEP executives and employees to assist him in deciding whether to recommend COLI to AEP's top management, Chairman of the Board and Chief Executive Officer W.S. White, Jr. and Chief Operating Officer and President Richard S. Disbrow.

Maloney sent the Alexander and Alexander materials on to L.V. Assante, AEP's assistant treasurer, for his comments. On November 27, 1989, Assante, with the assistance of an AEP accountant, Hugh McCoy, provided Maloney with a memorandum containing an analysis of the New York Life COLI plan, which indicated that the plan would produce income that would offset a significant portion of the accruals for postretirement benefits expense over a thirty-year period. The memo concluded with the following statement:

This appears to be a good funding vehicle over the long run but in the early years (and perhaps in total for AEP) it will not completely offset the increase in expense related to postretirement medical benefits that will result if the FASB's Exposure Draft becomes a final Statement.


In January, 1990, Maloney received another presentation on the use of COLI to fund postretirement medical benefits from James Campisi, who represented a joint venture between his firm, The Newport Group ("Newport"), and the insurance brokerage firm of Johnson & Higgins.

Campisi provided Maloney with information regarding a COLI plan offered by Mutual Benefit Life Insurance Company ("MBL"), which included schedules showing the projected performance of the MEL COLI plan based on 20,000 employees and 7 annual premiums of $16,667 per employee. The COLI plan proposed by Campisi contemplated substantial policy loans during the first three years, with the proceeds applied to the payment of premiums. The premiums for the next four years would be largely offset by dividends and withdrawals of cash value and no premiums would be due after the seventh year. The projected performance of the plan produced positive cash flow and earnings every year for the life of the plan.

Maloney and the members of his group met with Campisi on January 26, 1990. Campisi explained the benefits of the MBL COLI program. Issues discussed at this meeting included: (1) whether AEP had an insurable interest in the employees *771 sought to be covered; and (2) how the laws of the various states in which AEP had employees might affect that issue. Campisi pointed out that taxable income was necessary to realize the tax benefits contemplated by the program. Campisi offered to provide Maloney with a so-called "sweetheart letter" which guaranteed that the plan could be "unwound" any time before December 31, 1990, at no cost to AEP, in the event Congress and the President should approve tax legislation which would adversely affect the plan.

On January 31, 1990, Campisi sent Maloney a packet of materials which included policy applications and prepayment agreements. Maloney asked McCoy to review Campisi's proposal and McCoy provided him with a memorandum dated February 6, 1990, with attached schedules. See J752. McCoy pointed out that after the first three years of the program, policy loans would total nearly $1 billion and would generate interest expenses of approximately $130 million per year.

On February 7, 1990, Maloney forwarded Campisi's materials on to Peter J. DeMaria, AEP's treasurer and chief accounting officer, and M.R. Luis, AEP's senior tax counsel. In this memorandum, Maloney commented on various aspects of the plan including the following:

8. I noted that the "tax benefit" amounts that are projected are very substantial and could possibly be challenged in some future audit by the IRS. Mr. Campisi stated that this coverage must qualify as "life insurance" under relatively objective and quantifiable standards of Section 7702 of the Internal Revenue Code; while it is possible that there could be such a challenge by IRS, there could be a reasonable defense because of compliance with those standards.

J753. Maloney concluded his memorandum by stating that he had asked Mr. Luis to look into the questions of (a) insurable interest in the lives of retirees, and (b) exposure to a challenge by the IRS of the deductibility of interest charges on policy loans.

On February 8, 1990, Luis provided Maloney with separate memoranda concerning these issues. In his one-page memorandum on the issue of deductibility of interest on policy loans, Luis stated:

Since the pending COLI program, if adopted, will generate substantial interest expense deductions, the IRS will scrutinize the transaction. If the COLI policies meet the section 7702 definition of a "life insurance contract" and if the requirements of section 264 of the Code are met, the transaction along with its favorable tax attributes should not be challenged.

Defendant's Exh. D209. Maloney was satisfied with Luis's response.

The materials Campisi originally provided to Maloney were premised on a variable policy loan interest rate, which had a current rate of approximately thirteen percent. Maloney asked Campisi to provide him with additional schedules containing projections of performance based on lower policy loan interest rates. Campisi provided Maloney with this information in a letter dated February 7, 1990, which pointed out that the COLI plan provided a choice between 3 formulas for variable rates which, based on current rates, produced interest charges of 13.11%, 12%, and 10.91%. See J968. In each instance, the difference or "spread" between the policy loan interest rate and the crediting rate on encumbered cash value was fixed at one percent. Maloney selected the second option which produced a current policy loan interest rate of twelve percent. Maloney testified that he selected the middle rate *772 because it approximated the rate AEP was then paying on its highest quality debt securities.[4] He said he rejected the lowest rate because he wanted to maximize taxdeferred inside buildup and that he declined the highest rate because he thought it was too high and less defensible if questioned by the IRS. Maloney could have chosen an eight percent fixed rate but he gave that option no consideration because it would have been detrimental to the performance of the plan.

Maloney convened another meeting of his group on February 12, 1990. Maloney prepared a checklist for this meeting, which included an item entitled "IRS Disallowance of Interest Deduction." This item included a reference to obtaining an opinion by AEP's outside counsel, Simpson, Thatcher & Bartlett, followed by a question mark. Maloney testified that he ultimately decided to rely on Luis's one-page memo instead of seeking an opinion from outside counsel, believing that he would receive the same opinion from outside counsel. At the conclusion of the February 12th meeting, Maloney decided to recommend that AEP proceed with the COLI program and instructed Campisi to prepare to bind coverage. Maloney informed AEP Chairman and Chief Executive Officer W.S. White, Jr. of his recommendation and inquired whether board approval would be necessary. White accepted Maloney's recommendation and advised that board approval would not be necessary.

White signed applications and prepayment agreements for 8 AEP operating companies on February 16, 1990, and on the same date checks were written by the AEP operating companies in an aggregate amount in excess of $4 million. On February 21, 1990, Maloney and Campisi appeared before the AEP Board to explain the COLI plan the company had just purchased.

Maloney acknowledged that there was some urgency involved in consummating the purchase of the COLI plan because of the possibility that Congress might enact legislation restricting or eliminating the tax advantages of COLI. According to Maloney, however, everyone believed that existing plans would be "grandfathered" if that should occur.

Shortly after executing the policy applications, AEP, at the suggestion of Campisi, considered assigning the policies to a "grantor trust"[5] sited in Georgia in order to realize substantial savings on premium taxes which would generate significant additional cash flow during the first five years of the COLI plan. Luis was given responsibility for completing this project. An appropriate trust agreement was prepared and executed on March 21, 1590. Campisi had advised that this should be done no later than March 22, 1990 because of potential adverse tax legislation and fear that a later date might jeopardize "grandfathering" the plan.

On March 20, 1990, the chief executive officers of each of the AEP subsidiaries *773 executed a revocable trust agreement assigning the MEL COLI policies on the subsidiaries' employees to the Citizens and Southern Trust Company of Georgia, Trustee. The trustee accepted the assignments and executed the trust agreement on March 21, 1990.

The COLI policies were delivered on approximately March 23, 1990 and the balance of the cash portion of the first year's premiums was transmitted by the AEP subsidiaries to the trustee which, in turn, paid MBL. The policies were made effective on February 16, 1990, the date of the prepayment agreements. The policy loans, which were used to finance a substantial part of the first year's premium, were backdated to February 16, 1990, so that AEP could take advantage of the interest deduction for the entire year.

AEP's voluntary employees beneficiary association

The transition liability that FAS 106 required companies to report as a liability on their balance sheet can be reduced by the fair value of assets segregated and restricted to be used for post-retirement benefits. Prior to the adoption of FAS 106, AEP had funded health plans for active employees through contributions to a Voluntary Employees Beneficiary Association ("VEBA"). While it was considering the COLI plan, AEP also considered creating a VEBA to fund the FAS 106 transition liability, and it ultimately did so. This VEBA, actually a partition of AEP's existing VEBA, was funded through the purchase of trust-owned life insurance ("TOLI") written by the Prudential Insurance Company of Newark, New Jersey. At its inception, this plan insured 845 employees under a group contract which was owned by the VEBA. Contributions to the VEBA are tax deductible and earnings on plan assets within the VEBA are sheltered from income tax. The VEBA, however, is not sufficiently funded to cover the full extent of AEP's FAS 106 postretirement medical benefit liability. When AEP sought rate increases from the state regulatory agencies to cover its FAS 106 liability, it represented that it had mitigated its FAS 106 expense by funding a VEBA and by purchasing broad-based COLI. AEP represented that it would contribute the earnings realized by its COLI program to the VEBA. All of the regulatory agencies approved this approach and approved rate increases commensurate with the FAS 106 expense reductions which AEP projected would result from the VEBA and its COLI plan.

The money contributed to the VEBA, which was invested in the TOLI plan, was used to purchase variable term life insurance policies, wherein the cash values would reflect the performance of a bond fund and the Standard and Poor's 500 Stock Index. Mr. Maloney testified that the TOLI plan has performed very well, earning close to a twenty percent per year increase in the S & P fund component and approximately 7% per year in the bond fund component. Unlike the COLI plan, the TOLI plan does not rely on the tax deductibility of interest on policy loans.

Development Of MBL COLI Plan

The idea behind the MBL COLI Plan was conceived in 1985 by Henry F. McCamish ("McCamish"), a life insurance entrepreneur. McCamish retained Milliman and Robertson ("M & R"), a nationally prominent actuarial consulting firm, to develop a new type of COLI product. McCamish asked M & R to design a COLI policy for large corporations that had policy value at the end of the first policy year that exceeded the first year's premium, so that the policy value would support a policy loan large enough to cover most of the first-year premium. He also stipulated that the policy should generate positive earnings and cash flow in the first plan year. Stephen Eisenberg ("Eisenberg"), *774 Managing Consulting Actuary of M & R's life insurance practice, was the principal architect of this new form of COLI. He was assisted by Timothy Millwood, another M & R actuary. In designing the policy, Eisenberg had to ensure that it complied with all pertinent existing provisions of the Internal Revenue Code, including I.R.C. §§ 264 and 7702. Thus, Eisenberg had to design a plan in which the policyholder would not take policy loans in four out of the first seven policy years.

McCamish, who was a personal friend of the president of MBL, procured MBL's agreement to underwrite the new COLI product being designed by Eisenberg. James Van Etten, an actuary employed by MBL, collaborated with Eisenberg while he was designing the policy. McCamish formed two corporations that provided administrative and consultive services for the new MBL COLI policies, Integrated Administration Services, Inc. ("IAS") and IAS Development Corporation ("IDC"). As compensation for these services, MBL agreed to pay McCamish and his companies a policy administration fee for every COLI policy sold, as well as a percentage of the annual premiums generated by the policies.

By early 1986, M & R, MBL and McCamish had developed MBL policy form FA85, which was designated as "COLI I". Thereafter, policy form FA85 was modified at various times in response to changes in the tax laws and competitive pressures in the marketplace. The modifications were made by adding a series of endorsements or amendments to the form. Each new version of the policy was sequentially numbered in Roman numerals. The COLI III policy was created in late 1986 in response to the amendment of I.R.C. § 264, which limited the deductibility of policy loan interest to $50,000 per insured. The COLI III policy initiated the use of dividends to offset payments of premiums in four out of the first seven policy years. The COLI V policy modified COLI III by making premiums payable for only a defined number of years. It contemplated that premiums would stop once the policyholder reached the $50,000 loan limit of I.R.C. § 264(a) (4) (B) and the requirements of the four out of seven safe harbor of § 264(c) (1) had been satisfied. Thus, the premiums stopped after seven years, and thereafter, the policyholder would hold the policies under an extended term nonforfeiture option. Finally, the COLI VIII policy modified COLI V by reducing the amount of premium per $1,000 of death benefits. This was done to ensure compliance with I.R.C. § 7702(a), which placed a cap on the premium per thousand in life insurance contracts.

Features of AEP's COLI VIII policies and plan

The COLI VIII policies purchased by AEP were issued by MBL on policy form FA85 with endorsements EFA85-3, EFA85-4, EFA85-6, and EFA85-7. The policies are denominated on their face "Increasing Death Benefit Whole Life" with "premiums fixed and payable during lifetime of insured or until end of premium period" and "Eligible for Dividends." While the amount of the death benefit would vary depending upon the age of the insured employee, the remaining terms of each of the policies were identical. Each had a fixed annual premium of $16,667.

The policies provided for three interest rates: (1) the policy loan interest rate; (2) the loaned crediting rate, called the "Current Credited Loan Interest Rate," which was the rate at which the company would increase the amount of the cash value of the policy used to secure policy loans; and (3) the unloaned crediting rate, called the "Current Unloaned Interest Rate," which was the rate at which the company would increase the amount of the cash value not used to secure policy loans.

*775 The loan interest rate and the loaned crediting rate were set as follows. The policies permitted AEP to annually elect between a fixed or variable loan interest rate. The fixed loan interest rate was 8% per annum if charged in arrears or 7.4%, if charged in advance. If AEP selected the fixed loan interest rate, the loaned crediting rate would be the greater of a rate declared by MBL or 4%. If AEP elected the variable loan interest rate, the loan interest rate would be the greater of: (1) the Moody's Corporate Band Yield Average Monthly Average Corporate[6] ("Moody's Corporate Average") or (2) the loaned crediting rate plus 1%. Under the MBL COLI policy form, the loaned crediting rate would be equal to Moody's Corporate Bond Yield Average Monthly Corporate Baa.[7] ("Moody's Baa"). Moody's Baa has always exceeded Moody's Corporate Average by 0.60% to 1.00%. Accordingly, the second prong of the variable loan interest rate provision would always govern and the loan interest rate would always be equal to the loaned crediting rate plus one percent.

Under endorsement EFA 85-3, the loaned crediting rate was calculated according to the following mathematical formula:

             Baa   +   T%
             100%  -   Baa

This formula produces a rate higher than Moody's Baa even when T = 0. This rate (T = 0) will be referred to as "Moody's Baa enhanced." While AEP was considering the MBL COLI plan, it was given a choice of variable loan interest rates which were determined by adding 1% to 3 possible loaned crediting rates: Moody's Baa (based on the standard policy form FA 85); Moody's Baa enhanced (based on endorsement EFA 85-3 with T equal to 0); and Moody's Baa enhanced plus 1% (based on endorsement EFA 85-3 with T equal to 1). AEP chose Moody's Baa enhanced, which resulted in a loaned crediting rate of eleven percent and a policy loan interest rate of twelve percent. As a result, the policies were issued with endorsement EFA 85-3 and the factor "T" was set at 0.

Before it purchased the COLI policies, AEP was provided with sets of financial projections called "issue illustrations" which demonstrated how the COLI plan would perform in terms of cash flow and earnings. An issue illustration entitled "Scenario II" demonstrated the projected performance of the plan based on the policy loan interest rate AEP selected. See J968. This issue illustration demonstrated how the COLI policies were intended to be operated and the economic benefits they were expected to produce during the first 20 years of the plan, based on a corporate income tax rate of 37% and 20,000 insured employees. The following table (J968) shows the cash flow analysis presented in Scenario II for the first ten years of the plan.

          SCENARIO II
                                          (OOO's OMITTED)
_____________ USE OF FUNDS _____________                     _____________ SOURCE OF FUNDS ____________
                                                                         CASH        NET TAX
                      AT     GROSS     ANNUAL                            VALUE         FREE      ANNUAL
         CORP.     ADMIN     LOAN      POLICY     TAX       CASH      WITHDRAWALS   INSURANCE     CASH
1990     333,340     252          0   312,360    13,864           0             0      10,860     3,492
1991     332,900       0     37,440   334,360    28,686           0             0      11,220     3,926
1992     332,440       0     77,440   351,900    44,248      12,680             0      11,580    10,528
1993     331,940       0    119,500         0    44,182     295,860       115,400      12,380    16,382
1994     331,400       0    119,300         0    44,095     293,720       114,420      13,200    14,735
       _________   _____   ________  ________   _______   _________      ________    ________  ________
       1,662,020     252    353,680             175,075     602,260       229,820      59,240    49,063
1995     330,800       0    119,080         0    44,008     291,540       113,360      14,080    13,108
1996     330,160       0    118,860         0    43,921     290,980       112,240      15,020    13,141
1997           0     252    118,600         0    43,834           0        92,160      18,160    35,302
1998           0     252    118,340         0    43,725           0        87,600      22,660    35,393
1999           0     252    118,040         0    43,616           0        81,960      28,220    35,504
       _________   _____   ________  ________   _______   _________      ________    ________  ________
       2,322,980   1,008    946,600             394,181   1,184,780       717,140     157,380   181,513

The MBL COLI VIII policies were designed to have fixed annual premiums payable on the first day of each policy year. AEP selected the highest possible annual premium of $16,667 per policy, which permitted the accumulation of $50,000 in cash value in the shortest possible time, i.e. three years. The first year's total premium was calculated at $333,340,000. Premiums for the next 6 years were projected to be in the range of $331 million to $333 million. During the first through third years, the plan called for AEP to pay approximately ninety percent of the annual premiums through policy loans in simultaneous netting transactions in which the loans were offset against the premiums. Instead of paying $330 million in cash for the first year's premium, AEP would actually pay about $21 million (plus $2.5 million in first-year administrative fees.) This cash payment of approximately $23.5 million would be offset by the tax benefit of deducting the policy loan interest expense ($13.8 million) and by the receipt of taxfree insurance proceeds in the amount of $10.8 million. The net result would be positive cash flow of approximately $3.5 million the first year. Positive cash flow increased to approximately $3.9 million the second year, $10.5 million the third year and then quickly escalated to an amount in excess of $35 million per year by the eighth policy year.

After the third policy year, AEP could no longer finance the premiums with policy loans since this would violate the $50, 000 per policy loan limit of I.R.C. § 264 and forfeit the tax deductibility of the policy loan interest. The COLI policies were designed so that in years four through seven, most of the annual premiums and accrued loan interest would be paid by dividends and partial withdrawals in simultaneous netting transactions, which would occur on the first day of each of those policy years. In these annual transactions, approximately ninety-five percent of the annual premium was considered taken by MBL as an expense charge. After setting aside amounts sufficient to cover MBL's actual expense charges, the remainder was returned to AEP as a "loading dividend" offsetting ninety-five percent of the premiums due.

No additional premiums were due after year seven and the loading dividends ceased. The policy loans remained in effect, *777 however, generating interest expense in excess of $100 million annually, which was substantially offset by cash value withdrawals. Positive cash flows in the range of $35 million to $39 million were projected through year 20.

The AEP COLI VIII policies were projected to have a net equity equal to zero at the end of each and every policy year. The net equity, also known as cash surrender value, is equal to the total policy value, less policy loans and accrued interest. When a policy has a net equity equal to zero, all of the policy value is encumbered and it cannot support any additional policy loans. The designers of the MBL COLI plan used a sophisticated computer program to perform the calculations necessary to achieve this result.

AEP's MBL COLI program was designed to be "mortality neutral," to wit, the program's design was that the cumulative COI charges that AEP paid would equal the cumulative death benefits that it would receive, minus a profit margin to MBL of 20% of COI in the first year, 10% the second year and 2% each year thereafter. The concept of mortality neutrality was explained and described in M & R documents as follows:

Since the sale is tax driven, mortality should play a neutral role in performance. That is, clients usually do not expect to make or lose money on their mortality experience. They expect to realize the illustrated cash flows each year.


Policyholders typically do not buy these products for mortality gains. They only expect to be treated fairly with respect to mortality.
They expect COI rates will increase if experience so indicates. Likewise, they expect that favorable experience will result in increased policyholder dividends. In other words, COLI policyholders enter these contracts as a partnership with the insurance carriers with neither party taking advantage of the other party.

D580. MBL's commitment to mortality neutrality was reflected in a mortality dividend declared in 1991 for all plans issued through 1989 and a pledge to make annual mortality dividends thereafter, if mortality experience continued to prove more favorable to the company than expected. MBL's successor, Hartford, carried out this commitment to mortality neutrality by instituting experience rated dividends for the ten largest COLI plans, including AEP. Campisi advised AEP that "this mortality mechanism will result in a much closer match between expected cash flow from projected death benefits and claims that are actually paid. This will eliminate any volatility or variation in the cash flow that we are expecting in each year of the plan." D387.

The essential features of the AEP MBL COLI VIII plan can be summarized as follows. First, the COLI plan had a very high policy value on the first day of the first year of the policy. This was made possible by the highest possible premium per policy and the broad base of employees insured. The high first year policy values were also made possible by the agents foregoing first-year commissions. Second, the high policy values were used to secure maximum policy loans, which paid a substantial amount of the premiums due in the first three policy years and also produced substantial interest deductions that, when combined with projected tax free death benefits, fueled positive cash flows. Third, computer technology enabled maximum leveraging of the high policy values through programs designed to achieve zero net equity at the end of each policy year. Fourth, AEP had the right to select among crediting rates for the inside buildup on the loaned portion (i.e., virtually all) of the policy value, which automatically *778 determined the policy loan interest rate because of the one percent spread between the loaned crediting rate and the policy loan rate. Fifth, the one percent spread essentially fixed AEP's cost of borrowing with the counter-intuitive result that the higher the loan interest rate paid by AEP, the greater the cash flow to it as a result of higher interest deductions. Sixth, the COLI VIII plan had an extremely high expense load component in policy years four through seven, which was used to create first-day dividends, which were used to pay most of the premium charges for those years. Finally, the actual cost of providing the death benefits was carefully calibrated to equal the amount of death benefits which AEP would receive over the life of the plan so that, except for MBL's modest profit on COLI profit, there would be no gain or loss based on mortality experience.

Thus, through a preplanned, highly structured and calibrated combination of features, AEP's broad-based leveraged COLI plan was designed to produce positive cash flows in each and every plan year using the tax benefits of the deductibility of policy loan interest, deferral of taxation of inside buildup, and nontaxation of death benefits to transform paper losses into positive earnings and to generate substantial positive cash flows totaling over half a billion dollars at the end of twenty years.

The sham transaction doctrine

The government contends that the loading dividends, policy loans, and partial withdrawals that are all components of AEP's MBL COLI plan are factual shams. The sham transaction doctrine originated with the Supreme Court decision of Gregory v. Helvering, 293 U.S. 465, 55 S. Ct. 266, 79 L. Ed. 596 (1935). In Gregory, the Court affirmed the Commissioner's denial of deductions claimed by taxpayers for losses and expenses incurred in a corporate reorganization. Although the taxpayers had followed each step required by the Internal Revenue Code for the reorganization, the Court nonetheless held these losses nondeductible, finding that the transaction was a "mere device" for the "consummation of a preconceived plan" and not a reorganization within the intent of the code as it then existed. Id. at 469, 55 S. Ct. 266. Because the transaction lacked economic substance, it was not "the thing which the statute intended." Id.

The sham transaction doctrine requires the court to examine a challenged transaction as a whole and each element thereof to determine if the substance of the transaction is consistent with its form. See ACM P'ship v. Comm'r, 157 F.3d 231, 247 (3rd Cir. 1998), cert. denied, 526 U.S. 1017, 119 S. Ct. 1251, 143 L. Ed. 2d 348 (1999). If the form of a transaction complies with the Code's requirements for deductibility, but the transaction nevertheless lacks factual or economic substance, then expenses or losses incurred in connection with the transaction are not deductible. See Knetsch v. United States, 364 U.S. 361, 369, 81 S. Ct. 132, 5 L. Ed. 2d 128 (1960); United States v. Wexler, 31 F.3d 117, 122 (3rd Cir.1994), cert. denied, 513 U.S. 1190, 115 S. Ct. 1251, 131 L. Ed. 2d 133 (1995).

The burden of proof is on the taxpayer to show that the form of the transaction reflects its substance. National Starch and Chem. Corp. v. Comm'r, 918 F.2d 426, 429 (3rd Cir.1990) ("burden is on the taxpayer to show that the expenses are deductible") aff'd sub nom INDOPCO, Inc. v. Comm'r, 503 U.S. 79, 112 S. Ct. 1039, 117 L. Ed. 2d 226 (1992); Goldberg v. United States, 789 F.2d 1341, 1343 (9th Cir.1986). While a taxpayer can legitimately structure a transaction to minimize tax liability under the Code, the transaction must nevertheless have factual and economic substance. *779 See Gregory, 293 U.S. at 469, 55 S. Ct. 266.

AEP contends that the sham transaction doctrine includes a threshold requirement which requires the court to first determine "whether the thing done was the thing intended by the statute." Plaintiff's Post-Trial Memorandum at p. 32. According to AEP, if the court determines that the thing done was indeed the thing intended, the court need not and should not proceed to examine the issue of whether the transaction has economic substance. AEP argues that Congress has, over the years, carefully tailored the tax laws relating to the favorable treatment of life insurance, and since AEP's COLI plan satisfies all of the statutory requirements for the tax favored treatment of inside buildup, death benefits, and policy loan deductions, the plan satisfies the statutory intendment test and the judicial inquiry is at an end. AEP's statutory intendment test is taken from the following language in Gregory:

The legal right of a taxpayer to decrease the amount of what otherwise would be his taxes, or altogether avoid them, by means which the law permits, cannot be doubted. But the question for determination is whether what was done, apart from the tax motive, was the thing which the statute intended.

Gregory, 293 U.S. at 469, 55 S. Ct. 266 (citations omitted). AEP reads too much into this statement. Certainly, the Court did not mean that satisfying all of the statutory requirements precludes a court from examining either the factual reality or the economic substance of the questioned transaction. Indeed, in Gregory, the Supreme Court went on to note: "[n]o doubt, a new and valid corporation was created. But that corporation was nothing more than a contrivance," and "[t]he whole undertaking, though conducted according to the terms [of the statute] was in fact an elaborate and devious form of conveyance masquerading as a corporate reorganization, and nothing else." Id. at 469-70, 55 S. Ct. 266. AEP's statutory intendment test would effectively swallow up the economic substance test and vitiate the sham transaction doctrine. A transaction which has no factual reality or economic substance can never be the thing the statute intended even though the form of the transaction satisfies all of the requirements of the statute.

AEP argues that the Supreme Court's opinion in Hanover Bank v. Comm'r, 369 U.S. 672, 82 S. Ct. 1080, 8 L. Ed. 2d 187 (1962) and the Sixth Circuit's opinion in Humphreys v. Comm'r, 301 F.2d 33 (6th Cir.1962) are "clear examples of the application of the statutory intendment standard." Plaintiff's Post-Trial Memorandum at p. 35. However, neither of these cases support AEP's argument. In Hanover Bank, the question before the Court was one of pure statutory construction. The Court specifically noted that the "Government does not contend that the transactions entered into by the petitioners were a sham without any business purpose except to gain a tax advantage." Hanover Bank, 369 U.S. at 681, 82 S. Ct. 1080. In Humphreys, the Sixth Circuit specifically found that "[i]t cannot be controverted that the purchase of the bonds and the gifts to charity were real genuine transactions though motivated by tax considerations." Humphreys, 301 F.2d at 33-34. Neither case involved the application of the sham transaction doctrine.

The taxpayers in C.M. Holdings, and Winn-Dixie raised arguments similar to AEP's statutory intendment argument and they were rejected. See C.M. Holdings, 254 B.R. at 636 and Winn-Dixie, 113 T.C. at 290, 294. This court agrees.

Courts have recognized two basic types of sham transactions: (1) "shams in fact," in which the reported transactions never *780 occurred; and (2) "shams in substance" in which the transactions "actually occurred but which lack the substance their form represents." Kirchman v. Comm'r, 862 F.2d 1486, 1492 (11th Cir.1989); see also ACM P'ship, 157 F.3d at 247 n. 30; Lerman v. Comm'r, 939 F.2d 44, 49 n. 6 (3d Cir.1991). The government claims that the loading dividends, policy loans, and partial withdrawals under AEP's MBL COLI plan are factual shams and that the AEP MBL COLI plan as a whole was a sham in substance because it was devoid of economic substance apart from the policy loan interest deductions.

Sham in fact doctrine as applied to components of the AEP COLI PLAN Policy loans

The government contends that the policy loans were factual shams. Here, in addition to the decisional law relating to the sham transaction doctrine, the government relies on a separate body of case law defining interest and indebtedness for the purpose of determining the deductibility of interest under I.R.C. § 163. This line of cases stands for the proposition that interest and indebtedness must be real in order to support a deduction. See Knetsch, 364 U.S. at 369, 81 S. Ct. 132; Wexler, 31 F.3d at 122; Goldberg, 789 F.2d at 1343; Bridges v. Comm'r, 325 F.2d 180, 181-82 (4th Cir. 1963).

Section 163 allows a deduction for "interest paid or accrued within the taxable year on indebtedness." I.R.C. § 163. "Interest" represents compensation for the use or forbearance of money. Old Colony R. Co. v. Comm'r, 284 U.S. 552, 561-62, 52 S. Ct. 211, 76 L. Ed. 484 (1932). "Indebtedness" under I.R.C. § 163 is defined as "an unconditional and legally enforceable obligation for the payment of money." Autenreith v. Comm'r, 115 F.2d 856, 858 (3rd Cir.1940).

According to the government, AEP's COLI policy loans and their associated interest deductions lacked both factual and economic reality because: 1) the COLI loan interest rates were not determined by the market in an arm's-length transaction but were arbitrarily chosen in a collusive arrangement for the purpose of maximizing interest deductions; 2) they violated industry standards governing maximum permissible loan interest rates; and 3) the rates were excessive and unjustifiable from a financial perspective. Most of the arguments advanced by the government for the proposition that the policy loans were factual shams are more appropriately addressed to the issue of whether the policy loans were economic shams.

The court is satisfied by a preponderance of the evidence that the policy loans were not factual shams. The COLI policy loans were considered real debts of AEP, which had to be repaid when the insured employee died or if the policy lapsed. Whenever an insured employee died, the death benefit was reduced by the outstanding policy loan and unpaid interest. MBL and its successor, Hartford Life Insurance Company ("Hartford") carried the policy loans as assets on their books and recorded them separately from the receipt of premiums. In the aftermath of HIPA, AEP repaid the loans on all policies in excess of 20,000 employees with substantial amounts of cash.

All insurance policy loans seem rather unreal when compared to ordinary commercial loans. It is the longstanding custom and practice in the insurance industry that policy loans are deemed as made from the general funds of the insurance company, with the policy value serving only as collateral. They are not considered withdrawals of cash from the policy value itself. As the court observed in C.M. Holdings:

*781 One can think of the life insurance company acting somewhat like a bank, lending money to the policyholder from its general funds, with the policy value pledged as collateral.

C.M. Holdings, 254 B.R. at 602. The fact that the policy loans were made on the first day of each policy year does not deprive them of reality. There is precedent in the industry for first-day, first-year policy loans. It is not unusual for policy loans to be applied to offset premiums in a simultaneous netting transaction. First-day, first-year policy loans that were credited against premiums were not regarded as factual shams in Campbell v. Cen-Tex, Inc., 377 F.2d 688 (5th Cir.1967) and Woodson-Tenent Labs., Inc. v. United States, 454 F.2d 637 (6th Cir.1972).

The risk characteristics of the loans do not affect their factual reality. Indeed, all life insurance policy loans are devoid of most of the major risk factors which are present in commercial loans. It is the custom of the industry to set interest rates on policy loans without reference to the risk factors that attend ordinary commercial loans.

The government's argument that the policy loan interest rate was set at an artificially high rate in a collusive non-arms-length transaction does not implicate the reality of the transaction. Interest rates on insurance policy loans are always determined in private transactions between the insurance company and the policyholder, and are customarily determined by the terms of the insurance contract issued by the insurance company and accepted by the policyholder. There is no market rate for such loans, as there is in the case of real estate mortgages or commercial loans. There is ample precedent for variable rates based on standards such as Moody's Corporate Average. Fixed spreads between policy loan interest rates and cash value crediting rates are not unusual. The fact that the MBL COLI policies gave the policyholder a choice of rates higher than Moody's Corporate Average, by linking them to Moody's Baa (plus a choice of enhancements), does not affect the reality of the loans.

Backdating of the first-year policy loans

When AEP decided to participate in the MBL COLI plan, it entered into prepayment agreements, which were signed on February 16, 1990, wherein the AEP companies agreed to pay $200 per insured to purchase insurance from the date of the prepayment agreements to the date of policy issue. The prepayment amount was to be applied to the first premium due if the policies were issued. Under the terms of the prepayment agreements, coverage was to end on the earlier of: 1) sixty days from the date of the prepayment agreement; 2) the date the policies were issued; 3) the employer's failure to pay premiums on policy delivery; or 4) upon MBL's disapproval of the application for insurance. The prepayment agreements expressly provided that "accumulation or credit of interest on policy values and charges for policy loan interest as provided in the policy shall accrue only from the date of policy issue." J761. Thus, the prepayment agreements did not require AEP to pay any policy loan interest until MBL actually issued its policies.

The policies were issued and delivered on March 23, 1990. The policies and the policy loans were backdated to the date of the prepayment agreements, February 16, 1990, and AEP claimed interest deductions on the policy loans from that date. AEP was not obligated to pay any loan interest during the period of the prepayment agreement. No loans existed during that period. MBL did not provide AEP with any funds during that period, nor did any policies exist which could be have been used as collateral for policy loans. *782 McCamish's company, IAS, which was responsible for the administration of the MBL COLI policies, originally calculated the interest on the first year of the policy loans based upon an inception date of March 23, 1990, the date the policies were delivered. This was later changed at the urging of Campisi after AEP complained that it reduced the projected positive after-tax cash flow for the first year from $3,000,000 to a little over $1,000,000.

The illusory nature of the inception date of the policy loans and the fictitious obligation for the period from February 16, 1990, to March 23, 1990, is further revealed by the fact that the COLI policies were deemed to have been transferred to a grantor trust on March 21, 1990. Thus, by backdating the policy loans, the trust purportedly became liable for interest for a period predating its legal existence. The predating of the policy loans and the creation of an interest obligation for the period February 16th through March 23, 1990, were shams in fact.

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.

Loading dividends and premiums in policy years four through seven

The government contends that the loading dividends which offset most of the premium expense in policy years four through seven were factual shams. In C.M. Holdings, the court aptly described these transactions as follows:

In the simultaneous netting transaction involving the loading dividends, the premiums come from the policyholder, the loading charges come from the premiums, the loading dividends come from the loading charges, which dividends are then used by the policyholder to pay the premiums. Since the funds for each component is source from another component in a circular transaction, the entire transaction is simply a paper ruse.

C.M. Holdings, 254 B.R. at 608.

It is standard practice for life insurance policies to have loading charges to cover the insurance company's expenses in administering the policy. It is also standard practice to assess loading charges greater than the anticipated expenses in order to provide a reasonable margin of safety to cover unpredictable expenses, to make a permanent contribution to surplus, and to provide for the payment of dividends in the event of favorable experience. However, such charges customarily have a reasonable relationship to the insurance company's risk of incurring higher than expected expenses. The government experts provided credible expert testimony that margins of five to ten percent would be generous.

There is some degree of flexibility in industry practices with respect to the amount of loading charges, which relates to how a company wishes to present itself in the market place with respect to premiums and dividends. Higher premiums permit the payment of higher dividends and lower dividends result in lower premiums. The loading dividends in this case are nevertheless far beyond the range of standard industry practices.

The following table shows the expense loads assessed in the MEL COLI policies as a percent of gross premium and MBL's *783 actual anticipated policy expenses as a percent of gross premium for policy years one through seven:

         Contractual     MBL's Anticipated
        Expense Loads     Policy Expenses
Plan       (As % of           (As % of
Year    Gross Premium)     Gross Premium)
 1            10%               4.0%
 2            20%               8.8%
 3            30%               8.8%
 4            95%               8.8%
 5            95%               9.3%
 6            95%               9.3%
 7            95%               9.3%


In years four through seven, the contractual expense load was ninety-five percent of the gross premium, whereas MBL's actual anticipated policy expenses as a percentage of gross premium was less than ten percent. The excess was the source of the huge loading dividends, which offset nearly all of the gross premiums for those years. Thus, on the first day of each of the policy years four through seven, the entire premium and accrued interest were deemed paid by AEP. At the same time, eighty-eight percent of the premium was deemed paid back to AEP as a dividend in order to offset the premium obligation.

The New York Department of Insurance rejected MBL COLI policy endorsement EFA-4 which provided for the loading dividends. In a letter dated February 10, 1988, the New York Department wrote to M & R stating:

We need to turn attention to the expense loads ... and both the pattern and level of these charges are shocking[.] We do not see how these charges can be justified as reflecting real expenses reasonably expected to be incurred by the insurer and the use of such charges in the cost structure is misleading. The gross premiums should not be loaded up through the use of fictitious expense loads.


The loading dividends bear no relationship whatsoever to MBL's anticipated policy expenses. They exist only for the purpose of offsetting most of the premium charges ostensibly due in years four through seven. The premiums for these four years were set at an artificially high level, not because they were needed to cover the costs of operating the policy or to provide a reasonable margin for dividends, but to make sure that the policies did not violate the four out of seven safe harbor rule of I.R.C. § 264(c) (1).

Implicit in the four out of seven safe harbor rule is the requirement that the annual premiums must remain level for the first seven years of the policy. See C.M. Holdings, 254 B.R. at 645, 646. AEP acknowledges that this is the case. See Plaintiffs' Post-Trial Memorandum, p. 71, Doc. No. 124. The premiums for years four through seven had to stay at the same high level as years one through three in order to preserve the right to deduct policy loan interest expense. The premiums for years one through three were set high in order to support the maximum permissible policy loans of $50,000 per policy in the shortest time in order to generate the maximum amount of tax deductible interest expense as quickly as possible. The loading dividends were devised as a way to immediately return the unnecessary and excessive portion of the premiums in years four through seven. This so-called dividend was in reality a rebate or reduction of premium. As dividends, they were a sham in fact.

The excessive loading dividends in policy years four through seven cannot be justified under industry practices. See C.M. Holdings, 254 B.R. at 612-18. It is a standard of the industry that dividends are *784 paid from surplus. Surplus is calculated based on an insurance company's experience over the preceding year. Here, the source of the loading dividend was not surplus but the excessive loading charge built into the premiums. The dividend was credited at the beginning of the year, not the end of the year and was not based on experience in the usual sense of that term. The court cannot accept the testimony of AEP's experts who attempted to show that surplus could somehow be instantaneously created and immediately returned as a dividend in a simultaneous netting transaction. There is no credible evidence of an industry practice which would support this contention.

Dividends are by definition discretionary. Guaranteed distributions cannot qualify as dividends. The loading dividends were an integral part of the COLI plan. They were virtually guaranteed by being built into the premiums and they were represented as virtually guaranteed to prospective policyholders. See id. at 772.

State accounting regulations and industry practice require insurance companies to create a liability on their balance sheets for all dividends. Neither MBL nor Hartford set up a liability on their books for the loading dividends. During the design of the COLI VIII policy, MBL officers expressed concern about the company's failure to establish a dividend liability on the company's books and suggested that the loading dividends be relabeled as a "premium rate adjustment." Van Etten rejected this suggestion, saying it was "not viable" because "it would reopen Section 264 issues with all of our existing clients with potentially disastrous results." J46. Van Etten's comments reflect his understanding that if the loading dividends were acknowledged as premium adjustments rather than dividends, the gross premiums in years four through seven would not be level as required by I.R.C. § 264 and the policyholder's policy loan interest deductions would be disallowed.

Hartford's failure to establish a liability for the loading dividend was called into question by the Connecticut Department of Insurance during its examination of Hartford's 1993 financial statements. After examining the loading dividend, the Connecticut Department ultimately concluded that Hartford was not required to establish a year-end liability because the loading dividend was not really a policy dividend but rather a "premium refund." See D392 at p. 30; Gies Tr. 2988-89.

For all of the foregoing reasons, the court finds that the loading dividends in years four through seven were factual shams. It follows that the grossly excessive premiums in years four through seven were likewise factual shams.

First-year dividend

Part of the first-year policy loan was made possible by $26 million in dividends, which were instantly added to the first-year policy values and simultaneously borrowed as part of the $319 million in first-year policy loans.[8] These, dividends are indistinguishable from the loading dividends in years four through seven and were a factual sham for the same reasons given above.

Loading dividends as the economic equivalent of a non-guaranteed element other than a dividend

AEP argues that for federal tax purposes, dividends are broadly defined to include any distribution to a policyholder that is the economic equivalent of a dividend. AEP cites, inter alia, I.R.C. § 808(a) and (b), which provides in part:

*785 (a) Policyholder dividend defined. For purposes of this part, the term "policyholder dividend" means any dividend or similar distribution to policyholders in their capacity as such.
(b) Certain amounts included. For purposes of this part, the term "policyholder dividend" includes
(1) any amount paid or credited (including as an increase in benefits) where the amount is not fixed in the contract but depends on the experience of the company or the discretion of the management[.]

Section 808 relates to an insurance company's ability to deduct distributions to policyholders. AEP argues that this broad definition of dividends undermines the government's argument that the loading dividends are factual shams. The court disagrees.

Generally, a taxpayer cannot disavow the form of its transaction absent "proof which in an action between the parties to the agreement would be admissible to alter the construction or to show its unforceability because of mistake, undue influence, fraud or duress." Schatten v. United States, 746 F.2d 319, 321-22 (6th Cir.1984) (quoting Comm'r v. Danielson, 378 F.2d 771, 775 (3rd Cir.)) (en banc), cert. denied, 389 U.S. 858, 88 S. Ct. 94, 19 L. Ed. 2d 123 (1967); Spector v. Comm'r, 641 F.2d 376, 382 (5th Cir., 1981), cert. denied, 454 U.S. 868, 102 S. Ct. 334, 70 L. Ed. 2d 171 (1981); see also North American Rayon Corp. v. Comm'r, 12 F.3d 583, 586 (6th Cir.1993).

MBL chose to designate the return of eighty-eight percent of the premiums in years four through seven as dividends instead of partial withdrawals or nonguaranteed elements in order to take advantage of the reduction in premium tax which occurs when premiums are paid with policy dividends. Having accepted the financial benefit of that characterization, AEP cannot now argue that what was labeled as a dividend was really something else. In the final analysis, however, it makes no difference how the return of premium is labeled, it remains a sham because it was simply a paper ruse, engaged in to make it appear as though level premiums were being paid during the first seven years of the policy when, in fact, they were not.

Partial withdrawals

The MBL COLI VIII plan anticipated that partial withdrawals of cash value would be taken annually, commencing with the fourth policy year to be used to offset loan interest expense. The government initially claimed that these partial withdrawals were factual shams but did not include that claim in its post-trial brief. Be that as it may, the court concludes that these withdrawals were real. See C.M. Holdings, 254 B.R. at 618-19.

Sham in substance/economic substance doctrine as applied to the AEP COLI plan as a whole

The court will now turn to the question of whether the AEP COLI plan was a sham because it lacked economic substance.

The United States Court of Appeals for the Sixth Circuit has described the economic substance test as an inquiry into "whether the transaction has any practicable economic effect other than the creation of income tax losses." Rose v. Comm'r, 868 F.2d 851, 853 (6th Cir.1989); see also Pasternak v. Comm'r, 990 F.2d 893, 898 (6th Cir.1993); Smith v. Comm'r, 937 F.2d 1089, 1094 (6th Cir.1991); Bryant v. Comm'r, 928 F.2d 745, 748 (6th Cir. 1991). In Rose, the court said, "A taxpayer's subjective business purpose and the transaction's objective economic substance may be relevant to this inquiry." Rose, 868 F.2d at 853. Thus, in determining whether the questioned transaction has *786 economic substance, the court should look to both the taxpayer's subjective business purpose and the transaction's objective economic substance. It appears that the standard in the Sixth Circuit is consistent with the standard adopted in other circuits, including the standard enunciated by the Third Circuit in ACM Partnership. There, the court stated:

The inquiry into whether the taxpayer's transactions had sufficient economic substance to be respected for tax purposes turns on both the "objective economic substance of the transactions" and the "subjective business motivation" behind them. However, these distinct aspects of the economic sham inquiry do not constitute discrete prongs of a "rigid two-step analysis," but rather represent related factors both of which inform the analysis of whether the transaction had sufficient substance, apart from its tax consequences, to be respected for tax purposes.

157 F.3d at 247 (citations omitted). In discussing the objective aspect of the economic sham analysis, the ACM court noted that "the courts have examined `whether the transaction has any practical economic effects other than the creation of income tax losses[.]'" Id. at 248 (quoting Jacobson v. Comm'r, 915 F.2d 832, 837 (2d Cir. 1990)). The court went on to state that courts:

have refused to recognize the tax consequences of transactions that were devoid of "non-tax substance" because they "did not appreciably affect [the taxpayer's] beneficial interest except to reduce his tax."

Id. (quoting Knetsch v. United States, 364 U.S. 361, 366, 81 S. Ct. 132, 5 L. Ed. 2d 128 (1960)).

AEP argues that in determining whether its COLI plan had objective economic substance, i.e., whether it had any practicable economic effects other than the creation of income tax losses, the court should not consider the economic consequences of the transaction on a pretax basis since, unlike most other investments, life insurance has a built in tax preference in the deferral of taxation of inside buildup. It is clear from the evidence, however, that the COLI plan was marketed by MBL and its brokers and purchased by AEP with a primary focus on the benefits of the deductibility of policy loan interest. For example, in testimony before the Federal Energy Regulatory Commission, Hugh McCoy, an AEP accountant, explained the operation of the COLI plan as follows:

It's the tax deductibility, the tax advantage of the COLI program, which makes it work, which creates the net income and the cash flow.
The insurance program itself, prior to the tax effects, has a cost, a small cost, because it covers 20,000 people on a system-wide basis. But it has a cost and it's a tax benefit that turns it into a favorable item, both earnings and cash flow.

D570, pp. 478, 479.

It is the combination of tax-deferred inside buildup and the deductibility of policy loan interest that creates the opportunity for tax arbitrage, which was the real motivation for AEP's purchase of the COLI policies. The success of the plan depended not only on the interest deductions generated by policy loans but also on the deferral of taxation on inside buildup and the exemption of taxation of death benefits. The latter benefits are inherent in all whole life insurance policies, but, the benefit of tax arbitrage can be realized only if there are substantial policy loans which generate tax deductible interest expense.

AEP never considered financing its MBL COLI plan with cash but, instead, made its decision to purchase the MBL COLI VIII policies based on projections of *787 aftertax cash flow and earnings generated by the maximum amount of policy loans permitted under the law and the highest policy loan interest rate it thought would pass I.R.S. scrutiny. Therefore, it is appropriate to consider the economic substance of the COLI plan in the usual way by comparing the plan's economic effects on a pretax and posttax basis. The only recognized exception to this approach is found in cases in which courts have discerned a congressional purpose to encourage investment where an economic loss would be anticipated without tax benefits. See, e.g., Sacks v. Comm'r, 69 F.3d 982 (9th Cir.1995) (tax credits and accelerated depreciation to encourage investment in solar energy).

In the case at bar, there is no evidence of congressional intent to encourage corporate investment in life insurance. Indeed, to the contrary, the legislative history in the area of corporate owned life insurance, suggests a congressional intent to discourage tax driven insurance investments while, at the same time, preserving "the importance of being able to borrow on insurance policies ... for other than tax saving purposes." H.R.Rep. No. 749 (1963), reprinted in 1964 U.S.C.C.A.N. 1313, 1370; see also S.Rep. No. 830 (1964), reprinted in 1964 U.S.C.C.A.N. 1673, 1750; McLane v. Comm'r, 46 T.C. 140, 143, 1966 WL 1140 (1966), aff'd, 377 F.2d 557 (9th Cir.1967) ("The history of the provisions dealing with the deductibility of interest on indebtedness incurred or continued to carry life insurance reflects the constant `push' by Congress to keep abreast of the `pull' of ingenious schemes to take advantage of loopholes."); see also C.M. Holdings, 254 B.R. at 624 ("Congress appears to be in a perpetual game of catch up with the innovative geniuses of the life insurance industry, constantly trying to plug up perceived holes in the tax code regarding the deductibility of policy loan interest[.] ..").

The sales illustrations provided to AEP before it purchased the COLI program show that the financial results of the plan were economically dependent upon the tax benefits generated by the policy loan interest deductions. The economic projections based on the policy loan interest rates selected by AEP show that in each year, absent the tax savings from the policy loan interest deduction, AEP's cash flows would be negative. The total before tax cash flow at the end of 20 years, which included over $900 million in tax-free death benefits, was still a negative $268,854,000. The tax savings, which resulted from the deduction of the policy loan interest, totaled $820,573.00, producing a positive cash flow of $551,719,000. See J968. Without the policy loan interest deductions, the MBL COLI VIII plan would have been a financial disaster for AEP.

The objective element of the economic sham analysis requires the court to search the transaction to determine whether it has any practical or economic effects, other than the creation of income tax losses. This requires the court to examine the transaction to see if it appreciably affected the taxpayer's beneficial interest except to reduce its tax. An investment in whole life insurance offers two potential economic benefits: 1) a death benefit which exceeds the cost of insurance in the event of the premature death of the insured person and 2) the accumulation of tax-deferred inside buildup. Mortality neutrality and zero net equity eliminated both of these economic benefits in the AEP COLI plan.

Because the plan was designed to be mortality neutral, it was anticipated that over the life of the plan, the death benefits would equal COI charges, minus MBL's modest profit on COI. Thus, AEP could not anticipate any financial advantage from the payment of death benefits. Neither would AEP benefit from tax-deferred inside buildup, because the plan contemplated *788 that literally every penny of inside buildup would be used to support policy loans.[9] The court concludes that the AEP COLI plan did not appreciably affect AEP's beneficial interest except to reduce its tax.

AEP points out that over the period from 1990 to 1998, it was projected to receive death benefits in the amount of $222.7 million but that its actual death benefit receipts were $183.9 million a shortfall of $38.8 million. No doubt, most of this discrepancy occurred before MBL/Hartford instituted an experience rating system to "fine tune" the matching of death benefits to insurance charges, and while the shortfall experienced by AEP is a significant sum of money, it did not appreciably diminish the projected tax-generated benefits of the program.

AEP argues that its investment in the MBL COLI VIII plan had economic substance because, unlike the plans considered in Knetsch v. United States, 364 U.S. 361, 81 S. Ct. 132, 5 L. Ed. 2d 128 (1960), and other cases,[10] the economic consequences of the COLI VIII plan were not "locked" from the outset by fixed interest rates. While it is true that the variable policy loan interest rate and the loaned crediting rate in the COLI VIII policies were subject to change, the spread between the rates was "locked" at one percent, which guaranteed an aftertax gain. The amount of the aftertax gain and positive cash flow was not expected to precisely match the projections and AEP understood this when it purchased the policies. The fact that the amount of the tax savings could, and did in fact, vary, does not imbue the transaction with economic substance.

AEP argues that the independent status of its subsidiaries and the disparate economic impact of the COLI plan on the subsidiaries compels a finding that the plan was not a sham in substance. In support of this argument, AEP cites Moline Props. v. Comm'r, 319 U.S. 436, 63 S. Ct. 1132, 87 L. Ed. 1499 (1943) and Humana Inc. v. Comm'r, 881 F.2d 247 (6th Cir.1989). In Moline, the taxpayer asked the Court to disregard its corporate existence and treat gains on the sale of its real estate as individual income to its sole shareholder. The Court refused, holding that the corporation had a tax identity distinct from its stockholder. In Humana, the court held that sums paid by a parent corporation to a wholly-owned captive insurance subsidiary on behalf of the parent's other wholly-owned subsidiaries were deductible as ordinary and necessary business expenses because there was risk shifting between the wholly-owned subsidiaries. The parent corporation argued that the separate corporate existence of its subsidiaries must be respected for the purposes of the risk-shifting analysis and the court agreed.

Here, the taxpayer, AEP, chose to treat itself and its wholly-owned subsidiaries as a single entity with respect to the tax consequences of its COLI plan by filing a consolidated U.S. Corporation Income Tax Return. Furthermore, AEP's decision to purchase the COLI VIII policies was based on its assessment of the benefits of the plan to the AEP system as an economic unit without consideration of its potential impact on the individual corporate subsidiaries. It follows that an analysis of the economic substance of the plan should focus *789 on the economic consequences to AEP, not its wholly-owned subsidiaries.

Economic consequences to non-taxpayer parties

An analysis of objective economic substance includes an assessment of the transaction's economic consequences to nontaxpayer parties involved in the transaction. See Frank Lyon Co. v. United States, 435 U.S. 561, 580, 98 S. Ct. 1291, 55 L. Ed. 2d 550 (1978). MBL/Hartford reported premiums due under the COLI contracts for purposes of state and federal premium taxes. They also reported policy loan interest as income for state and federal tax purposes. None of this reporting altered the insurance companies' net economic position. The costs of state and federal premium taxes were included in the actual expenses of the COLI VIII policies and were reimbursed by AEP. The policy loan interest income reported by the insurance companies was offset by the amounts they contributed to inside buildup and, as a result, they paid income tax only on the one-percent spread between those rates. The economic consequences of the transaction to the insurance companies does not give the MBL COLI plan objective economic substance.

The policy loans and interest charges

As noted above, loans made by an insurance company to its policyholder, secured by the cash value of the life insurance policy, are distinctly different from typical commercial loan transactions. The unusual attributes of insurance policy loans are supported by longstanding customs and practices in the life insurance industry. There is also precedent for some of the more unusual features of the MBL COLI loans, such as first day, first-year loans, and the netting of loan proceeds against premiums. These considerations weighed heavily in the court's determination that the MBL COLI policy loans, while unusual on their face were, nevertheless, real. However, there are other unusual features of the MBL COLI loans, which place them far from the mainstream of traditional industry practices.

The designers of the COLI policy devised a mechanism by which policyholders, like AEP, through a range of available loan interest rates, could choose to inflate their own interest costs, thus inflating their interest deductions and, ultimately, increase the aftertax earnings and cash flow from the arrangement. They did this by inventing a loan interest rate formula which exploited a loophole in the National Association of Insurance Commissioners ("NAIC") model bill governing policy loan interest rates.[11] The section of the model bill relating to variable rates provides that the rate shall not exceed the higher of (1) *790 Moody's Corporate Average or (2) the rate used to compute the cash surrender value plus one percent. The designers of the COLI policy specified Moody's Baa as the base rate for computing the cash surrender value of the policy. Moody's Baa always exceeds Moody's Corporate rate, therefore, in the case of the MBL COLI policy the rate referred to in the second prong of the model bill will always be higher than the rate referred to in the first prong, effectively freeing the MBL COLI policy from the limitations of the model bill. See infra, pp. 56, n 14.

The NAIC model bill represented an industry-wide consensus on the appropriate maximum variable loan interest rate for life insurance policy loans and it became law in most of the states, including Ohio. See Ohio Rev.Code § 3915.051. The drafters of the model bill intended the Moody's Corporate Average to be the maximum legal rate. That average is based on the average rate of a composite of investment grade corporate bonds.

The designers of the MBL COLI policy circumvented the intended maximum policy loan interest rate of the model bill and established a rate based on Moody's Baa, the most risky component of the Moody's Corporate Average, a rate that has historically been 60 to 100 basis points[12] higher than Moody's Corporate Average. Then, they increased that rate by one percent. This resulted in a policy loan interest rate 160 to 200 basis points higher than the maximum intended by the model bill. Later, with the advent of endorsement EFA 85-3, they permitted the policyholder to chose even higher rates, resulting in "junk bond" rates for risk free policy loans. Camelot, the taxpayer in C.M. Holdings, chose the highest of the three options offered by MBL. AEP chose the second highest, constrained by the fear that the I.R.S. might disallow its interest deductions if it chose the highest rate.

The insurance department of MBL's state of domicile, New Jersey, rejected endorsement EFA 85-3,[13] stating:

This change to the current credited loaned interest rate has a direct impact on the policy loan rate. We feel that it is inappropriate for a policyholder to determine within a range what these two rates will be. The premise of the variable loan interest rate is to charge a rate dependent upon an outside index, and the methodology used to determine the credited interest rate should rely on company expectation, not policyholder discretion.

D84. When a transaction is structured so that the borrower actually benefits from a higher loan interest rate and the borrower is permitted to chose its own interest rate from a range of rates that begins with a rate that far exceeds the industry maximum, the interest rate component of the transaction lacks economic substance.[14]

*791 AEP argues that its motivation for choosing the Moody's Baa enhanced rate was to increase tax-deferred inside buildup. But the projections provided to AEP showed that all of the additional inside buildup would be encumbered by policy loans and accrued loan interest. Thus, AEP could not expect to derive any nontax economic benefit from the higher inside buildup. Instead of increasing inside buildup the higher interest rate was intended to produce larger loans and larger interest deductions. When Campisi presented the interest rate options to AEP, his presentation focused on the policy loan interest rate and the tax deductions it would generate, not the tax-deferred inside buildup. See J968. AEP's evaluation of the economic benefits of the COLI plan likewise focused on the tax benefits of the policy loan interest deductions.

What AEP received for its cash payments to MBL/Hartford

The amount of cash paid by AEP to MBL/Hartford in any given year was precisely equal to the sum of the COI charges; policy expenses (including brokers' commissions, fees to McCamish and his companies and state premium taxes); and the one-percent spread between the policy loan interest rate and the loaned crediting rate. Because the COLI VIII plan was designed to be mortality neutral and called for policy loan borrowings which produced zero net equity throughout the life of the plan, the plan functioned much like term insurance. Viewed in this light, AEP substantially overpaid for the insurance coverage provided by the COLI program. The evidence revealed that the net cost of AEP's death benefit coverage for a 42-year-old male employee during its fifth policy year was $502.07. Yet, AEP's fifth-year cash payment for this policy was $2,120.10, which is 257% more than the actual cost of the death benefit. The difference is attributable to broker's commissions, McCamish's fees, and MBL/Hartford's one percent profit on the policy loans. What AEP received for this 257% markup on the cost of the death benefits was a fee for the facade of policy loans that were projected to produce nearly one billion dollars in tax deductions during the first twenty years of the plan which were essential to achieving the projected positive cash flow of over a half a billion dollars.

AEP's business purpose for participating in the MBL COLI VIII plan

AEP has established by substantial, credible evidence that its purpose for entering into the MBL COLI VIII plan was to provide earnings and cash flow that would partially offset its increased employee benefits expense attributable to the implementation of FAS 106. AEP intended to use the COLI earnings and cash flows to offset reductions in earnings caused by FAS 106 and to fund the VEBA. This was a proper business purpose. However, this does not satisfy the subjective prong of the economic substance analysis, which requires the court to assess whether the transaction had a purpose other than creating tax deductions, not how the taxpayer intended to use the tax-driven proceeds. The proper test for determining whether a transaction is an economic sham "consists of an examination of the transaction, not the taxpayer." Thomas v. United States, 166 F.3d 825, 834 (6th Cir.1999) (quoting Illes v. Comm'r, 982 F.2d 163, 166 (6th Cir.1992)). If the transaction lacks economic substance, then the deduction must be disallowed without regard to the niceties of the taxpayer's intent. See Mahoney v. Comm'r, 808 F.2d 1219, 1220 (6th Cir.1987).

AEP's intended use of the cash flows generated by the MBL COLI VIII plan is irrelevant to the subjective prong of the economic substance analysis. If a legitimate business purpose for the use of the *792 tax savings "were sufficient to breathe substance into a transaction whose only purpose is to reduce taxes, [then] every sham tax-shelter device might succeed." Winn-Dixie, 113 T.C. at 287. AEP was free to use the earnings and cash flows generated by the MBL COLI VIII program for any purpose it wished. In a memo prepared early in the course of AEP's consideration of COLI, Maloney aptly described COLI as "an independent funding vehicle which could be applied to anything." D641. AEP's intent to use the tax-driven COLI earnings and cash flows to help defray increasing postretirement medical benefit expenses does not imbue the COLI program with economic substance.

Similarity between AEP's COLI plan and the transaction in Knetsch v. United States

When the AEP MBL COLI VIII plan is stripped down to its bare essentials, it bears a remarkable resemblance to the transaction before the court in the landmark case of Knetsch. Carl F. Knetsch, in 1953, at the age of 60, purchased ten thirty-year maturity-deferred annuity savings bonds from the Sam Houston Life Insurance Company. Each bond was in the face amount of $400,000 bearing 2.5% interest compounded annually. The total purchase price was $4,004,000, which Knetsch paid by issuing a check for $4,000 and signing non-recourse notes for $4 million, bearing 3.5% interest. The notes were secured by the annuity bonds and interest on the loans was payable in advance. On the same day, Knetsch prepaid the first year's interest on the loans, which was $140,000. Under the table of cash and loan values made part of the bonds, their cash or loan value at the end of the first contract year was to be $4,100,000. The contract permitted Knetsch to borrow any excess of this value beyond his indebtedness without waiting until the end of the year. Knetsch promptly borrowed back $99,000, giving his check in the amount of $3,465 for a year's advance interest on that loan. In the two succeeding years, interest was paid and the loan values were borrowed. If the loans had been repaid, Knetsch would have been entitled to a monthly annuity of $90,171 per month at maturity when he would have been 90 years old. If, however, he continued to borrow and to pay interest according to the pattern established in the first three years, the value of the contracts would never exceed $1,000 and would have produced an annuity of just $43 per month. Knetsch terminated the transaction in December, 1956, at which time his indebtedness totaled $4,307,000. The cash or loan value of the bonds was $4,308,000, which had been the basis of the loan of the prior year. He surrendered the bonds and his indebtedness was canceled. He received the difference of $1,000 in cash.

Knetsch, in his income tax returns for 1953 and 1954, claimed an interest deduction for the interest paid on the loans. The deduction was disallowed by the Commissioner and paid by Knetsch, who sued for a refund. The Supreme Court held that the indebtedness created by the borrowings was fictional and in reality a rebate of a substantial part of the so-called interest payments. The Court, likewise, held that the annuity contracts were fictional because their value was reduced to and by the annual borrowings, kept at "the relative pittance of $1,000." Knetsch, 364 U.S. at 366, 81 S. Ct. 132. The transaction did not, according to the Court, appreciably affect Knetsch's beneficial interest except to reduce his tax. Nothing of substance was to be realized by him beyond a tax deduction. The transaction was determined by the Court to be a sham. See id.

In Knetsch, the petitioners contended, as AEP does here, that Congress in enacting various provisions of the Internal Revenue *793 Code authorized the deductions. They pointed out that I.R.C. § 264(a) (2), at that time, denied a deduction for amounts paid on indebtedness incurred to purchase a single premium annuity contract but only as to contracts purchased after March 1, 1954. They attributed to Congress a purpose to allow the deduction of pre-1954 payments and the transactions of the kind carried on by Knetsch. The Court refused to attribute such an intent to Congress, stating: "`[t]o hold otherwise would be to exalt artifice above reality and to deprive the statutory provision in question of all serious purpose.'" Knetsch, 364 U.S. at 367, 81 S. Ct. 132 (quoting Gregory v. Helvering, 293 U.S. 465, 470, 55 S. Ct. 266, 79 L. Ed. 596 (1935)). In reaching this conclusion, the Court reviewed the legislative history of Congressional action with respect to the taxation of annuity contracts, which bears many similarities to the history of Congressional action in the area of taxation of life insurance policies. See id. at 367-370, 81 S. Ct. 132.

The similarities between Knetsch's annuity transactions and the AEP COLI VIII plan are striking. They include first-day, first-year loans, which paid for all but a small percentage of the total premium and generated substantial interest deductions. There was a pattern of annual borrowings, which consumed nearly all of the equity in the annuity bonds and produced even more tax-deductible interest expense. The potential economic benefit of the annuity bonds, substantial annuity payments thirty years hence, was wiped out by the borrowings. The only real benefit to Knetsch was the tax deductions.[15] Although three Justices dissented in Knetsch, it remains the law of the land and this court can find no principled way to distinguish the essential features of the AEP MBL COLI VIII plan from the essential features of the annuity transactions in Knetsch.

Sixth Circuit precedent relied upon by AEP

AEP argues that controlling Sixth Circuit precedent upholds the deductibility of the policy loan interest on its COLI policy loans, citing Woodson-Tenent Labs., Inc. v. U.S., 454 F.2d 637 (6th Cir.1972) and Campbell v. Cen-Tex, Inc., 377 F.2d 688 (5th Cir.1967). The court disagrees. In Woodson-Tenent, the Sixth Circuit expressly adopted the rationale of the Fifth Circuit in Campbell, and of the district court in Priester Mach. Co. v. United States, 296 F. Supp. 604 (W.D.Tenn.1969), finding Knetsch was not controlling for the reasons stated in those opinions. See Woodson-Tenent, 454 F.2d at 638-39.

The policies involved in Woodson-Tenent were key man life insurance purchased on the lives of two principal officers of the corporation. A substantial portion of the premiums were paid with policy loans. However, the court expressly noted that the death benefits of the policy were substantially in excess of the maximum indebtedness incurred to pay policy premiums. See id. at 640. Thus, unlike the instant case, the policyholder in Woodson-Tenent could potentially receive a substantial non-tax benefit from the policies. Campbell v. Cen-Tex, likewise, involved life policies on key employees and stockholders, which were purchased to assist the corporation in meeting its obligations under its deferred compensation plan and its obligation to purchase and redeem the stock of any deceased stockholder. There was a pattern of annual borrowings to pay the premiums on the policies. However, unlike the AEP COLI VIII plan, the policies *794 in Cen-Tex had substantial net equity after twenty years. The court noted:

[T]he policies would have, at the end of twenty years, with maximum loans effected, cash surrender values in excess of $200,000. These potential death benefits and cash surrender values cannot be brought within the Knetsch characterization of a "relative pittance."

Campbell, 377 F.2d at 693. Likewise, in Priester Mach. Co., the policies were key man life policies and, although there was a pattern of annual borrowing to pay premiums, the policies had substantial increasing net equity at the end of each year. See Priester Mach. Co., 296 F. Supp. at 606. The court distinguished Knetsch noting that:

Here[,] on the other hand, the death benefits of the insurance policies were and would always be substantially greater than the loan and, further, taxpayers were acquiring substantial and increasing net cash values in the policies.

Id. at 608. Thus, Woodson-Tenent, Cen-Tex and Priester Mach. Co. are distinguishable from the instant case because the policies in those cases, unlike the AEP COLI VIII plan, provided substantial nontax benefits to the policyholders.

Other cases relied on by AEP, including Shirar v. Comm'r, 916 F.2d 1414 (9th Cir. 1990); Coors v. United States, 215 Ct.Cl. 840, 572 F.2d 826 (1978); and Golsen v. U.S., No. 112-78, 1980 WL 4732 (Ct.Cl. Trial Div. Oct. 16, 1980), are likewise distinguishable. In Shirar, the policies would have built up significant net equity even with maximum policy loans. See Shirar, 916 F.2d at 1418-19. In Coors, the policyholder paid cash premiums for the first several years its policies were in force before taking out policy loans to pay premiums after falling into economic difficulties. See Coors, 572 F.2d at 829.

The four out of seven safe harbor

The Internal Revenue Code generally disallows deductions for interest paid on policy loans taken against a life insurance policy, "which contemplates the systematic direct or indirect borrowing of part or all of the increases in the cash value" of the policy. I.R.C. § 264(a) (3). Section 264(c) (1) provides an exception and allows deductions for interest paid on policy loans taken against a life insurance policy so 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."

Implicit in the four out of seven safe harbor rule is the requirement that the annual premiums must remain level for the first seven years of the policy. As noted earlier, AEP concedes that this is the case. In policy years four through seven, AEP purportedly paid the annual premiums on its COLI policy through a combination of partial withdrawals and loading dividends. However, the court has found that the loading dividends were factual shams. They were included in the premiums but immediately returned to AEP in a simultaneous netting transaction, which was a mere facade to make it appear as though the premiums in years four through seven were equal to the premiums charged in years one through three. In fact, the real premiums for years four through seven were a small fraction of the premiums charged in years one through three; thus, the MBL COLI VIII plan violated the seven year level premium requirement of § 264(c) (1). AEP's COLI VIII policies contemplated "the systematic direct or indirect borrowing of part or all of the increases in the cash value" of the policies. Accordingly, I.R.C. § 264(a) (3) disallows AEP's deductions for interest paid on the policy loans and AEP is unable *795 to avail itself of the exception provided by the § 264(c) (1) four out of seven safe harbor rule.


The court finds that the following components of the AEP COLI VIII plan were factual shams:

The first-year policy loans for the period February 16 to March 23, 1990; the first-year loading dividend and corresponding portions of the premium charges; and, the loading dividends and corresponding portions of the premium charges in policy years four through seven.

The court finds that the AEP COLI VIII plan as a whole was a sham in substance.

The court further finds that AEP's deductions for interest paid on policy loans with respect to the COLI VIII plan must be disallowed under the provisions of I.R.C. § 264(a) (3).

The Clerk shall enter final judgment in favor of the defendant at plaintiff's costs.

It is so ORDERED.


[1] Now codified as I.R.C. § 264(d) (1) but referred to herein as I.R.C. § 264(c) (1).

[2] Whenever the aftertax cost of a policy loan is less than the amount being credited to inside buildup, it becomes profitable to purchase life insurance and pay for it with policy loans.

[3] The New York Life COLI plan was similar in many respects to the Mutual Benefit Life COLI plan which AEP ultimately purchased.

[4] AEP's 1990 Annual Report indicates that its rate for first mortgage financial bonds was nine and one-eighth percent in September 1989. See J928.

[5] The term "grantor trust" is used by tax lawyers to describe a trust of the type described in sections 671-679 of the Code. The grantor of this form of trust retains sufficient powers over the trust, usually the power to revoke at will, that the trust is considered, for tax purposes, as the alter ego of the grantor. Although the trust is required to file a Federal income tax return, that return is essentially an information return that reports all income and expense as attributable to the grantor and includible by the grantor on its Federal income tax return. See Black's Law Dictionary 1516 (7th ed.1999).

[6] Moody's rates investment grade corporate bonds as Aaa, Aa, A, and Baa, with Baa being the riskiest and having the highest interest rate. The Moody's Corporate Average is the composite average of the interest rates on all of these investment grade bonds combined.

[7] Moody's Corporate Bond yield average monthly corporate Baa, is the composite average of the interest rates on Baa rated bonds.

[8] These dividends were "declared" by the board of MBL in November, 1989, three months before AEP purchased the COLI policies.

[9] MBL recognized that the zero net equity feature was a significant indicator of the COLI VIII plan's lack of economic substance and removed the net equity column from its sales illustration because it produced an "uninterrupted string of zeros." See C.M. Holdings, 254 B.R. at 632.

[10] See, e.g., Goldstein v. Comm'r, 364 F.2d 734 (2nd Cir.1966) cert. denied, 385 U.S. 1005, 87 S. Ct. 708, 17 L. Ed. 2d 543 (1967); Winn-Dixie Stores, Inc. v. Comm'r, 113 T.C. 254, 1999 WL 907566 (1999).

[11] The pertinent provisions of the model bill are as follows:

Section 2. Definitions

For purposes of this Act the "Published Monthly Average" means:

A. The Moody's Corporate Bond Yield Average Monthly Average Corporates as published by Moody's Investors Service, Inc. or any successor thereto; ...

Section 3. Maximum Rate of Interest of Policy Loans

A. Policies issued on or after the effective date of this Act shall provide for policy loan interest rates as follows:

(1) A provision permitting a maximum interest rate of not more than eight percent (8%) per annum; or

(2) A provision permitting an adjustable maximum interest rate established from time to time by the life insurer as permitted by law.

B. The rate of interest charged on a policy loan made under Subsection A(2) shall not exceed the higher of the following:

(1) The Published Monthly Average for the calendar month ending two months before the date on which the rate is determined; or

(2) The rate used to compute the cash surrender values under the policy during the applicable period plus one percent per annum.

Pl.Exh. P199.

[12] A "Basis point" is 1/100 of one percent.

[13] The endorsement was likewise rejected by Pennsylvania, New York, Texas and Oregon. Other states approved endorsement EFA 85-3 but it is not clear what level of scrutiny, if any, the endorsement received in these states.

[14] New York refused to approve the policy loan interest rate provision of the COLI policy because it exceeded the maximum intended by the NAIC model bill. An M & R actuary in New York reported to Eisenberg in September, 1986 that the New York regulators had "pointed out, correctly (and you and I have discussed this) that in theory if this approach is taken it is not necessary to use either a Baa index or for that matter any index at all; one could, for example, credit 24% and charge 25%." and "This got us into a discussion about what the IRS would or would not regard as a reasonable provision (as opposed to a sham) for tax purposes[.]" G73.

[15] The annuity bonds also provided a death benefit equal to the net cash value of the bonds at the time of death. See United States v. Bond, 258 F.2d 577, 579 (5th Cir.1958).